Power Distribution Engineering : Fundamentals and Applications Electrical

title:
Engineering and Electronics ; 88
author: Burke, James J.
publisher: CRC Press
isbn10 | asin: 0824792378
print isbn13: 9780824792374
ebook isbn13: 9780585157245
language: English
subject Electric power distribution.
publication date: 1994
lcc: TK3001.B87 1994eb
ddc: 621.319
subject: Electric power distribution.
 Page aa

Power Distribution Engineering

 Page ab
ELECTRICAL ENGINEERING AND ELECTRONICS
A Series of Reference Books and Textbooks
Executive Editors
Marlin O. Thurston
Department of Electrical Engineering
The Ohio State University
Columbus, Ohio
William Middendorf
Department of Electrical
and Computer Engineering
University of Cincinnati
Cincinnati, Ohio
Editorial Board
Maurice Bellanger
Télécommunications, Radioélectriques,
et Téléphoniques (TRT)
Le Plessis-Robinson, France
J. Lewis Blackburn
Bothell, Washington
Sing-Tze Bow
Department of Electrical Engineering
Northern Illinois University
De Kalb, Illinois
Norman B. Fuqua
Reliability Analysis Center
Griffiss Air Force Base, New York
Charles A. Harper
Westinghouse Electrical Engineering
and Technology Seminars, Inc.
Timonium, Maryland
Naim A. Kheir
Department of Electrical and
Systems Engineering
Oakland University
Rochester, Michigan
Lionel M. Levinson
General Electric Company
Schenectady, New York
V. Rajagopalan
Department of Engineering
Université du Québec
à Trois-Rivières
Trois-Rivières, Quebec, Canada
Earl E. Swartzlander
TRW Defense Systems Group
Redondo Beach, California
Spyros G. Tzafestas
Department of Electrical
and Computer Engineering
National Technical University
of Athens
Athens, Greece
Sakae Yamamura
Central Research Institute of
the Electric Power Industry
Tokyo, Japan
 Page ac
1. Rational Fault Analysis, edited by Richard Saeks and S. R. Liberty
2. Nonparametric Methods in Communications, edited by P. Papantoni-Kazakos and Dimitri Kazakos
3. Interactive Pattern Recognition, Yi-tzuu Chien
4. Solid-State Electronics, Lawrence E. Murr
5. Electronic, Magnetic, and Thermal Properties of Solid Materials, Klaus Schröder
6. Magnetic-Bubble Memory Technology, Hsu Chang
7. Transformer and Inductor Design Handbook, Colonel Wm. T. McLyman
8. Electromagnetics: Classical and Modern Theory and Applications, Samuel Seely and Alexander D. Poularikas
9. One-Dimensional Digital Signal Processing, Chi-Tsong Chen
10. Interconnected Dynamical Systems, Raymond A. DeCarlo and Richard Saeks
11. Modern Digital Control Systems, Raymond G. Jacquot
12. Hybrid Circuit Design and Manufacture, Roydn D. Jones
13. Magnetic Core Selection for Transformers and Inductors: A User's Guide to Practice and Specification, Colonel Wm. T. McLyman
14. Static and Rotating Electromagnetic Devices, Richard H. Engelmann
15. Energy-Efficient Electric Motors: Selection and Application, John C. Andreas
16. Electromagnetic Compossibility, Heinz M. Schlicke
17. Electronics: Models, Analysis, and Systems, James G. Gottling
18. Digital Filter Design Handbook, Fred J. Taylor
19. Multivariable Control: An Introduction, P. K. Sinha
20. Flexible Circuits: Design and Applications, Steve Gurley, with contributions by Carl A. Edstrom, Jr., Ray D. Greenway, and
William P. Kelly
21. Circuit Interruption: Theory and Techniques, Thomas E. Browne, Jr.
22. Switch Mode Power Conversion: Basic Theory and Design, K. Kit Sum
23. Pattern Recognition: Applications to Large Data-Set Problems, Sing-Tze Bow
24. Custom-Specific Integrated Circuits: Design and Fabrication, Stanley L. Hurst
25. Digital Circuits: Logic and Design, Ronald C. Emery
26. Large-Scale Control Systems: Theories and Techniques, Magdi S. Mahmoud, Mohamed F. Hassan, and Mohamed G. Darwish
27. Microprocessor Software Project Management, Eli T. Fathi and Cedric V. W. Armstrong (Sponsored by Ontario Centre for
Microelectronics)
28. Low Frequency Electromagnetic Design, Michael P. Perry
29. Multidimensional Systems: Techniques and Applications, edited by Spyros G. Tzafestas
30. AC Motors for High-Performance Applications: Analysis and Control, Sakae Yamamura
 Page ad
31. Ceramic Motors for Electronics: Processing, Properties, and Applications, edited by Relva C. Buchanan
32. Microcomputer Bus Structures and Bus Interface Design, Arthur L. Dexter
33. End User's Guide to Innovative Flexible Circuit Packaging, Jay J. Miniet
34. Reliability Engineering for Electronic Design, Norman B. Fuqua
35. Design Fundamentals for Low-Voltage Distribution and Control, Frank W. Kussy and Jack L. Warren
36. Encapsulation of Electronic Devices and Components, Edward R. Salmon
37. Protective Relaying: Principles and Applications, J. Lewis Blackburn
38. Testing Active and Passive Electronic Components, Richard F. Powell
39. Adaptive Control Systems: Techniques and Applications, V. V. Chalam
40. Computer-Aided Analysis of Power Electronic Systems, Venkatachari Rajagopalan
41. Integrated Circuit Quality and Reliability, Eugene R. Hnatek
42. Systolic Signal Processing Systems, edited by Earl E. Swartzlander, Jr.
43. Adaptive Digital Filters and Signal Analysis, Maurice G. Bellanger
44. Electronic Ceramics: Properties, Configuration, and Applications, edited by Lionel M. Levinson
45. Computer Systems Engineering Management, Robert S. Alford
46. Systems Modeling and Computer Simulation, edited by Naim A. Kheir
47. Rigid-Flex Printed Wiring Design for Production Readiness, Walter S. Rigling
48. Analog Methods for Computer-Aided Circuit Analysis and Diagnosis, edited by Takao Ozawa
49. Transformer and Inductor Design Handbook: Second Edition, Revised and Expanded, Colonel Wm. T. McLyman
50. Power System Grounding and Transients: An Introduction, A. P. Sakis Meliopoulos
51. Signal Processing Handbook, edited by C. H. Chen
52. Electronic Product Design for Automated Manufacturing, H. Richard Stillwell
53. Dynamic Models and Discrete Event Simulation, William Delaney and Erminia Vaccari
54. FET Technology and Application: An Introduction, Edwin S. Oxner
55. Digital Speech Processing, Synthesis, and Recognition, Sadaoki Furui
56. VLSI RISC Architecture and Organization, Stephen B. Furber
57. Surface Mount and Related Technologies, Gerald Ginsberg
58. Uninterruptible Power Supplies: Power Conditioners for Critical Equipment, David C. Griffith
59. Polyphase Induction Motors: Analysis, Design, and Application, Paul L. Cochran
 Page ae
60. Battery Technology Handbook, edited by H. A. Kiehne
61. Network Modeling, Simulation, and Analysis, edited by Ricardo F. Garzia and Mario R. Garzia
62. Linear Circuits, Systems, and Signal Processing: Advanced Theory and Applications, edited by Nobuo Nagai
63. High-Voltage Engineering: Theory and Practice, edited by M. Khalifa
64. Large-Scale Systems Control and Decision Making, edited by Hiroyuki Tamura and Tsuneo Yoshikawa
65. Industrial Power Distribution and Illuminating Systems, Kao Chen
66. Distributed Computer Control for Industrial Automation, Dobrivoje Popovic and Vijay P. Bhatkar
67. Computer-Aided Analysis of Active Circuits, Adrian Ioinovici
68. Designing with Analog Switches, Steve Moore
69. Contamination Effects on Electronic Products, Carl J. Tautscher
70. Computer-Operated Systems Control, Magdi S. Mahmoud
71. Integrated Microwave Circuits, edited by Yoshihiro Konishi
72. Ceramic Materials for Electronics: Processing, Properties, and Applications, Second Edition, Revised and Expanded, edited by
Relva C. Buchanan
73. Electromagnetic Compatibility: Principles and Applications, David A. Weston
74. Intelligent Robotic Systems, edited by Spyros G. Tzafestas
75. Switching Phenomena in High-Voltage Circuit Breakers, edited by Kunio Nakanishi
76. Advances in Speech Signal Processing, edited by Sadaoki Furui and M. Mohan Sondhi
77. Pattern Recognition and Image Preprocessing, Sing-Tze Bow
78. Energy-Efficient Electric Motors: Selection and Application, Second Edition, John C. Andreas
79. Stochastic Large-Scale Engineering Systems, edited by Spyros G. Tzafestas and Keigo Watanabe
80. Two-Dimensional Digital Filters, Wu-Sheng Lu and Andreas Antoniou
81. Computer-Aided Analysis and Design of Switch-Mode Power Supplies, Yim-Shu Lee
82. Placement and Routing of Electronic Modules, edited by Michael Pecht
83. Applied Control: Current Trends and Modern Methodologies, edited by Spyros G. Tzafestas
84. Algorithms for Computer-Aided Design of Multivariable Control Systems, Stanoje Bingulac and Hugh F. VanLandingham
85. Symmetrical Components for Power Systems Engineering, J. Lewis Blackburn
86. Advanced Digital Signal Processing: Theory and Applications, Glenn Zelniker and Fred J. Taylor
87. Neural Networks and Simulation Methods, Jian-Kang Wu
88. Power Distribution Engineering: Fundamentals and Applications, James J. Burke
 Page af
Modern Digital Control Systems, Second Edition, Raymond G. Jacquot
Integrated Circuit Quality and Reliability, Second Edition, Revised and Expanded, Eugene R. Hnatek
Handbook of Electric Motors, edited by Richard Engelmann and William H. Middendorf
Adaptive IIR Filtering in Signal Processing and Control, Philip Regalia
 Page i

Power Distribution Engineering

Fundamentals and Applications
James J. Burke
Power Technologies, Inc.
Schenectady, New York

MARCEL DEKKER, INC.

NEW YORK BASEL
 Page ii
Library of Congress Cataloging-in-Publication Data
Burke, James J.
Power distribution engineering: fundamentals and applications/James J. Burke.
p. cm.(Electrical engineering and electronics)
Includes bibliographical references and index.
ISBN 0-8247-9237-8 (alk. paper)
1. Electrical power distribution. I. Title. II. Series.
TK3001.B87 1994 94-11103
621.319dc20 CIP
The publisher offers discounts on this book when ordered in bulk quantities. For more information, write to Special Sales/Professional
Marketing at the address below.
This book is printed on acid-free paper.
Copyright © 1994 by MARCEL DEKKER, INC. All Rights Reserved.
Neither this book nor any part may be reproduced or transmitted in any form or by any means, electronic or mechanical, including
photocopying, microfilming, and recording, or by any information storage and retrieval system, without permission in writing from the
publisher.
MARCEL DEKKER, INC.
270 Madison Avenue, New York, New York 10016
Current printing (last digit):
10 9 8 7 6 5 4
PRINTED IN THE UNITED STATES OF AMERICA
 Page iii

To the memory of Aunt Marg & Uncle Rick

 Page v

Preface
The purpose of this book is to give the utility distribution engineer a document which is truly useful and up-to-date for today's
competitive environment. The book is based on over 25 years of experience in both performing distribution studies and in teaching
courses in distribution engineering. Too often the author has found that books on distribution engineering are either simply very good
reference data books or good academic textbooks. Neither document usually explains why utilities do things the way they do. It is
hoped that this book will bridge that gap.
Simplicity of concepts is emphasized. Complex mathematical concepts are not used since teaching experience has shown that
understanding concepts is far more valuable than being able to use a wide variety of mathematical techniques. The book's math is
straightforward and is used only to illustrate the point that many seemingly difficult concepts and calculations can be greatly simplified
using obvious assumptions.
This book is intended to cover a wide and comprehensive list of topics, from system protection to economic evaluations. Its chapters
reflect current and future thinking in these areas and are based on the author's work in IEEE Standards groups as well as many real
utility studies.
The material varies in depth. For example, Chapter 1 is meant to be a brief overview of a distribution system. This was done because
there are several good documents presently on the market that address the topic in greater detail. On the other hand, Chapters 4 and 5,
on overcurrent and overvoltage protection, are quite comprehensive. This was done for two reasons. First, philosophies in both these
areas are in a state of flux. Second, very little material on either of these areas is currently available in a comprehensive form.
Chapter 7, on power quality, appears near the end of the book since the concept of power quality can mean just about anything and as
such will be more meaningful if the earlier chapters are read beforehand. Typical topics, such as sags, swells, and harmonics, are
covered as are other areas such as electromagnetic fields and stray voltage, considered by only some distribution engineers to be
associated with power quality.
Each chapter has review questions meant to stimulate thought for
 Page vi
those readers who utilize the book as a self-teaching guide. Although this book can serve as a good day-to-day reference, gaining the
maximum value from its contents will be helped by attempting these exercises. The questions are meant to be simple and fun.
The author would like to acknowledge help over the years from Harold Campbell as well as the late Jack Easley, Norm Schultz, and
Bill Moody. Finally, the author wishes to thank Sherie Roseboom, who has worked so hard and so long typing and assembling the
many drafts and manuscripts.
JAMES J. BURLE
 Page vii

Contents
Preface v
1 1
Utility Distribution System Design and Characteristics
Introduction 1
Design 1
Overhead vs. Underground 12
Grounding 16
Load and Fault Characteristics 20
Questions 27
2 29
Transformers and Regulators
Transformers 29
Distribution Transformer Connections 47
Regulators 54
Questions 63
3 64
Application of Capacitors for Distribution Systems
Fundamentals 64
Voltage Rise and Voltage Drop 76
Losses 88
Questions 99
4 101
Distribution Overcurrent Protection
Introduction 101
Characteristics of Devices 101
 Page viii
Coordination of Devices 129
Questions 165
5 168
Surge Arresters
Characteristics of Lightning 168
System Overvoltage 174
Silicon Carbide vs. MOV Arresters 182
Classes of Arresters 184
Arrester Selection 185
Insulation Coordination 194
Traveling Waves 208
Line Protection 220
Induced Strikes 231
Reference 232
Questions 232
6 234
Distribution Reliability
Introduction 234
Fundamentals 234
Power Quality 246
Factors Affecting Reliability 252
Reference 268
Questions 268
7 270
Power Quality Fundamentals
Introduction 270
Definitions 270
Voltage Quality 272
Sensitive Loads 292
Power Line Conditioning 296
Electromagnetic Fields 305
Stray Voltage 312
Reference 316
Questions 316
 Page ix
8 320
Distribution Economics
Introduction 320
Time Value of Money 320
Annual Carrying Charge 328
Definition of Terms 332
Economic Evaluations 335
Reference 347
Questions 347
Index 349
 Page 1

1
Utility Distribution System Design and Characteristics
Introduction
The distribution engineer sometimes finds it difficult to define a typical distribution system. It is the purpose of this chapter to suggest
typical values of voltage, line lengths, load and fault levels, as well as types of system design and grounding which can be used as
background information for the more technical discussions later on in the book.
Design
The Utility System
The electric utility system is usually divided into three segments which are generation, transmission, and distribution. A fourth
division, which can sometimes be made is subtransmission, which can really be considered a subset of transmission since the voltage
levels overlap and operational and protection practices are quite similar. Figure 1-1, shown below, illustrates some of the major
components in these divisions.
The distribution system, which is our main area of interest, is commonly broken down into the following three components:
1. Distribution substation
2. Distribution primary
3. Secondary
Even on this greatly simplified one-line diagram, it can be seen that the distribution system consists of a much wider variety of voltage
levels, components, loads and interconnections than does the generation or transmission system.
 Page 2

Figure 1-1.
Typical Electric Supply System
Distribution Substation
The distribution system is fed through distribution substations. These substations consist of an almost infinite number of designs based
on considerations such as
Load density
High side voltage
Low side voltage
Land availability
Reliability requirements
Load growth
Voltage drop
Emergency conditions
Cost and losses
A typical substation is shown in Figure 1-2. This substation indicates the average arrangement and equipment ratings per an industry
survey. For example, the voltage of the high side bus can be anywhere from 34.5 kV all the way up to 345 kV and beyond. The average
or preferred high side voltage level is approximately 115 to 138 kV because this voltage
 Page 3
level is usually high enough to maintain a stiff enough source and low enough to alleviate the costs associated with high side
equipments. As shown, the average substation consists of two transformers rated 21/28/35 MVA (OA/FA/FOA) with an impedance of
approximately 10 percent. Protection of the substation transformer is usually attained via high side breakers or circuit switchers and
low side breakers used in conjunction with differential relays (overcurrent relaying is also used as backup and is not shown). The
transformer low side breaker, sometimes referred to as the substation secondary breaker, is also used to protect against low voltage bus
faults as well as back up the feeder breakers.

Figure 1-2.
Average Distribution Substation Arrangement
The low voltage bus in a multiple transformer substation is usually split (contains a normally open breaker or switch) to alleviate
circulating currents as well as reduce the short circuit current seen by the system. Two or more feeders are normally connected to each
bus through a feeder breaker. On smaller substations where short circuit levels are lower, a recloser is sometimes used instead of a
breaker. Short circuit levels at the
 Page 4
terminals of the low voltage bus are generally kept at 12,000 amperes or less although there are many systems where much higher
levels can be found.
Distribution Feeders
Figure 1-3 shows a primary distribution feeder with various equipment such as fuses, distribution transformers, reclosers, and switches
connected to it. Much of this equipment, such as a recloser, is utilized only at the distribution level. On the other hand, some of the
equipment such as capacitors, transformers, and arresters is also used at the transmission levels but with considerably different rules of
application. As shown, most distribution feeders are 3-phase and 4-wire. The fourth wire is the neutral wire which is connected to the
pole, usually below the phase wires, and grounded periodically.
A three-phase feeder main can be fairly short, on the order of a mile or two, or it can be as long as 30 miles. Voltage levels can be as
high as 34.5 kV, with the most common voltages being in the 15 kV class. While most of the 3-phase mains are overhead, much of the
new construction, particularly the single-phase lateral construction, is being put underground. Underground systems have the
advantage of immunity from certain types of temporary fault conditions like wind, direct lightning strikes, animals, etc. Permanent
faults, on the other hand, are much more difficult to locate and repair and have been the subject of much concern in recent years.
There are various feeder designs which can increase customer service. Five of these are shown below in Figure 1-4. These designs are
described as follows:
Radial System. It is obvious that the radial system is exposed to many interruption possibilities, the most important of which are those
due to primary overhead or underground cable failure or transformer failure. Either event may be accompanied by a long interruption,
given nominally by some utilities as 10 to 12 hours. Both components have finite failure rates and such interruptions are expected and
statistically predictable. The system will be satisfactory only if the interruptions frequency is very low and if there are ways to operate
the system without planned outages. Feeder breakers reclosing or temporary faults may affect sensitive loads.
 Page 5

Figure 1-3.
Primary Distribution Feeder
Primary Loop. A great improvement is obtained by arranging a primary loop, which provides two-way feed at each transformer. In this
manner, any section of the primary can be isolated, without interruption, and primary faults are reduced in duration to the time required
to locate a fault and do the necessary switching to restore service. This procedure can be performed either manually or automatically.
The cable in each half of the loop must have capacity enough to carry all the load. The additional cable exposure will tend to increase
the frequency of faults, but not necessarily the faults per customer. The addition of a loop tie switch at the open point also introduces
the possibility of a single equipment fault causing an interruption to both halves of the loop. Murphy's Law
 Page 6

Figure 1-4.
Five Basic Service Systems
generally applies to these situations. Once again, sensitive loads would be affected by reclosing under temporary fault conditions.
Primary Selective. This system uses the same basic components as in the primary loop, but arranged in a dual or main/alternative
scheme. Each transformer can select its source, and automatic switching is frequently used. When automatic, the interruption duration
can be limited to two or three seconds. Each service now represents a potential two-feeder outage (if the open switch fails), but under
normal contingencies, service restoration is rapid and there is no need to locate the fault (as with the loop) prior to doing the switching.
This scheme is in popular use on many underground systems. It also offers little remedy for computer problems caused by temporary
faults to the overhead system.
Secondary Selective. This is the first of the service systems using two transformers and low voltage switching. It is not in popular use
by
 Page 7
utilities for 480 volt service, but is common in industrial plants and on institutional properties. Primary operational switching is
eliminated and with it some causes of difficulty. Duplicate transformers virtually eliminate the possibility of a long interruption due to
failure. Load is divided between the two units and automatic transfer is employed on loss of voltage to either load. There must be close
coordination of utility and customer during planned transfers, and the split responsibility is probably the principal reason for its limited
use as a service system. Temporary faults on the primary feeders should have little if any effect on even sensitive computer loads.
Secondary Spot Network. Maximum service reliability and operating flexibility are obtained by use of the spot network using two or
more transformer/protector units in parallel. The low voltage bus is continuously energized by all units, and automatic disconnection of
any unit is obtained by sensitive reverse power relays in the protector. Maintenance switching of primary feeders can be done without
customer interruption or involvement. Spot networks are common in downtown, high density areas and are being applied frequently in
outlying areas for large commercial services where the supply feeders can be made available. This system also represents the most
compact and reliable arrangement of components and is the most reliable for all classes of loads.
Each of the five systems described can be evaluated in terms of reliability for traditional loads as shown below. As can be seen (see
Table 1-1), a radial system, which is used in some residential areas, can expect an outage almost once a year lasting approximately 90
minutes. On the other hand, a spot network, which many utilities use in downtown areas, will see only two outages for every 100 years.
Network Systems
Secondary ac network systems (grid networks) began around the year 1915 replacing the older dc networks which had problems with
cost of converters, copper costs and voltage difficulties. Most cities (over 260) in the U.S. have an ac network system. These network
systems are well known for their high reliability (see Table 1-1), although their cost differential has favored the primary selective
design for new construction in recent years.
 Page 8
Table 1-1. Measured Reliability of Different Distribution Systems
Primary
Type of System Radial Auto- URD Primary Secondary Grid Spot
Loop Selective Selective Network Network
Outages/yr 0.31.3 0.40.7 0.40.7 0.10.5 0.10.5 0.0050.0200.020.10
Average outage
90 65 60 180 180 135 180
duration, min
Momentary
510 1015 48 48 24 0 01
interruptions/yr

The major segments of a network system, as shown in Figure 1-5, are:

Primary feeder circuits
Network units (consisting of the network transformer and the network protector)
Secondary grid
The secondary grid is either 208Y/120 or 480Y/277 with virtually all spot networks favoring the higher voltage (see Figure 1-6).
Commonly used wire sizes range from 4/0 to 500 MCM AWG. The main protection of the secondary grid comes from the ability of the
system to burn off the fault. This, of course, meant that no protective device was required to operate. While this practice proved to be
highly successful on the 208Y/120 volt secondary network, there have been many instances of the 480Y/277 volt system not being able
to successfully burn itself clear, sometimes resulting in fires and considerable damage. A partial solution to this has been the use of
limiters, which are devices (really just restricted copper sections which act like a fuse) installed in the secondary main at each junction
point. It should be noted that the limiter is usually like an expulsion fuse and does not limit the magnitude of current like a true current
limiting fuse. In high short circuit areas, true current limiting fuses are now being used to limit damage due to these secondary faults.
The network protector is an electrically operated low-voltage air circuit breaker with self-contained relays for controlling its operation.
Its main purpose is to isolate the secondary from problems on the source side
 Page 9

Figure 1-5.
Single-Phase, Three-Wire Secondary System 240/120 Volt Service
 Page 10

Figure 1-6.
(network transformer, primary feeder, etc.). It is not its function to isolate faults on the secondary network itself. The network protector
performs the following three basic functions:
1. Provides automatic isolation of faults in the primary feeders or network transformers.
2. Provides ability to trip protector on reverse power.
3. Provides ability to close automatically when feeder is energized, and in-phase or leading.
The network protector is equipped with a set of fuses, one in each phase, located between the circuit breaker and the terminals for
connection to the system secondary voltage grid. The primary function of the fuse is to provide backup to the protector, not to clear
secondary faults. These secondary faults are either supposed to burn clear or be cleared by the limiters.
Secondaries
The purpose of the distribution transformer is to reduce the primary voltage to a level where it can be utilized by the customer. Three-
phase commercial distribution transformers range in size anywhere from about 75 kVA to over 2000 kVA. Single-phase transformers
range in size from
 Page 11
about 10 kVA to about 300 kVA with units in the 25 and 37.5 kVA size being the most popular for residential areas.
The secondary voltage level in the United States for residential service is 120/240. What this means is that a typical residence has a
choice of either depending on the requirements of the load. Figure 1-7 illustrates the connection of a typical single-phase, three-wire
secondary system used in most homes. Typically, lower wattage devices (lights for example) are connected line-to-neutral across both
sides of the transformer secondary. Higher wattage devices such as ovens, clothes dryers, etc., are usually connected across the 240
volt circuit since this has the effect of reducing voltage drop and losses.

Figure 1-7.
Single-Phase, Three-Wire Secondary System 240/120 Volt Service
Figure 1-8 shows the connections for a three-phase, four-wire secondary system rated 208/120 volts. This type of connection is used
where moderate three-phase loading as well as single-phase load is prevalent. For light industrial and commercial loads with higher
power requirements, 480/277 is utilized. Secondary network systems use either 208/120 or 480/277 depending on load requirements.
 Page 12

Figure 1-8.
Three-Phase, Four-Wire Secondary System 208/120 Volt Service

Overhead vs. Underground

Reliability
Underground distribution has become commonplace. The major drive from overhead distribution to underground distribution was
primarily a response to environmental pressures. Most utilities today put most of their single-phase residential developments (5 homes
or more) underground. Many utilities are also very gradually converting existing three-phase overhead to underground.
Most engineers agree that underground is more reliable than overhead. The argument is many times made that underground systems
fail less often but take much longer to fix. Also, the older URD cable has been failing at considerably higher rates than expected. The
bottom line is that average customer minutes outage (CMO), is considerably lower for most URD designs. Undergrounding also
eliminates almost all temporary faults which, for some systems, is 80% of their total. In these days of power quality, where momentary
outages are a large concern, URD can have a very desirable effect. A comparison of the failure rates (permanent faults) between
overhead and underground is shown below.
 Page 13
Table 1-2
Failure Rate (Failure/Year/Mile)
Voltage LevelOverhead LineUnderground Cable
5 to 11 kV 0.177 0.048
11 to 20 kV 0.130 0.097
33 kV 0.070 0.037
66 kV 0.059 0.028

Equipment
Most equipment used in an overhead distribution system utilizes air as its primary insulating medium (approximate 186 kV per foot for
an impulse). For example, most overhead lines are uninsulated or in the case of tree wire, minimally insulated. Most overhead switches
such as the standard disconnects rely on air to insulate the open contacts or even break very low level currents such as low charging
currents. Air-break switches use air to break load currents but these switches are essentially disconnect switches equipped with arcing
horns.
Underground equipment is different because the ground is in such close proximity and air insulation is now usually insufficient. Phase
conductors must consequently be insulated (cable) and certain other equipments such as switches must utilize a more effective
insulating medium such as vacuum, oil and most recently, SF6. A new term associated with underground distribution is deadfront. The
term deadfront (e.g., deadfront switch) means that there are no exposed connections. For example, when the cabinet door of a
deadfront switch is opened, no live parts will be exposed. The loadbreak elbow when connected into the bushing, shown in Figure 1-9,
would be classified a deadfront system. On the other hand, most overhead equipment and some URD equipment is classified as
livefront because exposed parts like spade connectors are clearly visible.
 Page 14

Figure 1-9.
Integrated Bushing
Some terminology common to overhead and underground switches which tends to cause confusion is defined unofficially (i.e., in
layman's terms, not IEEE) as follows:
a. Loadbreak. Loadbreak means the device has the ability to break load; usually 200 amps maximum for single-phase and 600 amperes
maximum for 3Ø phase. Many overhead switches have no loadbreak capability while most underground switches are, in part, rated for
breaking load.
The term loadbreak in URD is normally associated with the loadbreak elbow shown in Figure 1-10. The loadbreak elbow is actually a
mini, single-phase switch that allows the system to be sectionalized (opened) under energized conditions where loads up to 200
amperes on the laterals can be interrupted. A few utilities still use what is called a non-loadbreak elbow. This elbow looks similar but
cannot be switched under load (energized) conditions.
 Page 15

Figure 1-10.
Molded-shield Elbows
b. Continuous Rating. The maximum ampere rating of the device under continuous operation. If the device is a switch, continuous does
not mean that the switch can interrupt this load; it just means it can pass this load, in the closed position, without damage.
c. Momentary Rating. The momentary rating of a device is the amount of short circuit current it can pass, in the closed position,
without damage (still be operable). It does not mean the device can interrupt the fault current. For example, a loadbreak elbow has a
momentary rating and cannot be used to interrupt anything higher than load current (generally 200 amperes).
d. Short Circuit Rating. The short circuit rating of a device is the maximum current it is designed to interrupt. Examples of devices with
this rating are fuses, breakers and reclosers.
e. Close and Latch. The close and latch rating of a switch is the maximum ampere rating (fault level) that the switch can close into
successfully. It is never normal practice to close into a fault. However, by mistake, it is entirely possible that a switch will be closed
into a fault. The close and latch capability of a device is meant to protect the operator from this error. Even simple switches, like the
loadbreak elbow, have close and latch ratings. If an elbow is closed into a fault, within its rating, it will survive but it should be
replaced.
f. BIL. BIL (basic impulse level) is a rating which allows the user to assess the voltage impulse capability (ability to withstand
impulses
 Page 16
without failure) of the equipment. To establish this rating, the equipment is tested with a voltage impulse wave defined as 1.2 × 50 ms.
An example will help illustrate what all this means. Equipment on a 13.8 kV system normally has a BIL of 95 kV. This means that this
equipment has been impulsed with a wave (see Figure 1-11) which rises to a crest value of 95 kV in 1.2 microseconds and decays to
half its value in 50 microseconds.

Figure 1-11.
Standard BIL Wave

Grounding
Grounding on the primary distribution system (2400 volts to 34.5 kV) is usually either 4-wire multigrounded or delta. A 4-wire
multigrounded system, which is by far the most popular, means that the substation is grounded and a fourth wire, the neutral wire, is
carried along with the phase wire and grounded periodically (see Figure 1-12(a)). Some utilities ground as little as four times per mile
while others ground at every pole. Sometimes the substation's transformer is grounded through an impedance (approximately 1 ohm) in
an effort to limit short circuit current levels. Some of the more important advantages of a 4-wire multigrounded system (over a 3-wire
delta) are:
 Page 17
1. High short circuit currents allowing effective overcurrent relaying practice.
2. Much cheaper for single-phase service, especially underground, since only one cable, bushing, switch, fuse, etc., needs to be used as
compared to a delta system which needs 2 times as much equipment.
3. Lower rated arresters and lower rated BIL required.
Figure 1-12 shows the types of distribution circuits.
 Page 18

Figure 1-12.
 Page 19
The second most popular type of grounding for a distribution system is the 3-wire delta system shown in Figure 1-12(d). These systems
are generally older and lower in voltage than the 4-wire multigrounded type. They are also very popular on industrial power systems.
Although still in use, they are not being actively expanded since they lack some of the advantages shown for the 4-wire multigrounded
design. Some of the advantages of a 3-wire delta are:
1. Better phase balancing
2. Lower energy into fault
3. Less EMF.
The three other designs shown in Figure 1-12 (12(b), 12(c), 12(e) are all systems presently in use but only on a very limited basis.
Four-wire unigrounded systems (Figure 1-12 (b)) are systems where the primary neutral conductor is insulated at all points except at
the source. The neutral conductor in these systems is connected to the neutral point of the source transformer windings and to ground.
Distribution transformers usually are connected between phase and neutral conductors with the surge arrester connected between phase
and ground. Some four-wire unigrounded systems use an arrester between the neutral conductor and ground. A spark gap may also be
used at the distribution transformer between its secondary neutral and arrester ground to provide better surge protection to the
transformer windings. The principal advantage of four-wire unigrounded systems is the greater ground relaying sensitivity which can
be obtained in comparison to multigrounded systems.
On three-phase three-wire unigrounded primary distribution circuits (see Figure 1-12(c)), single-phase distribution transformers are
connected phase-to-phase. The connection of three single-phase distribution transformers or of three-phase distribution transformers is
usually delta-ground wye or delta-delta. The floating wye-delta or T-T connections also can be used. The grounded wye-delta
connection is generally not used because it acts like a grounding transformer.
 Page 20

Load and Fault Characteristics*

Typical Feeder Load Characteristics
While many distribution engineers are aware of the maximum constraints on system components (e.g., 3-phase underground equipment
is generally rated 600 amperes and single-phase underground equipment is generally rated 200 amperes maximum), some of the
average characteristics are somewhat less known. Figure 1-13, shown below, illustrates a system design which might be considered
fairly typical or average. Obviously, many systems have longer lines (up to 30 miles), higher voltages (up to 34.5 kV), larger fuses (up
to 200 amperes), and higher short circuit levels approaching 30 kA.

Figure 1-13.
Typical Distribution System
* These characteristics are developed for 4-wire, multigrounded systems only.
 Page 21
Fault Currents
Fault levels are easily calculated using standard computer programs available to almost all distribution engineers. A good calculation of
these levels and a proper understanding of what they mean is very important because fault levels are used to determine protective
coordination, switchgear duty, etc. The largest concern when making these calculations is whether to use bolted fault conditions (Z =
0) or a higher fault impedance to reflect the so-called worst case scenario. Studies have shown that there are two fairly distinct types of
faults: high impedance and low impedance. If the fault arcs to the neutral, the fault can be classified as low impedance reflecting a fault
impedance of less than two ohms and usually somewhat less. If, on the other hand, the phase wire lands on the ground and does not
contact the neutral conductor, then the fault impedance is very high (usually 300 ohms or more) and the current produced by them is
very low, too low to ever be seen by conventional protection practices. Based on the studies described below, experience would
indicate that coordination studies using bolted fault conditions is probably the best approach.
Low Impedance Faults. Coordination of protective devices on a distribution system is greatly affected by the short circuit levels on the
system. Fault levels are known to vary with distance as illustrated in Figure 1-14. The available short circuit current at the substation is
shown to be about 6500 amps, but could be as high as 30 or 40 kA in larger substations near metropolitan areas. The average level
reported by utilities at the substation was approximately 10,000 amperes. Many utilities subscribe to the concept that faults have an
impedance which consequently limits fault current values to levels considerably less than the calculated bolted fault magnitudes. To
compensate for this during coordination studies, some utilities add fault impedance to their short circuit calculations on the order of 5
to 50 ohms. This methodology, while conservative, not only limits the reach of many devices, but as such may cause miscoordination
if the actual fault impedance is different from the one used in calculations. There is no question that some faults are high impedance
faults and as such limit fault current to levels of approximately 50 amperes or less (i.e., hundreds of ohms fault impedance).
Two studies provide the raw data for this discussion. The first is work performed on the Virginia Power 34.5 kV system in 1977 and
the
 Page 22
second is edited from work completed for EPRI in 1983 which consisted of the instrumentation of 50 feeders in thirteen utilities.

Figure 1-14
In those cases studied where the location of the fault occurrence was known, it was possible for Virginia Power to calculate the
magnitude of fault current and compare this with the measured values. On the 34.5 kV circuits, Virginia Power made every effort to
include all the resistance and reactances of equipment, but did not put in any fault resistance. Figure 1-15 indicates the relative
accuracy of the calculated values as compared with the actual measured values. As can be seen, almost 50% of the calculated values
were within ± 5% of the actual implying that there was little, if any, fault impedance. Most of the error is where the actual currents
were less than the calculated, as would be expected.
In the late 70's, a large 4 year fault study was performed. Fifty feeders at various voltage levels were chosen to be instrumented at
thirteen utilities.
 Page 23

Figure 1-15.
Percent Differences of Actual Short Circuit Currents
from Calculated Values
Figure 1-16 indicates the relationship between utility-calculated fault current levels and actual recorded fault currents. In cases where
calculated and recorded fault currents are identical, the plotted data point will fall on the 45° line.

Figure 1-16.
Comparison of Actual vs. Calculated Fault
Current Magnitudes
 Page 24
In general, utility calculation of fault currents were extremely close to recorded values. No fault resistance was assumed in the utility
calculations, which proved to be a valid assumption, since essentially no fault resistance was seen in the recorded faults. The average
utility operated its feeders at approximately 7% above nominal voltage rating, with one recorded instance of a 17% steady-state
overvoltage noted. A small amount of fault resistance, on the order of 2 ohms, would explain those faults that fall below the 45° line.
Points falling above the 45° line reflect the operating voltage increase above normal or simply the slight inaccuracy of sequence
component calculations.
High Impedance Faults. While there have been many high impedance fault studies, many of these do not seem to deal directly with
magnitudes of currents during these tests. The information shown on the following page in Table 1-3 illustrates some of these values
and indicates the low levels of currents seen during this type of fault. Typically, tests have shown that faults to surfaces such as asphalt,
grass, and gravel are less than 50 amps. Faults to concrete, especially wet reinforced concrete, can be of the order of 200 amperes.
Table 1-3. High Impedance Fault Test Summary
Location Date Tests
1. 3/78 Normal system switching events were performed:
TESCO- capacitor bank feeder tie breaker operations, large
Handley motor start.
2. 7/11/78Twelve fault tests consisting of energizing
RG&E conductor on ground and lowering energized
conductor to ground. Conductor laid on dry sod,
dry and wet macadam. Arcing faults of 1050 A on
sod. No fault current on dry macadam, slight
smoke on wet macadam.
3. 8/23/78Two fault tests on soil each with about 20 A
TESCO- current. Normal switching events were also
Handley performed: capacitor bank, feeder tie, and breaker
operations.
4. Gulf 9/20/78Normal switching events were performed:
States capacitor bank and feeder tie operations.
5. 2/2/79 Five fault tests with lateral lowered on very wet
TESCO- and partially frozen oil and conductor lowered to
White touch a grounded metal post. Approximately 2040
A currents for faults on soil and 5060 A faults to
the post. Arcing occurred in each case.
 Page 25
Table 1-3. High Impedance Fault Test Summary - Con't
Location Date Tests
6. 5/15/79At White: four fault tests on relatively dry ground.
TESCO- Arcing occurred in each case with about 30 A
White current. At Handley: normal system data recorded.
7. 6/1/79 Thirteen faults on dry sod, reinforced and non-
RG&E reinforced concrete, wet and dry macadam. Several
tests with covered conductor produced small fault
current on sod. Small fault current on non-reinforced
concrete, higher magnitude on reinforced concrete.
Little or no fault current on macadam. Subsequent
tests on sod produced 50 A currents. Normal system
was monitored as were switching events: capacitor
bank and load tap changer operations.
8. 6/21/79Eight faults on dry and wetted sand and on a
TESCO- mesquite bush. Less than 1 A current on sand, 3040
Sand A current on the mesquite bush (due to deep roots).
Hills
9. PNM 7/23/797/2: Instrumentation located at a regulator bank 6
miles from substation. Faults at 1/8, 6, 11, 15, 18
miles from regulator
7/3: Instrumentation at substation, faults at 6, 12, 17
miles.
Fault current 1030 A, faults staged with capacitor in
and out and regulator in and out.
10. 2/2/80 Eleven fault tests on dry cement and very wet soil.
TESCO- Capacitor bank and load tap changer operations. One
Randol fault blew the fuse. Faults on cement had little
Mill current, on soil had high magnitude.

Ground Fault Current Flow. Good grounding has been credited with alleviating a variety of utility ills including line flashover, stray
voltage and reduced transformer failure rates. Utilities try to maintain 25 ohms or less when placing ground rods but in many utilities
obtaining such a low value is just about impossible, regardless of the number of ground rods, due to poor soil conditions. Several
utilities have major programs to reduce ground rod impedance. Shown below is a lattice diagram illustrating the flow of ground fault
currents for a line-to-ground fault at the end of the feeder. The ground rod impedance used for this calculation was 25 ohms at each
pole. It is interesting to note that if the
 Page 26
footing resistances are increased, the ground current will decrease, i.e., more current will flow in the neutral and shield wires.

Figure 1-17.
Ground Current Distribution
Inrush
Inrush current on a distribution system affects virtually all overcurrent protective device selection and settings. Although there are no
comprehensive standards to define inrush, the industry did perform a large study in an attempt to characterize this condition. Figure 1-
18 shows a distribution of inrush current magnitudes found in this study. The parameters chosen of peak magnitude, offset ratio, and
decay time constant were used primarily because of their importance in fault current characterization for circuit breaker applications.
Some general observations on inrush currents are as follows:
 Page 27

Figure 1-18.
Inrush Magnitudes
The peak instantaneous inrush current averaged 2448 amperes, with a maximum inrush current of 5700 amperes seen in one instance.
The rms magnitude of the first half-cycle of inrush current was, on average, 2.55 times the rms value of current over the first four
cycles. A peak offset ratio of 5.3 was seen in one case.
The average decay time constant was 53.9 milliseconds (about 3 cycles), indicating that the inrush transient is practically nonexistent
after 9 or 10 cycles.
Average fault current magnitude and average inrush magnitude were approximately the same.
Questions
1. The most common distribution voltage is 34.5 kV, 2.4 kV or 13.8 kV?
2. Why does short circuit level drop off so fast at relatively short distances from the substation?
 Page 28
3. Short circuit levels should be calculated from the (OA, FA, FOA) rating of a triple rated substation transformer?
4. What are some of the concerns associated with long feeders?
5. Laterals on a typical distribution system are usually fused (true or false)?
6. Name one advantage and one disadvantage of a secondary network system?
7. A typical load level for a distribution feeder is _____ amperes and a typical fuse size for that load is about _____. Why should the
lateral fuse size be higher than the load?
8. The most common type of grounding on a distribution system is ________.
9. High impedance faults refer to low current faults at the end of the feeder (true or false)? Explain!
10. Average inrush current levels are not a problem for relay engineers because relay settings are based on fault levels which are much
higher in magnitude (true or false)?
 Page 29

2
Transformers and Regulators
Transformers
Introduction
The distribution system is unique in that it uses not only many types of transformers but also many different transformer connections.
This by itself can make an otherwise simple concept seem more confusing than it should be.
The purpose of this chapter is to present some of the major concepts that concern the distribution systems engineer and operator.
Things like types of transformers, loading of transformers, construction and rating are discussed as they are applied to the distribution
system. Also, those special transformers, like autotransformers and grounding transformers, are presented since they, too, are common
to the distribution system.
Transformer Model
The basic transformer consists of two windings, which are electrically insulated from each other and wound around a laminated core of
magnetic sheet steel, as shown in Figure 2-1. Energy is transferred from the source E1, or primary winding, via magnetic flux to the
secondary winding. Assuming no transformer losses, power from the source, P1, must equal power to the load, P2. The secondary
voltage, E2, is a function of the turns ratio of the transformer, i.e., E2 = N2/N1*E1, where N1 is the number of turns in the primary
winding and N2 is the number of turns in the secondary winding. Of course this means that as the number of turns on the secondary
decreases relative to the primary, as in the case of a step-down transformer, the voltage goes down and the current must go up.
 Page 30

Figure 2-1.
Basic for a Lossless Transformer Construction and Relations
The primary and secondary windings are coupled by the transformer entirely through mutual flux, i.e., there is no direct electrical
connection since both windings are insulated. As shown in Figure 2-2, not all the flux links the two windings. Some of it leaks, hence
its designation as leakage flux. It follows that as the load increases and the currents go up so does the leakage flux and hence so do any
losses associated with this leakage.
The equivalent circuit for a two-winding transformer is shown in Figure 2-3. As can be seen, there are essentially the following two
impedances associated with a transformer.
 Page 31

Figure 2-2.
Flux Relations of Transformer

Figure 2-3.
Transformer Equivalent
Leakage Impedance. This impedance is composed of Rp+Xp+Rs+Xs or the winding resistances and leakage reactances of the primary
and secondary windings. This impedance is also called the through impedance and is the impedance used for the calculation of short
circuit
 Page 32
current and load losses. The impedance referred to on the nameplate of a transformer is this impedance e.g., the impedance of a
substation transformer is approximately 10%. This nameplate impedance is associated only with the leakage impedance (also called the
winding impedance).
Magnetizing Impedance. When a transformer is energized, it takes a small amount of energy to simply magnetize the iron core (make it
hum). The magnetizing branch of the equivalent circuit represents that part of the transformer that consumes energy even under no load
and consists of losses due to hysteresis and eddy currents as well as the inductive reactance required to produce the mutual flux. This
(magnetizing) impedance is generally quite high and in most cases is neglected since so little current goes through this branch. The
nameplate of a transformer normally makes no reference to this impedance since it is rarely used except for transformer loss
calculations.
The current through this impedance varies only slightly with load since E3 is relatively constant under variable load conditions. The
magnetizing branch results in losses which are relatively constant since both current in this branch and voltage across this branch are
essentially constant. Losses associated with this branch are called no load losses since they occur whether there is load on the
transformer or not. An equivalent circuit for no load is shown in Figure 2-4.

Figure 2-4.
Primary Circuit of a Transformer Under No Load
 Page 33
Autotransformers
The autotransformer is a special type of transformer that is used in various types of distribution applications, primarily where the
voltage transformation is small and cost is a major consideration. The equivalent circuit for a step-down autotransformer is show
below. As can be seen, the thing which distinguishes an autotransformer from a standard two-winding unit is the direct electrical
connection between the primary and secondary. While this connection has advantages in the areas of voltage regulation and energy
transfer, it has drawbacks concerning high short circuit currents (because the transformer leakage impedance is usually less than 5%)
and the fact that the primary is not completely insulated from the secondary.

Figure 2-5.
Circuit of Step-Down Autotransformer
Probably the best way to demonstrate the advantage of an autotransformer is with a hypothetical example. Suppose you purchased a 2-
winding, single-phase, 25 kVA distribution transformer rated 7200 volts on the primary and 240 volts on the secondary, as shown in
Figure 2-6a. We know that the power in (Vp*Ip) must equal the power out (Vs*Is) and both should equal 25 kVA.
 Page 34

Figure 2-6.
Conversion of 2 Winding Transformer to a Step-Up Autotransformer
If we reconnect the transformer so that point a of the primary is physically attached to point b and the load is connected between the
phase wire and ground, we now have a step-up autotransformer (see Figure 2-6b). The secondary voltage is now higher than the
primary by 240 volts, i.e., it's 7200 volts plus 240 volts or 7440 volts. The magnetic power transfer of the transformer is exactly the
same as it was since the currents through the windings are the same as they were (as are the
 Page 35
voltages across each winding). The voltage across the load and the current from the source, however, have both increased dramatically
to about 774 kVA. This means that the power transfer capability of this transformer has been increased by over 30 times, which is
mainly the result of the direct electrical transfer of energy now taking place.
Construction
There are two types of transformer construction commonly referred to in the utility. These are referred to as core type and shell type
and are shown in Figure 2-7. The obvious difference in these two types of construction is that the windings in the shell type
construction are surrounded by core (iron) material. Some manufacturers will make the claim that this makes the transformer inherently
more capable of handling the high mechanical forces imposed on the transformer during a short circuit. While in theory this may be
true, in actuality core type transformers produced by quality manufacturers have reliability rates as good or even better than most
manufacturers of shell type.

Figure 2-7.
Core and Shell Type Transformer Construction
 Page 36
Types of Transformers
Distribution Substation Transformers. Distribution substation transformers come in such a wide variety of ratings that it would be
impossible to cover them all. Figure 2-8 illustrates what might be considered a typical substation size and layout.

Figure 2-8.
Typical Distribution Substation
As can be seen, this 2-winding transformer has a 138 kV to 13.8 kV voltage transformation with an 8% transformer impedance. Some
substations use autotransformers which usually have less than half that impedance. Some typical characteristics (not all inclusive) that
the distribution engineer might find in a utility system substation are as follows:
High Voltage Winding - 230 kV to 34.5 kV
Low Voltage Winding - 34.5 kV to 4160 volts
Size in MVA (OA) - 2.5 to 75 MVA per transformer
Transformer Impedance - 5% to 12%
 Page 37
Number of Transformers in Substation - 1 to 4
Loading - OA, OA/FA, OA/FA/FOA, OA/FA/FA
High Side Protection - Circuit Switchers, Breakers, Fuses
Relay Protection - Overcurrent, Differential, Under-Frequency
Feeder Protection - Breakers, Reclosers
Most distribution substation transformers larger than 5 or 10 MVA are protected by breakers or circuit switchers and use both
differential and overcurrent relays. These transformers are usually regulated by LTC (load tap changers) which adjusts voltage level
depending on loading conditions. The secondary bus is usually split (normally open) to reduce the number of short circuits seen by
each transformer as well as to reduce the short circuit level of the system. Keeping this bus tie open also reduces any circulation
currents that might take place should the transformer impedances not be exactly matched.
Loading of the transformer is usually kept within its OA rating (passive convection), with the FA (fans) and FOA (fans and oil pumps)
ratings utilized for emergency conditions. Both the FA and FOA ratings increase the loading by about 33% each in transformers rated
10 MVA and greater.
Two calculations the reader should be familiar with in regard to a distribution substation transformer are load current and short circuit
current. They are as follows for the transformers shown in Equations 2-1 through 2-3.
Calculation of Full Load Current:

2-1
 Page 38
Calculation of Short Circuit (line-ground):

2-2
Alternate method:

2-3
*Only the OA rating should be used to calculate short circuit currents.
Distribution Transformers. The transformer that connects the high voltage primary system (4.16 kV to 34.5 kV) to the customer (at 480
volts and below) is usually referred to as a distribution transformer. The variety of ratings, transformer connections, tank
configurations, protection schemes, and loading practices are much too diverse to be covered by a single chapter so only some of the
more major areas can be discussed.
These transformers can be either single-phase or three-phase and range in size from about 5 kVA to 500 kVA. Table 2-1 shows some
of the standard kVAs and voltages for these units. Pad mounted transformers, used for underground service come in about the same
ratings and sizes with the exception that 3-phase padmounted units can be rated as high as 2500 kVA. Impedances of these 2-winding
transformers are generally quite low, ranging from about 2% for units less than 50 kVA to about 4% for units greater than 100 kVA.
 Page 39
Table 2-1. Standard Transformer KVAs and Voltages
KVAs High Voltages Low Voltages
Single-Three- Single- Phase Three- Phase Single- Three- Phase
Phase Phase Phase
5 30 2400/4160Y 2400 120/240 208Y/120
10 45 4800/8320Y 4160Y/2400 240/480 240
15 75 2400/4160Y 4160Y 2400 480Y/277
25 112½ 4800Y/8320YX 4800 2520 240X480
37½ 150 7200/12,470Y 8320Y/4800 4800 2400
50 225 12,470GrdY/7200 8320Y 5040
4160Y/2400
75 300 7620/13,200Y 7200 6900 4800
100 500 13,200GrdY/7620 12,000 7200 12,470Y/7200
167 12,000 12,470Y/7200 7560 13,200Y/7620
250 13,200/22,860GrdY 12,470Y 7980
333 13,200 13,200Y/7620
500 13,800/23,900GrdY 13,200Y
13,800 13,200
14,400/24,940GrdY 13,800
22,900 22,900
34,400 34,400
43,800 43,800
67,000 67,000

Pole type, as well as underground transformers, can be either single-phase, or three-phase depending on the requirements of the load
and the configuration of the primary supply. Figure 2-9 shows some of the more common pole type configurations. As can be seen,
single-phase transformers can have one or two primary bushings. If the primary system was delta connected (no neutral wire) then the
primary of the transformer would have to be connected phase-phase requiring two bushings. If the system was a 4-wire multigrounded
system, then single-phase loads would normally be connected phase-to-neutral (ground) and only one primary bushing would be
required (some utilities prefer 2 primary bushings even on these systems for reasons of flexibility).
Some of the more common high-voltage connections seen at the transformer are shown in Figure 2-10. As can be seen the winding
configuration affects the way the voltage rating is applied. This is primarily because grounded wye transformers may have tapered
insulation (less insulation near the neutral) and as such cannot be applied on an ungrounded system.
 Page 40

Figure 2-9.
Pole-Type Transformer Connections

Figure 2-10.
High Voltage Connections
Distribution transformers are assigned kVA ratings which indicate the continuous load the transformers carry and not exceed the
specified temperature rise of either 55°C or 65°C. In service, a distribution transformer is rarely loaded continuously at its rated kVA
and typically goes through a daily load cycle. Because the transformer has a relatively long thermal time constant, i.e., the oil
temperature goes up slowly relative to the load change, it is possible to load the transformer beyond its rating without serious effect on
the life of the unit. Table 2-2 shows
 Page 41
a typical loading guide. Many utilities have successfully used even more severe loading guides for their smaller single-phase
distribution transformers with no apparent problem with life expectancy.
Table 2-2. Permissible Daily Short-Time Transformer Loading Based
on Normal Life Expectancy
Maximum Load in Per Unit of Transformer
Rating
Period of Increased
Average Initial Load in Per Unit of
Loading, Hours
Transformer Rating
0.90 0.70 0.50
0.5 1.59 1.77 1.89
1.0 1.40 1.54 1.60
2.0 1.24 1.33 1.37
4.0 1.12 1.17 1.19
8.0 1.06 1.08 1.08

Distribution transformers are manufactured in a variety of ways from as simple as a 2-winding transformer with no protection to a self-
protected (CSP) transformer illustrated in Figure 2-11. The self protected distribution transformers arrive to the user equipped with a
primary surge arrester, an internal high voltage oil immersed expulsion (weaklink) fuse and a low voltage circuit breaker that provides
protection from some secondary faults and severe overloads.

Figure 2-11.
Self-Protected Distribution Transformer
 Page 42
The low voltage circuit breaker trips on the basis of temperature. This tripping temperature is caused by the combination of
surrounding oil and the current flow in the bi-metallic sensor. As an example, a 25 kVA transformer having 75% initial load, and a 35C
ambient would initiate an overload light and warning after approximately 2 hours of 2-per-unit overload and a secondary breaker trip in
approximately 2 1/2 hours at the same overload level. Finally, many utilities do not use secondary breakers because they find them
more of a nuisance than a help.
Network Transformers. Network transformers are large (300 kVA to 2500 kVA), liquid filled, three-phase transformers used to supply
secondary network systems (see Figure 2-12). The high voltage windings are generally anything from 4160 volts to 34.5 kV and the
secondary voltage rating is usually 216GrdY/125 or 480GrdY/277 volts. New systems and virtually all spot networks, used for highly
concentrated loads such as commercial buildings, use the higher secondary voltage rating. Since the units are usually installed under
streets, they must be sealed and corrosion resistant.
 Page 43

Figure 2-12.
Typical Secondary Network
The secondary network vault usually contains three fundamental components as shown in Figure 2-13:
1. High Voltage Switch. This switch allows the transformer to be connected and disconnected to the primary feeder. It does not have
 Page 44
load break capability and can be opened and closed only when the network protector is opened.
2. Network Transformer. These 3-phase transformers are liquid (PCB or silicon) filled, and connected either delta/GrdY or
GrdY/GrdY. Units 1000 kVA or smaller generally have an impedance of about 5% whereas the larger units, 1500 kVA and above,
have an impedance closer to about 7%.
3. Network Protector. The network protector consists of an air circuit breaker and relays that operate for reverse power flow (e.g.,
primary fault). The network protector also has a fuse whose purpose it is to backup the network protector breaker as well as provide
some thermal overload capability to the transformer. This fuse is not sized to provide protection against secondary faults. Secondary
faults are generally considered to be self-clearing, i.e., they burn themselves clear or are cleared by fuses at the cable's junctions called
current limiters. Because many faults at that level have been found not to burn themselves clear, 480 volt network systems pretty much
require current limiters.

Figure 2-13
Dry Type Transformers. Dry type transformers are transformers whose insulation medium is solid (glass tape, glass silicon, porcelain,
etc.) and not a liquid like oil. Dry type transformers are primarily used where safety is a major concern e.g., transformers in buildings
in close proximity to people. With the elimination of askarel (PCB's) as an insulation
 Page 45
medium for transformers, for environmental reasons, the importance of dry type transformers significantly increased. The kV and kVA
ratings of dry type transformers are very similar to oil filled transformers. Standard kVA ratings range from below 5 kVA up to 20
MVA with voltage ratings from less than 2400 volts up to 230 kV.
Dry type transformers are not considered as rugged as oil filled transformers of the same rating. To insure reliability most users have
found it prudent to never exceed the loading recommendations of the manufacturers. This is somewhat in contrast to the loading
practices of some utilities where overloading beyond the standards of liquid filled units is somewhat routine and has not produced
excessive failure rates. Also, a concern with dry type transformers is their lower dielectric strength. Table 2-3, shown below, illustrates
that a 13.8GrdY/7970, dry type transformer has BIL of 60 kV. This is only 63% of an equivalent oil filled unit which has a BIL of at
least 95 kV. Special dry type transformers are made with increased BIL.
Table 2-3. Insulation Levels for Dry Type Transformers
BIL BIL
Nominal Equipment (kV) Nominal Equipment (kV)
Voltage (1) (2) Voltage (1) (2)
1201200 12,000
1200GrdY/693 10 13,800
13,800GrdY/7970 60
2520 20
4360GrdY/2520 18,000
22,860GrdY/13,200 95
4160
7200 30 23,000 110
8720GrdY/5040 24,940GrdY/14,400
8320 45 27,600
34,500GrdY/19,920 125
34,500 150
 Page 46
Secondary Faults
The calculation of bolted fault currents on the secondary of the distribution transformer is sometimes necessary in order to select
service entrance equipment with adequate interrupting rating as well as to determine overcurrent coordination of secondary protective
devices. Since the fault can occur across either the 120 volt or 240 volt circuits, this calculation can sometimes get a little confusing.
The equations shown below can be used for this purpose as well as to calculate secondary faults on the secondary cable itself.

2-4

KVA =Transformer nameplate rating in kVA

WTOT =Transformer total losses at full load in watts
WNL =Transformer no load losses in watts
Z% =Transformer nameplate impedance in percent
RT =Transformer resistance in ohms at secondary terminals
X1X3
ZT =Transformer leakage impedance in ohms at secondary
terminals X1X3
L =Circuit length from the transformer secondary terminals to
fault in feet
 Page 47

2-5

I240=Available current for a bolted 240 volt (phase-to-phase) fault in

amperes rms symmetrical
I120=Available current for a bolted 120 volt (phase-to-neutral)fault in
amperes rms symmetrical
RS =Resistance of secondary circuit for a 240 volt fault in ohms per
1000 feet
RS1 =Resistance of secondary circuit for a 120 volt fault in ohms per
1000 feet
XS =Reactance of secondary circuit for a 240 volt fault in ohms per
1000 feet
SS1 =Reactance of secondary circuit for a 120 volt fault in ohms per
1000 feet.

Distribution Transformer Connections

The combination of lighting and large 3q power loads, brings up the question as to what is the best transformer connection? This is not
an easy question since there is such a wide variety of connections and considerations therein. Some of the more common areas of
interest are described as follows:
3q Secondary Voltage
An important consideration for supplying 3q load is whether it should be connected to a delta or wye connected secondary (see Figure
2-14).
 Page 48

Figure 2-14.
Secondary Loading of Delta and Wye Connection
It will be noted in the delta connection that all of the single-phase load is concentrated on one phase, whereas on the wye connection all
the single-phase (lights and small appliance) load is distributed on each of the three phases.
It is evident that if a large number of single-phase loads are to be served from the bank, the delta secondary connection is less suitable,
because of the unbalanced loading. On the other hand, if the single-phase load is small, as compared with the polyphase load, the delta
connection is preferable.
In urban distribution systems, particularly in heavy commercial areas, the trend is toward the four-wire, three-phase, 120/208Y volt
system. In rural distribution, where the need for three-phase secondary is limited to isolated loads, the four-wire, three-phase, 120/240
volt delta system is perhaps more common.
In selecting a secondary transformer connection, the performance characteristics of motors and other load devices at other than rated
voltage, must be considered. For example, with the delta connection, a 120 volt standard lamp and a 220 volt three-phase motor will be
applied at their name plate rating; whereas, with the wye connection, the 120 volt lamp will have name plate voltage applied but only
208 volts will be available for power purposes. Unless 208 volt motors are explicitly specified, standard 220 volt motors will not
perform as might be expected.
 Page 49
Wye-Delta
This bank can have the neutral grounded or isolated. With the neutral grounded the bank may be operated as open wye-open delta at
reduced capacity if one phase in the primary is lost but it also acts as a grounding bank for the system and as such will replace the main
source in feeding short circuit current to a line-to-ground fault after the protective devices have opened. To avert this and possible
transformer burnout, it is recommended that a three-phase wye-delta bank be ungrounded.
With transformer connected wye-delta and the neutral isolated, if one phase of the primary is opened by a fuse blowing or for any other
reason, the ungrounded bank ceases to operate as a three-phase bank and is unable to carry a three-phase load. It is therefore a safe
practice to have three element protectors on all three-phase motors fed from this bank.
Delta-Wye
This bank is similar to the wye-delta bank in many of its characteristics and it may also have the neutral grounded or isolated. With the
neutral grounded the bank is more flexible in the sense that it may consist of units of different rating as determined by 1/3 of three-
phase load plus whatever single-phase load is connected to it. Only 120/208 volt service will be available from this bank, compared to
120/240 volt three-wire service that may be obtained from wye-delta banks. If the neutral is isolated and is not carried through as a
fourth wire, the bank will not be expected to supply 120 volt loads.
This type of connection, i.e., delta-wye with the neutral grounded is the most popular connection in distribution substation banks where
transformation may be 69 kV to 7200/12470GrdY or 2400/4160Grd wye.
Delta-Delta
This type of connection should be used only when the percentage impedance and the ratio in all units are identical. Even a slight
difference in the voltage ratio among the units may cause large circulating currents to flow in the windings which can reach dangerous
proportions and heat up the transformers even under light loads. A different setting of taps is in its effect equivalent to units of different
transformer ratios; therefore, they should have not only the same voltage ratio but also the same voltage tap setting.
 Page 50
This bank connection will provide the possibility of serving a three-phase load with only two transformers connected open delta. This
is an emergency measure at which the total bank output will be 57.7% of the original bank capacity.
Wye-Wye
The most common connection that may be considered under this heading is the neutral grounded on both sides of the bank. In a three-
phase distribution system with a multigrounded neutral, the grounding of the primary and secondary neutral of a service bank appears
to be a good solution for feeding single-phase and three-phase loads with a 120/208 volt supply. One reservation to this type
connection is the necessity of using 208 volt motors instead of standard 220 volt motors, unless the 120/208 volt system is a part of a
secondary network system, with very close voltage regulation.
This type of connection is sometimes used on 14.4/24.9 kV systems where experience has shown that the ungrounded wye-delta bank
is cause for critical overvoltages when a phase is de-energized.
The wye-wye grounded neutral bank may be made up of transformers selected as the load on each phase requires without limitation in
their characteristics and kVA ratings.
Third harmonic currents will circulate freely in the primary and secondary circuits of the bank and may become a source of telephone
interference when telephone lines run close to or parallel to the power circuit.
Open Delta
The use of open delta or V connected transformers for a three-phase service is not to be considered as a standard practice and it is
recommended only in special cases as listed below. The bank must be operated at reduced capacity and its regulation is different in
each leg creating an unbalanced voltage condition on the secondary. Therefore, its use should be considered only under the following
conditions:
1. If one of the transformers of a wye-delta bank burns out or must be disconnected, the two remaining transformers can be operated to
furnish three-phase load at reduced capacity provided the primary neutral is connected to ground and common neutral wire.
 Page 51
2. If only two phases are available on the primary line and it is considered uneconomical to string the third phase to pick up a three-
phase load.
Transformers connected in open delta or V connected are able to carry 86% of their rated capacity. Voltage regulation in each phase
will be different for different values of load power factor.
Grounding Banks
An objection to a wye-delta connection is that it tries to act as a grounding bank during a fault at some other point on the system.
Figure 2-15 shows how fault currents flow from the substation transformer with a line-to-ground fault at X. The currents in the step-
down wye-delta bank where the high voltage neutral of the transformer bank is isolated, are not affected by the fault at X.

Figure 2-15
Figure 2-15b shows the same system, except that the neutral of the step-down transformer bank is connected to the neutral conductor.
With a fault at X, the same fault current flows from the substation transformer
 Page 52
as in Figure 2-15a. These currents are indicated by the solid arrows in Figure 2-15b.
The fault at X reduces the voltage across C'N which reduces the voltage across ac, causing circulating currents in abc that appear as
currents in A'n, B'N, and C'N. A very low resistance fault at X will result in a practical collapse of all voltage across C'N, producing
extremely heavy circulating currents in the transformer windings. These currents (dotted arrows in Figure 2-15b) add to the normal
load current in the transformer bank and are apt to burn it out.
Summary
A summary of the characteristics of single-phase transformers connected for combined 3<?> and 1<?> loads is shown in Table 2-4.
Page 53
 Page 54

Regulators
Introduction
The voltage regulator is simply a tapped autotransformer and as such works in a similar fashion, to either raise or lower voltage. On a
distribution system, regulators can be found in the substation where they are referred to as station or bus regulators, or out on the feeder
where they are referred to as supplementary or pole type regulators. Because an autotransformer uses both a direct electrical connection
as well as magnetic flux to transfer energy, the regulator nameplate rating in kVA will be much lower than the kVA of load it can
regulate. Typically, most of the energy transfer (approximately 90%) of the regulator comes via the direct electrical connection when
the regulator is either in the full boost or full buck mode. At nominal tap setting virtually all the energy is transferred by direct
electrical connection allowing some regulators to have higher kVA ratings for reduced tap settings.
Figure 2-16, shown below, shows a simple schematic of a single-phase voltage regulator. As can be seen, a typical voltage regulator
used out on the feeder (pole mounted), usually has a bucking range equal to the boost range or ±10%. Each tap normally gives a 5/8%
voltage change, so there are generally 32 taps.
A typical example of the use of a feeder regulator is shown in Figure 2-17. In this example, the supplementary feeder voltage regulator
is located at a point on the feeder where the voltage starts to dip below 119 volts. The voltage standards require a voltage to the
residential customer to be between 126 and 114 volts. This voltage is measured at the meter. In order to compensate for the voltage
drop from the feeder to meter, i.e., the drop through the transformer and secondary, the feeder voltage must be somewhat higher than
114 volts. Some utilities use a 5 volt drop to estimate this reduction thus requiring a 119 voltage (114V+5V = 119 volts) on the feeder
itself. Another point worth noting is that the feeder voltage can be higher than 126 volts as long as the voltage by the time it gets to the
customer meter is reduced to 126 or less.
Regulator kVA Rating
Single-Phase. The rating of a single-phase feeder voltage regulator is the product of the rated load amperes and the rated range or
regulation in kilovolts. The rated range of a regulator with a ±10% capability is 10%
 Page 55

Figure 2-16.
Schematic of Pole Mounted Single-Phase Regulator

Figure 2-17.
Supplementary Feeder Regulation
 Page 56
not 20%. For example, the minimum size single-phase regulator needed to supply 100 amperes on a 7620 volt line-to-ground system
(4-wire multigrounded system) would be calculated as follows:
1 Ø Regulator kVA = Range × System kVA
= .10 × 7620V × 100 amps 2-6
= 76.2 kVA (minimum)
For a single-phase delta rated 13200 volts, phase-to-phase, supplying a load of 100 amperes, the minimum regulator rating would be:
1 Ø Regulator kVA = Range and System kVA
= .1 × 13200 volts × 100 amps 2-7
= 132 kVA (minimum)
Three-phase. Three-phase voltage regulators can be applied either to three-phase, three-wire or three-phase, four-wire circuits. The
kVA ratings are determined in the same manner for each type system and the same regulators can be used. As shown in Figure 2-18,
the phases are normally wye connected internally. On the four-wire circuits the regulator neutral connection is connected to the fourth
neutral wire. On three-wire ungrounded circuits the neutral is normally connected through a surge arrester to ground.
 Page 57

Figure 2-18.
Three-Phase Regulator for Use on Three-Phase, Three-Wire or
Three-Phase, Four-Wire Circuit
When single-phase regulators are to be used on three-phase circuits, the regulator arrangement for a three-phase four-wire circuit is
quite different from that on a three-phase three-wire circuit.
On a three-phase, four-wire circuit, three, single-phase regulators, whose voltage rating is the line to neutral voltage rating of the
circuit, are connected in wye, as shown in Figure 2-19.
 Page 58

Figure 2-19.
Three Single-Phase Regulators Wye Connected on Three-Phase,
Four-Wire Circuit
On the three-phase, three-wire circuit the usual connection is to use two regulators, whose voltage rating is the line to line circuit
voltage in open delta as shown in Figure 2-20. Two ±10% regulators will provide ±10% regulation on all three phases as shown in the
phasor diagram of Figure 2-21.
 Page 59

Figure 2-20.
Two Single-Phase Regulators Connected in Open Delta on
Three-Phase, Three-Wire Circuit
 Page 60

Figure 2-21.
Phasor Diagram Showing Two Single-Phase Regulators
Connected in Open-Delta
It is a relatively simple matter to determine the kVA rating of a regulator if the necessary ground rules are kept in mind.
1. The kVA rating of a single-phase feeder-voltage regulator is the product of the rated load amperes and the rated range or regulation
in kilovolts.
2. The kVA rating of a three-phase feeder-voltage regulator is the product of the rated load amperes and the rated range of regulation in
kilovolts multiplied by 1.732.
3. The rated range of regulation of a regulator is the amount that the regulator will raise or lower its rated voltage.
When regulators are being considered for a given circuit, several pertinent facts must be known. These facts are the type of circuit
(single-phase or three-phase), the voltage rating of the circuit, the kVA rating of the circuit, and the required amount of voltage
correction. In general, the required correction will fall within the range for which standard regulators are designed. Although the basic
rules apply for any range, the standard range of plus or minus ten percent will be used for the following examples. From this
information, we can determine the rated
 Page 61
current to complete the information necessary to calculate the regulator kVA. Let us take several examples to show how the rules work.
Example 1. Three-Phase, Three-Wire or Four-Wire Circuit, with Three-Phase Regulator as Shown in Figure 2-18.
Assume: 7960/13,800Y-circuit, rated 4000 kVA for ±10% range
Rated Load Amperes =

Range in Kilovolts =
Range × kVL-L = 0.10 × 13.8 = 1.38
Regulator kVA =
167 × 1.38 × 1.732 = 400 kVA
Therefore, the next highest standard rating available will be a three-phase regulator rated 500 kVA, 13,800 volts, 209 amperes, 10%
raise and lower.
Example 2. Three-Phase, Four-Wire Circuit, with Three Single-Phase Regulators as Shown in Figure 2-19.
(This is the same as a single-phase regulator on a single-phase circuit.)
Assume: 7960/13,800Y-circuit, rated 4000 kVA with ±10% range
Single-phase regulators would be excited from line to neutral at 7960 volts with the rating calculated as follows:
Load Amperes =
167 (Same as Example 1)
 Page 62
Range in Kilovolts =
0.10 × 7.96 = .796 kV
Regulator kVA =
167 × .796 = 133 kVA
Each single-phase regulator should be rated 167 kVA, 7960 volts, and 209 amperes, 10% raise and lower
Example 3. Three-Wire Delta Circuit, with Two Single-Phase Regulators as shown in Figure 2-20.
Assume: 13,800 volt circuit, rated 4000 kVA with ±10% range
Load Amperes =
167 (Same as Example 1)
Range in Kilovolts =
0.10 × 13.8 = 138
Regulator kVA =
167 × 1.38 = 230 kVA
Each single-phase regulator would be rated 276 kVA, 13,800 volts, 200 amperes, 10% raise and lower as a standard unit of the next
highest rating to meet application requirements.
For feeder-voltage regulators in general, the actual kVA of regulation required bears the same relation to the circuit kVA as the
regulating range bears to the voltage. For example, as the examples show in Case 1 and 2, a 150 kVA three-phase regulator would
provide a range of 10 percent in a 1500 kVA feeder. Likewise, a 50 kVA single-phase regulator would provide a range of 10 percent in
a 500 kVA single-phase circuit. This is a useful rule-of-thumb that can be used for all normal feeder-regulator application. With the
regulator kVA requirement known, it is a simple matter to pick a standard rating equal to requirement or the next higher rating.
 Page 63

Questions
1. The impedance given on the nameplate of a transformer does not include the magnetizing impedance. Explain.
2. Explain the difference between the load loss and the no load loss of a transformer.
3. Why can an autotransformer be so efficient at carrying load?
4. Calculate the line-to-ground short circuit level for a substation autotransformer rated 20 MVA, 13.8 kV and having a 2% impedance.
Name one problem with using this autotransformer.
5. How are network cable faults cleared? And describe the function of a network protector.
6. Why are dry type transformers more prone to failure?
7. What is the major objection to the use of a wye-delta distribution transformer connection? Explain.
8. Which transformer connections produce high third harmonic voltages?
9. What is the normal range of a regulator?
10. Determine the three-phase regulator rating to be used on a four-wire distribution system, rated 19.9/34.5 kV, having a ±10% range,
and serving a load of 5 MVA.
 Page 64

3
Application of Capacitors for Distribution Systems
Fundamentals
Early distribution systems consisted of loads that were primarily lighting. Since lighting is a resistive type of load, the power factor of
these systems was inherently quite high. Today, however, much of the utilities' load is motor load, which contains reactive power
requirements. This low power factor reactive load can cause voltage drop and losses which are largely unnecessary.
With the higher cost of losses resulting from high prices for fossil fuels, utilities have found it beneficial to correct system power factor
to almost unity. Also, with the massive mandatory PCB replacement programs placed upon the utilities and sophisticated computer
programs becoming available, new concepts to maximize the effectiveness of capacitor placement are commonplace (see Figure 3-1).
Real and Reactive Power
The system shown in Figure 3-2 consists of an ac source connected directly to a lighting load. The effective voltage and current are
designated by E and I. In a resistive circuit of this type, the voltage and current would be in phase. A wattmeter connected to the line
would give a reading of P = E × I in watts (real or active power).
To get a better idea of the in phase interaction of E and I in a resistive circuit, the sinusoidal curves for E, I, and P are shown in Figure
3-3 below. The power factor in this example is unity (cos q = cos 0° = 1).
Since E and I are effective values, the peak values are the and the The power wave is a double frequency (120 HZ) flow
consisting of all positive loops where average P = PEAK POWER/2 This indicates that power is positive and always
 Page 65

Figure 3-1.
Typical Capacitor Bank Design

Figure 3-2.
Phase Relationship of Resistive Circuit
flowing from the source into the resistor. This is one of the basic properties of active (or real) power.
On the other hand, the circuit below in Figure 3-4 illustrates the concept of reactive power. This circuit is comprised of an ac source
feeding an inductance such as a motor load.
 Page 66

Figure 3-3.
Power in a Resistive Circuit

Figure 3-4.
Phase Relationship in an Inductive Circuit
In the case of an inductance, the current I lags the voltage by 90° (see Figure 3-5). The plot of the sinusoidal waveforms shown below
illustrates that the power P is positive for half a cycle and then negative for half a cycle. This means that power is oscillating back and
forth between the generator (or capacitor) and the inductor and that average real power is zero. This type of power is defined as
reactive power to distinguish it from the unidirectional active or real power previously discussed. The reactive power, given by the
product EI, is measured in vars.
 Page 67

Figure 3-5.
Power in a Totally Inductive Circuit
In a more realistic case, as shown in Figure 3-6, we will have both real and reactive power. Although the power is changing from a
positive value to a negative value, we can see that the average power is positive, indicating a power absorbed by a resistive load. Power
is defined as follows:
Real Power = P = E · I cos q
(where E and I are given in RMS)
Reactive Power = Q = E · I sin q 3-1
Total Power = Complex Power = S = P + jQ
i.e., S = VI cos q + jVI sin q
Relationship Between KVAR and Power Factor
The total power, given as KVA, delivered by the distribution system consists of both real (KW) and reactive power (KVAR). Reactive
power, while it does no useful work, must still be supplied. The vector diagram commonly employed to express the relationship of real
and reactive power is shown below in Figure 3-7.
 Page 68

Figure 3-6.
General Network

Figure 3-7.
Real and Reactive Power Relationship
 Page 69
Where:
KVA2 = KW2 + KVAR2
KW = KVA cos q
KVAR = KVA sin q 3-2
cos q = Power factor of the load
Figure 3-8 shows examples of the change in kVAR for a 1000 kVA load as the power factor changes from .90 to .80.

Figure 3-8.
Examples of Effect of Power Factor
At 90% power factor, the kVAR supplied is approximately half of the kW. When the p.f. is reduced 10%, to .80, the kVAR becomes
75% of the kW component. If the power factor is reduced by another 10%, as shown in Figure 3-9 below, the kVAR increases to 100%
of the kW component.
 Page 70

Figure 3-9.
Example of Low Power Factor
An observation that can be made from these diagrams is that to change the power factor from .70 to .80 requires 107 kVARs while to
change the power factor from .80 to .90 requires 160 kVARs. Likewise, it would take even another 440 kVARs to bring the load to
unity power factor. The conclusion to be gained is that it takes more kVARs (a capacitor) to change a high power factor system than a
low power factor system. This explains why it is almost always economical to add power factor correction on systems with power
factors of .85 or even .90 is justified since the kVAR requirement is so much higher. It should also be noted that if the load is constant
kW, then the effect of power factor correction is to reduce total kVA load. For example, if the 1000 kVA load originally had a power
factor of .9, then the addition of capacitors to correct to unity power factor (440 kVAR) would result in a total kVA load of 900 kVA.
Examples
Exercise 1. As power factor on a 100 kVA load is lowered from .85 to .65 or approximately 24%, what percent change is there in
kVAR?
Solution 1.

a. p.f. = .85
b. p.f. = .65
 Page 71

Figure 3-10.
Example of Effect of Power Factor
p.f. = .85 = q = 32°
p.f. = .65 = q = 49.5°

a.kW load = 100 × .85 = 85 kW

kVAR load = 100 × .53 = 53
kVAR
b.kW load = 100 × .65 kW = 65
kW
kVAR load = 100 × .76 = 76
KVA

\ kVAR increased twice as fast as p.f. decreases

Exercise 2.
a. How many kVARS of capacitors do I need to change the power factor from .8 to .85 on a 100 kVA load?
b. How many kVARS of capacitors do I need to change the power factor from .9 to .95 on a 100 kVA load?
c. What conclusions can you see?
 Page 72
Solution 2.

a.100 kVA at .8 => kW = 100 × .8 = 80 kW

kVAR = 100 × .6 = 60 kVAR
100 kVA at .85 => kW = 100 × .85 = 85 kW
kVAR = 100 × .5267 = 52.7 kVAR
Needed kVAR = 60 - 52.7 = 7.3 kVAR
b.100 kVA at .9 => kW = 100 × .9 = 90 kW
kVAR = 100 × .436 = 43.6 kVAR
100 kVA at .95 => kW = 100 × .95 = 95 kW
kVAR = 100 × .312 = 31.2 kVAR
Needed kVAR = 43.6 - 31.2 = 12.4
c.Need more kVAR to change high power factor than low
power factor.
kW goes up the same.

Exercise 3. The combination of active and reactive power is sometimes referred to as complex power, defined as follows:
S = P + jQ
= VI cos q + jVI sin q 3-3
where apparent power is defined as the absolute value of S, i.e., S = VI. Complex power S, can more conveniently be defined as
follows:

3-4
*implies the conjugate of a number, i.e., the same magnitude but an angle equal to an opposite.
Find the total complex power of the system and for each element in the system shown in Figure 3-11.
Suppose that, in Figure 3-11, the net impedance seen by the source E is
 Page 73

3-5

Figure 3-11.
Power in a Typical Distribution Circuit
I, the total current, is then

3-6
S, the total complex power, is

3-7
cos q1, the power factor seen by the source, is
cos q = cos 10.3° = 0.98 3-8
V, the voltage at the shunt load node, is
 Page 74

3-9
S2, the power consumed by X2, is

3-10
i.e., the var consumption turning out negative means that vars are generated by the capacitive element.
S3, the power consumed by X3, is

3-11
i.e., the reactive element is consuming vars. Note that while this var generation and consumption convention is widely accepted, an
alternative convention would result from defining S as

3-12
The concept of reactive losses in series elements is important too. In X1,
S = P + jQ = |I1|2 (jX1) = 0 + j15.8 vars 3-13
a purely reactive loss.
 Page 75
In R1,
S = P + jQ = |I|2 (R1) = 7.92 + j0 watts 3-14
a purely resistive loss.
The total losses for the circuit are then
P Q
R1 7.9 0
X1 0 15.8
X2 0 -39.8
X3 0 +39.8
R4 79.2 0
TOTAL 87.1 15.8

Note that the total agrees with the original solution for S, total complex power supplied to the circuit.
Exercise 4. What happens if we disconnect capacitor X2 from the circuit shown in Exercise 3?
Now,

3-15
I, the total current, is then
 Page 76

3-16
S, the total complex power is,

3-17
and V1, the voltage across the load, is

3-18
Some of the interesting observations which can be made and give some insight into the value of capacitor banks are as follows:
a. Current goes up after the capacitors are removed.
b. Watts or real power consumption goes down.
c. VAR consumption goes up.
d. Voltage to the load goes down.
Voltage Rise and Voltage Drop
Early distribution lines were generally voltage limited. This means that because the lines were usually long and the primary voltage
distribution low, the voltage drop became so great as to curtail further load additions to the line. Most of these lines relied heavily on
the use of voltage regulators to provide customer voltage within the prescribed limits.
 Page 77
Today, most distribution lines are thermally limited, i.e., the maximum current the line can carry without burning down is reached
before voltage drop became a problem. This is possible because modern distribution systems generally have shorter feeders and higher
voltages (15, 25, and 35 kV). Voltage drop on these lines certainly still exists, but with most utilities extensive use of capacitors limits
line loss. Many utilities have found that the judicious sizing and location of these capacitors will provide not only loss reduction but
also all the voltage control they need without the use of feeder regulators.
Voltage Drop Defined
When the electrical characteristics (R&X) of the line have been determined and the power factor of the load is known it is very
straightforward to calculate voltage drop. Let's take a look at the system shown in Figure 3-12.

Figure 3-12.
Voltage Drop in a Distribution Line
If we assume that the load is reactive, such as a motor, we would expect that current would lag the voltage.
The voltage at the sending end, ES, can be calculated as follows:
 Page 78

Figure 3-13.
Phase Relationship for an Inductive Load

3-19

Figure 3-14.
Vector Diagram for Voltage Drop
 Page 79
Actual voltage drop is equal to Es - ER. If we project ES to the ER axis, we see that the true voltage drop is almost equal to IR.R+Ix.X
and the out-of-phase voltage drop components (i.e., -jIxR and jIR.X) have virtually no effect on the total. For this reason, the following
equation is valid for most applications:
Voltage drop = I(R cos q + X sin q) 3-20
This formula gives the voltage drop on one conductor, line-to-neutral. The three-phase line-to-line drop is times the above value,
and the single-phase drop is twice the above value.
The addition of a capacitor to an inductive system decreases voltage drop since the in-phase drop produced across the line by the
capacitor is in the opposite direction to the voltage drop produced by load across the resistance and reactance.

Figure 3-15.
Vector Diagram for Capacitor Effect
 Page 80
Voltage Rise
Lagging, or reactive current, usually causes a greater voltage drop than an increase in active current. This is primarily due to the fact
that the reactance of a line is generally greater than the resistance.
Figure 3-16 illustrates graphically the relative effects of the active and the reactive components of the current on voltage. From this
figure, it is seen that over a five mile length of typical No. 4/0, 4160 voltage distribution circuit, 200 active amperes cause a voltage
drop along the circuit of 11.5% of the total voltage at the sending end. On the other hand, 200 reactive amperes lagging will produce a
drop of 28.6% of the total voltage at the sending end. Therefore, for this typical circuit, the destructive effect of lagging reactive
amperes on useful voltage is 2.5 times that of the active current.

Figure 3-16.
Comparative Voltage Drops Caused by
Active and Reactive Currents
 Page 81
Shunt capacitors, as explained earlier, are utilized to counteract this voltage drop by producing a voltage rise across the line. Shunt
capacitors applied to a feeder give a voltage rise which is fixed (for all practical purposes), and is independent of the magnitude or
power factor of the load. The formula for this rise is given as:

3-21
Where: CKVA = the 3q rating of the capacitor bank
X = the line reactance per mile
d = distance from substation
kV = line-to-line voltage
If we have a 34.5 kV system with a 900 kVAR bank of capacitors, for example, we would calculate the rise as follows:

3-22
Switched vs. Fixed Capacitors
If we have a distribution system with no load connected to it (and we neglect magnetizing and charging currents), we will have no
current and consequently no voltage drop. If we modify this system, as shown in Figure 3-17, by placing a shunt capacitor on it, we
have now created a distribution system with a leading power factor and a voltage rise.
Using the formula for voltage drop, we can verify that the voltage at the capacitor bank is actually higher than the source voltage ES.
 Page 82

Figure 3-17.
Effect of Leading Power Factor
ER = ES - Voltage Drop
V.D. = I(R cos q + X sin q)
q = -90° => cos (-90°) = 0 3-23
sin (-90°) = -1
ER = ES -I (-X) = ES + IX
If the IX calculated is large enough, we might have a voltage at the end of the system which is out of standards and too high at the
customer service entrance. This is one of the reasons that the distribution system should not allow the power factor to exceed unity.
On a typical distribution system, the load varies as a function of the time of day. Since the load at certain hours is very low, for
example, in the early hours of the morning and very high at other times (e.g., 9 AM and 5 PM), it is obvious that voltage drop is
different as are the VAR requirements.
Figure 3-18 shows the kW and MVAR requirements as a function of time for one particular utility.
The question which should be asked is: If the VAR requirements are always changing, how many kVARs should I put on my system to
maximize my investment and not create a voltage problem? The optimum solution would be to have a VAR supply system that would
track the VAR demand exactly. Although this is possible to do, it is much too expensive. What many utilities have done in an effort to
compensate for the changing VAR requirements is to utilize both fixed and switched
 Page 83
capacitor banks. For the load cycle shown in the previous figure, we might put in a fixed bank of capacitor of 300 kVAR to
compensate for light load and a switched bank of another 300 or 400 kVAR to compensate for heavy loading conditions. It should be
noted that even with this scheme, it is possible to have a leading power factor for short periods during the day.

Figure 3-18.
Daily kW and KVAR Load Curves
This is generally acceptable because the net voltage rise is usually not a problem during the peak loading hours.
 Page 84
Types of Capacitor Control
Once the system kVAR loading requirements in respect to time are known, it can be determined what amount of capacitive kVARs
should be installed as fixed banks to compensate for light load periods and what amount of kVAR loading can be balanced by
switching banks to compensate for peak load periods. What is attempted is to ideally obtain close to unity power factor during all
periods of time by adding capacitive kVAR to the system when required and to remove the capacitive kVARs from the system when
not required so as to prevent leading kVARs to be seen by the generator, which may cause overexcitation and possible instability. The
more modern generators are much less susceptible to this problem.
Reference to Figure 3-19 below shows adding capacitors to a feeder, or system, can reduce the lagging load kVARs which the supply
must deliver. In this case, source kVARs are never leading and during peak load periods the source is delivering approximately one-
half of the kVARs it would have had to supply if no capacitors had been added. It should be noted that the off-setting of each control is
lower than the reduction in kVARs obtained when the bank it is controlling is switched on. This is necessary in order to prevent
hunting or pumping. This is characteristic of controls that sense electrical conditions, such as voltage, current or VAR. Controls
operated by time or temperature are not subjected to this problem.
Where the capacitors are to be located on the system is determined by maximum loss reduction. How the capacitors are to be controlled
is, of course, dependent on the cause and nature of the load.
Switched capacitors installed on distribution feeders and laterals usually have less than 1800 kVAR ratings, 300 to 1200 kVAR being
the most popular. One purpose is to provide voltage regulation. Higher levels of voltage can be maintained with resultant higher
revenues. Lower voltages can be maintained if fuel conservation is of prime concern due to lower energy consumption by lighting
loads. In all cases, most public utility commissions limit the level of voltage supplied to customers to ±5%, based on 120 volts as
normal (114126 volts). Most utilities try to limit the variation in voltage to less than 8%, some to less than 6%, while still remaining
within the limits set by the PUC.
The change in voltage due to switching capacitor banks on and off the lines produces a ripple effect both upstream and downstream
from where the bank is located. Therefore, all controls in the area of a bank
 Page 85

Figure 3-19.
Effect of Capacitor KVAR's on Source Requirement
being switched will see that on and off settings are adjusted to take this ripple effect into consideration.
When it becomes difficult or impossible to set the operating points on a voltage control because of the voltage changes due to other
banks being switched, other means may have to be used such as using current sensing controls or time switches, etc. These will also
cause the voltage variation to be less. Reference to the figure below shows the effect of switching capacitors on a feeder, by whatever
means, and the resultant decrease in voltage variation. Figure 3-20 portrays the voltage profile at one point on the daily load curve, say,
at peak load, of a fictitious feeder.
Maximum permissible voltage rise due to switching a bank on varies from 2% to 5%, depending on the type of circuit and number of
switching operations per day. The total rise at the instant the bank is switched on is dependent on the circuit impedance all the way
back to the source and the decrease in line current due to the reactive current supplied by the bank.
 Page 86

Figure 3-20.
Feeder Voltage Profile at One Point on Daily Load Curve
There are several popular methods of controlling switched banks of capacitors. Each method has advantages and disadvantages and
many utilities use combinations of these basic methods to optimize capacitor performance. The standard methods are:
1. Voltage - this method is relatively inexpensive and works well where voltage varies with load. On short feeders where voltage drop
is not great or on feeders with regulators, this method may be difficult to coordinate. It is many times used to supplement other
methods for emergency conditions.
2. Current - this method responds to loading well and works well for systems having induction voltage regulators. It does, however,
require a current sensing device (expensive) and cannot differentiate between low power factor summer loads and high power factor
winter loads.
3. Kilovar - this method is most effective for minimizing IX losses and can also differentiate between summer and winter peak loading
conditions. It is, however, very expensive and provides only fair voltage control when used out on the system.
4. Time - this method is simple and inexpensive. It does not sense abnormal loads and can often get out of sync due to extended power
outages, holidays, etc.
 Page 87
5. Temperature - the advantage of this method is that it's simple and low cost. On systems where VAR load varies with temperature (air
conditioning) it works well. It does not, however, sense holidays and often requires additional controls to maintain acceptable voltage
levels.
Exercise 5. If voltage is defined as VD = I (R cos + X sin), what happens to the voltage when a 5 mile piece of conductor with a loop
impedance of .053 + J.095 ohms/1000 ft. sees the load changed from 200 amperes real to 200 amperes lag?
Solution 5.
R = .053 × 5.28 × 5 = 1.3992
X = .095 × 5.28 × 5 = 2.508 3-24
a. 200 Amps Active
VD = 200 (1.3992 × 1) = 280 Volts 3-25
b. 200 Amps Reactive
VD = 200 (2.508) = 502 Volts 3-26
Conclusion. The detrimental effect of lagging amperes is greater than real amperes. On a 4160 system, this is equivalent to:

3-27
 Page 88

Losses
Line Loss Defined
Over the years since the oil embargo, the subject of losses has received a great deal of attention. When the cost of electrical energy was
low, utilities did not concern themselves as much with the line loss (and transformers) of their system. The result of this was that many
distribution feeders were very long, relatively low voltage (5 kV) and were usually voltage drop limited. Today, with higher density
loadings, it is necessary to have higher voltages (15, 25, 35 kV) and shorter lines resulting in lines that are now thermally restricted.
Since losses are a function of the square of the current and are very costly, the application of capacitors to reduce these losses has
become imperative.
The peak of kW losses L are caused by the current I squared, and flowing through the resistance R, of the line.

3-28

where I = I cos q + jI sin q or

3-29
if we define L'as the real (in phase component) =R(I cos q)2
and L'' as the reactive (out of phase component) =R(I sin q)2
We can now determine the % load losses due to kVARS = L''/L.

3-30
 Page 89

3-31

3-32
A plot of this relationship is shown below:

Figure 3-21.
Losses Due to KVARS
 Page 90
As mentioned earlier, the most important formula to determine loses is L = I2R/1000 (in kW). Consequently, to reduce losses, you
must reduce the R (loop resistance 2 times conductor resistance) or reduce the absolute value of I. Since I is equal to I cos q + I sin q
we can see that the reduction of the reactive component of current has a direct effect on real power loss. Said another way, the higher
the power factor the lower the magnitude I.
For example, if we had a 10 mile line with a loop impedance of .3 + j.7 ohms/mile with a load located at the end of the line of 400
amps, the losses of the line would be calculated as follows:

3-33
If the load were evenly distributed, similar to the diagram shown in Figure 3-22, we would have much less loss because the current I
decreases with distance. The equivalent losses are when the total load is placed 1/3 of the distance from the substation. Total losses are
then calculated as:

3-34

Figure 3-22.
Uniformly Distributed Load
Another typical feeder configuration is shown below. This condition is normally found when the line coming out of the substation is an
express portion of line or cable feeding a load some distance away. In this instance the total loss is calculated as follows: Assume
 Page 91
L1 = 5 miles
and
L2 = 5 miles

3-35

Figure 3-23.
Uniformly Distributed Load Over Part of Line
In the previous example we did not assume a power factor because only the magnitude of the I determines the losses. However, if we
made the assumption that our 400 amperes had a power factor of .85 and we could use a capacitor to compensate for the power factor
we could see the value of this investment. For an uncompensated 400 amperes load at the end of our line the losses would be
 Page 92

3-36
If the load had a .85 p.f., the current is now comprised of both real and reactive components.
i.e., IREAL = 400 cos q = 400 × (.84) = 340 amperes
IREACTIVE = 400 sin q = 400 × (.52) = 208 amperes 3-37
Since our capacitors can eliminate all of the reactive component (i.e., the 208 amps) we are left only with the real component or the
340 amperes. The total loss would then be

3-38
Consequently, the reduction in losses due to the correction of power factor is 480-347 or 133 kilowatts.
Optimal Capacitor Placement
The placement of the capacitor on the distribution system is very, very
important. For example, a capacitor placed at the substation has no effect on
line losses since the reactive kVA must still be sent all the way from the
substation to the load as shown below:
 Page 93

Figure 3-24.
Substation Capacitors
For the condition shown, the optimum placement of the capacitor bank is obviously at the load. But where should the capacitors be
placed if the load is distributed? Suppose we assume a distribution line which has only resistance R and an evenly distributed totally
inductive load of I modeled at 4 locations (see Figure 3-25). We can easily calculate the I2R losses of this circuit as follows:

Figure 3-25.
Distributed Load (l)
 Page 94

3-39
If we look at the effectiveness of capacitor compensation, we see that it is very important where we place the bank of capacitors. For
example, let's assume that we place a capacitor bank 3/4 of the total load (I) at the end of the feeder:

Figure 3-26.
Effect of Capacitor at End of Line
The total loss =
O2R + (1/4 I)2R + (1/2 I)2R = (1/16 + 1/4)I2R
= 5/16 I2R 3-40
which is over a 50% reduction in losses.
If we move the capacitor bank, one section closer to the source we have the following situation:
 Page 95

Figure 3-27.
Optimum Location
The losses are:
0 + (1/4 I)2R + (1/4 I)2R = 1/8 I2R 3-41
or less than 20% of the original system losses for the same amount of capacitors. This example points out why it is so important for a
utility to consider where the capacitors are placed. It also shows, when taken to its limit (for a distributed load), the validity of the 2/3's
rule, i.e., to minimize losses we should place 2/3 of the needed kVAR 2/3 of the way down the feeder (see Figure 3-28). Any deviance
from this will only increase losses. Try it!

Figure 3-28
 Page 96
Load Factor and Loss Factor
Load factor is defined as the ratio of Average KW Demand divided by Peak KW Demand (Average/Peak). It is understandable that
there should be a relationship between system load factor which most utilities know and line losses, which are much harder to
calculate.
Line losses, which are the sum of the I2R, or resistance losses, can be determined when the currents at peak load are known. Since the
current in an electrical system varies with respect to time, there is no precise method to calculate losses over a period of time.
However, it is done by multiplying peak loss by loss factor. Loss factor is usually defined as the ratio of the average power loss over a
designated period of time, to the maximum loss occurring in that period. Definitions of loss factor and load factor are quite similar and
care should be taken that the matter is not used in place of loss factor when considering system losses. There is a relationship between
the two factors and the empirical relation at the distribution transformer is given by
LsF = 0.15 LdF + 0.85 LdF2 3-42
Suppose we calculate our peak kW line losses as 180 kW. If the system had a load factor of .7, then the average kW losses could be
calculated as follows:
180 × [.15(.7) + .85(.7)2] = 180(.5215) = 94 kW 3-43
Exercise 6. An 8 mile, 34.5 kV utility line with an impedance of .2 + j.6 ohm/mile serves an industrial customer with a peak load of 20
MVA at .8 PF and a .6 load factor. The customer has just purchased a 3 MVAR bank of capacitors and insists he be given full credit on
his annual bill for the returned line losses on the utility system. How much could he legitimately be entitled to at an energy cost of $.08
per kW hr?
Solution 6.

3-44
 Page 97
Loss Factor = .15(.6) + .85(.6)2
Avg. Loss = I2R × Loss Factor
= (334)2 × (8 × .2) × .396 3-45
= 70.682 kW per phase
Total Annual Losses =
3 × 8760 × 70.682 = 1,857,523 kwhr
Cost/Year = $148,602/yr 3-46
Losses with Capacitor Bank =
20 MVA at .8 PF = 16 MW
12 MVAR
with Capacitor Load = 16 MW and 9 MVAR
= New Load is 18.36 MVA

3-47

Avg. Loss = (307.25)2 (8 × .2) × .396

= 59.81 kW per phase
Total Annual Cost of Losses =
3 × 8760 × 59.81 × .08 = $125,745 3-48
 Page 98
Cost savings:
148,602
-125,345
$ 22,857
\ You could reduce his annual bill by $22,857 per year
Exercise 7. A distribution line 10 miles long has an impedance of .3 + j.7 ohms/mile. Other characteristics of this line are:
Load factor = 6
Load is distributed
Cost of energy is $.05/kWHR
Peak load = 400 amperes
Power factor = .8
How much can a utility save in energy costs by correcting the p.f. from .8 to unity?
Solution 7.
a. Calculate loss factor:
= .15 Load Factor + .85 Load Factor2
= .15 (.6) + .85 (.36) = .4 3-49
b. Calculate real and reactive amps:
Real amps = 400 × .8 = 320
Reactive amps = 400 × .6 = 240 3-50
c. Calculate average losses for present system:
 Page 99
= I2R = 4002 × .3 × .4 × 10
= 192,000 watts 3-51
= 192 kW*
* 3 phases × 1/3 factor for evenly distributed load = 1
d. Calculate I at unity power factor:
I at P.F. = 1.0 = 400 × .8
= 320 amperes real and 0 reactive 3-52
e. Calculate average new loss:
I2R = 3202 × .3 × 10 × .4 = 123 kW 3-53
f. Calculate cost of losses:
= (192 - 123) × 8760 × .05 = $30,222 3-54
Questions
1. Why does imaginary power do no work?
2. A load having 2 megawatts at a power factor of .87 has how many kWs and kVARs?
3. Why does it cost more to increase power factor by capacitors when the power factor is higher?
4. Voltage rise due to a capacitor bank is dependent on what two things?
5. Switched capacitors are used for what two reasons? Explain what happens when power factor is too high.
 Page 100
6. Reactive power can cause more voltage drop than real power (true or false)? Explain.
7. What is the reason not to use a switched capacitor bank controlled by voltage?
8. Why does reactive load cause losses?
9. Optimum capacitor placement for an evenly distributed load occurs when half the VARs from the capacitor flow toward the source
and half flow away from the source. Is this consistent with the 2/3's rule?
10. A 900 kVAR bank of capacitors could rarely cause excessive voltage rise on a 3 mile long feeder. Explain.
 Page 101

4
Distribution Overcurrent Protection
Introduction
The overcurrent protection of the distribution system is considerably different from any other part of the utility system. Unlike the
transmission and subtransmission system, the distribution system is usually radial in design. Also, where the protection of the
transmission system is usually performed by breakers with various types of relays, the distribution system utilizes breakers, circuit
switchers, load break disconnects, fuses, overcurrent relays, reclosers and sectionalizers (see Figure 4-1).
The fact that virtually all distribution systems are radial is what allows the protection system to work properly. A radial system is
coordinated on the premise that the fault current decreases as distance from the substation increases. A profile of fault current vs.
distance is shown in Figure 4-2 and illustrates this relationship. As can be seen, a line-to-ground fault at the terminals of the substation
would create fault currents on the order of 6500 amps. Just one mile away, this fault current is reduced to less than 3000 amperes. A
fault 10 miles from the substation will draw less current than full load conditions indicating that protection of faults far out on the
feeder cannot be performed at the substation.
Characteristics of Devices
Fuses
Fuses are the most basic and cost effective type of overcurrent device presently being used by the utility industry. They are also one of
the most reliable devices in that they can provide their function for over 20 years with essentially no maintenance.
 Page 102

Figure 4-1.
Primary Distribution Feeder
 Page 103

Figure 4-2
All overcurrent protecting devices carry the more commonly used name of current interrupter. Current interrupters do not actually
interrupt current. What they actually do is setup a high dielectric environment that prevents the arc from re-establishing when the
current passes through a current zero.
Fuses use the following mechanism to interrupt the flow of current:
1. Sensing .. heat and melt
2. Arc initiation .. separation
3. Arc manipulation .. stretching, cooling, deionizing, pressurizing
4. Current interruption .. current zero.
Although a fuse is deceptively simple in appearance, its function is complex. For a fuse to function properly, it must:
 Page 104
Sense the conditions it is trying to protect
Interrupt the fault quickly
Coordinate with other protective devices.
Expulsion Fuses. A distribution expulsion fuse link consists of three basic parts which are:
Button
Fusible element
Leader

Figure 4-3.
Fuse Link
Length and diameter of the element are the main determinants of the fuse characteristics. The longer the fuse, the faster it will operate
for low level short circuits. During high levels of short circuit, the temperature rise is very fast and heat cannot be conductored away
from the center of the fuses element. Consequently, melting time at the higher current range is not dependent so much on the length of
the element but rather by the diameter.
The expulsion fuse, as the name implies, expulses gases during its operation. Figure 4-4 shows a diagram of a typical expulsion fuse
cutout. The fuse itself is housed inside the fuse cartridge.
 Page 105

Figure 4-4.
A Distribution Expulsion Circuit
The typical expulsion fuse will utilize a relatively short fusible element section to sense the overcurrent and start the arcing required for
interruption. Attached to this short fusible element will be a larger type of conductor, commonly called a fuse leader or fuse link tail,
which then connects to the rest of the fuse hardware as required. During a fault, the fuse element will melt causing an arc inside the
fuse cartridge. When the arc is produced it will rapidly create gases from special materials (usually fiber) located in close proximity to
the fuse element. The primary function of the gases that are generated by the fiber is to deionize and remove arc generated ionized
gases and allow a rapid buildup of dielectric strength that can withstand the transient recovery voltage and steady state power system
voltage. Of course, consideration must be given to the fact that these gases will be produced and will be exhausting from the fuse
assembly. Some types, such as the boric acid fuses, can utilize condensers to prevent much exhaust beyond the fuse assembly.
One of the main advantages with typical expulsion fuses is that they can be reloaded with a relatively economical fuse link.
Furthermore, there are a wide variety of fuse link types and sizes that could be used in the
 Page 106
same fuse holder. This allows common usage of the same holder for a number of applications and also permits a wide latitude of
coordination possibilities.
EEI-NEMA standards divide expulsion fuse links into two types - fast and slow designated by K and T, respectively. K and T links of
the same rating have identical 300 second point or 600 second points. But as shown in Figure 4-5, they have different T/C curves, the T
link being slower at the high-current end than the same size K link.

Figure 4-5.
Minimum Melting Time Curves for Fast and Slow Links
of the Same Rating
 Page 107
The distinction between the two types is the speed ratio, which is the ratio between melting currents at 0.1 second and 300 seconds for
links rated through 100 amperes, and at 0.1 second and 600 seconds for links rated over 100 amperes. For example, a T link rated at 6
amperes has a 0.1 second melting current of 130 amperes and a 300 second melting current of 12 amperes, resulting in a speed ratio of
10.8. Slow links have speed ratios between 10.0 and 13.0. Fast links have speed ratios between 6.0 and 8.1.
To meet special requirements such as primary fusing of small-sized transformers, links rated below 10 amperes have been developed.
High-surge links rated 1, 2, 3, 5, and 8 amperes are in this category. They are specifically designed to provide overload protection and
avoid unnecessary operation during short-time transient current surges associated with motor starting and lightning.
Current Limiting Fuses. The two key components in the construction of a current limiting fuse are the fusible element, which is
normally of pure silver and quartz sand. Thus, the current limiting fuse is sometimes called a silver-sand fuse.

Figure 4-6.
Current Limiting Fuse
Silver is used in the current limiting fuse because it has the best balance of properties to produce the highest ampere rating for a
minimum physical size and also has the most desirable melting characteristic. The silver element will not become damaged under
overcurrent conditions until very close to its high melting temperature of 960°C. When it reaches this melting temperature, only a small
amount of additional heat is required to completely melt and vaporize the metal. This characteristic
 Page 108
is desirable to minimize the total energy released in the fuse during an interruption. In addition, silver has excellent current carrying
ability and immunity to oxidation.
The silica sand functions to absorb the heat given off by the fusible elements without becoming electrically conductive. Though this
heat absorbing function is basically simple, the presence of the sand is critical to the correct operation of the fuse. Since the
temperature of the arc approaches that of the sun, few materials can perform this heat absorbing function successfully. Even seemingly
incidental considerations such as the size of the sand granules and the method of packing can have significant influence on the
operating characteristics of the fuse.
More specifically, a current limiting fuse consists of one or more silver wire or ribbon elements (one manufacturer uses cadmium)
suspended in an envelope filled with sand. These silver elements are spirally wound on a temperature-resistant, non-tracking form or
core. By spirally winding the silver elements, the length of the fuse can be kept to a minimum. The core and element assembly is
mounted in a tube of temperature-resistant material and then filled with the high-purity silica sand and sealed. The degree of sealing
may vary with the application. For example, fuses designed for direct oil submersion must be carefully sealed to prevent the ingress of
oil at temperatures as high as 150°C. Fuses for outdoor application must be capable of withstanding the outdoor environments and be
sealed against possible entry of moisture over the service life, which is generally 20 years or longer.
Both wires and ribbon-type fusible elements or combinations of the two are used in the construction of current-limiting fuses. When
silver ribbons are used, each ribbon has small holes punched along its length forming a series of notches. These notches control heat
distribution and help shape the time-current characteristics. In normal operation, the I2R heat generated flows into the relatively large
and cool body of the element, then into the surrounding sand. When a fault occurs, heat is generated so rapidly that almost none can
escape. Instead, it remains in the elements and melts the silver elements.
On high-fault currents, the element melts instantaneously along its entire length to interrupt the current. Upon interruption, the heat of
the arc is transferred to the sand, transforming it into a crystalline structure known as a fulgurate. This transformation results in the
sudden insertion of additional impedance and the subsequent development of an arc voltage. When the arc voltage exceeds the system
(driving) voltage, the arc is extinguished and current-limiting action is accomplished. In the
 Page 109
figure below, note the magnitude of the developed arc voltage. Since it exceeds the system voltage, its effect on surge arresters must be
analyzed. The higher the generated voltage, the greater the current-limiting action.

Figure 4-7.
Voltage Current Relationships During Operation of a
Current-Limiting Fuse in a Zero Power Factor Circuit
Current-limiting fuses are available in two types: general purpose and backup. Both employ a silver fuse element wound on a
supporting core.
The general purpose fuse is defined by ANSI C37.40 - 1969 as follows:
A fuse capable of interrupting all currents from the rated maximum interrupting current down to the current that causes melting of the fusible element
in one hour. (25°C, ambient)
This definition usually covers those currents that are between 150 to 200% of the fuse's rating. Although this ability includes a wide
range of currents, it does not include all the possible currents that the fuse may be
 Page 110
called upon to clear. An example of this would be a fuse placed in a high ambient condition. It would be quite possible that a fuse
could be required to interrupt a current much lower than the value stated by the one hour criteria. Some examples of general purpose
fuses are:

RTE -- ELS*
McGraw -- Nx*
GE -- Surge Guard
GP
Westinghouse-- CX

The backup fuse is capable of interrupting all currents in a particular range. It is defined by ANSI C37.40 - 1969 as follows:
A fuse capable of interrupting all currents from the rated maximum interrupting current down to the rated minimum interrupting current (as specified
by the manufacturer).
Some examples of backup current limiting fuses are:

RTE -- ELSP, ELO

McGraw -- Companion (40k -
only)
GE -- OSP, ETP, EJO
S&C -- Fault Fiter
Kearney -- Type A
Westinghouse-- Type CL
Chance -- K Mate

The backup fuse is by design a fuse that clears only the high fault currents. The low current clearing is accomplished through the use of
an expulsion fuse connected in series with the current limiting backup fuse. Its construction is therefore simpler. Although the element
design is extremely important in controlling the let-through energy, no provisions are required for low-fault current clearing. When
such clearing is required, the time-current characteristics are a composite of the two fuses as shown in Figure 4-8 below, and each fuse
works in its optimum range.
 Page 111

Figure 4-8.
Time Current Characteristics for Backup Current Limiting
Fuse
Coordination. The initial operating mechanism of a fuse is the melting of the element. This mechanism depends on the following three
factors:
Magnitude of current
Duration of current
Electrical properties of the element.
It is the time current curve (TCC) which defines the characteristic of a fuse. More specifically, the fuse characteristic is defined by 2
curves, the minimum melt (mm) and the total clearing (TC).
The minimum melt curve is developed by electrical test. The magnitude of the current and the time it took the fuse to melt are recorded
and plotted. At this point, a curve is drawn through these points representing an average melt curve and from this curve, 10 percent is
subtracted and the resulting curve is called the minimum melt curve.
The fuse, however, has an arcing time associated with it. The arcing time is the time it takes the fuse to interrupt the circuit after the
fuse melts and it is also obtained by the test. The arc times, which are
 Page 112
recorded at different current magnitudes, are added to the maximum melt time (maximum melt time = 110 percent of average melt
time). The resulting curve is called the total clearing curve. The two curves (see Figure 4-9 below), i.e., minimum melt and total clear,
are the extremes of the fuse characteristics and are the curves published by each manufacturer.

Figure 4-9.
Typical Time Current Curves for a 10 K Line.
 Page 113
Relays
To some utilities, the distribution system starts at the primary of the distribution substation while other utilities classify their
distribution system from the feeder breaker to the customer. In either case, a knowledge of overcurrent relaying is necessary.
Required Characteristics. The required characteristics necessary for protective equipment to perform its function properly are:
sensitivity, selectivity and speed. This is especially true for relays.
Sensitivity: Sensitivity applies to the ability of the relay to operate reliably under the actual condition that produces the least operating
tendency. For example, a time-overcurrent relay must operate under the minimum fault current condition expected. In the normal
operation of a power system, generation is switched in and out to give the most economical power generation for different loads which
can change at various times of the day and various seasons of the year. The relay on a distribution feeder must be sensitive enough to
operate under the condition of minimum generation when a short circuit at a given point to be protected draws a minimum current
through the relay. (NOTE: On many distribution systems, the fault-current magnitude does not differ very much for minimum and
maximum generation conditions because most of the system impedance is in the transformer and lines rather than the generators
themselves.)
Selectivity: Selectivity is the ability of the relay to differentiate between those conditions for which immediate action is required and
those for which no action or a time-delayed operation is required. The relays must be able to recognize faults on their own protected
equipment and ignore, in certain cases, all faults outside their protective area. It is the purpose of the relay to be selective in the sense
that, for a given fault condition, the minimum number of devices operate to isolate the fault and interrupt service to the fewest
customers possible. An example of an inherently selective scheme is differential relaying; other types, which operate with time delay
for faults outside of the protected apparatus, are said to be relatively selective. If protective devices are of different operating
characteristics, it is especially important that selectivity be established over the full range of short circuit current magnitudes.
 Page 114
Speed: Speed is the ability of the relay to operate in the required time period. Speed is important in clearing a fault since it has a direct
bearing on the damage done by the short circuit current; thus, the ultimate goal of the protective equipment is to disconnect the faulty
equipment as quickly as possible.
Characteristics of Overcurrent Relays. The overcurrent relay is the simplest type of protective relay (see Figure 4-10). As the name
implies, the relay is designed to operate when more than a predetermined amount of current flows into a particular portion of the power
system. There are two basic forms of overcurrent relays: the instantaneous type and the time-delay type.
 Page 115

Figure 4-10.
Typical Type 51 Relay Mechanism with Standard Hinged Armature
Instantaneous Unit Withdrawn from Case
The instantaneous overcurrent relay is designed to operate with no intentional time delay when the current exceeds the relay setting.
Nonetheless, the operating time of this type of relay can vary significantly. It may be as low as 0.016 seconds or as high as 0.4 seconds.
The operating characteristic of this relay is illustrated by the instantaneous curve of Figure 4-11.
 Page 116

Figure 4-11.
Time-Current Characteristics of Overcurrent Relays
The time-overcurrent relay (IAC, CO) has an operating characteristic such that its operating time varies inversely as the current
flowing in the relay. This type of characteristic is also shown in Figure 4-11. The diagram shows the three most commonly used time-
overcurrent characteristics: inverse, very inverse, and extremely inverse. These curves differ by the rate at which relay operating time
decreases as the current increases.
Both types of overcurrent relays are inherently non-selective in that they can detect overcurrent conditions not only in their own
protected equipment but also in adjoining equipment. However, in practice, selectivity between overcurrent relays protecting different
system elements can be obtained on the basis of sensitivity (pickup) or operating time or a combination of both, depending on the
relative time-current characteristics of the particular relays involved.
These methods of achieving selectivity will be illustrated later. Directional relays may also be used with overcurrent relays to achieve
selectivity.
 Page 117
The application of overcurrent relays is generally more difficult and less permanent than that of any other type of relaying. This is
because the operation of overcurrent relays is affected by variations in short-circuit-current magnitude caused by changes in system
operation and configuration. Overcurrent relaying in one form or another has been used for relaying of all system components. It is
now used primarily on distribution systems where low cost is an important factor.
Figure 4-12 shows a family of inverse time curves of the widely used IAC relay, which is an induction disc type. The time curves for
the new design static overcurrent relays are similar.

Figure 4-12.
Inverse Time Curves
A curve is shown for each numerical setting of the time dial scale. Any intermediate curves can be obtained by interpolation since the
adjustment is continuous.
It will be noted that the curves shown in Figure 4-13 are plotted in terms of multiples of pickup value, so that the same curves can be
used for any value of pickup. This is possible with induction-type relays where the pickup adjustment is by coil taps, because the
ampere-turns at
 Page 118
pickup are the same for each tap. Therefore at a given multiple of pickup, the coil ampere-turns, and hence the torque, are the same
regardless of the tap used.

Figure 4-13.
Operating Time of Overcurrent Relays With Inverse Time
Characteristics
The time-current curves shown in Figure 4-12 can be used not only to determine how long it will take the relay to close its contacts at a
given multiple of pickup and for any time adjustment, but also how far the relay disc will travel toward the contact-closed position
within any time interval. For example, assume that the No. 5 time-dial adjustment is used and that the multiple of pickup is 3. It will
take the relay 2.45 seconds to close its contacts. We see that in 1.45 seconds the relay would close its contacts if the No. 3 time-dial
adjustment were used. In other words, in 1.45 seconds the disc travels a distance corresponding to 3.0 time-dial divisions, or three
fifths of the total distance to close the contacts. For the most effective use of an inverse-time relay characteristic, its pickup should be
chosen so that the relay will be operating on the most inverse part of its time curve over the range of values of current for which the
relay must operate. In other words, the minimum value of current for which the relay must operate should be at least 1.5 times pickup,
but not very much more.
Figure 4-13 shows the application of time-overcurrent relays to a radial feeder and the total tripping time characteristics for faults at
any location along a circuit. The figure shows the increase in the minimum tripping time as faults occur farther from the distribution
substation - an increase inherent with overcurrent relaying. It also shows the effect of the inverse-time characteristic in reducing this
increase. Obviously, the more
 Page 119
line sections there are in series, the greater is the tripping time at the source end. It is not at all unusual for this time to be as high as 2
or 3 seconds. This is not a very long time according to some standards, but it would be intolerable if system stability or line burndown
were an important consideration.
During light loads, some of the generators are usually shut down. At other times, the system may be split into several parts. In either
case, the short circuit current tends to vary with the amount of generation feeding it. This generally does not have much effect at the
distribution level. It should be appreciated that a reduction in the magnitude of short circuit current raises all of the characteristic
curves of Figure 4-13. For locations where inverse time-overcurrent relays must be mutually selective, it is generally a good policy to
use relays whose time-current curves have the same degree of inverseness. Otherwise, the problem of obtaining selectivity over wide
ranges of short-circuit current may be difficult.
Instantaneous or undelayed overcurrent relaying is used only for primary relaying to supplement inverse-time relaying and is presently
being used by most utilities. It can be used only when the current during short circuit is substantially greater than that under any other
possible condition - for example, the momentary current that accompanies the energization of certain system components. The zone of
protection of undelayed overcurrent relaying is established entirely by adjustment of sensitivity and is terminated short of the far end of
the line. For instance, the instantaneous-overcurrent relay is usually set so that its pickup is 25 percent higher than the maximum
current the relay will see for a three-phase fault at the end of the line. With this setting, the instantaneous relay will provide fault
protection for about 80 percent of the line section.
Undelayed (instantaneous) trips can frequently be added to inverse-time relaying and effect a considerable reduction in tripping time.
This is shown in Figure 4-14 where the two sets of characteristics are superimposed. The time saved through the use of the
instantaneous relays is shown by the shaded area. A reduction in the magnitude of short circuit current shortens the distance over which
the instantaneous unit operates and may even reduce this distance to zero. However, this fact is usually of no great importance since
faster tripping under the maximum short-circuit conditions is the primary objective.
 Page 120

Figure 4-14.
Reduction in Tripping Time Using Instantaneous Relaying
Instantaneous tripping is feasible only if there is a substantial increase in the magnitude of the short circuit current as the short circuit is
moved from the far end of a line toward the relay location. This increase should be at least two or three times. For this reason, it often
happens that instantaneous relaying can be used only on certain lines and not on others.
On systems where the magnitude of short circuit current flowing through any given relay is dependent mainly upon the location of the
fault to the relay, and only slightly or not at all upon the generation in service, faster clearing can usually be obtained with very-
inverse-time-overcurrent relays. Where the short circuit current magnitude is dependent largely upon system-generating capacity at the
time of the fault, better results will be obtained with relays having inverse-time operating characteristics.
However, towards the ends of primary distribution circuits, fuses are sometimes used instead of relays and breakers. In the region
where the transition occurs, it is frequently necessary to use overcurrent relays having extremely inverse characteristics to coordinate
with the fuse characteristics.
The extremely inverse relay characteristic has also been found helpful, under certain conditions, in permitting a feeder to be returned to
service after a prolonged outage. After such a feeder has been out of service for so long a period that the normal off period of all
intermittent loads (such as furnaces, refrigerators, pumps, water heaters, etc.) has been
 Page 121
exceeded, reclosing the feeder throws all of these loads on at once without the usual diversity. The total inrush current, also referred to
as cold-load pickup, may be approximately four times the normal peak-load current. This current decays very slowly and can be as
high as 1.5 times normal peak current after as much as three or four seconds. Only an extremely inverse characteristic relay provides
selectivity between this inrush and short circuit current.
Whenever service has been interrupted to a distribution feeder for 20 minutes or more, it may be extremely difficult to re-energize the
load without causing protective relays to operate. The reason for this is the flow of abnormally high inrush current resulting from the
loss of load diversity. High inrush currents are caused by:
a. Magnetizing inrush currents to transformers and motors,
b. Current to raise the temperatures of lamp filaments and heater elements, and
c. Motor-starting current.
Figure 4-15 shows the inrush current for the first five seconds to a feeder which has been de-energized for 15 minutes. The inrush
current, due to magnetizing iron and raising filament and heater elements temperatures, is very high but of such a short duration as to
be no problem. However, motor-starting currents may cause the inrush current to remain sufficiently high to initiate operation of
protective relays. The inrush current in Figure 4-15 is above 200 percent for almost two seconds.
 Page 122

Figure 4-15.
Five-, Ten-, and Fifteen-Minute Outage Pickup Curves for First Five
Seconds after Restoral
The magnitude of cold load current is closely related to load diversity, but quite difficult to determine accurately because of the
variation of load between feeders. If refrigerators and deep freeze units run five minutes out of every 20, then all diversity would be
lost on outages exceeding 20 minutes.
A feeder relay setting of 200 to 400 percent of full load is considered reasonable. However, unless precautions are taken, this setting
may be too low to prevent relay misoperation on inrush or cold load following an outage. Increasing this setting may restrict feeder
coverage or prevent a reasonable setting of fuses and relays on the source side of this relay.
A satisfactory solution to this problem is the use of the extremely inverse relay. Figure 4-16 shows three overcurrent relays which will
ride over cold-load inrush. However, the extremely inverse curve is superior in that substantially faster fault-clearing time is achieved
at the high-current levels.
 Page 123

Figure 4-16.
Comparison of Overcurrent Relay Characteristics
This figure, for the purpose of comparison, shows each characteristic with a pickup setting of 200 percent peak load and a five second
time delay at 300 percent peak load to comply with the requirements for re-energizing feeders.
It is evident that the more inverse the characteristic, the more suitable the relay is for feeder short circuit protection. The relay
operating time, and hence, the duration of the fault can be appreciably decreased by using a more inverse relay. Comparing the inverse
characteristic shows that the extremely inverse characteristic gives from 30 cycles faster operation at high currents to as much as 70
cycles faster at lower currents.
Unfortunately, the extremely inverse relay may not always take care of the problem. As the feeder load grows, the relay pickup must be
increased and a point may be reached at which the relay cannot detect all faults. At this time, it may be necessary to either move the
fuses or reclosers closer to the substation or use automatic sectionalizing.
 Page 124
Reclosers
An automatic circuit recloser is a self-contained device which can sense and interrupt fault currents as well as re-close automatically in
an attempt to re-energize a line. The operation of a recloser is very similar to that of a feeder breaker with a reclosing relay. The main
difference between the two devices is that the recloser has less interrupting capability and costs considerably less.
Recloser operation utilizes two inverse time curves. The first curve, referred to as the instantaneous or A curve (see Figure 4-17) is
similar to an instantaneous relay and is used primarily to save lateral fuses under temporary fault conditions. The second curve referred
to as the time delay or B (also C, D, or E) is used to delay recloser tripping and allow the fuse to blow under permanent fault
conditions. As can be seen in Figure 4-17, a 50 ampere recloser (hydraulic) has a minimum trip for both the instantaneous and time
delay currents of 100 amperes or twice its rating.

Figure 4-17
 Page 125
The reclosers time-delay curves (B, C, D, E) are fixed. This is in contrast to an inverse-relay curve which has an infinitely adjustable
time dial. On the other hand, a recloser may have more than one instantaneous operation in an attempt to dissipate a temporary fault
condition whereas a feeder relay can normally have only one. The most typical reclosing sequence for a recloser is two fast operations
followed by two slow (2A, 2B). This is illustrated below in Figure 4-18.

Figure 4-18.
Typical Reclosing Sequence
Major classifying features in automatic circuit reclosers are:
Single- or three-phase
Control: hydraulic or electronic
Interrupters: oil or vacuum.
Single- or Three-Phase. Single-phase reclosers are used to protect single-phase lines such as branches or taps of a three-phase. Also can
be used to isolate single-phase loads.
Three-phase reclosers are used where lockout of all three phases is required for any permanent fault. They are also used to prevent
single-phasing of three-phase loads such as large three-phase motors.
Controls: Hydraulic or Electronic. The intelligence that enables a recloser to sense overcurrents, select timing operation, time the
tripping and reclosing functions, and finally lockout, is provided by its control.
 Page 126
There are two basic types of control schemes used: an integrated hydraulic control or an electronic control located in a separate cabinet.
Hydraulic recloser control is used on all single-phase reclosers and in smaller ratings of three-phase reclosers made by McGraw (GE
did not use hydraulic controls). It is built as an integral part of the recloser. With this type of control, an overcurrent is sensed by a trip
coil that is connected in series with the line. When the overcurrent flows through the coil, a plunger is drawn into the coil to trip open
the recloser contacts. Minimum trip current for this type of recloser is two times the rating of the recloser.
The electronic control is more flexible, more easily adjusted and more accurate. The electronic control conveniently permits changing
timing, trip current levels, and sequences of recloser operations without deenergizing or untanking the recloser. Line current is sensed
by special CT's in the recloser. Minimum trip level is independent of recloser rating.
Interrupters - Oil, Vacuum, and SF6. Reclosers using oil (GE and McGraw) for current interruption use the same oil for basic
insulation. Vacuum (also supplied by GE and McGraw) provides the advantage of lower maintenance frequency. Depending on type, a
vacuum recloser may use either oil or air as the basic insulating medium. The newly introduced recloser by A.B. Chance uses SF6.
To properly apply automatic circuit reclosers, five major factors must be considered:
1. System voltage
2. Maximum available short circuit level
3. Maximum load
4. Minimum fault current (in protected zone)
5. Coordination with other devices.
Ratings of reclosers are generally considerably less than the ratings of popular feeder breakers in both continuous current and short
circuit interruption capability. Units made by GE (now Multiamp) some years ago were limited to 4000 amperes and less as shown in
Table 4-1.
 Page 127
Table 4-1. GE Recloser
Frame Size AmpskV RatingPhaseShort Circuit Interrupting
50280 14.4 1 & 3 1254000 A
100 24.9 1 & 3 2002500 A

A summary of McGraw's (now Cooper Power) capability is as follows:

Table 4-2. McGraw Recloser
Frame Size kV Rating No. of Phases Interrupting Amps
50560 2.414.4 1 12510,000
100 24.9 1 3008,000
100560 2.414.4 3 20020,000
560 24.9 3 3,00012,000
560 34.5 3 16,000

Sectionalizers
A sectionalizer is a protective device, used in conjunction with a recloser (or breaker with reclosing relay) which isolates faulted
sections of lines. The sectionalizer does not interrupt fault current. Instead, it counts the number of operations of the reclosing device
and opens while this backup device is open. After the sectionalizer opens, clearing the faulted section, the backup devices recloses to
return power to the unfaulted sections of the line. If the fault is temporary, the sectionalizer will reset itself after a prescribed period of
time.
A sectionalizer provides several advantages over fuse cutouts. In addition to application flexibility, they offer safety and convenience.
After a permanent fault, the fault-closing capability of the sectionalizer greatly simplifies the testing of the circuit, and if the fault is
still present, interruption takes place safely at the backup recloser.
Replacement fùse links are not required, thus the line can be tested and service restored with far more speed and convenience. Also, the
possibility of error in the selection of the right size and type of fuse link is eliminated.
 Page 128
Because it has no time-current characteristic, a sectionalizer has distinct application advantages.
1. It can be applied between two protective devices having operating curves which are close together. This is a vital feature in a
location where an additional step in coordination is not practical or possible.
2. It can be used on close-in taps where high fault magnitude prevents coordination with fuses.
Some of the disadvantages associated with sectionalizers are as follows:
1. Cost - The major reason utilities don't use more sectionalizers is that the advantages of the sectionalizer over a fuse are not justified
by the cost differential of over $500. Also, sectionalizers have technical problems of their own.
2. Memory Time - Hydraulically controlled and dry type sectionalizers have some problems with memory time. In a standard
sectionalizer, the reset time after a transient fault depends on the number of counts and the memory time selected. It can range from 5
to 22 minutes. Corresponding reset times for reclosers are from 10 to 180 seconds. Hydraulically controlled and dry type sectionalizers
do not provide a choice of memory times. Memory time is essentially a function of the viscosity of oil which in turn is dependent on
temperature. Consequently, accuracy predicting memory times is impossible and a too long memory time may result in
miscoordination during temporary faults (as well as to a lesser degree permanent faults). Areas with high isokeraunic levels are
particularly susceptible. The sectionalizer memory time must be sufficiently long such that the sectionalizer will retain its counts
throughout the entire tripping and reclosing sequence of the backup fault interrupter. The memory time of hydraulic and dry-type
sectionalizers vary with temperature, and this variable should be included in the calculation process. The consideration is not included
here since the process is dependent on the type and manufacturer of the individual sectionalizer.
3. Inrush - Inrush has been a very big problem for some sectionalizers. The problem has been that the sectionalizer is very fast and sees
inrush currents as fault currents. As such, it may misoperate (discussed in the next section on coordination).
 Page 129

Coordination of Devices
Fuse-to-Fuse Coordination (Expulsion)
Expulsion Fuse to Expulsion Fuse. Figure 4-19 illustrates the general principle of coordinating expulsion fuses in series. Fuse (1) is
called the protected fuse, and fuse (2) is called the protecting fuse. For perfect coordination, fuse (2) must melt and clear the fault
before fuse (1) is damaged. To ensure this, three things are done:
1. The maximum characteristic (the total clear curve) of the protecting fuse (2) is plotted to conservatively estimate the maximum
duration of the fault current.
2. The minimum (published) characteristic of the protected fuse (1) is plotted.
3. Seventy-five percent of the MM curve of the protected fuse is plotted to make sure that the fuse is not damaged and to account for
any degradation in the fuse characteristics. (This is sometimes called the damage curve.)

Figure 4-19.
Coordination of K Links
 Page 130
If the damage curve of fuse (1) and the TC curve of fuse (2) never cross, there is said to be perfect coordination. If they cross, however,
at some value of current, this is called the limit of coordination. For example, suppose the curves cross at 2000 amperes. This means
that coordination above 2000 amperes is unlikely. However, if the maximum available short circuit of the system at that location is
only 1500 amperes the fuses would be considered to be fully coordinated.
Current Limiting Fuse to Expulsion Fuse. The basic problem in using a full range current limiting fuses is that their normal TC
characteristics are different from most other devices making them difficult to coordinate.
A situation which commonly exists for a utility is the use of an expulsion fuse at a lateral tap and a CLF at a transformer as shown
below.

Figure 4-20
For a fault on the transformer, we want the CLF to clear the fault without damage to the 65 K fuse (protected). A plot of the fuse
characteristics using the 75 percent rule (i.e., 4 melt on 3 clear) is shown in Figure 4-21.
As can be seen, these two fuses coordinate quite well at least down to .01 seconds. As we know, however, the CLF is a current limiting
device and can in effect force a current zero which in effect means that the CLF can melt in less than .01 seconds. This energy limiting
characteristic actually protects the K link and full coordination down to 350 amperes is assured.
Another check on this ability to coordinate in less than .01 seconds can be illustrated by comparing the minimum melt (MM) I2t of the
65 K link to the total clearing (TC) I2t of the 10 ampere CLF. For example, the total maximum I2t of a 10 amp CLF is less than 4400
amp2 sec at any
 Page 131

Figure 4-21
voltage rating (per the manufacturer) whereas the minimum melt I2t of the expulsion is a calculated value at the .01 sec point, i.e.,
(3000)2(.01) = 90,000 amp2 sec. It is evident, therefore, that the CLF will always melt well before the K link and coordination up to
50,000 amperes is assured.
On the other hand, when the CLF is the protected fuse and the expulsion is the protecting fuse, coordination is limited. A lateral
protected by a CLF is shown below in Figure 4-22.

Figure 4-22
 Page 132
The reason for the limited coordination is that the expulsion fuse must wait for a current zero to interrupt. As a result, an asymmetric
current may flow in the device for up to .013 seconds. If we plot these fuses (see Figure 4-23) we can see that coordination exists up to
only 500 amperes. Above this fairly low level, it is probable both fuses would operate.

Figure 4-23
Current Limiting Fuse to Current Limiting Fuse. When a current limiting fuse is used both as the protecting and protected fuse, we
expect to see a coordination plot, now with fuses of similar time-current characteristics.
 Page 133

Figure 4-24
Since current limiting fuses are current limiting devices and can operate in less than .01 seconds, additional check is needed to make
sure coordination below .01 seconds takes place. All manufacturers of current limiting fuses determine from tests and publish
minimum melt I2t and total maximum I2t values for the purpose of coordination.
Listed below are the minimum melt as I2t and total let through I2t values for general purpose current limiting fuses rated 15.5 kV.
 Page 134
Table 4-3. CLF Characteristics
Fuse Rating I2t Min. I2t Max.
Amps Melt Total
6C 150 1280
8C 230 2500
10C 520 3200
12C 1160 9800
15C 1540 12000
20C 2690 16500
25C 4560 25000
30C 4560 16000
40C 10700 40000

For the 8C and 30C fuses in the coordination plot, we see that the MM I2t of the 30C fuse is almost twice the max total I2t of the 8C
protecting fuse. Consequently, coordination exists. Table 4-3 illustrates how a large fuse may not coordinate with a small rated fuse
even though they have similar characteristics. For example, any protecting fuse above 10 amperes will not coordinate with the 30C
fuse used in this example.
Backup Current Limiting Fuse Coordination - Two Fuse System. A backup current limiting fuse is always applied with a series-
connected expulsion fuse. The procedure mandates coordinating the time current characteristics so the expulsion fuse interrupts all
currents below the rated minimum interrupting current of the backup fuse. Two coordination choices are possible depending upon the
characteristics of the backup fuse:
Crossover coordination
Matched melt coordination
The expulsion fuse is chosen in the conventional manner based on inrush considerations and overload protection practice, if there are
no supplementary secondary protective devices. With primary fusing only, protection against overloading or low fault currents in the
region of 2 to 4 times full nameplate load current is most commonly used. When secondary breakers of fuses are involved, the primary
expulsion fuse must be coordinated with these devices.
 Page 135
Crossover Coordination. Figure 4-25 illustrates both the crossover and minimum melt I2t approaches.

Figure 4-25.
Time-Current Curves Showing the Effects of the Two Different Crossover
Regions Currently in Use
Matching the expulsion fuse with CLF A involves a crossover of characteristics in the time region between .01 and 100 seconds. This
is an approach sometimes preferred as an oil submersible protection (OSP) where the CLF is under oil in the transformer and the fuse
may be of a bayonet type. In this case it may be desirable to have the CLF operate only for very high fault currents that might be seen
in a transformer failure where the unit must be removed anyway. Secondary faults of lower magnitude would be interrupted by the
easily replaceable expulsion fuse. For successful coordination, the backup CLF must have a minimum interrupting capability which
includes the diamond shaped crossover region (plus some safety factor). In the case of fuse A in the illustration,
 Page 136
this would require a minimum current interrupting capability of approximately 700 amperes.
Matched Melt Coordination. This coordination method requires that the minimum melt I2t of the expulsion fuse be equal to or less than
the backup current limiting fuses at .01 seconds. This, in fact, means that the expulsion fuse will melt and arc for all levels of fault
current and will share the interrupting duty. Also, the interruption of the expulsion fuse will take voltage from the CLF, allowing
simpler design of the backup CLF. The relative amount of interrupting duty that will be imposed upon the expulsion fuse will depend
on how early it melts on the current wave. A comparison of the melting I2t for the expulsion fuse versus the backup current limiting
fuse will indicate relative melting times. Fuses must be matched carefully since too large an expulsion fuse will not properly protect the
current limiting fuse from fault currents below its interrupting capability, and too small an expulsion fuse may impose too much of the
interrupting duty on the expulsion fuse. This latter factor is of less importance for expulsion fuses mounted in cutouts, but certain types
of internal transformer expulsion fuses have a limited energy withstand capability. The combination of the expulsion fuse and the
series connected backup current limiting fuse should be tested to assure proper interrupting performance.
The total clearing I2t for the backup current limiting fuse used in this approach (fuse B in the illustration above) will be larger than that
of fuse A, even though it matches the same expulsion fuse. For many transformer voltage and kVA ratings, this higher let-through
current should not be a problem but testing will be required to verify this.
If the expulsion fuse of approach 2 is an internal transformer fuse, its arcing will add some energy to that generated by the transformer
internal arcing fault. Due to the presence of the current limiting fuse, this energy increase will be significantly less than it would
otherwise be. The resultant duty on the transformer is therefore increased somewhat as compared with approach 1 (fuse A).
In order to assure proper coordination between fuses, manufacturers prepare tables illustrating limits for coordination between
expulsion and backup CLFs. Shown in Table 4-4 is a typical table put together for use with externally applied backup CLFs.
Likewise, for an oil submersible (OS) fuse a table similar to the one shown in Table 4-5 is prepared by each manufacturer.
 Page 137
Table 4-4. Fuse Selection Data Summary
Fuse Rating Application Information
Recommended
Max System Peak Maximum Minimum Maximum Size of
Design Voltage Arc Continuous InterruptingMinimumMaximum Frequently Used
Volts K-Link Class Volts Current Current Melt i2t Total i2t Series-Connected
(kV) Coordination (kV) (kV) (Amps) (Amps) (A2SEC) (A2SEC) Expulsion Fuses
K T QA
8.3 12 15 26 20 415 3,200 10,000 12 8 15
8.3 25 15 26 40 500 11,000 33,000 25 15 30
8.3 40 15 26 70 750 28,000 80,000 40 20 50
8.3 50 15 26 80 850 39,000 120,000 50 25 60
15.5 12 27 49 20 440 3,200 10,000 12 8 15
15.5 25 27 49 40 580 11,000 33,000 25 15 30
15.5 40 27 49 65 850 28,000 80,000 40 20 50
15.5 50 27 49 75 1000 39,000 120,000 50 25 60
23.0 12 35 64 20 280 3,200 10,500 12 8 15
23.0 25 35 66 40 465 11,000 38,000 25 15 30
NOTE: All designs have a 50,000 amps rms symmetrical rating
 Page 138
Table 4-5. Coordination Tables for Expulsion Fuse/OS Fuse
Single Phase - Bay-O-Net Dual (Load) Sensing RTE Line
358C, GE Link 9F59LFC/Minimum Acceptable OS Fuse
OS Fuse Voltage Rating (kV)
8.3 15.5 23.0
KVA Transformer Voltage (kV)
2.4 4.16-4.8 7.27.96 1212.47 13.214.4 19.9
Link OS Link OS Link OS Link OS Link OS Link OS
5 C03 40 C03 40 C03 40 C03 40 C03 40 C03 40
10 C05 40 C05 40 C03 40 C03 40 C03 40 C03 40
15 C08 50 C05 40 C03 40 C03 40 C03 40 C03 40
25 C10 80 C08 50 C08 40 C05 40 C03 40 C03 40
37.5 C12 125 C10 80 C08 40 C05 40 C05 40 C03 40
50 C12 150 C10 80 C08 50 C05 40 C05 40 C05 40
75 - - C12 150 C10 80 C08 50 C08 40 C05 40
100 - - C12 200 C10 100 C08 50 C08 50 C05 40
167 - - - - C12 200 C10 100 C10 100 C08 50
250 - - - - - - C12 250 C12 150 - -
333 - - - - - - C12 250 C12 250 - -
 Page 139
Distribution Transformer Fusing
Purpose of the Fuse. With reference to Figure 4-26, which shows a transformer connected phase-to-neutral in a single-phase circuit of a
multigrounded system, the basic functions of a fuse are:
a. Isolate a transformer with an internal fault from the primary circuit so that only those customers served from the faulted transformer
experience a service interruption. Ideally this isolation process will be accomplished such that the transformer does not fail in a
disruptive fashion.
b. Protect the transformer against the effects of through fault currents from bolted faults at and downstream from the secondary (low-
voltage) terminals of the transformer. The fuse should operate before the transformer is damaged thermally or mechanically.
c. Protect, to whatever extent possible, the transformer for high impedance faults or arcing faults in the secondary circuits fed from the
transformer.

Figure 4-26.
Single-Phase Circuit With Line-to-Neutral Connected Distribution
Transformer
 Page 140
In addition to performing the functions given by items a, b, and c above, the fuse applied with the distribution transformer must satisfy
overcurrent coordination criteria. The two principal coordination criteria which the transformer primary fuse should satisfy are:
a. Coordinate selectivity with the upstream line fuse, line recloser, or circuit breaker such that only the transformer fuse blows for a
permanent fault downstream from the transformer fuse, thereby preventing blowing of the line fuse or operation of the recloser or
breaker on their time delay curves.
b. Coordinate selectively with the first downstream overcurrent protective device in the secondary system, which frequently is a main
breaker or fuse at the service entrance equipment or panel.
And finally, the fuse should not operate for any conditions, such as short time overloads or transient currents, which are not harmful to
the transformer or connected secondary system. The main conditions which the fuse should withstand without operating (melting) are:
a. Transient surge currents through the fuse due to lightning, magnetizing inrush current due to core saturation, and cold load pickup
currents following an extended outage.
b. Short time overload currents which are within the thermal capacity of the transformer.
Inrush/Cold Load. Minimum transformer fuse size is usually determined by the protection of the following three points:
a. One cold load pickup point (3 times full load for 10 seconds)
b. Two inrush points:
(12 times full load for 0.1 second)
(25 times full load for 0.01 seconds)
As an example you are asked to protect a transformer rated 500 kVA, 34,500/19,000 with K type fuses. Considering inrush and cold
load pickup, what is the smallest fuse you can use (neglect damage curve of fuse)?
 Page 141
Solution:
a. Full Load Current

4-1

b. Inrush Criteria
12 × 8.4 = 100.8 amps at .1 4-2
25 × 8.4 = 210 amps at .01
c. Cold Load
3 × 8.4 = 25.2 amps at 10 4-3
Plotting the points on a K link fuse curve would suggest a fuse size of at least 12K is required.
Twice the Load. Another way to select the minimum fuse size (used by the author) is to simply use a fuse size approximately 2 times
the full load current of the transformer. For the previous example, we would select a 20K load based on the calculation (2 × 8.4 = 16.8
amperes). The disadvantage to this higher rating is that some secondary protection is lost and there is no overload protection. The
arguments in favor of this approach are:
a. It greatly reduces the number of nuisance fuse blowings caused by multiple inrush during reclosing.
b. It reduces nuisance fuse blowings due to saturation of the transformer caused by lightning.
c. Inrush can be higher than previously thought.
d. Overload protection is not the purpose of the fuse.
e. Secondary protection is not the purpose of the fuse.
f. It's easy!
 Page 142
Capacitor Fusing
Capacitors should be fused as close as possible to minimize the chance of case rupture but not to close so as to create a problem with
nuisance fuse blowings. The standards recommend fusing to at least 135% of load (some manufacturers recommend 165%). It must be
kept in mind that some fuses have continuous overload capability.
Problem. Choose an expulsion fuse for a capacitor bank rated 150 kVAR, 3-phase, on a multigrounded 13.2/7.6 kV system. Assume
the fuse has a continuous overload capability of 150%.
Solution.

4-4

Fuse * 1.5 > 6.56 * 1.35 =>

Fuse must be > 5.9 amps
(choose a 6 amp fuse)
Setting an Overcurrent Relay
Setting an overcurrent relay is not particularly difficult, although it can be confusing. The main two items to be concerned with are that
the relay does not trip at too low a current so that false trips occur and that the relay coordinates with the downstream device. The
following two examples demonstrate some of these considerations. The second example coordinates two relays in series on a
distribution feeder. While this scenario is unusual, it is very similar to the case of having a recloser downstream and does demonstrate
the same exact basic principle of setting the substation relay with enough safety margin to minimize false operation.
 Page 143
Example: Setting an IAC Relay. Assume that the IAC relay should trip on sustained current at 450 amperes minimum and 3750 amps
in 1.9 seconds. Assume: CT is 60:1. In selecting an IAC relay, you must first select a tap and then select a time dial:
Step 1 - Current Tap

4-5
Since there is no 7.5 amp tap, use the 8 amp tap.
Step 2 - Time Dial. Need 3750 amps @ 1.9 seconds.

4-6

Example: Setting Two Devices (Breakers/Reclosers) in Series. The following is a simplified methodology to set devices in series:
1. Calculate short circuit levels

Figure 4-27.
Distribution One-Line
 Page 144
2. Calculate relay minimum pickup:

4-7

3. Assume margins of .3 sec. Time of relay 1, at a #1 T.D. (using Figure 4-28), for fault at breaker 1 line is found as follows:

4-8
.2 seconds (on #1 T.D.)
Relay 2 picks up at 1330 amps so the multiple of pickup for a fault at breaker location 1 is:

4-9
\ The time is .5 (.2 + .3) seconds (see Figure 4-29) and the multiple of pickup is 1.5. Pick a #1 time dial (or greater).
 Page 145

Figure 4-28.
Relay Curve

Figure 4-29.
Safety Factor
 Page 146
Relay-to-Fuse (Feeder Selective Relaying)
The most common philosophy of feeder protection is to use feeder selective relaying which means that the feeder breaker and lateral
fuse are coordinated in such a way that the lateral fuse only operates for permanent faults on the lateral. To accomplish this, the feeder
breaker must operate before the fuse is damaged, as shown in Figure 4-30.

Figure 4-30.
Coordination During Temporary Faults
 Page 147
Because the fuse is very fast at relatively high levels of short circuit current, it is sometimes impossible for the breaker to beat the fuse
and consequently both devices will operate. For example the limits of coordination for various types of fuses (i.e., the highest current at
which coordination can be expected), assuming a 6 cycle response of the relay and breaker, is as follows:
Table 4-6. Max Current at Which
Coordination Is Possible
Fuse Size Coordination Amperes
100K 1200
100T 2000
200K 3500
200T 5800

For permanent faults, the fuse is expected to operate before the relay disc completes its travel as shown in Figure 4-31. The most
common error when providing this type of coordination is to forget to consider the overtravel of the relay disc (assuming that an
electromechanical relay is being used).
 Page 148

Figure 4-31.
Coordination During Permanent Faults
While the theory of feeder selective relaying sounds good, it is difficult to implement because true coordination is limited to a very
narrow range of fault currents. Figure 4-32 shows a very simplified illustration of why this is. As can be seen, for very low levels of
currents the fuse may not operate as it is supposed to do for permanent faults. On the other hand, the fuse is too fast for high currents
and will always operate. This would be a problem for temporary fault conditions.
On a distribution feeder, it is conceivable that all three conditions could exist, i.e., there are areas where the fuse always operates, never
 Page 149
operates, and operates properly. This situation is shown in Figure 4-33. More realistically, the condition where the breaker or the
recloser is always too fast for the fuse rarely occurs.

Figure 4-32.

Figure 4-33.
Self-Extinguishing Lightning. The reclosing sequence of a feeder breaker is usually one fast trip followed by several time delayed trips.
The time between the reclosures, i.e., the breaker is open, is called the dead time. A typical sequence of dead times is 0,15,30 seconds.
This is
 Page 150
shown in Figure 4-34. The instantaneous tripping takes about 6 cycles, which includes 1 cycle for the instantaneous relay and 5 cycles
for the breaker. Some utilities in areas of high lightning activity have found that some lightning hits to the line are self-extinguishing.
What this means is that after the lightning hits the line and initiates fault current (flashover) the fault may sometimes go out by itself. In
most instances where this is successful, a utility has wooden crossarms. Wood has been found to possess good arc quenching
capabilities. This being the case, some of these utilities will delay the instantaneous relay, which is about a 1 cycle device, by a few
cycles thereby eliminating a needless trip of the feeder.

Figure 4-34.
Feeder Breaker Reclosing
Elimination of Instantaneous Reclose. An instantaneous reclose means that upon its first trip, the breaker is immediately closed back
into the circuit, i.e., no intentional time delay is introduced to the operation. The 0 second time shown in Figure 4-34 denotes such an
instantaneous reclose. In reality, the duration of the instantaneous reclose is about 20 to 30 cycles due to the inertia of the breaker
contact heads. Some utilities have found that the instantaneous reclosure is usually unsuccessful and the temporary fault reinitiates
upon reclosure. This fault is successfully cleared after the first time delay trip where dead time is usually 5 seconds or more. The
explanation given for the lack of success with instantaneous reclosing is that the ionized gases formed during the fault do not get a
chance to dissipate if reclosure occurs too quickly. This is especially a concern for higher voltage systems using compact designs.
Some of these utilities have found that by introducing a time delay of 2 or 3 seconds (up to 15 seconds) into the first dead time, the
reinitiation can be prevented.
 Page 151
Power Quality. For a temporary fault on the lateral, a utility using feeder selective relaying will expect the feeder breaker to open,
clearing the fault. The lateral fuse in this scenario will not be affected. The problem with this technique is that the entire feeder sees a
momentary interruption and the blinking clock syndrome is created. In an effort to reduce the number of momentaries a customer sees
and increase his so-called power quality, many utilities are eliminating the instantaneous trip from feeder breaker. This, of course,
means that temporary faults on the lateral now become permanent outages and as such are negatively reflected in reliability indices,
i.e., average customer minutes out per year will increase.
High/Low Scheme. One philosophy the author has never seen used but suggests should be considered is having two philosophies of
overcurrent protection on the same feeder, which we'll call the high/low scheme. Most feeders like the one shown below have areas of
high fault current and areas of low fault current. It is suggested that the feeder breaker just protect the area of high fault currents.
Because the fuse is faster than the breaker in this area, no instantaneous trip should be used since the fuse will operate anyway and the
feeder trip will only cause more blinking clocks.

Figure 4-35.
High/Low Scheme
 Page 152
It is suggested that for areas of the system where the fault currents are approximately 2000 amperes or less and coordination is possible
with 100 amperes* or smaller fuses, reclosers be utilized and selective coordination (fuse only operates for permanent faults) be
restored. This scheme, although requiring the addition of a recloser, reduces the number of momentaries as well as average customer
minutes outages.
Recloser-to-Fuse (Lateral)
There is no real philosophy when it comes to lateral fusing but a few trends and observations can be mentioned:
Standardized Fuse Size. Most utilities pick a fuse size, like 65K, and use it for virtually all their lateral taps. The reason given is that it
is easier for the crews to deal with one size. What it also says is the lateral loads are not really important and coordination is spotty. For
example, Figure 4-36 shows a recloser/fuse coordination example. The coordination plot shows that total coordination exists only for
points between a and b, which are fault current levels. If we consider the one-line diagram and assume that fault levels a and b occur on
the middle lateral, then we can conclude that laterals closer to and farther from this point will not totally coordinate. That is, since the
fuse is the same size, the coordination plot and hence the limits stay the same. This technique, however, is probably as valid as any.
*Use of a 200 ampere fuse would shift the recloser location to a point on the system with a short current available of about 4000 amps.
 Page 153

Figure 4-36.
Load Level. Some utilities fuse laterals depending on load. You can sometimes tell this because lateral fuses vary considerably and size
is not a function of short circuit level (higher fuse sizes near the substation might indicate an attempt to coordinate). Fuses used for
laterals should be rated at least twice the lateral load to allow for cold load pickup, inrush, and emergency backfeed. It should be noted
that fusing of laterals does little if anything to prevent overload. The fusing philosophy, here, is to take out the fault, not protect for
overload. Where lateral fuse sizes are still less than 25K or 15T, there is often a problem with lightning current blowing the fuse. Most
lateral fuse operations during lightning storms, however, are caused by line flashover (fault current) which will operate any fuse size.
 Page 154
Fusing a lateral on the basis of load makes little sense, in the author's opinion, unless there is a conductor burndown problem. This
should be a rare consideration on the newer designs. On some of the older construction, using small conductors, there is little choice
but to fuse as tightly as possible and this philosophy is probably the best.
Feeder Selective Relaying (FSR)/Coordination. Some utilities pick a fuse size to allow maximum coordination with the breaker or
recloser. Utilities using 100 or 200 ampere lateral fuses might possibly not be doing it because of lateral loading but rather because
short circuit levels are relatively high and that's the only way to slow the fuse down enough on temporary faults to allow the breaker to
operate.
Reclosers and Fuses. Figure 4-37 shows the time-current characteristic curves of the automatic circuit recloser. To these curves, the
time-current characteristics of a fuse C is superimposed. It will be noted that fuse curve C is made up of two parts, i.e., the upper
portion of the curve (low current range) represents the total clearing time T.C. curve, and the lower portion (high current range)
represents the melting T.C. curve for the fuse. The intersection points of the fuse curves with the recloser curves A and B illustrate the
limits between which coordination will be expected. Basically, this is correct within the interest of simplicity. However, to accurately
establish intersection points a and b, it is necessary that the characteristic curves of both recloser and fuse be shifted, or modified, to
take into account alternate heating and cooling of the fusible element as the recloser goes through its sequence of operations. For
example, since we want to protect the fuse for two instantaneous openings, as shown in Figure 4-38, it is necessary to compare the heat
input to the fuse during these two instantaneous recloser openings.
 Page 155

Figure 4-37.
Recloser and Fuse Time-Current Characteristics

Figure 4-38.
Fuse Link Heating and Cooling
 Page 156
One method used to account for the heating of the fuse is to shift the instantaneous curve to the right. For example, if the setting of the
recloser was two instantaneous shots followed by two time delays and the dead time between reclosings was instantaneous, then the A
curve would simply be doubled. Since there are usually about 2 to 5 seconds before reclosure, the fuse gets some chance to cool. To
account for this, a factor less than two is used.

Figure 4-39.
Recloser/Fuse Coordination
If we assume the system shown above and give the recloser a 2 second dead time, we would shift the instantaneous curve by a factor of
1.35 and have the coordination plot shown in Figure 4-40. The limit of coordination would now be about 1000 amperes for faraway
faults and over 5000 amperes for closed-in faults. If the system beyond the 30T fuse is within these limits, there is total coordination.
 Page 157

Figure 4-40.
Coordination Plot of Recloser/Fuse
Relay-to-Recloser
If a permanent fault occurs anywhere on the system beyond a feeder, the recloser device will operate once, twice, or three times
instantaneously (depending upon adjustment) in an attempt to clear the fault. However, since a permanent fault will still be on the line
at the end of these instantaneous operations, it must be cleared away by some other means. For this reason, the recloser is provided
with one-, two-, or three-time delay operations (depending upon adjustment). These additional operations are purposely slower to
provide coordination with fuses or allow the fault to self-clear. After the fourth opening, if the fault is still on the line, the recloser will
lock open.
Figure 4-41 represents the instantaneous time delay characteristics of a conventional automatic circuit recloser.
 Page 158

Figure 4-41.
Tripping Characteristic for Conventional Automatic
Circuit Recloser
At substations where the available short circuit current at the distribution feeder bus is 250 MVA or more, the feeder circuits are
usually provided with circuit breakers and extremely inverse-time overcurrent relays. The relays of each feeder should be adjusted so
that they can protect the circuit to a point beyond the first recloser in the main feeder but with enough time delay to be selective with
the recloser during any or all of the operations within the complete recloser cycle.
An important factor in obtaining this selectivity is the reset time of the overcurrent relays. If, having started to operate when a fault
occurs beyond the recloser, an overcurrent relay does not have time to completely reset after the recloser trips and before it recloses (an
interval of approximately one second), the relay may inch its way toward tripping during successive recloser operations. Thus it can be
seen that it is not sufficient merely to make the relay time only slightly longer than the recloser time.
It is a good rule of thumb that there will be a possible lack of selectivity if the operating time of the relay at any current is less than
twice the time delay characteristic of the recloser. The basis of this rule,
 Page 159
and the method of calculating the selectivity, will become evident by considering an example.
First, it should be known how to use available data for calculating the relay response under conditions of possibly incomplete resetting.
The angular velocity of the rotor of an inverse-time relay for a given multiple of pickup current is substantially constant throughout the
travel from the reset (i.e., completely open) position to the closed position where the contacts close. Therefore, if it is known (from the
time-current curves) how long it takes a relay to close its contacts at a given multiple of pickup and with a given time-dial adjustment,
it can be estimated what portion of the total travel toward the contact-closed position the rotor will move in any given time. Similarly,
the resetting velocity of the relay rotor is substantially constant throughout its travel. If the reset time from the contact-closed position
is known for any given time delay adjustment, the reset time for any portion of the total travel (when the longest time delay adjustment
is used) is generally given for each type of relay. The reset time for the number 10 time-dial setting is approximately six seconds for an
inverse type overcurrent relay, and approximately 60 seconds for either a very inverse or any extremely inverse type overcurrent relay.
The foregoing information may be applied to an example by referring to Figure 4-42. Curves A and B are the upper curves of the band
of variation for the instantaneous and time-delay characteristics of a 35 ampere recloser. Curve C is the time-current curve of the very
inverse Type IAC relay set on the number 1.0 time-dial adjustment and 4 ampere tap (160 ampere primary with 200/5 current
transformers). Assume that it is desired to check the selectivity for a fault current of 500 amperes. It is assumed that the fault will
persist through all of the reclosures. To be selective, the IAC relay must not trip its breaker for a fault beyond the recloser.
The operating times of the relay and recloser for this example are:
Recloser:
Instantaneous - 0.036 sec.
Time delay - 0.25 sec.
Relay:
Pickup - 0.65 sec.
Reset - (1.0/10) (60) = 6.0 sec.
 Page 160

Figure 4-42.
Relay-Recloser Coordination
The percent of total travel of the IAC relay during the various recloser operations is as follows, where plus means travel in the contact-
closing direction and minus means travel in the reset direction:
 Page 161
Percent of Travel Relay
Recloser Operation Travel
First Instantaneous (0.036/0.65) × (100) = +5.5%
Trip
Open for One Second (1/6) × (100) = -16.7%

It is apparent from this that the IAC relay will completely reset while the recloser is open following each instantaneous opening.
Percent of Travel Relay
Recloser Operation Travel
First Time-Delay Trip (0.25/0.65) × (100) = +38.5%
Open for One Second (1/6) × (100) = -16.7%
Second Time-Delay (0.25/0.65) × (100) = +38.5%
Trip

From this analysis, it appears that the relay will have a net travel of 60.3 percent of the total travel toward the contact-closed position.
From the foregoing, it is seen that the relay travel lacks approximately 40 percent (or 0.4 × 0.65 = 0.24 second) of that necessary for
the relay to close its contacts and trip its breaker. On the basis of these figures, the IAC will be selective. A 0.15 to .2 second margin is
generally considered desirable to guard against variations from published characteristics, errors in reading curves, etc. (the static
overcurrent relay Type SFC overcomes some of these problems since the overtravel of such a relay is about 0.01 seconds and the reset
time is 0.1 seconds or less).
If the automatic circuit reclosers are used at the substation as feeder breakers, it is necessary to select the proper size to meet the
following conditions:
a. The interrupting capacity of the recloser should be greater than the maximum calculated fault current available on the bus.
b. The load-current rating (coil rating) of the recloser should be greater than the peak-load current of the circuit. It is recommended that
the coil rating of the recloser be of sufficient size to allow for normal load growth and be relatively free from unnecessary tripping due
to inrush current following a prolonged outage. The margin between
 Page 162
peak load on the circuit and the recloser rating is usually about 30 percent.
c. The minimum pickup current of the recloser is two times its coil rating. This determines its zone of protection as established by the
minimum calculated fault current in the circuit. The minimum pickup rating should reach beyond the first-line recloser sectionalizing
point, i.e., overlapping protection must be provided between the station recloser and the first-line recloser. If overlapping protection
cannot be obtained when satisfying requirement (a), it will be necessary to relocate the first-line recloser to have it fall within the
station's recloser protective zone.
Sectionalizers
The following basic coordination principles should be observed in the application of sectionalizers.
1. The minimum actuating current of sectionalizers should be 80% of the minimum trip of the source-side device. For electronically
controlled units, the minimum actuating level is used directly. For hydraulically controlled units, the same series coil rating is used.
The minimum actuating current is 1.6 times the sectionalizer coil rating to provide proper coordination with recloser minimum trip.
2. Sectionalizers not equipped with ground fault sensing should be coordinated with the phase pickup minimum trip level of the backup
device. Setting the sectionalizer actuating level to coordinate with the backup device ground pickup level may cause erroneous lockout
operations due to inrush current.
3. The sectionalizer should be set to lockout in one less operation than the backup device. This general rule need not apply in the case
of several sectionalizers in series, where successive units may be set for one, two, or three operations less than the backup recloser.
4. The opening and reclosing times of the backup device must be coordinated with the sectionalizer's count retention time. The
combined tripping (except for the first trip) and reclosing times of the backup must be shorter than the sectionalizer's memory time.
If the backup operating time is longer than the sectionalizer's memory time, the sectionalizer will partially forget the number of
backup tripping operations. This will result in the backup locking out for a fault beyond the sectionalizer and may require an extra
 Page 163
backup tripping operation, and then both the backup device and the sectionalizer would be locked out.
5. Three-phase sectionalizers are limited to coordination with three-phase simultaneous opening backup devices. Nonsimultaneous
tripping of backup devices could result in an attempted fault interruption by the sectionalizer, which is not designed for such operation.
Table 4-7. Sectionalizer Counts
Sequence of
Comment
Events
Step S1 S2 S3
1 0 0 0 Fault initiation
2 1 0 0 Breaker opens
3 1 0 0 Breaker closes
4 2 1 1 Breaker opens
5 2 1 1 Breaker closes
6 3 2 2 Breaker opens and sectionalizer number 1
opens
7 3 2 2 Breaker closes but inrush again produces
pickup
8 3 3 3 Inrush produces a count and S2 and S3 try to
open under load.
The sequence of events shown above indicates one of the
problems sectionalizers suffer due to inrush. In this case
(and there are others) the sectionalizers beyond the fault
count incorrectly due to inrush. As can be seen, all
sectionalizers could indeed open if the magnitudes of
inrush were high enough.

One condition troublesome for sectionalizers without inrush restraint is shown in Figure 4-43. In this example a fault occurs on the
lateral protected by sectionalizer #1. After the feeder breaker opens, this
 Page 164
sectionalizer will count 1. The other sectionalizers will count 0, since they did not see fault current. If the fault is permanent the
recloser will again close and open. This time, sectionalizer #1 will count 2 but sectionalizers #2 and #3 will count 1 (see Table 4-7)
since inrush through them on reclose is of a magnitude similar to a fault current. This process continues until the breaker opens and
sectionalizer #1 counts 3 and drops open, isolating the fault. The other sectionalizers, which have counted to 2, see another inrush
during this successful reclose and try to drop open during a normal energized condition. Since some sectionalizers cannot interrupt load
current this could also result in failure.

Figure 4-43.
Sectionalizer Application
Example. The distribution feeder shown in Figure 4-44 illustrates a situation where a sectionalizer must coordinate with a 100 amp
hydraulic recloser. The recloser is set for 1 fast and 3 time delay operations and the sectionalizer size must be selected.
 Page 165

Figure 4-44
Minimum trip of recloser = 2 × rating
= 2 × 100 amps = 200 amps
Sectionalizer trip setting = 80% * 200 amps
= 160 amps

4-10
Sectionalizer must be rated 100 amps or less.
Questions
1. Give two reasons why distribution system protection differs from subtransmission or transmission protection.
2. Explain why the fault current level drops off so much only a mile or so from the substation.
3. Expulsion fuses do not limit short circuit current. Explain.
 Page 166
4. The general purpose current limiting fuse has more low current interrupting capability than the backup current limiting fuse but is
still not considered full range. An expulsion fuse, on the other hand, has considerably less interrupting capability but is considered to
be full range. Explain.
5. Why must the arc voltage of the current limiting fuse be limited?
6. What is a damage curve? And when is it used?
7. Will a 10C protecting fuse coordinate with a 20C protected fuse?
8. What are the three required characteristics of a protective device?
9. Name two types of overcurrent relays.
10. Why should the relay curve be at least .3 seconds greater (for all current levels) than the downstream device?
11. What is cold load pickup and how is it considered in the setting of a relay?
12. Discuss the difference between a recloser and a breaker with a reclosing relay.
13. Explain why sectionalizers are easy to coordinate.
14. What is fault selective feeder relaying and how easy is it to obtain?
15. The minimum time for a breaker to clear a fault is about 6 cycles. Explain.
16. Impedance of a fault Zf is generally less than _____ ohms.
17. Why is inrush such a problem?
18. What is notching? How can the problem be minimized?
19. Pick a fuse for a 25 kVA, 1-phase transformer used on a 13.8 grounded wye system.
 Page 167
20. Pick a fuse size for a 1200 kVAR capacitor bank installed on a 34.5 kv grounded wye system.
 Page 168

5
Surge Arresters
Characteristics of Lightning
In order to understand the effects of lightning, it is best to acquire some knowledge as to what lightning is, how it is caused, and where
it is most likely to occur. Terms commonly used when describing lightning which will be discussed in this section are as follows:
Stroke Leader
Time Duration
Current Magnitude
Rate of Rise
Multiple Strokes
Polarity
Isokeraunic Level
Stroke Leader
Under normal conditions, it is generally believed that clouds contain positive and negative charges that, being unlike, combine and
neutralize each other resulting in substantially zero charge and hence zero voltage difference within the cloud.
One explanation of lightning is that when moist air is heated, it rises rapidly and when it gets to the higher altitudes it begins to cool. At
very high altitudes (as high as 60,000 feet) precipitation particles are formed and begin to fall. This air going up and particles coming
down (as fast as 100 mph) create the mechanism of transfer of charge where the cloud polarizes (see Figure 5-1).
When the voltage gradient between clouds or between clouds and earth reaches the limit for air, the air in the region of high stress
ionizes and breaks down. The stroke leader, which is imperceptible to the eye, is a corona-like streamer that usually starts in the cloud
as an electrical puncture. This in turn establishes the downward path of the lightning stroke between cloud and earth. The leader
usually follows the direction
 Page 169

Figure 5-1.
Separation of Charges Within Cloud
of the highest concentration of voltage gradient in successive steps. These zig-zag steps are approximately 50 yards at a time with 30 to
90 microsecond hesitations between steps as shown in Figure 5-2.

Figure 5-2.
Propagation of Stroke Leader
As the leader approaches earth, negative ions progress downward along the leader path and positive ions begin propagating upward
(upward strokes). When the upward and downward leaders meet as shown in Figure 5-3, a cloud to ground connection is established
and the
 Page 170
charge energy from the cloud is released into the ground. This release of energy is the visible discharge we call lightning.

Figure 5-3.
Return Strokes
Time Duration
The duration of a lightning stroke is usually less than a couple of hundred microseconds as shown in Figure 5-4. The industry accepted
8×20 current wave shown in Figure 5-4 is a reasonable approximation of a lightning surge.

Figure 5-4.
Typical Lightning Surge
Some lightning strokes have relatively high discharge currents over fairly short periods of time. These strokes usually produce
shattering or explosive effects without much burning taking place. On the other hand,
 Page 171
other lightning strokes last up to thousands of microseconds with fairly low discharge currents (< 1000 amps). This type of stroke,
commonly called hot lightning, produces considerable burning, melting, fires, etc. Many lightning strokes are actually a combination of
both.
The following data shows one of the statistical distributions of stroke duration reported to the industry.
Single Stroke Duration (ms) Percent
> 20 96
> 40 57
> 60 14
> 80 5
Average time = 43 microseconds

The energy in a lightning stroke is not nearly so high as some people imagine because the duration of the wave is so short. For
example, a 43 microsecond wave lasts only .26 percent of the duration of a single cycle of 60 Hz current.
Current Magnitudes
The subject of current magnitudes is controversial and confusing. While most experts agree that stroke currents have been measured in
excess of 200 kA, it is questionable to many as to how much of that current is seen by shielded equipment or discharged through an
arrester.
Measurements of lightning currents over the years show that typical stroke currents fall into the following range:
Range of Stroke Currents
5% exceeded 90,000 amperes
10% exceeded 75,000 amperes
20% exceeded 60,000 amperes
50% exceeded 45,000 amperes
70% exceeded 30,000 amperes
 Page 172
Rate of Rise
It is interesting to note that, while the industry tests on an 8×20 ms wave, this wave shape is not at all substantiated by field data. Times
to crest shown in Figure 5-5 are much more representative than the 8×20 ms wave. All the investigators we have researched reported
rates of rise higher than 10 kA/ms for over 50 percent of stroke currents. Rates of 65 kA/ms for many stroke currents were reported by
several of the investigators.

Figure 5-5.
Time to Crest vs. Probability
Multiple Strokes
Over half of all lightning strokes are multiple and have from two to as many as forty-two strokes. Multiple strokes are caused by
quickly recharging the cloud area. After the stroke first occurs, some of the electric charges in other parts of the same, or adjacent,
clouds move in to replenish the discharged area. This replenishment occurs before the gaseous path of the first stroke has dissipated
and consequently makes the path of each succeeding stroke approximately the same. Some typical values for multiple strokes are as
follows:
50% of direct strokes had at least 3 components
24% of direct strokes had at least 4 components
 Page 173
15% of direct strokes had at least 6 components
Average duration for multiple strokes was approximately 1/10 second
Maximum duration for multiple strokes was approximately 1-1/2 seconds
Polarity
Polarity refers to the clouds and the earth's relationship in that the earth's charge is positive and the cloud's charge is negative in 90% of
the recorded measurements.
Isokeraunic Level
The isokeraunic map shown in Figure 5-6 is used to indicate relative frequency of lightning on a geographical basis. The number of
thunderstorm days in any particular location is known as the isokeraunic level. For example, Miami, Florida would expect to see
between 70 and 80 thunderstorms each year. Many believe that this is only part of the story because there is indication that higher
isokeraunic levels also result in more strokes per thunderstorm day.

Figure 5-6.
Isokeraunic Map
 Page 174
The United States takes a pretty bad beating with an average for the entire country of about 40 storms per year. The lightning season of
much of the northern U.S. extends from early May to October, but prevails increasingly longer for regions farther south. In the Gulf
states, lightning storms may occur from early March to December.
System Overvoltage
There are several sources of overvoltages which must be considered when applying arresters. Overvoltages caused by lightning, neutral
displacement during line-to-ground faults and current limiting fuse operation will be discussed later in the text. Several other sources of
overvoltages which will be discussed in this section are as follows:
Ferroresonance
Capacitor Switching
Current Chopping
Accidental Contacts with Higher Voltage Systems.
Ferroresonance
In three-phase circuits, single-phase switching, fuse blowing, or a broken conductor can result in overvoltages when ferroresonance
occurs between the magnetizing impedance of the transformer and the system capacitance of the isolated phase or phases.
A myriad of practical circuit situations can occur which may result in ferroresonance phenomena. Basically, the necessary conditions
can arise when one or two open phases result in capacitance being energized in series with the nonlinear magnetizing impedance of a
transformer, as in Figure 5-7, where the switches could be fuse cutouts at a cable riser pole, the capacitance could be that of a length of
cable connecting to the ungrounded winding of a pad-mounted transformer.
Ferroresonance cannot be entirely avoided. Conditions that are likely to produce ferroresonance are as follows:
Small transformer rating; the smaller the rating the greater the susceptibility. Bank ratings larger than 300 kVA are rarely susceptible.
 Page 175

Figure 5-7.
Single-Phase Switching in 3-Phase Circuit
No load; a load as small as 4% will result in near immunity.
Any 3-phase connection is susceptible. A single-phase transformer connected phase-to-phase on a grounded primary system is
susceptible.
Primary supply by underground cables; shielding increases the capacitance and susceptibility. Overhead primary cables generally
provide immunity unless voltage is higher than 15 kV. At 34.5 kV FR is definitely a possibility with overhead supply since internal
capacitance of transformer is sufficient for resonance.
Primary voltage above 5 kV. Voltages below 5 kV provide substantial immunity. Above 15 kV, FR is quite likely. Opinions differ on
susceptibility in the 5 to 15 kV range.
Secondary capacitor bank with floating neutral, even on a grounded Y-grounded Y connection if one phase of primary is de-energized
can energize magnetizing reactance of the de-energized phase through the capacitance and thus cause resonance.
At the present time, the most practical method of avoiding FR is by installation of wye-wye connected transformers with transformer
primary and secondary neutrals grounded and connected to the system primary neutral. The primary grounded wye neutral shorts out
the series connection of the transformer reactance and cable capacitance, thereby
 Page 176
preventing the establishment of a resonant circuit. Other mitigating techniques are as follows:
Single-phase transformers should be connected line-to-neutral.
FR can also be minimized by the installation of 3-phase switching and protective devices so that single phasing cannot occur. This may
not be possible in many cases and it may not be completely effective, but it is the single best preventive.
FR can be prevented if cables and transformers are never switched together. To accomplish this, the transformer switches need to be
located at the transformer terminals rather than at the riser pole. If switches are also required at the riser pole, interlocking is desirable
to ensure that upon energizing, all riser pole phases are first closed, and then all transformer primary switches are closed. Upon de-
energizing, all phases must first be opened at the transformer before opening at the riser pole.
The requirement for never switching primary supply cable and transformers together also applies to fuses and other protective devices.
This requires that fuses, reclosers or sectionalizers at riser poles and on the distribution feeder are coordinated to hold-in on a
transformer fault so the transformer primary protection will clear first. Of course, a primary cable fault would blow a fuse at the riser
pole first, but such a fault is most likely to short out the capacitance of the faulted cable section connected to the transformer, and thus
prevent resonance.
If a susceptible connection must be used, and if primary cable runs are long and must be switched with the transformer, and if 3-phase
switching and protection is not possible, then arrange the system to have all switching done with at least 5% loading on the
transformer.
Secondary capacitor banks should be connected with grounded neutral.
Capacitor Switching
Capacitor bank switching may cause an overvoltage upon either energization or de-energization. For example, consider the following
grounded neutral bank energization (Figure 5-8).
If the initial conditions (pre-closing) are such that the capacitor bank has no charge (no voltage) and the system voltage at contact
closure is at a maximum, the voltage will overshoot as shown in Figure 5-9.
 Page 177

Figure 5-8.
Capacitor Energization

Figure 5-9.
Capacitor Voltage During Energization
De-energization of a capacitor bank is even a greater concern. Going to the system representation in Figure 5-10, but now showing the
switch opening, we create conditions sometimes referred to as voltage magnification.
Assume that R and XL are very small compared to the capacitive reactance so that the capacitor steady-state voltage is essentially the
same as the source voltage. If the switch is assumed to have opened at some time shortly before time 0, then current interruption will
take place at a normal zero such as at time a in Figure 5-11.
 Page 178

Figure 5-10.
Capacitor De-energization

Figure 5-11.
Overvoltages Due to De-energization
By our assumptions, steady-state 60 Hz current leads the source voltage by 90 degrees as shown, so that the source (and capacitor)
voltage reaches its maximum value at time a. The result of interruption is that the capacitor voltage remains at peak value because of
the charge which is trapped on it. However, the source voltage continues its normal 60 Hz variation, and the voltage which gradually
appears across the switch is the difference between the fixed capacitor voltage on the one side and the source voltage on the other. As
shown, the switch voltage reaches a maximum of twice normal value at time c, one-half cycle following interruption.
If the switch can withstand twice normal voltage at this time, successful interruption has been achieved. Because of discharge resistors,
 Page 179
normally built into the capacitors, the capacitor voltage will drain off, ultimately to disappear.
However, if the switch does not achieve adequate dielectric recovery, the arc may reignite or restrike between the contacts sometime
during the period of a to c, which will re-energize the capacitor. Maximum transient voltages result if a restrike takes place at
maximum switch voltage, time c. When current is re-established at this time, the capacitor voltage which is at plus 1.0 attempts to
rejoin the system voltage at minus 1.0 or h. It must travel 2.0 to reach value h, and thus can overshoot point h by 2.0. Then the resulting
voltage at f is 3.0 times normal.
Since the capacitor current also undergoes a natural-frequency oscillation, it is theoretically possible that a natural-frequency current
zero may occur just after time c. A second interruption here could leave a trapped charge on the capacitor with voltage f of negative 3.0
per unit. As the system voltage again swings to plus 1.0, a maximum switch voltage of 4.0 could result, and a restrike at time g would
give 4.0 + 1.0 = 5.0 times normal voltage, etc. However, compounding of this nature is rarely, if ever, found in practice. Modern
switches generally do not restrike or restrike more than once during clearing. Voltages approaching 3 times normal will occur only if
restriking occurs at the worst possible time. Voltages in the order of 2.5 times normal are more typical of field measurements.
Current Chopping
Most fault-current interrupting devices, such as expulsion fuses, reclosers, circuit breakers, etc., accomplish their arc extinction by
waiting for a 60 Hz current zero. Transients produced in this manner are usually twice normal or less. It is possible under some
conditions such as current limiting fuse operation or breaker interruption of low currents where a current interruption occurs prior to
the normal current zero. This so-called current chop can cause exceptionally high voltages depending on the rate of current
interruption, the amount of current chopped and the system configuration.
Let us analyze abrupt current chopping by assuming that the current is forced instantaneously to zero from some finite value. If this
current is flowing in an inductance, it cannot change instantaneously, and it therefore follows that practically there must be capacitance
and/or resistance associated with the inductance if arc voltage is neglected.
 Page 180
Consider the circuit of Figure 5-12 where resistance is ignored and assume that the capacitive reactance is so much larger than the
inductive reactance that its normal current is negligibly small compared to i (i.e., wn = is very large).

Figure 5-12.
Circuit Illustrating Current Chopping
When the switch current suddenly changes from a value i to zero, i continues to flow instantaneously in L; therefore, it must also flow
in C. A natural frequency oscillation ensues in L and C. The maximum natural frequency voltage appearing across L (and C) is

5-1
Thus, the voltage is proportional to the magnitude of current chopped and to the surge impedance of the circuit being switched.
This equation can be manipulated as follows to express the transient voltage in a different way.
 Page 181

5-2

However, is the per unit natural frequency, wn/w.

Thus,

5-3
If XLi is the normal voltage, or some measure of the normal voltage across the inductance, then the transient voltage, e, is many times
normal by the ratio wn/w.
Theoretically, then, chopping can produce very high voltages. In practice, however, L is often the nonlinear magnetizing impedance of
a transformer. The magnetic characteristics of modern transformers coupled with typical switch performance do not usually give rise to
voltages more than 2 times normal.
Accidental Contact with Higher Voltage Systems
Often overhead primary distribution circuits are built underneath higher voltage circuits on the same pole. Broken high-voltage
conductors can fall upon the lower voltage circuit primary possibly causing the lower rated arresters to fail along the entire line or other
major equipment damage.
 Page 182

Silicon Carbide vs. Mov Arresters

The utility industry, like other industries has seen changes in arrester design over the years. While several older designs can still be
found on a few distribution systems, the vast majority of arresters now in use are either:
Gapped Silicon Carbide, or
MOVs (Metal Oxide Varistors).

Figure 5-13.
Silicon Carbide and MOV Arrester
Most of the arresters found on the distribution system are the older type gapped silicon carbide. The introduction of the metal oxide
arrester in the late 70's is one of the most significant advances in the utility industry especially since its acceptance by the industry
became almost universal by the early 1980's. At this time, manufacturers have even gone completely out of the business of making
intermediate and station class silicon carbide arresters and very few are making the distribution type.
A silicon carbide surge arrester has silicon carbide valve elements that are protected from continuous power-frequency voltage by a
series gap which acts as the insulator during normal voltage conditions and interrupts the power-frequency current that follows any
transient current discharged by the arrester. It does this by not restriking on subsequent half cycles of power-frequency voltage after the
first follow-current zero
 Page 183
has occurred. Voltage and current zeros occur simultaneously, permitting the gap to clear the circuit established through the arrester.
In a metal-oxide-varistor arrester, the metal-oxide disk insulates the arrester electrically from ground. The disk is composed of a variety
of materials in varying concentrations which determine the electrical characteristics of the varistor. Highly conductive particles
(usually zinc oxide) are suspended in a true semiconductor with characteristics close to that of a back-to-back zener diode.
The processing of metal-oxide-varistor disks is extremely critical. Purity of the materials and their uniform dispersion throughout the
disk must be carefully monitored. To demonstrate how critical monitoring is, an experiment was conducted in which the concentration
of one of the materials in a disk was increased by 50 parts per million beyond its specified parts-per-million concentration. The varistor
that resulted from this increase demonstrated a 15% improvement in protective characteristics, but the overall life expectancy of the
varistor was decreased by 90%.
A metal-oxide disk goes into conduction sharply at a precise voltage level and ceases to conduct when voltage drops below this level.
A series gap is not required to insulate the arrester from ground or to interrupt power follow current (which does not exist as long as
applied power-frequency voltage is below conduction voltage level).
The primary difference between silicon carbide and MOV arresters is that the MOV valve blocks are so nonlinear that no or at least
very little power current is drawn at normal line-to-ground voltage. The MOV arrester consequently does not require a series gap. The
MOV simply eases in and out of conduction. A comparison of the nonlinear characteristics shown in Figure 5-14 dramatically
illustrates the extreme nonlinearity of the MOV.
 Page 184

Figure 5-14.
Comparison of the Nonlinear Characteristics of MOV and
Conventional Silicon Carbide Valve Blocks

Classes of Arresters
There are 3 classes of arresters: distribution, intermediate and station. For the most part the major difference in these types of arresters
is the size of the block. A larger block reduces IR discharge voltage and greatly increases energy capability and thus reliability. All
three classes of arresters are available for distribution system as shown below. Distribution class arresters are as a rule utilized out on
the distribution feeder whereas intermediate and station class arresters are used in the substation.
 Page 185
Table 5-1. Voltage Ratings in Kilovolts
Distribution Intermediate Station
Arresters Arresters Arresters
1
3 3 3
6 6 6
9 9 9
10
12 12 12
15 15 15
18
21 21 21
24 24
25
27
30 30 30
36 36
39 39
48 48
60 60
72 72
90 90
96 96
108 108
120 120

Arrester Selection
Choosing an arrester rating for a distribution system is based on the system's line-to-ground voltage and the way it is grounded. The
limiting condition for an arrester does not usually have anything to do with the magnitude of the surges, (switching or lightning) that it
might see. This is in contrast to the selection of arresters for transmission. In distribution, rating of the arrester is based on the
maximum steady state line-to-ground voltage the arrester might see. This limiting condition is normally caused when there is a line-to-
ground fault on one of the other phases.
 Page 186
According to ANSI Standard C62.22, Guide for the Application of Metal-Oxide Surge Arresters for Alternating-Current Systems,
proper application of arresters on distribution systems requires knowledge of 1) the maximum normal operating voltage of the power
system, and 2) the magnitude and duration of temporary overvoltages (TOV) during abnormal operating conditions. This information
must be compared to the arrester MCOV rating and to the arrester TOV capability.
MCOV
The term MCOV or maximum continuous operating voltage sounds simple enough but has been difficult for many utilities to
determine. On a distribution system where the voltage is always changing due to varying load demands, and where the voltage on one
part of the system may be somewhat different to other parts (e.g., near the substation and at the end of the feeder), it is sometimes
impossible to define only one MCOV.
The MCOV of the arrester is, however, somewhat easier to define since it is approximately 84% of the arrester duty cycle rating as
shown in Table 5-2. What this means is that a 10 kV duty cycle rated arrester, typically used for a 13.2 kV system could be operated
continuously with a maximum continuous line-to-ground voltage of 8.4 kV or less.
Table 5-2. Arrester
Arrester Rating MCOV
3 2.55
6 5.10
9 7.65
10 8.40
12 10.2
15 12.7
18 15.3
21 17.0
24 19.5
27 22.0
30 24.4
 Page 187
Table 5-3. Commonly Applied Voltage Ratings of Metal-Oxide Arresters on
Distribution Systems*
Commonly Applied Arrester Voltage Ratings -
System Voltage (Kilovolts
kV RMS Duty Cycle Voltage Ratings
rms)
(MCOV)******
Three-Wire Three-Wire
Maximum Four-Wire
Nominal Low High
Voltage RangeMultigrounded
Voltage Impedance****Impedance****
B** Neutral Wye
Grounded*** Grounded
2400 2540 3 (2.55)
4160Y/2400 4400Y/2540 3 (2.55) 6 (5.1) 6 (5.1)
4260 4400 6 (5.1)
4800 5080 6 (5.1)
6900 7260 9 (7.65)
8320Y/4800 8800Y/5080 6 (5.1) 9 (7.65)
12000Y/6930 12700Y/7330 9 (7.65) 12 (10.2)*****
12470Y/7200 13200Y/7620 9 (7.65) 15 (12.7)*****
OR 10 (8.4)
13200Y/7620 13970Y/8070 10 (8.4) 15 (12.7)*****
13800Y/7970 14605Y/8430 12 (10.1) 15 (12.7)*****
13800 14520 18 (15.3)
20780Y/1200022000Y/12700 15 (12.7) 21 (17.0)*****
22860Y/1320024200Y/13870 18 (15.3) 24 (19.5)*****
23000 24340 30 (24.4)
24940Y/1440026400Y/15240 18 (15.3) 27 (22.0)*****
27600Y/1593029255Y/16890 21 (17.0) 30 (24.4)*****
34500Y/1992036510Y/21080 27 (22.0) 36 (29.0)*****
* Spacer cable circuits have not been included - there has been insufficient
experience with the application of metal-oxide arresters on spacer cable
circuits to include them in this table - refer to [30] for information on spacer
cable circuit overvoltages.
** See ANSI C84.1-1989.
*** Line-to-ground fault duration not to exceed 30 minutes. For longer
durations consult manufacturers' temporary overvoltage capability.
**** Low impedance circuits are typically 3-wire, unigrounded at the
source. High impedance circuits are generally ungrounded (i.e., delta).
Additional information regarding system grounding is contained in ANSI
C62.92, Part 1.
***** Individual case studies may show lower voltage ratings may be used.
****** For each Duty Cycle rating the Maximum Continuous Operating
Voltage (MCOV) is also listed.

Table 5-3, from ANSI C62.22, shows the commonly applied voltage ratings of metal-oxide arresters for distribution systems. All these
duty
 Page 188
cycle ratings are the same as the rating for the older gapped silicon carbide arresters except at the 13.8 kV level. Typically, a 13.8 kV,
4-wire, multigrounded system have used 10 kV gapped arresters. Today, most of these same utilities are still using 10 kV MOVs.
Some utilities, however, have recognized that the 10 kV arrester is very marginal and should possibly be replaced by a 12 kV rating to
be on the more conservative side.
TOV
How much voltage shift which will occur is a function of the type of system grounding. For example, on a delta system, a line-to-
ground fault will cause a full offset, i.e., the line-to-ground voltage will become the line-to-line voltage. Figure 5-15, shown below,
illustrates this condition. As can be seen, when a phase has a fault there is no current since the transformer is delta connected. In a
sense, ground at this point is A phase. The arresters connected from B and C phase to ground will now in effect be connected B to A
phase and C to A phase or line-to-line. This of course means that the voltage across these arresters will increase to 1.73 per unit. This is
the voltage upon which the arrester rating is based whether they be gapped silicon carbide or MOV's.

Figure 5-15.
Line-to-Ground Fault on a Delta System
Most distribution systems are classified as 4-wire multigrounded systems. The fourth wire is, of course, the neutral wire which is
grounded periodically at the pole. Figure 5-16 illustrates this type of
 Page 189
system under a single line-to-ground fault condition. As can be seen, for this solidly grounded system considerable fault current will
flow. If grounding was perfect, there would be no voltage difference between the substation ground and the point of the fault. If this
were the case, the voltage at the point of the fault would remain at 0 potential and arresters connected from the other two phases would
see no change in voltage. The ground, however, is not perfect and some rise does occur. For this type of system, the rise associated
with a single line-to-ground fault is considered to be 20% (or 25% if regulation is factored in). Consequently, the arresters on B and C
phases would see approximately 1.25 per unit across their terminals for this condition.

Figure 5-16.
Line-to-Ground Fault on a Grounded Wye System
Based upon some of the concepts just discussed as well as many years of successful operating experience, a proposal for the selection
of distribution lightning arresters was made some years ago by a working group in IEEE. This group proposed that the ratings of
gapped type surge arresters selected for open-wire, multigrounded neutral systems be equal to or greater than the nominal line-to-
neutral voltage multiplied by the product of the regulation factor 1.05 and the voltage rise factor 1.2. This is equivalent to 1.25 times
nominal line-to-neutral system voltage. For an MOV type arrester this voltage is compared to the TOV rating of the MOV. Because the
MOV arrester is more sensitive to poor grounding, poor regulation, and the reduced saturation sometimes found in new
 Page 190
transformers, it is generally recommended that a 1.35 factor be considered for MOVs.
A summary of this and other recommendations is as follows:
Open-Wire Multigrounded System
Rating = Nominal L-G voltage × 1.25 (Gapped)
Rating = Nominal L-G voltage × 1.35 (MOV)
Spacer Cable Systems
Rating = Nominal L-G voltage × 1.5
Unigrounded Rating = Nominal L-G voltage × 1.4
The difficulty now is to determine the duration of the TOV. If the assumption is made that the maximum duration of the fault is 300
seconds, then the maximum temporary system overvoltage could be 1.095 per unit of duty cycle rating (from Figure 5-17) or 1.3 per
unit of MCOV, since duty cycle rating is approximately 19% higher than MCOV.

Figure 5-17.
Example Curve - Consult Manufacturer for Exact Curve
An example of calculations to determine the arrester rating is as follows:
 Page 191
Table 5-4
Max I- Arrester Actual
System Min.
System Multiplication g Duty Arrester
Voltage Required
Grounding Factor Voltage Cycle MCOV
kV MCOV*
kV Rating
12.47 Multigrounded 1.35 9.72 7.5 9 7.65
13.2 Multigrounded 1.35 10.3 7.9 10 8.4
13.8 Multigrounded 1.35 10.76 8.3 10 8.4
23. Multigrounded 1.35 17.9 13.8 18 15.3
34.5 Multigrounded 1.35 26.9 20.7 27 22.0
34.5 Delta 1.82 36.25 27.9 36 29.0
34.5 Spacer 1.5 29.9 23. 30 24.4
34.5 Uniground 1.4 27.8 21.4 27 22.0
*Min MCOV = Max l-g voltage/1.3

Selection
As can be seen, when comparing Table 5-3 and Table 5-4, the arrester ratings for four wire multigrounded systems tend to be
determined by the MCOV while the less effectively grounded systems tend to be determined using TOV. A summary of recommended
arrester ratings for various system voltages and grounding practices and faults lasting less than 300 seconds is shown in Table 5-5.
Table 5-5. Arrester Rating (Duty Cycle)*
System Voltage 4-Wire
kV Multigrounded DeltaUnigroundedSpacer
12.47 9 12 10 10
13.2 10 15 10 12
13.8 12 15 12 12
23 18 24 18 21
34.5 27 36 27 30
*Rating based on fault lasting <300 seconds.
 Page 192
Considerations in the Applications of MOVs
Selection of an MOV arrester is primarily based upon the maximum continuous operating voltage (MCOV) that is applied across the
arrester in service (line-to-ground). For arresters on effectively grounded systems, this is normally the maximum line to ground voltage
(e.g., on a 13.8 kV system the maximum steady state line-to-ground voltage is calculated as 1.05 * 13.8/ or 8.37 kV). For an
ungrounded or impedance-grounded system, the MCOV should be at least 90 percent of the maximum phase-to-phase voltage.
Other considerations in the application of MOV's are temporary conditions on the distribution system that raise the voltage such as a
line-to-ground fault. MOV's have temporary overvoltage curves similar to the one shown below which must be considered for the
following conditions:
Voltage Regulation. Voltage standards call for a voltage at the meter no higher than 5% over nominal. This standard does not however,
limit the voltage fluctuation out on the feeder. For example, an Electric Power Research Institute study showed that the substation
voltage can be as high as 17% over nominal (the average was 7%). With average substation voltages, voltages out on the feeder are
generally not more than 5% over nominal. It is suspected that capacitor operation at light loads, or improper settings of voltage
regulators are resulting in system voltages 10% above nominal or even higher without the knowledge of the utility. The primary
concern regarding these quasi-steady-state overvoltages is that long term stability of metal oxide valve elements is normally
demonstrated at MCOV rather than at these higher voltages. A utility suspecting such overvoltages may do well to investigate going to
the next higher rated arresters.
Ferroresonance. Ferroresonance has long been a concern of the distribution engineer. Higher voltage levels, longer lines, and
underground cables have further increased this concern. Ferroresonance, when it occurs, does not always result in component failure.
This is due in part to the fact that the overvoltages can be fairly low and the condition may not last more than a few seconds, e.g.,
closing or opening a switch. Most ferroresonant overvoltages are in the range of 1.5 to 2.0 P.U. and as noted earlier, the power
frequency sparkover of gapped silicon carbide arresters is usually 2.0 to 2.4 P.U. The effect of
 Page 193
ferroresonance may then be negligible in many cases because the sparkover voltage of the arrester is rarely, if ever, exceeded. On the
other hand, since a metal oxide arrester has no gap and conducts current that is a function of applied voltage, it will make an
unsuccessful effort to reduce the ferroresonance overvoltage. The current conducted by a metal oxide distribution arrester at 2.0 P.U.
may be up to a few hundred amperes, depending on the impedance of the system, and the arrester may fail unless steps are taken to
limit the possible duration of the overvoltages.
Cogeneration. Overvoltages may occur in many circuits primarily because of problems associated with relay application and operation.
For example, cogenerators are often interconnected to the primary distribution system by means of a delta-connected transformer. This
connection utilizes only 3 wires of the standard four-wire system. During a feeder ground fault, the feeder breaker will open to separate
the faulted section from the system, but because of the delta connection, the overcurrent protection of the cogenerator will not operate.
After some delay, the underfrequency or undervoltage relay of the cogenerator will operate to disconnect it from the faulted section.
During the period of time between feeder breaker operation and cogenerator separation, the four-wire faulted section operates as a
three-wire system and the line-to-ground voltages on the unfaulted phases may reach 1.73 P.U. This overvoltage may have no effect on
gapped arresters, but the metal oxide arresters on this feeder will conduct current on the overvoltage. If the overvoltage is high, the
metal oxide arresters will fail unless the duration of the overvoltage is short.
Line-to-Ground Faults. Selection of rating for gapped silicon carbide distribution arresters is based on experience and on calculated
values of overvoltage on the unfaulted phases of a three-phase circuit during a fault to ground on one phase. The most commonly used
application rule for open-wire multigrounded neutral systems is that arrester rating be equal to or greater than the product of the
nominal [line-to-neutral voltage] × [the Range A factor 1.05] × [1.2] which is the maximum voltage rise on the unfaulted phases of a
loaded circuit. This is equivalent to 1.25 times nominal line-to-neutral system voltage. The rule cited above is very conservative for
silicon carbide arresters because the maximum calculated voltage is equal to arrester rating, and arrester sparkover is well in excess of
rating; therefore, another transient must be superimposed to cause sparkover during the short time that the overvoltage exists.
Furthermore,
 Page 194
silicon carbide arresters will generally tolerate at least a few operations at overvoltages as high as 1.2 times rating (1.5 × nominal line-
to-neutral voltage). For these reasons, the application rule given above has been satisfactory for silicon carbide arresters even though
many four-wire distribution systems are not effectively grounded. The same rule is not as conservative for metal oxide arresters unless
it is known that the system is truly grounded. Some metal oxide arrester failures are being experienced with the cause of failure
believed to be overvoltages in excess of arrester temporary overvoltage capability. In order to obtain an accurate estimate of system
overvoltage for comparison with arrester temporary overvoltage capability, it is necessary to take into account factors such as ground
wire size, ground rod spacing, fault resistance, earth resistivity and system impedance. Some utilities are using a 1.35 factor instead of
the usual 1.25 to accommodate for this.
Insulation Coordination
Margins for Overhead Equipment
From previous sections, we know how an arrester is rated, but how is it applied? It is important to note that application of arresters for
transmission and distribution is different. In transmission, lightning is of secondary concern in surge arrester application. Primary
concern is switching surges. On a distribution circuit, however, the relative low-voltage and short lines tend to make switching surges
minimal and, consequently, lightning is of primary importance.
Reflection of this fact can be seen in typical characteristics published for distribution-class arresters as shown in Tables 5-6A and 5-6B.
As can be seen, protective characteristics are shown for front-of-wave sparkover and IR discharges, but not for switching surge waves
(as shown for higher rated transmission arresters).
The two protective characteristics normally used for insulator coordination are:
Front-of-Wave Sparkover. This is the first thing that happens to the gapped arrester -- it sparks over. It is compared to the fast front
equipment insulation characteristics such as the chopped wave insulation
 Page 195
Table 5-6A. Distribution Arrester Characteristics per Handbook -
Silicon Carbide
Arrester
Maximum ANSI Front-
Rating Maximum Discharge Voltage
of-Wave Sparkover
(kV (kV Crest) at Indicated 8 × 20
(kV Crest)
RMS) Microsecond Impulse Current
With Externally 5000 10,000 20,000
Disconnector Gapped Amperes Amperes Amperes
3 14.5 31 11 12 13.5
6 28 51 22 24 27
9 39 64 33 36 40
10 43 64 33 36 40
12 54 77 44 48 54
15 63 91 50 54 61
18 75 105 61 66 74
21 89 72 78 88
27 98 87 96 107

Table 5-6B. Distribution Arrester Characteristics per Handbook -

MOV (Heavy Duty)
Arrester Front-of-Wave Maximum Discharge Voltage
MCOV
Rating Protective Level* 8/20 ms Current Wave
kV rms kV rms kV Crest 5 kA 10 kA 20 kA
3 2.55 10.7 9.2 10.0 11.3
6 5.10 21.4 18.4 20.0 22.5
9 7.65 32.1 27.5 30.0 33.8
10 8.40 35.3 30.3 33.0 37.2
12 10.2 42.8 36.7 40.0 45.0
15 12.7 53.5 45.9 50.0 56.3
18 15.3 64.2 55.1 60.0 67.6
21 17.0 74.9 64.3 70.0 78.8
24 19.5 84.3 72.3 78.8 88.7
27 22.0 95.2 81.7 89.0 100.2
30 24.4 105.9 90.9 99.0 111.5
36 30.4 124.8 107.0 116.6 131.3
*Based on a 10 kA current impulse that results in a discharge voltage
cresting in 0.5 ms
 Page 196
level of the transformer. An MOV has no gap but does have an equivalent sparkover as shown in Table 5-6B.
IR Discharge at 10 kA. After the arrester sparks over the gap, the lightning current discharges through the block material. Standards
recommend that a 10 kA discharge level be used for coordination purposes. Many utilities, however, are using a 20 kA discharge level
to gain some additional margin. (Discharge characteristics across a MOV are very similar so margin calculation is virtually identical.)
Distribution equipments are normally defined as being in a voltage class such as 15 kV, 25 kV, etc. Most utility equipment is operated
in the 15 kV class. A distribution transformer in the 15 kV class is defined by the following insulation characteristics:
60 Hz, one minute withstand = 34 kV
Chopped Wave (short-time) = 110 kV at 1.8 ms
BIL = 95 kV

Figure 5-18.
Characteristics of Transformer Insulation
Assuming a 12,470 volt, 4-wire system (7200 volts L-G), we would select the arrester rating based on the rules developed in the
previous section, i.e., a 9 kV arrester (gapped or MOV).
We can see from Table 5-6 that a 9 kV gapped arrester has a sparkover of 39 kV and an IR discharge at 10 kA of 36 kV. This could be
plotted with the transformer characteristics as follows:
 Page 197

Figure 5-19.
Insulation Coordination
Standards recommend 20% margins calculated by the formula:

5-4
Two margins are calculated, one for the chopped wave and one for the full wave (BIL) of the transformer. These calculations are
performed as follows:

5-5
As can be seen, these margins (182% and 164%) are greatly in excess of the recommended 20% and consequently show good
protection practice. If we were using an MOV we would simply use the equivalent sparkover or compare only the IR discharge and the
BIL since this is the lesser margin. The margins would be similar.
 Page 198
Margins for Underground Equipment
If the system is underground, we must be more concerned with the phenomena of traveling waves and the consequent doubling of
voltage surges at an open point. For example, a typical underground residential design is shown in Figure 5-20.

Figure 5-20.
Underground Lateral
A surge entering this cable will travel to the open point where its voltage will double, as shown in Figure 5-20 and start on its way
back.
This reflected wave plus the incoming waves impose approximately twice the normal voltage on the entire cable and all of the
equipment connected to it (see Figure 5-21). For example, if we had an arrester with a 36 kV IR discharge level (we are only
considering BIL margin), we would now expect to see 72 kV imposed across the insulation of this equipment. The new margin would
then be calculated as follows:
 Page 199

5-6

Figure 5-21.
Reflected Surge Voltage at Open Point
Since 32% is greater than the recommended 20% no problem would normally be anticipated. However, it is well known to many
utilities that impulse failures at this voltage do occur.
Similar margins can be calculated for other classes of voltages as shown in Table 5-7.
Table 5-7
Distribution Voltage 4- Arrester IR at Overhead Underground
BIL
Wire Multiground Rating 10 kA Margin % Margin %
12470 95 9 36 164 32
24940 125 18 66 89 -5
34500 150 27 96 56 -22

As can be seen, the only underground systems that can be protected using a standard distribution class arrester are those in the 15 kV
class and below. At the 34.5 kV and 25 kV levels the margins are actually negative. The general recommendations at these high
voltage levels are:
Use a better riser pole arrester (intermediate or station class).
Put an arrester at the open point.
 Page 200
Use of an intermediate class arrester at the riser pole can create slightly more margin but is usually not sufficient to maintain the
recommended 20% level. Open tie arresters on the other hand do not totally prevent doubling but are still quite effective in increasing
margin.
On 25 kV underground systems, many utilities still use only riser pole protection but put in an intermediate class arrester (IR 55 kV)
which provides approximately 14% margin and apparently considers it close enough. On the 34.5 kV systems, many utilities are now
putting in open tie protection and assuming that regardless of arrester characteristic or type of installation there is plenty of margin
because there is no doubling.
Factors Affecting Margins
Neither the arrester discharge voltage nor the equipment BIL are constant. This section will review some of the items which reduce the
margins we normally calculate.
Rate of Rise/Arrester Characteristics. The IR discharge of an arrester is found using the standard 8 × 20 ms wave. This wave has been
used for many years and has been considered to be representative of lightning. Experimental data has indicated that actual lightning
front times are much faster. Figure 5-22 showed data that is more representative of present industry thinking that front times of 2
microseconds or less are not unusual.
A time to crest of 1 microsecond would occur 17% of the time. If a wave with this rise time was impressed on an arrester, both the
sparkover and the IR of the arrester would change from the values published for a standard 8 × 20 ms wave. Industry curves show that
for such a wave the IR characteristic of a silicon carbide arrester increases approximately 30% and for an MOV the increase is about
10%. This is one of the main advantages of the MOV over the gapped silicon carbide arrester.
 Page 201

Figure 5-22.
Times to Crest of Lightning

Figure 5-23.
Voltage vs. Lead Length
 Page 202
Lead Length. The lead length of an arrester is the length of lead from the phase wire to the top of the arrester and the lead between the
ground of the arrester and the metallic sheath of the cable. These leads produce a voltage due to the rate of rise of current passing
through their inductance and added to the IR drop through the arrester in order to establish the total magnitude of discharge voltage
impressed upon the system.
It has been established over the years that the drop is approximately 2 kV per foot based upon a lead inductance of 0.4 microhenries
per foot, a 40 kA stroke, and an average rate of rise of 5 kA per microsecond.
Some time ago, tests were made using 7.5 feet of test lead and a current of only 9.5 kA crest and having a wave shape of 2.6 × 8
microseconds showing that the voltage appearing across the lead as 24.6 kV or 3.28 kV per foot. It is not always understood that the
initial rate of rise of current is usually much faster than the average rate. Even assuming that the wave front is sinusoidal (and it is
faster than sinusoidal) the initial rate of rise can readily be shown to be 1.57 (crest-current/time to crest). Figure 5-23 shows the arrester
voltage per foot of lead length for various currents and times to crest assuming a sinusoidal current and a lead inductance of .4
microhenries/foot.
While the ANSI guides do point out that the arrester discharge voltage should be considered to be the total of the arrester IR and the
lead voltage, it is rare when margins calculated for underground protection include this. Also, C62.2-1981, suggests that the acceptable
voltage per foot of lead length is 1.6 kV/ft. which might be considered low.
BIL Deterioration. The deterioration of BIL in transformers and cables has received considerable attention in recent years primarily
due to a much higher than expected failure rate. About 10 years ago, a research project sponsored by ERDA studied the effect of aging
and loading on distributor transformer BILs. The units tested were from several manufacturers and rated 25 kVA, 95 kV BIL, with a
65°C-rise insulation system. The diagram shown below in Figure 5-24 indicates the results of these tests for the aged units.
Some of the general conclusions drawn from this study are:
Almost 50% of the aged units failed below the 95 kV level.
Aging and loading reduced initial BIL to an average of 64%.
 Page 203

Figure 5-24.
BIL Deterioration vs. Aging and Loading
Reflections. Reflections at the open point can be reduced by using open tie arresters but are not eliminated. Studies indicate that the
maximum voltage that can be seen on an underground cable using gapped silicon-carbide for riser pole and open tie protection is equal
to IR discharge + 1/2 sparkover. Since sparkover (S.O.) is approximately equal to IR at 10 kA, the maximum voltage is approximately
150% of IR. The initial thought was that MOVs, having no sparkover, would eliminate this problem. Such is not the case. Studies
performed with MOVs at the open point indicate that reflections with peaks of approximately 130% of IR are typical. If the effect of
the increased rate of rise at the open tie is factored in, these estimates may be even higher.
Other. There are several other areas which also contribute to the uncertainty of the protective level. For example, while most utilities
calculate margins based on a 10 kA discharge current and the more conservative on a 20 kA discharge, there is certainly a considerable
body of evidence to suggest that the discharge currents could be considerably
 Page 204
higher. Also the effect of system voltage polarity, effectiveness of grounding, multiple taps, etc., can all have a major effect on
protective margins.
Considerations When Calculating Margins
The present use of the 20% margin factor does not properly address all the factors just discussed. Since lightning is a variable, so also
are the degrees of protection afforded the utility equipment. It is necessary and possible, however, to address the concerns of varying
arrester characteristics, lead length, BIL deterioration, reflections, magnitudes, etc., and ascertain whether these factors can all
reasonably be accounted for within the 20% value. While guidelines mention many of the items discussed, they are unclear as to
methods of quantifying their effects.
Table 5-8, shown below, is an attempt to put values on various degrees of concern. For example, a utility having a 13.2 kV system
located in a low isokeraunic level area and with a low historical failure rate might fit into the small concern category. A utility using
34.5 kV in a high isokeraunic level area and experiencing a high failure rate problem might fit into the extreme concern category.
Table 5-8. Change in Arrester BIL Characteristics
Open-Tie
Arrester Lead
BIL Arrester
Characteristics Length
Deterioration Reflection
Change Rate of Rise kV
Coefficient MOV
Small 5% 3.2 -10% 1.3
Concern
Moderate 10% 8 -20% 1.3
Concern
Extreme 15% 24* -30% 1.3
Concern
*6 ft. at 4 kV ft. = 24 kV

As an example, suppose we are trying to protect a 34.5 kV, 4-wire, underground system using 27 kV distribution class MOV arresters
at the riser pole and at the open tie. Figure 5-25 illustrates the system.
 Page 205

Figure 5-25.
34.5 kV Underground Lateral
Some standard parameters used in the calculation of margin are as follows:

BIL - 150 kV
IR @ 20 - 100 kV
kA
Lead - 6.4 kV (4 ft. of leads at 1.6
Voltage kV ft.)

The most common method of calculating this margin is to assume that the open tie arrester prevents doubling and the lead length
voltage is minimal. This calculation would be as follows:

5-7
This margin is, of course, almost double the value of the guidelines and might be interpreted as providing a very good level of
protection.
ANSI Guidelines do suggest that lead length voltage be considered as part of the IR discharge. If this is done, the new margin can be
calculated as follows:

5-8
So, as we can see, commonly used methods produce margin which would seem to provide very adequate margins of protection.
 Page 206
If, however, we considered the items illustrated in Table 5-8, we can begin to see why our assumed margins may be too optimistic. If
we factor in the three levels of concern, we can calculate the following margins:
Small
New BIL = 150 × .9 = 135
New IR = (100 × 1.05 + 3.2) * 1.3 = 141 kV

5-9
Moderate
New BIL = 150 × .8 = 120
New IR = (100 × 1.10 + 8) * 1.3 = 153 kV

5-10
Extreme
New BIL = 150 × .7 = 105
New IR = (109 × 1.15 + 24) * 1.3 = 181

5-11
Illustrated graphically in Figure 5-26, we can see the great disparity in margin calculation depending on what conditions are factored
into the calculation.
 Page 207

Figure 5-26
of Protection at 34.5 kV
Effect of Traveling Waves
The previous section dealt with conditions which are widely discussed and generally accepted as considerations in a complete
evaluation of insulation coordination. There are a number of other items, resulting from the fact that lightning is a complex traveling
wave, that a good protection engineer should be aware of, which may explain the cause of some failures where protection was thought
to be adequate.
Figure 5-27 shows a 12.47 kV system with a lateral underground tap of 400 meters. As was shown previously, the overvoltage
protection of this system is normally considered adequate by most utilities with the use of one arrester at the riser pole. In many cases,
the arrester at the riser pole is a riser pole class (non-official designation) arrester. These arresters have better characteristics than
distribution class arresters and as such give even more margins. This, of course, tends to support the argument that only one arrester is
necessary.
 Page 208

Figure 5-27.
12.47 kV Underground Lateral Tap
System conditions, lightning waveshapes, and the fact that lightning is a traveling wave, complicate any analysis. Add to this the fact
that the arrester is not perfect either and a strong case can be made that few underground systems have adequate protection. The
following are some examples of traveling wave phenomena which decrease protective margin. For all examples, a 1 × 20 msec surge is
used in the simulation since it is much more representative of actual lightning than the standard 8 × 20 ms. The arrester at the riser pole
is a riser pole class with a 30 kV IR discharge for a 10 kA discharge.
Traveling Waves
Voltage Doubling
Figure 5-28 shows the discharge voltage of the 9 kV riser pole arrester. This IR discharge is equal to approximately 30 kV and
approximately a dc discharge. This waveshape propagates down the 400 meter long cable where it is reflected at the open tie point
causing a doubling of the voltage to 60 kV as is seen in Figure 5-29. This doubling of discharge voltage would still allow a margin of
approximately 58% if some of the considerations like BIL deterioration, lead length, etc., are neglected.
Addition of an open tie arrester to this system will prevent doubling at the open tie as can be seen in Figure 5-30. While this form of
protection is very effective at the open point, it does not prevent all reflections. Figure 5-31 shows the cable midpoint voltage for this
condition. As can be seen, positive waves are reflected back which add to the incoming voltage and produce a maximum voltage of
about 40 kV or a margin of over 137%.
 Page 209

Figure 5-28.
Riser Pole Voltage

Figure 5-29.
Cable Endpoint Voltage
 Page 210

Figure 5-30.
Cable Endpoint Voltage With Open Tie Protection

Figure 5-31.
Cable Midpoint Voltage
 Page 211
Negative Trapped Charge

If a positive lightning stroke hits a distribution line at the precise time, the system voltage is at a negative peak ,
the arrester will not go into complete conduction until the voltage impulse has compensated for the negative system voltage of -10 kV
and the 30 kV normal discharge voltage. This equivalent impulse produces a 40 kV traveling wave at the riser pole as is shown in
Figure 5-32.
This voltage will, of course, be reflected at the open point to about 70 kV and effectively reduce the margin to about 35%. While it
could be argued that this margin is considerably greater than the recommended 20%, it should be pointed out that if even some of the
previous considerations are factored in, the 20% margin will not be attained.

Figure 5-32.
Riser Pole Voltage
 Page 212

Figure 5-33.
Cable Endpoint Voltage
Quadrupling
A number of lightning researchers have noticed that while the majority of lightning strokes are monopolar, some are bipolar, as shown
in Figure 5-34. Phil Barker's paper Voltage Quadrupling on a UD Cable was the first to point out that if the timing of the bipolar wave
and length of the cable were correct, voltage quadrupling could occur.
The wave of Figure 5-34 was simulated and discharged by the riser pole arrester. Figure 5-35 illustrates the discharge characteristic of
the arrester showing the dramatic effect the nonlinearity of the arrester has on the shape of the traveling wave entering the cable. When
this wave reaches the open point, it is reflected in a somewhat more complex manner resulting in a line-to-neutral voltage of
approximately 120 kV (see Figure 5-36) which is over 25% greater than the system BIL.
 Page 213

Figure 5-34.
Bipolar Surge

Figure 5-35.
Riser Pole Voltage
 Page 214

Figure 5-36.
Cable Endpoint Voltage
Tapped Lateral
The tapped lateral is difficult to generally assess because the traveling wave pattern is so much more complex. For this example, two
taps were added to a 200 meter cable, one 400 meters in length and the other 800 meters in length, as shown in Figure 5-37.
Under normal circumstances, the riser pole arrester voltage of 30 kV, would be doubled to 60 kV at the open point. Because of the
more complex interaction of the two open points, voltage at the end of the longer 800 meter branch reached approximately 80 kV (see
Figure 5-38) which gives less margin than the recommended 20%. Voltages at the tap point were almost 50 kV while voltages at the
end of the shorter branch were as high as 60 kV (see Figures 5-39 and 5-40). While it is difficult to predict the maximum overvoltage
seen on any tapped system, it is safe to say that the typical doubling philosophy should not be applied.
 Page 215

Figure 5-37.

Figure 5-38.
Voltage at the End of the Longer Branch
 Page 216

Figure 5-39.
Tap Point Voltage

Figure 5-40.
Voltage at the End of the Shorter Branch
 Page 217
Effect of Lateral Length
For simplicity, a traveling wave on an underground cable is typically shown as a square wave or in effect, a wave having an infinite
rate of rise. When this wave reaches the open point, it immediately doubles. The riser pole arrester has no effect on preventing this
doubling, as was shown previously. While this technique simplifies the explanation of voltage doubling, it does not tell the entire story
if the cable length is relatively short.
The front of the wave in reality looks more like a ramp or ramp function. As such, it takes time for it to reach its maximum value. For
example the standard 8×20 ms wave, shown below, takes 8 microseconds to reach its crest. While this may seem like a very short time
(and it is), from a traveling wave point of view it is quite long. If the assumption is made that the speed of the traveling wave is
approximately 500 feet per microsecond then the wave can be thought of as cresting in about 4000 feet. For cables less than 2000 feet
long, then the reflected wave has already returned before the crest current is discharged at the riser pole.

Figure 5-41
The surge reflection coefficient at the open point is positive because the open point represents an infinite impedance to the wave. On
the other hand, the surge coefficient at the riser pole is negative since the riser pole arrester has a lower impedance than the surge
impedance of the cable. This negative coefficient means that reflected waves will have their signs reversed (i.e., positive reflected
waves become negative reflected waves and vice-versa). Figure 5-42, shown below, illustrates the interaction of traveling waves on
short lines. As can be seen, the incident waves enters the cable. At this point, little if any arrester operation is taking place because the
current level and consequent voltage level are
 Page 218
still very low. The incident wave is reflected and adds to the incoming wave as shown in Figure 5-41. (It should be noted that the
incoming wave is still low in magnitude). The reflected wave is now re-reflected due to the negative coefficient of the arrester and in
effect lowers the voltage on the cable because it is negative as shown in Figure 5-42. The process continues as the incident waves
increases in magnitude. The net effect, however, is that while the voltage does indeed increase above the discharge level of the arrester,
it does not double.
While some engineers maintain that due to the above phenomena, short cables (the shorter the better) do not need open tie protection,
this may not be completely correct. The 8 × 20 ms wave is not truly representative of lightning. Lightning has much higher rates of rise
and can be expected to reach a crest in one microsecond or even less. While a cable of 2000 feet or less might see some benefit with an
8 × 20 ms wave and be classified as a short line, the same cable would see the full doubling effect from a 1 × 20 ms wave and be
classified as a long line.
 Page 219

Figure 5-42.
Graphical Description of Reflection Cancellation on a Short Line
 Page 220
Simulations have shown that a cable would have to be about 200 feet or less (for a 1 × 20 ms wave) in length before it could begin to
see any benefit attributed to a short line.
Summary of Recommendations
As has been demonstrated, considerations affecting underground protective margins are not constant. To summarize, some factors
which greatly affect margins are as follows:
Effect of ground potential rise
Rate of rise on current
Lead length
BIL deterioration due to aging
Reflections (with and without open tie protection)
IR discharge changes due to aging
Bipolar waveshapes
Tapped laterals
Negative trapped charge.
While it is not possible to elaborate on these items for all situations, it is the author's opinion after years of analysis that the following
general rules, shown in Table 5-9, for both straight and tapped laterals should be adhered to for proper underground protection.
Line Protection
Line Insulation
The primary insulation for distribution lines is air. Air, rather than other materials such as rubber or other solid dielectrics, is used
because it is a self-healing dielectric. If a dielectric breakdown occurs (flashover) and the arc is de-energized, with a self-healing
dielectric, the insulation is restored and the line can be re-energized without the insulation having to be repaired. With solid dielectrics,
once failure occurs, the insulation must be repaired or replaced before the line can be re-energized.
 Page 221
Table 5-9. Recommended Arrester Locations
Voltage Feeder Configuration Arrester Locations
15 kV Radial Riser Pole
Open Tie
25 kV Radial Riser Pole
Open Tie
35 kV Radial Riser Pole
Open Tie
Midpoint (near OT)
15 kV Tapped Lateral Riser Pole
Tap Point
25 kV Tapped Lateral Riser Pole
Tap Point
35 kV Tapped Lateral Riser Pole
Tap Point
All Open Ties

Lines must be supported in air and the supporting structures are usually electrically weaker insulators than the air. To maximize this
insulation strength, porcelain or polymer insulators are used, often in conjunction with wood.
The voltage level at which flashover will occur on distribution structures is a function of the basic insulation level (BIL) of the
structures. A direct lightning strike to a line will always strike the line not more than one-half span from a structure. For distribution
lines, this means that the strike point will generally not be more than 200 feet from a structure. A flashover will generally occur at the
structure closest to the strike. Flashover of multiple structures on both sides of the strike point are also common.
While accurate BIL for structures is obtained by testing the structure with a surge generator, estimates of BIL can be made. For
distribution lines, wet flashover values for negative impulses are used and added directly to the impulse flashover value for wood. BIL
for wood varies by type of wood but generally can be assumed to be about 100 kV per foot dry. Wet value is approximately 75 kV per
foot. For a structure with a
 Page 222
100 kV BIL insulator and a 3 foot spacing of wood, the BIL would be approximately 325 kV.
The primary concern when designing structures is to attain a 300 kV or greater BIL level to ensure that only direct strikes to the line
will cause a flashover. This is normally accomplished by utilizing the wood of the structure, itself. A standard 7.6/13.2 kV structure,
which is shown in Figure 5-43, has a BIL of approximately 300 kV. This structure uses an 8 foot crossarm with wood crossarm braces;
these wood braces are important in attaining the 300 kV BIL (note that if steel braces are used, part of the wood insulation is short
circuited, resulting in a much lower BIL). The goal in designing distribution structures should be to equal or exceed the BIL of this
structure.

Figure 5-43.
3q Pole Head
Figure 5-44 shows a typical armless construction found in many newer areas. The design uses 20 kV insulators and has a BIL of
approximately 150 kV.
 Page 223

Figure 5-44.
3q Pole Head with 3q Tap Armless Construction 20 kV
Insulators.
Types of Lightning Protection
No Protection. This type of protection is identified as one in which the line has exposed phase wires with the neutral, if used, located
beneath the phase conductors. No shield wires or arresters are installed for line protection. As such, this type of line represents the
worst case. Subtransmission lines with long runs may approach this type of construction and lack of lightning protection.
 Page 224
Distribution lines of this type will have equipments such as transformers, capacitor banks, reclosers, and regulators located along the
line. This equipment will usually be protected by lightning arresters. These arresters will protect the equipment adequately, but will
have only a modest effect on the flashover rate of the line itself. A distribution line having these equipments with associated lightning
arresters would have a lightning outage rate approaching that for one of the types of lightning protection to be described below. The
degree of protection would be proportional to the amount of equipment and the number of arresters.
Shield Wire. This type of protection is widespread in the transmission line area. On distribution lines it is used occasionally. It can be
the system neutral conductor if this is located overhead and if it is grounded frequently along the line. When properly installed, the
overhead ground wire will intercept nearly all lightning strokes which will terminate on the line. It is only effective, however, when
ground footing impedances can be maintained at a low value.
Occasionally, a stroke will terminate on a phase conductor because of a shielding failure. The shielding failures are a function of the
height above ground and of the shielding angle. For the conductor heights expected in the distribution area, shielding failures are not
expected to be a problem provided the shielding angle is less than 40 to 50 degrees. The shielding angle is the angle (measured from
the vertical) from the ground wire to the phase conductor.
Basically, as the impulse voltage between the two conductors (ground wire and nearest phase conductor) enters the breakdown range, a
sheet of prebreakdown currents erupts between the two conductors. The total current reaches thousands of amperes if the conductors
are of sufficient length, and causes a pronounced delay in breakdown by reducing the voltage somewhat in the same manner as a
lightning arrester limits voltage by drawing a heavy current. Breakdown can still take place, but because of the voltage reduction by
these prebreakdown corona currents, this breakdown may occur at 5 ms or more rather than at 1 or 2 ms which might be the classical
value. In the meantime, because of this long delay in breakdown, reflections have the opportunity to arrive from nearby grounded poles
and to reduce the voltage between conductors. A sufficient reduction in voltage within the breakdown time will prevent a flashover at
the stricken location.
 Page 225
Arrester on Top Phase. This type of protection is not considered to be effective although it is quite common. The combination of top
phase conductors and arresters act somewhat like an overhead ground wire system in which the footing resistance is increased by the
amount of the arrester resistance. Shown below is a simplified one-line of a pole having two isolated conductors. One of the conductors
(top phase) is protected by a lightning arrester.

Figure 5-45.
13.2 kV System
Now assume that a lightning stroke hits the line. This stroke will cause a voltage of the same polarity in adjacent lines due to pre-
discharge currents. The pre-discharge voltage is 186 kV per foot; so for a 5 foot spacing we would have an approximate 1,000 kV
voltage difference between these two lines. For this example, it can be assumed that the top phase surge voltage is 10,000 kV and the
conductor 2 voltage is 9000 kV or 10,000 kV - 1000 kV. Since the effective voltage difference between the lines has been reduced now
to 1000 kV due to pre-discharge, there is less chance of flashover out on the line.
When the two surges get to the pole where the top phase is protected, the arrester will spark over and reduce the top phase voltage to
approximately 40 kV for a 13.2 kV system. The lower phase however is still 9000 kV so the potential difference is now 9000 - 40 or
8960 kV making the probability of sparkover greater. Of course this situation will change with time because pre-discharge will take
place again until the
 Page 226
next arrester location where the process repeats itself at a lower energy level.
The next figure shows the same situation with arresters on the two conductors. Here it can be easily seen that when the surges arrive at
the pole both arresters will sparkover and the transmitted voltages to the next pole will be very low. This situation greatly reduces the
probability of flashover.

Figure 5-46
Arresters on All Phases. Prebreakdown currents have the effect of reducing the probability of flashover out on the span and of
increasing the voltage stress at the poles having grounds for the overhead ground wire. In addition, it is noted that the voltage at these
poles appears across the insulation of the conductors not involved in the stroke. It would seem likely, therefore, that arresters on all
phases would eliminate lightning flashovers at the poles having arresters, and result in a significant reduction in the total number of
flashovers.
With arresters installed on all phases, their ground connections would be tied together near the top of the pole and a common ground
connection installed to ground at the foot of the pole. The traveling wave voltages would be similar to those seen for the previous
figure illustrating two conductors and two arresters. The arresters, on all phases, must be located on the same pole to insure that the
voltage waves rise and fall
 Page 227
together, thus minimizing voltage stress between conductors. It is interesting to note that this scheme is immune to the effect of poor
footing resistances since the entire line will rise and fall together and the arrester voltage which is between each conductor and the top
of the pole, (approximately 40 kV) will be relatively constant. This is unlike the shield wire where a high footing resistance can cause
the shield wire to have a higher voltage than the phase wire and a backflash can occur.
Comparison of Line Protection Schemes
Today it is possible to evaluate the performance of specific line designs with the use of digital simulations. These simulations can
evaluate the effect of shield wires, arresters on each phase, spacing, rate-of-rise of current, grounding, neutral placement and other
important parameters. Figure 5-47, shown below, is a comparison of some of these parameters for a 13.2 kV system. As can be seen,
the best protective scheme is usually the use of arresters on all phases and the worst is typically having an arrester on the top phase
only. The effect of line BIL has also been evaluated and illustrates that for very effective protection, a BIL of 300 kV or above is
prudent regardless of the scheme used.
The types of construction used will be called flat, triangular, and post. Each type of construction has a probability of flashover which is
a function of the line BIL. Figures 5-48, 5-49, and 5-50 illustrate the probability of flashover as a function of line BIL and arrester
(ground) spacing.
 Page 228

Figure 5-47.
Comparison of Various Distribution Line Protective Measures
 Page 229

Figure 5-48.
Flat Construction
 Page 230

Figure 5-49.
Triangular Construction
 Page 231

Figure 5-50.
Post Construction

Induced Strikes
There is a considerable amount of evidence to suggest that many of the flashovers occuring on an overhead distribution line might be
the result of induced strikes, (i.e., nearby strikes that do not hit the line directly). The voltages caused by these strokes are considerably
lower than those for direct hits (usually less than 300 kV). The rise times are also less, making spacing of arresters less critical.
 Page 232
Protection for flashovers from induced strikes can, in some cases, become more effective by simply increasing the BIL of the line
design. This is normally done by increasing the insulator ratings, changing the position of the neutral wire, increasing the phase wire
spacing, etc. Arrester protection having spacing every quarter mile or so, while not very effective on a direct hit, can be effective for
induced strikes.
Reference
1. P.P. Barker, Voltage Quadrupling on a UD Cable, IEEE Transactions on Power Delivery, January 1990, Vol. 5 No. 1, Transmission
& Distribution Conference, April 1989.
Questions
1. What is the average duration of a lightning stroke?
2. Plot a 10 kA lightning stroke on a 10 kA fault. What does this tell you about energy content of lightning?
3. Why is the 8 × 20 ms wave not particularly representative of lightning?
4. Name 4 conditions that increase the possibility of ferroresonance.
5. What is restrike?
6. Why does current chopping create overvoltages?
7. If a silicon carbide arrester did not have a gap, how much current (approx.) would it draw at nominal voltage?
8. Why don't MOVs require gaps?
9. What is the major difference between the 3 classes of arresters?
10. Define MCOV and TOV.
11. Why are arrester ratings for delta systems higher than for 4-wire multigrounded systems of same voltage?
 Page 233
12. Why is the TOV curve virtually useless to distribution engineers?
13. Select arrester rating for 34.5 kV, 4-wire multigrounded system.
14. What is the relationship between arrester rating and MCOV?
15. Arrester rating is based on what?
16. What is insulation coordination?
17. What's one disadvantage of externally gapped arresters?
18. Do MOVs give more margin than gapped silicon carbide? Explain.
19. Calculate margin for an overhead 34.5 kV, 4-wire multigrounded system.
20. Why do margins go down as system voltage level goes up?
21. Why are margins for underground less?
22. Why don't we usually worry about doubling of voltage (due to traveling waves) on overhead systems?
23. What is the margin for a 34.5 kV URD system having only a rise pole arrester and a BIL of 125 kV?
24. Why can the arrester lead length be so detrimental to protective margins?
25. Why are protective margins much lower than we think?
26. Name 3 ways to improve URD performance?
27. Explain effect of trapped charge on URD protection?
28. What type of wave must you have to obtain quadrupling on a URD system?
29. It is well known that use of a single riser pole arrester can be very effective for short lines. What determines when a line is
considered short?
30. How many arresters should be used for 34.5 kV URD? 13.8 kV?
 Page 234

6
Distribution Reliability
Introduction
The great majority of service interruptions the customer sees are caused by problems on the distribution system. The overhead system
is primarily affected by meteorological conditions such as wind and lightning whereas the underground system, while it sees fewer
outages, is particularly concerned with service restoration time following a failure.
It is interesting to note that in recent years, with the emphasis on power quality to more sensitive loads, the meaning of system
reliability is changing. For example, the momentary interruption caused by the reclosing breaker to clear a temporary fault was only a
few years ago considered to be a minor annoyance. Today, this same operation can cause havoc in the form of computer shutdowns,
industrial process shutdowns and a multitude of blinking clocks.
Fundamentals
Calculating Reliability
Reliability is defined as the probability of a device performing its purpose adequately for the period of time intended under the
operating conditions encountered. This definition implies that reliability is the probability that the device will not fail to perform as
required for a certain length of time. The problem with this definition when it is applied to utility service is that it does not consider the
duration of the failure. It is conceivable that a bare bones distribution design (i.e., one with no switches, fuses, etc.) could have a very
high reliability (by this definition, anyway) but would, of course, have unacceptable repair time.
Reliability of a distribution system is really evaluated in terms of outage rate and outage duration. The basic formula for evaluating a
radial utility distribution system is as follows:
 Page 235
Components in Series

Figure 6-1.
Components in Series
A system consisting of two components in series with outage rates l1 and l2 and repair times r1 and r2 respectively has the following
reliability indices.
System outage rate,
lS = l1 + l2 6-1
System average outage duration,

6-2
System average total outage time,
US = lSrS 6-3
Components in Parallel

Figure 6-2.
Components in Parallel
 Page 236
A system consisting of two components in parallel has the following reliability indices.
System outage rate,
lP = l1l2 (r1 + r2) 6-4
System average outage duration,

6-5
System average total outage time,
UP = lPrP 6-6
Example No. 1

Figure 6-3.
Components in Series
Calculate:
1. Customer outages per year
2. Customer average outage duration
3. Customer average total outage time
4. Conclusions
 Page 237
Solution No. 1
l1 = 1.0 r1 = 5 6-7
l2 = .005 r2 = 4
lS = 1.0 + .005 = 1.005

6-8
US = lS rS = 1.005 (4.995) = 5.02
Example No. 2

Figure 6-4.
Components in Parallel
 Page 238
Solution No. 2
l1 = l2 = 1.005 6-9
r1 = r2 = 4.995
lP = l1l2 (r1 + r2)

6-10

6-11
Reliability Indices
There are two methods of calculation of reliability indices in the U.S. The majority of companies calculate indices based on the number
of customers per outage and the duration of the outage. There is another group that calculates indices based on the amount of load that
is lost. They do not keep track of the actual load that is lost, but rather use the peak load value of any given section in calculating their
indices. The companies that base their reliability indices on lost load are doing so as a precursor to the customer based index system. At
the present time, their data bases are not as extensive as the other companies.
Indices have been defined by various groups such as the IEEE (Institute of Electrical and Electronics Engineers), EEI (Edison Electric
Institute), EPRI (Electric Power Research Institute) and CEA (Canadian Electric Association) as follows:
 Page 239
SAIFI - System Average Interruption Frequency Index

6-12
SAIDI - System Average Interruption Duration Index

6-13
CAIFI - Customer Average Interruption Frequency Index

6-14
CAIDI - Customer Average Interruption Duration Index

6-15
ASAI - Average Service Availability Index

6-16
ATPII - Average Time until Power Restored

6-17
 Page 240
CMPII - Customer Minutes per Interruption

6-18
The load based indices were developed by the PEA (Pennsylvania Electric Association) and are defined as shown below in Equations
6-19 through 6-22.
These indices are based on the peak connected load during an interruption rather than on the number of customers interrupted. These
indices are used due to the limitations of the companies data base. The companies that are using this system plan to upgrade their
systems to customer based indices when their customer information systems track the number of customers interrupted. Typical
numbers for customer based indices are shown in Table 6-1.
ASIDI - Average System Interruption Duration Index

6-19
ASIFI - Average System Interruption Frequency Index

6-20
ACIDI - Average Circuit Interruption Duration Index

6-21
ACIFI - Average Circuit Interruption Frequency Index

6-22
 Page 241
Table 6-1. Customer Based Indices
SAIDI SAIFI CAIDI ASAI
95.9 min/yr 1.18 int/yr 76.93 min/yr .999375 int/yr

Survey Results
In 1990, a survey was conducted of utility distribution reliability practices throughout the United States. Among the most pertinent
findings were that the most commonly used indices are ASAI, SAIFI, SAIDI, and CAIDI as shown in Figure 6-5. In less technical
terms, most utilities are interested in the following:
Average Customer Minutes Outage per Year. This is the average cumulative amount of time a given customer would expect to be
without service in a one year period (CMO or SAIDI). A typical value for SAIDI is about 100 minutes.
Average Duration of a Given Outage. Some outages can be very short, others quite long. This number is simply the average duration
(CAIDI). A typical number for CAIDI is about 80 minutes.
Average Annual Number of Outages. This is the average number of times a given customer can expect an outage in a year (SAIFI).
The average customer in the U.S. can expect to see about 1.20 outages per year (SAIFI = 1.20).
Availability of Service. This is the ratio of the total number of customer hours that service was available to the total customer hours
demanded (ASAI). Although rarely used in discussion of distribution reliability, it is generally calculated for comparison to other parts
of the utility system.
Another important survey finding was that most utilities do not classify an interruption as an outage until its duration exceeds about 5
minutes (see Figure 6-6). This of course means that temporary faults and their associated breaker or reclosing operations are not
considered outages.
 Page 242

Figure 6-5.
Most Commonly Used Indices
 Page 243

Figure 6-6.
Typical Durations of Momentary Outages
Causes of Outages
It is very difficult to generalize on the causes of outages since they are somewhat particular to the geography of the region and the
practice of the utility. It is easy to assume, for example, that utilities in high lightning areas will see a large number of outages caused
by lightning. This, however, may not be true since many of these same utilities use overvoltage protection practices which greatly
mitigate the effects of lightning. Figure 6-7, shown below, illustrates results from a recent study. As can be seen, a large percentage of
outages cannot be identified. The most common reported causes of outages are lightning and tree contact. One utility in this same study
reported that virtually all their outages were animal related.
 Page 244

Figure 6-7.
Failure Rates
The largest problem in performing a reliability study is determining failure rates. Because of the newness of many system designs
being studied and their related equipment, there is many times very little data available. Good failure rate data takes many years to
develop. For example, it may take 10 or more years before a problem such as corrosion affects system operation. In coming up with
failure rates to be used in a study, a subjective judgement based on the data available, along with considerable advice from utility
engineers, in the specific area to be studied should be used. The failure rates shown in the table are thought to be quite representative of
what a typical utility would experience.
 Page 245
Table 6-2
Circuit Element Failure Rate
Primary Underground Cable 0.03
Secondary Underground Cable 0.11
Service Underground Cable 0.11
Cable Terminations 0.002
Distribution Transformer 0.003
Transformer Fuse 0.005

The second problem with regard to failure rate data is to determine how a failure rate is allocated to its failure modes. For example, a
primary cable can have a failure caused by the cable opening or a cable shorting to ground. Generally, all modes of failure are included
when a failure rate is given. Sometimes, only the predominant mode is included. For primary cable and terminations, the failure mode
for the vast majority of cases is a short to ground. On the secondary cable, past information shows that if the cable shorts, it usually
burns itself clear (thus creating an open); thus, the dominant failure mode for secondary cable is the open condition. Table 6-3 shows
the failure rates with both the predominant mode of failure and the secondary mode.
 Page 246
Table 6-3. Standard Failure Rates
Total Failure Failure Failure
Circuit Element
Rate/yr. Mode Rate/yr.
A. Secondary U.G. .11 Open .099
Cable
Short .011
B. Dist. .003 Open 0
Transformer
Short .003
C. Loop Switch .005 Won't 0
Close
Short .005
D. Service U.G. .11 Open .099
Cable
Short .011
E. Cable .002 Open 0
Terminations
Short .002
F. Primary U.G. .03 Open 0
Cable
Short .03
G. Transformer .005 Open .005
Fuse
Won't .0
Blow

Power Quality
Fault Selective Feeder Relaying (FSR)
The feeder relay is the primary device in the protection of the feeder main and also the protection of the lateral fuse during temporary
fault conditions. If a feeder is underground, it is usually assumed that all faults are permanent in nature. On the more predominant
overhead feeder, such is not the case and what's called a reclosing relay is used on the feeder breaker.
 Page 247
The reclosing relay recloses its associated feeder breaker at preset intervals after the breaker has been tripped by overcurrent relays.
Survey results indicate that approximately 70 percent of the faults on overhead lines are non-persistent. Little or no physical damage
results if these faults are promptly cleared by the operation of relays and circuit breakers. Reclosing the feeder breaker restores the
feeder to service with a minimum of outage time.
If any reclosure of the breaker is successful, the reclosing relay resets to its normal position. However, if the fault is persistent, the
reclosing relay recloses the breaker a preset number of times and then goes to the lockout position.
The reclosing relay can provide an immediate initial reclosure plus three time-delay reclosures. The immediate initial reclosure and/or
one or more of the time-delay reclosures can be made inoperative as required. The intervals between time-delay reclosures are
independently adjustable.
The primary advantage of immediate initial reclosing is that service is restored so quickly for the majority of interruptions that the
customer does not realize that service has been interrupted. The primary objection is that certain industrial customers cannot live with
immediate initial reclosing.
The majority of utilities use a three-shot reclosing cycle with either three time-delay reclosures or an immediate initial reclosure
followed by two time-delay reclosures. In general, the interval between reclosures is 15 seconds or longer, with the intervals
progressively increasing (e.g., a 153045-second cycle), giving an overall time of 90 seconds.
Fault selective feeder relaying allows the feeder breaker to clear non-persistent faults on the entire feeder, even beyond sectionalizing
or branch fuses, without blowing the fuses. (See Figure 6-8). In the event of a persistent fault beyond a fuse, the fuse will blow to
isolate the faulty section. Operating engineers report reductions of 65 to 85 percent in fuse blowing on non-persistent faults through the
use of this method of relaying.
 Page 248

Figure 6-8.
Distribution Feeder with Automatic Reclosers and Fuse Cutouts
Blocking the Instantaneous
The biggest drawback to feeder selective relaying is that the entire feeder sees an interruption every time a lateral experiences a
temporary fault. With today's sensitive loads, especially in high-tech commercial areas, the high number of interruptions caused by
successful reclosing operations to clear temporary faults can present more of a problem than fewer interruptions of much longer
duration. It is for this reason that some utilities have decided to block the instantaneous trip and NOT save the fuse for temporary faults
on the lateral. These, of course, now become permanent outages.
Many public service commissions (PSC) across the U.S. require utilities to report reliability indices. At the present time, most PSCs
require information about permanent but not momentary outages. While changing the protection from FSR to not saving the fuse has
the advantages just stated, the utility should be cognizant of the fact that by blocking the instantaneous trip, they will change their
indices. Allowing
 Page 249
the fuse to blow on temporary faults will result in permanent outages for that section of line. Thus, this change will ALWAYS make
their reported indices look worse to the public utilities commission. Even the utilities that calculate indices but do not report them
calculate the indices based on permanent outages.
Figure 6-9 illustrates this point. As can be seen, while the (unreported) momentaries go down less than 25%, the reported outages go
up over 50%.

Figure 6-9.
Example No. 3
Assume you are the only customer on a substation that has the failure rate and outage duration characteristics shown in Figure 6-10.
The PSC asks you to assess the effect adding a recloser to the mid-point of the feeder (See Figure 6-11). Calculate SAIFI, SAIDI, and
CAIDI before and after the system change.
 Page 250

Figure 6-10.
System Without Recloser
Solution No. 3
Before

6-23

6-24
 Page 251

6-25

Figure 6-11.
System With Recloser
After

6-26
 Page 252

Factors Affecting Reliability

Higher Voltages
The decision of a utility to standardize on a higher distribution voltage level is a very difficult one. While much of the decision is based
on load density, availability of right-of-way, losses and substation placement, a major item which must be considered is reliability.
Higher voltage levels, because they increase exposure, decrease reliability. It can be argued that by increasing sectionalizing capability,
higher voltage levels can be made as reliable as the lower voltages. This is easier said than done as is illustrated below:
Suppose a utility operating at 13.2 kV decides to convert to 34.5 kV and extend its feeder by 260% (34.5/13.2 × 100%) as shown
below. Assuming the 13.2 kV system served 1000 customers, it can be assumed that with the same load density (and lateral length) the
34.5 kV system can serve 2600 customers.

Figure 6-12
Assume:

CMO (on a = feeder failure rate · repair time · # of

feeder) customers
=f·m·r·n
 Page 253
where f = feeder failure rate/mile
m = miles
r = repair time (minutes)
n = number of customers
and that the feeder failure rate is .1/miles/year and the repair time is 120 minutes. The CMO's for the two designs can be calculated as:
@ 13.2
CMO = .1 · 5 · 120 · 1000 = 60,000 minutes/year

6-27
@ 34.5
CMO = .1 · 13.1 · 120 · 2600 = 408,720 minutes/year 6-28
=> CMO (average) = 157.2 minutes
If we attempt to make 34.5 kV more reliable by adding a midpoint switch, the CMO is calculated as follows:
CMO = .1 · 66 · 120 · 2600 + .1 · 6.6 · 120 · 1300
ie., CMO (feeder) = 205,920 + 102,960 = 308,880 customer minutes

6-29
Dividing the feeder into 3 parts; we calculate feeder CMO as follows:
 Page 254

6-30
= 136,240 + 90,824.9 + 45,415.1
= 272,480 minutes/year

\ since even 3 sections of 34.5 kV cannot approach the 60 minute criteria of the 13.2 kV, it could be concluded that similar reliability
can be difficult to achieve.
System Design
A three-phase feeder main can be fairly short, on the order of a mile or two, or they can be as long as 30 miles. Voltage levels can be as
high as 34.5 kV, with the most common voltages being in the 15 kV class. While most of the 3-phase mains are overhead, much of the
new construction, particularly the single-phase lateral construction, is being put underground. Underground systems have the
advantage of immunity from certain types of temporary fault conditions like wind, direct lightning strikes, animals, etc. Permanent
faults, on the other hand, are much more difficult to locate and repair and have been the subject of much concern in recent years. The
more complex and costly distribution designs shown in Chapter 1 have certain advantages in regard to reliability. Each of the five
systems previously described can be evaluated in terms of reliability for traditional loads as shown below. As can be seen, in Figure 6-
13, the primary loop scheme which is used for most residential areas in the United States will see mostly two hour outages (one every
ten years). On the other hand, a spot network, which many utilities use in downtown areas and some industrial use for important
processes, will see only two outages (one 12 hour and one 6 hour) for every 1000 years.
 Page 255

Figure 6-13.
Reliability Profiles
Automation
It is an interesting paradox that when utilities are asked what is the most important advantage of distribution automation they answer
reliability, while they also generally agree that their customers will not pay more for better reliability. Even though the inconvenience
to the average customer during an outage is generally small, the cost to the utility can occasionally be quite high. Interruption cost
assessment is a complex and often subjective task. Many of the effects resulting directly from interruptions are relatively easy to assign
a dollar value. Other direct impacts such as reduction of manpower efficiency, fear, injury, and loss of life are difficult to quantify.
Indirect effects such as civil disorder during blackouts or businesses moving to areas with higher reliability tend to be even more
difficult to predict and evaluate. Many utilities do not feel than an outage costs them any significant amount of money.
A review of the literature was undertaken by Roy Billinton of the University of Saskatchewan and presented in his paper Interruption
Cost Methodology and Results - A Canadian Residential Survey. A plot summarizing the results is shown below.
 Page 256

Figure 6-14.
Residential Interruption Cost Estimates
Using these curves, and if we assume that a typical feeder outage lasts approximately 2 hours, we might conclude that the average
interruption cost per kW is approximately $1.50. Assuming the typical distribution feeder shown below, we can begin to assess the cost
per year of outages.
If we assume that this 13.8 kV feeder is approximately 7 miles long, and that the connected load is at least twice the diversified peak or
approximately 20 MVA, we can begin to evaluate the value of increased reliability. Faults per year on the feeder would be calculated
by multiplying the failure rate of .15 faults/mile/year by the length.
 Page 257
Consequently, the average number of faults per year on a feeder would be about 1. If we further assume that automation allows the
fault to be isolated to 25% of the feeder main (assuming 3 sectionalizing points), a 75% reduction in outages can be recognized.

Figure 6-15.
Typical Distribution Feeder
Substation
The total savings per year per feeder based on the previous assumptions would then be .75 × 20,000 × $1.50 or $22,500 per feeder, per
year. This number is apparently reflective of some utilities' assessments of liability and indicates the strong possibility that many of
these utilities might justify automation on the basis of reliability only.
If, on the other hand, a utility is concerned only with the lost revenue costs because it believes that is its only cost, then the value of
increased reliability due to automation might be calculated as follows:
Assume:
 Page 258
Load factor = .5
1 fault/feeder/year
75% reduction in outage time due to automation [average duration = 2 hours]
10 cents per kWhr
Average MVA per feeder = 10 × .5 = 5.0 MVA/feeder
Savings = 5000 KVA × $.10/KWHR × 2 × .75 = $750.00/year/feeder
A savings of only $750.00 per year could be considered low and doesn't even take into account the fact that some of the projected
revenue would not have been lost anyway (e.g., electric heat). The actual increased revenue is probably reduced to less than half or
about $375/feeder/year.
System Modifications
There are several components that have successfully been used over the years to increase distribution reliability. A brief description of
some of them is as follows:
Arresters. The purpose of lightning arresters on distribution systems is generally to protect equipment against failure due to lightning
strokes (switching surges are usually too low to be a problem on a distribution system). Arresters can also be used to protect the
distribution line from flashovers, a practice called line protection. This practice can have a major effect in reducing the number of
momentaries and consequently increasing power quality. Many utilities, in areas of high lightning activity use line protection. The
problem with this concept is cost. In order for line protection to be very effective, arresters must be placed on all three phases of every
pole or every other pole. Putting an arrester only on the top phase is not effective although some utilities subscribe to this practice.
Table 6-4 illustrates the effectiveness of line protection practices.
 Page 259
Table 6-4. Line Comparison
Ground
Resistance Percentage Flashovers per
Design Protection (ohms) Flashovers 100 miles a year
None 100 100.0 13.73
Arresters on All
100 2.4 0.33
Phases
25 28.5 3.91
Arresters on
Every Other
CrossarmPole 100 24.8 3.40
250 27.5 3.78
25 11.1 1.52
Shield Wire 100 39.4 5.41
250 61.3 8.42

Fuses. All reliability studies have shown conclusively that laterals should be fused. The generally accepted rule of thumb is that fuses
larger than a 25K or a 15T should be used to avoid excessive nuisance fuse blowings. Nuisance blowing of transformer fuses can be
the result of lightning or multiple inrush due to reclosing.
Switches. The addition of switches to a distribution circuit does, in general, increase reliability by decreasing the duration of the outage
of many to the customers on the feeder. It is interesting to note, however, that the effectiveness of a switch is very much dependent on
the length of the section being switched, switching practices, switch failure rate, etc. It has been found in some studies that the effect of
a switch on system reliability can be very marginal and even have a negative effect. Table 6-5, shown below, illustrates the effect that
sectionalizing has on a 34.5 kV, 19 mile feeder. As can be seen, without any sectionalizing the average customer minutes outage is
approximately 51 minutes a year. With one switch (two sections), the average customer minutes outage is reduced to about 38 minutes
or about 25%. To reduce the outage time another 25% will take an additional 3 switches.
 Page 260
Table 6-5. Effect of Sectionalizing
THE FEEDER LENGTH IS: 19
THE AVERAGE CUSTOMERS MINUTES OUTAGE FOR 12 KV
IS: 17.5959
THE AVERAGE CUSTOMERS MINUTES OUTAGE FOR 35 KV
IS: 51.3
35 KV SYSTEM OUTAGE IMPROVEMENT
NO. SECTIONS OUTAGE AVG.
2 38.475
3 34.2
4 32.0625
5 30.78
6 29.925
7 29.31428
8 28.85625
9 28.5
10 28.215
11 27.98182
12 27.7875
13 27.62308
14 27.48214
15 27.36
16 27.25312
17 27.15882
18 27.075
19 27
20 26.9325

Automatic vs. Manual. The advantages of having automatic switching as opposed to manual switching is also dependent on a number
of factors including crew size, crew procedures, number of sections, feeder length, etc. Figure 6-16, shown below, shows the effect of
automatic switching for a 10 mile line. As can be seen, large numbers of automatic switches can prove quite effective (if the reliability
of the switch and communication schemes are neglected).
 Page 261

Figure 6-16.
Average Customer Minutes Outage vs. Number of
Sections - 10 Mile Line
Underground Reliability
The reliability of an underground system is generally felt to be better than its overhead counterpart. Most distribution engineers would
agree that while the frequency of failure is less for underground than overhead because of the lack of influence of most of the elements
like ice, wind, cars, etc., the duration of an outage can be longer since the failure tends to be more difficult to locate and repair.
The following is a computer analysis of an underground system being served by various voltage levels. As will be seen, the reliability
of an underground system is highly dependent on the system configuration, voltage (size) and the failure rates associated with the
components.
Failure Rates. Equipment failure rates, as discussed earlier, are extremely important in any reliability study. When evaluating the effect
that primary system voltage level has on failure rate the data becomes somewhat more difficult to interpret. While many utilities have a
general feeling that the failure rate of higher voltage underground equipment
 Page 262
may be somewhat higher, few have quantified the data. Several utilities, who have had particular problems involving customer
reliability have indeed quantified the failure rates at their various voltage levels and have shown failure rates to be considerably
different than those shown in Tables 6-2 and 6-3. Several of these have also shown underground equipment failure rates to be many
times higher at the higher voltage levels (25 kV or 35 kV) than at 15 kV. The baseline failure rates used for this analysis are shown in
Table 6-6.
Table 6-6
Failure Rate (per
Circuit Element year)
A. Primary .07
underground
cable (per mi.) .006
B. Elbow disconnect .001
C. Splice .003
D. Transformer .05
E. Switch .005

Load Area Optimization. In developing reliability indices for different voltage levels, it is necessary to know the load geometry of the
systems to be studied. No utility feeder main or lateral is the same and certainly great differences exist between utilities from the
standpoint of voltage levels and loading which greatly affect the areas that can be covered by a given main/lateral design.
When comparing feeder areas operating at different voltage levels, it is necessary to treat the highly complex subject of distribution
system planning in a generalized manner by making certain simplifying assumptions. In such a generalized study, the feeder load area
is assumed to have uniform load density and a regular geometric shape, such as a rectangle or a triangle. While the results of
generalized studies are not always directly applicable to a specific problem, they serve to illustrate the fundamental relations between
percent voltage drop, losses, load density, circuit voltage, etc. An example of an area analysis for 3 different voltage levels is shown in
Figure 6-17. As can be seen, the optimal way to increase system load at the higher voltages is to increase the length of both the feeder
main and the laterals.
 Page 263

Figure 6-17.
Load Area Optimization at Various Voltage Levels
Computer Reliability Analysis. Reliability analysis of the optimized load areas for different voltage levels was performed in this
example using two computer programs. The first analyzed the feeder main and the second the feeder lateral. Inputs to both are similar
in that restoration procedures and times, protective configuration, line length, manual or automatic operation, etc. must be specified.
The three measures of reliability used by these programs was:
1. Annual interruption time per average customer,
2. Interruption time to last customer restored to service,
3. Annual number of interruptions per average customer.
Figure 6-18 represents the main circuit configuration used in the first program. It contains N - 1 automatic interrupters which divides
the main into N equal-length sections. Each section contains M -1 manual sectionalizing points which divide the section into M is
equal to or greater than 2. A circuit breaker is located at the substation where the main circuit goes underground. At the remote end of
the main a normally open interrupter is located at an underground tie point to an adjacent similar circuit. The tie may be operated either
automatically or manually. Single-phase underground laterals are spaced at regular
 Page 264
intervals along the entire length of the main. These laterals are connected to the main at the manual sectionalizing points in a manner
which permits opening the main circuit on either side of the lateral connection.

Figure 6-18.
Main Current
Figure 6-19 shows the lateral circuit configuration used by the second program. Although it is not now common practice to use
automatic interrupters out on the lateral, the program has provisions for m automatic interrupters, including the one at the connection
point to the main.

Figure 6-19.
Lateral Circuit
For the purpose of this example, all reliability calculations were made assuming an underground system, use of fault indicators and
fuses at all lateral taps. Both programs use work functions or restoration procedures, that can be tailored to a particular utility's
experience and practice. A sample output showing reliability indices of feeder mains using standard failure rates is shown below. For
example, if an
 Page 265
optimized 23KV system had a feeder length of 1.81 miles, the system could expect an average customer minutes outage time for the
main of about 12.75 minutes, a time to last customer restoration of 172 minutes and .181 interruptions per year.
Inputs and outputs for lateral reliability are somewhat similar to those performed on the feeder main. However, since calculations
performed here had no intermediate protective devices in the lateral, the lateral reliability is simply a function of length. For example,
in Figure 6-20 shown below, if a utility had a lateral length of 1.32 (as in the case of 23KV), the average customer minutes outage
would be approximately 15.3 minutes.
AVAILABILITY CALCULATION FOR 23 KV
MAIN FEEDER WITH AUTOMATIC TIE
WORK FUNCTIONS
TIME TO ARRIVE AT DEVICE (A):? 20
TIME TO CALL REMOTE LOCATION (B):? 20
TRUCK IN/OUT TIME (C):? 10
TIME TO OPERATE SWITCH (D):? 5
RESTORE TIME (LOCAL) (E):? 10
TIME TO READ FCI (F):? 10
TRAVEL RATE mi./min. (G):? .25
TERMINOLOGY
NUMBER OF CUSTOMERS PER UNIT AREA (K):? 800
MAIN CIRCUIT LENGTH (X):? 1.81
LATERAL CIRCUIT LENGTH (Y):? 1.32
OUTAGES PER UNIT LENGTH (J):? .1
NUMBER OF SECTIONS (N):? 1
CONTROL CONSTANT (S):? 0
SECTIONS AVERAGE TIME TO LAST NUMBER OF
INT. TIME CUSTOMER INTERRUPTIONS
1 12.75 171.66 .181
2 6.38 168.33 .091
3 4.23 161.95 .060

Figure 6-20. Main Feeder Reliability Computer Output

 Page 266
LATERAL RELIABILITY PROGRAM FOR
UNDERGROUND DISTRIBUTION SYSTEMS
***** WORK FUNCTIONS *****
CASE NUMBER:? 1
TIME TO ARRIVE AT LATERAL (A):? 20
TRUCK IN/OUT TIME (C) :? 10
TIME TO OPERATE SWITCH (D) :? 5
RESTORE TIME (LOCAL) (E) :? 10
TIME TO READ FCI (F) :? 10
TRAVEL RATE mi./min. (G) :? .25
TIME TO OPERATE CONNECTOR (H) :? 10
NUMBER OF CUSTOMERS PER UNIT AREA (K) :? 800
OUTAGES PER UNIT LENGTH (L) :? .16
NUMBER OF SECTIONS (N) :? 1
NUMBER OF LAT. PER SEC. POINT (R) :? 2
CONTROL CONSTANT FOR TIE (0 for automatic) (S) :? 0
CONTROL CONSTANT (0 for radial) (V) :? 1
DISTANCE BETWEEN LATERALS (Z) :? .08
LATERAL LENGTH (Y) :? .1
LENGTH AVERAGE TIME TO
OF CUST. RESTORE
LATERAL MINUTES LAST NUMBER OF
OUTAGE CUSTOMER INTERRUPTIONS
.1 1.28144 110.6 .0192
.2 2.39648 111.2 .0352
.3 3.52112 111.8 .0512
.4 4.65536 112.4 .0672
.5 5.7992 113 .0832
.6 6.952639 113.6 .0992
.7000001 8.115679 114.2 .1152
.8000001 9.288321 114.8 .1312
.9000001 10.47056 115.4 .1472
1 11.6624 116 .1632
1.1 12.86384 116.6 .1792
1.2 14.07488 117.2 .1952
1.3 15.29552 117.8 .2112
1.4 16.52576 118.4 .2272
1.5 17.7656 119 .2432
1.6 19.01504 119.6 .2592
1.7 20.27408 120.2 .2752001
1.8 21.54272 120.8 .2912001
1.9 22.82097 121.4 .3072001
2 24.10881 122 .3232
2.1 25.40624 122.6 .3392
2.2 26.71328 123.2 .3552
2.3 28.02992 123.8 .3712
2.4 29.35616 124.4 .3872
2.5 30.692 125 .4032
2.6 32.03744 125.6 .4192
2.7 33.39247 126.2 .4351999
2.8 34.75712 126.8 .4511999
2.9 36.13135 127.4 .4671999
2.999999 37.5152 128 .4831999

Figure 6-21. Lateral Reliability Computer Output

 Page 267
Utilizing the optimized system configuration for 13.8, 23, and 34.5 kV systems, both programs were run and the systems total outage
time (main and lateral) computed. A summary of the outage times for the various voltage levels is shown above. As can be seen in
Figure 6-22, the inherent increase in customer minutes outage time by going from a 13.8 kV system to a 34.5 kV system is
approximately 62%. However, the customer minutes outage time of the 34.5 kV system is still considered very low when compared to
the overhead systems.

Figure 6-22.
Voltage vs. Outage Time
Some utilities have indicated that their failure rates with higher voltage components are even higher than those shown in Table 6-4. If
the assumption is made that the failure rate of 23 kV equipment is twice that of 13.8 kV and that 34.5 kV is three times the failure rate
of 13.8 kV equipment then a graph of the outage times vs. voltage (see Figure 6-23) would be considerably more dramatic. As can be
seen, the average customer minutes outage now increases to 105.5 minutes or almost 300%. This of course means that a utility,
unsuspectingly facing this high failure rate, may have to take drastic corrective action.
 Page 268

Figure 6-23.
Voltage vs. Outage for Higher Failure Rates

Reference
1. G. Wacker, E. Wojczynski, and R. Billinton, Interruption Cost Methodology and Results - A Canadian Residential Survey, IEEE
Trans. on Power Apparatus and Systems, Oct. 1983, pp. 33853392.
 Page 269

Questions
1. Average Customer Minutes Outage is defined by which indices?
2. Customers seeing an average frequency of interruption of 2 per year should consider themselves fortunate. (true or false?)
3. The leading cause of faults is __.
4. Transformers have a high failure rate. (true or false?)
5. Blocking the instantaneous improves power quality ________ (always, sometimes, never).
6. Higher voltage systems are usually less reliable. Explain.
7. Primary selective systems eliminate the blinking clock syndrome. (true or false?)
8. Laterals should only be fused if coordination with the feeder breaker (or recloser) is possible. (true or false?)
9. One way to decrease the number of momentary outages is to _.
10. Adding switches is a good way to increase reliability. (true or false?)
11. Blocking the instantaneous hurts reliability. Explain.
12. What are the advantages and disadvantages of 34.5 kV?
13. Explain the advantages and disadvantages of underground distribution.
 Page 270

7
Power Quality Fundamentals
Introduction
Good power quality is not easy to define because its measure depends upon the needs of the equipment which it is supplying; what is
good power quality for a refrigerator motor may not be good enough for a personal computer. For example, a short (momentary)
outage would not noticeably affect equipment such as motors or lights, but could cause a major nuisance to digital clocks or VCRs.
Definitions
Probably no other area in power quality has caused more confusion than that of definitions. Definitions have varied from author to
author. The biggest difference between definitions seems to exist between electronics engineers and power engineers. The definitions
we will follow in this chapter are shown in Figure 7-1 and defined as follows:
Outages (Interruptions). An outage is a complete loss of voltage usually lasting from as short as 30 cycles up to several hours, or in
some cases even days. Outages are usually caused by the fault induced operation of circuit breakers or fuses. Some of these
interruptions might be classified as permanent, while others might be classified as temporary (momentaries).
Surges (Lightning or Switching Surges). A surge is a transient voltage or current which can have extremely short duration and high
magnitude. Typically, surges are caused by switching operations or by lightning. Surges can be generated by customers through the
switching of their own loads, or they may be caused by utility switching operations (capacitors, breakers, etc.). Surges have always
existed on power systems, but it is
 Page 271
only in recent years, with the advent of extensive VCR and PC use, that they have received attention.

Figure 7-1.
Typical Voltage Disturbances
Undervoltage (Voltage Drop). A customer who experiences a long duration (several seconds or longer) service or utilization voltage
which is less than the proper nominal operating low voltage limit can be considered to be experiencing an undervoltage situation. (The
ANSI Range [A] service voltage and utilization low voltage limits are 114 volts and 110 volts respectively.) Two examples of factors
which may cause such a condition are: 1) overloaded or poor house wiring, and 2) poor connections and/or voltage drop on the utility
system.
Harmonics. These are the nonfundamental frequency components of a distorted 60 Hz waveform. They have frequencies which are
integral multiples of the 60 Hz fundamental frequency. Harmonics are not generally produced by the utility, but rather by the
customer's equipment. For example, a large nonlinear industrial load may produce harmonics which, if they are of sufficient
magnitude, can travel back through the power system and affect other customers.
Voltage Sags. A momentary voltage dip that lasts for a few seconds or less is classified as a voltage sag. Voltage sags may be caused
by faults
 Page 272
on the transmission or distribution system, or by the switching of loads with large amounts of initial starting/inrush current (motors,
transformers, large dc power supplies). Voltage sags may be sufficiently severe, especially in the case of faults, to cause sensitive loads
(computers, VCRs, clocks, etc.) to reset.
Voltage Swell. When a fault occurs on one phase of a 3-phase, 4-wire system, the other two phases rise in voltage relative to ground
(about 20%). This steady state rise in voltage is referred to as a swell. Voltage swells usually have durations of several seconds or less,
but can last as long as a minute or so.
Overvoltage. An overvoltage is classified as any steady state voltage (several seconds or longer) which is above the ANSI Standards'
upper service voltage limit of 126 volts at the customer's meter. Overvoltages usually occur as a result of improper regulation practices
(misadjustments of regulators and capacitors).
Voltage Quality
Voltage Drop
Almost all equipment connected to a utility system is designed to be used at a certain definite voltage. It is not practical, however, to
serve every customer on a power distribution system with a constant voltage corresponding exactly to the nameplate voltage. This is
because voltage drop exists in each part of the power system from the generator to the customer's meter. There is also considerable
voltage drop in the consumer's internal wiring. Since voltage drop is proportional to the magnitude of the load current flowing through
the entire power system impedance, the customer who is electrically farthest from the substation will receive the lowest voltage (see
Figure 7-2).
Since all customers have essentially the same utilization devices, it is necessary to provide each of them with nearly the same
utilization voltages. A compromise is needed, however, between the voltage range which the utility must supply, and the voltage range
within which equipment will operate satisfactorily. If the limits of voltage provided by the power company are too broad, the cost of
appliances and other
 Page 273
utilization equipment, especially computers, will be high because they will have to be designed to operate satisfactorily within these
limits. On the other hand, if the voltage limits required for satisfactory operation of the utilization equipment are too narrow, the cost of
providing power within these limits will be excessively high.
Over the years, electric utility companies and equipment manufacturers have cooperated in establishing standards for operating voltage
limits which have proved to result in satisfactory operation without excessive demands upon the design of the power system or on the
utilization of equipment.

Figure 7-2.
Voltage Profile - Residential Feeder
The American National Standards Institute (ANSI) has established Standard C84-1, Voltage Ratings for Electric Power Systems and
Equipment, which was formulated by both utilities and manufacturers, and its recommendations are followed by both. Figure 7-3
illustrates the standards for service (meter) and utilization (load) voltages. The utility must only meet the service requirements because
only the customer has control over the voltage drop in his circuit. The Range A values are defined as the limits within which systems
shall be designed and operated so that most of the service voltages fall within them (114126). Voltage
 Page 274
variances outside these limits are to be infrequent. Voltage Range B levels are allowable provided they occur infrequently and are of
limited duration (110127). When they occur, corrective measures are to be undertaken within a reasonable amount of time to provide
voltage within Range A limits.

Figure 7-3.
Voltage Ranges Specified By ANSI Standard C84. 1-1984
The utility has some control over voltage to the customer's load. This control, however, is not instantaneous and does not ensure a
constant voltage, but merely maintains the voltage between some desired limits. While standard voltage regulating practices may be
effective for gradual voltage changes, they have little effect on voltage flicker.
Typical voltage profiles for heavy and light loads are shown in Figure 7-4. These profiles show, in a conceptual way, the voltage drops
or rises that may occur on the overall system between the generators and the loads to be served. During any one day, there is a definite
difference in voltage levels between heavy load and light load periods. Under heavy load, the power flows (and thus the voltage drops)
are larger. Under light load, there will be less voltage drop. Notice that voltages are corrected throughout the power system, so that
excessive drops do not occur. For short distribution feeders, it is entirely likely that voltage
 Page 275
standards can be met by using only the voltage control equipment in the substation, such as LTC's (Load Tap Changing) or regulators.
However, for long feeders or feeders with extremely heavy loads, it may be necessary to augment the substation equipment with
voltage regulators located out on the feeder.

Figure 7-4.
Voltage Profile Across System During Heavy and Light
Load
In addition to LTC transformers and regulators, many utilities employ shunt capacitor banks to help control the feeder voltage profile.
Because capacitors supply reactive power (VARs) to a circuit, they can effectively reduce and even cancel out the VAR requirements
of inductive loads, such as motors. Placed strategically on the feeder, shunt capacitors will not only improve voltage profiles, but will
also reduce the resistive losses associated with the primary feeder and laterals.
Figure 7-5 illustrates the improvement in voltage obtained by connecting a three-phase capacitor bank in the primary feeder. Without
the capacitor (solid line), the service voltages range from 105% (first customer) to 90% of nominal (108 volts at the last customer).
With the capacitor connected from the feeder to ground, the voltage profile will be adjusted as shown by the dotted line. Voltages
along the entire feeder are now within plus or minus 5 percent of nominal.
 Page 276
Sometimes, especially during light load conditions, regulators or capacitors can overcompensate and create an overvoltage lasting from
a few seconds to many hours (see Figure 7-6).

Figure 7-5.
Voltage Profile With Capacitor

Figure 7-6.
Overvoltage Condition
Figure 7-7 shows the effect of putting a large bank of capacitors on a system at light load. As can be seen, the utility system voltage,
which
 Page 277
is usually limited to 127 volts (126 volts at the customer's meter), now approaches 130 volts and is above the ranges suggested by
ANSI C84.1.

Figure 7-7.
Effect of 2700 kVAR of Distributed Capacitors
Flicker
Many voltage problems associated with computers, however, are not necessarily related to simply high or low voltage, but rather to
rapid
 Page 278
changes in voltage. This is called voltage flicker. Utility voltage regulation equipment, which will be discussed later, will not
compensate for instantaneous voltage fluctuations caused by the sudden application of low power factor loads such as motors. To
illustrate, Figure 7-8 shows a power system supplying an arc furnace and some domestic loads where voltage flicker might be a
problem.

Figure 7-8.
Simplified Circuit Illustrating the Flicker Problem
Whereas the domestic load currents are practically constant, the arc furnace load undergoes rapid changes and generates a variable
voltage dip along the distribution feeder. The voltage of the feeder primary becomes variable, and disturbances may be noticed by
some of the domestic loads attached to it.
Figure 7-9 shows a curve of the maximum permissible voltage flicker in incandescent lighting before it becomes objectionable to the
customer (borderline of irritation). If a user, for example, is in the vicinity of a heat pump or air conditioner which fluctuates
approximately 6 times per hour, and notices that his incandescent lamps are flickering, he can assume, using this curve, that the voltage
change is on the order of about 4% (or more). In many cases, the computer may be far more sensitive to voltage change than the human
eye and consequently experiences problems the human eye cannot detect.
 Page 279

Figure 7-9.
Voltage Flicker - Border Line of Irritation Curve
Voltage Sags
A voltage sag is a short term decrease in steady state voltage. These sags, sometimes lasting many seconds, can either have very little
effect on sensitive loads if the drop in voltage is not more than 10% or 20%, or they can have a major effect (similar to an outage) if the
voltage sag is larger (e.g., 50%). Voltage sags in an industrial facility are generally caused by large loads switching on. This can also
happen on a utility, but more often, severe sags are caused by faults on the system. Figure 7-10 illustrates a severe sag caused by a
transformer fault. Most severe sags on a utility system are the result of faults on the same feeder, at the substation, or even on an
adjacent feeder, and usually, they are relatively easy to identify. A common mistake is to believe that a power line conditioner will
protect a sensitive load from sags. While this is true for some sags, it is not usually the case for severe sags where the PLC (power line
conditioner) cannot fully compensate for the voltage dip and the undervoltage relay of the PLC drops out.
 Page 280

Figure 7-10.
Sag Caused by Transformer Failure
Figure 7-11 illustrates another interesting situation. As shown, faults on the transmission and subtransmission system can affect
customer voltages over 50 miles away. Sometimes these voltage sags can be severe and cause equipment, such as PLCs or computers,
to trip out. When a customer calls to complain of poor power quality, the real problem for the utility is that they may have no local
record of a system disturbance because the fault was not in the distribution system, and possibly not even in the same district.

Figure 7-11.
Power Quality Disturbance: Voltage Sag due to
Subtransmission Fault
 Page 281
Voltage Swells
The occurrence of ground faults on 3-phase systems cause the voltages in the unfaulted phases to rise with respect to ground
(electronics engineers refer to these as surges) (see Figure 7-12). This voltage increase can be as much as approximately 30% for a 4-
wire, multigrounded system and over 70% for a 3-wire system. The duration of this overvoltage, or swell, is dependent on the system
protection, which can take as long as minutes or as little as a half cycle.

Figure 7-12.
Voltage Swell
The magnitude of overvoltages during ground faults are particularly important to the proper sizing and operation of surge arresters. The
standards in this area are hence developed by the Surge Protective Device Committee of IEEE. The maximum magnitude of
overvoltages (swells) occurring for various line-to-ground faults for different distribution system designs (per IEEE C62.92-1991) is
shown below:
 Page 282
System Overvoltage Magnitude
Ungrounded 1.82 × ELG
Four-wire multigrounded (spacer cable) 1.5 × ELG
Three- or four-wire unigrounded (open wire) 1.4 × ELG
Four-wire multigrounded (open wire-gapped) 1.25 × ELG
Four-wire multigrounded (open wire-MOV) 1.35 × ELG*
ELG = Nominal line-to-ground voltage of system
*Because the metal-oxide varistor (MOV) arrester is more sensitive
to poor grounding, poor regulation, and the reduced saturation
sometimes found in newer transformers, many utilities are using a
more conservative 1.35 factor

Harmonics
In the ideal power system, the voltage supplied to customer equipment and the resulting load current are perfect sine waves. However,
conditions are never ideal in practice, so these waveforms are often distorted. This deviation from perfect sinusoids is usually
expressed in terms of harmonics distortion of the voltage and current waveforms. Harmonics distortion problems are not new to
utilities. In fact, such distortion was observed by utility operating personnel as early as the 1920's. Typically, the distortion was caused
by nonlinear loads connected to utility distribution systems. For example, an arc furnace is nonlinear since it will draw a nonsinusoidal
current (rich in harmonics) when a sinusoidal voltage is applied (see Figure 7-13A). The distorted load current then causes distorted
bus voltages to appear throughout the system (see Figure 7-13B). In the past, such harmonic sources were not very widely used, and
harmonics were often effectively mitigated through the use of the grounded wye-delta transformer connection.
Today, however, additional methods for dealing with harmonics are necessary due to the influence of the following developments:
1. The recent proliferation in the use of static power converters,
2. Added network resonances, and
3. Power system equipment and loads which are more sensitive to harmonics.
 Page 283

Figure 7-13.
Harmonic Distortion
There are now two criteria that are used to evaluate harmonic distortion. The first is a limitation in the harmonic current that a user can
transmit into the utility system. Table 7-1 lists the harmonic current limits based on the size of the user's harmonic loads with respect to
the size of the power system to which he is connected. The ratio of ISC/IL is the short circuit current available at the point of common
coupling (PCC) to the nominal fundamental load current. Thus, as the size of the user load decreases with respect to the size of the
system, the larger is the percentage of harmonic current the user is allowed to inject into the utility system. This protects other users on
the same feeder as well as the utility, which is required to furnish a certain quality of power to its customers.
The second criterion specifies the quality of the voltage which the utility must furnish the user. Table 7-2 lists the voltage distortion
that is acceptable from a utility by a user. This table is similar to the one in the present edition of IEEE 519. Since the utility is the PCC
between users, it has the responsibility of monitoring the harmonic current each user injects into its system, and of ensuring that this
current does not cause its voltage to reach distortion levels higher than those listed in Table 7-2. The values of Table 7-2 are low
enough to ensure that equipment will operate correctly.
 Page 284

Table 7-1. Harmonic Current Limits (in %) from IEEE

Std. 519
ISC Total Harmonic
/ILOAD HARMONIC ORDER Distortion
<11 111617222324>35
< 20 4.0 2.0 1.5 0.6 0.3 5.0
2050 7.0 3.5 2.5 1.0 0.5 8.0
50100 10.0 4.5 4.0 1.5 0.7 12.0
1001000 12.0 5.5 5.0 2.0 1.0 15.0
> 1000 15.0 7.0 6.0 2.5 1.4 20.0
Where ISC = Maximum short circuit current at point of
common coupling.
And IL = Maximum demand load current (fundamental
frequency) at point of common coupling.
TDD = Total demand distortion (RSS) in % of maximum
demand

Table 7-2. Maximum Voltage Distortion per IEEE Std.

519
$Maximum Distortion (in
%) SYSTEM VOLTAGE
Below 69 69 138 > 138
kV kV kV
$Individual Harmonic $3.0 $1.5 $1.0
Total Harmonic 5.0 2.5 1.5
(For conditions lasting more than one hour. Shorter
periods increase limit by 50%)

Harmonics are not normally produced by the power system itself but rather by the loads connected to the power system. Some of the
more common sources of these harmonics are:
Static power converters
Overexcited transformers
Fluorescent lights
Solid state devices (computers, dimmer switches, variable speed drives).
 Page 285
For example, solid state switching devices may use diodes, thyristors, etc. The simple single-phase, full-wave rectifier, shown in Figure
7-14 illustrates the operation of this commonly used device. As can be seen, although the voltage input waveshape is a sine wave, the
current drawn by the device is rich in harmonics.

Figure 7-14.
AC Drawn by Consumer Equipment
Likewise, a transformer operating at a voltage higher than its normal operating point may be pushed into its saturation region (see
Figure 7-15) where large harmonic content is produced (see Figure 7-16).

Figure 7-15.
Transformer Excitation Current Plotted vs. Applied
Voltage
 Page 286

Figure 7-16.
Nonsinusoidal Transformer Excitation Current and its Harmonic Content
Other examples of wave distortion are shown below in Figure 7-17:
 Page 287

Figure 7-17.
Example Waveforms From Several Common Sources.
 Page 288
Harmonic problems on utility systems are generally difficult to diagnose since the level of harmonic buildup may be gradual or a
system change totally unrelated to a harmonic source may result in a new problem. The most common harmonic-caused problems
found on the distribution system would normally manifest themselves as one of the following:
Excessive capacitor fuse blowing
Motor and transformer overheating
Unexplained breaker tripping due to a ground fault relay
Complaints of telephone interference.
The solution to these problems range anywhere from simply changing the location of something like a capacitor bank to having to
replace a transformer or put in a filter. This is one area of distribution engineering where an ounce of prevention may be worth a pound
of cure, i.e., don't exceed IEEE 519 recommendations.
Example 1. Suppose a utility has a customer who wants to install a device, at 13.8 kV, where the continuous load is 100 amperes and
the total demand distortion (TDD) is 7%. The location of his business is at a point on the feeder where the short circuit current is 1600
amperes. Should the utility allow this load? And if not, what can be done to accommodate this customer?
Solution. ISC/IL = 1600/100 = 16
IEEE 519 indicates (see Table 7-1) that if the TDD is 7%, the ratio of short circuit to load current must exceed 20. Since the actual
ratio is only 16, the customer has two choices: relocate the business closer to the substation to increase the short circuit level to at least
2,000 amperes (very unrealistic) or put in filters to reduce the TDD to 5% or less.
Interruptions
Service outages can be caused by either planned or unplanned outages of equipment. When equipment outages are planned, such as to
increase the transformer size, the supply to a small number of customers may be intentionally interrupted. More likely, however,
service outages are caused by unplanned events (faults), such as lightning, wind, or ice. In
 Page 289
the case of either planned or unplanned outages, the system is designed and the protection schemes are operated so that:
1. The number of service outages is minimized;
2. The duration of a service outage is minimized;
3. The number of customers interrupted is minimized.
Faults on the utility system are classified as either temporary or permanent. A temporary fault may be due to a lightning stroke, an
animal, wind, or other natural occurrence. When the fault occurs, the line must be de-energized to stop the flow of fault current and
enough time must be allowed so that the faulted path can be de-ionized. To do this, the breaker in the substation will open immediately
to clear the fault, and then automatically reclose after some time delay. This reclosing can occur several times in an effort to re-
establish continuity of service for a temporary fault. Typical time delays between reclosures are shown in Figure 7-18.

Figure 7-18.
Current vs. Time
This opening and reclosing of the breaker is usually short enough that most customers may not even be aware that service had been
interrupted. A computer or other sensitive load, however, will see a total system shutdown unless steps have been taken
(uninterruptible power supply) to avoid this short outage.
A permanent fault will result in a service outage which may last several minutes to an hour or more. The average duration of a
permanent fault per a recent survey was approximately 2 hours. During a permanent fault condition, the breaker will usually operate
three or four times in an attempt to reestablish power before it locks open. A broken conductor,
 Page 290
a tree limb falling onto a conductor, or perhaps across two conductors, are all examples of permanent faults. In this case, the fault must
first be located and repaired before service is restored to all customers.
Most conductor-related outages on overhead distribution lines are of a temporary nature (roughly 75%). By contrast, most faults on
underground systems are permanent and take much longer to locate and repair. One consequence of having more sensitive loads is that
the 75% of overhead faults which were once classified as temporary are now classified as causing permanent equipment outages.
The frequency and duration of faults is to a limited degree a function of the utility system design. There are some basic system designs
used by utilities which are more reliable than others. Unfortunately, as the reliability goes up, so (usually) does the cost.
One major complication to reliability calculations is that, due to sensitive loads, what was once not considered an outage is now
considered an outage. This is certainly the case in successful reclosing on a temporary fault.
This re-evaluation caused by more sensitive loads has changed some utilities philosophy with regard to feeder selective relaying.
Feeder selective relaying simply means that for the system shown in Figure 7-19, the fuse should blow for permanent faults and the
breaker should operate faster than the fuse for temporary faults.

Figure 7-19.
Feeder Selective Relaying
Some utilities that have considerable penetration of computer loads on these laterals prefer not to reclose at all, even for a temporary
fault.
 Page 291
This is because the breaker operation affects the entire feeder, while the operation of the fuse affects only the lateral with the temporary
fault. This does, however, create a permanent outage for those on the faulted lateral tap.
Voltage Surges
Voltage surges on a power system are the most common sort of power problems seen by the computer user. These transients can be the
cause of lost data, false triggering, and equipment failure, and many of them are generated internally by the user himself. Other
transients are the result of some occurrence on the utility primary, such as lightning and equipment switching.
Lightning transients to the low voltage system can occur from either direct strikes to the secondary circuit or strikes to the primary
circuit where transient voltages pass through the distribution transformer. Although those transients which pass through the transformer
are reduced in magnitude, they are not reduced by the turns ratio of the transformer (approximately 60 to 1) because the transformer
windings appear, electrically, to be a capacitor. Lightning transient voltages on the utility primary are limited to equipment flashover
levels (approximately 95 kV-300 kV for a 15 kV class system) and by arrester protective levels which are approximately 40 kV crest
(see Figure 7-20).

Figure 7-20.
Lightning Induced Voltage Surge
 Page 292

Sensitive Loads
Household Appliances
Some engineers have argued that this present emphasis on power quality is the direct result of the sensitivity of modern digital
appliances. The blinking clock syndrome is certainly well known to all distribution engineers. The previous analog clocks were
somewhat immune to the successful reclosing sequence which transpired during a temporary fault. A digital clock without battery
backup, however, interprets a momentary interruption the same as it would a compete outage. Figure 7-21, shown below, illustrates the
sensitivity of 3 of our most common household appliances. As can be seen, virtually none of these appliances could operate properly
after even an instantaneous reclosure, since these take 20 cycles or more and most devices can't ride through even a cycle of no
voltage. A sag, on the other hand, is another story. All three curves indicate that a voltage sag that does not reduce the voltage to less
than 60% of nominal should pose no problem. Sags on the transmission system generally would not be expected to cause a sag of this
magnitude.

Figure 7-21.
Malfunction Range Summary For All Equipment
 Page 293
Computers
Computer equipment is more sensitive to power quality than most other equipment. The CBEMA curve (Computer Business
Equipment Manufacturers' Association) was created in order to define the transient and steady state limits within which the input
voltage can vary without affecting the proper performance of or causing damage to the computer equipment, as shown in Figure 7-22.
Referring to the figure, it can be seen that a range from +6% to -13% of the nominal voltage will allow the computer to function
properly. For shorter time spans, the voltage tolerances are larger. For example, voltage can fall to zero as long as it recovers with 0.5
cycles or 8.33 ms and it can drop to 30% below normal for up to 0.5 seconds. This is because stored energy is built into the computer's
power supply filters and motors.

Figure 7-22.
CBEMA Curve
Low Voltage Environment
Power quality problems at the customer level are the result of both utility created disturbances as well as those created by the customer
himself. The residential customer influences their own power quality either by the
 Page 294
design of their house wiring, their use of appliances, or the types of load used. Some of the more important findings concerning the
environment below 600 volts are the following:
Swells. Typical overvoltages in the range of 5% above nominal would not be expected to have any serious effect even on sensitive
loads. Swells, on the other hand, are severe steady state overvoltages on the unfaulted phases of a system experiencing a line-to-ground
fault. These overvoltages can be expected to approach 1.25 per unit on the typical 4-wire multigrounded system. Figure 7-23 shows an
actual case where the voltage measured in the home reached 150 volts. The duration of this overvoltage is a function of the speed of the
protection system. Tests by one utility showed that swells of 22% lasting 120 cycles caused no damage or misoperation of typical
sensitive loads. These same tests have not been performed repetitively or upon aged equipment so it is difficult to assess the overall
effect of this condition.

Figure 7-23.
Overvoltages During Line-to-Ground Fault
Harmonics. Harmonics standards (IEEE 519) have been discussed earlier and indicate that the utility system will tolerate only very
moderate amounts of current and voltage harmonic distortion. It should be noted that a household appliance, such as a stereo can
generate significant amounts of current distortion (see Figure 7-24). Although the percent distortion for this device is high, the voltage
distortion is low since the source is relatively stiff and the current consumption of such a device is low (see Figure 7-25).
 Page 295

Figure 7-24.
Current Measured During Stereo Operation.
Figure 7-24 shows the voltage waveshape measured at the same time as the current.

Figure 7-25.
Voltage Measured During Stereo Operation.
Sags. Sags in the home caused by the utility are generally infrequent but quite severe. For very sensitive loads, or for severe sags,
misoperation can occur. Sags caused by the home or industrial complex are more likely to be caused by motor startups and are
generally not severe enough to
 Page 296
cause any problem. Tests with air conditioners indicate that normal air conditioning operation can result in significant sags. For
example, in a test of a standard house window unit, the voltage sag measured at the meter during startup was only 3 volts while the
voltage sag due to the drop in the house wiring was over 30 volts.
Surges. Studies have shown that surges found in the home are almost all created by conditions within the home itself. These voltages
can be as high as 2000 volts but are more commonly in the range of 100 to 500 volts. They do not seem to cause any perceptible
problems.
Power Line Conditioning
Introduction
Utility engineers throughout the country are finding that knowledge of just their power system is inadequate when working in the area
of power quality. More and more, utility customers, who are experiencing power related problems for sensitive loads, are depending on
the utility for solutions, even when the only solution is on the customer's premises. For this reason, it is now important that distribution
engineers have a general understanding of the options available to the customer and the limitations of those options.
Power line conditioners mean different things to different people. The term power line conditioner can be applied to any device that
protects against one or more of the following power problems:
Overvoltages and undervoltages
Sags
Outages
Surges
Harmonics
There are some power disturbances which exist all the time but do not cause any malfunction or component stress on computers or
similar delicate electronic equipment. For example, there is always some harmonic distortion on the power line sinewave, and there are
occasional transients. These distortions generally cause no harm.
 Page 297
The price range of power line conditioners can vary from $25 to well over $10,000, and one cannot always be sure what is being
purchased for that price. Generally, complete specifications are not found on the product, and if an attribute is not described on the
packaging of the device, it will more than likely not have that capability. A line conditioner should solve most power quality problems,
and its specifications should cover each type of disturbance listed above. This section will briefly discuss the advantages and
disadvantages of the more commonly applied devices.
Surge Suppression
Surge suppressors are devices which conduct electricity when the voltage exceeds a limit. The device does not really suppress the surge
(like a filter) but rather diverts it (usually to ground). Figure 7-26 shows the effect of the surge suppressor known as an MOV (Metal
Oxide Varistor).
The amount of energy these suppressors can absorb is directly related to their size. If the device is overstressed, it will self destruct;
thus it is very important that the energy level of the transient for a particular location be known. For example, suppressors at the
panelboard will be much larger than those generally found in wall receptacles or on electronic devices.

Figure 7-26.
Normal Mode Surge Suppressor
 Page 298
Evaluating the performance of a surge suppressor is not an easy task. The configuration of components (line-to-ground, line-to-neutral,
neutral-to-ground), as well as the quality and characteristics of the components are not easily determined. Marketing tends to disclose
only what the devices will do and nothing about what they will not do.
The best surge suppression on the utility primary distribution system will not be adequate for the protection of a customer's sensitive
equipment. Voltage transients coming from the utility system, as well as those (over 90%) generated internally, can only be mitigated
on the secondary at the customer's premises (see Figure 7-27). This power conditioning is cheap and, in the author's opinion,
mandatory for good performance.

Figure 7-27.
Summary of Surge Suppressor Data Average Value of
Suppressed Transients
 Page 299
Constant Voltage Transformers (CVTs)
A ferroresonant transformer is essentially a transformer operating in a saturated mode. Ferroresonant transformers (see Figure 7-28)
regulate voltage and, to a degree, can perform wave shaping.

Figure 7-28.
Elements of Ferroresonant Regulation Using a
Transformer
While the CVT has many positive qualities which made it very popular for dc power supplies (IBM used nothing but ferroresonant dc
power supplies in their computers for many years), they do have the following disadvantages:
a. The transformer leakage impedance limits the short circuit current, which is good. This impedance however, can become a severe
limitation when it involves start-up currents of a motor and other electronic equipment.
b. If more than one device is on the transformer, start-up of one can cause the magnetic field to collapse. This can result in the
misoperation of the other(s).
c. Since the circuit (essentially a tank circuit) stores energy, it has some ability to ride through certain minor disturbances. If a tank
circuit loses power, then it must replace this power. Because of this, a one-cycle outage can look like a multicycle voltage sag, which
may have a greater effect on the proper performance of the equipment.
d. Devices are heavy, noisy, and cause output transients which can greatly interact with computers.
 Page 300
Isolation Transformers
Figure 7-29 shows how an isolation transformer was used to break the ground loop in order to eliminate common-mode problems (this
is now prohibited by Federal law). These types of problems are reduced because a common-mode pulse cannot develop a magnetic
field in the core of the transformer; yet the common-mode, high frequency impulses can still pass through to the secondary because of
the transformer coupling capacitance. Common-mode attenuation capabilities can be substantially improved in an isolation transformer
by using electrostatic Faraday shielding around the windings. The shields effectively add bypass capacitance which routes high
frequencies to ground.

Figure 7-29.
Isolation Transformer
Some of the drawbacks of isolation transformers are as follows:
Cannot remove any normal-mode signal or disturbance.
Shielding can form an LC circuit which is resonant at some high frequency.
Can cause severe power loss that causes clipping of the sinewave peaks that are above nominal rating.
 Page 301
Can saturate during transients and distort the next few cycles.
Saturation currents can result in nuisance fuse blowings.
Line Conditioners
While power line conditioning can, in most people's minds, refer to anything from a simple surge arrester to a complex UPS system, a
power line conditioner generally brings to mind a device which provides several functions, including impulse attenuation, filtering,
voltage regulation, and isolation. In truth, some line conditioners do provide these functions, and some do not so be careful. Two of the
more common types of line conditioners are described as follows:
a. Enhanced Isolation Transformer: The enhanced isolation transformer, shown in Figure 7-30, uses the inductance of the chokes to
attenuate higher frequencies, the MOV to attenuate impulses, and the capacitor to improve normal mode attenuation. Some are
available with single, double, or triple Faraday shielding to improve common-mode attenuation. The major drawback to this type of
line conditioner is that it does not regulate voltage.

Figure 7-30.
Enhanced Isolation Transformer
b. M-G Sets: An M-G set, or a motor-generator set, consists of a drive motor mechanically connected to an ac generator. M-G sets
generate
 Page 302
a new voltage waveform for the load and are therefore somewhat isolated from minor disturbances on the power system.
Motor-generators can shield the load from impulses and from voltage sags and surges. For power line voltage changes of ±20% or
more, voltage to the load is maintained at nominal. A useful feature of the motor-generator is its ability to bridge severe short term sags
or outages. The momentum of the rotating elements permits the motor-generator to span outages of up to 0.3 seconds (approximately).
This time span is mainly limited by the drop in frequency (speed) as energy is removed. The period can be extended by adding inertia
via a flywheel as shown in Figure 7-31. Through the combined use of variable-speed, constant frequency or quick-starting engine
generator techniques and flywheel inertia, it is now practical to extend the time an outage can be spanned so that it is long enough to
protect against the common problem of clearing and reclosure caused by feeder faults. Cost, however, is considerably higher with this
system than with conventional motor-generators.

Figure 7-31.
Motor-Generator With Flywheel
The problems with motor-generators are mostly on the output or load side. High generator output impedance can cause substantial
voltage
 Page 303
dips in response to sudden load changes, such as those resulting from large motor starting current. And in addition, response to load
changes can be sluggish. Also the drive motor may overheat under long term brownout or low line voltage conditions unless it is
oversized. Motor-generator efficiency is typically relatively low, so that electrical energy costs over its lifetime may be substantial.
Heat dissipation, weight and bulk, and the potential for audible noise are factors which must be considered in motor-generator
installation. Essentially silent machines are available at extra cost. Bearings must be inspected and periodically replaced and/or
lubricated in many cases, particularly when flywheels are used. Reliability potential, however, is very high.
Uninterruptible Power Supplies (UPS)
Uninterruptible power supplies are the only solution for continuous operation of computer or other sensitive systems when line voltage
interruptions last approximately 0.5 seconds or longer, as is common for a utility. A properly designed UPS system can provide
computer-quality power under essentially all normal and abnormal utility power conditions, including outages up to 15 minutes or
more
UPS systems are typically solid state, although some are currently made using rotating machinery in combination with solid state
conversion. UPS systems have three general configurations illustrated in Figure 7-32.
As can be seen, all the systems contain a battery. The inverter preferred system is essentially the standard configuration for critical
equipment because it provides full isolation and power conditioning. The line preferred system (standby) receives its normal flow of
power from the utility system and relies on a 1/4 cycle transfer switch to isolate utility system disturbances. This type of system is
obviously not as effective as the inverter preferred system and is generally used for small, low cost computer systems where operation
is not highly critical.
Line interactive systems, like line preferred systems, have found application in the smaller, less critical sensitive equipment systems.
They too change mode of operation upon power failure. They can provide somewhat more power conditioning than the line preferred
system, but are inherently the least reliable of the three systems. Normally, power flows through the single-throw static switch and the
inductor to the sensitive equipment load and to the converter which, in this mode, acts as a battery charger. By various techniques, the
voltage to the sensitive
 Page 304

Figure 7-32.
UPS Configurations
equipment can be conditioned similarly to the fast response or ferroresonant regulators. Upon failure of the line, the static switch is
opened and the function of the converter is changed to that of an inverter delivering power to the sensitive equipment, resulting in a
less than a 1/4 cycle disturbance. A line interactive system is vulnerable to loss of output if there is a failure of the converter at any
time. Also, taking the converter in and out of service cannot be done without incurring a momentary interruption of voltage to the load.
A more recent version of this system provides for isolation of a failed or serviced converter by use
 Page 305
of an additional static switch. These problems do not exist with the inverter preferred system.
Again, it should be pointed out that the vast majority of power quality related problems are not caused by the utility but rather are
generated on the customers' premises. Some of these, such as spikes or surges, are easily corrected. Others, such as those resulting from
power outages, may require costly M-G sets or uninterruptible power supplies (UPS). This so-called conditioned power is not perfect.
Even loads connected to the same conditioned power supplies interact (see Figure 7-33) and many users have found the need for
separate conditioning for each sensitive load.

Figure 7-33.
Realities of Conditioned Power

Electromagnetic Fields
Another area sometimes considered a power quality issue is magnetic fields. The question of the health effects of magnetic fields is
certainly very complex as well as very confused. It does not appear that conclusive answers will be forthcoming for some time and yet
utilities are being asked to investigate ways to mitigate these fields should they be deemed
 Page 306
a problem. Although all parts of the utility system create magnetic fields, it is probably the distribution system that causes the most
concern because of its proximity, its extensiveness, and the fact that the nature of its design exacerbates the condition.
Measurements of fields, as shown below in Figure 7-34, indicates that magnetic fields resulting from distribution lines are probably
lower than other areas of the utility system and considerably lower than those caused by certain appliances. The concern however, is
justified mainly on the basis that proximity of people and homes to distribution lines is considerably greater than that of the
subtransmission or transmission.

Figure 7-34.
Typical Magnetic Field Levels
There are several solutions that have been considered to reduce magnetic fields. While reduction of fields might be a positive outcome,
some of these ideas may actually create more severe problems. Some are discussed as follows:
 Page 307
Grounded vs. Ungrounded
Distribution systems are usually either delta (ungrounded) or wye (4-wire, multigrounded). The trend is, by far, toward the
multigrounded approach. Since a delta system does not utilize the earth as a return path, it has been suggested as one methodology to
reduce magnetic fields. While this point cannot be argued, it does present some serious concerns since the present predisposition of
utilities toward the grounded system design results from three major factors:
Overcurrent Protection. Overcurrent protection of a distribution system is premised on the fact that short circuit current levels go down
as distance from the substation increases. These levels vary from as high as 30K amperes (more likely about 10 kA) to quite low values
in the hundreds of amperes. Most of these faults occur as line-to-neutral or line-to-ground faults and fault currents are very high
because of the multigrounded neutral design.
A delta system on the other hand produces very little fault current for a line-to-ground fault because there is no zero sequence ground
path, and as such produces very low levels of fault current (this due to line capacitance). In any case, it now becomes very difficult for
the relay to see many of these faults.
Overvoltage Protection. The selection of an arrester rating is based on the system grounding. The system BIL must be, by industry
guidelines, at least 20% higher (I believe it should be at least 50% higher) than the protective level of the arrester. The degree of
system grounding determines the maximum line-to-ground voltage an arrester will see due to a fault on another phase. At the present
time, guidelines suggest the following multiplication factors:
Table 7-3
Overvoltage
System
Magnitude
$Ungrounded $1.82 × ELG
4-Wire Multigrounded (open wire- 1.25 × ELG
gapped)
$ELG = Nominal Line-to-Ground Voltage of System.
 Page 308
An example probably best illustrates the problem. If we had a 34.5, multigrounded system, we would calculate the arrester rating as
follows:

7-1
On the other hand, if the system were a delta, we would calculate the arrester rating in this manner:

7-2
BIL must be, at the very least, 20% greater than the protective level (IR discharge at 10,000 amperes) of the arrester. So we would
determine the system BIL as follows:
34.5 kV - Grounded Wye:
108 × 1.20 = 130 kV => use 150 kV BIL 7-3
34.5 kV - Delta:
156 kV × 1.20 = 187.2 kV => use 200 kV BIL 7-4
As can be seen, the BIL of the delta system must be higher which of course means higher costs of transformers, lines, cables, switches,
etc.
Costs. Besides the cost of higher BIL, there is another major cost that must be considered. Residential load is single phase. To serve
this single-phase load requires two wires. With a multigrounded system, these two wires are a phase wire and a neutral wire. On a delta
system, two-phase wires must be used. When the residential subdivision is underground, which it usually is these days, the delta
system requires two insulated and very expensive cables where the multigrounded system requires only one (and an uninsulated
concentric neutral). Other considerations such as more bushing, more disconnects, etc., all increase cost.
 Page 309
Single-Phase vs. Three-Phase
There are utilities in Europe who serve residential loads with three-phase transformers. In the United States, we serve only our large
industrial and commercial loads in this manner. It is likely that such a system design reduces magnetic fields because individual phase
currents would tend to be lower and the fields would tend to cancel. Current unbalance would also tend to be less. On the other hand,
this system would be far too costly to justify in the United States where new residential construction tends to be underground (URD).
A comparison of some of the fundamental differences between the two systems is shown below in Table 7-4.
Table 7-4. 3-Phase vs. 1-Phase Residential Distribution
U.S. EUROPE
120/240 380 Wye/220, 4-wire 416 Wye/240,
4-wire (UK)
1-phase transformers heavily Less load per home than U.S.
overloaded - 25 kVA typical
4 homes/transformer fairly 3-phase xfrms >>$ 1-phase
typical
Higher load density Residential units in 300500 kVA
range
Fuses are typically expulsion 5 to 10 radial, 3-phase, 4-wire
secondary feeds, per transformer
No overload
Fuses are current limiting
100 to 200 dwellings per
transformer.

Underground
Underground distribution cables are generally enclosed in plastic duct (CIC) or direct buried. It is of particular interest to note that the
earth is not a good magnetic field shield at 60 Hz. IEEE Task Force conclusions relative to this subject are:
Buried pipes often carry sufficient 60 Hz current to substantially change the ground level magnetic field.
The magnetic fields of buried cables exceed those of an overhead power line carrying the same currents.
The pipe of a pipe-type cable may be an effective magnetic field shield at 60 Hz.
 Page 310
Line Compaction
Decreasing the spacing between conductors (see Figure 7-35) has been shown to reduce magnetic fields. Another advantage to this
compaction on distribution lines is esthetics.

Figure 7-35.
Example of Line Compaction
While compaction may present certain advantages, it also presents several of the following disadvantages:
Overvoltage. Bundling or compaction of conductors, such as spacer cable, increases the line-to-ground voltage seen by the unfaulted
phases during a line-to-ground fault. This factor changes from 1.25 to 1.5 per unit. This will usually have the effect of causing the
multigrounded system to utilize a higher arrester rating, especially for MOV applications. This, in turn, means that the protective
margin is decreased should the same BIL be possible to use. Or, in some cases, a higher system BIL may be required.
Overcurrent. Some utilities have found that an instantaneous
recloser on a system having reduced phase spacing may not be successful for
 Page 311
temporary fault conditions. These utilities have consequently been forced to omit the instantaneous reclosure and utilize a time delayed
reclosure after the first instantaneous trip. While this is not necessarily a problem, it does subject the customer to a longer outage (2 to
15 seconds instead of a half cycle) and has been a cause of concern in these days of higher power quality and more sensitive loads.
Line Flashovers. Lightning causes many of the temporary faults seen on the distribution system. These faults are a result of the
lightning stroke (induced or direct) which causes the insulation (usually air) to break down. A typical overhead line has an insulation
level of about 300 kV BIL. Some utilities, in high lightning areas, are seriously considering going to even higher line BILs, such as 450
kV, to decrease the occurrence of line flashover and increase power quality.
Unbalance
Residential loads are single-phase loads and connected phase to neutral on a 4-wire multiground system. In an effort to balance the
three-phase loads, single-phase lateral connections are rotated between the three phases. Since it is virtually impossible to completely
balance the three phases at all times, unbalances will exist. This three-phase unbalance shows up as residual current and returns to the
substation via the neutral wire and the earth. This, of course, adds to the magnetic field since phase cancelization is not as complete.
Typically, this phase unbalance can be significant. Some utilities, for example, set their ground fault relays between 25% and 50% of
full load current so as not to misoperate due to system unbalance. This can mean that up to 200 or 300 amperes of unbalanced current
can be accommodated before tripping.
Up until recently, it was rather difficult for a utility to monitor unbalance for an extended period of time. Today, however, devices are
available which are accurate and convenient to use. Such a device was used to obtain the data shown in Figure 7-36. As can be seen,
the feeder monitored here shows very good phase balance for most conditions but some phase unbalance during heavy load periods. It
is doubtful whether this condition could be improved to any significant degree by shifting loads amongst phases.
 Page 312

Figure 7-36.
Main Feeder Load Cycle

Stray Voltage
Introduction
Stray voltage is a small voltage of normally less than 10 volts that is usually measured between the earth and neutral. While this
voltage could exist on virtually any electrical facility, it is normally associated with dairy farms because farm animals, especially cows,
appear to be more sensitive to it then do human beings. Many experts feel that farmers seldom have a stray-voltage problem that affects
the health of the animals and/or milk production. They feel that these farmers' problems are generally caused by poor management,
improper feeding, poor cleanliness of the animals, improper breeding, defective installation and operation of milking equipment, and
poor sanitary conditions. However, very large stray voltages, during experimental tests, have been shown to affect animal behavior.
Since utilities can, under certain conditions,
 Page 313
contribute to stray voltage, law suits have been common for utilities serving dairy farms.
Sources of Stray Voltage
There are many sources of stray voltage which is one of the reasons it can be so difficult to locate and correct. Several of the more
common sources are described as follows:
Bad Connections. Bad connections, such as those caused by corrosion, looseness or even a break, are a primary cause of stray voltage.
The reason for this is that this otherwise low impedance path may cause the current to take an unintended path and produce stray
voltage.
Improper Wiring. The interconnection of the neutral (white) conductor and the equipment grounding conductor (green or bare wire) is
generally required by electrical codes. Compliance with the code implies that the neutral and ground will be at the same potential at
these points of interconnection. Improper grounding practices, such as the use of the neutral wire (white) also as the grounding
conductor or the use of a ground rod instead of a conductor leading back to the service entrance, may result in stray voltage.
Coupled Circuits. An ac electric field from a high voltage line can capacitively couple to a metallic object, like a metal fence, giving it
a measurable voltage.
Utility System Neutral. The neutral of a 3-phase, 4-wire multigrounded (typical distribution system design) utility system will produce
a neutral voltage as result of either a line-to-ground fault or unbalanced current on the line. If this system were balanced (same load on
all phases), virtually no current would flow in the neutral wire and the earth. However, because this is never the case, some of the
current will return via the earth and neutral conductor. A single-phase tap can be considered to be 100% unbalanced since all the return
current must return via the earth and neutral conductor.
Figure 7-37 shows how this neutral voltage occurs. As can be seen, the current in this single-phase transformer (or in a single-phase
tap) returns to the source via the neutral and earth. Some of the current returns through the earth (we have assumed 40%) and the rest
returns
 Page 314
through the neutral. The current flowing in the neutral between points A, B, C, and D will produce a voltage on this neutral as a result
of Ohm's Law. The voltage at A is then transferred to the secondary neutral via the distribution transformer. This neutral voltage can be
different from the customer's earth potential, even though the distribution transformer is grounded, primarily because the ground
footing impedance can be high (generally 25 ohms or more). The differential between the neutral voltage and the earth voltage, at the
customer's premises, is generally the stray voltage which concerns the utility engineer.

Figure 7-37.
Interconnection of Primary and Secondary Neutrals of Distribution Transformers. The practice of connecting the primary and
secondary neutrals on a 4-wire multigrounded system is a matter of safety. If a primary-to-secondary fault developed within the
transformer and the neutrals were not connected, the resistance of the return current path could be so high that not enough fault current
would flow to enable the primary protective device to clear the fault. High voltage would then be
 Page 315
impressed on the secondary for an extended period of time, posing a risk to humans, animals, and equipment. The practice of isolating
the secondary neutral is being used on some unigrounded and ungrounded systems. This practice is based upon a concern that a
primary fault to the transformer tank could cause a high voltage to be introduced into the secondary utilization services if the primary
and/or tank grounds are interconnected to the secondary neutral. This practice is especially not recommended for 4-wire systems.
Mitigation of Stray Voltages. The control and mitigation of stray voltages in livestock facilities demand careful consideration of both
animal sensitivity and electrical sources. All mitigation approaches have their own advantages and disadvantages and any approach
must consider the alternatives and constraints of a given situation.
Approaches for controlling neutral-to-earth voltages can be categorized as follows:
Voltage Reduction - Reduction of grounding resistance on the distribution neutral will reduce the neutral-to-earth potential due to
system loading. Corrective action such as correcting bad neutral connectors, removing faulty loads, balancing loads, improving
grounding, or increasing the size of the neutral can be helpful.
Active Voltage Suppression - If modification of the basic electrical system is difficult, one method of mitigation is active voltage
suppression. This method utilizes a second source of current to cancel the original source. Disadvantages include the possible
maintenance that may be involved, the cost, and the possibility of creating more voltage than needs to be mitigated.
Gradient Control - Gradient control is the same methodology used in substations to control step and touch voltages during faults. Such
control by equipotential planes can negate the effects of all neutral-to earth voltages in livestock facilities if it reduces the potential
differences at all possible animal contact points to an acceptable level. If an equipotential plane is installed (it should include
equipment grounding, metalwork bonding, conductive network in the floor bonded to the electrical grounding), the only possible
concern is that animals will receive an electrical shock when they move on or off the plane. One simple and apparently effective
solution where it is difficult to retrofit a concrete floor to establish an equipotential plane is the use of insulating paint.
Isolating the Primary and Secondary Neutrals - If the neutral voltage is determined to come from the utility distribution system, one
 Page 316
possible solution is a neutral isolation device. These devices (see Appendix B, IEEE C62.92.4-1991, IEEE Guide for the Application of
Neutral Grounding in Electrical Utility Systems, Part IV - Distribution) isolate the primary and secondary neutral for normal load
conditions but instantly and solidly connect them together in the event of any system disturbance that causes the voltage of one neutral
to rise above a predetermined threshold relative to the other neutral, as could be the case of an internal primary to secondary fault in the
transformer.
Reference
1. L.M. Anderson and K.B. Bowes, The Effects of Power-Line Disturbances of Consumer Electronic Equipment, '89 TD 423-5
PWRD, Transmission & Distribution Conference, April 1989.
Questions
1. Name three conditions that cause power quality problems.
2. A UPS system will solve most power quality problems. (true or false?)
3. Name 5 utility power quality problems.
4. The most important power quality parameter to a utility is:
a. primary voltage
b. primary current
c. secondary voltage
d. secondary current
Explain.
5. Voltage drop from the feeder to the meter can be as high as ___ volts.
 Page 317
6. A secondary voltage of 100 volts will cause most VCRs, microwaves and digital clocks to misoperate. (true or false?)
7. UPS systems can cause more problems than they solve. (true or false?) Explain.
8. Identify the wave (sag, swell, impulse, etc.), state approximate duration, give a possible cause, and suggest a possible solution for the
following:
 Page 318
8. (continued)
 Page 319
8. (continued)
 Page 320

8
Distribution Economics
Introduction
It is no longer possible for a utility engineer to justify a design or product change simply on the basis of technical superiority. Today
engineers are more often than not placed in the position where they must explain the economic payback of their ideas. In some cases,
this is fairly straightforward and conclusive but in others the economics is not that easy to assess.
It is the purpose of this chapter to demonstrate economic techniques that are useful specifically to a utility and are simple to use. These
techniques will be developed around such concepts as present worth, carrying charge, cost of losses, and operating costs as well as
customer satisfaction.
Time Value of Money
The time value of money means that a dollar today has a different value than a dollar a year from now. This is true even without
inflation because that dollar given to you today could be invested at a rate higher than inflation. People will pay to use your money.
Taking dollars at one point in time and finding its value at some other point in time is called present worth arithmetic. The following is
a review of some of the concepts associated with this type of analysis.
Future Worth
The process of taking money and finding its equivalent value at some point in the future is called future worth. For example, one dollar
today
 Page 321
would be worth 1 dollar and 12 cents a year from now if the established interest rate was 12%. If I wanted to find the value of one
dollar two years in the future I would just take the $1.12 I had after the first year and add 12% interest to it. I would now have $1.25 (I
would have $1.24 if this were simple interest). This concept of future value can be illustrated mathematically as follows.
Here are the principal symbols:
i = Interest rate
n = Number of interest periods
P = Present sum of money
S = Sum of money n interest periods from now.
Then formulas for calculating future worth of money at end of year are:

Year S = P + iP = P (1 + i)
1
Year S = P (1 + i) + iP(1 + i) = (P + iP) (1 + i) = P (1 + i) (1
2 + i) = P (1 + i)2
Year S = P (1 + i)n
n

The amount P is the present value of today's investment while the amount S is the future value. This concept can be shown by a simple
diagram, Figure 8-1.

Figure 8-1.
Future Worth
An example of the use of the present worth factor follows:
 Page 322
Example 1. How much would you have to invest at 12% interest on January 1, 1985, in order to accumulate $1,791 on January 1,
1991?
Solution 1

Figure 8-2.
Present Worth
The process of taking future dollars and bringing them back in time, for example back to the present, is called present worth. Present
worth is just the opposite of future worth, so if we take the future worth formula we just developed, we would have the following:
S = P(1 + i)n 8-5
so

8-6
 Page 323
An example of the use of the present worth factor (for a single payment follows:
Example 2. What is the present worth on January 1, 1984, of $1,263 on January 1, 1991, if the interest rate is 12% (use Table 8-1)?
Solution 2

Figure 8-3.
Equal Payments
Utility equipment, not unlike our own possessions, has costs (and savings) associated with them which are calculated on an annual
basis. When these costs are uniform, like our mortgage or car payment, we can estimate the value of these payments at some time by
using compound interest factors for a uniform series of payments. The terms generally used for equal payment economic evaluations
are:
1. Sinking fund factor
2. Capital recovery factor
3. Compound amount factor
4. Present worth factor
 Page 324
A graphical interpretation of these formulas is shown below in Figure 8-4. These figures are very useful in determining which formula
to use. For example, suppose you knew that the application of a capacitor bank could save you $300 per year. If you wanted to know
today's value of those yearly savings, you would go to Figure 8-4 and find the graph which fits this evaluation. In this case, we have
uniform payments (R) and wish to know the present value (P). We can see that we need to use the present worth factor for a uniform
series. The other terms become self explanatory when viewed in relationship to their graphical interpretations.

Figure 8-4.
Graphical Interpretation of Uniform Series Payments
Example 3. Assume a low loss transformer design can save an estimated $114.10 each year. How much would be saved at the end of
10 years at an interest rate of 12%?
 Page 325
Solution 3

Figure 8-5.
Example 4. How much would you need to deposit at 12% on January 1, 1991, in order to draw out $179.20 at the end of each year for
7 years, leaving nothing in the fund?
Solution 4

Figure 8-6.
 Page 326
TABLE 8-1. 12% COMPOUND INTEREST FACTORS
SINGLE PAYMENT UNIFORM SERIES
Compound Present Sinking Capital Compound Present
Amount Factor Worth Factor Fund Factor Recovery Amount Factor Worth Factor
n caf' pwf' sff Factor crf caf pwf n
Given P To Given S To Given S To Given P To Given R To Given R To
Find S (1 + i)n find P find R find R find S find P
1 1.120 0.8929 1.00000 1.12000 1.000 0.893 1
2 1.254 0.7972 0.47170 0.59170 2.120 1.690 2
3 1.405 0.7118 0.29635 0.41635 3.374 2.402 3
4 1.574 0.6355 0.20923 0.32923 4.779 3.037 4
5 1.762 0.5674 0.15741 0.27741 6.353 3.605 5
6 1.974 0.5066 0.12323 0.24323 8.115 4.111 6
7 2.211 0.4523 0.09912 0.21912 10.089 4.564 7
8 2.476 0.4039 0.08130 0.20130 12.300 4.968 8
9 2.773 0.3606 0.06768 0.18768 14.776 5.328 9
10 3.106 0.3220 0.05698 0.17698 17.549 5.650 10
11 3.479 0.2875 0.04842 0.16842 20.655 5.938 11
12 3.896 0.2567 0.04144 0.16144 24.133 6.194 12
13 4.363 0.2292 0.03568 0.15568 28.029 6.424 13
14 4.887 0.2046 0.03087 0.15087 32.393 6.628 14
15 5.474 0.1827 0.02682 0.14682 37.280 6.811 15
16 6.130 0.1631 0.02339 0.14339 42.753 6.974 16
17 6.866 0.1456 0.02046 0.14046 48.884 7.120 17
18 7.690 0.1300 0.01794 0.13794 55.750 7.250 18
19 8.613 0.1161 0.01576 0.13576 63.440 7.366 19
20 9.646 0.1037 0.01388 0.13388 72.052 7.469 20
 Page 327
Table 8-1. 12% COMPOUND INTEREST FACTORS - CON'T
SINGLE PAYMENT UNIFORM SERIES
Compound Present Sinking Capital Compound Present
Amount Factor Worth Factor Fund Factor Recovery Amount Factor Worth Factor
n caf' pwf' sff Factor crf caf pwf n
Given P To Given S To Given S To Given P To Given R To Given R To
Find S (1 + i)n find P find R find R find S find P
21 10.804 0.0926 0.01224 0.13224 81.699 7.562 21
22 12.100 0.0826 0.01081 0.13081 92.502 7.645 22
23 13.552 0.0738 0.00956 0.12956 104.603 7.718 23
24 15.179 0.0659 0.00846 0.12846 118.155 7.784 24
25 17.000 0.0588 0.00750 0.12750 133.334 7.843 25
26 19.040 0.0525 0.00665 0.12665 150.334 7.896 26
27 21.325 0.0469 0.00590 0.12590 169.374 7.943 27
28 23.884 0.0419 0.00524 0.12524 190.699 7.984 28
29 26.750 0.0374 0.00466 0.12466 214.582 8.022 29
30 29.960 0.0334 0.00414 0.12414 241.332 8.055 30
31 33.555 0.0298 0.00369 0.12369 271.292 8.085 31
32 37.582 0.0266 0.00328 0.12328 304.847 8.112 32
33 42.091 0.0238 0.00292 0.12292 342.429 8.135 33
34 47.142 0.0052 0.00260 0.12260 384.520 8.157 34
35 52.799 0.0189 0.00232 0.12232 431.663 8.176 35
40 93.051 0.0107 0.00130 0.12130 767.088 8.244 40
45 163.987 0.0061 0.00074 0.12074 1358.224 8.283 45
50 289.001 0.0035 0.00042 0.12042 2400.008 8.305 50
0.12000 8.333
 Page 328

Annual Carrying Charge

When you buy a car, you not only concern yourself with the cost of the car but also with other things like insurance, interest, gas, and
maintenance. Likewise, when a utility purchases equipment, it not only must consider the original cost of the equipment, but also must
consider the following:
1. Return on investment
2. Depreciation
3. Income tax
4. Property tax
5. Insurance
6. Operation and maintenance.
The combination of all these costs is called the carrying charge. A brief description of each of these costs is as follows:
Return on Investment
This is the part of the carrying charge which pays investors for the use of their money. It does not consider inflation. A utility has the
following two sources of borrowing the money:
Bonds (or debt money) - repayment of which is called interest
Stock (or equity money) - repayment called equity return.
The capitalization structure of a utility is the percentage of debt and equity by which the utility borrows its money. For example, if a
utility has a capitalization structure of 52% debt (bonds) and 48% equity (stock), its overall return on investment might be calculated as
follows:
Example
Given:

Cost of debt money (bonds) = 9.2%

Cost of equity money = 15.0%
(stocks)
 Page 329

Capitalization = 48% Equity

structure
52% Debt

The overall ROI (return of investment) = .52 × 9.2 + .48 × 15.0 = 12%
Depreciation
This second major cost component is some times referred to as recovery of capital or the return of investment. Investors not only want
interest on their investment but also want that investment back at the end of its life.
There are many different methods that utilities use to depreciate their plant and equipment. The most common is called straight line
where a fixed percentage of the investment is returned each year until the entire investment is returned. Table 8-2, shown below, is an
example of straight-line depreciation for an investment of $10,000 and a life of 5 years.
Table 8-2. Straight-Line Depreciation
10,000 Investment 5
Year Life Straight Depreciation is returned annually
Line
Annual Total
Year Investment ROI
Depreciation Depreciation
1 10,000 2,000 2,000 1,200
2 8,000 2,000 4,000 960
3 6,000 2,000 6,000 720
4 4,000 2,000 8,000 480
5 2,000 2,000 10,000 240

Federal Income Tax

Equity return (stock dividends) are a cost of borrowed money because they are not deductible in computing taxes. When a utility
invests in a plant it, in a sense, incurs a tax obligation for future years on that plant. So, when a utility borrows money from investors, it
must pay not only the
 Page 330
investors but also the federal government almost as much, which can be illustrated as follows:
P = Profit = Total revenue required to pay investors and taxes
Assume tax rate is 48% of total profit.
Assume investors (equity) earn $1, then
P = Equity Return + Taxes
P = 1 + .48 (P) 8-3
= > P = 1.92
OR
That for every dollar we pay investors we need $1.92 in profit
OR
For every $1 of equity return I need $.92 for the feds.
Property Taxes
Annual property taxes may be expressed as a percentage of the capital investment. This is of the order of 2.5%.
Insurance
Plant insurance is also an expense that may be directly related to the capital investment in a plant and may be expressed as a percentage
thereof. Insurance might be 0.25% per year.
Operating and Maintenance Expenses
The calculation of this cost component varies with the nature of the project. It is usually not a direct function of the capital invested and
may have an inverse tendency. That is, alternatives often exist for higher capital expenditures to reduce operating costs. Therefore, it is
not expressed as a percent of capital investment in most cases.
 Page 331
Calculation of Annual Carrying Charges
The annual carrying charge is the summation of all these costs. It is most convenient to have a single percentage number to represent
the annual cost of return on investment, depreciation, federal income tax, property tax, and insurance. Such a number is called the
levelized annual carrying charge. Multiplying the capital investment by the levelized annual carrying charge will give the annual cost
of the investment in the system.
To illustrate the calculation of carrying charges, the following table is proposed:
Table 8-3. $100 Investment - 5 Year Life, Straight Line
Depreciation, 48% Federal Income Tax
YEAR
1st 2nd 3rd 4th 5th
Investment for the year 100 80 60 40 20
Return Equity (7.2%) 7.20 5.76 4.32 2.88 1.44
Interest (4.8%) 4.80 3.84 2.88 1.92 .96
Depreciation 20.00 20.00 20.00 20.00 20.00
Income Tax 6.65 5.32 3.99 2.66 1.33
Yearly Charge 38.65 34.92 31.19 27.46 23.73

To convert these yearly charges to a single annual carrying charge, the present worth of the five yearly charges must be calculated
(using a 12% rate of return) and then converted to an equivalent uniform annual carrying charge. This is shown as follows:
= (38.65 × .8929 + 34.92 × .7972 + 31.19
× .7118 + 27.46 × .6355 + 23.73 × .5674) × .27741 8-4
= (34.51 + 27.84 + 22.20 + 17.54 + 13.46) × 0.27741
= 32.03
 Page 332
To this must be added the annual cost of property tax and insurance. The total annual carrying charge for the capital investment is:

Return on Investment
Depreciation 32.03
Federal Income Tax
Property Tax 2.50
Insurance .25
34.78

Annual carrying charges get smaller as equipment life gets longer. The total carrying charges for various service life and 12% return
are:
Years Total Levelized Annual
of Life Carrying Charge
5 34.78%
10 24.71%
15 21.87%
20 20.78%
25 20.35%
30 20.19%

Definition of Terms
Diversity
Diversity is a term used to cover the fact that individual loads occur at different times. This means that if the maximum load of two or
more loads are added, their sum will generally be greater than the true sum because these peaks occur at different times. Figure 8-7,
below, shows a typical diversity for three types of residential loads. These loads have been labeled high use or all electric, medium use
or full use where the customer has all the major appliances, and low use like you might see for an apartment dweller. Use of these
curves is fairly straightforward. For example, suppose a transformer is needed to supply 6 all electric or high use homes. The maximum
demand per home is about 20 kVA.
 Page 333
The diversified demand per home for 6 homes is about 13 kVA, so we would need a transformer that could handle 78 kVA at peak.

Figure 8-7.
Typical Diversity Curve
Load Factor
Load factor is defined as the ratio of the average load over a given period of time to the peak load during that time. This factor can be
determined very simply by dividing the number of kw-hr metered during the period by the product of the number of hours in the period
times the peak kilowatts. This is an important characteristic for the economic well-being of the electric utility company.
In effect, the peak load determines how much system capacity and, hence, system investment is required to serve a particular load or
group of loads, where as the average load determines the kw-hr billing revenue which will be obtained from serving that same load.
Hence, high load factor is favorable and low load factor unfavorable to the utility's economic welfare.
 Page 334
Overall system load factor of individual companies vary significantly from the national average of about 60%. Figure 8-8 shows actual
load factor curves for two utility companies. Company A enjoys annual load factors higher than the national average, while Company
B has load factors markedly below the national average.

Figure 8-8.
Individual Company Load Factors
Loss Factor
Loss factor is defined as the ratio of the average losses over a specified period of time to the peak losses during that same period. While
it is relatively easy to determine load factor, it is difficult to determine loss factor because average system losses cannot be easily
determined. The reason for this is, in part, due to the fact that system losses vary as the square of the current. The evaluation of energy
losses is of prime importance, for it represents the loss of a real, saleable product.
As would be expected, loss factor varies as some function of load factor. The following empirical relationship between load factor (Ld
F) and loss factor (Ls F) has been determined to be a relatively good approximation for calculations on a distribution system having a
typical load cycle:
 Page 335
Ls F = 0.15 Ld F + 0.85 Ld F2 8-5
The cost of conductor losses per year = (loss factor) (kW loss at peak load) (cost per kW-hr) (8760).
Transformer core losses, while relatively small in comparison with the total conductor losses, should not be neglected, and should, of
course, be evaluated on the basis of 100 percent loss factor or equivalent hours.
Responsibility Factors
Responsibility factor is the ratio of that part of an individual load that occurs at the time of the system peak load to the individual load
peak. It is thus a measure of how much the individual load contributes to the system peak load. Typical peak responsibility factors
which can be used for evaluating the cost of losses in various parts of the system are shown in Table 8-4.
Table 8-4
$LOSSES
Primary Transformer Secondary
To Production Peak 0.58 0.56 0.56
To Transmission & Station 0.60 0.58 0.58
To Primary Peak 1.00 0.85 0.85
To Transformer Peak -- 1.00 1.00
To Secondary Peak -- -- 1.00

Economic Evaluations
Economic evaluations are made in as almost endless variety of ways. The approaches for making economic evaluations used in this
text follow the philosophy below:
a. Estimate the duration of the comparison. If two devices are being compared and they have the same life, then only one life cycle is
normally used. If they have different service lives, then it is
 Page 336
necessary to go through enough life cycles until the alternatives terminate at the same time. If this doesn't happen, then taking the
evaluation out to 60 years or more is sufficient.
b. Calculate the annual costs associated with each device. For example, if we purchased a piece of equipment, we would have to pay a
levelized annual carrying charge on this equipment.
c. Calculate their present sum and sum up for each scheme.
Examples are shown below to illustrate different approaches for making economic comparisons:
Switches
Suppose you had a choice between two manufacturer's switches. The first manufacturer offered an oil switch for a cost of $12,000 and
you knew it had an annual maintenance cost of $50 per year and an average life of 20 years. The second manufacturer offered an SF6
switch for $15,000. He claimed that this switch had an expected life of at least 30 years and an average annual maintenance cost of
$25. The cost of money is 12%. You are asked to determine which switch is the least expensive.
The total time period for the evaluation is 60 years because this number is divisible by both the 20 year life and the 30 year life. The
evaluation is as follows:

Figure 8-9.
Life Cycle for Economic Comparison
 Page 337
Solution
1. Calculate Cost of Oil Switch
The carrying charge for a 20 year life has been previously calculated to be 20.78%. The present worth of the 3 switches over the 60
year period is calculated as follows:

020 (Cost)(Carrying Charge)(pwf) ($12,000)(.2078) $18,624.70

(7.469) =
2040(Bring to present from year 20) $18,624.70 * 1,931.38
.1037 =
4060(Bring to present from year 40) $18,624.70 * 199.28
.0107 =
Switch Total$20,755.36
P.W. cost of maintenance over 60 years $50 × 416.65
8.333 =
Total Cost of Oil Switch$21,172.01

2. Calculation of SF6 Switch

030 ($15,000)(.2019)(8.055) = $ 24,394.57

3060 ($24,394)(.0334) = 814.76
0+M $25 × 8.333 = 208.33
Total Cost of SF6 Switch $ 25,417.66

\ The oil switch would be more cost effective.

Capacitors
A 10 mile, 13.8 kV utility line with an impedance of .2 + j.6 ohm/mile serves an industrial customer with a peak load of 8 MVA at .8
PF and a .6 load factor. The customer has just purchased a 3 MVAR bank of capacitors. Is the cost of losses reduced at a rate of $.08
per kHR?
 Page 338
Solution
1. Losses without caps

8-6

8-7
Total Annual Losses =
3 × 8760 × 88.35 = 2,321,838 kWHR 8-8
Cost/Year = $185,747/Year
2. Losses with capacitor bank
8 MVA at .8 PF = 6.4 MW 89
= 4.8 MVAR
With capacitor load
= 6.4 MW and 1.8 MVAR 810
= New Load is 6.65 MVA

811
Avg. Loss = (278.22)2(10 × .2) × .396 812
= 61.31 kW Per Phase
 Page 339
Total annual cost of losses =
3 × 8760 × 61.31 × .08 = $128,898 813
Cost Savings:
$185,747
-128,898
$ 56,849
Transformers
Economic evaluation of transformers involves several of the concepts discussed earlier. To properly evaluate the cost of owning a
transformer, the following three annual costs must be considered:
1. Annual carrying charge of the unit
2. Cost of no-load losses
3. Cost of load losses.

Figure 8-10.
Transformer Model
 Page 340
The losses in a transformer occur under both loaded and unloaded conditions. These losses can be further broken down into two major
parts:
1. Energy component or production cost to generate kWhr losses.
2. Demand component or annual costs associated with system investment required to supply the peak kW of loss.
As we can see in the figure above, the core of the transformer always has approximately the same voltage across it and consequently
the I2R losses in the core (magnetizing branch) are constant, occurring 24 hours a day, 365 days a year. Since these losses occur even
when there is no load on the transformer, they are called no load losses. Also, since these losses occur at system peak, the utility must
install a kW of system capability for every kW of no load loss (responsibility factor is 1).
The load losses are the losses in the leakage impedance related to load conditions. These losses, sometimes referred to as copper losses
even though most windings are made from aluminum, go up as the square of the load. Because the peak load on a transformer may
occur at a different time than the system peak, it is not usually necessary to build a kW of system for every kW loss due to load.
Consequently, the responsibility factor is usually less than 1.
Example. There is a 5.6 kVA single-phase secondary load with a growth rate of 6% to be served. A 15 kVA distribution transformer is
readily available. How much extra in transformer cost and loss evaluation would it be to use this transformer rather than a 10 kVA?
Both transformers have a 20 year life, and operating and maintenance expenses are assumed to be the same. Assume the following
values:

System cost = $800/kW

Energy =
$0.03/kWh
Responsibility factor = 0.9
Loss factor = 0.15
Transformer costs 10 = $150
kVA
Transformer costs 15 = $200
kVA

Losses = TRANS. LOAD NO LOAD

10 kVA 125 W 59 W
15 kVA 179 W 76 W
 Page 341
Formulas for determining transformer loss evaluations are:
Annual Cost No Load Power Loss
= (P + 8760E) No Load Loss 814
Annual Cost Load Power Loss
= K2 (P × RF + 8760 × LsF × E) Load Loss 815

K = Equivalent annual peak

load
RF = Responsibility factor
LsF= Loss factor
P = Annual cost of system
capacity
E = Energy cost ($/kWh)

The first thing to do is to determine K, which can be done as follows:

 Page 342
Table 8-6. 10 kVA Transformer Loading
Year PU PU2 12% PVF PV
0 0.560 0.314 1.0000 0.314
1 0.594 0.353 0.8929 0.315
2 0.629 0.396 0.7972 0.316
3 0.667 0.445 0.7118 0.317
4 0.707 0.500 0.6355 0.318
5 0.749 0.561 0.5674 0.318
6 0.795 0.632 0.5066 0.320
7 0.842 0.709 0.4523 0.321
8 0.893 0.797 0.4039 0.322
9 0.946 0.895 0.3606 0.323
10 1.003 1.006 0.3220 0.324
11 1.063 1.130 0.2875 0.325
12 1.127 1.270 0.2567 0.326
13 1.194 1.426 0.2292 0.327
14 1.266 1.603 0.2046 0.328
15 1.342 1.801 0.1827 0.329
16 1.422 2.022 0.1631 0.330
17 1.508 2.274 0.1456 0.331
18 1.598 2.554 0.1300 0.332
19 1.695 2.873 0.1161 0.334
6.47

Uniform Annual Series = 0.13388 × 6.806 = 0.8662

Therefore, for the 10 kVA transformer K2 = 0.8662

Since per unit loading of the 15 kVA transformer is of the 10 kVA transformer, K2 for the 15 kVA transformer is 0.3850 (i.e.,
(2/3)2 × 0.8662).
 Page 343
10 kVA Transformer

Annual Cost of Transformer 150 × .2078 =

31.17
Load Loss = 0.8662 (0.9 × 800 × .2078 + 8760 × =
0.15 × 0.03)0.125 20.46
No Load Loss = (800 × .2078 + 8760 × 0.03)0.059 =
25.31
Total =
76.94

15 kVA Transformer

Annual Cost of Transformer 200 × .2078 =

41.56
Load Loss = 0.3850 (0.9 × 800 × .2078 + 8760 × =
0.15 × 0.03)0.179 13.02
No Load Loss = (800 × .2078 + 8760 × =
0.03)0.076 32.61
Total =
87.25

It would cost $9.66 more per year to use the 15 kVA transformer rather than the 10 kVA.
Automation
The methodology utilized to evaluate the value of a function is normally performed with consideration of both the carrying charge
associated with the additional necessary equipment and the rate of return associated with the appropriate savings in losses, etc. The
following is an example of a methodology used by the author to evaluate these relative values to indicate whether a function is
justified.
Assume that there are two feeders which are equally loaded for approximately 21 hours of each day but feeder A has a 3 hour peak
which occurs when feeder B is off peak. The normal configuration of this feeder for this peak period on feeder A is shown in Figure 8-
11. Also assume that the 5 mile section of line at the end of feeder A contains 200 amps of load.
As can be seen, the feeders for this hypothetical (and greatly exaggerated) case are unevenly loaded during this period. An additional
breaker could be installed 10 miles out on feeder A to balance these loads as shown in Figure 8-12.
 Page 344

Figure 8-11.
Normal Configuration at Feeder A Peak

Figure 8-12.
Automated Configuration at Feeder A Peak
If the loads are considered to be evenly distributed and the wire resistance is the same in all sections of the system, the percent
reduction in losses can be calculated as follows:
Old System:

817
 Page 345

818
New System:

819
% Change in Losses =

820
Let R be equal to 5 ohms (.5 ohms/mile × 10 miles). The change in losses is then equal to:

821
Assume that the load factor for the peak 3 hour period is 90%. Then the average losses can be calculated as follows:
 Page 346

822

The cost savings per year at 5 cents per kilowatt hour is equal to $12,841.
Suppose the cost to implement this additional function of load switching is $50,000 for the additional breaker and modifications to the
existing open tie point. The question is, how do we estimate whether this additional automated function makes economic sense?
Straight present worth analysis is not completely accurate since the total revenue requirements of the new equipment must include such
considerations as:
Return on investment
Depreciation
Income tax
Property tax
Insurance
Operating and maintenance expenses.
These costs change from year to year but are most commonly represented by a levelized value normally referred to as an annual
carrying charge. If we assume that the additional equipment has a life of 20 years, a levelized annual carrying charge of 20%, and that
the cost of money (return) is 12%, then the total present worth cost of ownership can be calculated as follows:
 Page 347
Annual Carrying Charge = $50,000 × .20 823
= $10,000
P = Present worth
R = Uniform annual cost
h = Years

824
Where i = 12% and h = 20 Years
P = 10,000 x 7,469
P = 74,690 = Cost of System Changes
The present worth value of the losses saved each year can be calculated in a similar manner as follows:
P = $12,841/yr. × 7.469 825
= $95,909
Consequently, the addition of this function will save the utility approximately $21,000 over the life of the equipment.
Reference
1. Campbell, H.E., Distribution Economics, Chapter 3, Power Distribution System Course, PTI-1982.
Questions
1. If we have no inflation, then $500 today is equal to $500 in 5 years. (true or false?)
2. If 12% is the established annual simple interest rate, then $100 today is equivalent to $____ 3 years from now.
 Page 348
3. If 12% is the established annual compounded interest rate, the $100 today is equivalent to $____ 3 years from now.
4. Why must a utility maintain a good ROI (return on investment)?
5. Return is the total depreciation value of the plant. (true or false?)
6. The cost of stock money is called interest. (true or false?)
7. The return of capital or return of investment is referred to as _____.
8. Carrying charge is composed of what areas?
9. Load factor is defined as the ratio of the average power loss over a designated period of time to the maximum loss occurring in that
period. (true or false?)
10. The cost of transformer losses can be broken down into what two categories?
11. The no load losses of a 10 kVA transformer are _____ (higher, lower) than a 37.5 kVA transformer?
12. The total losses of a 10 kVA transformer are always less than a 37.5 kVA transformer. (true or false?)
13. What three factors should be considered in the evaluation of the total investment in a distribution transformer?
14. When economically comparing two different schemes having different service lives, what must you do?
 Page 349

Index
A
Arresters 182-193, 258-259
Classes 184-185
Considerations in the Applications of MOVs 192-194
MCOV 186-188
Selection of Arresters 185-192
Silicon Carbide vs. MOV Arresters 182-184
TOV 188-191
Automation 255-257
Autotransformers 33-35
B
BIL Deterioration 202-203
C
Capacitor Application 64
KVAR and Power Factor 67-76
Load Factor and Loss Factor 96-99
Losses 88-91
Optimal Capacitor Placement 92-95
Real and Reactive Power 64-67
Switched vs. Fixed Capacitors 81-83
Types of Capacitor Control 84-87
Voltage Drop 77-79
Voltage Rise 80-81
Capacitor Control 84-87
Capacitor Fusing 142
Capacitor Placement 92-95
Capacitor Switching 176-179
Capital Recovery Factor 323-327
Carrying Charge 328-331
Cold Load Current 122-123
Compound Amount Factor 323
Compound Interest Factors 326-327
Constant Voltage Transformers (CVTs) 299
Coordination of Devices 129
Backup 134-138
Capacitor Fusing 142
 Page 350
Current Limiting Fuse 132-134
Current Limiting Fuse to Expulsion Fuse 130-132
Distribution Transformer Fusing 139-141
Expulsion 129
Instantaneous Reclose 150-151
Relay-to-Fuse 146-148
Relay-to-Recloser 157-162
Sectionalizers 162-165
Setting an Overcurrent Relay 142-145
Coordination of Fuses 111-112
Current Chopping 179-181
Current Limiting Fuse 107-111, 132-134
D
Depreciation 329
Design 1
Distribution Substation 2
Feeders 4-7
Grounding 16-19
Network Systems 7-9
Overhead vs. Underground 12-16
Primary Loop 5
Primary Selective 6
Radial 4
Secondaries 10-11
Secondary Selective 6
Spot Network 7
Utility System 1
Distribution Substation 2
Distribution Transformer Connections 47
Delta-Delta 49-50
Delta-Wye 49
Grounding-Banks 51
Open Delta 50
3q Secondary Voltage 47
Wye-Delta 49
Wye-Wye 50
Distribution Transformer Fusing 139-141
Distribution Transformers 38-42
Diversity 332-333
Dry Type 44-45
 Page 351

E
Economics 320-347
Annual Carrying Charge 328-331
Capital Recovery Factor 323-327
Compound Amount Factor 323-327
Compound Interest Factors 326-327
Depreciation 329
Diversity 332-333
Equal Payments 323-327
Future Worth 320-321
Insurance 330
Load Factor 333-334
Loss Factor 334-335
Operating and Maintenance Expenses 330
Present Worth 322-323
Present Worth Factor 323-327
Responsibility Factors 335
Return On Investment 328-329
Sinking Fund Factor 323-327
Tax 329-330
Effect of Lateral Length 217-220
Electromagnetic Fields 305-312
Grounded vs. Ungrounded 307-308
Line Compaction 310-311
Single-Phase vs. Three-Phase 309
Unbalance 311-312
Underground 309
Equal Payments 323-327
Expulsion 129
F
Failure Rates 244-246, 261-262
Fault Characteristics 20
Fault Currents 21-26
Fault Selective Feeder Relaying 246-248
Feeders 4-7
Ferroresonance 174-176
Flicker 277-279
Fuses 101, 259
Coordination of Fuses 111-112
Current Limiting Fuses 107-111
Expulsion Fuses 104-107
 Page 352
Future Worth 320-321
G
Ground Fault Current Flow 25-26
Grounded vs. Ungrounded 307-308
Grounding 16-19
Grounding-Banks 51
H
Harmonics 271, 282-287
High Impedance Faults 24-25
Higher Voltages 252-254
I
Induced Strikes 231-232
Inrush 26-27
Instantaneous Reclose 150-151
Instantaneous Relays 119-120
Insulation Coordination 194-219
BIL Deterioration 202-203
Effect of Lateral Length 217-220
Effect of Traveling Waves 207-220
Factors Affecting Margins 200-204
Margins for Overhead 194-197
Margins for Underground 198-200
Negative Trapped Charge 211-212
Quadrupling 212-214
Rate of Rise 200-202
Tapped Lateral 214-216
Voltage Doubling 208-210
Insurance 330
Interruptions 288-291
Isokeraunic Level 173-174
Isolation Transformers 300
L
Leakage Impedance 31-32
Lightning Characteristics 168-174
Isokeraunic Level 173-174
Magnitudes 171
Multiple Strokes 172-173
Polarity 173
Rate of Rise 172
Stroke Leader 168-170
 Page 353
Time Duration of Lightning 170-171
Line Compaction 310-311
Line Conditioners 301-303
Line Insulation 220-223
Line Protection 220-232
Comparison of Line Protection Schemes 227-231
Induced Strikes 231-232
Line Insulation 220-223
Types of Lightning Protection 223-227
Load Factor 96-99, 333-334
Load and Fault Characteristics 20
Fault Currents 21-26
Ground Fault Current Flow 25-26
High Impedance Faults 24-25
Inrush 26-27
Low Impedance Faults 21-24
Typical Feeder Load Characteristics 20
Loss Factor 96-99, 334-335
Losses 88-91
Low Impedance Faults 21-24
M
Magnetizing Impedance 32
Margins for Overhead 194-197
Margins for Underground 198-200
MCOV 186-188
Multiple Strokes 172-173
N
Network Systems 7-9
Network Transformers 42-44
O
Outages 243-244, 270
Overcurrent Relays 114-118
Overhead vs. Underground 12-16
Overvoltage 174-181, 272
Capacitor Switching 176-179
Current Chopping 179-181
Ferroresonance 174-176
P
Polarity 173
Power Factor 67-76
 Page 354
Power Line Conditioning 296-305
Constant Voltage Transformers 299
Isolation Transformers 300
Line Conditioners 301-303
Surge Suppression 297-298
Uninterruptible Power Supplies 303-305
Power Quality Definitions 270-272
Harmonics 271
Outages 270
Overvoltage 272
Surges 270
Undervoltage 271
Voltage Sags 271-272
Voltage Swell 272
Power Quality vs. Reliability 246-251
Blocking the Instantaneous 248-251
Fault Selective Feeder Relaying 246-248
Present Worth 322-323
Present Worth Factor 323-327
Primary Loop 5
Primary Selective 6
R
Radial 4
Rate of Rise 172, 200-202
Reclosers 124
Controls 125-126
Ratings 126
Regulators 54
Regulator kVA Rating 54
Single-Phase 54-56
Three-Phase 56-62
Relays 113
Characteristics 113-123
Cold Load Current 122-123
Instantaneous Relays 119-120
Overcurrent Relays 114-118
Reliability 234-238
Reliability Factors 252
Arresters 258-259
Automation 255-257
 Page 355
Fuses 259
Higher Voltages 252-254
Switches 259-261
System Design 254-255
Reliability Fundamentals 234-246
Calculating Reliability 234-238
Causes of Outages 243-244
Failure Rates 244-246
Reliability Indices 238-241
Survey Results 241-243
Reliability Indices 238-241
Responsibility Factors 335
Return on Investment 328
S
Sags 279-280
Secondaries 10-11
Secondary Faults 46-47
Secondary Selective 6
Sectionalizers 127-128, 162
Selection of Arresters 185-192
Sensitive Loads 292-296
Appliances 292
Computers 293
Sinking Fund Factor 323-327
Spot Network 7
Stray Voltage 312
Mitigation 315-316
Sources 313-315
Stroke Leader 168-170
Substation Transformers 36-38
Surge Suppression 297-298
Surges 270, 291
Swells 281-282
Switches 259-261
System Design 254-255
T
Tax 329-330
Time Duration of Lightning 170-171
TOV 188-191
Transformers 29