Wagner Evan

t
SYMMETRICAL
COMPONENTS
AS APPLIED TO THE ANALYSIS
OF UNBALANCED ELECTRICAL CIRCUITS
BY
C. F. WAGNER
R esearch Enoineer,
Westinohouse Electric and Manufacturino Company; Member A.I.E.E.
AND
R.D.EVANS
Manaoer, Transmission Enoineerino,
Westinohouse Electric and Manufacturino Company; Member A.l.E.E.
WITH AN INTRODUCTION BY
C. L. FORTESCUE
USED AS A TEXT IN THE DESIGN COURSE

OF THE WESTINGHOUSE GRADUATE SCHOOL
FmsT EDITION
N INTH lMPRESSION
McGRAW-HILL. BOOK COMPANY, INc.

NEW YORK AND LONDON
1933
CoPYRIGHT, 1933, BY THE
McGRAW-HILL BooK CoMPANY, INc.

PBINTl!lD IN THEl UNITEID 8TATE8 0:1' AMl!lBICA
All rights reserved. This book, or

parts thereoj, may not be reproduced
in any form without permission of
the publishers.
PREFACE
The solution of unbalanced electrical circuits has been found
to be practicable only by the method of symmetrical components.
It is a very powerful analytical tool and is based on sound theory.
The method ha~ been applied to advantage in the solution of
practically all phases of power system engineering, particularly
in the investigation of conditions resulting from unsymmetrical
transient disturbances.
The value of the method was recognized immediately upon
its publication in 1918. There is now considerable literature
on the subject of the solution of unbalanced polyphase circuits
by this method. In spite of this there is no text which gathers
together the fundamental theory on which the method is based
and discusses the severa! principal applications. In the hope
of meeting this need the present volume has been prepared.
·\. The method of symmetrical components was discovered by
~. Dr. C. L. Fortescue while investigating problems of single-phase
railway systems. The early applications of the method were
made principally by Dr. Fortescue and his associates, Messr~.
Gilman, Peters, Chubb, and Dr. Slepian of the Westinghouse
' Electric and Manufacturing Company. During the past severa!
years the authors have devoted their efforts principally to
application problems of power systems. In this work they have
·¡ found it necessary to extend the method particularly in connec-
tion with the determination of fault currents under unbalanced
conditions and of stability under unsymmetrical faults.
This book is intended to meet the needs of the practicing
engineer and the engineering student. It presupposes a knowl-
edge of fundamental single-phase and polyphase circuit theory
and the operating characteristics of the more important types of
alternating-current apparatus. It is, therefore, intended for
senior students and graduates in electrical engineering. Because
of its recent development the practicing engineer frequently will
not have had an opportunity to study the subject in formally
organized classes. Special attention has therefore been given
y
vi PREFACE
to working out, as a part of the text, illustrations of the appli-

cation of the method. In addition, problems have been included
in order to provide exercises and a convenient means for testing
the amount of information gathered from study of the individual
chapters. The problems have been drawn principally from those
encountered by the authors in connection with practica! engineer-
ing work. It is believed, therefore, that they will be of special
interest to the practicing engineer.
Chapters I to X cover the mor~ essential phases of the funda-
mental theory and the principal application of the method, which
is the determination of voltages and currents under unbalanced
faults. The remainder of the volume takes up more highly
specialized problems. For those interested only in the subject
of measurements, Chaps. I, II, XIII, XIV, and XV will be of
greatest value. The Appendix to this volume contains much
convenient reference material. It includes a review of the
principal equations upon which the method of symmetrical com-
ponents is based and also a number of convenient formulas for
network simplification and solution. Particular attention is
called to the tables of conductor characteristics, since they
include not only the ordinary (positive-sequence) constants
but also the ground return or zero-sequence constants. It is
hoped that the Appendix will provide all of the additional
reference material required for ordinary unbalanced fault calcu-
lations by the method of symmetrical components.
In presenting this volume the authors wish to acknowledge the
inspiration and help which they have received from Dr. C. L.
Fortescue, Mr. J. F. Peters, Dr. J. Slepian, and Mr. C. A. Powel;
and the use of both published and unpublished work, including
particularly an extensive memorandum prepared by Mr. Peters
for use in connection with the Westinghouse Design School.
The authors wish also to express their appreciation of assistance
from their associates, particularly from Dr. W. A. Lewis for many
general suggestions and especially for the use of certain tables
given in the Appendix; Mr. S. B. Griscom and Mr. E. L. Harder
for general suggestions and for assistance in proofreading; and
Mr. W. R. Ellis for the preparation of the majority of the illus-
trations and for the numerous calculations required.
EAST PITTSBURGH, PA. c. F. WAGNER.
May, 1933. R. D. EVANS.
CONTENTS
PAGE
PREFACE . . . . . . . . . V
lNTRODUCTION, BY c. L. FORTESCUE . . . . xiii
CHAPTER I
!NTRODUCTION AND HISTORICAL DEVELOPMENT. . . . . . . . . . . 1
What Is Symmetrical Components-Short Historical Review-
Application of the Method
CHAPTER II
FUNDAMENTAL PRINCIPLES. . . . . . . . . . . • . . • . . . . . 9
Laws of Complex Numbers- Vector Representation of Alternating
Quantities-Symmctrical Component Systems- Resolution of
Three Vectors into Symmctrical Components- Star-dclta Trans-
.formations-Independence of the Sequcnces in Symmetrical Sys-
tems-Sequence Impedances
CHAPTER III
CALCULATION OF UNBALANCED FAULTS . . . . . . . . . . • . . • 26
Kirchhoff's First Law-Kirchhoff's Second Law-Sequence Net-
works-Sequence Networks of a Particular System Including
Shunt Loads-Connection of N etworks to Represent Faults-
Single Line-to-ground Fault-Double Line-to-ground Fault-
Line-to-line Fault- Three-phase Fault-Application to Calcu-
lating Boards-Representation of Fault Impedances-Short-circuit
of Generators-Numerical Example of Generator-fault Calculations
CHAPTER IV
!LLUSTRATION OF UNBALANCED F AULT CALCULATIONS . . . . . . . . 52
Conventional Simplifying Assumptions in Short-circuit Calcu-
lations-Layout of Assumed System-Single-line Diagram of the
System and the Thrce Sequence Nctworks-Outline of Short-circuit
Calculations-Equivalent Impedances and Distribution Factors-
Single Line-to-ground Fault Calculations-Detcrmination of
Sequence Currents-Phase Currents from Sequence Currents-
Determination of Sequence Voltages-Line-to-neutral Voltages
from Sequence Voltages-Line-to-line Voltages-The Conversion
of Currents and Voltages Due to Transformers- Double Line-to-
. ground Fault Calculations-Line-to-line Fault Calculations-
Comparison of Short-circuit Calculations and Test Results
Vll
viii CONTENTS
PAGE
CHAPTER V
CONSTANTS OF SYNCHRONOUS MACHINES. . . . . . . . . . . . . . 74
Three-phase Short-circuits of Generators-Alternating-current
Components- Time Constants-Direct-current Components-
R.M.S. Total Current-Short-circuit of Loaded Generator-Ex-
ternal R eactance- Example of Decrement Calculations-Nega-
tive-sequence R eactance-N egative-sequence Resistance-Zero-
sequence Impedance-Typical Values of Synchronous Machine
Constants-Unbalanccd Faults-Decrement Curves-Interna!
Voltage Method- Effect of Varying Exciter Voltage
CHAPTER VI
CoNSTANJs OF TRANSFORMERS . . . . . . . . . . . . . . . 111
Two-winding Transformers, Series and Shunt Impedances-Three-
winding Single-phase Transformers- Three-phase Transformers-
Autotransformers-Inclusion of Series Impedance in Autotrans-
former Circuits- Interconnected Star-delta Transformers
CHAPTER VII
ÜONSTANTS OF SHORT TRANSMISSION LINES WITHOUT ÜROUND WIRES 136
Positive- and Negative-sequence Impedance of Single Circuits-
Positive- and Negative-sequence Reactance of Parallel Oircuits-
Zero-sequence Impedance-Carson's Formulas-Physical Con-
ception and Factors Affecting Earth-return Circuits-Simplified
Formulas for Zero-sequence Impedances-Self Impedance of
Parallel Conductors with Earth Return-Mutual Impedance
between Two Circuits with Earth Return-Self Impedance of Two
Identical Parallel Polyphase Circuits with Earth Return-Equiva-
lent Circuits for Parallel Lines with Mutual Inductive Coupling
CHAPTER VIII
CoNSTANTS OF SHORT TRANSMISSION LINES WITH GROUND WIRES . . 163
Positive- and Negative-sequence Impedance-Application of Car-
son's Formulas to the Calculation of Zero-sequence Impedance-
Practical Calculation of Zero-sequence Impedance-General
Method for Zero-sequence Calculations-Steel Ground Wires-
Copper-clad Ground Wires-Sequence Impedance Constants of
Typical Circuits
CHAPTER IX
CoNSTANTS oF LoNG TRANSMISSION LINES . . . . . . . . . . . . . 182
Positive- and Negative-sequence Characteristics-Zero-sequence
Susceptance-Mutual Susceptance between Two Parallel Circuits
-Zero-sequence Equivalent Network for a Single Transmission
Circuit-Parallel Circuits with Distributed Inductive and Capaci-
tive Coupling-Susceptance Calculations for Circuits with Ground
Wires
CONTENTS lX
PAGE
CHAPTER X
CONSTANTS OF CABLES. . . 198
Positive- and N egative-sequence Resistance-Zero-sequence Re-
sistance--Positive- and Negative-sequence Reactance-Zero-
sequence Reactance-Shunt Capacitive Reactance-Shielded or
Type H Cable-Submarino Cable-Constants of Typical Cables
CHAPTER XI
PowER SYsTEM VoLTAGES AND CuRRENTs UNDER FAULT CoNDITIONS 222
Assumptions-Formulas for Line Currents and Voltagcs between
Line Conductora and to Ground-Basis for Cnrrent and Voltage
Curves-Fault Current and Voltage Curves-Discussion of Curves
-Systeros Giving Maximuro Voltage or Current Ratios-Effect
of Method of Grounding. Zero-sequence Impedance Ratio-
Positive-sequence Voltages at the Fault
CHAPTER XII
SIMULTANEOUS FAULTS . . . . . . . . . . . . . . . . . . . . . 238
Simplest Equivalent Network-Dctermination of Branches of the
Forro Za.,-Circuit Equations for Simplest Equivalcnt Network-
Impedance-drop Constants of the Form Du and D., 11-Discussion
of Network Restraints-Two Simultaneous Faults with Two
Sources of E.M.F.-Restraints Imposed by the Sequence Network
Constants-Restraints Iroposed by the Type of Fault-Solution
by Corobination of Restraints Due to N ctwork Constants and
Types of Faults-Simultaneous Faults at More than Two Loca-
tions-Siroultaneous Faults for the Case of a Single Source of
E.M.F.-Current and Voltage Distribution for Siroultaneous
Faults-Equivalent Circuits Representing Fault Conditions-
Single Fault for the Case of Several Sources of E. M.F.
CHAPTER XIII
DETERMINATION OF SEQUENCE QuANTITIES FROM PHASE QuANTITIES 259
Determination of Vectors from Amplitudes-Analytical Expres-
sions- Graphical Constructions-Charts- Unbalance Factor
CHAPTER XIV
MEASUREMENT oF SEQUENCE VoLTAGES AND CuRRENTS . . . . . . . 271
Measureroent of Zero-sequence Voltage-Measurement of Positive-
and Negative-sequence Voltages-Common Form of Positive- and
Negative-sequence Voltroeters Applicable in the Abscnce of Zero-
sequence Voltage--Second Forro of Positive- and Negative-
sequence Voltroeters Applicable in the Absence of Zero-sequence
Voltage--Red.uction and Elimination of Voltage Transformers for
Sequence Networks-Simultaneous Measurement of Sequence
Voltages-Measurement of Zero-sequence Current-Measurement
of Positive- and Negative-sequence Cu'rrents-Coromon Form of
Positive- and Negative-sequence Ammeters Applicable in the
X CONTENTS
PAGE
Abscnce of Zero-sequence Current-Bridge Type of Positive- and
N egative-sequence Ammeters-Elimination of Zero-sequence
Current- Simultancous Measurements of Sequence Currents-Fre-
quency and Transient Errors-Adjustment and Testing of
Segregating N etworks- Application of Sequence Voltage and
Current Devices-Positive-sequence Graphic Voltmeter-Posi-
tive-sequence Voltage Regulator-Negative-sequcnce Overcurrent
R elay-Polyphase Sequence Networks and lnstruments
CHAPTER XV
MEASUREMENT OF PowER QuANTITIES ON PoLYPHASE CIRCUITS . . . 306
Power Expressed in Terms of Voltage and Current Vectors-
Reactive Volt-amperes Expressed in Terms of Voltage and Current
Vectors- Power Quantities on Single-phase Circuits-Power Quan-
tities on Polyphase Circuits-Measuring Devices for Single-phase
Power Quantities-Total Power Quantities in Terms of Sequence
Quantities-Power Quantities on Unbalanced Three-phase Cir-
cuits Analyzed by Sequence Componcnts-Analysis of Meter
Connections for Power Quantities on Three-phase Circuits- Cross
Connections for Measuring Reactive Volt-amperes- Meters for the
Measurement of Sequence Power Quantities-Flow of Power
Quantities Dueto Unbalance
CHAPTER XVI
M ULTIPHASE SYSTE:..IS. . . . . . . . . . . . . . . . . . . . . . 328
Resolution of Multiphase Systems into Symmetrical Components-
Two-phase and Four-phase Systems-lrregular Systems-General
Treatment of Polyphase Circuits-lmpedances of Symmetrical
Polyphase Systems-Harmonics
CHAPTER XVII
INDUCTJON MOTORS. . . . . . . . . . . . . . . . . . . . . . . 345
Characteristics with Application of Positive-sequence Voltages to
the Stator-Characteristics with Application of Negative-sequence
Voltages to the Stator- Characteristics with Application of Both
Positive- and Negative-sequence Voltages to the Stator-Appli-
cation of Unbalanced Voltages to the Stator of a Particular Motor
-Operation with Open-circuited Stator Phase-Unbalanced Im-
pedances in the Rotor Circuits-Operation with Open-circuited
Rotor Phasc
CHAPTER XVIII
UNSYMMETRICAL SYSTEMS AND PHASE-BALANCERS. . . . . . . . . . 363
Voltage Drops in Three Unsymmetrical Self Impedances-Dis- ·
cussion of the Impedances Bo, ,81, and ,82 of Unsymmetrical Self
Impedances-General Static Network Containing Unsymmetrical
Self and Mutual Impedances-Symmetrical Transformations-
General Case of Unsymmetrical Systems-Independence of the
CONTENTS
.
XI.
PAGB
Sequence Components in a Symmetrical General Network-
Systems Combining Symmetrical and Unsymmetrical Pbase Im-
pedanccs- Unsymmetrical Series Impedances-Power Supply to
Single-phase Loads-Single-phase Generation-Phase-balancing
of Single-pbase Loads- Rotating Balancers- N egative-scquence
E.M. F.- type Phase-balanccr- Im pedance-type Phase-balancers-
Phase-converter-Balanced Polyphase Voltages from a Single-
phase Source
BmLIOGRAPHY . . . . . . . . . . . . . . . . . . . . . . . . . 391
Englisb Language References- Foreign Language References
APPENDIX
I. NOTATION . 403
II. GENERAL FORMULAS. 404
III. APP ARATUS CoNSTANTS 406
IV. TRANSMISSION LINE CoNSTANTS. 407
V. CABLE CONSTAN'l'S. 409
VI. EQUIV ALENT NETWORKS 414
VII. CHARACTEH.ISTICS OF CONDUCTORS. 419
VIII. EXPONENTIAL FUNCTIONS 426
INDEX. 427
INTRODUCTION
The complex number arose from attempts by early mathe-
maticians to solve algebraic equations, sorne of which had
known geometrical solutions, in terms of algebraic operations.
The first to make use of the minus sign under a radical was
the Italian mathematician, Cardan, and the dual numbers
he obtained were the same as our present complex numbers.
Although many geometric solutions of algebraic equations
soluble algebraically only by such numbers had been obtained
in the past, it was not until the nineteenth century that the
correspondence between the system of real and complex numbers
to points of the two-dimensional Cartesian continuum was
realized. Until this time such numbers were considered impos-
sible or imaginary numbers as stated in Euler's algebra as late
as 1770. It is peculiar that mathematicians befare this time
did not realize that the system of positive and negative real
numbers w~s not a closed system, since in that domain the square
root of a negative quantity had no meaning. With the inclusion
of the complex and so-called imaginary nnmbers the system of
numbers was closed for all algebraic operations.
In spite of the fact that until the nineteenth century no
interpretation had been given to the complex number, its use
had become quite general among mathematicians. So we find
DeMoivre introducing it into trigonometry through bis famous
theorem
(cos e + i sin e)n = (cos ne + i sin ne)
:. while Euler introduced the identity
eiB = COS e + Í sin 8
and
é'll" = -1
It is remarkable that these mathematicians shot so close
· without hitting the bull's-eye. Apparently, in spite of the
, development of trigonometry and its relations with algebra, the
·ooncept of a vector to define a point in a plane had not yet
been realized and the symbol eiB was still regarded as a rather
bsurd algebraic deduction which nevertheless gave useful
·resulta. The remarkable beauty of the conception of eie as a
Xlll
xiv INTRODUCTION
unit vector or coordinate oriented at the angle () from the refer-
ence line or vector had not yet dawned on them and therefore
the full meaning of DeMoivre's theorem had not been realized.
Otherwise Euler from his identity
é1r = -1
would have at once deduced the geometric meaning of i = .y'-=1
by the relation
ei~ =y -1
which is a unit vector at right angles to the line of reference.
The validity of these interpretations and the correspondence
-- of the system of real and complex numbers to points on the two-
dimensional Cartesian continuum were established by Gauss,
Abel, Cauchy, and Weirstrasse, and the representationof complex
numbers by vectors in a plane was introduced by Argand, and
diagrams showing the composition and resolution of such num-
bers in a plane are known as Argand diagrams. These dis-
coveries were the basis of the theory of functions of a complex
variable established by Cauchy, Weirstrasse, and Riemann.
Gauss, Abel, Cauchy, and Weirstrasse showed that the system
of numbers extended to embrace complexas well as real numbers
is closed to infinite processes of analysis as well as to purely
algebraic operations. Consequently, all linear operations per-
formed on a complex magnitude will give a number in this
system. Two complex numbers are said to be conjugate to
each other when their real parts are equal in magnitude and sign
and their imaginary parts are equal in magnitude but opposite
in sign. If a linear operation is performed on a complex number
and its conjugate, the resulting numbers will be conjugate to
each other.
The result of this extension of the domain of number was
far-reaching; it brought plane geometry under the domain of
algebra, and trigonometry became merely a particular branch of
algebra dealing with the relation of the component parts of a
complex unit number to its direction with respect to sorne fixed
direction in the complex number plane.
The use of complex numbers in the theory of alternating
currents was first introduced by Dr. A. E. Kennelly and the late
Dr. C. P. Steinmetz. The treatment of periodically varying
functions in terms of vectors very greatly simplified the solution
of all of the alternating-current problems.
INTRODUCTION XV
In balanced polyphase problems the phase displacement

between currents in the ·various phases and the voltages in the
various phases are equal, and because of this symmetry it is
possible to reduce the solution of the problem to that of an
equivalent single-phase problem. Unbalanced current and
voltage conditions, even in a polyphase system in which the
constants are symmetrical, offer very considerable difficulty
when attacked by classical methods. In such cases it is necessary
to solve the phase currents and voltages simultaneously introduc-
ing self and mutual constants between phases. In addition,
where rotating machines are involved, it is necessary to introduce
impedances relating the rotor and stator circuits. In my early
investigation of phase balancers I observed that certain sym-
metrical relations between phase currents and also between
phase voltages recurred frequently, which led meto the investiga-
tion of the general problem of unbalance. This investigation
finally led to the discovery of the fundamental principies of the
Method of Symmetrical Coordinates which I published in 1918.
In this paper it is shown that unbalanced problems can be .solved
by the resolution of the currents and voltages into certain sym-
metrical relations. When the constants are symmetrical, that
! is, when the system viewed from any phase is similar, then the
r· symmetrical components of currents do not react upon each
t..·other so that it becomes possible to eliminate the mutual relations
~ with their attendant complication in the solution of the problems.
f' These relations are discussed in detail in this volume.
~.· In many respects the method of symmetrical coordinates
~ bears the same relation to the solution of unbalance in polyphase
~ alternating-current problems that the complex variable bears to
the solution of single-phase and balanced polyphase problems.
The method of symmetrical coordinates is analogous in sorne
respects to the resolution of a periodic function into its funda-
mental and higher harmonics by Fourier series. By the method
of symmetrical coordinates a set of unbalanced currents may be
resolved into systems of balanced currents equal in number to
the number of phases involved.
The expression "symmetrical coordinates" is preferable to
the more commonly used expression "symmetrical components."
The former is more logical from a mathematical sense in that it
refers to the axes of reference, but the latter name is probably
more descriptive of the process to the laymen,
xvi INTRODUCTION
1 have had the pleasure of close association with the authors

in their work, and the technical development of the system of
symmetrical components owes a great deal to their initiative,
enthusiasm, and untiring work during the past several years.
This book will be welcomed by engineers as it is the first text-
book brought out entirely devoted to the subject of symmetrical
components.
C. L. FoRTESCUE •
•
r
¡
~
~·
r
¡
\'
SYMMETRICAL
. COMPONENTS
¡
1
t
~
CHAPTER I
1
!·• · INTRODUCTION AND HISTORICAL DEVELOPMENT

~
The solution of balanced polyphase circuits is usually accom-
plished by converting the constants and the applied voltages to
per phase values and solving for one of the phases in a manner
similar to single-phase circuits. The mutual reactions between
phases may be represented by equivalent self impedances because
the symmetry of the pro blem determines the magnitu de and
phase position of the other phase currents. The currents and
voltages for the other phases are equal in magnitude to those of
the first but displaced symmetrically in phase position.
The solution of unbalanced polyphase circuits . or balanced
circuits with unbalanced terminal conditions does not permit
of the same simplification. By the older methods of analysis
it was necessary to assign symbols to the quantities in all the
phases and carry t hrough the phase solutions simultaneously.
This complicates the problem enormously. The modern method
of analysis presented in this volume reduces the solution of such
problems toa systematic form and results in considerable simpli-
fication, especially for balanced systems operating under sorne
kind of unbalance such as a line-to-line, single line-to-ground, or
double line-to-ground fault. This new analytical tool is called
the method of symmetrical components.
1. What Is "Symmetrical Components"?
The fundamentals of the method of symmetrical components
are in reality quite simple. No very great amount of mathe-
matics is required for its understanding. It is on t he contrary
characterized by a large number of relatively simple steps, and
once the fundamentals are mastered the application of the
principies becomes relatively simple. One can in a very short
time indicate these fundamentals and point out their field of
application.
.[
2 SYMMETRICAL COMPONENTS
Consider for a moment the balanced vectors in Fig. 1 (a)

of which Eal, Eb1, and Ecl are the line-to-neut ral voltages of
phases a, b, and e, respectively, of a three-phase system. The
instantaneous values of these vectors are represented by the
projection of the vectors upon the X-axis. With the conven-
"' b e ~ b e
(01) Posii'ive-Se'fuence (b)
. t l l l eJ S
b a5e " '
b
• x-axis
;.....;.:::-=----'-'-__:.;.;_;.:.=.__
(e) Ne~c:ftive- seqvenc:e (ol)
\~l~o
) (e) 'C7(f) 'C7
Zero-seotuenee
FIG. 1.-Sequence components and their combinations to form unbalanced phase

quantities.
tional rotation of vectors in the counter-clockwise direction the

lnstantaneous values of the voltages can be developed as shown
in Fig. l(b). It will be observed that the arder of maxima occur
in the sequence abcabc and for this reason they are said to be
a "positive-sequence" set of vectors. Note that this sequence
has no referencé to the direction of rotation of the vectors
themselves.
i
INTRODUCTION AND HISTORICAL DEVELOPMENT 3
In Fig. l(c) is shown another system of balanced voltages.
Again assuming the conventional direction of rotation of the
vectors the instantaneous values of the voltages may be developed
as shown in Fig. l(d). For this case the sequence of the maxima
of the instantaneous values occurs in the order acbacb or cba
which, since it is opposite to the previous one, is called the
"negative-sequence" set of vectors. Note again that the
sequence of maxima has no relation to the arbitrarily chosen
direction of rotation of the vectors.
Consider lastly another type of three balanced vectors, such
as that shown in Fig. l(e), in which all three vectors are in phase.
The maxima of the instantaneous valu es as shown in Fig. 1(f)
must of course all occur simultaneously, and this set of vectors
has been given the name "zero-sequence."
These voltages may exist in separate systems or simultaneously
in the same system. In the latter case phase a will be composed
of Eal from the positive-sequence set of vectors, Ea2 from the
negative-sequence set of vectors, and Eao from the .zero-sequence
set of vectors. The resultant value for this phase is shown in
Fig. l(g). Phase b will consist of the corresponding b components
from the three sequences, and the summation for this phase is
also shown in Fig. l(g). The total voltage for phase e is obtained
in a similar manner. It may be seen therefore that the simul-
taneous presence of three sets of balanced voltages of the type
described results in a set of unbalanced voltages. Currents can
of course be analyzed in a similar manner.
One purpose of the method of symmetrical components is
to show how the three unbalanced voltages can be built up in
an analytical manner from the three fundameTttal sequences.
Another purpose is to show how any three unbalanced voltages
can in turn be resolved into three sets of three balanced or
symmetrical components. The question naturally arises, why
should the resolution of three vectors into nine vectors necessarily
simplify the solution. The answer líes in the fact that the
resolution results in three symmetrical systems, each one of
which in a balanced system can be treated separately just as
the balanced polyphase problems in the past have been solved
by reducing the constants and voltages to per phase values and
solving on a single-phase basis. In symmetrical circuits, currents
and voltages of different sequences do not react u pon each other;
currents of the positive-sequence produce only voltage drops of
4 SYM METRICAL COMPONENTS
positive-sequence, currents of the negative-sequenc~ only voltage

drops of negative-sequence, and similarly, currents of the zero-
sequence only voltage drops of zero-sequence. This fortunate
circumstance results in considerable simplification of all kinds of
problems involving asymmetry such as that introduced by
short-circuiting conductors of a system either together or to
ground, singly or in pairs, or by the open-circuiting of a conductor.
The resolution of the problem into the sequence components
has a further advantage in that it isolates the quantities into
components which represent a better criterion of the controlling
factor or factors in certain phenomena. In stability investiga-
tions the synchronizing force between machines is affected
principally by the positive-sequence quantities. The demagnet-
izing factor of armature current in synchronous machines is
measured by the positive-sequence component of current only
so that it is this component of current which determines excita-
tion requirements. The performance of damper windings,
as to both h~ating and torque, is responsive to the negative-
sequence component only. Ground relays ·and grounding phe-
nomena in general are very closcly associated with zero-sequence
components. Power quantities can be resolved into components
associated with the different sequences. These and many other
subjects attest to the value of resolving the problem into
the sequence components; the solution of the problem itself
is not only simpler, but the results are resolved into factors which
can be investigated separately. While the foregoing considered
only three-phase systems, the fundamental conceptions are
applicable to systems of any number of phases.
Most of the apparatus used in practice such as generators
and condensers, induction motors and transmission lines and
cables (to within a sufficient degree for practica! purposes)
are of the symmetrical type. In fact special precautions are
taken to insure that they are symmetrical, otherwise undesirable
featurcs, such as telephone interference and extra losses, may be
introduced. It may be seen, therefore, that this simplification
is applicable to the vast majority of systems in commercial use.
In unsymmetrical systems currents of one sequence produce
not only voltage drops of the same sequence but may produce
voltage drops of the other sequences as well. In the method
of symmetrical components, rules are developed for solving the
fundamental relation between the sequences in such systems.
IN7'RODUCTION AND HISTORICAL DEVELOPMENT 5
2. Short Historical Review.
The germ of the idea may be traced to the early analysis
of single-phase motors by Ferraris, Lamme, and others, about
1895. As a part of this analysis it is shown that the field set
up in the single-phase motor can be resolved into two revolving
flux fields rotating in opposite directions. Somewhat later,
unbalanced currents from three-phase machines were resolved
into the two sets of components that are now known as the posi-
tive- and negative-sequence components and their effects were
analyzed in terms of the positively and negatively rotating
fields in the machine associated with the positive- and negative-
sequence components, respectively. This conception was utilized
by E. F. W. Alexanderson in connection with his work on phase-
balancers,<l)* a qualitative treatment of which was published
in 1913, and by L. G. Stokvis in the determination of generator-
voltage regulation in terms of the phase currents, a mathematical
analysis of which appeared in 19120°1> and 1915. <2 >
Stokvis appears to have approached his analysis from the
machine point of view. He was thus limited in his resolution
of a system of vectors into sets which produced certain e:ffects
within the machine and chose for his purpose:
l. A set which produced a positively rotating field.
2. A set which produced a negatively rotating field.
3. A set which produced a pulsating field.
He failed to recognize that what was needed was a new kind of
component (the zero-sequence component) which produced
neither a rotating nor a pulsating field in a symrnetrical machine.
He thus lacked the essential element required to set up compo-
nents which did not react upon each other in symmetrical parts
of systems.
It remained for C. L. Fortescue approaching the pro blem from
a di:fferent point of view to perceive the beautiful generality
of the method as applied to all kinds of polyphase systems.
The general concept was developed by Fortescue and his asso-
ciates, R. E. Gilman, J. F. Peters, J. Slepian, and others, in
studying problems of unbalanced circuits and in analyzing the
characteristics of single-phase motors, polyphase motors with
unbalanced voltages, and synchronous motor-generator sets and
phase-balancers for single-phase railway electrification. In
* Superior numerals in parentheses refer to items in the Bibliography, page
391.
this way a new, simple, and complete method was developed for
handling the problem of unbalanced circuits, which was pub-
lished by C. L. Fortescue in his classic paper<5) in 1918.
3. Application of the Method.
In studying the method developed by Dr. Fortescue, it i~
convenient to consider its application, first, to commercial
three-phase systems which are symmetrical except for the unbal-
ance at a particular point, such as the point of fault; and, second,
to the more general case of a system which may be unsymmetrical
throughout. As applied to the first type of system, three princi-
pal developments may be recognized:
l. Introduction of zero-sequence. This device makes it possi-
ble to resolve any three unbalanced vectors into three sequence
components, namely, positive-, negative-, and zero-sequence.
2. Demonstration that in those parts of a system which are
symmetrical, the currents and voltages of one sequence have no
influence upon those of another sequence. Recognition of this
fact is important as it forms the basis for all the subsequent work
on the determination of short-circuit currents of unbalanced
faults on commercial systems.
3. Assignment to lines and apparatus of distinct impedances
for each sequence that are fixed quantities independent of each
other and of the character or amount of the unbalance.
The general method of analysis is also applicable to systems
which are unsymmetrical throughout. This case requires addi-
tional constants and somewhat more complicated relations to
define the reaction of one sequence upon another in the unsym-
metrical parts of the system. Dr. Fortescue showed how such
calculations could be simplified and systematized by means of
the "sequence operator."
He further demonstrated the generality of the method by
showing that it could be applied to systems of any number of
phases, with similar results. The same general method has
even been used to salve the cubic equation and to reduce the
degree of a higher order equation. <151 ) It is this generality of
the method which caused Dr. Fortescue to select for the unit
sequence vectors the more general term "coordinates" instead
of the term "components" which is commonly used in electrical
engineering.
INTRODUCTION AND HISTORICAL DEVELOPMENT 7
SillCe the presentation of Dr. Fortescue's paper in 1918,
the work along this line has been directed principally toward
the application of the method to the solution of problems com-
monly encountered in commercial systems, the most important
of these being the determination of system currents and voltages
under unbalanced fault conditions. This work has been much
facilitated by the concepts of sequence networks and the use
of equivalent circuits connecting these networks to represent
fault conditions. These short-circuit studies may make use
of the direct-currcnt or alternating-current calculating boards,
the first application of which was described by R. D. Evan(in
1925. The method was next applied to system-stability studies
by R. D. Evans and C. F. Wagner in 1926. Also in 1926,
A. P. Mackerras collected and presented in two articles the most
important features of the method as applied to the determination
of single-phase short-circuit currents on three-phase systems.
Another important phase of the subsequent work on sym-
metrical components has been the determination by calculation
and test of the sequence impedance constants of lines and
apparatus. This work has given considerable ímpetus to the
further studies of the short-circuits of synchronous machines
as evidenced by articles and papers by Bekku, Wagner and
Dovjikov, Park and Robertson, Doherty and Nickle, Kilgore,
and Wright. Intensive study has also been given to the zero-
sequence impedance of transmission lines and cables which is
of particular importance in connection with the problem of
inductive coordination. The fundamental work of Carson and
the theoretical and experimental work of the Joint Sub-committee
on Development and Research of the N .E.L.A. and Bell Tele-
phone System have been very valuable.
A knowledge of the method of symmetrical components is
essential for an adequate understanding of application problems
involving the magnitude a·n d phase relation of voltages and
currents under unbalanced conditions and is therefore necessary
for the design of power systems · from the standpoint of circuit-
breaker application, relay protection, and stresses, in electrical
machinery. The method is particularly suitable for the analysis
of the performance of rotating machinery for single-phase or
unbalanced polyphase operation and has been used quite exten-
sively in connection with the design of machinery on single-phase
railway ~-.:;ystcms, particularly from the standpoint of phase-

converting apparatus. A number of new schemes have appeared
commercially which make use of conceptions arising from the
method of symmetrical components, such as the negative-
sequence protective relay and the positive-sequence voltage
regulator. Symmetrical components is also advantageous in
analyzing metering means, particularly polyphase instruments on
unbalanced loads.
In recent years stability of power systems has been recognized
asan important problem, and it has been shown that the limiting
condition arises at times of faults, hence the method of sym-
metrical components ha·s been of very considerable assistance in
the analysis of stability problems.
Problems*
l. Given the self and mutual reactances between three long parallel con-
ductora located at the apexes of an equilateral triangle, show that for
balanced operation the reactances can be represented by an equivalent self
reactance expressed in terms of ohms to neutral per phase.
2. A 625-kva. 2,30ü-110-volt single-phase transformer has a reactance of
3 per cent. Determine the reactance in ohms on both the high-voltage and
low-voltage bases.
3. A 20,000-kva. bank of transformers is connected to a 132,000-volt
system. The reactance of the transformers is 9 per cent. Convert to per
cent on 100,000-kva. base. Determine the reactance in ohms per phase on
the 132,000-volt base and a 220,000-volt base.
4. A 10,000-kva., 132,00ü-13,200-volt transformer bank having a reactance
of 10 per cent is connected on the low-voltage side toa 20,000-kva. generator
having a reactance of 30 per cent. A three-phase short-circuit occurs on the
high-voltage side of the transformer. Assuming a generator interna!
voltage equal to rated volts, determine: (a) the short-circuit current on the
high-voltage side of the transformer, (b) the short-circuit current on the low-
voltage side of the transformer, and (e) the voltage to neutral on the low-
voltage side.
*A knowledge of the methods for handling this group of problems dealing
with symmetrical three-phase systems is essential for the ready understand-
ing of the treatment presented in subsequent chapters. De:ficiency in pre-
vious training in this regard may be disclosed at the outset and promptly
corrected.
••
CHAPTER II
FUNDAMENTAL PRINCIPLES
The method of symmetrical components requires no additional

knowledge of mathematics beyond that commonly used in the
solution of the ordinary alternating-current problems. A work-
ing knowledge of the fundamentals of the algebra of complex
numbers is, however, essential, and consequently this subject
will be reviewed briefly.
4. Laws of Complex Numbers.

Any complex number, such as a + jb, may be represented
by a single point, p, plotted on
y
Cartesian coordinates, in which a p
is the abscissa and b the ordinate.
This is illustrated in Fig. 2. It is b
often convenient, however, to
represent complex numbers by
another method. Referring to 0
Fig. 2, let r represent the length of Fro. 2.-A complex number.
the line connecting the point with
the origin and O the angle measured from the X -axis to the line r.
It is apparent that
.1'
a = r cosO
'·
·,
b = r sin O
or
a + jb = r ( cos O +j sin O) (1)
This is called the polar forro, in which 8 is the argurrumt or

amplitude and r the modulus or absolute value. Increasing 8 in
a positive sense rotates op counter-clockwise. Similarly
a - jb = r( cos O - j sin O) (2)
Increasing O in a positive sense in this case produces clockwise
rotation.
9
.....
If E8 be expanded into an infinite series,
Similarly,
Separating the real and imaginary components,

.
E18 =
( ()2
1- -
2!
+ -4!()4 - -6!()6 + )
+J
·( ()3
e-3,+5!-7!+
()5 () 7
·)
These expansions* are equal to cos O and sin O respectively, so
that,
Eio = cos O + j sin O (3)
Similarly,
cio = cos O - j sin O (4)
This suggests another form to represent a complex number.
Substituting (3) in (1) and (4) in (2),
a+ jb = rEi8 (5)
a- jb = rci8 (6)
These forms will be found valuable for operations involving
multiplication and division. Unless otherwise stated this angle
will be expressed in degrees throughout this volume.
Returning to equations (1) and (2) and (5) and (6), it will be
noted that the real components are equal, but that the imaginary
components have opposite signs. These numbers are called
conjugates.
In the notation used in this volume an ordinary complex
number will be represented by a capitalletter and its conjugate
by the same letter with a circumflex, for example, I and l,
respectively. If
I =a+ jb
iihen its conjugate is
l = a-- jb
* These relations may be verified by reference to any standard college
algebra.
FUNDAMENTAL PRINCIPLES 11
The absolute value of the vector will be designated by a bar
above the capitalletter, thus
1 = Va + b
2 2
The real and imaginary components may be expressed in

terms of the complex number and its conjugate as follows:
a = !(I + 1) (7)
}b = !(I - 1) (8)
The law for the addition of exponents holds equally well for
both real and imaginary quantities. Just as 103 X 104 = 107 so
from which it follows that the product of two complex numbers,

E = Eé01 and I = iE 1o2
is equal to
(9)
It will be noted that the absolute value of the product is the
. '
product of the absolute values of the components and that the
argument is the sum of the arguments.
Multiplying by é 0 merely rotates the modulus through an
angle Oin a counter-clockwise direction. This point is important
as it forms the basis for the proof of quite a number of circle
diagrams.
If a complex number be multiplied by the conjugate of another,
the argument of the product is the difference in arguments of the
components.
El = EE10¡Jé- Oz
= EJE1( 0¡- 0z) (10)
Resolved into real and imaginary components,
El = Ellcos (01 - 82) +}sin (81 - Bz)] (11)
Squaring a complex number merely increases the modulus as
the square and rotates it through double the angle.
(12)
Division, the inverse of multiplication, can be accomplished

in a similar manner.
(13)
-
12 8YMMETRICAL COMPONENTS
The absolute value of the resultant is equal to the quotient of

the absolute value~, and the argument is the difference of the
arguments of the co'mponents.
As squaring a number merely increases the modulus to the
square of the modulus and doubles the argument, extracting the
square root, the inverse, is accomplished by taking the square
root of the modulus and halving the argument. Generalizing,
the modulus of the nth root of a number is the nth root of the
modulus of the number and the argument is 1n of the argument
of the number.
~¡- ,r¡- , (J
vJei6 = vi/;, (14)
If a complex number be multiplied by its ~onjugate, the
product is a real number of magnitude equal to the square of the
absolute value of the number, for example.
II = Jei6Je- i6
= J 2eiC6- 6)
= ]2 (15)
5. Vector Representation of Altemating Quantities.

The instantaneous value of a simple harmonic function, such
as an alternating electromotive force, may be represented by the
equation
e = V2E cos (wt + a)
where
w = 27rj.
f frequency.
=
E
= · r .m .s. magnitude.
V2E = crest value.
a = angle which determines the value of e for t = O.
By the addition of equations (3) and (4) it may be seen that
cos (} = !( ef6 + €iD) and substituting wt + a for O, gives
-
v'2E cos (wt + a) = V
- JO -[ eftwt+a)
2E
+ e-i(wt+a)]
2
= V2E[ •'"é"' +2el••-"'']
so that, if we let V2Eeia = v2E and V2Ecia = V2E
e = V2(Eeiwt
2
+ Ee-iwt) (16)
It will be noted from this equation that the instantaneous

value can be represented as the sum of two oppositely rotating
vectors which are conjugate at any
y
instant, the imaginary components
canceling out. This relation is
shown in Fig. 3, in which the full
lines indicate the positions of E and
E for t = O and the dotted lines at
an instant later. It will be noted
that the two vectors are always
symmetrically disposed relative to
the X -axis so that their sum is
always real and does not contain
an imaginary component. Fur-
thermore, the projection of either
vector on the X-axis is always equal Fw. 3 .-Simp~~on~armonic func-
to one-half the sum of the vectors,
and the instantaneous value can be represented by the real com-
ponent of either vector. The same relations apply equally well
to the instantaneous values of current.
6. Symmetrical Component Systems.
Any point on a line is completely defined by the distance
from a given reference point, in either the positive direction or
the negative. Similarly a point on aplane is completely defined
by the distance from two reference axes and a point in space by
the distance from three reference axes. A point on a line, being
restrained to lie in the line, is said to possess but one degree of
freedom ; the point in the plane two degrees of freedom; and the
point in space three degrces of freedom. So also with a co-planar
1
vector which, since it is determined by the position of its terminal,
is said to possess two degrees of freedom.
Similar considerations apply to systems of vectors. For
example, a system of three co-planar vectors is completely
defined by six parameters; the system possesses six degrees of
Jreedom. When, however, one applies the restriction that the
system be symmetrical, the added restraints reduce the system
to one of two degrees of freedom. Going a step further, it is

quite conceivable that a system of three co-planar vectors with
six degrees of freedom can be defined in terms of three symmetri-
cal systems of vectors each having two degrees of freedom.
Positive-sequence System. Consider first the symmetrical
system of vectors Eat, Ebt and Ect in Fig. 4(a). Being balanced,
the vectors have equal amplitudes, are displaced 120 deg.
relative to each other, and
Eal = Eal
1240
Ebt = E E at = a 2Eat
Ect = é 120Eat = aEat
'in which
a js the unit vector, é 120 = -0.5 + j0.866
and
a 2 is the unit vector, Ei240 = -0.5 - j0.866
Because of the frequent use of this vector it will always be
designated by the smallletter a. Sorne of the properties of this
vector are given in Table I.
TABLE I.-PROPERTIES OF THE VECTOR a
1 = EiO = 1.0 + jO.O

a = él2o = -0.5 + j0.866
a2 = é 24o = -0.5 - j0.866
aa = éa6o = éo = 1.0 + jO.O
a4 = Ei4so = él2o = a
a& = ésoo = é240 = a2
1 +a 2
+a= O
a - a2 = v'3é9o = iv'3
a2 - a = V3E-;go = -IV3
1 - a = v3E-jaO = 1.5 - j0.866
1 - a 2 ·= v3é 30 = 1.5 j0.866 +
This system of vectors is called the positive-sequence system,
because as explained in Chap. I the order of the sequence of their
maxima occur abe. The system must always be treated as a
unit. In a three-phase balanced electrical system, Ea 1 would
represent the voltage of phase a; a 2Ea 1 the voltage of phase b;
and aEat the voltage of phase c. Fixing the phase position and
magnitude of Eat immediately determines Ebr and E ct· '
N egative-sequence System. Likewise, consider the balanced
system of vectors Ea2, Eb2, and Ecz in Fig. 4(b).
•
Ea2 = Ea2
Eb2 = Ei 120Ea2 = aEa2
E c2 -_ "';24oE
'-
_ a2E a2
a2 -
This system of vectors is called the negative-sequence system.

A physical picture of the significance of these systems may be
obtained by considering the fields which result when voltages
of the two systems are applied to a three-phase machine with
distributed windings, such as an induction motor. If the a, b,
... ~----¡<-o"
/ '
'' \ ,E"'<)
' ~Ebo
~-Eco
(e)
Fro. 4.- Symmetrical vector systems.
and e phases of the positive-sequence voltages be applied to

the terminals a, b, and e, respectively, a magnetic field will be
produced which will revolve in a certain direction. If now the
voltages to terminals b and e are changed by interchanging the
leads to terminals b ande, it is well known from induction-motor
theory that a magnetic field will be produced which rotates
in the opposite direction. A little consideration will show that
for this condition the re]ative phase positions of the voltages
applied to the motor are the same as for the negative-sequence
system. It- follows, therefore, that the negative-sequence set
of voltages produces a field rotating in an opposite direction to
that of the positive-sequence.
Phase sequence should not be confused with the rotation of
the vectors. In this volume the standard convention of counter-
clockwise rotation has been followed for all vectors; both positive-
r and negative-sequence vectors will rotate in the same direction,
J but the resultant fields in machines will have opposite rotations.
' Zero-sequence System. Finally, consider the system Eao,
Ebo, and Eco in Fig. 4(e) which represents three equal vectors.
•
Eao = Eao
Ebo = Eao
E co = Eao
This system is called the zero-sequence system.
Combination of Sequence Quantities to Form Phase Quantities.
In all three systems of the symmetrical components, the subscripts
denote the components in the different phases. The total voltage
of any phase is then equal to the sum of the corresponding com-
ponents of the different sequences in that phase. Therefore,
any three arbitrary vectors E a, Eb, and Ec may be equated.
Ea = Eao + Eal + Ea2 (21)
Eb = Ebo + Eb1 + Eb2 (22)
Ec = E co + Ecl + Ec2 (23)
Or substituting their equivalent values
E a = E ao + E + E a2
al (24)
Eb = Eao + a2Eal + aEa2 (25)
E c = Eao + aEal + a 2Ea2 (26)
The arbitrary system is thus de:fined in terms of three balanced
systems.
.. 7. Resolution of Three Vectors into Their Symmetrical Compo-

nents.
The individual components may be obtained as follows.
Zero-sequence. Adding the three equations,
Ea + Eb + Ec = 3Eao
+ (1 + a2 + a)Eal
+ (1 + a + a )Ea2 2
But since (1 + a + a)
2
= O,
Eao = i(Ea + Eb + E c) (27)
Positive-sequence. Multiplying equation (25) by a and
equation (26) by a 2 and adding these twd equations to (24), there
results
or
(28}
Negative-sequence. Multiplying equation (25) by a 2 and
equation (26) by a and adding these two equations to (24),
there resul ts
Ea+ a 2 Eb + aEc - (1 + a + a)Eao
2
+ (1 + a + a )Eal + 3Ea2
2
or
(29)
Ee !ello
o~..._~.,.
OrigÍt'IC'I) Zero -sequence

wc+o,..
~tive-seCljuence
Poaitiw-MclJuence
FIG. 5.-Grapbical construction for tbe determination oí sequence components.
Thus these three fundamental equations defining t he sequence

quantities in terms of the phase quantities are grouped together
in Fig. 5.
lllustration. Suppose the given voltages are balanced, for example,

E a= Ea
Eb = a 2Ea
Ec = aEa
---
then
Eao = (1 + a2 + a) Ea = O
3
Eat = ( + aa + aa) Ea = Ea
1
3
Ea2 =
1 + a3 + a 2Ea
4
= O
The zero- and negative-sequen ce components of voltage in a balanced

three-phase system disappear, leaving only the positive-sequence.
The graphical construction for the determination of the sequence com-
ponents will aid further in elucidating the operations. In Fig. 5 are shown
three vectors E a, E b, E c. The zero-sequence component E ao is obtained
by direct addition. The positive-sequence component Eal is obtained by
adding to Ea, Eb rotated counter-clockwise 120 deg., and E c rotated counter-
clockwise 240 deg., and taking one-third of the sum. Similarly Ea2 is deter-
mined by adding to E a, Eb rotated counter-clockwise 240 deg., and Ec
rotated counter-clockwise 120 deg., and taking one-third of the sum.
Numerical Example. The method may be further illustrated by means of
the analytical calculation of the components. Consider the three vectors
Ea = 60 +jO
Eb = 45- }75
• E. = -21 + j120
From equations (27), (28), and (29),
Eao = MEa + Eb + E .]
= ![(60 +jO) + (45 - j75) + ( -21 + j120)]
= 28 + jl5
Eat = ![Ea + aEb + a Ec] 2
= ![(60 +jO) + ( -0.5 + j0.866)( +45 - j75) +

( -0.5 - j0.866)( -21 + j12Q)]
= 72.2 + jll.5
Ea2 = ![Ea + a Eb + aE.]
2
= ![(60 +jO) + (-0.5 - j0.866)(45 - j75) +

( -0.5 + j0.866)( -21 + j120)]
= ( -40.2 - j26.5)
These three values give the symmetrical components of the three original
vectors. Using these same values of components the determination of the·
original vectors in terms of their components will be illustrated.
Ea = Eao + E al + Ea2
= (28 + j15) + (72.2 + j11.5) + ( -40.2 - j26.5)
= 60 +jO
Eb = Eao + a 2Eal + aEa2
= (28 + j15) + (-0.5 - j0.866)(72.2 + j11.5) +
( -0.5 + j0.866) (- 40.2 - j26.5)
= 45 - j75
•
Ec = Eao + aEal + a 2Ea2
= (28 + j15) + (-0.5 + j0.866)(72.2 + j11.5) +
( -0.5 - j0.866) ( -40.2 - j26.5)
= -21 + j120
8. Star-delta Transformations. *
Problems sometimes arise involving both star and delta
currents and voltages. Sorne conditions are such that the equa-
tions are set up more easily with sorne star and sorne delta
quantities. It is convenient for the solution of such problems
to be able to convert readily from star to delta and vice versa
in terms of the symmetrical components.
Current Relations. Let 1 A, 1 B, and 1e, be the delta currents
in the delta-connected windings of a machine and 1 a, h, and le,
the currents in the lines as shown in Fig. 6. At the junctions
the following relations are satisfied:
l a =lB-le
h =le- lA
l e = l..t - lB
The zero-sequence component of the star currents, l ao, is, by
equation (27),
lao = i(Ia + l b + l e)
= i[(IB+le +lA)- (lc+lA+lB)] = 0 (31)
This result shows that the zero-sequence current of a polyphase
circuit feeding into a delta connection is always zero, which
may be verified physically by the fact that for zero-sequence
currents to fiow in the line a neutral return circuit must be
,, available, a condition impossible of fulfillment in the delta
;. connection. The converse of determining 1 Ao in terms of 1 ao is
indeterminate. Currents of zero-sequence may circulate within
the delta without getting out into the line.
1 For the positive-sequence, by equation (28),
lal . i(la +
alb +a2l c)
= i[(IB +ale+ a 21A)- (le+ alA+ a 21B)]
= i[ (a 2JA + + +
1 B al e) - (al A a 21 B 1e) J +
By factoring a 2 out of the first term and a out of the second
term,
a2 a
lal = 3(1A + alB + a 2le) - 3CIA + alB + a 2lc)
* An application of these principies to transformers is given in Sec. 35.
20 SYMMETRlCAL COMPONENTS
By reference to equation (28), it may be seen that

I ai - (a 2 - a)IAI
- -jvf3JAI (32)
or
JAI (33)
•11
1
1
1
1
1
1
""1
1
1
1 1
1
Ec:1 lEa
1
1
1
1
1 1
1 1
1 1
1
la = lB - le EA = Ec- Eb
h = l e - lA. EB = Ea- Ec
le = fA. -lB Ee = Eb- Ea
1~ 1
Positive sequence
~ 1~ 2
Negative-sequence
EA1
Positive- sequence
EA2
Negotive-sequence
E.,.2
lA. o = indeterminate =0
j = jyi3Eal
lA. t = -v:/al = -Jv3E a2
J = indeterminate
lA 2 = - V3la2
= - -J-EAt
lao = O y3
lat = -jyi31At
la2 = +j'\1'31A2
Baz = + ~'3EA2
Fw. 6.- Star-delta current Fw. 7.-Star-delta voltage
transformations. transformations.
Similarly it may be shown that

J a2 = jV3JA2 (34)
and
(35)
These relations are illustrated in Fig. 6. They could have been
obtaincd more simply by the following reasoning. Assume that
positive-sequence currents alone flow in the delta. Positive-
sequence currents alone flow in the line, and the values of these
line currents can be obtained by applying the relations expressed
in equations (30), obtaining thereby the current triangle of Fig. 6
directly. The relations between the 1al and 1 Al can then be
,¡
obtained from this figure.
Voltage Relations. Now let EA, EB, and Ea of Fig. 7 be the
delta voltages and Ea, Eb, and E c, the star voltages. What are
·- the corresponding equivalences relating the star and delta
sequence voltages? By definition:
EA= E c - Eb
EB =Ea- Ec
Ea= Eb- Ea
Since the delta voltages must by their very nature form a

closed triangle, EA + EB + Ea must always equal zero and
EAo = O. The delta voltages, therefore, can never contain a
,. zero-sequence component. The star voltages, on the other hand,
may contain a zero-sequence component. It follows then that
Eao cannot be determined from EAo but must be determined by
sorne other relation in the problem.
The relation between the star and delta volt1ges of the positive-
and negative-sequence voltages are shown in Fig. 7. These
relations may be verified by applying equations (36). The
a-phase components of the different sequences are related by the
· expressions shown also in Fig. 7.
Nomenclature and Convention of Current Flow. A word of
warning may be injected at this point concerning the conventions
used in the nomenclature and in the direction of current flow.
There is no generally accepted convention relating the desig-
.~ nation of line conductors and delta phases. The authors have
therefore standardized upon that given because of the rather
simple relation to remember; conductors and windings of like
' designation are opposite, thus the a conductor is opposite to the
A winding, etc. The choice of positive sense of current flow is
.. entirely arbitrary. If the opposite directions had been chosen
in one set t he signs must be reversed in the equations relating
the. quantities of t he two sets. Care must be taken to insure
·.,
that the directions of current flow are consistent with the fore-
going assumptions.
9. Independence of Sequences in Symmetrical Systems. *
In symmetrical systems the di:fferent sequences do not react
upon each other; positive-sequence currents produce only
positive-sequence voltages, negative-sequence currents produce
only negative-sequence voltages, and zero-sequence currents
produce only zero-sequence voltages. While Fortescue gave a
rigorous proof of this statement as applied to both static devices
t1
-
le~~
rrW\N
1
1
1
Z"'b: <
1
1
1
ZCIIOI
\1
1
1
1
1
"'1
1
1
1
.. 'lb
~ :1 :''VVVV'
z"'"' 1
1 i
..
t
•l1 1 1 1
1
1
1
"'
1
1
1
1
Ze~~b\ ~.
'
: ' 1
1
1
l1
1
1
1
1
1
1_..
le : ¡ z"'"' e 1
1
1
1
i1 ''W\N 1 ;¡ 1
1
1
1
1
1
1
1
'+
1
1
1
'
1
1
1
11
1
1
1
1
1
1
1
1
1
1
1 1 1 1 1 1
1 1 1 1 1 1 1
' l$
111 '.0
LLI Uf '1'
1
> (Zt?ln
1 1
~
LLI
.o
1&1
u
1&1
1 1 1 : 1
1
1
'
1 1 1 1
1
1
1 1
-
1
1 1 1 1
1 11 'V\IW) <
1
,1 .t
-(J.,¡+ lb• le:) Znn
FIG. 8.-Symmetrical static network element.
and rotating machines, the present discussion merely undertakes

to show that these conclusions are reasonable.
Static N etworks. Consider :first the static network shown
in Fig. 8 which may representa perfectly transposed transmission
or distribution line in which capacity e:ffects may be neglected.
If only balanced positive-sequence currents be made to flow
through the line conductors, it follows that no current flows
through the neutral and no voltages are induced in the line con-
ductors or impedances due to the mutual coupling between the
line conductors and the neutral circuit. Since the mutual cou-
pling between the neutral circuit and the di:fferent line conductors
are all equal, the total voltage induced in the neutral circuit is
equal to zero. Otherwise, the induced voltages in each phase
are equal except for the phase displacement of 120 and 240 deg.
*A. mQre extended discussion is ~iven in Chap. XVIII, Sec. 167.
The positive-sequence currents thus produce only positive-

sequence voltage drops. It can similarly be shown that negative-
sequence currents produce only negative-sequence voltage drops.
If only zero-sequence currents fiow, equal currents fiow in each
line conductor and the combined currents of the three line
conductors return through the neutral. In this case equal
voltages will be induced in all three conductors, and including
the drop in the neutral impedance the drops will be the same in
all three phases. These equal voltages constitute a zero-sequence
from which it may be concluded that zero-sequence currents
produce only drops of zero-sequence.
This discussion has considered only series impedances. It
should be apparent that similar conclusions may be drawn
for the characteristics of symmetrical shunt impedances. Since
the steady-state characteristics of lines with distributed con-
stants can be represented by the well-known equivalent 1r or T,
the conclusion may be further extended that the three sequences
may be considered independently for static networks with
distributed constants.
Rotating Machines. As remarked previously, positive-se-
quence currents in the stator of a symmetrical machine produce
a rotating field which rotates in the same direction as the rotor.
This field naturally produces only voltages of the same sequence
across the terminals. The direct-currents in the rotor of syn-
chronous machines and the slip frequency currents in the rotor
of induction machines likewise produce only synchronously
rotating fields in the same direction as the rotation of the rotor
so that only positive-sequence voltages are produced on the
termina.ls of the stator due to these currents. I t is apparent
then that under normal operating conditions with positive-
sequence voltages applied to the stator of rotating machines
only positive-sequence currents are produced.
If negative-sequence voltages only are applied to the stator
of synchronous or induction machines, a synchronously rotating
field is produced which rotates in a direction opposite to the
rotation of the rotor. This field induces currents in the rotor,
which in turn produces a synchronously rotating field in a
direction opposite to that of the rotor. Thus all the currents
and voltages in the stator would be of negative-sequence.
Because of the 120-deg. space displacement of the windings,
the zero-sequence currents, which are in phase with each other in
the three phases, produce no flux in the air gap. Hence, these
currents can produce only voltage drops of the zero-sequence.
This analysis indicates that in symmetrically wound machines
currents of the different sequences do not react upon each other.
10. Sequence lmpedances.

It has been shown that in symmetrical networks the com-
ponents of current of the different sequences do not react upon
each other. When voltage of a given sequence is applied to a
piece of apparatus, a very definite current of the same sequence
flows. The apparatus may be characterized as having a definite
impedance to this sequence. Special names have been given
to these impedances, namely: the impedance to positive-sequence
currenM, the impedance to negative-sequence currents, and the
impedance to zero-sequence currents but have been contracted
to the simple expressions, positive-sequence impedance, negative-
sequence impedance, and zero-sequence impedance.
The impedances of symmetrical static networks are the same
for the positive- and the negative-sequences but may be different
for the zero-sequence. For rotating machines the impedances
will in general be different for all three sequences. Methods
for determining these impedances by calculation and test will
be given in subsequent chapters.
Problems
. l. With a defined as ( -0.5 + j0.866) or ei 120, evaluate the following:
1 -a a - a 2
(1 - a)(1 - a2), (a - a.2)(a2 - 1),
1
_ a 2 ' a 2 _ , (1
1
+ a)2, (1 +a) (1 + a2),
1 +a , 1__
__ - a, 1 + az .
1 + a2 1 +a 1 -a
2. Determine the six roots of {11.
3. Given: l a = 8 - j6; h = -13 - j10; l e = 2 + j10. Find: /1, lz,
and 1 o, and ch eck.
4. Given: Ea = 150.4; Eb = 220é 250 ; Ec = 220E-i 110• Find : E1, E2, and
Eo, and check.
5. Given: l a = O + j20; h = 20 +jO; l e = O. Find: l1, l2, and lo,
and check.
6. Given: l a = O + jlOO; h = - 86.6 - j50; l e = +86.6 - j50. Find:
11, 12, and 1 o.
7. The a phase of a three-phase circuit is open-circuited, and currents
h = 1 and l e = - 1 fiow in t he two other phases. Determine /1, l z, and lo.
8. The b phase of a three-phase circuit is open-circuited, and currents
le = 1 and l a = - 1 fiow in the two other phases. Determine /1, lz, and
I 0, and compare with Prob. 7.
FUNDAMENTAL PRlNClPLES 25
9. Assume that la = 100 amp., and l b = le = - 50 amp. What are
the sequence currents?
10. The line-to-line voltagcs of a three-phase system measure 100, 150,
and 200 volts. Find the magnitudes of the positive- and negativc-sequence
.. components of the delta voltagcs and the star voltages .
11. The loss in three resistanccs of R in cach phase is equal to R multiplied
by the sum of thc squarcs of the currents in each of thc phases. Show that
thetotalpower lossPisP = R (la 2 +ib2 +ic2 ) = 3R(i 1 2 +1 2 2 +1 0 2 ).
•, 12. A groundcd-neutral systcm has a positivc-sequence voltage of Eali
,,
} show that if thc ground be removed and one phase wire be grounded, the
t- positive-scquence voltagc remains unchanged.
13. A 13,200-volt delta-connected distribution system is groundcd on one
phase. What is the magnitude of the zcro-sequcnce voltage?
14. Three resistors of 5, 10 and 20 ohms are connected in delta across
phases A, B and C, rcspectively, of a balanced 100-volt system. What are
the sequence components of currents in the resistors and in the supply lines?
..
1
CHAPTER III
CALCULATION OF UNBALANCED FAULTS
Probably the most important application of symmetrical

components is the calculation of unbalanced faults on commercial
systems. These systems are usually symmetrical except for
the fault itself, so that the sequences do not react upon each
other except as related to each other by the particular char-
acteristics of the fault. The calculation of the currents and
voltages under unbalanced conditions requires a knowledge
of the generated voltages and the impedances of the network
elements to the different sequences. For the present, it will
be assumed that these constants are known. Their deter-
mination will be taken up in detail in subsequent chapters.
· The present chapter extends the single-phase method of cal-
culating polyphase systems to t he calculation of unbalanced
faults utilizing the principies of sym-
metrical components.
11. Kirchhoff's First Law.
The a na 1y si s so far considered
only one branch or network element.
In this element t he voltage drops of
the various sequences are dependent
only upon the currents of the indi-
vidual sequences and the correspond-
Fw. 9.-Phase a currents at a
junction point. ing impedances. What will be the
result when a group or network is
considered?
At each junction point of a network, for example, the junction
in Fig. 9 in which three circuits converge, by Kirchhoff's first
law the sum of the currents in all of the conductors equals zero:
la' +la" +la 111
= O
lb' + lb" + h"' = o
le' + le'' + le'" = o
These equations may be expanded by substituting the sequence
26
CALCULATION OF UNBALANCED FAULTS 27
components for the phase values in accordance with equations
(24), (25), and (26), with the result that
(lao 1 + lao + lao
11 111
) + (la/+ la1 11
+la/") +
(la2 1 + la2 11 + la2 111 ) = O (38)
(l ao
1
+ 1ao + 1ao
11 111
) + a 2(1 al + 1a1 + 1at ) +
1 11 111
a(la2 1 + 1a2 11 + 1 a2 111 ) = O (39)

(lao' + lao" + lao'") + a(lal '+ la1 11 +la/")+
a 2(la2 1 + 1a211 + 1 a2"') . o (40)
The addition of equations (38), (39), and (40), remembering
that 1 + a 2 + a = O, gives
(41)
Upon multiplying equation (39) by a and equation (40) by a 2
and adding them to equation (38) one obtains the following
result
(42)
Similarly, multiplying (39) by a 2, (40) by a and adding to (38),
we find
(43)
The equations (41), (42), and (43) show that the various
sequence components at any junction must individually add up
to zero; each sequence of currents obey Kirchhoff's :first law
separately. This proof has been given for three circuits, but
it is apparent that it is a· perfectly general relation which must
always hold at any junction and may be extended to any number
of conductors.
12. Kirchhoff's Second Law.

Consider a general symmetrical network. Since the voltage
drops of each sequence, as has been shown in C~ap. II, may be
considered separately, the sum of the voltage drops of any
sequence around any closed circuit must be equal to the e.m.fs.
of that same sequence. The network phenomena therefore
obey Kirchhoff's second law. This holds true for all points
in the system including the point of fault if this point be con-
sidered as a source of zero-, negative-, and positive-sequence
voltage. The system may therefore be considered as three
separate and distinct networks: one for positive-sequence, one
for negative-sequence, and the third for zero-sequence. The
tie between these systems will be the terminal conditions at the

point of fault.
13. Sequence N etworks.
It has been shown that each of the individual sequences
may<be considered independently and since each of the sequence
networks involves symmetrical currents and voltages, and
impedances in the three phases, each of the sequence networks
may be solved by the single-phase method that is commonly
used for balanced three-phase systems. The conception of
the sequence networks thus makes it possible to solve unbalanced
three-phasec problems by methods similar to the single-phase
solution of balanced problems. These sequence networks
may be represented by single-line diagrams of the conventional
type in which t he currents represent line currents, the voltages
line-to-neutral voltages, and the impedances the corresponding
star impedances. Each network will thus have its own line
conductors and neutral point between which the line-to-neutral
voltages are measured and shunt loads are connected. Lines
with distributed constants may be represented by series imped-
ances in the lines and shunt impedances between the lines and
the neutral of the individual networks.
The positive-sequence network is in all respects identical with
the usual networks considered; the resistan ces and re.a ctances are
the values usually given to calculate line regulation. Each
synchronous machine must be considered as a source of e.m.f.
which may vary in magnitude and phase position, depending
upon the distribution of power and reactive volt-amperes just
previous to the application of the fault. The positive-sequence
voltage at the point of fault will drop, the amount being condi-
tioned u pon the type of fault; for three-phase faults, it will be
zero; for double line-to-ground faults, line-to-line faults, and
single line-to-ground faults, it will be higher in the order stated.
The exact value is calculable and will be determined.
The negative-sequence network is in general quite similar to
the positive-sequence network, in fact the number of branches is
the same. Because positive-sequence voltages only are generated
in the synchronous machines, this network will contain no
sources of e.m.f. except the :fictitious one at the point of fault.
The zero-sequence network will likewise be free of interna!
voltages, the flow of current resulting from the voltage at the
point of fault. The impedances to this sequence are radically

different from those of either the positive- or negative-sequences.
The line im pedances are those obtained by imagining the three
conductors connected together, the ground forming the return
conductor. In the zero-sequence network, part of the impedance
may be in the outgoing conductor and part in the earth return.
However, the total impedance is the only factor affecting current
distribution and line-to-ground voltages at the terminals of the
network. For the calculation of these quantities the total zero-
r sequence impedance may be inserted in one conductor of a single-
¡.
line diagram which becomes similar to the corresponding
diagrams for the positive- and negative-sequence networks, the
ground constituting a bus of zero potential. The transformer
and generator impedances will depend upon the type of connec-
tion, whether delta- or star-connected; if star, whether grounded
or not. The calculation of the impedances to negative- and
zero-sequences will be considered in a subsequent chapter.
The number of branches which require consideration in the zero-
,.. sequence network is usually less than for the positive- and
negative-sequence networks, because of the presence of delta-
connected transformers. For the time being, let it be assumed
that these quantities are known.
14. Sequence N etworks of a Particular System Including Shunt
Loads.
Asan illustration of the sequence networks, consider the system
shown in Fig. IO(a). Two generators A and B at the sending
end of a line are connected to two delta-star transformers (only
one of which is grounded) which are bussed on the high side. At
the receiving end there are two star-star transformer banks
(both grounded on the high side, but only one on the low side)
connected to separate generators C and D, one of which is
grounded. *
The load is assumed concentrated at the receiving end and
1' is connected as an equivalent star with the neutral point
· ungrounded. It is represented in the positive-sequence network
of (b) and the negative-sequence network of (e) by shunt imped-
ances' cónnected to the respective neutral buses, but since the
* The unusual connections shown in this diagram are used merely to illus-
trate the method as applied to different transformer connections. The
mutual impedances between lines in the zero-sequence n etwork are neglected.
ao SYMMBTIUCAL COMPONENTS
neutral is ungrounded, the load impedance does not appear in the

zero-sequence network of (d). If the neutral point of the load at
,D were connected to ground, it would be necessary to add a shunt
branch connected to
the neutral bus in the
zero-seq uence net-
work.
Assume the fault to
occur at the point F
midway on one of the
lines. The single-
line im pedance dia-
gram for the positive-
sequence is shown in
Fig. lO(b) and those
Nevf_(gf. bus_qff.!f!UEi'Xf.!J..s_g_q:l.f..I?.!J.f.lU1ff:t!2!:./s. _--.,
for the negative- and
r -
: ~ ~· zero-sequences in Fig.
lO(c) and (d). The
synthesis of these re-
spective networks will
be di se us sed later.
/(!.!!!!E!. É!:!~ 2L'L~!tz.~f!.Cf!l!'!ff.!!!'t...Vfil.!~, It is importánt to re-
•___,.. - 1 tain the positive fiow
''1 of ~e u r r e n t, as indi-
cated by the arrows,
the same in any
branch for all three
F1o. 10.-Diagrammatic representation of the three networks, as the fun-
. sequence networks for a three-phase system.
damental equations
postulate this condition. E', E", E"', and E" " are the in-
terna! voltages of the four generators A, B, C, and D,
respectively. Since positive-sequence voltages only are gener-
ated within the machines, the corresponding internal voltages for
the negative- and zero-sequence networks are zero. The points
associated with the internal voltages may therefore be connected
to the neutral buses of the corresponding sequences.
15. Connection of N etworks to Represent Faults.

The only tie between these otherwise independent networks
is the terminal conditions at the point of fault. The nature of
this tie will vary with the character of the fault. Four cases
will be considered at this time:
l. Single line-to-ground fault.
2. Double line-to-ground fault.
3. Line-to-line fault.
4. Three-phase fault.
Imagine three short conductors of zero impedance connected
to the three line conductors at the point of fault. The terminal
conditions imposed by the different ~----r------
types of faults will be applied to
these imaginary leads, the potentials b--~--~---
to ground of which will be Ex, E,11
and Ez, respectively, and the currents e --I-x-,~--t----r--
1 x, I '" and I z· These imaginary con-
ductors are shown i~ Fig. 11.
't Iy Iz t t
Ey
16. Single Line-to-ground Fault. FIG. 11.-Imaginary leads
brought out at point of fault.
For this case suppose the a phase
to be grounded as indicated in Fig. 12(a). The terminal condi-
tions in this case are:
From equations (27), (28), and (29) of Sec. 6, the symmetrical

components of current are:
lxo = !(lx + 111 + lz) = i(lz) (45)

lz1 = !Cix + alv + a2Jz) = !(Ix) (46)
I z2 = i(I :¡; + a Iv + al z) = !(1z)
2
(47)
or
1 zO = 1 x2 = 1 zl (48)
Also
(49)
N ow from equation (48) it can be seen that the currents at

F 1, Fz, and Fo (Fig. 10) in the three networks for the three
sequences are equal, and if the point F 1 of the positive-sequence
network be connected_to the neutral bus of the negative-sequence
network, and the point Fz in the negative-sequence network be
connected to the neutral bus of the zero-sequence networkt ae
.
shown in Fig. 12, the three networks will be connected in series
and the same total current must fiow through the three networks.
Since the current distribution within each of the networks is
determined sol el y
:==_Ix=:=-+---ty~-~~
lzl\ ==_
by the im pedances
in t he individual
networks, this com-
E;t" Ey Ez bination of the net-
(01) works enables one
" to d e t. e r m i n:e the
e u r re n t distribu-
tion in any branch.
However, for this
connection, to de-·
termine the current
completely, the re-
Nef!_f_!q_l.P.!!~ E[!!'!.{13!!'!.e_ Sf!.CfY~'!_c_e_ f!.f!f~_!Jf:k_ __ -. lation expressed in
r- 1
! - - --1 equation (49) must

also be satisfied.
1;2 tl;:- Now starting
F"z from any one of the
Neutral bus of zero-sec¡vencenefwork
interna! voltages_of
r------------)---------------,:
¡-)o- Fig. 12, say E', and
1
1 tracing one's way
1
through the net-
work by any path
to F 1, it can be seen
(b)
Fw. 12.- Interconnection of the three sequence net-
that the positive-
works for a line-to-ground fault on one conductor. sequence vol tage
Exl at Fds equal to
E:&! = E' - 1; positive-sequence voltage drops from E' to the
point F1 (50)
Similarly with the negative-sequence network, starting at any
point of the neutral bus and tracing one's way to F2 it will be
found that
Ex2 = O- 1; negative-sequence voltage drops to the pointF2 (51)
and for the zero-sequence network
Exo O - 1; zero-sequence voltage drops to the point F 0 (52)
=
Adding these three voltages and equating to zero as indicated
by equation (49):
O = E' - ~ voltage drops from E' to F 1 in positive-
sequence network
- ~ voltage drops from neutral bus to F2 in nega-
(53)
tive-sequence network
- ~ voltage drops from neutral bus to Fo in zero-
sequence network
or
E' = ~ voltage drops from E' to F 1 in positive-sequence
network
+ :2; voltage drops from neutral bus to F2 in negative-
(54)
sequence network
+ :2; voltage drops from neutral bus to Fo in zero-
sequence network
1. If the point F 0 be connected to the neutral bus of the positive-
sequence network, it can be seen that equation (54), and con-
sequently also equation (49), will be satisfied, and the connection
of Fig. 12 will completely determine the flow of current. The
positive-, negative-, and zero-sequence currents in any branch
are equal to the actual currents flowing in the respective branches
in the three networks.
The determination of the sequence and phase voltages and
currents at any point in the system will be discussed in connection
with the application of the calculating board. This will insure
a better pict.u re of the correlation of the different networks and
provide a better conception of the physical significance of the
various quantities.
Since the only e.m.fs. generated in tb.e synchronous machines*
are those of the positivc-sequence, the question is frequently
raised asto the origin of the negative- and zero-sequence voltages
that appear throughout the network. Any unbalanced condition
such as a line-to-gr<~und fault requires that positive-sequence
currents and currents of other sequences flow. The negative-
and zero-sequence currents then distribute themselves throughout
the system within their respective networks. In each network,
as for example the negative-sequence, the impedance drops
produced by these currents give rise to voltages of the correspond-
* For special applications such as ccrtain types of phase-balancers, a
negative-sequence e.m.f. may be generatcd. Such cases may be solved by
íncluding the negative-sequence generated e.m.f. in the negative-sequence
.network as i.s done for the positive-sequence network.
!
a
~
~
•
0--------~----------------
b-------+--~-----------
e--------~---+----~-------
~
t)l 1AI lz!
Ex Ey .l._ Ez
(01) ~
~
......
~
r---~~!.cq!..P~.!..!J.!..P.!J.§i..fJo.:...t'::_~eg~~!!~e_!!~t.~~c_k___ .. r J!.f!_lJj~~!_I¿_IJ_§_Q(.ll~t.!..·'(!::§!!f!!~I'IE!_n.ft_f'!!J~~ r!!f.l!!_!q!_q_V.§_o!J?!'!!'~/!!!l'!...e!E'_~~_r;
1 1 1 1
8
:-oE' - --- - - E"'....t ~~ ·~
~.
_.,.
'VVv 1 1' ~
~------ El------ ~ 1 1
- ~
1
l., E" ----~
- _____ .J:
r~- ---->
1;2 hx7" I;o I¡Ix:-
F¡ F2 Fo ~
(b)
Fw. 13.-lnterconnection of the three sequence networks for a double line-to-ground fault.
ing sequence. These voltages are in general a maximum at the
fault point and decrease as the neutral bus i~ approached.
17. Double Liue-to-ground Fault.
For this case assume both phases b and e faulted to ground
simultaneously as shown in Fig. 13(a). The terminal conditions
for this case are then
E 11 =O
Ez =O
lx =O
Exo = i(Ex +E¡¡+ Ez) = iEx
Exl = i(Ex +
aE11 +
a 2Ez) = tEz }(56)
Ez2 = !(Ex +
a 2E 11 +
aEz) = iEz
or
Exo = Ezl = Ez2 (57)
Also
l x = l xO + l xl + l x2 = O (58)
·'
Equations (57) and (58) define the conditions which must be
fulfilled at the terminals for this case.
Connecting the three sequence ne'tworks of Fig. 10, as shown
in Fig. 13(b), it will be seen that the two above conditions are
.1 fulfilled; equation (58) is fulfilled at the junction of the three
networks, and equation (57) is fulfilled because the three voltages
are measured across the same terminals and must therefore be
equal to each othe~.
18. Line-to-line Fault.
Assume the fault to occur between phases b and c. The
terminal conditions for Fig. 14(a) are then
lz =O
'
The sequence components are then
Izo= !(Ix + 1 11 + lz) = !(Ó- lz + lz) =O (60)
Iz1 = !(Ix + aly + a 2lz) = t(O- alz + a 2l,)
(61)
=!(O- a 2l11 +al,) -
(62)
Since I xo = O, it follows that Exo must also be equal to zero.

Furthermore,
and
Then, substituting in these equations the relations, E 11 = E-z and

Exo = 0:
Exl = MEx + (a+ a 2)Ez] (64)
Ex2 = MEx + (a + a 2)Ez] (65)
= Exl (66)
:=_-==-~x~:=r~-y=:)GJ=-~:z:=_,_l=--==-
X Ey E:z:
(o\
/'!..e_l!...IT::!.I_!:..u_s_ q!_p_o~!!!!" ::_~e_CfY!!!:!~e_ :!..f!!:YECk__ 1

' , ~ ~ ....,_~E"w
kE · ,
~----- E; ------- Posi five-seqvencenefwork +
- ---~:,J
J
+--
J ,
LoE _____ ..J i
(b)
FIG. 14.-Interconnection of the three sequence ne tworks for a line-to-line fault.
Because both I xo and E xo are zero, the zero-sequence network

can be eliminated from consideration. The only conditions
which must be fulfilled by the equivalent network are those
expressed by equations (62) and (66). By connecting the
positive- and negative-sequence networks, as shown in Fig. 14,
it may be seen immediately that equation (62) is satisfied because
I xr must always cqual - f x2· Equation (66) is satisfied because
the equal voltages expressed thereby are measured between
the same points and hence must be equal.
19. Three-phase Fault.
This case is the simplest and the one most familiar to persons
calculating short-circuits for the application of circuit-breakers
and relays. The equivalent diagram reduces to that shown in
Fig. 15. Since the system remains balanced, the negative- and
zero-sequence networks do not enter the problem.
CALCULATION OF UNBALANCED FA ULTS 37
20. Application to Calculating Boards.
The calculation of system faults by the methods just described
may be grea tly
facilitated by t he Q
b
use of calculating
boards. * Figure
lx~ Iyt l.z.~
16 shows a typical
alternatin g- e u r- Ex Ey Ez
(Q)
rent calculating
~----------------~
board use d f o r
power- system
studie s . This
board is made up
of a large number
of adjustable
resistance, react-
ance, and capaci- Fw. 15.- Interconnection of the three sequence networks
for a three-phase fault.
tance branches to
!1<1'
.·.;.:
'
"
i
~
Fw. 16.- Alternating-current Calculating Board: 179 circuit board of the
Commonwealth Edison Company as built by the Westinghouse Electric and ·
Manufacturing Company.
r epresent the corresponding elements of the actual power system.
Special branches are designed to represent inductive coupling
* TRAVERS, H. A., and W. W. PARKER, An Alternating-current Calculat.
ing Board, Elec. Jour., pp. 266- 270, May, 1930.
between branches. Several sources of e.m.f. are provided with

means for adjusting their magnitudes and phase relations.
Metering equipment is provided which is suitable for measuring
the magflitude and .relative phase position of voltage and current
vectors as well as power and reactive volt-ampere quantities.
The switching arrangements make possible the convenient connec-
tion of network branches to represent a network or the combina-
tion of the three sequence networks.
By appropriate connection, the alternating-current calculating
board may be used to represent the different types of faults.
Regardless of the manner in which the sequence networks
are connected, the current in phase a at any point can always
be obtained by adding the currents of the three networks at the
corresponding points. Suppose it is desired to determine the
current in the left-hand section of the faulty line in Fig. 10.
The current in phase a is determined by adding directly the
current 1 ml from the positive-sequence network, 1m2 from the
negative-sequence network, and 1 mo from the zero-sequence
network. This can be accompHshed most conveniently by
plugging into the three networks by means of current trans-
formers and totalizing on the secondaries. For phases b and e,
networks can be devised which shift the positive- and negative-
sequence components 120 and 240 deg.; the shifted currents
can then be added to the zero-sequence current. In the absence
of such a device, the phase currents can be obtained by adding
the quantities analytically after ·operating, by the unit vectors
a and a 2 , using the relations
h = Imo + a 2Iml + almz
l e = Imo + afm1 + a 2 Imz
The sequence voltage for any point in a network appears
only in the corresponding sequence network, so that the manner
in which the sequence voltages are measured is independent
of the particular way in which the networks are connected
together for the representation of the different types of faults.
In the positive-sequence network the potential at any point is
measured between that point and' the neutral bus, indicated
by the dotted line in the positive-sequence network of Fig. 10.
Expressed differently, it is equal to one of the generated e.m.fs.
minus the drop from the particular e.m.f. to the point con·
sidered. If the value of the particular e.m.f. be zero, the poten·
1
••
tial at the point will be the negative of the drop from the neutral
bus to the point. In the negative-sequence network, since
normally no generated e.m.fs. of negative-sequence exist, the
negative-sequence potential at any point will be merely the
negative of the drop to the point from the negative-sequence
neutral bus. Similar considerations apply to the zero-sequence
network, the potential at any point being the negative of the
drop to the point from the neutral bus of the zero-sequence
network. For example, the voltage of phase a on the left-hand
bus is equal to the vector sum of the voltages E/, E2', al\d E 0',
which are indicated in Fig. 10. Note that this voltage is equal
to the generated positive-sequence voltage E" minus the drops
in the generator and transformer due to the positive-, negative-,
and zero-sequence currents. Thus it may be observed that
at the interna! voltage of the machine all the drops are zero and
the voltage of phase a is the generated positive-sequence voltage.
Also, for the case of a single line-to-ground fault, as the point
under consideration approaches the fault point, the voltage of
phase a approaches zero. On the alternating-current calculating
board the phase a voltage can be obtained by direct addition by
means of potential transformers, and the voltages of phases b and
e by phase-shifting devices or analytically by means of the funda-
mental relations connecting phase and sequence voltages.
For short-circuit studies in which only the magnitudes of cur-
rents are required it is usually permissible to assume that the
generated e.m.fs. are of the same magnitude and in phase so that a
single source may be used to represent the generated e.m.f. A
further simplification results when it is found suffi.ciently accurate
to assume all the impedances of the same phase angle. In this case
the networks may be set up on a direct-current board, the elements
of which are pure resistances.
When it is desired to obtain only the zero-sequence currents,
or the residual* or ground currents, the positive- and negative-
sequence networks are required only for determining the magni-
* The sum of the zero-sequence currents in the three lines gives 3lo; this
quantity is sometimes called residual current. Part of the total zero-
sequence current may :flow in a neutral- or ground-wire or cable sheath
mounted on the same structure with the line conductora, and part in the
ground or remote conductora. The latter is commonly referred to as ground
current and occasionally as residual current. Zero-sequence is invariably
used in fault-current calculations. Residual current in the past has bee:Q
used principally in inductive coordination work,
tu de of the fault current; the distribution of ground currents

01 is determined of course only by
b
.. the zero-sequence network.
Faul t im pedances m ay be
e introduced in the connections
~~¡ of the sequence networks fur
F different types of faults and
Z <
..':• e-;
;'>' therefore, as will be shown in
G - - - - ___ _), ___ J,___ _ the next section, do no( affect
the method just described for
(DI)
measuring sequence and phase
voltages and currents.
21. Representation of Fault
Posifive - sequence network
Impedances.
In many cases the f ault im ped-
ance forms a considerable pro-
z portian of the total impedance
so that the resultant short-
circuit current is much smaller
than that which would occur if
the fault impedance were zero.
Negafive- sequence nefwork
Inaddition, the fault impedance
may affect thc distribution of
the current between the nega-
tive- and zero-sequence net-
works, such as occurs for a
double line-to-ground fault.
The following discussion shows
how the equivalent diagram of
Zero- seqvence nefwork.
the system should be altered to
include the fault impedance.
Single Line-to-ground Fault.
z The boundary conditions for
this type of fault are indicated
(b) il} Fig. 17(a), in which Z is the
Fw. 17.--Method of connecting fault impedance ofthefault including
impedance into network for linc-to- the are, tower footing imped-
ground f aul t.
ance, etc.
Equations (44) then become:
Ex= Zfx (67)
·'.
in which Ex is the voltage of the line conductor x.

f
1 The current conditions have not been altered, so that the
1
relations between the different sequence currents obtained for
Z = O still apply. Equation (49), however, showing the relation
between the sequence voltages, becomes
Ex = E xa + E xl + E x2 = Z1x
and since
1 x = 31 xl
Ex=3Z1xl (68)
Equation (53) bccomes
¡ 3ZI .el = E' - 1; voltage drops from E' to F1 in positive-
rt sequence network
~ -1; voltn.ge drops from neutral bus to F2 in ( )
• 69
~ nega ti ve-sequcnce net work
¡•
l~·
-1; voltn.ge drops from neutral bus to Fa in
~· zero-sequence network
¡ ~··¡;.
E' =
+
~ voltage drops from E' to F 1 in positive-sequence
network
L voltage drops from neutral bus to F2 in negative-
sequence network (70)
+ ~ voltn.ge drops from neutral bus to Fa in zero-
sequence network + 3Z1xi
The three networks must therefore be connected to satisfy
equations (48) and (70). This will be accomplished by con-
necting the networks as shown in Fig. 17 (b).
The three impedances Z can be lumped into one impedance1
but by so doir~g the identity of the points F a!ld G which represent
the conductors n.nd the ground is lost. Since no current flows in
conductors y and z, the fiction of imaginary impedances Z in
these conductors may be carried out to extend the limits to
which the system may be considered symmetrical. When this
is done each sequence network may be extended by the value
of the impedance Z which will be common to each of the networks.
The voltages ofF and G in each network may then be obtained
by merely determining the voltage to the neutral bus in the
respective networks.
Double Line-to-ground Fault. The boundary conditions

and position of faült impedance for this case are indicated in
Fig. 18(a). Equations (55) expressing these boundary conditions
become
E 11 = Z(l11 lz) +
Ez=Z(l¡¡+lz)=Eu
fx = Ü
Applying these relations, equations (56) become

E xo = ![Ex+ 2E11]
+
Ext = ![Ex+ (a 2 a)E11] = }(Ex - E 11)
Ex2 = ![Ex+ (a+ a 2)E11] = }(Ex - E 11)
a ---,------
b --+-..,.-----
G --------
(DI)
( o () ,•
_______ l ___ ----
fbsifive-sequence nefwork
------- ------
Negative -ser¡vence nefwork
------- ------
Zero·set¡~Jence nefwork
F, F2 Fo !>
~3Z
G2. Go~
G,
(b)
Fxo. 18.-Method of connecting fault impedance into network for double line-to-
ground fault.
These equations supply two voltage conditions which must

be satisfied, namely,
Ex2 = Exl (73)
Exo - Ext = E¡¡ = Z(I" + 1 z) (74)
The voltage E 11 from equation (74) may be evaluated in terms
of sequence components of currents as follows:
!y= lxo + a 2lxl + al,_2
lz = lxo + alxl + a2fx2
Adding
~-
Now from the last of equations (71)

1 :¡; = 1 :&o + 1 + 1 :&2
:&1 = O (76)
or
1 :&1 + 1 :&2 = -1 :&o
Substituting this relation in (75)
111 + 1z = 31 :&o (77)
and (74) becomes
E:&o- E:&l = 3ZI :&o (78)
Now, since the three networks are sufficient in themselves,
but their manner of connection is dependent upon the boundary
condition, the necessary and sufficient conditions to be met by
the equivalent networks are those expressed by equations
(73W76), and (78). These conditions are satisfied when the
networks are connected as shown in Fig. 18(b).
4'f - - - . - - - - - -
()
b--t--~---
------- ---------
Posifive-seqvenr;;e ntlfwor!t:
------- _j_ ---- -~-
Negufive -seqvence nefwol'k
Fl
~ z F1
.1
: z
G,
~ G2 2
(b)
Fw. 19.-Method of connecting fault impedanco into n etwork for a line-to-line ·
fault.
Line-to-line Fault. The terminal conditions for this particular

fault are indicated in Fig. 19(a). Expressed mathematically
they are
¡:¡; = o
111 = -Jz
Ell = E z - Z1z
in which E 11 and E z are the potentials of the conductors y and z.
Since the current relations are unaltered, equations (60),
(61), and (62) still apply. From equation (61)
3
1z = - a + a l:&l
so that
(80)
From t he fundamental relations expressed in equations (63)

which connect the phase and the sequence quantities
4'
Exl - Ex2 = i(a- a 2)E i(a 2 - a)Ez11 +
Substituting E 11 from equation (80)
1 (a - a 2 )Z
Exl - Ex2 =
3
2
- (a - a )E z - -a "I xl
a~
+ + -31(a 2
- a)Ez
= ZI xl (81)
From equations (62) and (81) the two necessary and sufficient
conditions which must be met are
J x2 = -J xl
}(82)
Exl = Ex2 +ZI xl
These are satisfied by the connection shown in Fig. 19(b).
22. Short-circuit of Generators.
To illustrate the application of the more general relations
which have been developed, a simple analytical example will
be considered. For this purpose let it be desired to determine
t he voltages and currents during the short-circuit at the terminals
of a symmetrical star-connected generator under different types
of unbalanced faults. The impedances to positive-, negative-,
and zero-sequence currents will be Zt, Z2, and Zo, respectively.
This particular example was chosen not only because it is an
interesting problem in itself, but also because it furnishes the
simplest or elemental forrn of network which still retains unequal
and finite impedances to the three sequences.
b
e
Fw. 20. -Line-to-ground fa ult on a generator.
Single Line-to-ground Fault. The equivalent diagram for

this condition is illustrated in Fig. 20. As can be seen, the
currents of the three sequences are all equal to
E
11 = l z =lo = z 1 + z 2 + Zo (83)
CjiLCULATION OF UNBALANCED FAULTS 45
The current in phase a (i.e., the short-circuit current) is

la = (lo+ l1 + lz)
3E
(84)
-: Z1 + Zz + Zo
1b = lo + a 2l 1 + al 2
= (1 + a 2 + a)lat = O
le = lo + al 1 + a 2 I 2
= (1 + a + a 2 )lal = O
The positive-sequence component of the terminal voltage
is the voltage between F1 and F o (Figs. 12 and 20), namely,
E1 =E- Z1I1
After substituting the value of l 1 from (83),
E1 = [1- Z1 + zl
Z2 + Za E
J
or
(85)
Similarly the negative- and zero-sequence voltages are

E2 = -Z2l2
_ z2 E (86)
Z1 + Z 2 + Zo
Eo = -Zolo
_ Za E (87)
Z1 + Z 2 + Zo
The voltage of phase a is then
E a= E o + E1 + E2
= E Z2 + Zo - Z2 - Zo = O
Z1 + Z2 + Zo
and of phases b and e
Eb = Eo + a E1 + aE22
E
a E- z + z + z (Zo + a Z1 + aZ2)
2 2
=
1
- d~ f.~
2
E
= z +z +z [(a 2 ' - a)Z2 + (a 2 - , l )Zo] (88)
Ec
1 2
= Eo + aE 1 + a 2E 2
0
-
1
.,... -/O, i : 1.
= aE .- Z¡ + ~2 + z/Zo + aZ + a Z 1
2
2)
'
•'
,.
Double Line-to-ground Fault. The equivalent diagraÍn for

this case reduces to that shown in Fig. 21. The solution of this
network gives
(90)
(91)
(92)
"'
E
b
I,! z2. lo! Zo
e
F, F2. Fo
-
Fxo. 21.-Double line-to-ground fault on a generator.
from which
la =lo+ l1 + l2
E
ú - ZoZ1 + Z1Z2 + Z2Z} -Z 2 + Zo + Z 2 - Zo] = O
1b = 1o + a 2l1 + al 2
'·
E
= ZoZ1 + Z1Z2 + Z2Zo[- Z + a2(Zo + Z aZo]
2 2) -
- ZoZ1 + z~2 + Z2Zo[(a2 - l)Z2 + (a2 - a)Zo] (93)

le = 1o + al + a'}¡12
1
E
ZoZ1 + Z1Z2 + Z2Zo[ -Z + a(Zo + Z
2
- 2 a Zo] 2) -
E
- ZoZl + z1z2 + z,z}(a - l)Z2 + (a - a2)Zo] (94)
Since the three sequen ce componente of terminal :roltage
are equal, i.e:,
(95)
and
(96)
then
(97)
Line-to-line Fault. The circuit condition to which this case

reduces is shown in Fig. 22.
Fw. 22.-Line-to-lino fault on a generator.
E
11 = Z1 + Zz (98)
E
lz = - Z1 + Zz (99)
lo= O
Therefore
la= lo+ l1 + lz
=O+ E E =0
Z1 Zz + Z1 Zz +
1b = 1o + 2
+
a 11 al 2
-O+ E E
a z +z a z + z~
2
- 1 2 - 1
- . -).. E
= (a - a) z 1 + Zz
2
(100)
1e = 1 o + al + a 12 1
2
-O+ E - z E
- aZ1 + Zz a Z1 + Zz
E
= (a - a z + Zz = -lb
2
) {101)
1
,.
48 SYMMETRICAL COMPONEN'fS
Also
E1 = E2 = -Z2l2
_ z2
E (102)
z1 + z2
Therefore
Ea = Eo + E1 + E2
_ 2z2 E (103)
z1 + z2
Eb = Eo + a2E1 + aE2
2
= (a + a)Z2E
z1 + z2
= _ z2 E (104) "
z1 + z2
Ec = Eo + aE1 + a 2E2 1
= (a + a2)Z2E = Eb (105)
1
z1 + z2
23. N umerical Example of Generator-fault Calculations.
i
To aid in a more thorough understanding of the different
types of generator faults, c·a lculations will be made of an assumed
machine for sustained short-circuits. Consider a machine in
which ,
Z1 = O + jl.O (synchronous impedance)
= 0.1 + j0.3 (impedance to negative-sequence current)
z2
Z 0 = O + j0.1 (impedance to zero-sequence current)
E= 100 +jO
Single Line-to-ground Fault. Referring to Fig. 20:
100 100
lo = 11 = l 2 = Zo + Z1 + Z2 - 0.1 + j1.4
= 5.08 - j71.1
la = lo+ l1 + !2 = 3(5.08 - j71.1)
= 15.24 - j213.3
E1 =E- l1Z1
= 100 - jl.0(5.08 - j71.1)
= 28.9 - j5.08
E2 = -I2Z2
= - (5.08 - j71.1)(0.1 + j0.3)
= - 21.84 + j5.59
Eo = -loZo
- ....:. (5.08 - j71.1)(0 + jO.l)
= -7.11 - .f0.508
CALCULATIUN OF UNBALANCED FAULTS 49
Eb = Eo + a E1 + aE2
2
= ( -7.11 - j0.508) + (- 0.5 - j0.866)( +28.9 - j5.08)

+ (-0.5 + j0.866)( -21.84 + j5.59)
= ( -7.11 - j0.508) + (-18.85 - j22.49) + (6.08 - j21.70)
= -19.88 - J.44.70
Ec = Eo + aE1 + a 2
E2
= ( -7.11 - j0.508) + ( -0.5 + j0.866)( +28.9- j5.08)
+ (-0.5 - j0.866)( -21.84 + j5.59)
= ( -7.11 - j0.507) + (-10.05 + j27.57) + (15.76 + j16.1)
= -1.40 + j43.16
Double Line-to-ground Fault. For this case, referring to

Fig. 21,
ZoZ2 j0.1(0.1 + j0.3)
Zo + Z2
- j0.1 + 0.1 + j0.3
_ - 0.03 + j0.01
- 0.1 + j0.4
= 0.00588 + j0.0765
E
11 = -----==---c=--
z + ZoZ2
Zo + Z2
1
100
jl.O + 0.00588 + j0.0765
= ~~~~~~~~
100
0.00588 + jl.0765
- -=---::--=-=-::::-=--,--------:-..,..--::,--:~
= 0.50 - j92.8
ZoZ2
12 = _Zo + Z2¡1
z2
= - (0.00588 + j0.0765) (O _ .92 )
0.1 + j0.3 .50 J .
8
= -5.58 + j21.81
ZoZ2
lo= _Zo + Z211
Zo
= - (0.005!8j:1 j0.0765) (0.50 - j92.8)
= 5.07 + j71.02
50 SY M METRICAL COMPONENTS
Combining the sequence components, the following phase

currents are obtained:
la= O
lb = lo + a2J 1 + al 2
= (5.07 + j71.02) + (- 0.5 - j.866) (0.50 - j92.8)
+ (-0.5 + j.866)( -5.58 + j21.81)
= -91.8 + j101.3
le = lo + al 1 + a2J 2
= (5.07 + j71.02) + (-0.5 + j.866)(0.5 - j92.8)
+ (-0.5 - j.866)( -5.58 + j21.81)
= 106.9 + j111.8
The sequence voltages and phase a voltage are /
E1 = E2 = E o
E2 = -Zolo
= -j0.1(5.07 + j71.02)
= 7.10 - j0.507
Ea= 3E2
= 3(7.10 - j0.507\
= 21.3 - jl.521
Line-to-line Fault. Referring to Fig. 22:

E
11 = zl + z2
100
j1 + 0.1 + j0.3
- -=-:---:---:::--:--~-;:::--::::
= 5.88 - j76.5 \
l2 = -l1
= -5.88 + j76.5
lo= O
la = lo+ l1 + l2 = O
lb = lo + a2l 1 + al 2 = O + (a 2 - a )l 1
= -j1.732z
1
!
z2 = -j1.732(5.88 - j76.5)
= -132.5 - j10.2
le = I o + al 1 + a2l 2 = O + (a - a 2 ) l 1 = -lb
= 132.5 + j10.2
E1 = E2 = -Z2l2
= - (0.1 + j0.3) (- 5.88 + j76.5)
= 23.54 - j5.89
Eo =O
..
•,
Ea - Eo + E1 + E2
- o + 2(23.54 - j5.89)
- 47.08 - j11.78
Eb - Eo + a 2E1 + aE2
- O + (a 2 + a)E1 = -E1
- -23.54 + j5.89
E e - Eo + aE1 + a 2 E2
- O + (a + a 2 )E 1 = -E¡
- -23.54 + j5.89
Problems
1. Assume a system with a sustained supply voltage of 2,300 volts from
line-to-neutral and with linc impedances Zt = +}10, Z2 = +}8, and Zo =
10 + j12. Find the currcnts in thc different phases for different types of
faults as follows: (a) single linc-to-ground; (b) double line-to-ground¡ (e)
line-to-line; (d) three-phase.
2. Assume that a gencrator with constant interna! voltage is subjected
to different types of faults and that all rcsistances are negligible. Also
assume that the absolute values of fault current for thc diffcrent types are as
follows: (a) three-phase, 1,000 amp.; (b) line-to-line, 1,400 amp.; (e) line-to-
ground, 2,220 amp. If the positive-sequence voltage to neutral is 2,000
volts, what are the positivc-, ncgative-, and zero-sequence impedances?
3. A generator is suhjectcd to a sustained fault with phases b and e short-
circuited together and to neutral. Show that thc zero-sequence impedance
may be detcrmined from the following rclation: Zo = : : where Ea is the
voltage from the unfaultcd conductor to neutral and In is the neutral cur-
rent, i.e., the sum of the cUl-rents h and le.
Ze
Actual Network Equivodent Ci rcuit
( 01 ) ( b)
4. Assume a generator in which phases b and e are short-circuited and
connected through impedance z.to phase a as illustrated in (a) of the accom-
panying figure. If the generator impedances are Z1 and Z2, and the interna!
voltage is E t, show that the equivalent circuit is ~,?;iven in (b ).
..
CHAPTER IV
ILLUSTRATION OF UNBALANCED FAULT CALCULATIONS
The calculation of system performance at times of unsymmetri- 1

cal faults is one of the most important applications of the method
of symmetrical components. Such calculations are important
from the standpoint of circuit-breaker and relay application,
mechanical and thermal limitations of apparatus, and system
stability.
In Chaps. II and III, a method was developed for the calcula-
tion of unsymmetrical faults on systems in which the impedances
may be considered as symmetrical in the three phases. It
was shown that the three sequences could be represented individ-
ually by separate networks, the connection between these
networks being determined by the particular type of fault
considered. · An example has already been given showing the
application of these methods to the simplest type of network,
namely a generator, which was completely soJved analytically
and then numerically by substitution in the derived formulas.
l
1
Such analytical expressions for practica! networks, however,

become so unwieldy that recours~ is usually taken to the solution
of the netwurk with numerical ~rcuiL consLanLs. The present
chapter presents the solution of a network of sufficient com-
plexity to illustrate sorne of the necessary steps involved in
-1
the simplification of the network and also the manipulation
of the sequence networks to obtain the current and voltage
distribution in the different phases.
24. C?lhventional Simplifying Assumptions in Short-circuit
Calculations.
A large percentage of the problems, within the accuracy
ordinarily required, can be solved by making the simplifying
assumptions enumerated below. These simplifying assumptions
are those which are normally employed for the calculation of
short-circuit currents by means of t he direct-current calculating
ooard . ~
ILLUSTRATION OF UNBALANCED FAULT CALCULATIONS 53
l. All generated e.m.fs. are of equal magnitude and in phase. *
2. All resistances are neglected, only the reactance of im-
pedances or an equivalent reactance being used.
3. All shunt-impedance branches, such as the equivalent
circuits for representing magnetizing currents of transformers
and charging current of transmission lines, as well as equivalent
impedance branches representing normalloads, are neglected.
4. Mutual t impedance between lines is neglected.
Consideration will be given subsequently to methods for
'
avoiding the limitations imposed by the above assumptions.
20,0~ ~' 66l<v. Y.. ~
30miles Vo copper . ~ ts.ooo'i{V:01
6ftspotce
20
·~ Y F Y.. sl1. J.:.
~ ~~
(ot) 15,000 l<v""'
fief~f!lt'lc~ ~ pry_~-v~t§: Q.+..j~JOQ _____ • .,
n •t · : "65 "20 22 30 '50:
r'OSr IVe-.......,IVV'--.JVV~---"""'v-----...."MM--.~......,..
Sequence,
Network•
Fxo. 23.-Single-line diagram of a network and the corresponding positive-,

negative-, and zero-sequence networks.
25. Layout of Assumed System.
To illustrate the method for the calculation of faults in whicb
these assumptions are justified, a numerical example will be
worked out. For this purpose the networkt shown in Fig. 23
* The case of non-identical generatcd e.m.fs. is treated in Sec. 113.
t The zero-sequence mutual impedance between lines, if the fault occurs
between bussing points, must be taken into consideration forfaults to ground.
t This network corresponds to Fig. 10 of Chap. III, with the omission of
the shunt loads.
is selected as having a degree of complexity sufficient to illustrate

sorne ·of the necessary steps of network simplification. This
network consists essentially of two generating stations connected
t hrough two 30-mile, 66-kv. transmission lines. The lines
may be considered as on separate rights-of-way, to justify t he
assumption t hat mutual impedance may be neglected. F aults
of different characters will be considered as occurring at t he
point F, about eight miles from the left-hand bus. The unusual
transformer connections are employed merely to illust rate the
manipulation of the zero-sequence currents.
26. Single-Une Diagram of the System and the Three Sequence
Networks.
The first step is to set up the network impedances in the form
of a single-line diagram for thc: three separate sequences. All the
impedances should be reduced to a common voltage, which in
this case is the 66-kv. transmission voltage. It will be assumed
t:Rat these impedances are known and of the values indicated
in Fig. 23(b), (e), and (d). The negative-sequence network is
identical with the positive-sequence network except for the
impedance of the synchronous machines. With the exception
of the transformer impedances, the zero-sequence impedances
are, in general, different from the impcdanccs of either the posi-
tive- or negative-sequence. Also, t he zero-sequence diagram
generally possesses a smaller number of branches. From what
has already been said, branches which do not permit the flow of
neutral or ground current can be neglccted in the consideration
of the zéro-sequence network. Also, in the solution of all net-
work problems, one is at liberty to choose an arbitrary direction
of current fiow in the individual branches in setting up the
problem. This is a freedom permissible in these problems with
t he limitations that the current jlow must be the sarne in the
corresponding branches of all three sequences. This point was
discussed in a previous chapter, but it is again introduced because
of t he ease with which this very important point might be
overlooked.
27. Outline of Short-circuit Calculations.
The general outline of the calculations is somewhat as follows:
l. The determination of the single equivalent impedance
of the positive-sequence network and the distribution factor~
í·
¡ ..:.J.
giving the current in the individual branches for unit positive-
sequence current at the point of fault.
2. The corresponding equivalent impedance and distribution
factors for the negative-sequence network.
3. The corresponding equivalent impedance and distribution
factors for the zero-sequence network.
The above calculati~:ms will be common to all types of faults.
The remainder of the calculations are different for the different
types of faults, but in all cases the necessary steps are as follows:
4. Grouping of equivalent impedances and the determination
of the sequence currents at the fault.
5. Determination of the sequence-current distribution through-
out the three networks by the application of the distribution
factors to thc sequence currents at the fault.
6. Synthesis of pnase currcnts from thc scquence components.
7. Determination of sequcnce voltages throughout the three
networks from the currenL distribution and branch impedances.
8. ·Synthesis of phase voltages and line-to-line voltages from
the sequence cornponents.
9. The conversion of the equivalent voltages a,nd currents
from the common voltage base to the actual voltage correspond-
ing to the actual transformer connections.
28. Equivalent Impedances and Distribution Factors.
The first step is the determination of the single equivalent
impedance and the current distribution factors for each sequence
network. Thesc two calculations are conveniently rnade together
taking the three sequences in turn.
The calculation of the positive-sequence equivalent impedance
from the sources to the fault will now be undertaken. Figure
24(a) shows the positive-sequence network of Fig. 23(b) redrawn
avd sirnplified by combining the generator and transformer
impedances. Figure 24(a) may be simplified to (b) by reducing
by conventional methods;* the two parallel impedances at the
extreme right to a single impedance, the two parallel impedances
at the extreme left to a single impedance, and the remaining
impedances, which are connected in delta, to an equivalent group
of star-connected impedances. By combining impedances in
series or in parallel successively, it is possible to reduce the net-
work of Fig. 24(b) to the form shown in (e) and (d), giving finally
* See Appendix, Section VI.
..,. the equivalent impedance

lA .._~NV'---. ..---h.I-·NV'-~ e
1 .__--.NVvvv----.1 of the positive-sequence net-
18 -...VV.NV'-_. D
1 work shown in (e) and hav-
ing a value of j25.6.
F The distribution factor of
1 j42.5 any branch for the positive-
1
1 sequence network is defined
1 (b) as the ratio of the positive-
1 j2.18
sequence current in that
§•
branch to the positive-se-
~r---~----------------~
:¡ j45.5 quence current at the fault.
&~--~~~----r---~~--~
The most convenient way
.t j2.18
:¡
(C:) to obtain these factors is to
~
é.3 as su m e unit posi tive-se-
1 quence current at the point of
1 lj23.4 fault and determine the
1 <on
1
1 tj 2.18
distribution of current by
working back through the
1 • network. The various steps
1
1
<e>
! j25.6=
}Equivalent impeclolnce in this procedure are indi-
l cated in Fig. 24(e) to (i).
The division of current in
0.514tilmp. 0.486amp.
~(f) and (g) can be obtained
~ \ormp. from the familiar fact that
(f)
the current divides inversely
as the impedances.
In the transition from
Fig. 24(g) to (h), it is neces-
sary to find the current dis-
tribution in a delta group of
impedances from the known
current distribution in the
equivalent star group. This
may be accom plished by
making use of the fact that
the potential difference be-
tween any two terminals of
the delta group is the poten-
t e'lmp. oi~p. (j) ~1amp. ~~~- ~~p. tial difference between the
. · d
F IG. 24 .- Cu cmt re uctwn an
. d d ' 'b .
1stn utwn
two corresponding terminals
factors for the positive-sequence network. of the star group. The
potential difference between any two terminals in the star group
can be calculated from the impedances and the known current
distribution, after which, by dividing the potential difference
between any two terminals by the impedance of the delta group
connected across these terminals, the current in the particular
impedance branch can be determined. The remaining current
divisions may be computed in a similar manner or by equating
the currents at various junction points. For example, the
potential difference between the left-hand bus and the point F,
from Fig. 24(b) and (g), is
(l.O)(j2.18) + (0.514)(j3.0) = j3.722 volts
from which the current in the delta branch between these two
points in Fig. 24(h) is
3 722
Í · - +0.621 amp.
j6.0
For clearness, the distribution of the positive-sequence current

throughout the system for unit positive-sequence current at the
fault, or the distribution factor, is shown in Fig. 24(i).
The equivalent impedance and the distribution factors for the
negative-sequence network, shown in Fig. 23(c), may be obtained
in a similar manner. The principal steps involved are indicated
in Fig. 25, of which (e) gives the equivalent impedance and (f)
gives the distribution factors. Similar data for the zero-
sequence network are shown in Fig. 26.
The distribution in all three sequences is based on a system
voltage of 66 kv. and therefore does not take into consideration
transformer ratios or phase shift in star-delta transformers.
These are discussed subsequently.
29. Single Line-to-ground Fault Calculations.

Having obtained the equivalent impedance and the distribution
factors for the different sequences, the next step is to determine
the value of the sequence current at the fault. This will vary
with the particular type of fault under consideration. Probably
the most important type of unbalanced faults on transmission
systems is the line-to-ground fault. As this fault involves all
three aequences, the calculations will be carried out in detail in
order to illustrate the steps essential to the general method.
30. Determination of Sequence Currents.

For a line-to-ground fault the three sequence networks must
be connected in series,* and as corollary of this connection the
j764
(b)
(e)
"S&
~
inLJj2.1& .
j30.5
(C) j7.64
j 73
(Di)
tj22.95 (ti{) r2L51
j2.18 j7.64
<e> { . j25.13=
1Equtvorlet1t impeol(itnCe
Disfn'bufion
(e)
ij 29.15.
~qu/va/enf
1mpedance
.Pacfors
~6
o 0.,705 .-SQ!.!.
Fw. 25.-Circuit reduction and Fw. 26.- Circuit reduction and

distribution factors for the ncgative- distribu tion factors for the zcro-
sequence network. scquence nctwork.
three sequence currents at the fault are necessarily equal. The
value of these currents, since only the terminal currents are
required, may be obtained by replacing the three sequence
networks by their equivalent impedances. The equivalent
network for the determination of the sequence currents at the
fault, therefore, reduces to that shown in Fig. 27, from which
E
I1 = lz =lo= z 1 + Zz + Zo (106)
Now for the particular case under consideration, in which
all the impedances are purely reactive, the equivalent impedances
* This was developed in Chap. III and is illuatrated in Fi.g. 12.
will be purely imaginary. If the phase of the voltage E be taken
as purely imaginary so that E = jE, the values of I 1, I z, and
[ 0 are all real and the currents in all
branches are real. This arbitrary EGJuiv~!Qnt Posi-t-ive-

Sequence lmF_>eolornce
choice of phase of the rcference volt- z, · j25.6
age is taken to avoid the awkward Equiv~len+ Ne9ortive- 1¡..
Sequence lmpeolornce l' •
negativc imaginary terms for currents z 1 =j '25.13
that result if the reference voltage Eqoivollen+ Zero- 1
10
were located along the real axis. U sing Sequence lrn_peolornce
z0 • j29.15 l'
this assumption gives idcntically the Fw. 27.-Combination of
type of solution obtained with the equivalent impedances for a
direct-current calculating board, in single line-to-ground fault.
which the systcm rcactanccs are replaced by resistances in the
board. With this voltage reference
j38,100
I = lz = lo = j(25.6 + 25.13 + 29.15) = 477 amp.
1
The current distribution in the individual networks may now

be obtained by the application of thc distribution factor to
the sequence current at the fault . . For example, the positive-
sequence current in the sound line is equal to the positive-
sequence current at the point of fault (477 amp.) times the
distribution factor for thc positive-sequence current [0.107 from
Fig. 24(i)], which is equal to 51.1 amp. Following this method
the distribution of current for thc three sequences is obtained,
the results of which are shown in Fig. 28.
31. Phase Currents from Sequence Currents.
In the development of the general method in the preceding
chapters, the a phase was used as the reference phase and the
line-to-ground fault was applied to this phase. The currents
in the different phascs may therefore be obtained by application
of the following formulas:
la=lo+ll+l2 }
l b = lo+ a2J1 + al2 (107)
le = lo + al 1 + a212
These formulas give correct results only if the arbitrarily
chosen directions of positive current flow are identical f or the three
sequence networks.
As an illustration of the procedure which may be employed
for each branch, the currents flowing in the individual conductors
of the faulty line from the right-hand bus will be determined.

From Fig. 28 the sequence currents in this lineare
l o = 135.3 amp.
l1 = 180.7 amp.
l2 = 209.0 amp.
These currents are all in phase and, because of the choice
of the phase position of the generated voltage, are all real.
~ ..lll 11s.9 ~e The current in phase
~r-~--~~--~~~~ aisthen
Posi+ive-
sequenc~ ~f.:\D la = 135.3 + 180.7
&--5 ~'-+==-.:---:-:-=:--=-~~ ~ + 209.0
= 525 amp.
~ ....=94;::::::.5=.----=l:::9.0=--~~~"=4~ ~
'--'5c-- ~ The currents in
Neg01+ive- h b d b
sequenc~ 144 ~-L~f"'\ p ases an e are o -
\.-15 >=~-=~-==---....;:::::: ..=< ~ tained in a similar
477
t manner by applying
~ 336.4 ~ .$_ ~ the above relations,
Zero- r-~---''---~~'-< ~ but, since the se-
sequenc~
5~6,.....,. , quence components of
\.-15 r---+--~__;:..__;;:...__...:::.==< r-u current are all real, a
simpler form may be
FIG. 28.-Current distribution in the three sequcnce
networks. used to advantage.
Rewriting the expression for l b and l e and remembering that
lo, 11, and 12 are all real, there follows that
lb =lo+ ( -0.5 - j0.866)l1 + (-0.5 + j0.866)l2
le = lo +(
-0.5 + j0.866)1 1 +(
-0.5 - j0.866)l 2
and
lb = lo - 0.5(1 1 + 1 2) - j0.866(1 1 - 1 2) (108)
le = l o - 0.5(1 1 + 1 2) +j0.866(I 1 - 1 2) (109)
It will be observed that lb and fa
are conjugate, i.e., that
their real parts are identical and their imaginary parts are equal
but of opposite sign. Substituting the numerical values given
lb = 135.3 - 0.5(180.7 + 209.0) - j0.866(180.7 - 209.0)
= -59.6 + j24.5
le = -59.6 - j24.5
In this manner the phase currents throughout the entire
system are determined. The results may be conveniently
JLLUSTRATTON OF UNBALANCED FAULT CALCULATIONS 61
tabulated in the form shown in Fig. 29. The currents shown
are expressed in amperes on the 66-kv. base throughout. They
must be modified by
~11
. r .~ ~_f,;\c
the introduct ion of f.\A 2
~ }==r---..:::s=----r=s=-< ~
transformer ratios to Phase(a)
show the actual phase ® ~ ~ r==.J..-=~----==::..__...¡;::s:~ ~
currents.
32. Determination of
Sequence Voltages.
The sequencc com-
ponent of line-to-
neu tral vol tage m ay f3Q¡j14.b
be obtained from a
knowledge o f t h e r-:::=....-==r-----==:::--~~ ~
sequence-current dis-
tribution and the Fw. 20.- Phasc current distribution.
branch impedances. In particular, the positive-sequence voltage

at a point is equal to the generated voltage minus the
positive-sequence drop from the source to the point under
lntern01l Yolts ~ r----.-----+-~ ~ lnternOtl Volts
•j38,100 ~ ~ •j38,100
lnter-noll Volts~ ~ lnteYnaiYolts

ao ~~-.-----+--~~ :Q
-j 6,730 j 6,330
Internos! VoltsQ----Li -·13,910 ~
· 2110 Jnternod Volts
cO r-_.~~~--+-~ =o
Zero-
seotuence
~. Fw. 30.- Voltagc distribution in thc t hrce sequence networks.
consideration. For example, the positive-sequence voltage
(
of the left-hand bus is equal to the generated voltage (j38,100)

minus the drop due to the positive-sequence current (122.6
amp.) in the generator and transformer (j85 X 122.6), which
is equal to j27 ,680 volts.
Similarly, the negative-sequence voltage at any point is
equal to the generated negative-sequence voltage (which in
this case is zero) minus the negative-sequence drop from the

source to the point under consideration. For the left-hand
bus this becomes
O - j110 X 94.5 = -j10,390 volts
Similarly, for the zero-sequence network the voltage of the
left-hand bus is
O - j20 X 33.64 = - j6, 730 volts
The voltage distribution for the three sequences for the
entir~ system is obtained in this manner, the results being
shown in Fig. 30.
For cases in which extensive voltage calculations are involved,
the voltage distribution factor may be obtained for different
points throughout the system for each of the three sequence
networks. These may be obtained and applied in the manner
analogous to that described for currents.
33. Line-to-neutral Voltages from Sequence Voltages.
The line-to-neutral voltages are obtained from the sequence
voltages by applying the fundamental conversion formulas
(24), (25), and (26), which for ready reference are repeated
below
Ea = Eao + Eal + E a2
Eb = Eao + a 2Eal + aEa2
E c = Eao + aEal + a2Ea2
These expressions may be simplified by rewriting in the following
form
E a = Eao + Eal + Ea2 (1 10)
Eb = -j0.866(E1 - E2) +Ea- 0.5(E¡ + E2) (111)
E c = +j0.866(E¡ - E2) + Ea - 0.5(E¡ + E2) (112)
Since, for this particular case, all the voltages are purely imagi-
nary, it will be observed that the first terms of Eb and Ec are
real and the negative of each other, and that the second terms
of Eb and E c are purely imaginary and equal. Voltages Eb
and Ec are then mirror images of each other about the imaginary
axis.
Inserting the values of positive-, negative-, and zero-sequence
voltages for the left-hand bus, the following values of line-to-
netitral voltages are obtained.
Ea = j( -6,730 + 27,680 - 10,390)

= j10,560 volts
Eb = -j0.866 (j27,680 + jl0,390)
. + [-j6,730 - 0.5(j27,680 ·_ jl0,390)]
= 32,970 - j15,370 volts
Ec = -32,970 - j15,370 volts
The line-to-neutral voltages for the entire system are given in
Fig. 31.
32,970-.,15,370 32,450-j 16;4-10

32¡330-"20,860 f--t-()33poo-j19,050
'J 31,720~j 16,090
FIG. 31.- Line-to-neutral voltage distribution.
34. Line-to-line Voltages.

The line-to-line voltages may be obtained from a knowledge
of the line-to-neutral voltages by applying the relations of
equations (36) and Fig. 7. These equations are repeated for
convenience.
EA = Ec Eb ___.,!
En= Ea- Ec
Ec = Eb- Ea
Applied to the voltages at the left-hand bus
EA = ( -32,970 - j15,370) - (32,970 - j15,370)
= -65,940
En = j10,560 - ( -32,970 - j15,370)
= 32,970 + J"25,930
Ec = (32,970 - j15,370) - (j10,560)
= 32,970 - j25,930
When the line-to-neutral voltages are not desired, the line-

to-line voltages may be obtained directly from the sequence
components of line-to-neutral voltages by means of the following
relations:
EA = jyi3(Ea1 - E a2)
En = 1.5(Eal + E a2) -- j0.866(Eal - Ea2)
Ec = -1.5(Eal + Ea2) - j0.866(Eal - E a2)
The foregoing equations, it must be remembered, are valid

only for the particular assumptions which for the sake of emphasis
will be repeated. The resistances of the network branches
are negligible and the generated voltages are of the positive-
sequence only. In addition, the voltage of the reference phase
(phase a of the star gro u p) is purcly imaginary.
35. The Conversion of Currents and Voltages Due to Trans-

formers.
To this point all calculations have been made on the equivalent
star basis corresponding to the transmission voltage of 66 kv.
For the transmission line itself this gave the actual currents
and voltages of the three conductors, but for the low-voltage
portion of the system ccrtain transformation constants must be
applied. Hence, the determination of the actual currents and
voltages where trn,nsformations have taken place requires
conversion in terms of the new base. The phenomenon is com-
plicated to a certain extent by the fact that associated with
every transformation there must necessarily be sorne impedance
and magnetizing curren t. In most cases the magnet izing current
is so small in comparison wit}\ the load current under short-
circuit conditions that it may bé neglected, as has been done in
this particular case. The problem may be further simplified
by divorcing t he im pedancc and conversion aspects by considering
the actual transformer replaced by a perfect transformer with
an impedance equal to the transformer impedance in series
with it. In this case the impedance will be considered as being
in series on the high side, because the voltage of the high side
was chosen as the voltage base. It can be shown that the
impedance could be inserted on either side.
Determination of Actual Currents. For star-star transformers,
the currents on the low side are equal to the actual turns ratio
1LLUSTRAT!ON OF UNBALANCED FAULT CALCULAT10NS 65
times the currents on the high side. For example, the actual
phase currents in generator C are from Fig. 29
66,000
la = 259.9 = 1,559 amp.
11 000
'
lb=~~,~~~( -130 + j24.3) - -780 + j146 amp.
'
le = r~,~~~( -130 - j24.3) -780 - j146 amp.
'
Star-delta transformers introduce not only a change in magni-
tude of voltages and curtents but also a change in phase angle.
For this reason it is necessary that sorne convention be adopted
1 al " =N
J 1 al 1
J1'
1a2" -- -N a2
1ao" = indeterminate
FIG. 32.-Relation between atar currenta in atar-delta or delta-atar tranaforma-
tiona.
for designating the phases on the delta side with reference to

those on the star side. Figure 32 shows the convention that
has been chosen. It will be noted that on the delta side the
subscript for the current in any terminal is the same lette:r
as the subscript of the current in the winding opposite. In
the notation small subscripts refer to line currents, large sub-
scripts to delta currents, and single prime and double primes to
assumed line currents on the delta and star sides, respectively.
This notation conforms to that chosen in Sec. 8.
Calling n the turns ratio of the transformer and N the line
voltage ratio (N = v3n), the following nilations may be derived
by inspection of Fig. 32:
•
' - -1n
1a - n(lb
' - le' - " -e
l ") -- N (lb l ")
" -c
0
h' =le' -l./= n(lc''- la")= N (le''- la")
V3
I c' -
- A
I ' -n
1 '-
- n(l a
" -b
l ") = N (l " - lb")
VSa
Following this method the actual currents in all parts of the

system can be determined, taking both turn ratio and trans-
former connections into account.
Illustration. To illustratc thesc relations, lct it be desired to determine
the currents in generator A. Thc star currents on thc 66-kv'. pase may be.
obtained from Fig. 29, from which /
la' = [( -108.6 - j24.3) - ,e-108.6 + J24.3)] j168.4

~~;~~~.}a = -
lb' =~~;~~~.}a [( -108.6. + j24.3) - (217.1 )] -1,128 + j84.2 =
le' = ~~;~~~.}a [(217.1) -· ( -108.6 - j24.3)] = 1,128 + j84.2

An alternative method for determining the phase currcnts, when star-delta
transformations are involved, is to consider thc cffect of the transformation
upon the sequence componcnt of currents individually. Since the equations
(114) hold for any values of current, then for the positive-sequence com-
ponent alone, from the first equation
l at·· = ~(a2 - a)l at'' = -jNial" (115)
Similarly, for the ncgative-sequencc componcnt alone
la2' = N (a - a 2 )Ia2" = jN/a2" (116)

V3
and for the zero-sequencc alone
lao 1 = ~(1 - 1)lao11 = O (117)
Conversely, the\ne currents on thc delta sidc may be obtained in terms

of the line currentJ on the star side.
¡ al " -- N
J Ial'
I a2 " = i¡a21
-N
I a o" = indeterminate
The above rclations hold whether the transformer is grounded or not.
In general, n and N may be greater or less than unity.
With cascade transformers either of the methods just described
may be applied to each transformation successively._ However,
in general, it will be found more convenient to follow the effect
of the various transformations upon each of the phase-sequence
components of current. Proceeding from any point at which
the actualline currents are known, the rule is as follows.
In passing through a star-delta or a delta-star transformation
the positive-sequence components of line or phase currents on the
star side lead the corresponding quantities on the delta side by
90 deg.; similarly the negative-sequence components on the star
side lag the corrcsponding quantities on the delta side by 90 deg.
In addition, the ratio of transformation must be taken into
account. The zero-sequence current will be blocked at the first
star-delta transformation. Applying this rule to the transformer
e Hz
~~~-~
1(11' ~
..__ __ __,
b
(DI)
1(11'~()1 lm"~ t1t be
le/'~
~-12
"'
i)l;-~é,
b
¡,
(b)
Fra. 33.-Cascade transformer connections.
combination in Fig. 33(a), it will be observed that the following

relations hold.
1 al 11 = - J·N¡ 1 al
111
1 a2 11 = J·N 11 a2,,
and
1 al ' = - J.N21 a l " = - N1N z.t' a1'"
1a2 1 = jN2l a2 11 - -N1N2lq,2"'
and to Fig. 33(b)
1al1 = +jN21at'' = N1N21 a1"'
1a2 1 = -jN21a211 = N1N2!a2"'
A double star-delta transformation reverses both the positive-
and negative-sequence components of current and ~onsequently
the phase currents, but a star-delta, delta-star t.ransformation
retains the original sense of the sequence components and phase
currents without change in phase relation.
It is important that the notation indicated (see also Fig. 32)

be strictly adhered to, for otherwise the reference phases become
confused.
Determination of Actual Voltages. The voltages on the low
side of a star-star transformer are equal to the voltages on the
high side less the drop through the transformer branch on the
high-voltage base divided by the ratio of transformation.
The effect of star-delta transformers upon the phase position
of the voltages to neutral must be the same as that upon the
phase position of the line currents, for otherwise the values of
power and reactive volt-amperes on both sides of the ideal
transforrner would not be equal. The,refore, the positive-
111 tsequence star voltage
in going through a star-
Et~t" delta transformation is
rotated backward 90
deg., and in passing
e b through a delta-star
transformation forward
---------'N·v'Jn 90 deg. Similarly, the
E , .Eat" E a1 " = jNEa 1 ' negative-sequence volt-
al = - J--¡¡- Ea/' = - jNEa/ age-to-neutral in pass-
, . Eaz" E ao" = indetermi- ing through a star-delta
E a2 =J - -
E ao' =
0
N nate transformation is rotat-
FIG. 34.-Relation between star volta~es instar- ed forward 90 deg. and,
delta or del ta-star transformatiohs. passing throu~h a delta-
star transformation, backward 90 deg. These re1.ations, with
the notation corresponding to that used in Fig. 32 are expressed
in Fi\34. Since _
EA1 = jy3Ea11
1
it follows that
EA1' = jv3(- jE;;") -
Also, since
EA2' = -jv3Ea2'
it follows that
EA,' ~ -jvi3(l;{) ~ '{}E.,'' ~ E:"
Cascaded transfonners rnay be handled by considering the
transformations successively just as outlined in the treatment
of the current relations.
36. Double Line-to-ground Fault Calculations.

The calculation of double line-to-ground faults differs from
that for single line-to-ground faults only in the arrangement
of the equivalent impedances. As developed in Sec. 17, the
impedances should be grouped as shown in Fig. 35. It is impor-
tant as emphasized and explained before, that the positive
direction of current fiow be maintained as indicated. Solving,
it is found that
/¡ = j38,100
Z oZ2 z +
Zo + Z2
1
j38,100
- ------~~~---==-.~
. + .29.15 X 25.13
J 25 .6 J 54.28
= 975 amp.
Zo ( 1
12 = z +z
o 2 - ¡)
~29.15( -975)
=
J54.28
- -523.5 amp.
lo - -451.5 amp.
E E
Equ1'ot'J1/mfposit1~ Equ!WJ/Itnf ~N!!?J~Xfi fqul'ot'JI!tnf uro- [Jfvivtl/6nf III!Jorfi.-.-
seque~ impemnct uei1C4 1mperiance sequenct imped11nat uence impedet11e1
l, ¡ Ztj25.6 .~~~ Z1•j25.13 10~ Z0 j29.15
Fw. 35. -Combination of cquivalent Fw. 36.-Combination of e qui v-

impeda n ces for a double line-to-ground alent impedances for a line-to-
fault. line fault.
It will be observed that both 12 and 1 0 are of negative value.

Now applying again the distribution factors from Figs. 24, 25,
and 26 ' to these sequence components of current at the fault,
the sequence currents, and, after combining, the phase currents
at all points in the system may be determined. Voltages may
be calculated in the same manner as described for single line-to-
ground faults.
37. Line-to-line Fault Calculations.
The zero-sequence network does not enter into the calculation
of the line-to-line fault. This is to be expected, since the bound-
ary conditions at the fault require only positive- and negative...

sequence voltages to define them. The arrangement of the
equivalent network for this type of fault was developed in Sec. 18
and is as indicated in Fig. 36. Solving the network,
j38,100
11 = j(25.6 + 25.13) = 751 amp.
!2 = - ] ¡ = -751 amp.
From this point the procedure is the same as for the calculation
of the two other types of faults.
38. Comparison of Short-circuit Calculations and Test Results.
It is of interest to compare the results of calculations by the
method of symmetrical components of fault currents and voltages
on a complicated system with the results obtained by actual
tests. In 1925, during an extensive stability investigation, an
opportunity was afforded to obtain a very complete record of
power-system voltages and currents at different points in the
system. In addition the systcm characteristics were accurately
known because of the stability investigation, and calculations
were made for the conditions existing at the time of the fault.
The power systeJJ}. is shown diagrammatically in Fig. 37. *
The transmission circuits included a 220-kv. transmission line
202 miles long paralleled by a similar line operated at 110 kv.
The short-circuit calculations require consideration of
l. Distributed capacity of the 202-mile transmission line.
2. Inductive~upling to zero-sequence between the 110-kv.
and 220-kv. lines.
3. Unequal transformer ratios in the different circuits.
The results of calculation are shown by means of voltage and
current vectors at different points in the system diagram of Fig.
37. All the vectors are given on the comrnon voltage base of
110 kv., and to obtain the actual values of the currents and volt-
ages in any particular part of the circuit it is necessary to take
into consideration the transformer ratio and phase shifts dueto
star-delta transformations.
Figure 37 shows the radical distortion of voltages and currents
that obtain under unbalanced faults. The distortion is accen-
tuated by the relatively small value of load current in comparison
with the fault current and even with the charging current on
* Figure 37 and Table II are taken from "Studies of Transmission
Stability.'' (28 >
'-
~
t-t
~ ~
""'3
~
¡.,. ~
1-'3
Conoletlser .......
1 o
, le ~
"~('
le 1 -~
'"i
lb._,
lb ~
' ~ 1
'
1 lb 1 le 1
1
e::
1
\ : IQII
1
1 lb 1 ~
: la 1 1 - t:r.:!
Ec .. 1' ~ 1 V01ca substahon ~
'
l
1
1
1 t-<
1
1 220Kv. 1 1 ~
l 1
, / ~
~
~""/, t_,;j
/
t:;,
~ ~
~
IIOKv.
1
1 1 1
81-'3
Eb 1 1
1
Eb 1
1
Hor+ Creek No.1112 1 1
1 ~
1 1
Volt01~e SCOIIe !..,) 1 1 1 1 1 1 1 1 ' 1 1 1 1 1 1 1 ~
1 10 O 10 20 3040.50 100 ISO K\!
~~
t
1 1 t-<
Currerr;- ~>COl le lw·l 1 1 1 1 11 1 1 1 1 1 1 1 1 1 1 1 1 1 11 1 1 1 1 1 ~
1 1 le 1
100 o
. ~ 500 1000 1500 2000 2.500 1
I~
8
..),
IQII
lb
(A JI voHcrges crnd curren-ts on 110 Kv. br;rse) lb{ 110
lt~~
,,,,b
1
~
o~
IDI
IDI
Condenser ~
F10. 37.-Voltage and current distribution on the transmission system of the Pacific Gas and Electric Company with
-a. single line-to-ground fault on the 220-kv. line near Vaca substation-calculated. ~
¡.......
..
the 220-kv. line. The effect of the charging current is brought

out by the change in the currents between the sending and fault
ends of the 220-kv. line. It may be pointed out that the vector
diagrams show that the proportion of zero-sequence current is
relatively large in the 220-kv. line and relatively small in the
110-kv. line. This is due to the fact that the zero-sequence
current induced in the 110-kv. line by the current in the 220-kv.
line is opposite in direction and approximately equal to the zero-
sequence current that would tend to fiow conductively through
the 110-kv. line to the Vaca bus and through the autotransformer
to the fault.
In Table II is given the comparison of the calculated voltages
and currents and the corresponding quantities observed during
TABLE II.-COMPARISON OF MEASURED AND CALCULATED lNS TANTANEOUS
SYMMETRICAL VoLTAGES AND CuRRENTs FOR LINE-TO-GROUND FAULT
AT THE V ACA-DIXON SUBSTATION
Currents and voltages 1 Measured 1 Calculated
Zero-sequence current in 220-kv. circuit-breaker at

Vaca ... ........ .... . . .... . . . . ... ... ....... . 1 , 020 1 , 130
Voltage, e phase at Vaca .... . ................. . 60,000 59,300
Zero-sequence voltage, 110-kv. bus at Vaca ... .. . . . 14 ,200 13,300
Zero-sequence current, 220-kv. line at Pit l ...... . 140 138
Current a phaso, 220-kv. line at Pit l. .... .. ..... . 235 280
Voltage A phase, Pit 1 generator ............ . ... . 4 ,840 4 ,996
Current A phase, 110-kv. line on low voltage side at
Pit l ................... . .................. . 1,070 1 ,030
the test. The calculated values are based on the instantaneous

symmetrical value of the short-circuit currents and the test
results are based on the values during the fourth cycle after the
application of the fault, this cycle being chosen to minimize
the effect of the rapidly decaying subtransient components of
current. The close agreement obtained, considering the factors
involved, serves as a satisfactory check on the general method.
Problems /
1. For the system of Fig. 23 (a), determine the actual currents and voltages
of the generators C and D for the single line-to-ground fault at F.
2. Assume that faults of different types occur on the right-hand bus of
Fig. 23(a). Determine the fault currents for (a) single line-to-ground fault;
(b) double line-to-ground fault; (e) line-to-line fault; (d) three-phase fault.
·~

3. For the same system and faults as Prob. 2 determine the corresponding
fault voltages.
4. In Fig. 29 it is shown that for a single line-to-ground fault at .F the
currents in the b and e conductora of the transmission line are not zero.
Explain wl'\y this condition exists. State the conditions in regard to network
impcdanccs which are necessary in ordcr to eliminate the phase b and e
currents in generator C and in gcncrator D.
6. Consider the system of Fig. 23(a) and assume that a single line-to-
ground fault occurs on phase a through a resistance of 115 ohms. Determine
the voltages and currents for the point F, thc fault location.
6. Determine the ground currents of all transmission circuits of the system
of Prob. 5.
7. For the system of Fig. 23(a), if generator A is disconnectcd, how will
thc fault currcnts and ground currents be affected? In case generator D is
disconnccted? How are the current distribution factors affected in both
cases?
8. Assume a gencrator with an interna! voltage Eat of 7,000 volts line-
to-ncutral under sustained single-phasc load conditions on phases a and b,
If the gcncrator resistanccs are negligiblc and the rcactan~es are X1 = 1.0.
x2 = 0.15, and the load impedance ZL = 0.8 + j0.6, what are the delta
terminal voltages EA, En, and Ec'?
9. Assume a gcncrating station of n similar units, each with negligible
resistancc and with reactances of X 1, X 2, and X o, and that each machine is
opcrating with an interna} voltage of Eai· Calculations are to be made to
determine the relative current stress in the generators for different methods
of grounding the generator neutrals. Find the analytical expression for the
phase currcnt in each machine for (a) the neutral grounded on all machines;
(b) the ncu tral grounded on only one machine. If the number of ungrounded
machines, in parallel with the grounded neutral machine, approaches
infinity, what value does the current in the grounded neutral machine
approach? Compare this expression with that for the currents on a three-
phase fault.
CHAPTER V
CONSTANTS OF SYNCHRONOUS MACHINES*
Only positive-sequence voltages are generated in most com-

mercial systems. System short-circuit phenomena are so closely
related to the characteristics of synchronous generators, motors
and condensers that their performance under short-circuit
conditions will be reviewed at sorne length.
39. Three-ph~se Short-circuits of Generators.

Assume that an instantaneous three-phase short-circuit is
applied across the terminals of a synchronous machine operating
·~=
IJ0NV'N ........... " .... " ' " ' "
WVVVVVWVVVVVVVWVlJ VVIfVVIfV'r¡--•""'v.....
vm
. v
l
FIG. 38.- Three-phase short-cir cuit on a 75,000-kva. 11 ,600-volt- 60-cycle 1,800
r. p.m. turbo-gcnerator.
with. constant excitation and no load. The characteristic current

waves for the three phases are shown in Fig. 38. It will be
observed that in general each current wave consists of two kinds
of components, namely,
l. Alternating-current components.
2. Unidirectional or direct-currcnt components.
The former are equal in all three phases. The latter are depend-
ent upon the particula r point of the cycle at which the short-
circuit occurs.
* The treatment in this chapter presupposes an elementary knowledge
of synchronous-machine theory as well as simple transients. For the latter
the reader is referred to "Electric Transients" by C. E. · Magnusson,
McGraw-Hill Book Company, Inc., New York, 1926.
74
CONSTANTS OF SYNCHRONOUS MACHINES 75
40. Altemating-current Components.
Figure 39 shows an enlarged view of the alternating-current
component from which it may be observed that this wave may
in turn be resolved into thrce components.
a. Sustained Current. For a prolonged short-circuit the
armature current :finally attains t he sustained value I indicated
by the cycle shown in ·dotted lines. The magnitude of this
Jl ~ ~)
368 ( ~- Xéf
""l- ---- - - /
1
, ~ ~
/
~
Subfrans1enf 036B(ji..~- X.J
__ -·Tirans1enr
Al"
componenf = .u . .L E )\
11 .~ •• ., ,
;
•
:
~~~~ ~
·r··
¡' 1
~~~-·~tT'fl-~r -~
- -i:i
co~m;'Po:n~en
~ f=.::d~I~r--..--.....
tA '"" - - +¡_ --- -- - TJ -- --- -· -- -- ·- -/F~. ,..,/,/
: r=XDI r- J.. - ~- - _;;,; - - -- - - - r-!::..-~~-r-
:
1
: 1\ !É. S h
' l . Xa= ync ronO/J.S
~ t
.
¡ 1.¡.. componenf
1 1
1-- - -- - - - -· o.-~ - - ..__

___V_
-~-
~
Fro. 39.-Curvc of symmetrical (alternating-currcnt) component oí short-circuit
curren t frorn no loac~.
quantity is determined by the expression E, where E is the

Xd
synchronous internal voltagc which in this case is equal to the
air-gap voltage corresponding to the no-load excitation, and xd* is
the constant of the machine known as the synchronous reactance.
This reactance is the value that is commonly used for all problems
involving steady-state calculations of machine performance and
includes armature-leakage reactance and the reactance equivalent
to armature reaction. It is usually obtained from the saturation
curve and the three-phase short-circuit curve (see Fig. 40) and
is defined as the ratio of the field current at rated armature current
on sustained symmetrical short-circuit to the field current at normal
* The notation used is based on Blondel's two-reaction theory of syn-
chronous machines in which Xd and Xq represent reactances in direct and
quadrature axes, respectively . For most problems the quadrature-axis
reactances may be ignored and thc direct-axis qua ntities only are required.
For a discussion of reactances refcrence may be rnade to A.I.E.E. papers by
R. H . Park and B. L . Robertson, <41 > L . A. Kilgorc, (72 > and S. H. Wright. <73 >
'
open-circuit voltage on the air-gap line (that is, the extended straight
part of the magnetization curve).* The definition gives the result
in the per unit t quantity which from Fig. 40 may be seen to be
3,200
í
~
..,.........-
r
3.000 .'
2J300 / c1.lvy ~
2,600 &' ~v<~,¡p E
Cl)
0...
~~ e}' <t:
·o¡ ..~;
2,400 x.~ v' "/ S:
lLl'
,) -,c..; - - - - r- Normal vo!fage +S:
z:zoo ~,,
~ '/ ~
~
r ,0f(/
u
~
1..
;:¡
~2.,000
~
~~y Cl)
11)
] ~~ V ~
.I:.
o.
·~j'
1,800
(S
S: 1,600
"t/
r
.¡e V ~
::l
+¡:¡
E
L.
1,400
j
.\;\
·o<' ~ E
1..
~
1,200
1 {(}
e;IJ~v olp
i}
Norm01! curren-f
~
!5C
e.\~1'
:l· !
1,000 1 . íc,
e) el' -
11. r-IJ<P
i·/
e}' e.e
-~'(\
800 100
(~'(<(
v~~<7
600
400 50
1V 1
200
y )
' o
o 2 4 6 8 10 12 14 !6 18 2D 2Z 2.4 26 30
Fielc:t Current in Am~res
Fw. 40.- Saturftion and synchronous impedance.
H unity. This unit) may be easily converted to per cent
=
values or to actual ohms at any voltage base. If E is given as a
r.m.s. value, 1 will also be expressed as a r.m.s. value.
* Excerpt from R eport of Subject Committee on Definitions of T erms Used
in Power System Studies. <78 )
t The per unit quantities are similar to per cent quantities except that
100 per cent is equal to one per unit; rated armature current is 100 per cent
armature current or one per unit armature current. This general system
has most of the advantages of the per cent system in comparing performance
of machines or lines of different rating plus the advantage that it is unn~·<)es
sary to carry through awkward figures that result from multiplying an imped-
ance by a current expressed in per cent to obtain a voltage· drop in per cent.
Thus 100 per cent reactance times 100 per cent current is equal to 10,000
per cent voltage which, of course, must be corrected to 100 per cent voltage.
b. Transient Current. If the envelope of the current wave

be projected back to zero time neglecting the firstfew cycles having a
very rapid decrement, the value thus obtained for zero time is
caBed the transient current 1'. This value is determined by
div:iding the transient interna! voltage E', in this case the open-
circuit voltage before the short-circuit, by the transient reactance
xi. The portion of the current to neglect in obtaining the
envelope may be made more evident by plotting the envelope of
~
1 1 1 1 1 1
,--Envelope or currenf f"rom Ftg..39 1
-
SS S! sfait ecl wlve
A ~
¡......_
""'
--
1
.-.., ¡....,_
~
\\
\
r--
r--
-- r--
<-T~ J\ t
1\! IA:Jrlve orA
'\: 1
1\ '
1\Y
~ 1\
~
...... \
-~
-~
-~
\
~
~ \
C;:j
~
1
1 '1\
T1me ~
Fw. 41.- Dctermination of T l'.
thc excess of current over the sustained value on semilogarithmic

paper, as shown in Fig. 41.
At the instant of sudden short-circuit the demagnetizing effect
of the armature current tends to decrease the flux linkages with
the field winding. However, the flux linkages with any inductive
circuit cannot be changed instantly but tend to remain constant.
This results in an induced current in the field winding of such
magnitude as to just annul the demagnetizing effect of the
armature current. If the leakage flux associated with the field
winding were zero, the air-gap flux would remain constant and
the initial armature current would be determined by the arma-
ture-leakage reactance alone, and the transient reactance would
be equal to the leakage reactance. Actually, however, the field
winding does possess a certain amount of leakage, and with the
increase of field curren t the leakage flux also increases. In
order to maintain constant flux linkages with the field winding
the air-gap flux must decrease. Since the armature current is
equal to the air-gap voltage divided by the armature-leakage
reactance, the initial armature current will be smaller than if
the field leakage were zero. This effect may be lncluded by
íncreasing the transient reactance by an amount proportional
to the field leakage. The transient reactance thus includes the
effect of both armature and field leakage::;. *
c. Subtransient Current. Thcrc remains yet to consider the
subtransient component, the value in Fig. 39 between the actual
current and the transient component. The total initial current
is determined by dividing the subtransient interna! voltage E",
which in this case is equal to the open-circuit volt-age, by the
subtransient reactance xi'. The increment of armature current
over that of the sum of the transient ·component and the sustained
value, namely, the subtransient component, is dueto the induced
currents in the damper windings or any similar circuit. If only
the phenomenon associated with the field winding were consid-
ere<;!, the air-gap flux would decrease upon the application of a
short-circuit, but actually the close proximity of the damper
winding i:rnbedded in the surface of the pole piece prevents the
air-gap flux from changing instantaneously. In the damper
winding currents are induced whic~ tend to maintain the air-
gap flux and thus increase the initial value of armature short-
circuit current. For a perfect damper winding, in the sense
that it does not possess any leakage flux, the transient reactance
of the machine is therefore equal to the armature-leakage react-
ance. Actually, however, damper windings possess sorne leak-
age so that the subtransient reactance is somewhat greater than
the armature-leakage reactance.
* A more complete discu<;;sion of thc physical significance of transient

reactance in terms of thc armaturc and field leakages is given in a paper
by C. F. Wagner, Effect of Armature Resistance upon Hunting, Tran.'l.
A.l.E.E., vol. 49, p. 1011, July, 1930.
Subtransient and transient reactances can be determined from
short-circuit tests. An alternative method for determining
the subtransient reactance is to apply voltage across any two
terminals (excluding the neutral) with the rotor at rest and short-
circuited on itself through an ammeter. One half the voltage
req uired to circulate ra ted curren t is eq ual to xi 1 in per unit
values. For this test the rotor must be in the position of maxi-
mum induced field current (the direct axis). If the rotor is in
the position of mínimum induced field current the quadrature
subtransient reactance x/ 1 is obtained.
41. Time Constants.

The initial value of the transient component of armature
current is produced by the induced cnrrcnt in the field winding.
The field winding and the coupled armature circuits act sub-
stantially as a simple resistance and inductance. The induced
portion of the field current and conscquently the transient
component of armature current therefore decay in an exponential
manner. Any exponential curve may be expressed in the from
t
le - T in which 1 is the magnitude for t = O and T is the time
constant. These curves have the property of decaying in any
interval of time t to ~E or 0.368 of its value at the beginning of the
interval. The time Ti in Fig. 39 required for the transient
component to decay to 0.368 of its initial value is thus the time
constant of the transient component and is called the transient
short-circuit time constant.
The transient open-circuit time constant of a synchronous
machine is the time constant of the exponential curve of armature
voltage obtained when the exciting voltage is suddenly removed
from a machine operatirig at no load, the field circuit of course
being maintained. The transient short-circuit time constant
Ti may be obtained in Lerms of the transient open-circuit time
constant T 1do by the expression <34 >
1
T d1 = -Xd T'dO (119)
Xd
The subtransient current is maintained by currents flowing

in the damper windings or any other similarly located circuit.
·This circuit is similar in nature to the :field circuit ex0ept for the
80 SYMMETRICAL COMPONEN'l'S
relative values of inductance and resistance. Because of the

relatively high value of resistance the time constant is usually
very small, which accounts for, tbe rapíd decay. In general,
t herefore, it will be found that the subtransient short-circuit
time constant Ti' is very small and is of t he order of 0.1 or 0.2
sec. or smaller. Figure 39 shows the significance of the time
constant Ti' and Fig. 41 indicates how it may be obtained from
the alternating-current component of short-circuit current.
42. Direct-current Components.

Considering only the alternating-current components of cur-
rent in the individual phases of the armature requires t hat the
armature currents in at least two, and possibly three, of the
phases increase instantly fro~ zero at times of three-phase short-
circuit. Because of the presence of the inductance in the indi-
vidual phases of the armat"'ure the currents cannot change
instantly from zero to a finite value. A direct-current or
equalizing current therefore flows in each phase of such magnitude
at t = O to bring the current equal to zero. Thereafter t he indi-
vidual direct-current components decay in an exponential manner
corresponding to the direct-current time constant 'Pawhich is equal
for all three phases. The valu e of this time constant is dependent
u pon the ratio of inductance to resistance in the armature circuit.
As will be shown later the negative-sequence reactance of the
synchronous machine x2 is a sort of ave:rage of t he armature
subtransient reacta¡nces so that the direct-cu!:rent time constant
may be written <4 1) "
T - ~ (120)
a - 21rfra
where ra is the armature resistance, R.nd X2 and ra must be expressed

in the same units, eitber per unit or ohms per phase.
43. R.M.S. Total Current.

The r.m.s. total current at any instant is equal to the square
root of the sum of the squares of the alternating-cm:rent and
direct-current components; the mínimum currents t hus occur
on that phase in which the direct-current component is equal
to zero and the maximum currents on t h at phase in which
the direct-current component is a maximum. Since the maxi-
CONSTANTS OF SYNCHRONOUS MACHJNES 81
Jnum value that the dircct-current component can attain 1s
V2I", the maximum value of thc .r;..m.s. total current is
'\)¡" 2 + (V2J") 2
= V3I".
44. Short-circuit of Loaded Generator.
The foregoing assurned that the generator under consideration
had been operating at no load just previous to the short-circuit.
Such is not usually the case. It is necessary therefore to be
able to calculate short-circuits for the more usual case of the
loaded machine. <73 > This involves principally the determinatioo
~~J
Et
)(." 1
á
/
/
/
/
', ' ,-90°-- /
//
'<' ~
' V/ /
/
Fw. 42.-Construction to determine E' and E".
of the synchronous, transient, and subtransient interna! voltages.

The latter two may be obtained by merely adding vectorially
the transient and• subtransient reactance drops, respectively,
to the terminal voltage E t· An example of this construction *
is shown in Fig. 42. In determining the sustained value of
short-circuit current the interna! voltage used is that fictitious
value corresponding to the point on the air-gap line for the
particular excitation. This is pcrmissible because armature
reaction decreases the air-gap flux below the saturation point.
Saturation is an important factor, however, in determining the
excitation for the particular load carried by the machine before
the short-circuit. A convenient method for obtaining this
excitation is the A.I.E.E. method, which utilizes the no-load and
the zero power-factor saturation curves shown in Fig. 43. To
obtain the excitation at any other power factor for rated current
the distance. ab between the no-load saturation curve and the zero
* This construction is not strictly cov-ect for salient pole machines but is
sufficiently accurate for practical purposes.
power-factor curve for the excitation Oc is laid off along the line
bd (Fig. 43) which makes an angle equal to the power-factor
angle with the X-axis. A line equal to the distance ac (the
no-load voltage corresponding to the excitation Oc) is then scribed
from the point a intercepting the Y-axis. The horizontal line
..
,."
Power
mdor
angle
FIG. 43. -D et ermination of excitation for loa d p ower factors other than zero by
the A.I.E.E. method.
ce is then the terminal voltage for the particular excitation.

Following this procedure another excitation is chosen and the
construction repeated from which the dotted line ef, shown in
Fig. 43, is obtained. The intersection of this line with the
normal voltage gives the excitation for the particular power
factor at rated load. If the machine is not operating at rated
load, the zero power-factor curve corresponding to the particular
current should be used. The synchronous interna! voltage
expressed in per unit is the ratio of the excitation under considera-
tion to the excitation required to generate rated voltage at no
load, measured on the air-gap line. .
The initial values of I' s.nd I" are then equal, respectively, to
E' E"
- ., and - 11 • The maximum value of the direct-current com-
Xd Xa
ponent Iac is obtained by taking v2 times the negative of the

' 1
/·/nrerna/ volf'f9e
--80% P.F.cmgle /
--Loaol currenf bef=ore shorf· clf-r:uif /
' 1
' 1
' ......., 11
' 1
',._. , ..g()o-, _~1
--3.33 ',¿/ /
....... /
' 1
',/
r.r
XJ
F10. 44.-Constructio n to obtain the maximum value of 1 de for a three-phase
short-circuit on a loaded machine. ·
vector difference between J" and the load current before the short-
circuit. This construction is shown in Fig. 44.
46. Externa! Reactance.

Machine short-circuits with externa! reactances may be
treated after determining the interna! voltages in all respects
as though the machine reactances were increased corresponding
amounts; thus calling X e the externa! reactance, Xd + X e, xl + Xe,
and xa" + Xe should be used as the synchronous, transient, and
subtransient reacta nces, respectively. The time constants vary
in a similar mai).ner, th us
Ti - xi ++ XeT'
Xa X~
dO· (121)
,,
46. Example of Decrement Calculations.

To illustrate the foregoing -discussion, let it be desired to
calculate t he three-phase short-circuit of a large typical turbine
generator whose characteristics are given in Fig. 45. An externa!
or feeder reactance of 0.145 will be assumed. This odd value
,..
.3
/
1.2.
1 V
--1
Air-gap /ine - /
L -No load sodurafion curve_
-~
'\
1.1
f.O
~
1\..--- V/ 80 Percenf PF- .,.,
~
o9 1V /
/
/
V
~-
~
~o.8 1/ /
/
1./-zeroP.F.
~~
g )
~o.7
/
~ 1
t:
-~
¡!!!o.6
1
o.5 ~ 1 1
o.4 \ 1 1
o.3 X V
o.2
;~
1 - \
:1 1 ~ V
o 1 ~~ 1 1.4 1.6
~
O.B t.O 1.2 I.B 2.0 2.2 2:4 2.6
F"ielcl Current
80 Perc x:;, = o.t45 T~= 10

x::,. =o.24
1
JCc¡¡ = l. tO TO.= 0.034
x2 '"' o.r4s
r Z ..'> 0..30
T.PI = 0. 2.6
t (l '
Fw. 45.-Ch aracteristics of turbine generator.
)~
is chosen as i t is eq u al to the nega ti ve-seq uence reactance; the
results of this calculation can therefore be used la ter in illustrating
unbalanced short-circuit calculations. The machine is assumed
to be fully loaded at 80 per cent power factor before t he short- 'l
circuit and t he short-circuit to occur on an unloaded feeder.
Interna! Voltages. Thc construction of the 80 per cent power-
factor regulation curve is shown in Fig. 45. From t his curve
it will be seen that t he excitation for rated voltage is 2.19 in
'
per unit terms using the excitation for normal voltage for the air- ·
gap line as unity. Therefore, E, = 2.19. Using the values
xi = 0.24 and xi' = 0.145, the corresponding interna! voltages
are E/ = 1.165 andE/' = 1.095.
Initial Altemating-current Components. For the short-circuit
condition, the load will be represented by a reactance which takes
the same reactive kilovolt-amperes as the load before the fault.
This reactance is j1.67. The externa! or feeder reactance is by
assumption j0.145. During the fault condition these rcactances
will be in parallel, so that the net externa! reactance is
= (0.145)(1.67) = o 134
Xe 0.145 + 1.67 ·
';['he initial machine currents then become
!" = E/' _1.09.5 3 93
X/
1
+ Xc
0.145 + 0.134 = .
E/ 1.165
!' = - = 3.12
x/ + Xe 0.24 + 0.134
I = Ei _ 2.19 = 1.
77
Xd + Xe 1.1 + 0.134
A portion of the above currents is by-passed through the load
reactance, so that the initial feeder currents are
!" = 6:~!:3.93 = (0.924) (3.93) = 3.63

!' = (0.924)(3.12) = 2.88
I = (0.924) (1. 77) = 1.64.
It will be observed from the above that the currents taken by
the feeder and the machine vary only about 8 per cent.
Direct-current Components. The maximum direct-current
component of the machine current is obtained by taking t he
vector difference between !" and the load current. Figure 44
has been laid out to scale for the particular case und~r considera-
tion, from which the difference is 3.33. The direet-current com-
ponent is then
I ac = v2
X 3.33 = 4.72
For the feeder, since it had becn unloaded,
Idc = v2 X 3.63
= 5.1
Total I nitial Current. For the machine
!total = v(4.72) 2 + (3.93) 2
= 6.15
· For the feeder

1 totaz = v3 X 3.63 - 6.29
10 r--
9
8
7
6
5
4
1\
\
3
r\
\
2
i\
.. \ ~
·¡:
t
:::1
'-
&.
S:
1
0.9
~· ~1'
1\..
K i\
\.
"\. \
t 0.8
~
d 0.7
~'\.- -- f-~ 2TDI --- R\
0.6
~ \
0.5
'\ 1\
0.4
0.3
1
~ :.11" "
¡'\ ""
\
1\.
\
'1
\ 0.1
~-- -- -- r---
\
\
P\
"" !'\. \
---
"
11~
f\
2Td
1-<--f-- - t - -
~
2Td 1-- - 1---
O.f ~
O 1 2 3 4 5 Scale for ~I'
O 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.1 Sea le t'or A 1"
O 0.1 O.z 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 SCQIIe for lk
Time in seconds
FIG. 46.-Transient components of three-phase short-circuit curr ent of turbina
generator with a fault reactance of 0.145 per unit.
Instantaneous Values. Tbe instantaneous values of the feeder

currents only will be determined. Tbe initial subtransient com-
ponent ~I" (see Fig. 39) for the feeder is
~I" = I" - I' = 3.63 - 2.88 = 0.75
~I' = I' - I = 2.88 - 1.64 = 1.24
The t~ansient time constant is, from equation (121)

T , = 0.24 + 0.1347 =
21
á 1.1 + 0.134 .
and from Fig. 45,
0.034 Tá'' =
It will be assumed that the ratio of the Iiegative-sequence react-
ance of the externa! impedance to its resistance is the same as
6.0 1---
5.5 t - -
~
5.0
4 .5
5
E4.o
o
z \ ..,..f----1 Toral
~ 3.5
E
¡.:: 3.0
i\ .~
'-. '
.E .
t
~
2.5
5 2.0
u
-:r- ~ ......
\
' 'lot-e - ¡..._
'
1.5
-- --·
10
0.5
0.0
o 02 0.4 1.0 1.2. 1.4 1.6 t.a 2.0 2.2 2.4 2.6 2.8 .J.O.
0.6 0.8
Time in Seconás
Fw. 47.-Curves illustra ting example of decrement current calculation.
that of the generator so that the direct-current time constant

remains the same as that of the generator, namely,
Ta = 0.26
The instantaneous values of the exponentials may be obtained
with the assistance of the single logarithmic plots of Fig. 46.
These may be drawn as follows. Plot the initial value for t = O,
and (0.368) 2 or 0.135 of the initial value for t equal to two times
the time constant. The straight line connecting these points
gives the value of the particular component as a function of
time. Adding !11', !11", and the sustained value, 1.64, gives the
total alternating-current component; and combining the alternat-
ing-current and direct-current components by taking the squarb

root of the sum of the squares gives the total current. Both of
these are plotted on Fig. 47. Thus, at 0.05 sec., the currcnts
in the feeder are
I = 1.64 ,.
!::..]" = 0.17
b..l' = 1.21
lac = 3.02
Jdc = 4.20
R.m.s. total current = v3-= .o:-::2:-::-2 -+-:--4::-.2-=-o=2 = 5.18
47. Negative-sequence Reactance.

The negative-sequence impedance of a machine is the imped-
ance offered to the flow of negativc-sequcncc current, i.e.,
it is the negative-sequence voltage across the machine when
one unit of negative-sequence current :flows through the machine.
As stated previously, when negative-sequence voltage is applied
to the armature, a field is set up which rotates with synchronous
velocity in 3 direction opposite to that of the rotor. This field
sets up currents of double system frequency in the rotor. The
subtransient reactances may be measured by blocking the rotor,
with the field winding short-circuited, and applying a single-
phase alternat}Ilg voltage across two terminals of the armature.
The reactance per phase measured in this manner varíes with the
position of the rotor. I,f the machine is a salient pole machine
without damper windings tbe variation is very great, for when the
axis of the rotor coincides with the axis of the pulsating field, the
field winding constitutes a short-circuited seconda.ry producing
a low impedance, but when the two axes are in quadrature, the
impedance is ·merely that determined by the exciting current
and is quite high. The upper curve in Fig. 48 shows these
subtransient reactances plotted as a function of the angular
position of the rotor. For tbe negative-sequence measurement
a similar pbenomenon is involved except that the rotor is rotating
with double frequency witb relation to the field set up by tbe
applied voltage and is taking successivély all the possible posi-
tions used in determining tbe subtransient reactance. One
would expect, tberefore, tbat the imaginary component of z2,
namely, the negative-sequence reactance x2, be sorne kind of a
mean between the maximum value of subtransient reactance xa''
and the mínimum value xi'. Park and Robertson< 41> give for
this value, when the circuit has a large value of reactance in
series, i.e., when the current wave form is essentially sinusoidal,
Xz = Hxi' + xq'') (122)
Distortion ansmg from saliency effects results in somewhat
diff"erent values depending upon whether sinusoidal voltage is
applied or sinusoidal current is · circulated. The dotted line
1.1
1.0 ~
1 1
t- Qvadmtvre Axis X~
1 1 H Direcf 1 X O.
1Axis
V
0.9
"' Blockeol rofor me~hool V
0 .8 1\ ~o darnJ)ers /
,___ ,____ _\~- _¡_ -,__1 --
0.7
1\ l..-- -
Negarive sequence
¡_ _
1
-
0.6
\ V
r 0.5
0 .4
1\.
'\ ..__ ....-

V
)
0.3
0.2
Copper chrnpers
0.1
o.o
o 15 30 45 60 75 90 105 120 135 150 165 180
Angular Position of R otor in Degrees
FIG. 48.- Relation between subtransicnt and negative-sequence reactance.
midw'l,y between the maximum and mínimum values gives the

test result for the negative-sequence reactance of the same
salient pole machine without damper windings.
The addition of low-resistance damper windings has different
effects u pon the reactances in the two axes. In the direct axis the
close proximity of the damper windings to the air gap and hence
1 to the armature winding results in a smaller leakage and thus
'
l
:·
¡
xi' is reduced somewhat. A machine without damper windings
has no windings on the field structure in the quadrature axis
and thus xq'' is determined by the exciting reactance in this
axis. A damper winding would, however, constitute a sbort-
circuited turn and reduce the reactance from one of an exciting
reactance to one approaching a leakage reactance. Actually,
the reactances in tbe two axes for a machine with damper wind-
ings are very nearly equal, as is shown by the curve for this case
.. in Fig. 48. It follows, therefore, that the negative-sequence
reactance is also very nearly equal to these values.
Method of Test. In addition to the met'-lod outlined whereby
X2 is defined as the arithmetic mean of xa" and xq'', x 2 may be
determined directly either by applying negative-sequence voltage
or by the method<34>·<73> shown in Fig. 49. Witb the machine
driven at rated speed, and with a single-phase short-circuit
e
-1~
a2 a
l~ =o lb = l le = -l la z = t(o + a2l - al) =
-
3
l
a2 a ·
EA = o EB = E Ec = -E EAz = t(o +
a2E - aE) =
-
3
E
E _ . EA2 _ .a 2 - aE E az jE
Zz=-=--
a2 - J -\13 - J 3VS la2 VSl
If rp = cos-1 Jr wher~ P = wattmeter reading,
then, Z2 = _E <.
J';) - sm rp + J. cos </>) = T2 +.JX2
v3l
Fw. 49.-Determination of the negative-sequence impedance of symmetrically
wound machines.
applied between two of its terminals (neutral excluded), the

sustained armature current and t he voltage between the terminal
of the free phase and either of the short-circuited phases are
measured. The reading of a single-phase wattmeter with its
current coil in the sbort-circuited phases and with the above
mentioned voltage across its potential coil is also recorded. The
negative-sequence impedance equals the ratio of the voltage to
the current so measured, divided by the square root of three. •
The per unit negative-sequence reactance equals this value of
impedance multiplied by tbe ratio of power to the product of
voltage and current.
,
48. N egative-sequence Resistance.
The power associated with the negative-sequence current may
be expressed as a resistance times the square of the current.
This resistance is designated the negative-sequence resistance.
For a machine with no dampers the only source of loss is in the
armature and field resistances, eddy currents, and iron loss. The
copper loss in the armature and field is very small as is also
the iron and eddy loss in the armature, but the iron and eddy loss
in the rotor may reach quite high values. Copper damper wind-
ings provide a lower impedance path for the eddy currents and
hinder the penetration of flux into the pole structure. The
relatively low resistance of this path results in a smaller negative-
·-·
~r,.
FIG. 50.-Equivalent circuit of induction motor.

·sequence resistance. For higher resistance dampers, the nega-
tive-sequence resistance increases to a point beyond which the
larger resistance diminishes the current in the rotor circuits
su:fficiently to decrease the loss.
Induction Motor Diagram. The nature of the negative-
sequence resistance is best visualized by analyzing the phe-
nomena occurring in induction motors. Figure 50 shows the
usual enuivalent Circuit* of an induction motor in whicb
T 8 = stator resistance.
Xs = stator-leakage reactance at rated frequency.

Tr = rotor resistance.
Xr = rotor-leakage reactance at rated frequency.
Zm = shunt impedance to include the effect of magnetizing
current and no-load losses.
Es = applied voltage.
ls = stator current.
Ir = rotor current.
s = slip.
* For more detailed description of this circuit see any standard text book
such as Ralph R. Lawrence, "Principies of Alternating Current Machinery,"
McGraw-Hill Book Company, Inc., New York.
All of the foregoing quantities are assumed to be given in per unit

quantities.
The justification for this diagram may be seen briefly as
follows: The air-gap flux dueto the currents l s and Ir induce the
voltage E 0 in the stator and sEu in the rotor. In the rotor the
impedance drop is
rrlr + jsxJr
since the reactance varies with the frequency of the currents
in the rotor. The rotor current is therefore determined by the
equation
or
(123)
It follows from this equation that the rotor circuit can be com-
pletely represented by placing a circuit of impedance 0:.
S
+ jxr
across the voltage E 0 • The total power absorbed by ~ must be
1 S
the sum of the rotor losses and the useful shaft power, so that,
1
resolving 0:. into the resistances rr and - srr, the power absorbed
S S
by rr represents the rotor copper loss and the power absorbed by
1- S
- - r r represents the useful shaft power.
S
N eglecting rs and the real part of Zm, the only real power is that
concerned in the rotor circuit. N ow assume that the induction
motor is loaded by means of a direct-current generator connected
to the shaft. At small slips the electrical input into tne stator
is equal to the copper loss, i.e., the I;rr of the rotor plus the shaft
load. With the rotor locked the shaft load is zero, and the total
electrical inpqt into the stator is equal to the rotor copper loss.
At 200 per ccnt slip, i .e., with the rotor rotating at synchronous
speed in the reverse direction, the copper loss is I;rr, the electrical
input into the stator is Ir~rr, and the shaft load ; rrlr 2 o"r
1 2
- I~r. A negative shaft load signifies that the direct-current

machine instead of functiontng as a generator is now a motor.
Physically that is just what would be expected, for as the slip
increases from zero the shaft power increases to a maximum
and then decreases to zero for 100 per cent slip. A further
increase in slip necessitates mntion in the opposite direction,
which requires a driving torque. It will be observed, t herefore,
that at 200 per
r,.
cent slip the elec-
trical input into
the stator is equal
to the mechanical
input through the
shaft; half of the (a)
copper lo ss is
supplied from the Jx.a Jxr
-~~~J-·"_m_'_oo_o_o_o ---'~ ~·
stator and half
through the shaft. o-:--r-oo_o_o_o_'i _,
This is the condi-
tion o bt a ining (b)
with respect to
t h e n e g a t i v e-
sequence in which
the rotor is rotat-
=
1
00000'
ing at a slip
of 200 per cent
(e)
relative to the
F10. 51.-Development o f n egative-sequence resistance
synchronously and reactance from equivalen t circuit of induc tion motor.
rota ting nega- (a) Negative-sequence diagram for induction motor;
(b) neglecting armature and no load l osses; (e) simplified
tive-sequence network-negative-sequence resistance a nd r eactance.
fi e ld in the
stator. * Half of the machine loss associated with the negative-
sequence current is supplied from the stator and half by shaft
torque through the rotor.
The factor of fundamental importance is the power supplied
by the stator and through the shaft, which can always be deter-
mined by solving t he equivalent circuit involving t he stator
and rotor constants and the magnetizing-current constants. A
more convenient device, since s is constant and equal to 2 for t he
negative-sequence, is to reduce the equivalent network to a
simple series impedance as shown in Fig. 51(c). The components
of this impedance will be called the negative-sequence resistance
* For a more complete discussion of induction-motor performance with
the application of negative-sequence voltage refer to Chap. XVII.
r2 and the negative-sequence reactance xz. The values of these

constants are also given in Fig. 51. The current fiowing through
the negative-sequence impedance is the current fiowing through
the stator of the machine, and the power loss in rz is equal to the
loss supplied from the stator of the machine and the equalloss
supplied through the shaft.
The total electrical effect of the negative-sequence resistance
is obtained by inserting the negative-sequence resistance in
the negative-sequence network and solving the network in
the usual manner. All three of the sequence currents are thus
affected by a change in the negative-sequence resistance. The
total electrical output of a generator is equal to the total terminal
power output plus the losses in the machine. However, since
the negative- and zero-sequence power outputs are merely the
negative of their losses, the. contribution to the electrical output
by the negative- and zero-sequences is zero. The total electrical
output is therefore that due to the positivc-sequence and to
include the positive-sequence armature-resistance loss it is only
necessary to use the positive-sequence interna! voltage in the
calculations. Or viewed differently, since there are no interna!
generated voltages of the negative- or zero-sequence, the corre-
sponding internal power must be zero. In addition to this
electrical output which produces a torque tending to decelerate
the rotor, there also exists the negative-sequence shaft power
supplied through the rotor. It was shown that the value of this
power tending to decelerate the rotor is numerically equal to
the negative-sequence power supplied to the stator which is equal
to the loss absorbed by the negative-sequence resistance. There-
fore, tbe total decelerating power is equal to the positive-sequence
power output plus the loss in the negative-sequence resj'3tance.
The assumption was made that the stator resistance and
the losses in the magnetizing branch were neglected. For greater
refinements, the stator resistance and the losses in the magnetiz-
ing branch can be taken into consideration by substituting
them in the equivalent circuit and reducing that circuit to a
simple series resistance and reactance, wherein the resistance
becomes the negative-sequence resistance and the reactance
the negative-sequence reactance. The ratio of the negative-
sequence shaft power to the negative-sequence stator power
is t.i.len equal to the ratio of the power loss in~ for unit negative--
sequence current in the stator to r2. This ratio can be obtained
very easily by test by measuring the shaft torque and the nega-
tive-sequence input when negative-sequence voltages only are
applied to the stator.
While this analysis has premised induction-motor construction,
tbe conclusions may also be applied to synchronous machines.
E¡
lz
Fxo. 52.-Negative-sequence resistance of a synchronous machine.
3(P. - PF+W)
T2 = 2J2
in which
P 8 = shaft input.
PF+W = friction and windage losa.
Method of Test. While r2 and x2 can be determined by apply-

ing negative-sequence voltage from another source of supply
to the armature, the following method <58> has the advantage that
the machine supplies its own negative-sequence voltage. Two
terminals of the machine under test are short-circuited and the
machine driven at rated frequency by means of a direct-current
motor. The equivalent circuit and vector diagram for this
connection are shown in Fig. 52. The positive-sequence power
96 SYMMETRI CAL COMPONENTS
at the terminals is equal to the product QÍ E1 and 11 and the

cosine of the angle cf>. It will be observed that this power is
positive. However, the negative-sequence power output is
equal to the product of E2, 12, ::md the cosine of the angle between
E2 and 12, and since 12 = -11, andE1 = E2, the negative-sequence
power output is the negative of the positive-sequence power out-
put, which, of course, must follow since the output of the machine
is zero. A negative output is equivalent to a posit ive input.
This input is equal to r21 2 2 • Therefore, the positive-sequence
terminal output is r21 2 2 , and adding to this the copper loss due to
I 1, gives the total shaft power due to the positive-sequence as
r21 22 + rJ1 2. Now from Fig. 51(a), if Zm be neglected, the
negative-sequence input is equal to
( rr +r8 - ~)l 2 2 or (~ + r )122,

8
from which it follows that

(124)
As shown previously the negative-sequence shaft power is equal to

~1 2 2 , which on substituting ~ from (124) reduces to (r2- rs)12 2•
But since rs = r1, the expression for the negative-sequence shaft
power may also be written (r2 - r1)1 22 . Therefore the total
shaft input into the alternating-current machine is equal to
r 2122 + rJ1 2 + (r2- r1)l2 2 · and, since 11 = 12, reduces to
2r212 2·
lncluding the effect of friction and windage, P <F+w), and
calling P s the total input into the alternating-current m achine
from the driving tool,
Ps- P (F+WI
r2 = '- (125)
212 2
~
and, since 12 =
3 where 1 is the actual measured phase current,
(126)
49. Zero-sequence Impedance.

The zero-sequence impedance is t he impedance offered to the
flow of unit zero-sequence current, i .e., the voltage drop across
\
any one phase (star-connected) for unit current in each of the
phases. The machine must, of course,
be star-connected for otherwise the
term zero-sequence impedn.nce has no
significance as no zero-sequence cur-
rent can flow.
The zero-sequence impedance of
synchronous machines is quite variable
and depends largely upon pitch and
Rotor at synchronous spccd
breadth factors. In general, however, (or lockccl)
the values are much smaller than the Zcro-sequence impedance,
P os i ti ve- and negative-sequences. zo = 3!
E
The nature of the im pcdn.nce may be Fw. 53 .-Connection for
SUggested by COnSidering thu.t, if the IDC'asuring zero-sequence im-
armature windings wcrc infinitely pcdance.
distributed so that each phase produced a sinusoidal distribution
0.60 of the m.m.f., then the
0.55
- J superposition of the
three phases with equal
0.50
ébuafion --- _
Xd'= .02+1.4X0 '' ' ( instantaneous currents
OA-5
/_ cancel en.ch other and
..'"CS produce zero field and
X0:40 1 consequently zero
tl V reactance except for slot
. ~0.35
u
es V a·n d en d-e o n n e e ti o n
fluxes. The departure
~0.30
a:
'1: Q25
Q)
1/
from this ideal condition
introduced by chording
VI
~0.20 J and the bread th of the
1-
0.15
1 phase belt determines
the zero-sequence
0.10 / impedances.
/
Q05
The zero-seq u en ce
resistance is equal to,
V
• oo 0.05 0.10 0.15 020 0.25 0.30 0.35 0.40 or somewhat larger than,
Subtrcmsient Reactance Xcf'
the positive-s e que n e e
FIG. 54. -Relation betwcen x/ and xd'' (satu- resistance. In general,
rated values) for threc-phase synchronous however, it is neglected
machines in per unit quantities.
in most calculations.
Method of Test. Thc most convenient method for test is to
connect the three phases together, as shown in Fig. 53, with the
13
12
/
(/)
o 11
vv<v/
z
o /}'
----
~ 10
rv"'
(/)
/
V
z .9 ¡......-
.~
z 8
V Z'A
,6 <t>cp~
~ c;3:3;_..:;..---~
(/)
z 7 1 ,..,......,
eo_
8 V
w 6
~
... V ...
~ 5
5
li 4
V
i3
zw 3
1 '
i
Q.
o 2
o
l
-~ 1
o ~~
o 10 20 30 40 50 60 70 80 90 100 110 120 130 140 ISOTGW
RATING
Fw. 55.-0pen-circuit time constante of turbine gcnerators.
13
(/)
12
/{JI
•' 0>v ~V
ú1
-.JI
s:_f v'
, (}
lf/- .
.!:'
~ cY v'-'~
~...,'V
~
<v
0
<v
~
~, ~· ¡:;:;ro
~ 11 . '----....,-·ro o e,
ti'_L-t¿-~17:P:·, ~
~~ ~·
~·
8
w 10
Cl:l Q.· Q.•,' Q.·
q,·-<?-<?·/
<1) 01 o o, o o / c,'v
.;¡11 o
o,-¡
"j , /
'
"'V t>~ e<
~ 9 100
...z _,;, //V/ o~~
~ 6 p.O (,~
<1)
V j ,// / ~ ~~·~·eo
a 7
l'1Z
.,.:'
u
' v V,/ / /V __)-~ 160~~~
~ 6
¡:
V v_v~/ · /V / V
/~ :#'~
....
3
~
5
-~1V
/~
-- V : V / / V V v
v--
1 ~·
1 f¿-. -
:-j~~V
4
u
zw 31 / / /V ~~~---
o 2 /i ~ ~ ~
Q.
,,....
r::-V --
o
lf/J~f.-'
¡r
o
o 2 3 4 5 10 20 30 40 50 60~
RATINGt 1000
Fw. 56.-0pen-circuit time constants of alternating-currcnt gencrators and

motors.
,CONSTANTS OF SYNCHRONOUS MACHINES 99
field short-circuited on
itself. This connection ~
..... o C'l lC
.....
C"l C'l
ci <O o ci ci
insures equal distribu- +-
o 1 o 1
C':) t--
o
00
H
o
1 o 1
tion of current between o ..... o C"l o ~
o C"l
o o o o
the three phases and for ci o ci ci
this reason is preferable - - .---..
>,
......
to connecting the three 1'-<
~
o
t-- t--
Q,) ~
phases in parallel. The *~ H"'
¡:::1
'--' olC
ci
~
o
zero-sequence i m pe d- ~ ::.... ::::
.... o o
1 C"l
ci
1
ance is then equal to ~

H H t--
C':) .....
11 11 ci o
Zo = ~ n:s indicated in ~ t--
..... lC
C':)
00
C':)
z C"l ci C"l
ci lC ci
the figure. ¡¡:¡ '.,.
' ..... 1 C"l
1 ' .... C"l
1
C)
H ci ci H ci
< t-- ~
..... lC 00
.....
60. Typical Values of ::;:g o 00
o o ci o
Synchronous M achine o~ .---..
,_;
Constants. ~ .p
;::l
~ ~
en
The constants of § ' .... ce .....
'--'
tt:> lC
synchronous machines ~
m
H C"l
ci
lC
o
~
o o
lC
O)
..... C':) tt:> t--
vary over a wide range r:.1
o 1
C':)
1 C':)
1
C':)
1
.... o
o cío o
with the particular type ~ C"l
..... t--
of machine, rating, and ¿; --

o
C"l
ci ló C"l
ci
speed. As a result of <D

lC lC tt:> o
~ ~ ~ C"l
averaging the data on a

r.1
~ o..... ..... o ..... ..... C'i
.... ..... o..... ~ 1
~ H 1 1 1
large n u m ber of m a- ~ ..... lC
..... o ..... o ..... o
r.1 O) <O <O lC
chines the straight-line r-1
~
o ci ci ....-4
relation between xi and ~ ..e::

.p
.---..
en
xi' in Fig. 54 was found ...:l
<
.......
~
'--'
bl)
.s'"d
to hold remarkably well. ~
.S
o~
p.
Tbe relation between the ~ .p
~
~
1'-<
~ Q,)
time constan ts and the 1'-<
Q,) 0..
rating can also be repre- ~
:::1
¡:::1
Q,)
8~
bl)
en '"d
sented by curves when r:.1
Q,)
'"d o
...:l .S ¡:::1 S
segregated into 25- and ~ ..e::
Q
~
en
60-cycle turbine gen- ~
c3
:g
en
1'-<.---..
. o1'-<
~
o
.pbD
en .p
c3
erators and salient pole .p
o o .S
1'-<
Q,)
c3 S'"O ¡:::1
Q,)
generators and motors of 1'-<
Q,)
¡:::1
Q,).s
bl)
......
different speeds. These Q,)
bl)
-~
o 1'-< Q,)
Q,) ~
Q,)
~
o..Q,)
Q,) en
are shown in Figs. 55 .S
.Po.. ~
..0
¡:::1 8c3 Q,)
Q,)
'"d
and 56. Other con- 1'-<
Q,)
:.::l'"d
-+"
c3
¡:::1
stants classified as to ~ c3
OO. ¡:: 8 1
100 .OYMMETRICAL COMPONENTS
type of machine are given in Table III, the figure above the line
representing average values.
51. Unbalanced Faults.
It was shown in Chap. III tbat unbalanced short-circuits on
synchronous machines can be represented by equivalent circuits
involving the positive-, negative-, and zero-sequence impedances.
From these circuits it may be seen that the relations between
positive-, negative-, and zero-sequence currents are depend-
ent only upon the character of the fault and the negative- and
zero-sequence impedances. It follows therefore that the nega-
tive- and zero-sequence currents vary proportionately with the
positive-sequence current, and in determining the negative-
and zero-sequence components of current it is necessary to
obtain merely the positive-sequence time variation of current.
I t m ay further be seen from these diagrams that the effect
of the unbalance upon the positive-sequence current may be
completely simulated by placing an impcdance in series relation
with the machine; for a line-to-ground fault the impedance is
Z 2 + Z 0 , for a line-to-line fault Z2, and for a double line-to~
ground fault ZoZoZ2 + z • T he time variation of t he positive-
2
sequence current can then be determined in the usual manner

after which the negative- and zero-sequence currents are deter-
mined by applying the proportionality constants.
,, The foregoing applies only to the alternating-current compo-
l!:ent of current. Tbe direct-current component is obtained as
for the balanced short-circuits by taking the negative of the
difference between the initial instantaneous alternating-current
component a;nd the instantaneous current in the different phases
before the fault.
52. Decrement Curves.*
The commercial application of relays and circuit-breakers
demands a simpler means than that described for calculating
the short-circuit current at any time. By making certain
assumptions whicb re~resent typical operating conditions and
machine characteristics it is possible to compute a set of curves
to evaluate the short-circuit current which are useful over a wide
range of machine constants. These curves are called standard
* This section is based on data and curves prescnted in an A.I.E.E. paper
by W. C. Hahn and C. F. Wagner. <75 l
decrement curves. They are applicable not only to single
machines but to any number of parallel machines whose con-
stants are sufficiently similar. The assumptions upon which
they are based are :
l. The generators are assumed to be operating at rated
voltage and kilovolt-amperes at 80 per cent power factor imme-
diately preceding the short-circuit.
2. No automatic voltage regulator is used.
3. The actual systcm subjectcd to fault may be represented
by a single equivalent generator of the same total rating and an
externa! reactance.
4. The load is assumed to be located at the machine terminals,
and the machine reactance is taken as 15 per cent unless the total
reactance is lcss than 15 per cent. In this event all the reactance
is assumed to be in the machine.
5. The short-circuit is assumed to occur on an unloaded
feeder.
6. The short-circuit is assumed to occur at the point of the
voltage wave which corresponds to maximum possible instan-
taneous current.
7. All resistan ce in the circuit including the resistan ce of the
fault is neglected.
8. All the machine e.m.fs. are assumed to be in phase.
9. The machine reactances and time constants are taken as
representative of modern machines. Of particular importance
is the relation between transient and subtransient reactance of
machines (see Fig. 54).
Description of Decrement Curves. The decrement curves of
Figs. 57 and 58 give the r.m.s. total current expressed in terms of
times normal current with total connected synchronous capacity
in kilovolt-amperes as a base. In Fig. 57, the elapsed time is
the abscissa and reactance is the parameter. In Fig. 58, react-
ance is plotted as the abscissa with elapsed time from the begin-
ning of the fault as a parameter for the curves.
The reactance to be used with the curves for any kind of a
fault may be obtained either by means of analytical calculations
or by use of the calculating board. In order to choose the proper
decrement curve, it is essential that the reactance used to select
a curve in Fig. 57 ora line in Fig. 58 be expressed as a per cent
of the total connected synchronous capacity in kilovolt-amperes
rather than, as is common practice, an arbitrary value con-
veniently chosen to expedite system calculations.
102 SYMMETRICAL COMPONE NTS
If the system open-circuit time constant referred to the point

of fault is known to be other than 5.0 sec., the time scale of the
curves may be corrected to conform to it by means of the formula
given on t he curve sheets.
Three-phase Short-circuit. The reactance to be used to
select t he proper decrement curve for a three-phase fault is
z 20.0
~
o
~
z
w
a:
~ 10.0
u :g 9.0 \\\\o 8
<( 8 .0
~ 10 7.0
1\ \\\l 10
o1 6.0
\\\\ ~12
-' ~ 5.0 1\ \ \ "
-' "~
\ '\.
-'"""'
'-'15
s::E 4.o
1\\ .......... ~
IL.~
l\ \.
..._ 1""""'-o...
(/) ~ 3.0
UJ(/)
::E
~
IL.
o
a:
2.0
\\
1\
\
\"- ..__
1\ '\.
"
-
~
30
40
50
- 10 hl5-
l-t2'l 8
\"- 75
~
::E
;:)
z
1.0
0.9
0.80
' lOO
0.5 1.0 1.5 2.0 2.5 3.0

TIME IN SECONDS
Fw. 57.-Short-circuit decremcnt cur ves.
System reactance must be based on system kilovolt-amperes and not on the
particular kilovolt-ampercs chosen as a base for calculations.
System kilovol t-amperes is connected synchronous capacity in kilovolt-
amperes.
Note: These cur ves ar e based on T' do = 5 sec.
For other values of T' do. multiply actual times by T~ ' to get equivalent times
dO
to use in curves.
H ow to U se Curves:
For three-phase shm·t-circuit:
Reactance: Use system reactance to the point of fault.
Times full l oad (normal) scale: U se se ale reading.
For line-to-line short-circuit:
Reactance: Use two ti\nes the system reactance for the t hree-phase fault.
Times full-load (normal) scale: Multiply scale reading by y3,
For line-to-ground short-circuit:
R eactance: Use (ZX1 Xo) +
Times full-load (normal) scale : Multiply sca.le reading by 3.
what is commonly spoken of as t he system reactance referred to

or viewed from the point of fault. I t is the reactance used
with the decrement curves formerly published. When deter-
mining this reactance it is important t h at synchronous machines
be represented by their subtransient reactances, and also that
al1 loads other than synchronous be neglected. Proper account
has been taken of these loads for the average system in deriving
the decrement curves.
The system reactance referred to the point of fault enables
the proper decrement curve to be .selected which gives the three-
phase r.m.s. short-circuit current or kilovolt-amperes in times
normal at any time after the occurrence of the short circuit,
subject to the assumptions previously given.
Line-to-line Short-circuit. For a line-to-line short-circuit,
the fault current is determined by the use of the standard decre-
20.0
\
w
l1l
e(
CD
,.,,
e(
>
1 ,\
X:
::l 10.0
w 9.0
'l\\
1-
<ll \
>- 6 .0
<ll \
7.0
z \ \ \ \.
o 6.0 \ \ \ \.
1- t--- 1\ \ \
z \ \ 1\\ \
w 5.0
\ \
a::
a::
B 4.o 1\.~ 1\\. 1\.
~
"
r\'\.
;r.
I~C
"" ~"'
"'\.
o o 011.
""'- ·oo8 <:)-~'
e(
o
..J 3.0
¡..-.., ' r\"\
...... ['._ qos "['-.._
f'\."' ""'
..J
..J
ql '
)~ ~" ~ ~~~ ~ r:--....
::)
1.1.
C/'1
w 2.0 "' ..........: r--. "~"'q~ ~ ~
.........
r.....
:l ........
¡:::
~~'
<?.~ :S"
r-.. !"......
¡....... .............
1.1.
o
::-....... r-......'
a::
w ~~ ~ ~ ~ ~
~
m
~
::) 1.0 ~~~
z 0.9 ~
0.80 lO 20 30 40 50 60 70 80 90 100
SYSTEM REACTANCE IN PER CENT
Fxo . 58.- Short-circuit decrement curves.
For way to use curves, see Fig. 57.
ment curves of Fig. 57 or Fig. 58 in the same manner as the three-

phase short-circuit current by using a reactance equal to twice
the system reactance (the value used for three-phase short-
circuit calculation) and multiplying the resulting current in
times normal read from the curves by V.3.
Where the negative-sequence reactance of the system is known,
more accurate results may be obtained by using a reactance
equal to the sum of the system reactance x1 and the negative-

sequence system reactance X2 instead of twice the system react-
ance X¡. The negative-sequence reactance of the system is
obtained in the same manner as the system reactance x1 except
that the negative-sequence reactance of the machines is sub-
stituted for the subtransient reactance of machines.
Line-to-ground Fault. The value of reactance to be used
with the curves fo r a line-to-ground short-circuit is the sum of
the positive-, negative-, and zero-sequence reactances of the
system (x1 + +
x2 xo) as measured from the point of fault. For
ordinary calculations it is sufficiently accurate to use a reactance.
equal to (2xl + xo), where x1 is the system reactance used for
three-phase short-circuit calculations and xo is the zero-sequence
reactance of the system. The line-to-ground current at any time
will be equal to three times the value of times normal current
read from the decrement curves using the above values for systcm
reactances.
Minimum Fault Current. For certain relay applications it
is sometimes necessary to know the alternating component
of the fault current, i.e., with no direct-current component,
rather than the maximum current at any time as given by the
standard decrement curves which are based on an asymmetrical
fault. The initial r.m.s.- alternating (symmetrical) currents
for values of system reactance are given in Table IV. As the
asymmetrical component is negligible after 0.3 sec., the r.m.s.
alternating (symmetrical) current for any time less than 0.3
sec. may be obtained by interpolating between the initial valuc
obtained from Table IV and the curve values at 0.3 sec. Sn.tis-
factory results can readily be obtained by sketching in a curve
starting from the 1initial r.m.s. alternating (symmetrical) current
and making it tangent to the corresponding standard decre-
ment curve at 0.3 sec.
TABLE IV.- INITIAL R.M.S., ALTERNATING (SYMMETRICAr,) CuRRENT
Reactance on d ccrement
curve ................. . . 8 10 12 15 20 30 40 50 75 100
Initial symmctrical current . 12 .3 9.96 8.38 6.88 5.11 3 . 37 2.52 2. 01 1.33 1.0
R.M.S. Total Current for Salient Pole Machines without

Dampers. Where salient pole machines without damper wind-
ings supply the short-circuit current the standard decrement
curves give good results after 0.2 sec., provided that for these
machines thc equivalent subtransient reactances are obtained
from the known lransient reactances by means of the relation
given in Fig. 54. The initial values indicated by the standard
decrement curves may be 30 per cent too high. If more accurate
valucs of iniLial current are desired, it is necessary to make a
separate calculation of the system reactance using the actual
value of the subtransient reactance of the salient pole machines
without damper windings instead of the equivalent value obtained
from the transient reactance by means of the relation given in
Fig. 54. A curve may readily be sketched between the new
initial point and the 0.2-sec. point on the standard decrement curve
corresponding to the equivalent subtransient reactance of Fig. 54.
The subtransient reactance of machines without damper
windings has generally been taken as equal to the transient
reactance. Howcver, there is an appreciable subtransient
component in the shorL-circuit current of such a machine due
to the damping action of the po]c riveLs anrl. other closed circuits
and also due to thc rapidly disappearing effects of saturation.
The test daLa available indicatc that the subtransient reactance
of a machine wiLhout damper windings should be taken as about
85 per cent of the transient reactance.
53. Interna! Voltage Method.

The interna! voltage method * of calculating short-circuit
currcnt variation was developed to allow solution of short-
circuit problems of systems that cannot be replaced by a single
equivalent generutor, i.e., systems in which machines are located
unsymmetrically with respects to the fault or have different
time constants and reactances. The method is adapted to the
use of the calculating board, which use reduces considerably
the time required for solution. It is also easily extended to
include the effect of variation in exciter voltage.
It may be seen from Fig. 47 that the contribution of the sub-
transient component is practically negligible, 0.1 sec. after a fault
occurs, and since one is usually interested in a breaker opening
time of either one cycle (practically instantaneous) or something
over 0.1 sec., the subtransient component may be neglected
exc~pt as it influences the evaluation of the direct-current com-
ponent at zero time.
* For a more detailed discussion including an illustrative example of this
method refer to bibliography itero 76.
Fundamental Assumption. The variation of the transient

component of the alternating current is determined from the
variation of internal transient voltage. The assumption is
made that the internal transient voltages of the individual
machines vary in an exponential manner between their initial
values and their final values. The initial values are obtained
from the loading of the machines previous to the inception of
the fault, and t he final values are obtained by using the syn-
chronous internal voltages of all the machines, calculating the
sustained current in all of the machines and then subtracting
from the synchronous internal voltages of the individual machines
4.0
3.6 1--~;;::r--.----r~~
3.2 ~.l-l,.-J...=:::t;::=F""'
0
1 2 3 4 5 6 o 1 2 3 4 5 6
Time in Seconols Time in Seconols
---lb appro)(.ime~te
0o 1 2 3 4 5 6
Time in Seconols
FIG. 59.- Comparison be tween actual short-circuit currents and those obtained
by using "interna! voltage " m ethod of calculation. Subtransient and direct-
current components have bcen neglected.
the difference in the synchronous and the transient reactance

drops. This will give the sustained transient internal voltage
back of the transient reactance for the individual machines.
The time constant for the exponential curve connecting the
initial and final values of transient interna! voltage is determined
as though the particular machine were the only one connected
to the system, the other machines being replaced by their respec-
tive synchronous reactances. The accuracy of this assumption
may be gauged from Fig. 59 which gives the transient alternating-
current component for systems shown in the inserts as determined
in two ways: (1) by the internal voltage method, and (2) by an
accurate differential-equation solution. Two of these systems

consist of two generators each having different time constants.
The shunt reactances may represent either the reactance of a
high-reactance three-phase fault or the equivalent negative-
sequence reactance of a line-to-line fault. The error involved
may be seen to be very small and becomes still smaller as the
shunt reactance is decreased. The left-hand figure represents
the current resulting when a line-to-line fault is applied to the
system shown, which consists of two generators connected to
an in:finite system. It is evident that the accuracy of the
interna! voltage method is quite sufficient for most practica!
purposes.
Altemating-current Component. Having given the instan-
taneous value of the transicnt interna! voltage for the individual
machines, the posiLive-sequence component of the transient com-
ponep.t of current in any branch of the network can be obtained for
any instant by inserting the particular values of transient interna!
voltages obtained from the exponential curves for the particular
instant. For the different kinds of faults the negative- and
zero-sequence components can be obtained by calculating the
currents in the particular branches of the negative- and zero-
sequence networks. For line-to-line faults the· maximum
positive- and negative-sequence components of current in the
reference phase a are in phase opposition, so that the maximum
current occurs in phases b and e and the sequence components are
60 deg. out of phase. If I 1' and I 2' be the positive- and the
negative-sequence currents in any branch, then the total current
is given by
•
(127)
For a line-to-ground fault the maximum current occurs in the
particular phase upon which tbe fault occurs and tbe value of
this current is given by
I' = I 1' + I + I o'
2' (128)
Direct-current Component. The initial value of the sub-
transient current !" is necessary in thc determination of Idc and
is obtained by computation or measurement on a calculating
board by using the initial value of subtransient interna! volt-
ages and the subtransient reactances in the various machines.
The initial value of Idc is obtained from the vector difference
108 SYMMETRICAL COMPONE_NTS
between the initial value of 1" and the load current in a manner
analogous to the case shown in Fig. 44.
Calculations have shown that the direct-current component
will increase the r.m.s. value of short-circuit current less than 5
per cent after 0.1 scc. if the direct-current time constant of the
circuit is 0.05 sec. Hence for such machines the contribution
of the direct-current component may be neglected after 0.1 sec.
For faults involviag only synchronous machines and transformers
the time constants are usually greater than 0.05 sec., and there-
fore the direct-current component will be of importance beyond
0.1 sec.; but for circuits involving transmission or distribution
lines the time constant is less than 0.05 sec., and therefore the
contribution of the direct-current component is negligible after
0.1 sec.
Summary. To summarize the method, the subtransient
component because of its rapid attenuation is negligible after
0.1 sec., and therefore the total alternating-current component is
equal to the transient component which is obtained by calculating
the time variation of the transient interna! voltage of the indi-
vidual machines and using these voltages in connection with the
network which is set up with the transient reactanccs of the
machines. .The direct-current component is obtained from a
knowledge of the initial subtransient alternating-current com-
ponent and the load current before the fault. The initial value
of total current is then obtained by taking the square root of the
sum of the ~\tares of the initial direct-current and the initial
subtransient current. The total current, when the fault involves
transmission or distribution lines, is equal to the transient alter-
nating-current component after 0.1 sec., and it is only when the
fault involves only transformers and generators that the direct-
current component must be taken into consideration after
0.1 sec.
54. Effect of Varying Exciter Voltage.

The interna! voltage method is easily extended to include
the effect of alteration in exciter voltage, <76l in fact this is
one of the principal advantages of the method. The direct-
current and subtransient alternating-current components are
affected only slightly by exciter-voltage variation and may there-
fore be computed in the usual manner neglecting the variation in
exciter voltage. The problem then reduces to the determination
•
r
1

of the instantancous values of transient interna! voltage of each
machine, after which the proccdure is identical with that for con-
stant cxciter voltage as discussed previously.
In brief, the method consists of the following. The sustained
valucs of transicnt internal voltages corresponding to the instan-
taneous values of exciter voltage are computed. In this way
curves are obtained of the valucs thc transient interna! voltage
would attain if the time constant of the machine were zero.
The transient time constants
of the individual machinesare
then o btained as described
under the interna! voltagc
method. The actual instan-
taneous transicnt interna!
voltage for a particular
machine is then obtained by
plotting t hes e sustained
values displaced a distance
eq u al to the transien t time
constant from the point of I•'ro. üO.- Dctcrmination of e' with variable
cxcitation.
zero time as shown in Fig. 60.
Starting with thc initial value of transient internal voltage which
is givcn by the point a, the actual locus is built up in small
incremcnts by following along the lines ab, cd, etc., as shown iri the
figure. This must be repeated for each machine, after which the
instantaneous values of transient component of current are easily
obtained.
Problems
1. A thrce-phasc gcncrator operating at normal voltage no load is sub-
jectcd to a thrce-phase short-circuit. From the oscillograms thc current
in pcr unit valucs aftcr climinating thc dircct-current componcnt is found
to be as follows:
Time J Time J Time I
0 .00 4.48 0.25 2.47 1.50 1.27

0.05 3.41 0.40 2.19 2.00 1.13
0.10 2.96 0.50 2.04 2.50 1.05
0.15 2.72 0.70 1.80 3.00 1.01
0.20 2.56 1.00 1.55 .. .. 0 .96
Find the values of X i, X i', X d, Ti, and Ti'.

2. The direct-current component of current in the three phases of a short-

circuited generator are 1,060, 1,932, and 516 amp. In this test the short-
circuit occurred at such an instant as not to produce thc maximum
dissymmetry. Determine (a) the maximum possible value of the direct-
current component; (b) the peak value of the alternating-current component;
(e) the maximum possible r.m.s. total current.
3. A three-phase 50,000-kva., 11,000-volt, 60-cycle generator is subjected
to different kinds of short-circuits while operating at ratcd voltagc no load.
The sustained short-cirruit currents are found to be: three-phase fault,
2,000 amp.; line-to-line fault, 1,800 amp.; line-to-neutral fault, 2,200 amp.
The instantaneous symmctrical three-phase short-circuit current is found
to be 20,000 amp. Determine the values of Xi', Xd, X2, and X o in per cent
and in per unit values.
4. Compute the value of the fault currcnt at the point F of Fig. 23(a),
Chap. IV, at 0.2 sec. after the inception of thc fault for different types of
faults as follows: (a) line-to-ground; (b) line-to-linc ; (e) three-phase. Use
the standard decrement curves assuming that the reactances given include
the subtransicnt reactance X i' for each machine and neglect the effect of
are resistance.
6. Assume the same system as in Prob. 4 but that the generators are all
of the salient pole construction without damper windings. Determine the
value of the instantaneous symmetrical short-circuit current at 0.05 sec.
after the inception of the fault for (a) line-to-ground fault¡ (b) line-to-line
fault; (e) three-phase fault.
6. Assume that an exciter builds up its voltage immediately upon the
occurrence of a fault to a magnitude and at a rate given by the expression
/
t
E.,= 3- 2E-T
where E., is expressed in per unit excitation, t is the time as measured from
the application of the fault, and T is the time constant of the exciter build-up
curve. Compute the alternating-current component of the armature
current as a function of time for the case of the generator subj ected to a
three-phase short-circuit from no-load normal voltage. Neglect the
subtransient effects and assume the constants
Xi = 0.40 T = 0.3
xd = 1.00 = 2.0
Tdo 1
CHAPTER VI
CONSTANTS OF TRANSFORMERS
The sequence impedances of transformers and similar devices,
including series and shunt impedances, two- and three-winding
transformers, and autotransformers, will now be considered.
Since this apparatus is non-rotative, and since symmetry between
the different phases has been assumed, it follows that the imped-
ancc of these devices is independent of the phase rotation of the
e.m.fs. applied to the terminals, and that the positive- and nega-
tive-scquence impedances are identical. The zero-sequence
impedances of thcse devices are, in general, different from the
positivo- or negative-sequence impedances. While the imped-
an.ce of the device itself may be represented by the same equiva-
lent circuit, the actual impedance is dependent upon the externa!
connections, which may be different for the zero-sequence.
55. Two-winding Transformers, Series and Shunt Impedances.
Series-impedance branches which are symmetrical in the
different phases have the same value for positive-, negative-,
and zero-scquence, if there is a neutral return as illustrated
in Fig. 61(a). If there is no path for zero-sequence current,
the corresponding zero-sequence impedance is infinite. Each
device may be represented in the zero-sequence network by a
suitable connection, as shown in the right-hand column of
Fig. 61 with the impedance the same as for positive- or
negative-sequence.
Shunt-impedance branchcs which are symmetrical in the
different phases have the same value for positive-, negative-,
and zero-sequence, if there is a neutral connection as illustrated
in Fig. 6l(c). However, if the shunt branches are connected
in delta or in star without neutral connection, as illustrated in
Fig. 61(d), the impedance to zero-sequence is infinite and is
represented by an open-circuit.
Neutral impedance in a neutrai-wire or ground connection
ii1. of course, wholly zero-sequence. The value of this impedanoe
111
should be multi plied by three to obtain the equivalent value for

phase-to-neutral. This is illustrated in Fig. 61 (e), which shows
a star-connected group of shunt impedances with a neutral
impedance in the ground connection. It will be noted that the
equivalent circuit for zero-sequence includes two impedances
in series, the first impedance having the same value as the actual
impedance per phase in the group of star-connected impedances,
whereas the second impedance has three times the value of the
impedancc in the neutral connection.
Three-phe~se
connection Connedion for
~~~TI~
(f)
(01) z zero·SeCJuence De liCI-DeltCI
Series impeobnce --'i--,.,~ ~
.w1lh neulrGOI connedion
(b)
--"-~-
Nev+~<>l
--=z'"""""'~- Ungro~~~ecl star-~

ungrounoleol st01r ~
E
mT
Series impedf.lnce ~
without neutro.! connedio~
GrounS:~ s+Cir- ~e=
grounoleolstor ~ ~
iln edC~nce
Shunt (e) z
~~
with neutrclconnection
Grounol~cl
ro T
z z (') sfOir-
oleltCI ·
Shunt (d)impecle~nce z
(')
Ungrou~cleol stw- 3~
oleltCI
without neutro~! conneclion z z
3E
ttr
y (k) ste~r-
Ungrounoleol
grou n cle ol stCI r
Shuni (e)
impeclCince
z z
Z
lnterc~~nedeel-:Jfs
.··-S01me
with neutml impeolcmce
FIG-
ZN
s t01 r gro u ncl i ng
+rG~nsfo rmt>r
., .
cor~;
1/
61.- Equivalent circ uits for zero-sequ ence of transformers , series and ahunt
T
impodances.
The two-winding transformer can always be replaced by an

equivalent circuit consisting of two series impedances represent-
ing the primary and secondary leakage impedances, anda shunt
branch representing the exciting admittance. For certain
applications, the exciting-admittance branch is required; but
for short-circuit calculations it may be neglected and the equiva-
lent circuit reduced to a single series impedance. Typical
values of transformer impedance are given in Table V, these
values represeni¡ing the impedances to positive- or negative-
sequence. These impedances are usually expressed in per cent
on a given kilovolt-ampere base, or in ohms on a given voltage
base. In any event, they must be reduced to the common base
chosen for the calculations.
CONSTANTS OF TRANSFORMERS 113
TABLE V.-TYPICAL VALUES OF TRANSFORMER REACTANCES
Per Cent
Distribution. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
N etv,rork. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
Power
Up to 66 kv............... . ....... .. .. . . .. ... . . 5-7
88 and llO kv..... . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-9
132 and 154 kv..... . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8- 10
187 and 220 kv .......... . ........ . ........ .. ... 1ü-14
The direct-current resistance varíes from 0.35 to 0.50 per cent.
The impedance to zero-sequence of a bank of three two-

winding transformers* is either t he same as for positive- or
negative-sequence, or is infinite depending upon the type of
winding. The connections of the equivalent circuit of a two-
winding transformer to the externa! circuit for zero-sequence
are illustrated in Fig. 61(! to l). The most common connections
are delta, star, and grounded star. In the case of any delta
winding, it may be seen (1) that no .zero-sequence current can
flow from the delta winding to t he externa! circuit or vice versa;
and (2) that zero-sequence current can circulate in the delta
winding, without flowing through t he externa! circuit. The
zero-sequencc connections for the eq uivalent circuit of a delta
winding are t h erefore an open-circuit between t he delta winding
and the externa! circuit, and connection to ground of t he delta
side of t he equivalent impedance, as illustrated in (f). In the
case of a star winding with free neutral, no path is provided
for the flow of zero-sequence current either from the externa!
circuit or in the t ransformer winding itself. Hence, for the
zero-sequence connection, a star winding with a free neutral is
represented by an open-circuit of the star end of the branch repre-
senting the transformer for both t he externa! circuit and to
ground, as shown in (g). In the case of the grounded star
winding, zero-sequence current can flow through the transformer
winding only when the externa! circuit provides a completing
circuit. H ence, the connections of the equivalen t circuit for
grounded star winding are as shown in (h) . Other combinations
of ungrounded star, grounded star and delta windings are shown
in (i) (j) and (k) .
The interconnected star or zigzag grounding transformer,
illustrated in (l), is the only other form of two-winding trans-
* For three-phase transformers see Sec. 57.

former commonly cncountered in a three-phase bank. This

transformer has two separate windings on a common core, which
are interconnected between phases, the windings on the same core
being represented by parallel branches in the diagram. This
transformer has an open-circuit or, more accurately, an exciting
admittance for positive- or negative-sequence e.m.f. supplied
to the terminals. For zero-sequence, however, the currents
in allline terminals are of the same value. Hence, the impedance
per phase to zero-sequence is the leakage im pedance from one
winding to the other winding on the same core. The zero-
sequence connection is therefore a simple shunt impedance to
ground, as shown in (Z). The interconnected star-delta trans-
former is discussed with the three-winding transformers.
56. Three-winding Transformers.
For stability or short-circuit calculations the exciting current
of three-winding transformers may be neglected and only the
ZAs = ZA + .Zs
Zsc = Zn + Z c
Z cA = Zc + ZA
ZA = H ZAB + ZAc - Zsc)
Zs = H ZAs +Zsc - ZAc)
Zc = HZcA +Zsc - ZAs)
FIG. 62.- Equivalent star representation of three-winding transformar in

terms of impedance betwecn windings. Impedances ZAs , etc., are the short-
circuit impedances m easured between the windings corresponding to the sub-
scripts, with the other winding open. All impedances must be exprcssed in
the same kilovolt-ampcre and voltage base.
series drop characteristics taken into consideration. The char-

acteristics of such transformers are usually given in terms of the
reactance lJetween two windings taken at a time, the third being
open-circuited. While in two-winding transformers both wind-
ings have the same rating, the difference in rating of the three
windings in three-winding transformers requires that special
care be exercised to specify the particular base to which the
impedance refers. It has been shown* that when the magnetiz-
ing current is neglected, the performance of the transformer can
* PETERs, J. F., and M. E. SKINNER, Transformers for Interconnecting
High Voltage Systems, Trans. A.l.E.E., Vol. 40, p. 1181, 1921.
be represen ted by an equivalent circuit consisting of three star-
connected impedances. In Fig. 62(a), let the impedance between
windings A and B be designated ZAB; that between B and e,
Znc; and that between A and e, ZAC· The equivalent star
impedances are shown in Fig. 62(b). For winding open- e
circuited, it follows that
ZAn = ZA + Zn (129)
and similarly for the other windings
Znc = ZB + Zc (130)
ZAc = Zc + ZA (131)
It also follows from the equivalent diagram that the voltage
across winding B with voltage applied to A and with e
short-
circuited is the drop across the impedance Zc. Expressions for
converting these im pedances from one form to another are given
in Fig. 62. It should perhaps be pointed out that the star point
of the equivalent circuit in Fig. 62(b) is a fictitious point and does
not represent the system neutral and that loads or short-circuits
can be applied only to terminals.
Example. As an cxamplc of thesc relations consider a three-phasc trans-
former bank having the following characteristics:
Winding A - 7,500 kva. (2,500 kva. per phasc) star-connected for
66,000-volt line
Winding B- 10,000 kva. (3,333 kva. per phase) star-connected for
11,000-volt line
Winding C- 5,000 kva. (1,667 kva. per phase) delta-connected for
2,300-volt line
The per cent resistance drop for this transformer may be assumed to be
as follows:
Winding A--ü.75 per cent resistance drop on 7,500 kva.
Winding B- 1 pcr cent resistance drop on 10,000 kva.
Winding C- 1 pcr cent resistance drop on 5,000 kva.
The per cent rcactance drop betwcen the various windings may be assumed
to be as follows:
Winding A-B- 10 per cent reactance drop on 10,000 kva.
Winding B-C- 5 per cent reactance drop on 10,000 kva.
Winding C-A- 10 per cep.t reactance drop on 5,000 kva.
In obtaining the data for making the substitution of the star group of
impedances for the three-winding transformer, it is important to express
the impedances in terms of the same voltage and kilovolt-amperes per wind-
ing in order to avoid error. For this purpose express the transformer
impedances in terms of a 10,000-kva. bank for a system voltage of 11,000
1
volts. The phase voltage is l_j~0 = 6,350 volts, and the phase current is
10,000,000 - 525
3 X 6,350 - amp. Hence the equivalent winding resistances in ohms
per phase are as follows:
X ~
0 01 6 350
ra = 0.75 per cent on 7,500 kva. = 1 per cent on 10,000 kva. -
-
·
525
= 0.12 ohm
rb = 1 per cent on 10,000 kva. = 0.12 ohm
re = 1 per cent on 5,000 kva. = 2 per cent on 10,000 kva. = 0 02
· ~:~ 3 50
= 0.24 ohm
Similarly, the winding reactances in ohms por phasc are as follows:
0.10 X 6 350
Xab = 10 p er cent on 10,000 kva. = ' = 1.21 ohms
525
X 6 350
Xbe = -? p er cent on 10,000 kva. = 0.05 525 ' = 0.60 ohm
0.20 X 6 350
Xea = 10 per cent on 5,000 kva. = 20 per ccnt on 10,000 kva. = '
525
= 2.42 ohms
The transformer impedance bctween windings A, B, and C may now be
written ·
ZAs = ra + rb + jXab = 0.24 + jl.21 ohms
+
Zsc = rb +re jXbc = 0.36 + j0.60 ohms
[ ZcA = ra +re + jXac = 0.36 + j2.42 ohms
The above values of transformer impcdance may be substituted in the
equations of Fig. 62 to obtain the equivalcnt star-connected impedances
with the results.
zA = Z AB + Z AC - Z BC
2
(0.24 + jl.21) + (0.36 + j2.42) - (0.36 + j0.60)
2
+ j3.03 = o. 12 + J"1 .or.: l. oh ms
-- 0.24
2
Zs =ZAs + Znc- ZcA
2
(0.24 + jl.21) + (0.36 + j0.60) _.:: (0.36 + j2.42)
2
0 ·24 ~ jO.Bl = 0.12 - j0.305 ohms
Zc = ZcA + Zcs -ZAs

2
(0.36 + j2.42) + (0.36 + j0.60) - (0.24 + jl.21)
2
= 0. 4 81 jl.S1 = 0.24 + jO.ü05 ohms
NOTE. The resistances of the r)l·anches ZA, Z s, and Zc, are identical
w;th the valucs of r 0 • rb, and rr, rcspectively. This will always be the case
for transformers wíth separate windings, so that the resistance terms of
ZA, ZB, and Zc may be determined in this manner.
The impedance of three-winding transformers for the positive-
and negative-sequences may be represented by the groups of
three star-connected impedances just described.
The equivalent circuit of three-winding transformers to zero-
sequence involves the same complication due to the connection
of windings, whether they are Connection for
delta-,' or grounded- or un-
grounded-star, as arises in the
case of the two-winding
transformer. The same con-
ceptions with respect to the
connection of the im pcdances
to the external circuit or to
ground as were used in connec-
tion with the two-winding
transformer diagram in Fig. Fw. 63.- Comrnon three-winding
61, may be applied to any type transformer connections with equivalent
of three-winding transformer. circuits for zcro-sequence.
The results for three common types are shown in Fig. 63. The
impedances in these diagrams are, of course, the per phase values.
For the delta-connected winding they may be visualized best by
regarding the winding as a fictitious star-connected winding of
the same capacity and same percentage impedance.
67. Three-phase Transformers.
The preceding discussion applied directly to three-phase
banks made up of single-phase units. Three-phase transformers
require further consideration because of the fact that flux paths
,,
for the several windings may be common to each other.
The iron circuit of a common form of the three-phase shell-
type transformer is illustrated in Fig. 64(b). The middle wind-
ing is reversed to decrease the flux in the section between any
two windings. For convenience in comparing the iron circuits
the three-phase bank of single-phase transformers is illustrated
in Fig. 64(a). For positive- or negative-sequence voltages
applied to terminals a, b, and e, of either the single-phase bank
or the three-phase transformer, the impedance is very high since
it is that corresponding to the exciting impedance. In the case
of zero-sequence voltages applied, it will be observed tha.t the
outer legs provide an iron return path for the flux set up by the
zero-sequence currents. Thus the zero-sequence exciting imped-
ance of the single-phase transformer bank of Fig. 64(a) or the
three-phase shell-type transformer is the same as for positive- or
negative-sequence voltage of the same value across each phase.
The presence of secondary windings has the same effect upon the
short-circuit impedance for either the single-phase bank or th~
three-phase shell transformer.
(e) -:::-
(b) ~
FIG. 64.-Diagrams to illustrate the zcro-sequence impcdances of three-phase
transformer banks. (a) Three-phase bank of single-phase transformers; (b)
three-phase shell-typc transformer; (e) three-phase corc-type transformer.
In the case of the three-phase core-type transformer with

three legs as illustrated in Fig. 64(c), it will be ob~erved that
for positive- or negative-sequence currents a return circuit is
provided through the other legs and the exciting impedance
is thus very high. For the case of zero-sequence voltages,
however, it will be noted that there is no iron return path for
the flux set up by the currents in the windings on the three legs.
Consequently, zero-sequence currents will produce relatively
little flux within the transformer and the reactanee to zero-
sequence will be low compared with the ordinary exciting imped-
ances, though the value will be high when compared with the
ordinary series leakage reactance of the transformer. In com-
mercial three-phase core-type transformers the zero-sequence
exciting impedance of a grounded-star winding varies from 30 to

300 per cent, the higher values applying for the larger size units.
The presence of the secondary windings has practically the same
effect upon the short-circuit impedance as for a bank of single-
phase transformers. However, because of its lower exciting
impedance, the zero-sequence impedance of the core-type trans-
former will vary from 90 to 100 per cent of its positive-sequence
impedance.
The core-type three-phase transformer may be considered
"in the same manner as the other three-phase banks if the effect
of the magnetic circuit be replaced by a fictitious delta tertiary
winding of very high impedance. This tertiary winding tends
to reduce the flux in the transformer due to the fl.ow of
zero-sequence currents from the value which obtains for positive-
sequence currents and makes it possible to obtain the inter-
mediate value of impedance as described in the preceding
paragraph. This tertiary winding should be considered in
addition to any other delta-connected windings that may exist
within the transformer, for example, a three-phase core-type
transformer with the primary winding connected in grounded-
star anda secondary connected in delta would havean impedance
to zero-sequence corresponding to Fig. 63(a) . Thus it appears
that the zero-sequence impedance of grounded star-delta trans-
formers of the three-phase core construction is somewhat less
than the positive-sequence impedance.
68. Autotransformers.
Autotransformers encountered in power systems usually con-
sist of two windings, parts of which are common, and a third
independent winding called the tertiary winding. The common
windings of three-phase banks are connected in star, and the
tertiary in delta. Exceptions to this type of autotransformer
are to be found in applications of single-phase transformers to
railway work. The present analysis was developed formally
for the three-winding autotransformer, permitting tbe two-
winding autotransformer to be evaluated as a special case.
When the magnetizing current is neglected, it is possible to
select three pairs of terminals and completely represent the
characteristics of the transformer by a set of three star-connected
impedances between these terminal pairs. A choice exists as to
the appropriate pairs of terminals involved in the common wind-
ing. Figure 65(a) and (b) shows the two combinations that are
usually specified. For convenience, the left-hand combination
will be referred to as on the winding b·a sis and the right-hand
combination, as on the circuit basis.
The equivalent circuits of the autotransformer for the positive-
and negative-sequence are shown in Fig. 65(c) and (d) for the
DI
h
S
-
~ .--1-
1-
1-
1-
~
~
- r-- ~
e e f:: ~
-~
(01) {b)
S H
Ib(Giclool) lc(cdual)
Zs = Iq(actual) ZH •lb(adual)
+le (actual) -ICII(aclual)
ZT ZM
e "·
1 T M ~ 1
L
Ic(equiv.)"lccactual) l¡,(equiv.)" I¡,(aduCII) / (d)
(e)
Zr = t(Zrc + Zrs - Z cs) Z L = H ZLM + Zu¡ - ZMu)
Zc = HZcr + Z cs - Z sr) Z M = HZML + ZMH - ZLH)
Zs = HZsr + Z sc - Zcr) zll = HZHL + ZuM - ZLM)
(e) (j)
ZML = Zcr Zcr= ZML
z uM = (n : 1
Yzsc Zsc = (n~ 2
1) ZHM
Zrc n 1 n
ZH L = -
n
+ n--Zrs---Zcs
·- 1
n
n-1
n 2 Zrs = -
n--1ZuL - -
n--1ZML+ (n- l) 2 ZnM
(g) (h)
n
Zc = - - ZM
n - 1
ZM
Zr = Z L- - -
n- 1
· ( n
Zs = n - 1 ZH
)2 + (n n
- 1) 2Z~
(i)
FIG. 65.-Conversion formulas for autotransformer with tertiary winding. e
and M windings used as base. Ratio H / M = n.
winding and circuit bases, respectively. It will be noted that

the form is identical with that of the transformer with separate
windings.
The equivalent circuit to zero-sequence for autotransformers
involves the same connections of the equivalent circuits as for
transformers with three separate windings. This is illustrated in
the example given in a later paragraph of this section.
Conversion of Autotransformer Impedances from Winding
Basis to Circuit Basis. Frequently the autotransformer imped-
ancés are given on one basis and it is desired to use impedances
in the network on the other basis. It therefore becomes neces-
sary to derive expressions whereby t he impedances on one basis
may be converted to the impedances on the other basis.
Before going directly to these dcrivations it may be well to
specify more clearly the currents in the different circuits. The
impedances shown in Fig. 65 and consequently the currents
fiowing therein are all based on a common voltage. The actual
currents in the various circuits can be obtained from the currents
in the equivalent circuits by converting the currents from the
common voltage base used in the calculations to the actual
voltage base of the particular circuit. The currents in branches
e and S of Fig. 65(c) represent, after converting to their respec~
tive voltage bases, the actual currents in conductors e and a,
respectively, of Fig. 65(a) . The actual current flowing in the
common conductor b of Fig. 65(a) cannot be represented in the
equivalent diagram, but is the sum of the actual currents in
conductors a and c. Similarly, currents in branches H a:r..d M
of Fig. 65(d), after reducing to the proper voltage base, represent
the actual currents in conductors a and b, respectively, and the
. actual current in conductor e is the difference of the actual
'·· currents in conductors a and b.
For the purpose of this analysis the winding .M will be used as
the refcrence winding, and the impedances will be expressed
in terms of this voltage as the base. A ratio n will be taken to
give the turns ratio of the high-voltage to the medium-voltage
winding, that is
En Ec +Es
n = - = ---=-- (132)
EM Ec
a:~ad
(133)
In addition, windings T and L may be assumed for the present,

without limiting the general application of the formulas, to have
e
the same voltage base as windings and M.
Let it be assumed that all impedances appearing in the equa-
tions of Fig. 65(e) are given on the same kilovolt-ampere base
and the voltage base of the e winding, and it is desired to convert
these impedances to the terms appearing in the equations of

Fig. 65(!), all of which are expressed on the same kilovolt-ampere
base and voltage base of the M winding.
Determination of ZHL· Apply voltage to the H winding
and circulate unit current through this winding with the L
E = Es(actual) +
Ec( actual)
'---------' 1
f"or / amp in a, ZHL on C windjng base·~
= n 2ZT +
(n - 1) 2Zs +
Zc
E
S - =
ZuL =n2 z7' +
/ampj Ac.tuol
· Curren+s
e
n- 1
- --
n2Z c s
Fw. 66.- Determination of ZFTL in terms of s. e and T constants of transforrner.
E = (n - 1) 2 (Zs + Zc)
= (n - 1) 2Zsc volts
,
n
E
ZHM = - 2 = --
(n- 1)2
n
Z .sc
Fw. 67.- Determination of ZH.M in t erms of S, C, and T constants of transformer.
winding short-circuited as shown in Fig. 66. For this condition

the actual currents flowing in circuits S and e are the actual
currents in the terminals a ande, namely, unity in both terminals.
These currents in the equivalent circuit correspond to (n- 1)
in circuit S and unity in circuit e, t he former flowing into and
the latter flowing out of the star point. The equivalence
of these currents in the actual and in tbe equivalent windings
may easily be verified by consideration of the corresponding
magnetic effect as inftuenced by the relative number of turns.
The current in the T winding must equal the difference of these,
namely, (n - 1) - ( -1) = n ftowing from the star point.
Since the T winding is short-circuited, its voltage is zero. The
equivalent voltage across the S winding is then the sum of the
drops in the T and S impedances, thus nZr +
(n - 1)Zs, and
the equivalent voltage across the e winding is the difference in
drops in the T ande windings, thus nZT - ( -1)Zc = nZr + Zc.
The actual voltages are obtained by multiplying the equivalent
voltage of S by (n - 1), the ratio of the turns of Sto those of e,
e
and the equivalent voltage of by unity. The actual voltage E
across 11 is the sum of these two voltages. Since by assumption
unit current is flowing in these windings, the value of E is equal
to the impedance ZuL in terms of the voltage base of 11. Divid-
ing by n 2 converts this impedance to the common voltage of the
e winding with the final result
ZuL = Zr + n -
( n
1) Zs + nZc
2
2 (134)
Inserting the values of ZT, Zc, and Z s from Fig. 65(e) gives
Zrc
ZuL = -
n
+ n-
n
1
Zrs -
n- 1
n2
Zcs (135)
Determination of ZHM· The impedance ZHM between the

high- and medium-voltage windings in terms of the impedances
of the other combination may be obtained in a similar manner.
The details of the steps involved are indicated in Fig. 67, giving
finally
2
1
ZHM = ( n n ) Z se (136)
Determination of ZML. It is evident that

ZML = Zcr (137)
as it involves merely a change in notation.
The results of these transformations are tabulated in Fig.
65(g) for ready reference. The converse equations defining
Zcr, Zsc, and Zrs in terms of ZML, ZnM, and ZuL are also given
in Fig. 65(h). To make the conversions complete, the star
impedances are also defined for the two combinations. These
are tabulated under .(i) and (j) of Fig. 65 and were obtained by
124 SYMMETRieAL eOMPONENTS
substituting equations (g) and (h) in (e) and (f), remembering

tbat ZTs = ZT + Zs, etc.
For tbe above development, the voltage of the C and M
windings has been chosen as the base, and all impedances must
of course be expressed in terms of this base. When any of the
other windings are chosen as base, it is necessary to multiply
all the in:tpedances given by the ratio squared of the respective
voltage bases.
Example. To illustrate the use of these relations consider a three-phase
autotransformer bank whose H, M, and L windings are conncctcd for 220,
150, and 11 kv. respectively. The reactances between the various com-
binations of windings are
ZnM - 9.5 per cent on 35,700 kva.
ZML - 9.2 per cent on 11,000 kva., or 29.85 per cent on 35,700 kva.
ZnL - 14.0 per cent on 11,000 kva., or 45.4 per cent on 35,700 kva.
or reduced to ohms on thc 150-kv. base
_ .(0.095) X 150,000 X 150,000 _ . h
ZH M J
- 35 700 000 - J 59 ·9 0 ma
Z11tL = j188 ohms
' '
ZnL = j287 ohms
The resis~ance of the windings is neglected.
The star impedances, by application of the equations, Fig. 65(!), are
ZL _= ~(188 + 287 - 59.9) = j207.6
ZM = ~(188 + 59.9 - 287) = -j19.6 ohms on 150-kv. base
Zn = ~(287 + 59.9 - 188) = j79.5
If the nature of the problem requires the determination of the star imped-
ances in the S, e, and T combination, they may be obtained by substituting
the above numerical values of ZL, ZM, and Zn in equations Fig. 65(i),
remembering that n = H8 = 1.467.
Zc =
ZT =
1.i6~ ~ ,1 (- j.19.6)
6
119 6
= -j61.6
J"207 . 6 - 1( -467 -· )1 = J"249 . 6

t
_ ( 1. 467 ·) 2 . r::' ~ ohms on 150-kv. base
Zs - 1.
467
_
1
(J79.v 1 +
(1.4~7 ~ 1)2( -j19.6)
4
= j652.6
Similarly, given thc lcakage rcactances between two windings at a time
in the H, M, and L combination, the leakage reactances between winding::,
for the S, e, and T combination may be obtained by substituting the
numerical values of ZHM, ZML, and ZHL, in equations Fig. 65(h).
Zcr = j188
= (1.16i ~
6 2
Z sc 1) (j59.9) = j591
1.46; - 1 (j188) + (1.4~;6~ 1)2()"50.9)

= j902
59. Inclusion of Series Impedance in Autotransformer Circuits.

In sorne cases it is desirable to include series impedances into
•
the equivalent network. It will be assumed that these imped-
ances are expressed in terms of the e
and M winding voltage
base. To be more specific, impedances in T or L windings are
reduced to the e and M base by dividing the actual impedance
in ohms by the square of the ra tio of turns; impedances in the
a conductor are reduced to the e and lvf winding base by dividing
the actual ohms by n 2 ; and impedances in the b ande conductors
are already on the M and e winding turns base. Series imped-
ances in either the T or L windings or the a conductor offer no
difficulty, because they are merely added externally to their
respective windings. Impeda nces in conductors b and e may
require sorne manipulation depending upon the combination
of windings chosen. It is more convenient, for transmission
work, to use the combination H, M, and L, Fig. 65 (b). Assume
then the impedances S a, Sb, and S e in conductors a, b, and e,
respectively, Fig. 68(a). It should be evident that, expressing
the transformer impedances in the H, M, and L combination,
Sa and Sb can be treated as an impedance in series with the H
and M windings, respectively. The impedance Se should be
treated as an impedance in series with the e winding of the S, e,
and T combination, Fig. 65(a). Therefore, the same coefficients
should be applied to this impedance as was applied to Zc in
the equations in Fig. 65(j), in obtaining the star impedances
n-11 n-1
for the II, M, and L combination, namely, , - and -
n n n2
for ZM, ZL, and ZH, respectively. The result of this conversion
is shown in Fig. 68(b).
Tbe most important application of these considerations is
to the determination of the zero-sequence impedance. A form
which the problem sometimes takes is shown in Fig. 69, in
wbich a star-delta connected machine, which may be a trans-
former or synchronous machine, and whose neutral is grounded

through an impedance, is connected to the low-voltage side of the
common winding of the autotransformer bank. · The autotrans-
former itself may be connected to ground through an impedance.
s~..
The tertiary of the autotrans-
SOl f o rme r will be delta-con-
et
-+- ¡-
¡-
- - nected. The problem is to
H
b sb ...r- ¡-
M{;- Se - ¡-¡-r- - l.¡ determine the zero-sequence
-
¡ .....
¡..... impedance to ground of this
combination. The notation
(01)
H and the current flow assumed
are shown in Fig. 69(a). The
single-line diagram for this
combination reduces to that
shown in Fig. 69 (b). In
particular, the proQ.lem re-
duces to that of finding the
..,-\
.
impedance of the H winding
-nSc
S ~IM(ec¡uiv.)'=
with the two other windings
M lc(ec¡uiv.)" 1 c(oduol) L
short-circuited as indicated.
lc(Oicruo) =Ib(oc+uol)- ICII (Oiduc:~l)
Applying the relations indi-
Fw. 68.-Equivalent circuit of auto- cated in Fig. 68(b), the
transformer with series imp edancc.
Winding M used as base. Ratio H 1M equivalent circuit, using the
= n. M winding as base, becomes
that shown in Fig. 69(c) . Now the impedance desired is merely
the impedance of the H winding branch in series with the M
and L branches in parallel, which is
Z 0 -_ z H -
3(n - l)z
n2 n2
+
(zL + ~)[ Zu + Z + 3Znl + a(n;:

0 l)Zn2 ] (138)
( ZL + 3 ~n 2) + [ ZM + Z0 + 3Znl + a(n;: l)Zn2]
This is the value of impedance to use in the zero-sequence

network when using winding M as base. When using the H
winding as base, multiply this impedance by n 2•
The actual currents in the different windings in terms of
the equivalent current in the H winding (the current which one
will obtain from the solution of the complete network) are
developed in Fig. 69. The important idea in this development
is to keep in mind the conversion coefficients from equivalent
to actual currents. It is interesting to observe in passing, as
inspection of the equation for the neutral current in Fig. 69
•
discloses, that the current in the neutral under sorne conditions
is of t he same sign as the high-tension current, under other
Tirfiary
- fo(eq<~iv)• n
nla~fvat)f - ;; JZn2
lc(acfm!)•
ZH lb(acluol)
-Ia(ac/uol)
. _ (e;)
L
H Zg
e
(b)
Zo = Zu -
3(n- 1)
nVJ. Z,.2 +
( ZL + 3:" 2
) [ zltf + + 3Znl +
Zo
D.
3
(n;: l)Zn2]
in which
~ = ( ZL + 3:"2) + [ ZM + Z + 3Z,.t + 3 (n n- l)Zn2]
0
J 1u (equiv.)
a(actual) = 1 ll(uctual) = n
l n1(actual) = 3h(actual) = 3/ltf( equiv.)
_ JH(.quiv.)
1n2(actual) = 31c(ac tual) = 31 M( equiv. ) 3
n
3[(n - l)ZL - Zu - Z 0 - 3Znd¡
= nb.. H(equiv.)
Fw. 69.-Zero-sequencc impedancc and current distribution for autotrans-

former with M winding connected to star-connected grounded machine. Neu-
trals of stars grounded through different impedances. All impedances expressed
in terms of M winding.
conditions is of opposite sign, and under certain other conditions

is zero. This is, of course, due to the fact that two paths are
provided in the transformer for the fl.ow of current, these paths
causing current of opposite signs to flow through the transformer
neutral connection. Consequently, the actual current will have
a magnitude and sign which depend u pon the relative impeda.nce
of the different paths. This is an important point in connection

with certain relay schemes.
Ungrounded Autotransformer with Tertiary Winding. The
ungrounded autotransformer with delta-connected tertiary wind-
ing may be considered as a special case of the foregoing in which
7: 1 :::35% on35,700Kv~a.
j220.5 j 63 j50 j 59.9

Posifive·seqvence
(b)-....MM{Wo-----"WV>NWv-----'V'v'IWWv----'VWNVN--- nefwork
~j""'15M7.1N5'l'-------'>lvWjM63'1'v----...¡.jÍIV'5MONVv--~jM5M9.""
(e) ....
9
,.___ Neqcdive·seqvence
nefwork
Generafor Transformer Line Aufofransformer
H
-j13
-'19.6
~ 1 3(j20), J
jl13
M L
(e) (f) (g)
--
(cA)
@
--
-68.1
-68.8
' 0.7
Acfucrl zero-sequence curronf disfribufion
(n)
Fw. 70.-Single line-to-ground fault calculation. Quantities in the circles
represent the same positive sense of curr ent flow as in positive- and negative-
sequence networks.
Zn2 becomes infinitely large. The value of Zo for this case may
be obtained by converting the expression for Zo to a simple
fraction and then letting Zn2 = oo, thus obtaining
Zo = ZH + (n- 1)2zL + ZM + Zo + 3Znl (139)

n n2
Using the relations given in Fig. 65 this can be reduced to
z, = ( n: l)'z•• + z, ~,3z.. (140)
It will be recalled that the impedances are expressed in terms
of the M winding. To transform to the H winding turns it is
necessary to multiply through by n 2, after which it will be
observed that the impedance consists of two parts, that due to
the impedance between the S and T windings and that due to
the generator and the neutral impedance, the latter impedances
being their actual impedances since the same current flows through
the generator and neutral impedance as flows through the high-
voltage line.
Example. To illustrate the application of these relations to a practica!
problem, let it be desired to calculate the current distribution for a single
line-to-ground fault on the 220-kv. circuit for the system shown in Fig. 70,
consisting of a generator and two transformers connected by a section of
150-kv. line. The autotransformers will be assumed to have the same
characteristics as those illustrated by thc example under Autotransformers,
namely,
ZL = j207.6¡
ZM = -j19 .6 ohms on 150-kv. base
ZH = j79.5
¡ The positive- and ncgative-sequence networks for this system consist of
¡
f
merely the generator, transformcr A, the line and autotransformer imped-
ances, reduced to a 150-kv. base, connccted in series. The autotransformer
f
r impedance for this case is the impedancc ZHM (59.9 ohms). Thc impedance
1
~· diagrams for the two networks are thc same exccpt for the generator imped-
t ances which are diffcrent for the positive- and negative-sequences. Mter
.r adding, these impedances become j393.4 and j330.4 for the positive- and
negative-sequences, respcctively.
The zero-scquence diagram reduces to the same kind as has just been
under considcration, and which is shown in Fig. 69. The series impedance
Z 0 of Fig. 69 in this case equals the sum of the line impedancc j50 ohms
(which is already on the 150-kv. base) and the transformer impedance j63
ohms. Since transformer A is solidly grounded, Z nl is equal to zero. The
neutral impedance Zn2 is equal to j20 ohms. The equivalent diagram of
Fig. 69(c) with thesc values substitutcd is shown in Fig. 70 (d), which through
successive simplification reduces to j143.8 ohms.
The sequence components of current n.t the point of fault (on the 150-kv.
base, of course) are then equal to the line-to-neutral voltage divided by the
sum of the three impedanccs, namely,
10 - 1 1 - l - jl50,000 99 8
.y3"j(393.4 + 330.4 + 143.8)
2
- - - = · amp.
A word of cau tion is necessary at this point. The theory, u pon which the
method of symmetrical components as applied to single line-to-ground faults
l.
!
1
was developed, presupposed that the positive direction of current at any
point in the system is the same for all three sequences. It will be observed
that in Fig. 70 (a) the positive direction of current in the 220-kv. line, for
example, was taken from left to right, but in Fig. 69 and Fig. 70(d) the
•
positive sense was taken oppositely. This difference in the positive sense of
current flow, which was arbitrarily chosen, does not alter the value of the
zero-sequence impedance and consequently the value of the three sequence
eomponcnts of current at the point of fault, but it does mean that the curren t.
TH (Fig. 70(d)) is the ncgative of the zero-sequence current at the fault or
-99.8 amp. The zero-sequence current in the M winding [terminal b,
Fig. 69(b)] is then
248.4
112 .5 + 248 .4 ( -99.8) = -68.8 amp.
1 IH <equiv.) =
f Ia(~quiv.) =
nla(actual)
( 01)
(e)
3Zn1
Zo
ln (equiv.)
I a( actual) = I H (actual) = n
3Zn2 lu( equiv.)

Jnl( actual)
Znt + Zn2 n
3Zn l ln( equiv. )
J n2( actual)
Z nt + Z n2 n
Fw. 71.-Zero-sequence impedance and current distribution for autotrans-
iormer with tertiary, connected toa star-connected grounded machine. Neutrals
of sta,rs tied together and grounded through different impedances. Al! imped-
ances given in terms of the low voltage (M) winding of autotransformer. (a)
System lay-out; (b) single-line schematic diagram; (e) equivalent circuit.
This is the actual zero-sequence ~omponent of current in the 150-kv. line
in a direction toward the generator and in the neutral of transformer A in a
direction into the ground. The total neutral current in transformer A is
then 3 X 68.8 or 206.4 amp. out (because of the negative sign) of the ground.
The actual current in the grounding reactor is
3(/M(actual) - /Ji ( actual)) =
a(- 68.8 - ( ~!:;,

8
))J = 0.7 amp.
in a direction up from the ground. The actual current distribution in the
zero-sequence network is shown in Fig. 70(h), the values in thc circles
corresponding, in the positive sense of current fl.ow, to that of the positive-
and ncgative-sequcnce networks. The phase currents at any point can be
obtaincd from the sequence components by the usual methods.
In Fig. 71 is shown another combination which is sometimes

met in practice. The figure is self-explanatory in depicting
the development of the zero-sequence network corresponding
to the circuit condition given.
60. Interconnected Star-delta Transformers.

An interesting case involving three-winding transformer con-
siderations is the intcrconnected star-delta transformer connec-
tion. The usual method of designating this connection 1s
shown in Fig. 72(a), but to aid in the
~(a) 6
analysis the complete wiring diagram
is shown in (b), which also shows
the nomenclature which will be used.
6,----+--,~ 1 td
It will be assumed that equivalent
star impedances for the three wind-
ings of each transformer are given,
designated as ZA for the delta-
connected winding, Zn for the star
portian next to the high-tension side,
and Zc for the star portian next to
the ground side. This notation will
be clenr from Fig. 72(c) , which shows
the diagram for the left-hand trans-
former. The current flow through- Fw. 72.- Interconnected star-
out the en tire bank is shown in (b), delta transformer.
from which it is seen that current lx flows out of the Zn branch
and - l z out of the Zc branch in (e). With these currents flowing,
the following equations may be written:
(141)
and
(142)
Similar expressions may be written for the two other transformers
giving
E2 = Eb - ZA(J ti - 1 z) - Znl ti (143)
v2
= Eb- ZA(ly- lz) Zolx + (144)
and
E 3 = E e - Z A(l z - J y) - Z BJz (145)
Va= Ec- ZA(lz- 1 11 ) +
Zcl 11 (146)
Now from Fig. 72 (b)
Ex = E1 - Vz (147)
and substituting the above values of E1 and Vz
Ex= Ea- Eb- Z A(2l x - l z- ly) - ZBl x- Zclx (148)
Similar expressions can be obtained for the two other phases,
but because of syrnmetry all of the necessary information can
be obtained from equation (148) alone.
First let it be desired to obtain the positive-sequence imped-
ance. For this case
Eb = a 2Ea (149)
111 = a 2I x (150)
lz = al X (151)
substituting
Ex = (1 - a 2 )Ea - ZA(2 - a - a 2 )l x - Znl x - Zcl x
= (1 - a 2 )Ea - (3ZA + Zn + Zc)l x (152)
This equation indicates that on the basis of equal turns on all
windings when no current is fiowing, namely, when I x = O, the
star voltage on the interconnected star side is (1 - a 2 ) times the
line-to-line voltage on the delta side, and when current flows
the star voltage on the star-connected side decreases by the
amount (3ZA + Zn + Z c) per ampere fiowing in that side. It
follows, therefore, that, given the transformer impedances
ZA, Zn, and Z c in ohms on the voltage base corresponding to the
operating voltage across the individual windings of the two in
series, the quantity (3ZA + Zn + Zc) is the series impedance in
ohms per phase for the positive-sequence when using the voltage
of the interconnected star side as base. A concrete example
which will be worked out later will serve to illustrate these
relations.
The apparatus being non-rotative the negative-sequence
impedance is equal to the positive-sequence impedance.
For the zero-impedance, all similar voltages and currents in
the three phases are equal, so that letting
Eb =Ea (153)
l 11 =lz=lx (154)
and substituting in equation (148) gives
Ex= Ea- Ea- ZA(2lx- lx- lx) - ZBlx- Zalz
= -(Zn Zc)lx +
= -Znclx (155)
The impedance Zl:Jc in ohms on the voltage base of each of the
individual windings, therefore, represents the impedance per
phase for the zero-sequence. It is interesting to observe that
the impedance of the delta winding does not enter this relation.
This results from the fact that the two sections of the inter-
connected star winding were assumed to have the same number
of turns, and since equal and opposite currents fl.ow through these
windings because of the series relationship, the m.m.f. conditions
within the transformer are satisfied and no current flows in A
branch, Fig. 72(c).
Example. In order to illustrate the calculation of the sequence imped-
ances of an interconnected star-delta transformer, refer to Fig. 72(b) and
assume a bank of three 2,000-kva. transformers with Ea as 11 kv. on the
delta side, and E1 and V 1 as 22 kv., each corresponding toa line-to-line volt-
age of 66 kv. on the interconnected star side. Further, assume that trans-
former resistances may be neglected and the reactances are as follows:
X AB = 15 per cent on 2,000 kva.
X Ac = 15 per cent on 2,000 kva.
Xnc = 10 per cent on 1,000 kva. = 20 per cent on 2,000 kva.
The equivalent star impedance diagram for tbe single-phase transformer
may be calculated in the usual manner, using the notation of Fig. 72(c).
ZA = J.( 15 + 15 - 20) per cent = +J.5 per cent

2
Z B = J·(15 + 20 - 15) per cen t = +J'10 per cent
2
Ze = J·(20 + 15 - 15) per cent = +J'10 per cent
2
The above impedances may be converted into ohms at 22 kv. (line-to-
neutral) with the following results:
2
ZA . 5 (22,000) '12 1{
= JlOO 2 000 X 1 000 = J .
z = ._!Q ' (22,000)' 2 = .24 2) ohms at 22-kv. (line-to-neutral)
B JlOO 2,000 X 1,000 J .
Zc = j24.2
The J?Ositive- or negative-sequence reactance of this path of three 2,000-
kva. transformers may be computed from the reactance of tbe single-pbase
transformer by the use of tbe expression
(3ZA + ZB + Zc) = 3(j12.1) + j24.2 + j24.2 = j84.7 ohms
This is the proper vp1ue of reactance for the 38-kv. line-to-neutral voltage
corresponding to a 66-kv. line on the interconnected star side of the trans-
former. The impedance of this transformer on the 11-kv. side can readily
be obtained by taking thc ratio of transformation into account, with the
following result:
j84.7(!!:~~~)
2
= j2.35 ohms.
The zero-sequence impedancc of thc three 2,000-kva. transformers from
the interconnected star side is
Zsc = j48.4 ohms
which is, of course, the value to be used with the zero-sequence voltagc to
neutral of 38 kv., corresponding toa line-to-line voltage of 66 kv.
Problems
1. Determine the equivalent star impcdancc diagram for a bank of three-
winding single-phase transformers whose connections are as follows: high-
voltage winding, groundcd star, 66 kv. between lino conductors; low-voltage
winding, grounded star, 11 kv. bctwecn line conductors; tertiary winding,
delta, 2.2 kv. The reactances between transformer windings for the bank
ratings are as follows: H- L, 10 per cent on 30,000 kva.; H-T, 6 per cent on
10,000 kva.; L-T, 14 per cent on 15,000 kva. Also determine the reactances
in ohms on: (a) 66-kv. base; (b) 11-kv. base; (e) 2.2-kv. base.
2. If the transformer of Prob. 1 is supplied by a generator having a
zero-sequence reactance of 10 por cent on 30,000 kva., what is the zero-
sequence reactancc of the combination as mcasured from the 66-kv. side?
What is the corrcsponding reactance when the gcnerator is disconnected?
3. Assume an autotransformor for connection bctween the 220-kv. and
110-kv. linos, with an 11-kv. dclta-connected tertiary. If all the reactances
are reduced toa 20,000-kva. base, they may be expressed as follows: H- M,
10 per cent; H-L, 50 per cent; L-M, 20 por cent; as given in Fig. 65(b).
Determine the reactanccs for the equivalent star diagram. Find the zero-
sequence reactance as viewed from (a) the high-voltage winding and (b) the
medium-voltage winding, assuming transformer neutral is solidly grounded.
4. If the transfor,mer of Prob. 3 is grounded through a résistor whose
zero-sequence resistancc per phase is 30 per cent on the 20,000-kva. base,
find the equivalent circuit for zero-sequence.
6. A power system is supplied by a generator through a bank of delta
grounded-star step-up transformers and a three-phase transmission line.
At the receiving end, a bank of three two-winding transformers is connected
in grounded-star on the transmission side and the secondary windings are
kept separate. If one of the transformer secondaries is short-circuited, cur-
rent will fiow through the ground connections. If the sccondary windings
are connected in delta and a short-circuit is placed on one phase as before,
will current fiow through the transformer neutrals? Explain your answer
in terms of zero-sequence. Also give the explanation in the terms of the
single-phase solution.
6. A system is supplied by a grounded generator through a circuit-
breaker to a transformer which has a grounded-star low-voltage winding, a.
delta-connected tertiary winding, and a grounded interconnected-star
high-voltage winding. It is assumed that the component windings of thc
interconnected star are symmetrically coupled with the delta-connected
tertiary and the grounded-star low-voltage windings. What is the zero-
scquence impedance of the system as viewed from the high-voltage tcrminals
with the circuit-breaker closed and also with the circuit-breaker open?
Will current flow through the generator neutral?
7. Two parts of a system are supplicd by delta-connected transformers
and are connccted through a zigzag autotransformer. This autotransformer
employs an interconnccted-star connection, and the intermediate tap is
taken off the junction points of the two windings. Does this autotrans-
formcJ· ground the system on both sides of the transformer? How is the
answer affected by transformer construction whether made of single-phase
units or three-phase core-type units.
CHAPTER VII
CONSTANTS OF SHORT TRANSMISSION LINES
WITHOUT GROUND WIRES
For the purpose of determining their characteristics trans-
mission lines may conveniently be divided into three general
groups, (a) short lines without ground wires, (b) short lines with
ground wires, and (e) long lines. The treatment of long lines
di:ffers from that of short lines only because of the effects of
distributed capacity. The present chapter is concerned with
the first classification only, after which the two other classifica-
tions will be considered in succeeding chapters.
61. Positive- and N egative-sequence Impedance of Single
Circuits.
The impedance of symmetrical non-rotative apparatus is the
same for both t he positive- and negative-sequences. For short
transmission lines, in which the distributed capacitance may be
neglected, t he impedance is equal to R + jX, in which
R = totalline resistance = rl.
X = total line reactance = xl.
l = length of line.
r = resistance per unit length óf line, single conductor.
x = reactance per unit length of line.
The resistance r and the reactance x may be obtained from
tables which take into consideration the effect of stranding,
twist, skin e:ffect, and effect of steel cores. Such tables are
given in the Appendix.
In determining t he reactance of symmetrically disposed three-
phase conductors, the flux linking the conductors is integrated
from the center of the conductor to the center of t he two other
conductors. For convenience t~üs integration may be divided
into two parts: (1) t hat including the flux within the conductor
and also that externa! to the conductor to a radius of 1 ft.,
and (2) that including the flux between a radius of 1 ft . and the
center of the two other conductors. W. A. Lewis has made use
136
SHORT TRANSMISSION LINES WITHOUT GROUND WIRES 137
of this conception in the preparation of tables of reactances by

which the reactance of any conductor for any spacing may be
obtained by adding the reactance for the particular conductor
for 1-ft. spacing and the reactance corresponding to the flux
between 1 ft. and the actual spacing.
For properly transposed unsymmetrically spaced conductors
the effect of the dissymmetry may be taken into consideration by
introducing the concept of the geometric mean of the three
spacings. This quantity is usually abbreviated G.M.D. and for
three-phase circuits is equal to the cube root of the product
of the three separations. This conception will be discussed
in more detail in connection with the determination of the zero-
sequence reactances.
N eglecting "skin effect," the flux within a conductor per
unit current is dependent only upon the configuration of the
cross section and can therefore be represented by a constant K.
The reactance due to the externa! flux within a radius of 1 ft.
is equal in ohms per mile at 60 cycles to 0.2794 log !,a where a
is equal to the radius of the conductor in feet. By adding the
reactance due to the interna! flux, the reactance due to all flux
within a radius of 1 ft. is
0.2794(log"~ +K)
or
0.2794[ log10~ + log10 (10)K J
or
IOK
0.2794 log1o -
a
The expression ~K is commonly called the geometric mean radius

1
and is designated by the abbreviation G.M.R. Observe that
it is a fictitious radius which when inserted in the logarithmic
term enables one to include the effect of both the internal flux
and the external flux within a radius of 1 ft. For an infinitely
thin tu be the G.M.R. is ·equal toa and for a round homogeneous
conductor, to 0.779a. Table VI gives the G.M.R. for the more
common strandings and conductor patterns met in practice.
Tests have shown that for multilayer A.C.S.R. conductors the
TABLE VI.-GEOMETRIC MEAN RADII* AND D ISTANCES

Solid round conductor. . . . . . . . . . . O. 779a
Full stranding
7 ... ... .. . ........ . ....... O. 726a
1.9 .. . .... . .. . .............. O. 758a
37 ... .......... . .... ....... O. 768a
61 ....... ......... .... .... . O. 772a
91 ......................... 0 .774a
127 ..... ... .. .. ........ .. ... O. 776a
Hollow stranded conductors and
A.C.S.R. (neglccting steel strands)
30-two layer. . . . . . . . . . . . . . . . . . O. 826a
26-two layer ........ . ......... O. 809a
54-three layer .... . . .. ...... . . 0.810a
Single layer A.C.S.R. . . . . . . . . . . . O. 35a--o. 70a
Point within circle to circle. . . . . . . a
Point outside circle to circle. . . . . . distan ce to center of circle
Rectangular section of sides a and
{3 . . . • . • . . . . • . . . . . • . . • . . . . . . • O. 2235 (a + {3)
Circula r tube
tOO
J
0.95 J
V
1:5 0.90
/
-;?
~
V
0 0.85 V
V
0.80
V
V
1---V
0.750 2
O.t O. 0.3 0.4 0.5 0.6 0.7 0.8 0.9 tO
. inner raolius
RCl t 10 oufer raolius
a = ! outside diamcter
* The G.M.R. of standard conductors are given in thc impedancc tablcs of the Appendix
steel strands have an inappreciable effect upon the reactance so
that the G.IVI.R. is dependent only upon the configuration of the
aluminum strands. For single-layer aluminum over steel this
is not the case, and the G.M.R. varies considerably with the
different conductors and also varies somewhat with the current
flow.
Thus the total effect of unequal spacings and conductor strantl-
ing may be represented conveniently by means of the formula.
x = 0.2794 logto ~:Z::~: ohms per mile per phase at 60 cycles (156)
Both of these dimensions must be expressed in the same units.
62. Positive- and N egative-sequence Reactance of Parallel
Circuits.
Due to the close proximity of parallel circuits, especially
those on the same tower, transposition will not entirely eliminate
the effect of mutual inductance between circuits. For the usual
transposition and the configuration shown in Fig. 73 the reactance
of the paralleled circuits is
X = 0.2794[! log¡o V dabdbcdca
2 G.M.R.
_ _!_ log daa' 4 db6 12 ] ohms per mile per phase (157)
1
12 12
lO dab dca dac' dba
12
at 60 cycles
in which t he distances refer to distances between conductors in
the first section.
The first term is merely t he reactance of t he combined cir-
cuits neglecting mut u al effects in which .,Ydabdbcdca is the G.M.D.
or the geometric mean scparation of the
cond uctor s. The second represents the
correction factor and may reduce the reactancc
lb ~--;;t-----'~-'=--70 b
3 to 5 per cent. The formula assumes
transpositions in t he order abe, cab, bca, c'b'a',
b'a'c' and a'c'b', thus pcrmitting each con- o
ductor to occupy each of the six positions in " First
turn. The formula also assumes symmetry g Section
o
about the vertical axis but not necessarily e ~·
a bout t he horizontal axis. o
e
o
b'
o Second
01
63. Zero-sequence Impedance. o
Section
o
b e'
The zero-sequence impedance of a trans- o
b
mission line is of an entirely different character 0 o,
Thi rcl e
from either the positive- or the negative- e Section
sequence impedance. By its very nature it ~ b'
involves the impedance to currents that are FIG. 73.-Configu-
ration and u sual order
in phase in the three conductors, which neces- of transposition of
sitates a return path either in the earth twin circuits.
orina neutral or ground wire. For the important case in which
the ground constitutes the only return path, the problem involves
•
the determination of the current distribution in the earth.

This problem has been attacked, among others, by Rudenberg, 0 37 )
Mayr* and Pollaczekt' in Europe, and by Carson<32) and Camp-
belU in this country. Carson and Pollaczek considera conductor
parallel to the ground and assume the earth to have uniform
resistivity and to be of infinite extent. No other assumption is
used. Rudenberg assumes the conductor to be placed at the
center of a hollowed
cylindrical depression
in the earth's surface
of radius equal to
the height of the con-
ductor above ground
and that the earth
has a uniform re-
sistivity. Mayr uses
a radically different
assumption-that the
ground current flows
in a very thin stratum
next to the surface
of the earth. His
Fw. 74.- Current flow in determination of self and a r g u m en t for this
mutual impedances in ground return circuits.
assumption is that on
penetrating the earth's surface solid granite is soon reached.
Current cannot penetrate into or beyond this stratum and must
therefore flow next to tbe surface. Carson's formulas will be
considered in detail.
64. Carson's Formulas. ,
The problem will take two phases: :first the determination of
the self impedance Zu of conductor a, Fig. 74, with earth return
(the voltage between a and earth for unit current in the con-
ductor), and, second, the mutual impedance Zum between con-
ductors a and b with common earth return (the voltage between
b and earth for unit current in a and earth return). Reference
*Die Erde als Wechselstromleiter, E T Z. vol. 46, pp. 1352- 1436, Sep-
tember, 1925.
t Elekt. Nachrichten-Tech., vol. 3, p. 339, 1926.
¡Mutual Impedance of Grounded Circuits, B ell System Tech. Jour., p. 1,
October, 1923.
should be made to Fig. 74 for the geometry of the system. All
distances will be measured in feet, including the radius of the
conductor, and e will be measured in radians. The conductor
will be considered of such length that the end effects may be
neglected. Let
p = resistivity of earth in ohms per meter cube.
f = frequency in cycles per second.
Te = resistance of conductor in ohms per mile.
a = radius of conductor in feet.
w = 2nf.
Carson in his development shows that the self and mutual
impedances for ground return circuits are equal to the self and
mutual impedances for a perfectly conducting earth plus a term
t involving P + jQ for both impedances, which is a function of the
¡
f variables p and e, which will be defined. This function is the
same for both the self and mutual impeda nces, but the variables
p and e are different. Carson then develops formulas equivalent
to the following
l. Self impedance of ground return circuits = Zu.
e= o.
p = 1.713 X l0- 3h1~~ = 8.565 X I0- 4 Daa-J! (158)
Z 0 = Te + j0.004657f log10 G.:.k. + 0.004044f(P + jQ) in ohms

per mile (159)
2. Mutual impedance of ground return circuits = z0 m
e= sin- 1 ~ = tan-1 X (160)
Dab h1 + h2
p = 8.565 X 10- 4Dab jp (161)
Z0 m = j0.004657f log10 Ddab

ab
+ 0.004044f(P + jQ) in ohms per
mile (162)
For the evaluation of P and Q, Carson has derived formulas
for three ranges of the values of p as given in Fig. 75. For the
intermediate range (0.25 < p < 6.0) the curves may be used,
and for the highest (p > 5.0) and lowest (p < 0.25) ranges the
series given in the figure may be used. For most power-system
calculations, p falls within the lowest range. To form an idea
0.40 General Formulas:

0.38 Zu = r. + j0.004657j log1o G.~~R.
0.36 + 0.004044j(P + jQ)
0.34 ohms per mile
0.32 Z0 m = j0.004657j log 1o Ld·\ ,.b
ab
0.30 + 0.004044j(P + jQ)
0.28 ohms per mile
026 For Z0 , p = \[?Daa
8.565 X 10-4
p
0.24
0.22 For Z11 m, p = 8.565 X 10- \[?Dab 4
p
po.zo Dry earth, p = 1,000
018
Damp earth, p = '.1.00
Sea water, p = 1
0.16 f--'\~~.__-......--,
...~
0.14 r-----~----r-~~~~/~.~~---,
0.12
010
0.08
006
0.04
0.02
o
o 2 J 4 5 6
p
For p = 0.25 to 5, use
curves.
For p < 0.25:
1
P
71'
= - - =-P cos () +
8 3V'2
f~ cos 2e(0.6728 + logE ~)
p2 .
+ 16 e sm 2()
1 2
Q = -0.0386 + 2 logE P+
1
a-yf2P cos ()
of the maximum probable value of tbis quantity consider the
case in which f = 60, h = 50, and p = 10, for which
p = (1.713 X 10- 3 )(50)~

= 0.22
For lower frequencies, smaller heights, and larger values of
earth resistivity, p will be still smaller than the above value.
An important simplifying approximation, which ordinarily
introduces but a small error, is to neglect the terms in the series
for (p < 0.25) which contain e. The errors introduced by
this approximation will be discussed in the next section. N eglect-
ing tbe e terms, p and Q become
7r
p = 8= 0.3927
Q = -0.0386 + 21 1o~ p
2
and since ( -0.0386) is equal to 1.151 log10 1.~79

2 1.853
Q = 1.151 log10 l.079P = 1.151 log1o - P-
Substituting these values of P + jQ in the expressions for
Zu and Zum (from Fig. 75), there results
Z0 = re + j0.004657f log10 G.::.R.

+ 0.004044!(0.3929 + j1.151 log 1.~53 ) 10
= re + 0.00159j
+j0.004657{log10 G.r::.k. 1.~ 53
]
and substituting the value of p from equation (158), there results
. 2,160~¡.
Zg = re + 0.00159! + J0.004657f loglO G.M.R. ln ohms per
mile (163)
. 2,160-Jj .
Z0 m = 0.00159! + J0.004657f Jogw ·· ··· d 1n ohms per
t"'J
mile (164)
For p > 0.25, the self and mutual impedances of earth return
circuits may be obtained by 'the use of the curves in Fig. 76
which give the increases in resistance and reactance due to the
005 015
1 1 1 V
6,000
1 j V
1 L
1 ILI::l 1 1 1
~ 1
~ ~ ~ 1/
/
004 Q10
Q_ .1l'(. Q¡ 1 V
5,000
~ 1 f::J/ 1/
~/ \1~~} 1~ o/ V
~ ~~ V ~ ,\0
V
\.. 1 IV' ,\\'u V r - 4,000
8. 0.03 QIS 1 1 V\ \\' ~ / V
1 1 \ ~ 1\ \ ~
r/)
E N
"' L / 'S
v ¡..\
V
.. ..
..e ¡f
o "' 1/ 1 ~ V
"V
Q.
a.
·-- -'§
e:
o
e o J '\ \ 1\ >( 0 V .a
E~ ~ 4000 r1
11 1 \
~3
<JV> 1/)
111 1\/ \ '\
!'.."'-.,/
K "" ~
L.
o
"" " ""'

L. Q02 0.10
o 1 1 1/ 1 \ '\ ['-. ........ ~~o.
~_/I:_B =!JO" /
~
..........
~ /\ ......... 2,000
<J
N ./ / / r\ ~ ........ ...........
"' "
/
~6'o. -.. -..
........
o-o~?- ~0~60/
Jr/ 1 'f ~ /
/
\ """ ..................... ¡-.. ...._
0.01 0.05
1 V
---
1 ....._~S·
/ ...._ 1,000
r¡l/1 / / ·""" .................
¡_llv / ........... ~~so.
VI ~
r
- --r--
o o o
o 2 3 4 5 6 7 8 9 10
p
Z = re + j0.2794log G.r;;."a. +
0 (ó..r0 + jtlx0 )
Zom = +i0.2794log Ddab + (tlr0 ,. + jtlx0,.)

ab
where (tlr0 + jtlx 0) and (tlr0 n, + jtlx
are determined from curves a and b 0 ,.)
60
for 60 cycles. For other frequencies use the fictitious value of p = f Pact .. al and
multiply ordinate by f0
·
FIG. 76(a) .-Curves of the in crease in eartñ r eturn circuit resistan ce and react-
ance due to the resistivity of the earth.
departure of earth resistivity from zero. The caption gives the

reactance for p = O, and correction terms are obtained from
the curves. These curves are laid out on a 60-cycle basis, so that
p can be determined from equations (158) and (161) when p
and Daa (or Dab) are given. This is accomplished graphically
j
SHORT 'l'RANSMISSION LINES WITHOUT GROUND WIRES 145
in Fig. 76 by using the right-hand scales and the straight lines
labeled for the corresponding values of p. Mter p is determined,
l:lr0 or D.x(J is read from the curve for e = O and the determined
values of p. Similarly, A1'gm and D.Xgm are used from the curves
labeled for the appropriate angle.
Olb
()~ . . . . ....
0.07 0.3 5 fí r··

Daot
DOib
\
~wtvm~w::~
\1 1 \
\
006 0.30
\ 1\
\ l\
1 7,000
1-,:.. \ 1\ \ 1 V
~--~
(;:)
\ l/ 1 /
~
" \f\ 1
005 0.25
~l
,, \ \ s:s .el 1
~r \ 1y ' 1'< 6POO
1¡
il.L
1/ '1/\ 1\ IY ~'¡
1/ \
1 V
spoo
1\/ 1\ X \ 1\ 1/
j _ 17
1
1 S::,
1'7
~ ¡........:
k-
1\
~
k1\ 1/
/
9\~ k:: 1--

+-
Q)
tf
1/ ~~ \ 1/ \ \I/ 1~ 4,000 e
1 \ l'v 1'\ V .a
1/ 1 \ lj 1\ .'\ ~ V es
1.. o
o \ / K 3,000 ~
:f
~d
1 \- f- X ""'
t--.... . 6"-0•
~' lS
<1 0.02 0.10 J
lf~ t
~.lo-;;-.
¿'\ V\ ". " ~"''S?st'--.. .........
~ '> • ¡:::::::: 2,000

Hl~ ./ 1'-.... ¡-.....
k\ /'\ 1'""'61.,.6'r o t-- t-.
001 0.05 rr lC' 1\ x ['-_ ¡--. ¡-.....
1[1 ~ ~F? / .........

~<-->s· --~---
1,000
'/ 1 ;~~ J-- t--
1/F;r/ \~ =¡.:.::::.: f-.
t--
V/1,/ V<.Po
1--0
o o o
o 2 3 4 5 6 7 8 9 JO
p
Fw. 76(b).
It will be observed from equations (158) and (161) that p

varíes directly with v] and inversely as VP· Since the curves
are laid out for f = 60, they may be used for other frequencies
by using a fictitious value of p equal to ~O times the actual
value. The increment in impedance varies directly with the

frequency so that it is necessary to multiply the ordinates by [ .
0
65. Physical Conception and Factors Affecting Earth Retum
Circuits.
A very useful physical conception for use in the analysis of
earth return circuits is that of concentrating the current in a
fictitious return conductor at sorne considerable depth below
10.000
V ,.,., ¡;....-
/' ~
V ~
.,....!-" ~
,.,., ~~~
/'
V
~ ~~~
~ 1.000
,...!-' ~
rJ
(\)
~
e: ~
~
,....... V: V "'
~
V"
/'
~
............
"'
'
4 6 10 20 40 60 100 200 400 600 1,000

p
Fw. 77.-Equivalent depth of ground return circuit as a function of the earth
resistivity p in meter-ohms and frequen cy.
the outgoing conductor. The ~quivalent depth of earth retum *

De is the distance from the outgoing conductor to the fictitious
conductor and is the value to be used in the equation
x = 0.004657/ log1o G.~~R. ohms per mile (165)
This conception has been used for sorne time with good results
because the equation takes the same formas that of the simplified
form of Carson's. Thus, by equating the logarithmic expressions
of (163) and (165), there results
De = 2,160~ (166)
* The distance to the equivalent earth plane assuming infinite conduc-

tivity is half of D•.
l
The values of De in feet for different values of p and f are given in
Fig. 77.
]
o
~
zoS
Q)
...d
E-4
Earth Resistivity. During the past few years considerable
data have been collected regarding the values of p met in prac-
tice. Most of these data have been obtained by measuring the
self impedance of a ground return circuit of known geometrical

configuration and determining the equivalent p which when used
in Carson's formula gives the same impedance as that measured.
The map shown in Fig. 78 has been prepared as a result of work
carried out by the Joint Research and Development Subcom-
mittee of the N.E.L.A. and the Bell Telephone System. <79 >
Table VII gives the resistivity for various constituents of the
earth's surface. Comparison of the values of p from Table VII
TABLE VII.- EARTH RESISTIVITY IN METER-OHMS
Average of large number of determinations* .. ... 100
Sea water............ . ................ ..... . O. 01- 1. O
Sw::tmpy ground............. .. ............ . ... 10- 100
Dry earth . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 , 000
Pure slate.. . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107
Sandstone...... . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109
* See insert, Fig. 77.
and the insert of Fig. 78 shows that the average value corresponds
to clamp earth. This comparison also shows that the resistivity
of the earth is not uniform, a result to be expected from t he
presence in sorne localities of coal or mineral veiris, surface or
subsurface strearns, railroad tracks, pipe lines or other under-
ground structures. Also it is known that in sorne rnountainous
localities the geological formation consists of a solid granite
base with rnerely a relatively shallow layer of soil. This fact
forrns the basis of an advanced method of ground return circuit
analysis in which the earth is assurned to be cornposed of two
homogeneous layers of different resistivities. The condition
of the soil as influenced by clirnatic changes and seasons also
affects the conductivity. Sorne localities such as deserts and
prairies becorne extremely dry and offer high resistance to current
flow. Then, again, in sorne districts the resistivity of the soil
may be uniforrnly low in the winter and spring, but during the hot
season the soil rnay dry out to a considerable depth, increasing
its resistivity and, in effect, increasing the effective height of
the conductors above true ground. Of course, in such cases it
is recognized that the earth cannot bt considered homogeneous,
but that the conductivity increases with the distance from the
surface. The value of the Carson form of the expression is that
it enables one to evaluate the infiuence of peculiar local condi-
tions and thus to estímate these reactances more accurately.
Current Distribution. The current density in the earth is
greatest at the surface directly underneath the conductor and
decreases gradually as the distance from the conductor increases.
The curves in Fig. 79, showing the current density at thc carth's
surface in terms of the density directly under the conductor,
were calculated for the two more common frequencies met in
practice and for damp earth. It will be observed that at a
distance of 1 mile, the current density has reduced to 6 per cent
for the 60-cycle case and to 9.5 per cent for the 25-cycle case.
For smaller heights the current is still more concentrated. The
significance of these curves is, as Rudenberg0 37> points out,
~lOO
-iS
.3
i
:::>
90
=f
-~.,so
{
Jl10
15 l.
:f
Jj60
1:
~~zca.:C! t!&J·
~
~so
'5
E \\
~4o
.E
8
\'
~ 30 \ 1\
"'
<1)
1\1'\
..
;E
..32o
' ..... ~
1i "'ro.. !"--~ s-,e~
~ No~
~ 1o
e¡
1:
~...
}'r- ~
- 1--:--
! oo '2 :; 4 50 7 B
X- Horizont01l Ois+01nce From Conductor in Thou~omol Fee~
9 10
F10. 79.- Current d ensit y at earth's surface as a function of distance from t he

conductor for p = 100 (clamp earth).
that the ground current for an irregular line, such as indicated in

Fig. 80(a) and (b), will follow the irregularities of the line rather
than a path across country. As a matter of fact, even for a
line such as (b), ground currents of the valu e calculated exist
under the line.
Effect of Resistivity and Height. It may be noted from the
approximate formulas, equations (163) and (164), that for the
range of earth resistivity met in practice and for commercial
frequencies both the self and mutual impedances are independent
of height above ground. It is interesting to note that the
resistance component of the self impedance is independent of

the resistivity of the earth. This paradox is explained by the
fact that at high resistivity the current spreads out over a
larger area and at low resistivity is restricted to an area near the
conductor. Thus the ground resistance is for 25 cycles 0.04,
and for 60 cycles 0.095 ohm per mile.
It should be remembered, however, that the approximate
formulas neglected certain terms including e in t he expressions
for P and Q. These terms
include resistivity, frequency, and
certain geometrical factors (h or
Dab) . The effect of these terms
\b} will be to decrease the resist-
ance as indicated by the curves of
Fig. 81. Except in very rare
cases such as river crossings
Fw. 80.-Ground current follows (which may represent a very
irregularities in the line.
small portian of the total length
of the line under consideration), the highest average height
between towers is about 50ft. For p = 10, the resistance of the
0.10
0.09
..... ¡ ----.¡._:
·6ofyclef
-- r---r--- r--
--
p=/,000
p = /00
-
0.08 .. -p ·lO
p- 1,000
C1l ,;~l.;t--= p~ 100
:= 0.07
l:
lo..
~0.06
- p~to
IJ)
.E
L 0.05
o
e
-~ 0.04 - p=/,000
p = IOO
(
()
e '25 cycles p -10
~ 0.03
.,
·;;
& 0.02
0.0 1
0
0 10 20 30 40 50 60 lO 80 90 tOO
Conoluc+o r Heigh+ in Fea+
Fxo. 81.-Effect of conductor height, frequency, and earth resistivity upon
resistance of ground return circuits. Zero-sequence resistance is three times
the ordinate.
ground return is 0.086 ohm per mile. The use of 0.095 instead
of 0.086 therefore represents an error of 10 per cent. The
equivalent resistance per mile contributed by the resistance of
the conductor re, the tower-footing resistance, and the station
ground will be of the order of 0.1 ohm. per mile, so that the error
in terms of the resistance will be about 5 per cent. In addition
the reactance is usually larger than the resistance, and since
they add at right angles the effect of this error in the resistance
term upon the magnitude of the current is still further reduced.
If the phase position of the zero-sequence current is desired,
the error is of course still effective. In view of the variable
character of the tower footing and station grounds, and also
the departure of the earth from the assumption of uniform
conductivity, this error is quite tolerable.
The effect of neglecting the cos e term of the Q expression
upon the reactance of the ground return circuit may be obtained
by multiplying the cos eterm (3 ~p cos e) by 0.004044/ and
inserting the value of p, giving the term
1.63 X 10- 6hf,Jf (167)
which must be added to equation (163) to give the self reactance
of the ground return circuit and the term
0.82 X 10- 6Dabf,Jf cos e (168)
which must be added to equation (164) to give the mutual
reactance of ground return circui.ts. For the self-reactance
calculations this corn~ction is always less than 1 per cent, but
for mutual reactances in extreme cases, such as p = 10, f = 60,
and Dab = 200ft., the error may become as great as 4 per cent.
Therefore, for practically all power work, expressions (163)
and (164) are sufficiently accurate. For inductive coordination
work the distances and frequencies involved are greater so that
p is frequently larger than 0.25. For this field of application
it is necessary to use the curves of Fig. 75 or 76. The Joint
Development and Research Subcommittee of the N. E. L. A.
and the Bell Telephone System have prepared elaborate curves<79>
for this purpose which will save much time if many calculations of
induced voltage are to be made.
66. Simplified Formulas for Zero-sequence lmpedances.
Using the formulas involving the equivalent depth of earth
return current, the expressions for self and mutual impedance
beco me
.......
~
~
t:s:._ .
1--.. :......... ~t--....
.........
N"""' ~
2r-....: K r-.... t-........: 25 cycl es
~~ r--..1--¡-... ~ r---.;.' V;¡
~ :---- ~ ....... ,.. ~ ~

~~ '"<
:-... ti-- ...... ~
""'f'b-.. ¡-... ~
¡-......_¡--... ~
'
' ~ ¡--..... r---¡...... ~ : t:z;j
a.l 5
:E: '
¡.....¡ :
...... N'- ....... ........_:......... -.:::t- ¡-....._l ~
' ........... ~
L 1 ~ ¡-....._: r-..... N ~~ '-.....;.._
............. ¡......; ¡-..._N .......
~
Q
"' ' ' : N--: ........... -.........! K r-........_ ~: ;;t..
E
.<:::
o ''
: ' : r-- ~~ r--..1-- ~1'--L t:'-1
.S l '
'' '
' ~ ;--.. ¡::::: ..... íi~ ~ .::::--r--.. Q
..,
Q¡
1
i ~
;........ 0 ¡-.... ~ ¡......... ~ 1
a
r;: ' p ~
Cl
i ' ' ' ''' ¡-...._ tt-- ~ ¡......._ !'--¡-...
~ 10,000
+-
u
C5
"'
o:
' .
'
'
'' ''
'
'
.'
'
'
' .......
~!'-.....
........ '
....... ¡-.... ¡...ro-
!"--... :......, 5,000 a
'"ti
~
! '' ' ' ~ r......' ~ 1,000
t:z;j
' r;-.:: r-.....¡......
'' 500
'
' '
''
: N. ~
o.0.5 '
' ; ' : '
' '' : : "-¡-.... ~ 50
100
~
: ' '' : '' '' '
' ~
' ' ; ;-....._¡ 510
: '' ' ' '
'
' : '' : ' ' :
' '
'
o ! : ''
''
Olll 0.01 0.04 O.Oó 0.1 0.2 0.4 0.6 1 2 4 6 10 10 40 60 100 200
Geometric Mean Radius or Separation in Feet
F.w . 82(a).-Zero-sequence self and mutual reactance per phase oí .single equivalent conductor at 25 cycles.
Zer o-sequence resistance = 3(resistance of conductors considered as a group) + 0.12 ohm per mile per phase
am: u -!!u••- 94i<ttFC.w:.. :qe::z;_ ..,,_p_40 $4.%1""""#-. ,+::c:osv..~'194"'!. ... •:-~~~ ------··-r"f"o,_.,.___ .....,._ -- '-'y- -'
V.:¡
!'...... ::t:
o
~~ ~
1-'3
4 N. ~ ~~ ....... 1-'3
.... "-¡::j... ~
~" ~
~
~ r-..: f'~ 1
~
~
......... ""
¡..., ........ N~
~ ... __ ¡..., l........
~
......... r-.~
~ ~ N_
i
Q)
r-....
¡...,_1 ......
1 r--;..: r--... ..........:. r-.,..;: ¡:..... ~~ t'.......,
~3 1 r-.... ~¡... ~~ ~~ ......~ ~~
L. 1 50 cycles o
Q)
0..
1
1
1
~ ~
¡.....,:t:i---
~ ~ ...
~ ~ r:::: r--. ~
~"' ~~ r-.
en 1 1 ~ ~
E .....
~
f' r-.r-. ~ r r-.... ~1'
~
.t::. 1 1 1
o
.S 2
1 1
1
1 """ " ~ ¡.. N. l ........
~~ ~"!'-¡.. ~ r-....
¡....,__ l ........
....... ~'r-o-
Q)
1 1
1 1
~ ,...., 1
]'...
r:::: ~--~ ... ~ ~
,......_
~
e
(.)
1
1
1 1 1 K ¡-....,...., ¡..
~ t"¡-...." ro,-. t--........, p .....
~ 1
1
1 1
1
1
: 1 )'..1"
~~ r-. ~ ....¡.., "'N 10,000 " ~
o
~ 1
1
1 1
1 1 1 " ~ ...........
~ ~""' 1'-~
5,000
1,000
o
e:: ~
1
1 1 1 1 1 "':' ~¡....., ......!' .... ~
"' .... ~ 500 1-'3
1
1
:
1
1
1
:
1
1 1
1
1
1
r-.... c:"r--
¡...!'-
~ 100
50 ~
1 1 t-...... 1""1 o
1 1 1
1
1 1
~
.............. 10 ~
~
1
1
1
1
1 1 1
1
:
1
1
1
1
1
5
t:::1
o 1 1 1 1 ' .
~
0.01 0.02 Q04 Q06 0.1 0.2 0:4 0.6 1 Z 4 6 lO 20 40 60 100 - 200 .....
Geometric Me~n Radius or Sepcmiltion in Feet ~
FIG. 82(b).-Zero-sequence self and mutual reactance per phase of single equivalent conductor at 50 cycles. ~
Zero-sequence resistance = 3 (resistance of conductora 0onsidered as a group) 0.24 ohm per mil e per phase + ¡.-¿
e.n
~
...
~
C11
~
~ ~
5
R: -,..... ..............
~~ ..............
~1 ~ !"'- ¡.. r--.....
~ ..... ¡....., ¡... ~....,...,.
t'-..... 1 ~~ ¡.. r-...._1 ~~

~ 4
::E
~" -1 r--. 1"1111~ ......................¡..2 ¡...._1"¡...
r U:l
¡--..;.::t'-. ¡.. ......... 1 r--.;: ~
--:-
.... r-..... "~ ~
L
1 r"-¡... ~ ~to..,l ......¡...,. r--¡... ·····"···1..···~~ ~
8. 1
1 """" ¡....,_ 1 ~ ~ r....c ........ ~

.,........ ....t....~ 60 cycles ~
1/)
o
E
..e 3
1
1
1
1
.............
1
1 """"
~ ¡....,_
1 !"'- ~ .....
¡... .....
~
~ .........
........t.... ¡....¡...
......
¡... ¡..t....._
~-.......
""3
~
......
; 1 ~ 1 ....¡....
e ¡.....,.""""" ...... ¡... 1 ..... r......--.¡....,_ ~
1 1 1 1 too.
.........._....,......._ p..
Q)
u
T 1 1 1 r--..: 1"'111""" ¡.......,_'i...... r-....!""'¡...., t:""'
e 1 1 1
1 l......
~'
¡..,N ..... j""'oo.~¡... ¡....NN
loo. p ~
~ 2 í r-.... r'-¡.._ ¡......,_~ 1 t.... o
o
{S
Q)
1
1
T
1
1
1
1
1
1
i
1' 1
1
¡... ¡.. 'j'....; ~
t......··•;;;;¡:::
"" t....
¡...
1"" t....:: ,.....,lto..
¡..
~'-.......~
~
10,000
"!';;;¡ 5,000
~
1
0::
1
1
1 1
1
1 0 ::--¡.... ¡... 1,000
~ 500
o
~
1
1 ; 1 1 1 1 1 i""- """""-- t-... ~

1 1 1 1 1 1 1 i """'¡..... too ~ 100 ~
T T 50
~
~
1 1 1 1 1 1
1 10
i 1
1
1
1
1
1 i 'S 5
1 i 1 1 1 1
i 1 1 1 _j_ 1 .L..... "- .1 -
O.oJ 002 O.o4- 0.06 0.1 0.2 0.4 0.6 1 2 4 6 10 20 40 60 100 200
Geometric Mean Radius or Sepamtion in Feet
FIG. 82(c).-Zero-sequence self and mutual reactance per phase of single equivalent conductor at 60 cycles.
Zero-sequence r esistance = 3(resistance of conductors considered as a group) + 0.286 ohm per mile per phase
----
Zo = re + 0.00159!, + j0.004657f loglO G.:;.~R. ohms per

mile (169)
Z0 m = 0.00159f + j0.004657f log 10 ~~ ohms per mile (170)
These equations can also be applied to multiple-conductot

circuits if re, the G.M.R., and dab refer to the conductor or con-
ductors as a group. This point will be amplified later.
Unit zero-sequence current consists of 1 amp. in each phase
conductor and therefore of 3 amp. in the earth. In replacing
the three conductors by a single equivalent conductor, 3 amp.
fiow in the equivalent conductor for every ampere of zero-
sequence current. Hence the corresponding zero-sequence
impedances are three times the above values. Calling the
zero-sequence impedances zo and Zom,
Zo = 3re + 0.00477! + j0.01397f log10 G.~~R. ohms per phase

per mile (171)
Zom = 0.00477! + j0.01397f log10 ~~ ohms per phase per mile
(172)
The resistance re in this formula is, of course, the resistance
of the equivalent conductor and is equal to one-third the resist-
ance of a single conductor when a single three-phase circuit is
being considered; 3re in this case thus becomes the resistance
of one phase conductor.
The substitution of the value of De from equation (166) in
equations (171) and (172) gives the reactance of earth return
circuits. This reactance is plotted for the three frequencies
25, 50, and 60 cycles, in Fig. 82(a), (b), and (e), respectively,
as a function of the G.M.R. or separation for different earth
resistivities. In the absence of definite information regarding
the value of p in the particular geographical location under
consideration, the values corresponding to the heavy lines
;, (p = lOO) may be taken as representative of average conditions.
'
67. Self lmpedance of Parallel Conductors with Earth Return.

So far the self and mutual impedances between single cylindri·
cal conductors with circular cross section only have been con-
sidered. In practice, however, the problem usually arises to
determine the self impedance of three similar conductors con-

nected in parallel. Consider any three conductors, such as
a, b, and e, in Fig. 83, whose geometrical
dispositions are as indicated. With con-
ductors transposed, the current distributes
b
uniformly between conductors, so that·
for a total current equal 'to unity the
current in each conductor is one-third.
The voltage drop in conductor a for the
position indicated is
WA\W~@W'~
Fw. 83.-Notation for Z aa + Zab + Zac
the determination of the 3 3 3
zero-sequence reactance. for conductor b
and for conductor e
in which Zab, Zbc, and Zac are the mutual impedances between
the conductors represented by tbe subscripts and Z aa, Zbb, and
Z cc are the self impedances of the tbree conductors with ground
return. Since conductor a takes all these positions successively,
the drop per conductor is
i (Zaa + Zbb + Z cc + 2Zab + 2Zbc + 2Zca)
Mter inserting the values of · impedances from equations
(169) and (170), this drop, or impedance, since unit current is
fiowing, becomes
Zg = 1[ 3rc + 9(0.00159f) + j0.004657J

{ 3 log 10 G.~e.R. + 2 log10 f: + 2 log1o f~
+ 2 log" ~:}] ohms per mile
Com bining terms
Zg = ~ + 0.00159f + j0.004657flog 10 V" De

(G .M.R.) 3d2abd2bcd2ca
ohms per mile (173)
J
SHORT TRANSMJSSION LINES WITHOUT GROUND WIRES 157
and multiplying by 3, the zero-sequence impedance per phase

becomes
Zo = re + 0.004 77f + j0.01397f
l l)e
10
og v(G.M.R.) 3d 2abd 2bcd 2ca
ohms per mile per phase (174)
The ninth root in the denom-
inator of the logarithmic ·term
is the G.M.R. and is equal to
the radius of an infinitely thin
tube which would have the
same inductance as the system
representcd. Equation (173)
reduces to (169), re in (169)
being the resistance of the three
conductors in parallel, and re in
(173) and (174) the resistance
of one conductor. In deter~
mining the reactance for the
positive- and negative-sequence
impedances the G.M.l). of the Fw. 84. -Configuration of typical twin-
conductors was used. This circuit transmission line.
equivalent separation was 397,500 cir. mils A.C.S.R. - 60 cycles
Conductor radius = 0.403 in.
obtained by taking the cube Conductor G.M.R. (30 strand-two !ayer)
root of the three separations. = 0.826 X 0.403 in.
From equation (174)
G.M.R.eircuit = v(G.M.R.) 3conduetord 2abd 2bcd 2ca
- ..if(G.M .R ·) conduetor(G .M ,1)•) 2 sepa.ration (175)
Example. Let it be desired to determine the zero-sequence impedance
of one circuit of the double-circuit line shown in Fig. 84, assuming damp soil
(p = 100 meter-ohms).
Assume a resistance of each conductor = 0.235 ohm per mile
G.M.D. = ~8.54 X 8.54 X 16 = 10.5 ft~
G.M.R. = {}o.826 X
0
·{g3 (10.5) 2 = 1.45 ft.
De (from Fig. 77) = 2,800 ft.

zo = 0.235 + 0.00477 X 60 + j0.01397 X 60 log1o 2Í~~
= 0.521 + j2.76 ohms per mile per phase
This same result may be obtained by referring to Fig. 82(c) for G.M.R. =
1.45 and p = 100 and picking off the reactance which is equal to 2.76 and
adding the resistance of the conductor to the resistance of the ground return
circuit [0.286 also from Fig. 82(c)].
68. Mutual Impedance between Two Circuits with Earth Return.

The mutual impedance between two circuits with common
earth return may be derived in a similar manner and is found
to be
Zmo = 0.00477! + j0.01397f log¡o a.i!r~D. ohms per mile per
phase (176)
where the G.M.D., in the case of the three-phase circuit, is the
ninth root of the product of the nine possible distances between
the conductors in the two circuits.
Example. Let it be desired to calculate the mutual impedance between
the two circuits in Fig. 84.
De (from Fig. 77) 2,800 ft.
=
G. M. D. = y ...,-
( 1-8,.-
) 2...,.,
(2--=-2---
.5,.....,
) 4--:-:(2,.--,-
4.,--)a
= 21.9 ft.
j0.01397 X 60 log1o 2~~
2
Zmo = 0.00477 X 60 +
= 0.286 + jl.77 ohms per mile per phase
Or using Fig. 82(c) for G.M.D. = 21.9 ft. and p = 100, the mutual reactance
is 1.77. The mutual resistance is 0.286.
69. Self Impedance of Two ldentical Parallel Polyphase Circuits

with Earth Return.
The impedance of two parallel identical circuits is
Zo = ;e+ 0.00477f + j0.01397f log10 _1 De

~ (G.M.ll.)(G.M.D.)
ohms
per mile per phase (177)
in which G.M.ll. is the geometric mean radius of one set of
conductors and G.M.D. is the geometric mean distance between
the two sets of conductors.
Example. To illustrat e the application, calculat e the impedance of botb
circuits in Fig. 84 in parallel. From a previous calculation
G.M.R = 1.45 ft.
G.M.D . = 21.9 ft.
(G::-::
V -:-;: .M
:-:-.-=R--:-
.) -:-::
(G::-::
.M .Do-7
:-::-: .) = 5. 64 ft.
Substituting in formula (177)
zo = 0.4035 + j2.26
This should check against 1(zo + z,.o) from previous calculations.
Zo = !(0.521 + j2.76 + 0.286 + jl.77)
= 0.4035 + j2.26 ohms per mile per phase
For non-identical circuits it is better to compute the mutual
and self impedance for the individual circuits.
70. Equivalent Circuits for Parallel Lines with Mutual Inductive
Coupling.
Frequently it is required to calculate faults at locations
between bussing points on physically parallel circuits. The
coupling between thcse circuits for the positive- and negative-
011
E"
-Yo.
lb
F~vlf
ZOI 101 ZOI
2b
( 01)
EOI EOI
~ 2m El"
Faulf ( b)
Fault at One End Fault at an Intermediate Point
Fw. 85.-Zero-sequence equivalent networks for parallellines after one circuit-
breaker has opened.
sequences is small, and it is further reduced by transpositions.

For the positive- and negative-sequence networks, then, this
effect is negligible. Such is not the case for the zero-sequence,
and transpositions do not reduce the effect. One important
case involving these considerations is illustrated in Fig. 85(a),
in which it is assumed that the breaker at the left bus has already
opened for the fault which is assumed to have occurred just
outside of the breaker. If the problem is such that capacitance
effects may be neglected, either because of the shortness of the
line or because of the accuracy required, the actual network
may be replaced by the equivalent shown in (b). The values
of the constants are given in the figure and the proof is indicated
subsequently in the discussion of the more general case involving
distributed capacitance. For faults at intermediate points the
circuit condition and equivalent circuit are shown in Fig. 85(c)
and (d), respectively.
Another important case is that of a fault on a two-circuit line

at sorne point between buses, illustrated in Fig. 86. The
impedances Za, Zb, and Zm are the zero-sequence impedances
for the whole line. It will be observed that if the two circuits
are divided into two parts at the fault point, each side can be
represen+,ed by the equivalent circuits of the type shown in
Fig. 85. This has been done in Fig. 86(b). For the usual case
Za = Zb = ZL and, using this assumption, the circuit can be
successively simplified as indicated in (e), (d), and (e). The
(1-n)Z a A nZa
<a> (á)
~r. ( Fredion oftot~l

<'Q'< length)
»,)
(b)
orZa= Zb =Zt. (á)

{1-n)Zm ZL-Zm
<l.±.n>¡ +ni
2 Q 2 b
Cun•ent Distl'"ibution
Fw. 86.- Zero-sequence equivalent network for a two-circuit line with a fault
atan intermcdiate point.
current distribution in the faulted and unfaulted lines in terms

of the current input into the network is given in (f).
Example. Let it be desired to determine the zero-sequence network
for a 25-mile double-circuit line whose configuration is shown in Fig. 84.
The self and mutual capacitances wlll be neglected. Further, let the fault
occur at a point 10 miles from the right-hand bus, Fig. 86(a), so that
??. = U = 0.4. The self and mutual impedances for the zero-sequence have
already been calculated for this configuration and were found to be
ro + jxo = 0.521 + j2.76 ohms per mile per phase

rmo + jzmo = 0.286 + jl.77 ohms per mile per phase
Therefore
ZL = (ro + jxo)l = 13.02 + j69.0 ohms
ZM = (rmo + jzmo)l = 7.15 + j44.25 ohms
and
n( 1 - n) (ZL - ZM ) (0.4) <1 - OA) (13.02 + j6u.o - 7.15 - j44.25)

2 2
0.704 + j2.97
SHORT TRANSMJSSION LJNES WJTHOUT GROUND WIRES 161
~(ZL + ZM)
0
=2\13.02 +j69.0 + 7.15 +j44.25) = 4.034 +j22.65
1
(1 ; n) (ZL + ZM) - 0.4(13.02 + j69.0 + 7.15 + j44.25) =
2
6.05 + j34.0
Thesc valucs, substituted in thcir respective positions in Fig. 86(e), thus
givc the cquivalcnt network. The current distribution is givcn in (J) in
terms of the externa! currents.
Problems
1. Assume that the ground-rct.urn-circuit solution given by Carson and
the normal values of earth resistivity apply for radio frequency. Determine
the resistance of the earth return circuit per mile for 1,000 kilocycles. assum-
ing p = 100 mcter-ohms and h = 50 feet.
/Oohms
2. Two diffcrent power stations and thc stcp-up transformcrs are located
at the samc point as indicated in thc accompanying figure. Each generator
and cach transformer has a reactancc of 10 pcr cent on a 20,000-kva base.
Each transmission lino is 40 miles in lcngth and consists of 4/0 stranded
copper conductors spaced triangularly 10 ft. apart and transposed at
intervals. The transmission lino is operating at its no-load normal voltage
of 66 kv. At thc gcnerating station the transformer neutrals are connccted
and then grounded togcther through a connection that has a resistance of
10 ohms. The ground connections for cach receiver transformer has a
resistancc of 10 ohms. Determine thc amount of current that fl.ows in the
various transformer neutrals for a line-to-ground fault at the receiver end
of onc transmission 1ine assuming that thc transmission lines do not parallel
each othcr. The system frequency is 60 cycles per second.
3. Assume the same systcms as in Proh. 2 except that the ground connec-
tions are assumed to be of zero impcdancc, but that the transmission lines
parallel each othcr throughout their en tire lcngth ata separation of 100ft.
Determine the current fl.owing through each of the four transformer neutrals
for the same fault conditions.
4. Assumc that a 33-kv. line forming a ten mile square is paralleled atan
average separation of 50 feet by a 132-kv. system as shown in the accom-
panying figure. The 33-kv. circuit is made up of 266,800 A.C.S.R. con-
ductors spaced 6 ft. delta. If the 132-kv. transmission circuit is opcrated
at 60 cycles and is subjected to a line-to-ground fault which draws 1,000
amp. through the parallel, determine the zero-sequence current fl.owing in
the 33-kv. system.
Explain why zero-sequence current can flow in the 33-kv. lines which have
no ground connections.
6. Assume a source of negligible impedance to positive-, negative-, and
zero-sequence in comparison with the impedance of a 40-mile transmission
line made up of 3/0 stranded copper conductors spaced 10 ft. apart and
symmetrically transposed. Assume that one conductor is grounded at the
far end. Determine thc positive-, negative-, and zero-sequence impedances
of the line by the method of symmetrical components and the line-to-neutral
fault current for a normal voltage of 110 kv. between lines. Determine the
impedance of a single-phase circuit with one line conductor and ground as
t
1
~ t
~
V ~
!Omlles - f--+
Gener()lfor 3JKv.
Generedor 132Kv.
the return and the resultant fault current by the "single-phase" method.
Analyzed by thc single-phase method, it is necessary merely to determine the
supply frequency, the resistivity of the carth, and the conductor material,
stranding, and diametcr. By the method of symmetrical components it is
necessary to determine, in addition, the spacing from the faulted conductor to
the other conductors. Show that the single-phase method and symmctrical
components will give the same analytical expression for both the resistance
and reactance components.
6. T ests were made to determine the resistivity of the earth in a particular
area. A long, insulated cable was laid on the surface of the earth and care-
fully grounded at each end. A test loop was formed by laying an insulated
wire parallel to the cable at a distance of 100 ft. and grounding at one end.
The power cable is su:fficiently long so that end effects may be eliminated
from the test loop. A current of 100 amp. at 60 cycles is caused to flow
through the grounded conductor. The voltage induced in the test loop is
found by test to be (9.5 + j 36) volts. Determine the equivalcnt dcpth of
the earth return current and the resistivity of the earth. Check the induced
voltage by the geometry of the circuit and the value of the earth resistivity
derived from test results.
CHAPTER VIII
CONSTANTS OF SHORT TRANSMISSION LINES WITH
GROUND WIRES
Previous chapters have emphasized that, in the symmetrical
portions of a system, the positive-, negative-, and zero-sequence
quantities are independent of each other; whereas, in unsym-
metrical portions, the sequence quantities are not independent.
When the impedances are unsymmetrical, positive-sequence
currents may produce negative- and zero-sequence voltages,
and similarly for the other sequences. Due to the physical
arrangement of conductors, all transmission lines are unsym-
metrical to sorne extent. Fortunately, the error introduced
by dissymmetry is negligibly small for the arrangements of con-
ductors and ground wires usually met in praetice if the geometric
mean distances (G.M.D.) are used. This phase of the problem
will be illustrated by working out several examples, using both a
rigorous method and the simplified method of symmetrical
components, which assumes that the phases are balanced, and
comparing the results. For extreme cases, investigation of this
error should be undertaken befare using the simplífied method.
The detailed assumptions involved will be specified in the discus-
sion of the determination of the different sequence impedances.
71. Positive- and N egative-sequence Impedance.
The ground wires may be neglected in the calculation of the
positive- and negative-sequence impedance. This assumption
is rigorously accurate for the case of perfectly transposed lines
with the ground wires grounded only at barrel points, and line
faults occurring at barrel points, because the sum of the voltages
induced in the ground wires by positive- and negative-sequence
currents is zero. In practice, however, the ground wires are
grounded at every tower, so that the induced currents in the
ground wires will have a certain definite but negligibly small
effect upon the positive- and negative-sequence impedances
of the line. The error is quite small also for faults at intermediate
points in a transposition barrel.
163
72. Application of Carson's Formulas to the Calculation of Zero-

sequence Impedance.
It wa.s shown in Sec. 65 that, as a result of Carson's work, the
self impedance of a conductor and the mutual impedance between
two parallel conductors, with earth return (neglecting end
effects), can be calculated by equations (169) and (170) and Fig.
75. The corresponding impedances in terms of zero-sequence
-~~WD~ ·'~~~.-..
(01) (b)
Ib Leak01ge impeol01nce
, / 'óf grovnof wire
Leakage ímpeclance -~
of'conolucfor')
. y
Q
¡1
HufuorÍ impeclomce befween
conclucfor 01nt?l grovnd wire
FIG. 87.-Equivalent network for two inductively coupled aerial circuits with
ground return.
(a) Physical arrangement; (b) equivalent circuit of two acrial condu ctors with
ground return; (e) equivalent circuit whcn one conductor is grounded.
Do = equivalent dcpth of ground rcturn
G.M.R. = geometric m ean radius
For 60 cycles 0.004057/ = 0.2794
For zero-sequence constants, multiply by 3.
are given by equations (171) and (172) and Fig. 82(a), (b ), and (e)
for 25, 50, and 60 cycles, respectively.
The disposition of two such conductors is shown in Fig. 87(a).
Conductor a is grounded at the far end to show the manner of
obtaining the simplified network. This would actually be the
case only when the circuit is grounded for test through a ground-
ing device having zero impedance to zero-sequence current.
The resultant impedance, converted toa zero-sequence base, is the
impedance which is to be connected in series relation between
SHORT TRANSMJSSION LINES WITH GROUND WIRES 165
the two points in the zero-sequence network of the system under

consideratio-n.
If current la flows into conductor a and returns through the
ground, and lb
flows into conductor b and returns through the
ground, then the voltages of a and b, above ground per mile of
length may be written:
E. ~ [ r. + 0.00159/ + j0.004657f log1o (G..L1~\.).}•

+ [ 0.00159! + j0.004657f logw ~:]lb (178)
Eb = [ rb + 0.00159! + J·0.004657j logw (G.~~R.)Jh

+ [ 0.00159J + j0.004657f ~og10 f~]la (179)
The equivalent circuit for such a combination is shown in
Fig. 87 (b). It can readily be seen that the branch between o
and ground is the mutual impedance, and that the remaining
impedances are associated with the circuits a and b and are the
differences between the self and mutual impedances (a sort of
leakage im pedance, such as between the windings of a transformer).
This equivalent circuit is of course independent of the voltages
impressed across the terminals. Applied to the ground wire
case, one of the conductors, say the b conductor, is at ground
potential, which requires that b be connected to ground in the
equivalent network shown in Fig. 87(c). The total impedance
of the ci~cuit between a and ground is found by solving for the
total impedance, composed of ao in series with the parallel
circuit composed of the mutual impedance and the leakage
impedance of the ground wire.
The current distribution between the ground wire and ground
is obtained by solving for the current distribution between the
two parallel branches.
Application of the curves of Fig. 82(a), (b '¡, and (e) will reduce
the labor involved in the determination of the reactances in the
three branches in Fig. 87. The procedure is indicated in Fig.
88, which is largely self-explanatory. The curve gives the zero-
sequence reactance per phase as a function of the G.M.R., for
a particular value of p. The sum of the reactance in the a
branch and of the mutual reactance is equal to the reactance
read from the curve for an abscissa equal to the G.M.R. of the
a conductor. The mutual reactance is the value read from

the curve for an abscissa equal to the geometric mean separation
between the two sets of conductors. It follows therefore that
the reactance in the a branch is the difference between these
two quantities, read from the curve. Similar relations apply
to the b branch. Since these values have already been reduced
to per phase quantities, it is unnecessary to multiply them by 3.
Curve from Fi~s.B2(A)

<b~cmcl(c)
,..~,t:IF.<w (G.M.R.)Cll" '·(G M.R.)b ·..olorb

3 ti mes resisto111ce of p01rodleletl conoluctors of "b"~roup
3 times re:!istotnce of potr"'llcleol conoluctors of "~"9roup
FIG. 88.-Calculation of the equivalent zero-sequence network of two induc-

tively coupled aerial circuits with earth return.
The resistance component for the a and b branches are, respec-

tively, equal to three times the resistance of each set considered
as a group.
73. Practica! Calculation of Zero-sequence Impedance.
The foregoing considered only the impedance between one
conductor and the combined ground wire and ground. Prac-
tically, the zero-sequence circuit necessarily involves three
or more conductors. These cases may be solved by resorting
again to the idea of the G.M.D. of combined conductors. Several
different cases will be considered.
Single-circuit and One Ground Wire. The disposition of
the conductors for a typical case is shown in Fig. 89 together
with the detailed calculations. The G.M.R. of the conductors
is found by taking the ninth root of the product of the nine
possible distances between the three conductors comprising the
polyphase circuit, including the G.M.R. of each of the three
conductors, which may be considered as the G.M.D. from the
conductor to itself. As st.ated in the preceding chapter, this
SHORT TRANSMISSION LINES WITH GROUND WIRES 167
is also equal to the cube root of the G.M.R. of a conductor times
the square of the geometric mean separation of the three con-
ductors. The geometric mean separation of the conductors
GW
8 ~-¿.
za. ~ti.
to -~
G. ~
/_G/1. 1?•
~/8 ~9-L9~
11'!1{"1". 00277ff.
_,_...._/
Con d.('-_) G.H.R. • 2.43 ft.
lih.,Yf/.SY/&íAWf\\III..VI~lfX\'l/J\WMII&'IlM'~
(b)
I n
Anodyt1cod Methocl Use of Curves Figs. 82 01ncl 88
(see Fig.87)
eond. G.W Cond. G.W.
0.0783; o ---0235 ~Res. of G.W
Res. of 3 é-bnds.
'<·-0.2794 /og~ ~0.781
<--- 0095 ~ Hufual res1sfance
0.705 ;j2.35
~
~ 0.46 +j !.75
lero-.se9uence impeclance •3(0./53+j0.585) lercrsec¡uence impedance
•0.46 +,J 1.76 ohms permde per phasP.
Conductor:::. a11d ground wire = 397,500 cir. mils. A.C.S.R. (30 X 7)
G.M.R. of conductor and ground wirc = 0.0277 ft.
Resistance pcr mile of conductor = 0.235 ohm
Frequency = 60 cycles
Damp earth- p = 100 meter-ohms. D. = 2,800 ft.
Mutual resistance = 0.0159! = 0.095 ohm per mile
G.M.R. = -\1 (dau) (dbc) (dca ) (G.M.R.)
2 2 2 3
= -\1 (18) (18)2(36) (0.0277)3 =

2 2 2.43 ft.
G.M.D. upara tiu,. = ~ (d.,a)(d.,b)(d xc)
= v'C28.8)(13.45) 2 = 17.3ft.
FIG. 89.- Calculation of zero-sequence impedance for a single circuit with a
single ground wire. Forty-three per cent of the current returns in the ground
Wlre.
and ground wire is merely the cube root of the three separations.
With these quantities obtained, two procedures are available.
The left-hand column illustrates the analytical method, which
follows the outline indicated in Fig. 87. The treatment is
carried through in terms of actual impedance, and to convert
to zero-sequence impedance per phase it is necessary to multiply
by 3 at the end. The procedure outHned in the right-hand

column of Fig. 89 is carried through on the basis of z~ro-sequence
impedance per phase, as in Fig. 88, and the reactances are
obtained from the curves of Fig. 82(c).
To give an idea of the error involved for the particular con-
figuration shown>. the current was calculated by the above
method for a line-to-ground fault at the open-·circuited receiving
end of the transmission line, assuming the voltage maintained
Groond wif?!s
Conductor Ground wlres
~
~
"' 18 b /él e tiroond wlres (~-, ú.MR. •0706/f.
;::::
~ lió
Conductor r,_',6.M.R.=2.43ft
~.~-
~vvvv~
~·~~~~ ~ 1.43 Ohmspermi/e

("') (b) ~perpnase
62 percenf currenf
re/ums in groundwiro
(e)
Conductors and ground wirc = 397,500 cir. mils. A.C.S.R.
G.M.R. of each conductor and ground wire = 0.0277 ft.
Neglect resistance of wires.
Damp earth, p = 100 mctcr-ohms. D e = 2800 ft.
G.M.R. ground wiro = V (dxy) (G .M.R.)oround wire
= v (18) (0.0277 ft.) = o.706 ft.
G.M.R .conductors = V (dab)2(dbc) (dca) 2 (G.M.R.)3
2
v
= 4(18) 6 (0.o:.>.n ftY = 2.43 ft.
G.M.D. separation = V (dax) (bbx) (dcx) (day) (dby) (dcy)
= {/(13.45) 4 (28.80) 2 = 17.3 ft.
Fw. 90.-Calculation of zero-sequence r eactance for a single circuit with two
ground wires. (a) Physical configuration of conductora; (b) cquivalcnt con-
ductora; (e) reduction of equivalent network.
at the sending end and the line transposed. These results

gave, at 60 cycles, a current which was 0.2 per cent too high.
In general, the effect u pon the reactive component of neglecting
the resistances of the line conductors, ground wires, and mutual
branch is very small. In the particular case just calculated
the total reactance, neglecting the resistances in the branches, is
J·o.71 j(+
(j 2_·35)+
(jl.SS)
l.SS) = J·1 .745 . Th.IS va1ue compared WI"th
2 35
j1.75 constitutes a substantial check. For more ground wires,
the check should be closer. An important exception to this
case is the use of steel ground wires.
In many cases short-circuit currents are calculated by using
the reactance alone, and ncglccting thc rcsistance. In the
example just discussed the magnit ude of the zero-sequence
impedance is 1.81 ohms as compared with the reactance of 1.75
ohms, giving a difference of 3.5 per cent. Hence the error in
current magnitude introduced by neglecting the resistance
cannot exceed t his value.
Conductor ároundwire
Groundwire ~~-)6.M.R.=QtJ/6ff.
~~
Conductor ( '¡u.M.R.•5.64 ff
,,.,
!.// Ground
~ires
l/0 1.05 .
~Ohmspermile
~ (b) 63per cenfcl/rrenf perphase
( ) r efl/rns in grounclwires ( C)
01
Co nductors and ground wircs = 397,500 cir. mils. A.C.S.R.

G.M.R. of each conductor a nd ground wire = 0.0277 ft.
Ncglect resistancc.
p = 100 mcter-ohms.
. M~.R
G·M· R ·eo" d"e tora = ?{/[:-:e-:::G:-c =-:-
.)-eo-,.d-..-cl-ur-.,e:-:U-ab'7"")-:-
ed.,...a....,.
c)...,.e1...,-la-d-:-)e..,. d.,. .u. . ,)...,.e
.• d-::-a-1.,-,-)
]4
( eG. M .R. )conductora edba) edbc) edbd) edbc) (db¡ ))2
= ~ eo.0277) (8.544) 4 e16) 2 (24.08) 2 e18) 2 e22.47) 4 (24) =
3
5.64 ft.
G.M.R. ground wires = .,Y(G.M.R.)oround wireedxv ) = ..;! e0.0277) e24)
= 0.816 ft.
G.M.D.separation = ~ edxa) edxb) (dxc) ed.ed) edxe) edx¡) ed¡¡a)
~~~~~~77~
edyb) (dyc) (dyd) (dy.) (d¡¡¡)
f
= {!= .8=3~)(-:-;
e5,--,: 13""")=e2:-: -1.-: :-2-::-<1)-;-:e2::-:::9,. -7=
::. 0~)(=27=-.3:;::-;0,..,.-)=e2:-::-1.-=579)7 .
i
¡,
= 17.44 ft.
¡ Fw. 9 1.-Calculation of zero-sequence r cactance for a twin circuit with two
r ground wircs. (a) Physical configur ation of conductors ; (b) equivalcnt con-
ductors; (e) rcduction of equivalent nctwork.
l1 Single-circuit and Two Ground Wires (Resistance Neglected).

¡ The configuration of a typical single-circuit line equipped with
t two ground wires is shown in Fig. 90. Calculating t he equivalent
radii as indicated givcs an cquivalcnt system in which the
f conductor group has .a radius of 2.43 ft . and the ground-wire
i
group O. 706 ft. The equivalent separation of the two groups
is 17.3 ft . The remaining steps in the calculation of the zero-
sequence reactance are shown in t hc figure, t he resistance being
neglected in this case.
To test the order of the error in using the simplified method,
the line alone was considered, with the receiving end open-
circuited and constant balanced voltages maintained . at the

sending end. Assuming a transposed line, a comparison between
a rigorous solution and the result obtained by the simplified
method of symmetrical components showed that the error in
the current was but 0.18 per cent for a single line-to-ground
fault at the receiving end. For a double line-to-ground fault
on phases b and e at the receiving end, the rigorous calculation
gave
lb = -0.4194 - j1.0923
and
le = -0.4194 + j1.0923
Corresponding quantities obtained by the simplified method
are
lb = -0.4155 - jl.0856
l e = -0.4155 + j1.0856
indicating an error of about 0.6 per cent.
Twin-circuit Line and Two Ground Wires (Resistance Neg-
lected). The configuration for this case is shown in Fig. 91,
together with .the details of calculation. The general method
is essentially the same as for the previous cases. The fault
current was calculated for a line-to-ground fault at the receiving
end of an open-circuited transposed line, assuming the voltage
maintained at the sending end. A comparison of the results
obtained by the rigorous method with that obtained by the
simplified method of symmetrical components indicates an
error of 0.25 per cent.
In the determination of the errors involved for the three
cases just cited no cognizance was taken of the symmetrical
series impedances introduced by the transformers and machines,
which are usually present in practice and whose effect would
still further reduce the errors.
74. General Method for Zero-sequence Calculations.

In the foregoing every effort has been made to represent the
circuits under consideration by means of equivalent circuits.
This has the advantage that it tends toward the elimination of
errors by the progressive inspectional checks of the computations
in the successive simplification of the network. More com-
plicated problems than those heretofore treated become quite
involved when this is attempted. The following method is

always applicable, regardless of the complexity of the problem.
Given any number of physically parallel conductors, a, b,
e, . . . n, and designating the currents and voltages to ground>
la, lb, In, and Ea, Eb, . . . En, respectively,
Ea - Zaala + Zablb +
Eb = ZaJa + ZbbÚ +
Znnln
These equations should preferably be set up on a zero-sequence
basis, in which the self impedances Zaa, Zbb, . . . Znn, and the
xo Ground wires oy
(~\ o o o
1 \
1 1
¡o l o o o
1 1
\ ,_O_, ¡ O o o
(01 )
faulfon
fhis c¡rcuif
xO
1
.......
,
......,
1 J
'(; b (b)
Fw. 92.- Equivalent zero-sequence network for four inductively coupled

earth return circuits. (a) Physical configuration of actual conductora; (b) equiv-
alent conductora; (e) equivalent network; (d) equivalent network when x and y
are regarded as grounded.
mutual impedances Z ab, . . . Z cn-l)n are three times the values

determined by equations (169) and (170), or three times the
values obtained from Fig. 75.
As a particular example consider the two twin-circuit lines
of Fig. 92(a) which are equipped with two ground wires. To
determine the zero-sequence network for a fault on the left-hand
circuit of the left-hand tower, the conductor configuration may
be replaced by the equivalent configuration shown in (b). It is
well to retain the identity of the two ground wires because the
great difference in their distances from the faulted circuit
probably results in considerable difference in current through
them. The twin circuit on the right-hand tower, on the other
hand, can reasonably be replaced by a single equivalent circuit.
A set of five simultaneous equations of the form shown by equa-
tion (180) may now be set up. These equations will involve
the variables la, lb, l e, representing currents in the. conductor
rtroups; and 1~ and 1117 representing currents in the ground wires
The physical conception of these constants is shown in Fig.

92(c) in which the self impedances are represented by series
impedances in the conductors and the mutual impedances are
indicated by the brackets connecting the respective conductors.
For this particular case the potential of the ground wires is zero,
so that the voltage equations for these two conductors are
O = Zaxla + Zbxlb + Zcxlc + Zxxl + Zxyly:¡; }
181
O = Zavla + ZbyÚ + Z cvlc + Zx 1 + Zyyly
11 11 ( )
By solving these two simultaneous equations for 1 x and 1 11
in terms of l a, lb, and le, and substituting in the remaining
equations, three new voltage equations of the form of equations
(180) will be obtained. These new equations, however, will
have but three terms in each.
The resultant networkreduces
to that shown in Fig. 92(d)
in which the primed constants
represent the new values. By
this process the network has
been replaced by an equiv-
alent network without
ground wires.
To illustrate further how
Fw. 93.-Inclusion of the equivalent this circuit fits into the zero-
circuit of a transmission line into the zero- sequence network, assume
sequence network.
that circuit e is grounded
through zero-sequence impedances Zc and Zc' at the near and far
ends, respectively; that circuit b is grounded through the zero-
sequence impedances Zb and Zb' at the near and far ends, respec-
tively; that circuit a is grounded only at the far end through the
zero-sequence impedance Za', and that the zero-sequencc imped-
ance of the network for a fault at the near end of circuit a is
desired. The above impedances are inserted in the network as
shown in Fig. 93. The resultant equations take the form
Ea=Z'aala+Z'ablb+Z'aJc +Z'ala }
O = Z'abla + Z'bblb +
Z'bclc + (Zb + Z'b)h (182)
O = Z ael a + Z bch
1 1
+
Z 1 ccl e + (Z e + Z 1e)1e
The answer sought is, of course, ~a or the value of Ea for
l a = 1, which can be obtained most conveniently by solving
the last two equations for lb and l e and inserting in the first.
Example. To illustrate the use of this method, let it be

desired to determine the zero-sequence impedance of the 50-mile
66-kv. transmission system of Fig. 94(a), including two twin
¡<-------------100'- ----- ---~
· -------¡;- •
1
r ---18' --i ------i~

~
1
o o
1
¡<--J~~~ k-3'-~ <o
b 6t 1
o o
~
1
o o- - L o o
Conclvcfors omcl grovncl w/res=397,.{00 C.M.,A.C.S.R,G.M.R. ofe()fch
conclvcfor omcl grovnol wire "'0..0277ff.
Neglecf restsfPtnce Fret>j.umcy =60 cycles
Lengfh "'50 miles p ~ 20
\Ot}
-,
j.~~~~~~------~9~0-'------------~~
\ <\,,
/ -"""-....1.45 ' /"'-S.61'
-----------------no~-------------' -
(lo}
Xxy=j0.92
...,
11
~
-
><
.....
,,-L.::rn--:---..-:~~---:-:~~:=-.~-----;:-"""'1 e \1
tOl,....
..... 1
/
L;.;::la%!~-Jt.=;;__-----------.!:::.::......L.::._;_;__ _ _ _ _...J ( _X ce• j2.00
(e}
(a) Circuit configuration and conductor data
(b) G.M.D. and G.M.R. data i n ! eet
(e) Self and mutual reactances per phase p er mile
Fw. 94.-Transmission line circuit wi th ground wires used to illustrate zero-
sequence impedance calculations.
circuits with solidly grounded transformers at each end. The

fault will be assumed to occur on the extreme left-hand set of
conductors. For this case, the zero-sequence current disiri-
i1'
1
1
'
~
!
f
bution between the three conductors of this circuit should not

depart far from uniformity, so that the three conductors can
be represented by a single equivalent conductor. Similarly,
the second circuit from the left can be represented by an equiva-
lent conductor. Because of the wide separation between the
two towers supporting the twin circuit and assuming that the
two circuits are grounded in a similar manner, the current
distribution between the six conductors of the right-hand
twin circuit should also be substil.ntially uniform, so that the
six conductors may be represented by a single equivalent con-
ductor. The two ground wires cannot be combined, except
as an approximation, into a single equivalent conductor because
of the difference in current in the two conductors. The identity
of the two ground wires will be retained to further illustrate the
application of the method. With these simplifications the
system may be represented by its equivalent conductors shown
in Fig. 94(b). The G.M.R. and G.M.D. are obtained in the
manner explained previously. For the wide separations it is
sufficiently accurate to take as the mean separation the distance
between the centers of gravity of the two systems under con-
sideration. Probably the easiest way to determine the center
of gravity is by the intersection of the medians of the triangle
formed by the three conductors as apexes.
With the G.M.D. available, the self and mutual impedances
to zero-sequence can be obtained from Fig. 82. These are
shown in Fig. 94(c).
Since the ground wires are assumed to be grounded at each
end through zero resistance their potential is zero and the follow-
ing equations can be written by using the self and mutual
impedances per mile from Fig. 94(c).
Ex = j[1.50fa + 1.50Jb + 0.91Jc + 3.90J x + 0.921 11 ] = O
E 11 = j[0.88la + 0.95h + 1.50fc + 0.921 x + 3.901 11 ] = O
Solving for 1 x and 111
Jx = -[0.352fa + 0.347Jb + 0.151Jc]
111 = -[0.143Ja + 0.162h + 0.349Jc]
For circuits a, b, ande, the voltage equations are
Ea = j2.50fa + j1.48Jb + j0.89J~ + j1.50J x + j0.88Jy
Eb = j1.48fa + j2.50h + j0.96fc + j1.50J x + j0.95Jy
Ec = j0.89fa + j0.96h + j2.00fc + j0.91J + j1.50fy
:r;
Substituting the values of 1 x and 1 v in these equations, the
results per mile are
Ea = j1.83la + j0.82lb + j0.35le
Eb = j0.82la + j1.83lb + j0.40le
E c = j0.35la + j0.40lb + j1.341e
The above impedances are for unit length of line. By mul-
tiplying by 50 they may be converted to those of the 50-mile
line represented by the equivalent circuit shown between the
dotted lines in Fig. 95(b) from which the ground-wire circuits
have been eliminated.
(ot) (b)
F10. 95.-Zero-sequcnce equivalent network of circuit whose physical configura-
tion is given in Fig. 94, without bussing on high-voltage sidcs of transformers.
Beyond this point the solution is determined by the system

bussing arrangements and whether certain circuit-breakers are
open or closed. For the system layout shown in Fig. 95(a),
in which the high-voltage circuits are not bussed and the breaker
in the a circuit is open, the transformer impedances may be
included as shown in Fig. 95(b). The voltage equation for the
b and e circuits may then be written
Eb = j41.0la + j131.5h + j201c = O
E e = j17.5la + j20h + j87le = O
and solving for lb and le
h = -0.292la
,.
le = -0.134la
!·
For the a conductor
!~ Ea = jlll.5la + j41.0lb + j17.51c
and substituting the values of lb and l e
~: = Zo = j97.2 ohms
If the high-voltage lincs had been bussed at both ends, as

shown in Fig. 96(a), the equivalent circuit for solution would
be changed to that shown in (b).
If the breaker in the a circuit had been closed, the equivalent
circuit could have been obtained by merely closing the corre-
sponding gap in the equivalent circuits. Of course, if this is the
only condition to be analyzed, i.e., with this breaker closed,
(a) (b)
Fw. 96.- Zero-selluence equivalent network of Figs. 94 and 95 when the lines
are bussed at both ends.
the equivalent circuit can be obtained easier by considering

the a and b circuits as a unit, as was done for e, assuming equal
current distribution between the six conductors.
7 5. Steel Ground Wires.

The zero-scquence impedance of circuits with steel ground
wires is obtained in a manner similar to that indicated for
circuits equipped with copper and aluminum ground wires,
except that, whereas the resistance for the latter may usually
be neglected, the resistance must be considered when dealing
with steel ground wires. Data on the characterisiics of steel
conductors are given by Prof. H. B. Dwight. * The curves in
Fig. 97 were calculated from these data and give the resistance
and the G.M.R. of steel conductors of diffcrent grades as affected
by current, for frequencies of 25 and 60 cycles.
Prof. Dwight's data are presented in the form of resistance
and reactance (due to interna! flux only) per mile. However,
a knowledge of the equivalent G.M.R. fits into the present cal-
culation more readily. These values can be obtained in the
* Resistance and Reactanco of Commercial Steel Conductors, Elec. J our.,

p. 25, January, 1919.
1
~Íe~de E . B. B.Seven-Stre~nd 1
C"ble 60 Cvcles 1
l'~b.M.R.
,, -~
y., .
'4 ln.c(¡"q
.L • ~et-er- 1
1
1
~-'"if~)' "X 9/nm. ~ 1
''·\V ~>": '--'.-..¡;,!... Jie m. .- ~~ 1
''· V ./
\1'.. "'f ;;;;;;:¡¿.';f,;n.
../\ ~-v ->'- !":K
1 ~. .X \ / Y§ 0 :4:r::::. r--.:
~~ltl X ~ ..:-0., m ::::"" ~
A r.~ '1,.. r.:~6'
s ~-"" J.1 -, ,
, ...
6 /,1\1 '' .....
--
9-32 1n. /
-- .... r'
7 Ulf\ /
a Resisfomce' f----
1
f4in.7
10 20 30 40 50 60
Amperes p~r CQble
t-Or~.i ne~ry Steel Nineteen StranGl

1 Cocl;>les, OCycles 1 f-Q¡·c 'nc.ry Stee 1 Naneteen S rc.nd
--
11 C<>bles, -~ Cvcles
~ 10~ o 1
50 :E
:::¡: 50 9::E 9 :i:
~ 8 ... 8 r-
o 60 ~
¡.......- 7 lt 7 60 "
G
<t; 80 6 "'G r-
S ¡: 5 ¡..-
80<t
100 'lS
1100 ¡.....- 4<34 150 §
~ ISO 3 .;3 200 ..
lOO 2 g2
~881
::>
~ ~gg <S 1
J~o t-
o 50 100 150 ~ o 50 100 150
Ampere.s per Coable er: Amperes per Cotbli
Fw. 97.- Electrical characteristics of steel ground wires.

following simple manner. Let x be the reactance in ohms per

mile at 60 cycles due to the interna! flux only, then
radius
x = 0.2794 log10 (G.M.R.)
radius
(G.M.R.) (183)
antilog10 _;
0 2 94
For 25 cycles 0.2794 should be changed to 0.1164.
As an exam ple of the importan ce of the resistance of ground
wires when steel is used, the same arrangement of conductors
as given in Fig. 90 was chosen, except that the ground conductors
From Fig. 97 assuming ! in.

diamctcr Grade E.B.B.
seven-strand cable and 60
amp. per wire at 60 cycles
G.M.R. = 10-s ft. p = 100
Conductor Grouncl wire: G.M.Roround wirr• =
Resi.s famce of
J conclucfors
..-Resisfcmce of 2 V (dxy) (G.JI!J.R. )urut<n d wirc•
groundwtres•24ohms
=~= 0.0783/ per m /le (see F.ig.97) = y(18)(10- 8 ) = 4.24 X
,¡1
10- 4 ft.
! 02794 /og4_~;to"" =1.29
0.279-f/ogJ.~j =0.238·/ G.M.D. of scparation from
\02794/og ~~~l =0.616 Fig. 90 = 17.3 ft.
G.M.R. of conductors from
Fig. 90 = 2.43 ft .
Zcro-scquence impedance = 3
(0.244 + 0.752) = 0.73 + 2.27
24+JI.29
ohms per milc por phasc. 14
per cent of total current flows
through ground wire
0. 244 +jo 752
~
Fw. 98.-Calculation of zero-sequence impedanco of a single-circuit transmission
line with two steel ground wires.
were assumed to consist of two !-in. stecl conductors. The

details of calculation for the case are shown in Fig. 98. N eglect-
ing the resistance components in this case gives a value of
reactance equal to
·o 238 + (j1.29 ) (j0 ·616) = ·o 655
J . j(1.29 + 0.616) J .
which is considerably smaller than the value (j0.752) given in
Fig. 98. The use of steel conductors also increases the zero-
sequence reactance over that obtained with aluminum or copper
conductors as evidenced by the comparison with a similar layout
using A.C.S.R. conductors (see Fig. 90) which gave a reactance
of j1.41 ohms per mile per phase.
76. Copper-clad Ground Wires.

Copper-clad steel cables are sometimes used for ground wires.
The increase in resistance* at 25 and 60 cycles may be taken as
from 5 to 10 per cent over the direct-current resistance value for
currents up to 200 amp.
The G.M.R. of round iron conductors of constant permeability
¡¡.
assuming uniform current distribution is a€ -;¡. For ¡.¡, = 1, this

11-
reduces to the well-known figure O. 779. When alternating
current fiows through steel conductors, because of magnified skin-
effect phenomena, most of thc current fl.ows near the surface.
Nevertheless, it is always possible to express the result as an
equivalent G.M.R. of this type, using an effective permeability.
This effective permeability is not related directly to the actual
permeability and is a function of current and frequency.
Sorne of thc availablc test data* on copper-clad conductors
express the results as a ratio of the interna! inductance of the
copper-clad conductor to the interna! inductance of a copper
conductor of the same diameter. It muy be shown that this
ratio is equal to the effective permcability defined above. The
test results show a range for this value of from 4 to 20, but for
most practica! calculations a single value equal to 7 may be used.
This gives a G.M.R. equal to 0.17a.
77. Sequence Impedance Constants of Typical Circuits.

In order to develop a sense of proportion as to the order of
magnitude of transmission-line impedances and the effect of
adding ground wires, the constants of a number of typical trans-
mission lines were calculated and tabulated in Table VIII.
The two columns undcr thc hcading of positive-sequence imped-
ance represent the ordinary impedance of a single and two
parallelcd circuits. The next general heading under zero-sequence
impedance gives the accurate constants including resistance
for the circuit conditions indicated at the head of the columns.
The addition of ground wires may decrease the zero-sequence
* "Ovcrhcad Systcm Reference Book," pp. 85 and 98; published by
N ational Electric Light Association.
T ABLE VIII.-CONSTANTS OF TYPICAL TRANSMISSION LINES ......
00
_..,- o
!> Zcro-sequence reactance ..,;lO-+='
·;:; '3 g g
·- ·-
o>
~
i!l
-~
Positive-sequence impedance Zero-sequence impedance neglecting resistance in
computations ...o ...o
·o ·s
1)C)
- ·-
_
L..
~ _~
...
:Q "O -- ..e ..e

¡:; e
" 'ª -~
ococ ""
~~
-cJ- 1 ~1-D-
Configuration of conductora 1 ·; E Z ao
~
oc oc c...'-"
o o
........
o o
!l o...
o> Single circuit / Double circuit ~~ ~~
"O "'
..0
Double ._3 .Si
"' S Double circuit 1 Zao
z ... Vl
~ z" Zuo ZmJ 1 circuit
~ ~ ~
o--.- 1
~
~
7' O Copper 441 o 10.56 +10.76 / ............. ¡0.84 +i2.58 2.58 3.38
9-·--I- p = 20
o ......... o. o • •
, ' 7'
---~4.50..- ....,J._ ~
0 - · 7<
----· ·-----·----- ~
~
/6' Q
?--~,
0000 Copper 1101 o 10.28 + j0.81 / ..... o. o o o o ••• /0.56 + j2.3 ... o o o o. o.'. o •• 2.34 2.89 p..
p = 20
t:'-4
··-1 9'o-.- _x_
-- 3.52 ,_ _ Q
__......_ 18 ,_~-- T Conductors o ,0.24 +i0.81 , ........... ... ,0.52 +i2 .85 • • • • o • • • • • • • • • ..... . ........ 2.85 . .. . . a
~
,-- • , and gruund
Grouncl wires -' 10 wires 397,500 1101 1 0. 24 +i0.81 ... ........... 0.46 +il.i6 . ········ ...... . . . .. .. .. .. . . l. 76 .. ... . .... 2.17
()4·/8'~!8'--o...L ~.m.AS.C.R. 1 -12_10. 24 +i0 .81 .. .. ... .. ... .. ,0 . 41 +il.43 . .. . . . ····· .. ..... . ...... .. .. 1.43 ... .. . .... __
1.77 __ a ,
~
-
1G-I"{~JF*t2~. Conductors o 0.24 +i0.72 0.12 +i0.3500.52 +i2.75 0. 29 +il.77 0.40 + j2.26 2.75 1.77 2.26 3.82 6.45
26 ws·- ··o: -.-;o~
18 and grmwd ~
O.· 24'···><>+
wires 397,500
c.m. A.C.S.R.
661 1 0.24 +i0.72 0. 12 +i0.3500.46 +i1.92 0.23 +jO. 94 0 .40 + j1.43 l. 92 0.94 1.43 2.67 4.09 ~
8' 100 2.18 3.12
0./8'..0 .....i.
p = 2 0. 24 +i0.72 0.12 +i0.3500.42 +il.57 0. 17 + j0.62 0.29 + jl.09 1.57 0.62 1.09
Conductors 795,000 c.m. A.C.S.R.-two

- - ----
layers
Ground wires 203,200 c.m. A.C.S.R.-one
• ·----.. - --
!6.75 '
Ot-25S~!.~~~¡icJ;~?~9-~25.5.:0
!ayer
• 220
fl ,0.117 +jO. 814 O. 059 + j0.407,0 .403 + j2. 446 0.286 + jl. 02
' 4 ¡0 . 117 ..Lj0.8140.059+i0 . 407 ¡0.28 +11.42* O.llO+i0.212*0 . 195+j0.81 1.42 0 .233
O. 345 + jl. 73 2. 45 l. 02 l. 73
0.826
3.00
l. 74
4.25
2.00
------------------------~--~----------~------
• Four ground wires.
reactance as much as 50 per cent. As shown in the next set of
columns the inclusion of the resistances of conductors has a
negligible effect upon the value of zero-sequence reactance
obtained. A convenient measure of zero-sequence reactance
that is sometimes used is to ex¡:1t:ess the value in terms of the
positive-sequence reactance. These raLios are tabulated in thc
last two columns, from which it may be seen that 3.5 represents
an approximate value to express this ratio for single circuits
without ground wires, but for circuits with ground wires this
value is much smaller, decreasing to 2.7 or 1.7 depending upon
the effectiveness of the ground-wirc system. For double circuits
this ratio varíes over a still greater range being 6.5 or 4.2 for
lines without ground wires and 4.1, 3.1, or 2.0 for circuits with
ground wires.
Problems
1. Assume the transmission system of Fig. 90 except that the ground
wire is of 000 copper and that the valuc of p cquals 10 meter-ohms. Com-
pute thc zero-sequcncc impcdance of the circuit and the division of the
return currcnt betwecn thc ground wircs and thc earth.
2. Assumc that the transmission system of Fig. 91 is equipped with grade
EBB scvcn-strand steel cables and that ti1e current is 50 amp. in each cable,
assuming a wire of Yz inch diamctcr. Determine the zero-sequence imped-
ance and division of current between ground wires and .earth.
3. A powcr system with a transmission lino as in Fig. 89 but without
ground wircs is parallcled at 100-ft. separation for 1 mile by a telephone
circuit of two wires located 20 ft. abovc the ground. (a) Determine the
voltage induccd in the ground-rcturn circuit of the telephone lino due to
electromagnctic induction resulting from thc flow of 4,000 amp. through the
parallel ata frequency of 60 cycles. (b) What rcduction in voltage will be
accomplished by the installation of a 4/ 0 eopper ground wire on the power
circuit as in Fig. 89 providcd thnt the short-circuit current is not increased
by thc reduction of zero-sequence im pcdancc? (e) If a 4/ 0 coppcr ground
wire (shield wire) is located in a position 1ft. above and midway between the
telephone wires, which a re located 10 in. apart horizontally, what 'per cent
reduction will be accomplished if the connections to ground are of negligible
impedance?
4. For the power system of Fig. 95 determine the impedance to zero-
sequence as measured ·at the point a if the adjacent circuit-breaker is closed.
Also d etermine the current division for zero-scquence.
5. For thc case of Fig. 95, if the transformar terminals are bussed as in
Fig. 96, determine the zero-sequence impedance for a fault at a with the
adjacent circuit-breaker open.
•
CHAPTER IX
CONSTANTS OF LONG TRANSMISSION LINES
In the preceding chapters consideration was given to the

characteristics of short transmission and distribution lincs in
which t he effects of charging currents and the distributed nature
of the constants could be neglected. The present chapter will
discuss how these factors may be included in the characteristics
of longer lines.
78. Positive- and N egative-sequence Characteristics.

For longer lines in which the capacitance is not negligible the·
positive-sequence characteristics may be specified by t he ABC
constants, which are defined by the equations
Es= AEr + Blr
(184)
f a= CEr + Afr
where Ea a nd Er are the voltages to neutral at the sending and
receiving ends, respectively, and I s and I r are the correspondin g
line currents. These coeffi.cients are the result of the hyperbolic
solution of t ransmission lines taking into consideration the dis-
tributed character of the constants. *
- [ 2 2
Y3Z 3
A = cosh VZY = 1 + -2 + -- + -720 + · · · ]
YZ Y Z
24
B =
/Z .
~y smh yZY = z
[1
+6 +
YZ Y 2Z 2
120 +
Y 3Z 3
5040 + ]
e =
(Y -
~z sinh VZY = y
[
1 +6 +
YZ Y 2Z 2
120
Y3za
+ 5040 + ]
(185)
* Various methods for calculating these constants are given by William

N esbit, on p. 57, "Electrical Ch.a racteristics of Transmission Circuits,"
3d ed., Westinghouse Technical Night School Press, East Pittsburgh, Pa.
These constants have been calculated and tabulated for different conductors,
spacings, and lengths of line in the same book.
182
CONSTANTS OF LONG TRANSMISSION LINES 183
in which
Z = R + jX = totalline impedance = (r + jx)l.
Y = G + jB = totalline susceptance = (g + jb)l.
g = conductance to neutral in mhos per unit. length of line
(usually neglected).
b = capacitive susceptance to neutral in mhos per unit length
of line.
The value of b may be obtained by taking the reciproca! of the
capacitive reactances given in tables in the Appendix.
In the absence of tables, b may be obtained by means of the
following formula:
14.64 X 10- 6 •
b = D mhos per m1le per phase at 60 cycles (186)
log w -
a
in which D is the equivalent separation or G.M.D. of the con-
ductors * and a is the radius of the conductors.
Equivalent 1r. For many purposes, such as setting up the
equivalent of the transmission
line on calculating board s, it is Zeq
desirable to know the equivalent
Yeq
1r of the transmission line. This
is a circuit of the form indicated
in Fig. 99 having a series impedance Fro. 99.- Equivalent circuit of
a transmission line.
and a shunt susceptance at each
end. It can be easily shown that
z[
~(187)
Zequiv. = B = 1 + YZ + Y2Z2 + yaza +
6 120 5040
Y equiv. _ A B- 1 _ y[_!2 _ YZ
24
+YZ
240
2 2
_ •
])
For short lines Yequiv. may be neglected entirely; for longer
lines the first term, namely ; should be used; and only for the
very long lines is it necessary to use more terms of the series.
Example. An idea of the relative importance of the terms may be
obtained by calculating Zequi v . and Yequiv. for a 300-mile, 795,000 A.C.S.R.
conductor line with 25-ft. spacing, 60 cycles.
* D is numerically equal to the cube root of the product of the three
distances between conductors. It should not be confused with D used with
the ABC constants.
Zoquiv. = Z[ 1.0
-0.0631 + j0.0093
+0.0012 - j0.0004
+0.0000 + jO.OOOO + · · · ]
= Z[ 0.9381 + j0.0089]
Yequiv. = Y[ 0.5
+0.0158 - j0.0024
+0.0006 - j0.0002 + . . . ]
= Y[ 0.5164 - j0.0026]
It will be observed that for ordinary accuracy in short-circuit

studies it is necessary, in calculating Z equ iv. and Y equiv. for a
300-mile line, to consider only two terms and since Z and Y
both vary linearly with the length of line it is necessary to
consider only the first term for lines below 100 miles. In
economic studies involving the determination of losses it may be
necessary to carry the computation one term further.
79. Zero-sequence Susceptance.

In the calculation of the capacitive susceptance the surface
of the earth may be considered an equipotential plane. The
charge distribution on the surface will be such as to produce a
field in the air equal to that produced by the charge on the
conductor and an opposite charge on its image, at an equal
distan ce below the earth's surface. U nder these conditions the
susceptance is
0.2440f10- 6
b = h mhos per mile (188)
2
log1o-
a
When the susceptance of three parallel conductors is desired,
the G.M.D. may be determined and inserted in formula (188).
The derivation of these G.M.D. for the ground return case
follows:
Referring to Fig. 100(a), for unit positive charge (in absolute
units) on the cylindrical conductor a and a unit negative charge
upon its image, the potential of a above ground is
Va = 2 logE Ddaa
aa
and the potential of the point b is

Dah
vb = 2 logE dab
CONSTANTS OF LONG TRANSMI8SION LINES 185
olbb=radius
(01)
Dbb Dab
0.244¡10- 6 loglO -¡¡;;;; 0.244!10- 6 l og1o -d
ab
bab =
~
2
0.244f10- 6 1og 10 -d
Daa Daa l Ogto-d
t.= logto-d Dvb - (log¡o-d
Dab)
aa aa bb ab
01
olcc=radius b
01
Dee
De,.O!
V•
(o!)
6 2
b aa =
0 .244j10-
.ó. [ ( Dbb) ( 1Ogtod-
1Ogto-d Dcc) - ( 1oglod-
Dbc) ]
bbce be
6 2
bbb = 0.244/10-
A
[ Dcc) (log10-d
(Iog10-d Daa) - (log10-d
Dac) ]
u r.c aa ac
bCC -_ 0.244/10-
A
U
6
(l0g10d-
[ Daa) (t0g10d-
aa
Dbb) - (toglod-
bb
,Dab)
ab
2
]
0 24
bab = · ~10 -T (1og 10 ~:e) (1og 10 ~:e) - (log 10 ~:b) (log 10 ~:e)]
fbc = · y -T(log 10 ~:b)(log 10 ~:c)- (log 10 ~:c)(log 10 ~:a)]
0 24 10
b ac = · y -T(1og 10 ~:b)(log 10 ~:c)- (log 10 ~:c) (log 10 ~~)]

0 24 10
t. = (] og10 -dDaa 1ogto-d

Dbb log1o -d
Drc) + 2 (log10-d Dab 1og10-d
Dbc 1og10-d
Dac)
aa Lb ce ab be ac
(1 Dbb) Ogto dar
oglo dbb (1 2
Dac) - (1
Dcc) ( loglO dab
Ogto dcc
2
Dab) - ( lOgto daa
Daa) oglO dbc
Dbc) (1 2
FIG. 100.- Equivalent susccptance networks for two and three parallel circuits.
Constan ts are in mhos per mil e. For zero-sequence in mhos per phase per mile,
divide the above values by 3.
App1ying these same relations to the system of cylindrical

conductors in Fig. 101, for unit charge on the three parallel
conductors (one-third of a unit charge on
each conductor), the potentia1s of the three
conductors are
2l D aa 2l D ab 2 Dae
Va = 3 ogE daa +
3 og. dab 3 1ogE d ac +
2 D ab 2 D bb 2 Dbc
vb = 3 log. dab + 3 log. dbb + 3 1og. dbc
2l D ac 2 Dbc 2 D ce
Ve = 3 og. dac + 3 1og~ dbc + 3 1og. d cc
The distribution of charge between the three

·-Dcc conductors in practica! prob1ems will not befar
---D01c
... obc from uniform, but, to average out any difference,
the potential of the three paralle1 conductors
shall be taken as the average of the three
potentials Va, Vb, and Ve.
(,_,
lm01ges V= Va+ Vb +Ve

3
FIG. 101.-Nota- _ 2 D aaDbbD rrD ab 2Dbr 2D ac 2
tion for determina- 1
- -9 og. d d d d 2d 2d 2
tion of zero- aa bb ce ab be ae
sequence
tance.
suscep- = 2 lo
g
V DaaDbbD ccD ab2Dbc2D ar2 (189)
t V daadbbdccdab 2dbe 2da c 2
The denominator in this expression is t he G.M.R. for sus-

ceptance and is the same as the G.M.R. used for inductive
reactance ca1culations, except that it does not contain the factor
0.779 which took the interna! flux into consideration. For
identical conductors daa = dbb = d cc = a, and the denominator
may therefore be expressed as
G.M.R. = VOJ52 (190)
The numerator, the G.M.D., represents the ninth root of the
product of the nine distances bctween the conductors and their
images. For most work it is found sufficiently accurate to use
as the G.M.D. twice the arithmetic mean of conductor heights
h1, h2, and ha •
G.M.D. = 2h1 ~2 + hs + (191)
Expressing the susceptance in practica! units
o. 244f1 o-s . (192)
b = G.M.D. mhos per mlle
loglO G.M.R
or expressed in mhos per phase for the zero-sequence
0.0813! X 10- 6 •
bo = G.M.D. mhos per m1le per phase (193)
loglO G.M.R.
Example. For onc of the circuits shown in Fig. 84 of Chap. VIII,
G.M.R. = ~ 0 ·:~ 3 (10.5)2 = 1.55 ft.

G.M.D. = i(45 +53 + 61) = 106ft.
NOTE. Using the correct G.M.D. gave a figure of 105.6 ft.
0.0813 X 60 X 10- s .
bo = = 2.66 X 10-6 mho per mlle per phase
106
logto 1.
55
80. Mutual Susceptance between Two Parallel Circuits.

It is sometimes desirable in making calculations on parallel
circuits to express the charging-current characteristics in terms
of self and mutual susceptance as in the equations
J a = jbaaEa - jbabEb (194)
Jb = ,ibbbEb - jbabEa (195)
In these equations the coeffi.cients baa and bbb determine the
charging current ftowing in their respective circuits due to
the potential on itself, whcreas the coefficient bab determines the
charging current ftowing in one circuit when the other circuit
alone is energized. These coefficients may be determined as
follows:
Referring to Fig. 100(a), for charge qa (in absolute units)
on the cylindrical conductor a, the potential of a above ground
is
Va2qa log. Ddaa
=
aa
And the potential of point b above ground is
Dab
V b = 2qa l og. -d
ab
in which daa is the radius of conductor a.
For a charge qa on a and qb on b
Va =
Daa
2qa log. -d
aa
+ 2qb log. -dDooab
(196)
Dbb
vb = 2qb loge -d
bb
+ 2qa log. -dDoo ab
Solving for qa and qb in terms of Va and Vb
l)bb l)ab
2 1o~ d 2logE d
qa = G bb Va - G ab Vb
l)aa
2 logE -d 2 1 l)ab
ogE -d
qb = G aa Vb - G ah Va
in which
2
l)aa 1
G = 4 1ogE -d l)bb
og. -d - 4(log. l)ab)
-d
aa bb ab
Converting to susceptances in practica! units
l)bb
0.244f10- 6 log 10 -d
baa = bb m h os per illl·¡e (197)
~
0.244j10- 6 log1o ~aa

bbb - ________
aa mhos per mile (198)
~
0.244/10- 6 log10 ~ab

bab = ~ ab mhos per mile (199)
in which
For multiconductor circuits the G.M.l). may be used with

sufficient accuracy. Zero-sequence quantities are one-third
of the above, for example,
0.0813/10- 6 logw ~ab

babO = ab mhOS per mile per phase (200)
~
The equivalent network for the circuit as shown in Fig. 100 (a)
is given in (b). Its derivation may be obtained from equations
(194) and (195) as follows. The current flowing in b with one
end grounded is equal to jbab for unit voltage on a, so that the
susceptance connecting a and b must be equal to b ab· , The current
flowing into a with b grounded is equal to jbaa for unit voltage on
a. Since a susceptance b ab is already connected betwecn a and
b, the susceptance (baa - bab) must be connected between a and
ground to permit current jbaa to flow in a. Similar considerations
CONSTAN'i'S OF LONG TRANSMISSIÓN LINES 189
prove that the susceptance between b and ground must be equal
to (b,,,, - bab).
, The equivalent circuit for susceptance may readily be extended
to take care of additional conductors or circuits. The value
r of the sus·ceptances are determined by setting up a group of
equations similar to (196). Thus the case of three conductors
is shown in Fig. 100(c), and the corresponding equivalent circuit
is given in (d). The accompanying equations define the suscep-
tances for (d) and also the coefficients of the equations of the form
of (194) and (195).
Example. As an example of such calculation , consider the twin-circuit
line whose configuration is shown in Fig. 84. From previous calculations,
daa = 1.55 ft. = dbb
dab = 21.9 ft.
Daa = 106 ft. = Dbb
Taking the ninth root of thc product of the nine possible distances between
one group of conductors and the images of the other group of conductors,
Dab = 107.7 ft.
NOTE. In general for most practica! cases it is sufficiently accurate to
take Dab = D aa·
2
106 )
t.. = ( log 1 o 1. log1o 107.
- ( 7) 2
55 2 1. 9
= (1.835) 2 - (0.692) 2 = 2.88
0.0813 X 60 X 10- 6 107.7
babO = 2.88 1Og¡o 21.9
= 1.17 X 10-s mho pct-mil e per p hase
81. Susceptance Calculations for Circuits with Ground Wires.

For positive- and negative-sequence susceptance, the presence
of the ground wires may ordinarily be neglected because the
ground wire is usually located at such a position that the resulting
dielectric field is very low. The effect of ground wires may be
taken into account in the manner described below in connection
with zero-sequence.
For zero-sequence, the presence of the ground wire should
usually be taken into account. This may be done by setting up
the equivalent circuit for zero-sequence susceptance as discussed
in the prec~ding section in connection with Fig. 100. Thus, if
conductor e of Fig. lOO(c) is the ground wire which is taken as
short-circuited to ground, then the equivalent circuit of Fig.
lOO(d) reduces to three terms, namely, the two self susceptances
(baa - bab) and (bbb - bab) and the mutual susceptance (bab) .
82. Zero-sequence Equivalent N etwork for a Single Transmis-

sion Circuit.
For short lines the charging current may be neglected and the
characteristics determined by the series resistance and reactance
alone; but for longer lines it is necessary to take the distributed
capacitance into consideration just as for the positive- and nega-
tive-sequence. The characteristics may be expressed either in
terms of t he ABC constants as in equations (185), or in terms
of the equivalent series impedances and shunt admittances,
as in equation (187) and Fig. 99.
Example. For one of the single circuits shown in Fig. 84 the zero-sequence
constants for a 100-mileJine are from Sec. 67 and 79:
Zo = 100 (0.521 + j2.76) = 52.1 + j276 ohms
Yo = 100(j2.66 X 10- 6 ) = j2.66 X 10- 4 mho
YoZo = -0.0734 + j0.01386
(Y 0Z 0 ) 2 = 0.0052 - j0.0020
z . - z[
equ•v. - O
1 + -0.0734 + j0.01386
6
+ 0.0052 - j0.0020 + . . . ]
120
= Z o[1 + (-0.0122 + j0.00231)
+ (0.0000 + jO.OOOO) + · · · ]
= (52.1 + j274)(0.9878 + j0.00231)
= 50.8 + j271 ohms
. _ Y [o _ -0.0734 + j0.01386
Y equ>v. - O
. •5
24
+ 0.0052 - j0.0020 + . . . ]
240
= Y o[0.5 - ( -0.0031 + j0.0006)
+ (0.0000 - jO.OOOO) + · · · ]
= (j2.66 X 10-4) (0.5031 - j0.0006)
= (0.0016 + jl.34)10- 4 mho
The above indicates that for alllines below 100 miles it would
be permissible for the degree of accuracy normally expected in
short-circuit studies to disregard all but the first term, thus
reducing the equivalent circuit to a series reactance equal to
the total reactance and a shunt branch at each end equal to
half the total susceptance. In certain cases the susceptance
branches may be neglected entirely. The line may be so short
that for all conditions the charging current will be negligibly
small, or the fault may occur at such a point that the shunt
susceptance is directly in parallel with the line impedance,
in which case the impedance of the shunt susceptance, even for a
100-mile line, would be so large in comparison with the series
impedance of the lineas to by-pass an extremely small current.
83. Parallel Circuits with Distributed Inductive and Capacitive
Coupling. *
For sorne cases it will be necessary to include also the effects
of self and mutual capacitances as well as mutual inductances.
The type of circuit under consideration is shown schematically
in Fig. 102, which shows only a finite number of branches. The
cases which usually occur in practice involve similar lines, and
Fw. 102.- Schematic diagram of two parallel lines with distributed self and
mutual constants.
for this case the equivalent network reduces to that shown in
Fig. 103. A discussion of the derivation of this simplification
follows:
In Fig. 102 letea and eb be the voltage to ground at any point
on lines a and b, respec-
tively, ir¡, ib the current at
any point in lines a and b, .
t~
respectively, and
Za = self impedance of
line a, per unit
length of line.
Zb = self impedance of
line b, per uni t
length of line.
Fra. 103. -Zero-sequence eq ui va l en t
Zm = mutual impedance network of two parallel t r ansmission lines
between linea and including distributed inductiva and capaci-
line b, per unit tiva couplings.
length of line.
* This section d cals
with a problem of infrequent application ; its review,
therefore, may well be postponed until t he practica! need arises. The prob-
lem and the m ethod of attack employed are of interest from the theoretical
point of view.
Ya = self admittance of line a, per unit length of line.

Yb = self admittance of line b, per unit length of line.
Ym = mutual admittance between line a and line b, per unit
length of line.
The conductance components of the three admittances Ya, yb,
and y m may usually be neglected. The susceptance com-
ponents are defined in accordance with equations (196), (197),
and (198).
The voltage to ground of any point, at a distance x from the
receiving end, is equal to the receiver voltage Er
plus the self
impedance drop dueto ia and the mutual impedance drop due to
'l-b. Expressed analytically:
(201)
The voltage of line b can be expressed similarly:
eb = Er + fox ÍbZbdX + fox ÍaZmdx (202)

The current at any point in linea is equal to the load current
la plus the charging current to ground and the charging current
to the other conductor.
ia = 1 a+ fox eayadx - fax eby mdx (203)
Similarly for the current in b:

Íb = 1b + LXebybdx - fox eay mdx (204)
These equations may be solved by a method of successive

approximation that is frequently found of value in the practica!
solution of equations. Arguing from the short line in which the
current is constant throughout the entire length but the voltage
of which varies linearly, the first logical approximation is to
assume
Íal = la (205)
Íbl = lb (206)
Substituting in (201) and (202)
eal = Er + l aZO:X + lbZmX (207)
%1 = Er + lbZbX + l aZmX (208)
and substituting (207) and (208) in (203) and (204), the second
approximation is obtained.
• x2 x2 x2
'ta2 = la + ErYaX + laYaZa2 + lbYaZm 2 - ErYmX - lbYmZb2 -
x2
laYmZm2 (209)
. x2 x2 x2
%2 = h + ErYbX + lbybzb2 + l aYbZm2 - ErYmX - l aYmZa2 -
x2
lbYmZm2 (210)
and calling Za, Zb, Zm, Y a, Yb, and Y m, the corresponding con-
stants for the whole length of the line,
.
'l.a2 = las = [Y a - y m]Er + [1 + -Ya.Za
2- - Y mZm]l
2 a
+ [Ya; m _ Y zb ]lb (2 1l)

2
ib2 = lbs = [Yb - Y m]Er + [ Y ~m - Y ;Za ]1a
+ [ 1 + y;z, - y2Zm }• (212)
Substituting (209) and (210) in turn in (201) and (202), integrat-

ing, and inserting the constants for the wholc line there is obtained
(213)
'1
(214)
These series may be expanded by the method indicated to

any degree of accuracy desired . It is probable, howevcr,
that the expansion has already been extended sufficiently for ·
all practica! applications. The solution in the form given,
however, does not lend itself well for coordination with other
parts of the network. The desirable form is an equivalent
network which represents the actual system, just as the equiva-
lent 1r represents the simple transmission line. The equivalent
circuit may then be set up on a calculating board or solved by
the usual analytical processes.
It is known that any three-terminal network can be repre-
sented accurately by means of three impedances (or admittances)
connected between terminals and an admittance or impedance
connected to ground at each terminal.* Such a network is
shown in Fig. 103. The value of the network constants can be
obtained by applyi~g unit voltage to one terminal with the two
other terminals grounded. From a knowledge of the currents
fl.owing in the three terminals, the impedance of each of the three
branches connected to the energized terminal can be determined
at once. This process rcpeated at the other terminals enables
one to determine all of the impedances. This general method
will be applied to our present proble.m for the particular case in
which the lines are similar.
For this particular case Za = zb = ZL and Ya = yb = YL,
and equations (211) to (214) may be rewritten in the following
form
Ea = K1Er + K2la + K3Ib (215)
Eb = K1Er + K 3I a + K2lb (216)
1as = K 4Er + K 5Ia + K6h (217)
ha = K 4Er + K6Ia + K5Ib (218)
J
J (219)
* A more general discussion of the method of setting up equivalent networks

is given in Chap. XII.
To determine Zar, Yaa, and Y ah of Fig. 103, let
Ea= 1
Eb = Er =O
Substituting these values in (215) to (218), there is obtained
1 = Kda+ Kah
O= Kala + K2h
l as = Kda + K6Jb
h s = K6la + Ksh
Remembering that la+ lb = Ir, these e quations m ay be
solved with the following results
(221)
In terms of the equivalent constants, it may be seen from

Fig. 103 for the same conditions
1
I =-
r Zar
1 (222)
las = -z
Uar
+ Yab + Yas
Equating equations (221) and (222) and solving for the

unknown equivalent constants
Zar= K2 +Ka (223)
y _ K2K6 - KaK5
ah - Ka 2 - K2 2 (224)
y _ K5 + K6- 1 (225) .
as- K2 +Ka
By symmetry
Zbr = Zar (226)
Ybs = Yas (227)
In a similar manner, with the assumptions that
Er = 1
and
it may be shown that

Yr = 2(K¡- 1) (228)
K2 +Ka
It may be observed from equations (219), that
K5 + K6 - 1 = K1 - 1 (229)
It therefore follows from equations (225), (227), and (229),
that
(230)
The equivalent network thus obtained for the coupled circuits

may be inserted in the complete system network and the currents
at the terminals l as, hs, and I r, determined. During the simpli-
fication process the identity of the currents l a and l b in the line
at the bussing point was lost. If these currents are desired,
they may be obtained by solving equations (217) and (218),
remembering that l a + l b = Ir, with the result that
la
= Ir + I as - I bs (231)
2 2(K5 - K6)
Ir
l as - lbs
h =- - (232)
2 2(Ks- K6)
The expansions expressed by equations (219) have been
carried out to include the first-order terms only. This should
be sufficiently accurate for all practica! problems, since bussing
points are usually employed on the longer lines.
Problems
1. D etermine the positive-sequence constants per mile, z1 and y1, for the
60-cycle transposed transmission line shown in Fig. 89 ncglecting the ground
wire. Assuming that the transmission line is 200 miles in lcngth, determine
the A 1, B 1, and C1 constan ts.
2. Find the equivalent circuit in the 1r and T forros for positive-sequence
for the system describcd in Prob. l.
3. For the case of Prob. 1, determine the zero-sequence constants per mile,
zo and yo, and also the Ao, Bo, and C 0 constants, assuming the average height
above ground is 40 ft., p = 100 meter-ohms.
4. Find the equivalent circuit in the 1r and T forros for zcro-sequence for
the system dcscribed in Prob. 3.
6. Determine the zero-sequence susceptance per mile for the case of Prob.
1 taking into account the prcscnce of the ground wire.
6. Assume that two idcntical transmission lines on separate rights-of-way
be parallel at each end. Neglect the inductive coupling between circuits,
and assume equivalent T networks for each lineare as follows:
CONST ANTS OF LONG TRANSMISSION LINES 197
Positive- and negative-sequence constants
Z equi11. 1 = 20 + j200
and
Yequiv.l = +jO.OOl
Zero-sequence constants
Zequiv.O = 50 + j600
and
Yequi v. O = +j0.0005
If a constant positive-sequence voltage of 69 kv. from line-to-neutral is
maintained at the sending end, find the phase b voltage-to-ground at the
receiver for a fault on phase a at the same point, assuming the load is dis-
connected from the transmission line.
7. With the same system and fault as described in Prob. 6, assume that
the breaker in the faulty line is opened at the sending end leaving the faulty
line connected to the receiver bus. Determine the value of th~ voltage-to-
ground on phase b at the sending end of the faulted line.
8. Considera section of the line illustrated in Fig. 91 and assume that the
left-hand circuit has impressed upon it 30-kv. zero-sequence. Determine
the value of the voltage impressed on the right-hand circuit as the result of
electric induction.
·'
CHAPTER X
CONSTANTS OF CABLES
Practically all of the present-day commercial cables used for
power transmission may be classified in one of the following
groups:
l. Single-conductor.
2. Belted three-conductor:
a. Round conductor.
• b. Sector conductor .
3. Shielded three-conductor (type H):
a. Round conductor.
b. Sector conductor.
The characteristics, both electrical and thermal, of the first
two groups for normal operation are discussed in a very complete
manner by D. M. Simmons in an article* from which much of
the material presented here has been obtained. This volume is
concerned only with the electrical characteristics which are
required for the application of the method of symmetrical
components. The properties of the single-conductor and belted
three-conductor cables will be treated together, and later the
properties of type H cable will be discussed.
An important consideration that requires repetition in con-
nection with cable systems is the principie that regardless of
the complexity of mutual inductive relations between component
parts of individual phases, provided only that the three phases
are symmetrical, the three sequences do not react upon each
other. Three-conductor and type H cables by the nature of
their construction inherently satisfy this condition; single-
conductor cables may or may not. Unsymmetrical spacing
and change in permeability resulting from different phase currents
when certain devices to eliminate sheath currents are used may
introduce di'ssymmetry. In general, the effect has a negligible
* Calculation of the Electrical Problems of Transmission by Underground
Cables, Elec. Jour., p. 366, August, 1925. Subsequently a revision of thi~
article appeared in the Elec. Jour., May to November, 1932.
198
CONSTANTS OF CABLES 199
influence upon short-circuit currents. When single-conductor
armored cables are used, the effect of current upon the per-
meability of the steel may also affect symmetry.
84. Positive- and N egative-sequence Resistance.
The effective resistance to alternating current is larger than
the resistance to direct current. The resistance to direct current
may be taken as the rcsistance of a solid rod of the same cross
section and thc smne lcngth, increased by 2 per cent to take into
account the effccL of spiraling the strands. In three-conductor
cables the twisting of the three conductors about each other
produces an additional increase of about 2 per cent.
Skin Effect and Proximity Effect. The interior filaments
of a conductor are surrounded by a larger number of magnetic
lines of force than ihe filaments nearer the surface. For alternat-
ing current this rcsults in an unequal distribution of current, a
forcing of thc current toward the surface of the conductor, called
skin effect. I'his effect results in a largcr resistance for alter-
nating current ihan for dircct current, the ratio of the two being
called the ski..'l-effect ratio. For small conductors this effect is
negligible, but 1or larger conductors eveu at 60 cycles it may be
quite pronounccd. Table IX, taken from Simmons' article,
T ABLE IX.-DIMENSIOKS AND 60-CYCLE SKIN-EFFECT RA'l'IO OF STRANDED
CoPPER CoNDUCTORS AT 65°C.
Conductor size
Sk in-
Diam- effect
Circular ratio
eter,
mils
in ches
3 , 000 , 000 1 .998 1.439

2 ,500 , 000 1.825 1.336
2,000,000 1.631 1.239
1,500,000 1.412 1.145
1, 000 , 000 1.152 1.068
800,000 1.031 1 . 046
600,000 0.893 1.026
500 ,000 0.814 1.018
400,000 0.728 1.012
300 ,000 0.630 1.006
shows the magnitude of this ratio for a selected number of cables

at 65°C. The ratio varies slightly with temperature.
The alternating magnetic flux in a - conductor due to the

current in a neighboring conductor gives rise to circulating
currents, the effect of which is reflected in an apparent increase
in resistance of the conductor. This effect is known as proximity
effect. Only for the very largest sizes of cables is this effect
important.
Sheath Currents of Single-conductor Cables. The flow of
alternating current in the conductors of single-conductor cables
induces alternating voltages in the cable sheaths. When the
cable sheaths are solidly bonded, the voltages give rise to sheath
currents and therefore additional losses. These losses may
most conveniently be represented by correspondingly increasing
the value of resistance assigned to the conductor and then
ignoring the actual sheath currents. The increase in resistance
for a typical case may be found as follows.
Assume three single-conductor cables arranged at the apexes
of a triangle and let
Xm = mutual reactance in ohms per mile per phase between
conductors and sheaths equal to the voltage per mile
induced in a sheath by unit balanced three-phase
currents in the conductors.
rs = sheath resistance per phase in ohms per mile.
X8 = sheath reactance per phase in ohms per mile.
re = conductor resistance per phase in ohms per mile.
The induced sheath voltage per mile of cable for unit conductor
current is then Xm, and the sheath current resulting therefrom is
VXs2 + rs2
The l 2 R loss in watts or the increase in conductor resistance is
!:ir = Xm2rtt (233)
Xtt
2
+rtt
2
The determination of the mutual reactance Xm is very similar·

to the determination of the ordinary self reactance of three
aerial conductors, the only difference being that for aerial
conductors the flux ·is integrated within the conductor and
between the surface of the conductor and the center of the two
other conductors, whereas for this case the flux is integrated
between the sheath and the centers of the two conductors.
This change in the integration limits results in the expression
CONSTANTS OF CABLES. 201
Xm = 0.2794 log10 ~ in ohms per phase per mile at 60 cycles

T4 rs ·
(234)
in which
S = spacing between conductors in inches.
r 4 = inner radius of lead sheath in inches.
r 6 = outer radius of lead sheath in inches.
The quantity r 4 ~ rs is the mean radius or approximately the

G.M.R. of the sheath.
The self reactance of the sheath x8 is also determined by the
\.
flux between the lead sheath and the centers of the conductors
of the two other phases, since the disposition of the flux outside
the sheath is the same whether unit current flows in the conductor
or the sheath. Hence, the limits of integration are the same
and it follows therefore that, to quite a high degree of accuracy,
XB = Xm (235)
The resistance per phase of the sheath is given by the formula
r8 = 7 ,936s X lo-s
rs2 - r.z
7 9368
_ ~ X 10-6 ohms per phase per mile
(r5 + r4)(r 5- r4)
in which s = resistivity of the lead sheath in microhm per
centimeter cubed; equal at 50°C. to 25.2.
Inserting this value of s
r8 = (
rs
+ 0.2000
)(
r4 rs - r4
) oh ms per p h ase per m1·¡e (236)
Since Xa = Xm the increment in resistance due to sheath currents

18
(237)
in which Xm and T 8 are determined from (234) and (236),

respecti vely.
Thus the resistance to positive- or negative-sequence is
(238)
These considerations assumed a triangular disposition of the

three individual cables. · For other arrangements the geometric
mean separation may be used with results sufficiently accurate

~or most practica! purposes.
Example. Let it be desired to determine the resistance at 60 cycles of
three 1,000,000-cir. mils single-conductor cables having a conductor insula-
tion of H- in., and i in. of lead sheath, run individually in fiber conduits
4.125 in. between centers, all in the same horizontal plane. The geometric
mean separation is
S = ~ (4.125) 2 (8.25) = 5.20 in.
From conductor tables (see Table IX), it may be found that the outside
radius of the conductor is equal to 0.576 in. and, adding the insulation
thickness,
r4 = 1.138 in.
r5 = 1.263 in.
From equation (234)
2(5.20)
Xm = 0.279 log1o 1.
138 + 1. 263
= 0.178 ohm per mile
From equation (236)

0.200
r. = (2.401)(0.125)
= 0.667 ohm per mile
Substituting in equation (237)
(0.178) 2 (0.667)
llr = (0.178)2 + (0.667) 2
= 0.044 ohm pcr mile
From wire tables it may be found that the direct-current resistance at
65°C. is 0.0655, but due to skin effect (sce Table IX) this should be increased
to (1.068) (0.0655) or 0.070 ohm per mile. The total resistance including
the sheath loss is then
r = 0.070 + 0.044 = 0.114 ohm per mile per phase
The sheath loss can conveniently be eliminated by interposing
insulation in the sheath, to prevent a complete sheath circuit
being formed. This may give rise to quite large induced voltages
in the sheaths which introduce a hazard to life or difficulties
dueto electrolysis. Sorne cables are broken up at several points
to limit the induced voltage per scction. Other devices are also
utilized to prevent the flow of the sheath currents without
producing dangerously high voltages. The reader is referred
to cable specialists for more detailed information. P apers by
Halperin and Miller, * and Arnold t will be found useful for this
* HALPERIN and MILLER, Reduction of Sheath Losses in Single-conductor
Cables, Trans. A.I.E.E., vol. 48, p . 399, April, 1929.
t ARNOLD, A. H. M., The Theory of Sheath Losses in Single-conductor
Lead-covered Cables, Jour. I .E.E., vol. 67, p . 69, 1929.
CONST ANTS OF CABLES 203
purpose. Such schemes do not invalidate the use of the simplified
system of symmetrical components as long as the constants
of all the phases are symmetrical. Caution must be exercised
in those schemes involving reactances for limiting the current
in which the permeability is altered by the current strength.
The reactances in the different phases may not be equal in such
cases.
Sheath loss for three-conductor cables is negligibly small
except for the very large cables. The external magnetic field
is almost completely annulled by the close proximity of the three
conductors, but for the larger sizes local fields produce eddy
currents in that portion of the sheath adjacent to the conductor.
85. Zero-sequence Resistance.
The zero-sequence resistance of the conductors alone (neglect-
ing the sheath and ground circuits) is equal to three times the
resistance of the three conductors in parallel. In this case
the total currents in the individual conductors are in phase, so
that the adjacent conductors tend to produce still greater
distortion (proximity effect) of the current distribution. How-
ever, according to unpublished data of F . Wollaston, the net
effect of the proximity effect is very small. This is probably
due to the high contact resistance arising from either surface
films or imperfect mechanical contact which prevents current
flow between adjacent layers. Within individual layers the
stranding also prevents the flow of current by a continuous
transposition of the individual strands. The zero-sequence
resistance of the conductors, per phase, is then equal to the direct-
current resistance of one conductor multiplied by the skin-effect
ratio of one conductor. Currcnts are produced in the sheath
both for single-conductor and three-conductor cables when zero-
sequence currents flow. Sheath losses and the resistance of the
earth-return path, which are involved in the total zero-sequence
resistance, will be discussed in conncction with the zero-sequence
reactance.
86. Positive- and N egative-sequence Reactance.
The positive- and negative-sequence reactance of three-
conductor cables is determined in a manner analogous to that
for aerial conductors. Assuming uniform current distribution
throughout the secti0n of t.be conductor~ the formula becomes
,.
¡
~
TABLE X.-CHARACTERISTICS OF ROUND THREE-CONDUCTOR CABLES (60 CYCLES) o
~
POBitive-negative Zero-sequcnce Sheath POBitive-negative Zero-sequence Sheath

sequence sequence
Circular Resist-
mils Diam- ance, • G.M.R.,
or eter, ohms one Series Shunt ca- Series Series Shuntca- Resist- Series Shuntca- Series Series Shuntca- Resist-
A.W.G. inches per con- react- pacitivet G.M.R., react- resist- pacitivet Thick- ance, react- pacitivet G.M.R., react- resist- pacitivet Thlck- ance,
(B & S) mil e ductor ance, react- three- ance,t ance,§ react- ness, ohms ance, react- three ance.t ance, § react- ness, ohms
ohms ance, con· ohms ohms ance, per ohms ance, con- ohms ohms ance, per
in ches mil
per ohms ductors per per ohms inches mil e per ohms ductors per per ohms e ~
mile per mile mile mile per mile 50°C. mile per mile mile mile per mile 50°C. ~
~
a'\-in. conductor insulation
4\ -in. belt insulation
(1.0 kv.) /r-j.n. conductor insulation
h-in. belt insulation
(1.0 kv.)
~
~
~
~
~
6 0 .184 2 .50 0.067 0 . 177 6, 530 0. 185 0. 307 8.38 12,360 h l. 96
4 .232 l. 58 .084 . 167 5,700 .220 .282 6. 62 10,850 l .. l. 68
3 .260 1.25 .094 .162 5,300 .241 .271 5.99 10, 100 h 1.58 ~
2
1
.292
.332
0.987
.787
.106
.126
. 157
. 152
4,950
4,550
.263
.297
.264
.247
5.40
4.90
9 ,500
8,800
e"i.
8
44.
1.47
1.37 8
00
o .373
.418
.618
.494
.141
.159
.150
.146
4,250
3.950
.327
.361
.239
.230
4.43
4.03
8,250
7, 750
d~
4~
1.27
1. 18
~
o
()()()
0000
.470
.528
.391
.310
.178
.200
. 143
.140
3,700
3,350
.398
.439
.222
.217
3 .66
3.31
7,200
6,620 ..,
l.t 1.09
1.00
<:
~
250,000 .575 .266 .221 ..... . .. .. . . . .. . .... . . .. ... .. . .. . ... 0.144 3,800 0.490 0.223 2.53 7,560 a~ 0.755 <:
300,000
350,000
.630
. 681
.220
. 190
.242
.262
.....
.. ...
. ....
.. ...
. ....
... ..
. ....
. ....
. ...
. ...
......
. .. ...
..
..
. ...
. ...
.141
.140
3,600
3,450
.531
.568
.217
.213
2.43
2.28
7,120
6,800
e'\
h
. 737
.698 ~
400,000 .728 .166 .280 . ... . . .... . .... . .... . ... . ..... .. . ... .138 3,250 .601 .210 2.16 6, 510 ñ .664
450,000 .772 .148 .297 . .. .. . .... . .... . .... . ... ...... .. . ... . 137 3,150 .634 .204 2.05 6,220 -h .636
500,000 .814 . 134 .313 . ... . . .... . .... . .... . ... . ..... .. . ... . 136 3 ,000 .665 .201 l . 97 6,000 ñ .611
• Alternating-current reslstance based upon 100 per cent conducthity at 65°C. lncluding 2 per cent allowance for splral of strands and 2 per cent allowance for splral of
conductora.
t For specific inductive capacity = 3.5.
i Reactance based upon all return current In the sheath; none In the ground.
§ Resistance based upon all return current In the sheath; none In the ground.
~4&!, 32! ...~~~. :=:~~~--,~~@ )4~-AZQ~~-~%WKX~~~- --~-- ~-·----· - ·-~--~-
T ABLE X.-CHARACTERISTICS OF RouND THREE-CONDUCTOR CABLES (60 CYcLEs).-(Continued)
Positive-negative Positive-negative
Zero-sequence Sheath Zero-sequence Sheath
sequen ce sequen ce
Circular Resist-
mils Diam- ance,• G.M.R..
or eter, obm.s one Series Shunt ca- Series Series Shuntca- Resist- Series Shuntca- Series Series Shunt ca- Resist-
A.W.G. inches per con- react- pacitivet G.M.R., react- resist- pacitivet Thick- ance, react- pacitivet G.M.R., react- resist- pacitiver ance,
ance,§ reactance, reactance,§ Thick-
(B & S) mile ductor ance, react- three- ance,t ness, obm.s three ance,t react-
ness,
ohms
obm.s ance, con- obm.s obm.s ance, per ohms ance, con- ohms obm.s ance, per
per obm.s ductors per per obm.s
inches mile per ohms ductors per per ohms
inches mil e
mile per mile mil e mil e per mile 50°C. mil e per mile mile mile per mile 50°C.
1 ~
a
8°,¡ -Ía.
conductor insulation (2.0 to 3.0 kv. incl.) ll i.-in. conductor insulation ( 4.0 to 6.0 kv. incL) ~
j¡-in. belt insulation &4.¡-Ín. belt insulation V:¡
1 ~
~
6 0.184 2.50 0.067 0.200 8,350 0 .208 0.326 7.21 14 ,950 ~~ 1.57 0 .21J 9,150 0 .220 0.339 6 . 79 16, 470 t. 1.43 ~
4 .232 1.58 .084 . 186 7,250 .245 .301 5. 87 13,350 {'¡ 1.43 .195 8,000 .257 .315 5.51 14,800 ll l. 31 ~
3 .260 1.25 .094 . 180 6,800 .266 .291 5.33 12,600 ~'\ 1.36 .188 7,500 .278 .304 5.00 14,000 ~fl;¡ l. 25 V:¡
2 . 292 0.987 . 106 . 174 6,400 . 289 .282 4. 8.3 12 , 000 611_¡ 1.28 .182 6,950 .302 .293 4 .53 13, 100 e"• 1.18 a
~
1 .332 .787 .126 .168 5, 850 .323 .264 4 .39 11 , 120 /¡ l. 20 .176 6,450 .336 .276 4.15 12,400 l'• 1.12
o .373 . 618 .141 .163 5,450 .353 .256 3.98 10,360 118_¡ 1.12 . 171 6,000 .367 .266 3.77 11,650 &S. 1.05 ~
~
üO .418 .494 .159 .160 5,050 .388 .246 3.64 9, 700 t. 1.05 .166 5, 650 .401 .256 3.46 10,900
~··
0 . 99 t:::o
000 . 470 .391 . 178 .155 4 , 750 .426 .237 3.33 9,100 t .. 0.98 .162 5.250 .439 .248 3. 15 10,200 .92
~
88_¡
0000 .528 . 310 . 200 .152 4 , 400 .467 .231 3.04 8,550 •
114 .91 .157 4 ,900 .481 .239 2.89 9,550 e\ .86
250,000 .575 .266 .221 .149 4,150 .504 .223 2.85 8.150 •
~¡ .86 .154 4,650 .518 .235 2.43 9,130 ñ .72
300.000 . 630 .220 .242 .145 3, 950 . 544 .221 2.35 7. 700 -h .71 .150 4 , 400 .558 .229 2.26 8,650 ñ . 68
350,000 . 681 .190 .262 .143 3,750 .582 .216 2.23 7,330 -h .68 .149 4,200 .596 .223 2.11 8,300 ñ .64
400, 000 .728 .166 .280 . 142 3,600 .616 .212 2.09 7,000 ñ .64 .147 4,000 .630 .220 2 . 03 7,960 -Ir .62
450, 000 .772 .143 .297 .141 3,450 .648 .209 2 .01 6.800 ñ .62 . 145 3,900 .663 .217 l . 92 7, 700 -Ir .59
500, 000 .814 .134 .313 .140 3,350 .578 1 .206 l. 90 6,500 ñ .59 .143 3,750 .693 .213 1. 84 7, 450 ñ .57
- - - - - -- - - - - - --------- -
• Alternating-current reslstance based upon 100 per cent conductivity at 65°C. including 2 per cent allowance for spiral of strands and 2 per cent allowance for spiral of
conductora.
t For specific inductive capacity = 3.5.
t Reactance based upon all return current in the sheath: none in the ground. ~
§ Resistance based upon all return current in the sheath: none in the ground. 0
c.n
~
TADLE X.-CHARACTERisTrcs OF RouND THREE-CONDUCTOR CABLES (60 CYCLEs).-(Continued) o
O')
Positive-negative Zer<H!equence Sheath 1 Positive-negative Zero-sequence Sheath

sequen ce sequence
Circular Resist-
mils Diam- a nce, • G.M.R.,
or eter, ohms one Series Shunt ca- Series Series Shuntca- Resist- Series Shunt ca- Series Series Shuntca- Resist-
A.W.G. in ches per con- react- pacitivet G.M.R., react- resist- pacitivet ance, react- pacitivet G.M.R., react- resiBt- pacitivet ance,
Thick- ance, ance,§ Thick-
(B & S) mile ductor ance, react- three- ance, t ance,§ react- neas, ohms react- three ance,t react- ness, ohms
ohms ance, con- ohms ohms ance, per ohms ance, con- ohms ohms ance, per
inches per per per in ches l'-2
per ohms ductors per per ohms mile ohms ductors ohms mi le
~
mile per mile mil e mile per mile 50°C. mile per mile mile mile per mile 50°C.
6"4 -in.
l~-in.
conductor lnsulation
belt insulation
(7.0 to 7.5 kv. incl.) H-in. conductor insulation
h -in. belt insulation
(15.0 kv.;
~
~
~
.......
6 0 . 184 2 . 50 0 .067 0 .219 9,800 0.233 0 .340 6.55 16.950 684 l. 35 0.242 11,500 0 .264 0 .359 5.89 18,900 h 1.13 ~
~
4 .232 l. 58 .084 .203 8,650 .270 .317 5.30 15,350 8
64 1.24 .224 10,200 .303 .335 4.73 17,300 e"i 1.05
3 .260 1.25 .094 .195 8,050 .291 .306 4. 82 14,500 •
84 1.19 .216 9,600 . 325 . 323 4 . 28 16,500 .~ 1.01
2
1
.292
.332
0.987
.787
.106
.126
. 189
.182
7, 500
6,950
.315
.349
.295
.278
4 .38
3.97
13,700
12,850
u
R
l4
1.13
1.06
.209
.200
9,050
8,400
.350
.385
.313
.297
3.90
3.55
15,650
14,750
h
8~
0 . 97
.92 8
00
o .373
.418
.618
.494
.141
.159
.176
.171
6,450
6,000
.380
.414
.269
. 260
3.65
3.34
12, 200
11,450
•••
•
1.01
0.95
.194
.188
7, 800
7,250
.417
.451
.284
.275
3.26
2.98
13 ,940
13,200
l\
.fl;¡
.88
.83
~
a
64
()()()
0000
.470 .391 .178 .166 5,650 .452 . 251 3 .06
2 .50
10, 750
10,200
8
••
l.
.89
.73
.181 6, 750
6, 300
.490
.533
.268
.259
2.46
2.26
12, 480
11,800
-h .69 <:
~~
.528 .310 .200 .161 5,250 .495 .245 . 175 -h .65
250,000 .575 .266 .221 .158 5,000 ..132 .238 2.37 9, 550 .¡. .70 .171 6,000 .572 .250 2.16 11 ,250 .¡. .63
300,000 .630 .220 .242 .154 4.750 .572 .232 2.20 9,100 .¡. .66 .167 5,650 .612 .244 l. 99 10,600 .¡. .59 l'-2
350,000 .681 .190 .262 .151 4,500 .610 . 226 2.05 8,700 .¡. .62 .164 5, 400 .650 .238 l. 90 10 ,200 ñ .57
400,000
450,000
500, 000
.728
.772
.814
- -
.166
.148
.134
.280
.297
.313
. 149
. 148
. 146
1
4.300
4,150
4,000
--
.644
.677
.708
------
. 222
.219
.215
1.97
1.89
1.81
8,400
8,100
7,850 ..
.¡.
-h
•
.60
. 58.
. 56
.161
.160
. 157
5.200
5.000
4,850
. 684
. 717
.748
.234
.230
.224
l. 82
l. 74
1.66
9,800
9.550
9,250
-h
.¡.
-h
-
. 55
.53
.51
-- -
* Alternating-current resistance based upon 100 per cent conductivity at 65°C. including 2 per cent allowance for spiral of strands and 2 per cent allowance for spíral of
conductora.
t For specific lnductive capacity = 3.5.
t Reactance based upon all return current in the sheath; none in the ground.
§ Resistance based upon al! return current in the sheath; none in the ground.
0.2794 log1 o 0.~

x =
9a ohms per mile per phase for 60 cycles
(239)
in which a = outside radius of the conductors and 0.779 = factor
to convert to G.M.ll. * For cables it is convenient to express S,
a, and the G.M.R. in inches.
The departure from uniform distribution, while important in
its effect upon the resistance, is of little consequence in its effect
upon the reactance. In Table X are evaluated the reactances
for the most common range of cable sizes met in practice. These
reactanccs were calculated by equation (239) for a frequency of
60 cycles.
The positive- and negative-sequence reactances for single-
conductor cables, when sheath currents are present, may be
determined from the following expression:
X = 0.2794 log10 _
S
0 779
a Xm 2
X 3
+
r 82 ohms per mile per phase
(240)
The last term represents the correction for the presence of
sheath currents. The negative sign arises from the fact that
the current in thc sheath is in a direction opposite to that in
the conductor, thus tending to limit the flux to the region
between the conductor and the sheath. The last term corre-
sponds to equation (238) with Xm substituted for rs and is derived
by considering the current in the sheath and the component
of voltage which it induces in the conductor in quadrature to
the conductor current.
For sector-shaped conductors no accuratc data are available,
but Simmons may be quoted as authority for the statement
that the reactance is from 5 to 10 per cent lower than for round
conductors of the same area and same thickness of insulation.
87. Zero-sequence Reactance.

Cable sheaths commonly have definite connections to ground
at several points and considerable leakage to ground throughout
the length of the cable ducts. Hence it is obvious that the
* For most underground cahles the numbcr of strands is so large that the
factor O.779, b ased on solid conductors, may be used without appreciable
error. For cables with a small number of strands the factors from Table
VI, Sec. 61, may be used or the G.M.R. may be used directly in inches from
Table X.
method of making these connections and their resistances will

have important effects upon the zero-sequence path. Calcula-
tions of the zero-sequence impedance can be made practically
for two cases as follows:
Case l. Return current in sheath and earth in parallel.
Case 2. Return current in sheath alone.
The actual value will lie between the limits given by these
cases, depending upon the length, the grounding, and leakage.
In Case 1 it is assumed that the ground connections and length
of cable are such that maximum earth current
~~\~ obtains; in other words, the calculations are based
1
Nof greafer upon the distribution of current between sheath and
fhcm !Orf.
earth which will occur in the middle of a long section
of cable with well~grounded sheath. In Case 2 the
return current may be confi.ned to the sheath because
of the relatively high resistance of ground connections
FIG. 104.-
and the small effect of leakage in short lengths of
Cable in a cables in well-dri.ed cable ducts. The method
buried duct. of analysis to be employed will make use of an
equivalent circuit from which the zero-sequence impedance

can be obtained for both cases.
For the analysis of the zero-sequence rea\!tance of cables,
consider the conductor shown in Fig. 104, completely buried
in the ground with a non-magnetic non-conducting materia]
insulating the conductor from ground. The radius of the
conductor will be denoLed by a and the raJdius of the intervening
space by b. Since the conductors are usually located near the
surface of the earth, one would expect considerable dissymmetry
in the current distribution in the earth. Carson * has shown,
however, that for the cases met in practice the effect' of the
changed current distribution is not very great, increasing the
earth-return part of the impedance ·on the arder of 5 to 10 per
cent over the value that would be obtained if the earth extended
indefi.nitely above as well as below the conductor. In view
of the fact that cables are usually supplied with lead sheaths
which are then in parallel with the earth circuit whose impedance
is much higher than the resistance of the sheaths, it follows that
it is unnecessary to determine the earth-return circuit precisely.
Hence, the above correction can in general be neglected.
* CARSON, JoHN R., Ground Return Impcdance: Undcrground Wire with
Earth Return, Bell System Tech. Jour., p. 94, January, 1929.
Carson's results are surprising, in that one would at first
expect that the impedance of the conductor when near the sur-
face of the earth would be much larger than if completely sur-
rounded by earth of infinite extent in all directions. However,
Rudenberg0 37> has arrived at this same result by a different
method of reasoning. Rudenberg has further shown that the
impedance of such a symmetrical circuit due to the flux in the
earth is, for commercial frequencies and normal ground resistance,
and for unit length, equal to
.-'f + j2wlog.
0 17
( - :~ in absolute units (241)
The reactance due to the flux in the air and in the conductor is
or
(242)
Adding the conductor resistance to expressions (241) and

(242), the total self impedance of the conductor and ground
return is
- 2 • 0.178-J.j
Z11 - Tconductor + 7r 'f + J2w logE 0. 779a
in absolute units.
= Tconduc tor + 0.00159/ + j0.004657J log10
2,160-J.j
(G.M .R. )conauc tor in ohms per mile (243)
This expression is identical with that* which determines the

impedance of aerial wires with ground return. The latter is
practically independent of the height of conductor above the
ground, and this expression is independent of the radius of air
section in which the conductor lies. Equation (243) can m
turn be reduced to the form of equation (169) giving
* See equation (163), Sec. 64.
Za - rconductor + 0.00159j + j0.004657J log1o G.~~R.

ohms per mile (244)
in which De = 2,160~ [Eq. (166)]

In a similar manner it may be shown that the mutual imped-
ance between two conductors in the cylindrical air tunnel can
be reduced to the form of equation (170).
Zum = 0.00159f + j0.004657f log1o dDe ohms per mile

ab
(245)
The expressions (244) and (245), changed to per phase values,

are represented in curve
form in Figs. 82(a), (b) and
(e), in which De, G.M.R.,
and dab are expressed in
feet, and p in meter-ohms.
With these re la t io ns
esta blishe d, the zero-
seq u en ce impedance of
cables can be determined
("') Geomefric 0\sposition in a manner analogous to
that used for the case of
the transmission line with
ground wires. *
The geometric configu-
ration of the con.ductors of
a three-conductor cable in
(b) E:qu•v"\ent Diagncom
an underground duct is
shown in Fig. 105(a) . The
distance from the surface
is immaterial. The equiv-
(o) Equiv<>len+ Di<>gmm when Shec>th is Grounded alent circuit representing
Fxo. 105.-Calculation of earth or sheath the impedance between con-
return impedance of buried cable. For
sheath or ground return circuits use these ductors, sheath, and ground
values directly and for zero-sequence per phaso is indicated in Fig. 105 (b).
multiply by 3. Resulta are in ohms per mile.
The mutual impedance
between the sheath and conductors considered as a unit is given
by equation (245), in which the geometric mean spacing 1s
* For a discussion of the equivalent circuits used, refer to Sec. 73.
approximately r 4 ~ r
6
This determines the star branch next
to ground. The self impedance between conductors and ground
is given by equation (244). This self impedance is equal to the
¡.,
sum of the star branches in the conductor and ground legs, so
that the impedance associated with the conductor leg is the
difference between the self and mutual impedances, or
rconductors + 0.00159j + j0.004657j log10 (G.M.~econductor/J

2
- 0.00159j - j0.004657j log1o
r4
~er6 = rconductm +
. r4 + r5
J0.004657j loglo 2(G.M.R)conductors (246)
Similar considerations apply for the impedance of the sheath
branch, but in this case the self and mutual impedances are
equal, except for the a512sh«xff>!Jra~~cll
sheath resistance.
With the sheath
grounded at both ends,
the impedance between
conductor and ground r
is obtained 'by solving (01)
the circuit shown in Fig. FIG. 106.-Calculation of the earth and sheath
105(c). The distribu- return impedances of a special 450,000-cir. mil
threo-conductor cable. The resistance of each
tion of current between conductor at 65°C. is 0.146 ohm per mile. For
sheath and ground can 60 cycles and damp earth, De = 2,800 ft. For
sheath or earth return use the foregoing values
also be determined from directly. For zero-sequence multiply by 3 to
this circuit. obtain results in ohms per mile per phase.
If the sheath is grounded at one end only, the voltage between
sheath and ground per ampere per mile is equal to the mutual
impedance.
The zero-sequence impedance is obtained by multiplying the
above quantities by 3.
Example l. As an illustration of such a case, let it be desired
to obtain the impedance of a single three-conductor cable, whose
characteristics are shown by the section of cable in Fig. 106,
when buried in damp earth: De = 2,800 ft.; resistance per
conductor = 0.146 ohm per mile, 60 cycles.
G.M.R. of three conductors = ~(0.779)(0.386)(1.210) 2 =
0.761 in.
.
G.M. spacmg, h h
conductor to s eat =
r4 + rs -
1.5 + 1.625
2 2
= 1.5625 in.
2,800 X 12
Mutual reactance = 0.004657 X 60 log1o 1. = 1.21
5625
Reactance of conductor branch =
1.5625
0.004657 X 60 log10 . = 0.0874
0 761
0
Resistance of conductor branch = ·~46 = 0.0487
3 125 0 . 125
. °~ 0
Resistance of sheath branch = = 0.512
Mutual resistance = 0.00159 X 60 = 0.095 ohm.

The equivalent network is shown in Fig. 106(b).
Paralleling the sheath and ground branches and adding the
impedance of the conductor branch,
0.5 12 (0 ·095 + jL 210 ) + co .0487 + J·o .0874) = o.474 + J·o .260

0.512 + (0.095 + jl.210)
The zero-sequence impedance is then
3(0.474 + j0.260) = 1.42 + j0.78 ohms per mile per phase
The absolute value of this impedance is 1.62 ohms per mile per
phase
Comparing this value of zero-sequence impedance with the
positive- and negative-sequence impedances per phase
(0.146 + j0.169),
it will be observed that for this particular case the ratio of zero-
sequence resistance to positive-sequence resistance is 10 and the
ratio of zero-sequence reactance ·to positive-sequence 1eactance
is 4.6.
If the cable is not grounded or the ground connections are of
such high resistance that no appreciable current flows through
the ground, then the zero-sequence impedance is the sum of the
impedances in the conductor and shcath branches, namely,
(0.0487 + j0.0874) + 0.512 = 0.561 + j0.0874
or in terms of ohms per mile per phase, (1.68 + j0.262). The
absolute value of this impedance is 1.70 ohms, which is practically
equal to that for the ground connections of zero resistance.
The impedance diagram in which the two conditions are repre-

sented by closing and opening the switch, and the vector repre-
sentation of these impedances, are shown in Fig. 106(b).
Example 2. To illustrate further these calculations, determina-
tion of the zero-sequence impedance per phase of the four three-
0.128 0.137 +j 0.037

ohms per mile
~
0.095+ji.029
G.M.R. oí three conductors in one cable = ~(0.779)(0.386)(1.210) 2 =
0.761 in.
G.M.R. of conductors of four circuits = '{/(0.761) 4 (7.5) 6 (15) 4 (22.5) 2 =
5.775 in.
. r4 + r5
G.M.D. spacing three conductora . to surroundmg sheath = =
2
1.5 +2 1.625 = 1 .563 m. .
G.M.D. spacing of conductora and sheaths ~(1.563) 4 (7.5) 6 (15) 4 (22.5) 2 =

=
6.913 in.
2 800 X 12
Mutual reactancc = 0.004657 X 60 log 10 ' . = 1.029 ohms
6 913
6.913
Reactancc of conductor branch = 0.004657 X 60 log1o _ = 0.0218 ohm
5 775
Resistance of conductor branch = ~· ~: = 0.01217 ohm
X .~2~ ~ _
3
0
Resistancc of sheath branch = = 0.128 ohm
4 0 125
Mutual resistancc = 0.00159 X 60 = 0.095 ohm
Zero-sequence impedance (no ground conncction) =
3(0.01217 + j0.0218 + 0.128) = 0.421 + j0.066
Zero-sequcnce impedancc (with ground conn ection) = 3(0.137 + j0.37) =
0.411 + j0.111
Fw. 107.-Calculation of zero-sequence impedance of four three-conductor
cables in parallel. A description of the individual cable is given in Fig. 106.
conductor cables in the conduits shown in Fig. 107 will be made.

The calculations in the figure are self-explanatory. The results
show that if the ground connections have zero resistance the zero-
sequ ence impedance per phase is 0.411 +
jO.lll ohm per mile,
the absolute value of which is 0.425. If the connection resistance
is very high the cable impedance approaches the value corre~

sponding to the condition in which no current flows through the
ground. The cable impedance for this condition is
0.421 + j0.066,
the absolute value of which is 0.426. Here again the absolute
value is practically unchanged by the nature of the ground
connection.
From Fig. 105 it may be seen that the zero-sequence impedance
for lead-covered cables with no ground return circuit (sheath
only) is
r 4~( num b er of sh eaths) + 3(rparalleled conductora)

60
Zo = ( r5 + r 4)(r 5 -
+ j0.014J loglO (G.M.D.)separation (24 7)
(G.M.R.)conductors
This formula is usually suffi.ciently accurate for practical
purposes because the nature of the ground return circuit for
cable systems is usually quite indefinite, being mixed up with
water-pipe circuits, and the difficulty of assuring a ground
connection which is low relative to the sheath resistance.
In Table X are given the zero-sequence series reactances for
the commonly used three-conductor cables assuming no earth-
return currents. The sheath resistances are also given which
will be of assistance in calculating the zero-sequence impedances,
for example, assuming no ground return circuit the zero-sequence
resistance of a single three-conductor cable is equal to the resist-
ance of one conductor plus three times the given sheath resistance.
The G.M.R. of the individual conductors and of the group will
also be of assistance in these calculations.
The zero-sequence impedance of solidly bonded single-con-
ductor cables can be obtained in a manner analogous to that
described for the four three-conductor cables of Example 2.
Where special devices are employed to reduce sheath losses, the
zero-sequence impedance is largely dependent upon thc degree to
which this is accomplished. When the sheaths are completely
open-circuited, the return current must flow through the ground.
lmpedance devices have been used which represent a mean
between solidly bonding and open-circuiting. Other devices are
arranged to insert high impedance in the sheath circuit for the
flow of positive-sequence current and low impedances for the flow
of zero-sequence current.
88. Shunt Capacitive Reactance.
In Fig. 108 are given curves for the determination of the posi-
tive-, negative-, and zero-sequence shunt capacitive reactance
for single-conductor and belted three-conductor cables. The
results in each case are given in terms of ohms to neutral per
phase. This was done for the two-fold purpose of providing ·
a convenient scale and to present the data in the most usable
form for the most common application of symmetrical compo-
nents: short-circuit studies. These particular curves were
calculated for 60 cycles and for a uniform specific inductive
capacity of 3.5. For other values of these constants the capa-ci-
tive reactance can be calculated by the formula indica~ed in
the caption under the figure. The data* from which these
curves were prepared are based u pon semigraphical and analytical
calculations of geometric factors of cables which indicate an
\
accuracy to within 1 per cent.
For three-conductor cables the positive- and negative-sequence
values differ from the zero-sequence values, but for single-
conductor cables these values are all equal. The capacitance
of sectored cables is larger than for the equívalent round con-
ductor cables, so that Xc is smaller. The correction factor
to be applied to the constants for round conductor cables of
the same conductor section and insulation thickness to obtain
the corresponding constants for the sectored belted cable is
given by the bottom curve in Fig. 108.
The foregoing premised a uniform insulating material, which
may not always be true. Further, a value of k = 3.5 was
assumed. Most paper-insulated cables run about 10 per cent
less than this value. The usual range is given in Table XI.
T ABLE XL-V ALUES OF SPECH'IC lNDUCTIVE CAPACITY

Material Value of k
Impregnated paper. . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 . O to 4 . O
Varnished cambric ........ ... . . ..... . .... . . ... 4.0 to 6 .0
Rubber . ...................... .. ............. 4.0 to 9.0
In Table X are given the positive-, negative-, and zero-

sequence shunt capacitive reactances for the commonly used
three-conductor cables.
* Calculation of the Electrical Problems of Transmission by Underground
Cables, Elec. Jour., p. 366, August, 1925.
t
T
.8
.6
.2
t
T
.o
.4
o.
l.
"'
~ 16
~
o 14
tí
::1
'tl
g 12
u
~10
e
u;
... 8
.E
.!! ,_o
i 6
~~
;;:;" Coáection actor for Sector Cable
To be pplie~ to ound Cond6ctor ~ables
[ ,.._q, of Equal opper Are and sula'tion hickn'ess
4
"'E
.e
Cj<:-
o 2
S
u
X o 0
o.o 0.2 0.4 0.6 0 .8 1.0 1.2 1.4 1.6 1.8 2.0 2.2 2.4
T +t
Ratio - d -
Fw. 108a.-Shunt capacitive r eactance to neutral of single- and three-conductor
cables in ohms per mile for 60 cycles.
X, _ 3.5 60 1 X
e - k f length in miles e
89. Shielded or Type H Cable.

In typc H cable, copper tape is wound around all or a portion
of the conductor insulation. A typical construction of this
cable is shown in Fig. 109. The purpose of the tape is to control
the electrostatic stress, reduce corona formation, and decrease
15,000
14,000
Ql
:0
8 12,000 1---+--+--hmrmrrll:
L.
-t 11,000 l---+--+--lftl!li!1~
::>
1:1
fjiO,OOO
u
Q)
~ 9,000
J5
c:s
F
u J!!
L. ~
~
u
::>
"U
S: t/1
o E
u .S::
1
QJ
o
g>10,000
lñ
~ 8,000
i:
ti; 6,000
0..
lf)
j4,000
o
.!; 2,000
u
X
o
W ~ ~ M M ~ ~ M ~ W
T+t
Ratio· 7
Fw. 108b.- Shunt capacitive reactance to neutral of single- and three-conductor
cables in ohms per mile for 60 cycles.
X, _ 3.5 60 1 X
e - k f lcngth in miles e
thc thermal resistance. The tape is made very thin in order

to minimize the circulating current under normal operating
conditions and thus minimize the loss.
The effect of these circulating currents upon the positive-
and negative-sequence reactance and resistance can be calculated
by the method discussed for determining the effect of sheath
currents for three single-conductor lead-covered cables. For

short-circuit calculations these currents may be neglected.
The zero-sequence impedance is likewise affected to only a very
small extent by the circulating currents in the shields. Calcula-
tions of the impedance between the parallelcd conductors with
the paralleled sheath and shields as return for a three-conductor
450,000-cir. mil type H cable show that only 1 per cent of the
total current flows through the shields. This effect may there-
fore be neglected and the impedance
calculated as though the shields were
not present.
In practice the shields and sheath
l are connected together at f re que n t
points along the cable so that the
shunt capacitive reactance may be
calculated on the basis of three single-
conductor cables in which the shields
represent the sheaths. The shunt
capacitive reactances per phase will
be the same for the positive-, negative-,
and zero-sequences. In the absence
Fw. 109.-Three-conductor of constants for sectored cables, the
type H cable. shunt capacitive reactance of a single-
conductor cable having the same insulation thickness T and
the same mean conductor insulation periphery is a very
close approximation. If the mean periphery of the conductor
insulation is designated P, then the diameter of the round
conductor having the same insulation thickness is
p
d = 7r--T
and
(248)
The shunt capacitive reactance for the positive-, negative-,

and zero-sequences can then be determined from the curve for
single-conductor cable in Fig. 108. The same data are replotted
in Fig. 110 directly as a function of ~·
90. Submarine Cable.
For cables lying on the sea floor, at considerable depth below
the surface, the return current path includes not only the sheath
and earth but also the sea water. Using Rudenberg's assump-
tion of radial symmetry of current fl.ow in both the sea water
and the earth, the current distribution will not differ from that
for the current flow xt,ooo
in either the sea water 14
1_1
~
or the earth alone.
' \~·: ~T
13
The voltage drops in 12 - (
both mediums must be 11 ~~y
equal at all points of
their surface contacts, ~ 9
10 -
from which it follows :i:L. 8
that their curren t ~7 Ul
densities at the contact
E6
surfaces, and hence the .5o 5
division of current '><4 ()
between the two medi-

3
ums, must vary
2
inversely as their
respective· resistivities.
0
Since from Table VII 0 0.02 0.04 0.06 OD8 0.10 0.12 0.14 0.\6 0.18 0.20 0.22 0.24
T
.
1
the resistivity of sea 1i
water is approximately Fw. 110.-Positive-, negative-, and zero-se-
one-hundredth that of quence shunt capacitive r eactance to neutral for
type H cables in ohms per mile for 60 cycles
damp earth, practically and k = 3.5.
all the current fiows through the sea water. Thus the impedance
may be determined by using the value of p corresponding to sea
water in the formulas previously outlined:
91. Constants of Typical Cables.

In Table XII are shown the constants, calculated by the
methods outlined, of a number of typical cables. These values
should be useful as an approximate check upon calculations, as
they enable one to orient his sense of values. The effect of the
elimination of sheath currents is reflected principally in the values
of the positive-sequence resistance and is important only when
the spacing between the single-conductor cables is large.
TABLE XII.-CONSTANTS OF TYPICAL CABLES
(In ohms per phase per mile) ~
~
o
Zero-sequence, sheath
Positive- and negative-sequence return only, solidly
bonded sheaths
Fre- Volt-
Cable A-c. resistancet 1 Series reactance Shunt Shunt
quency age
capa- R esist- Series capac- V;¡
citive react- itive ~
No Include No Include ancet ~
sheath sheath sheath sheath reactance react-
current current current current ance ance ~
~
~
......
Single conductor; 1,000,000 cir. mils;
U-in. insulation; i-in. sheath; three
cables spaced 4.125 in. horizontally o .. 60 26,000 0.070 0.114 0.297 0 . 285 5,800 0.737 0 . 119 5,800
~
~
Single conductor; 500,000 cir. mils; 670 C>
mils insulation; l 4 -in. sheath; three
cables spaced 6, 6, 8.49 in ..... ...... 60 138,000 0.134 0.209 0.386 0.358 8,300 0.752 0.158 8,300
~
C>
Single conductors; 250,000 cir. mils; ~
tz;j
1'8"2"-in. insulation; -l4 -in. sheath; three ~
~
sheaths in contact and 0000 copper V;¡
neutral wire . ............ o •••. • o o . . 60 210 0.264 0.239 0.181 0 .180 2,400 1.45* 0069* 2,400
Three conductor; type H round; 500,000
. m ils,. 61 Í -m.
crr. . msu
. l a t'1on .... , o • . • o. 60 15,00Q 0.134 . .. .. 0.168 . .... 3,600 l. 74 00220 3,600
Three conductor; belted round; 500,000
cir. mils; / 4 in. conductor insulation;
6\-in. belt insulation ..... o . o ..•...•. 60 7,500 0 . 134 ...... 0.146 . .... . 4,000 1.82 0.215 7,850
--- - -- - - - - - - ·· - - -·· - -- - -- - - - - -
*Neutral wire included in sheath circuit.

t Conductor at 65"Co, sheath at 50°Co
Problems
l. Assume three single-conductor cables of 4/0 stranded copper con-
ductors with H-in. insulation and a t-in. lead sheath. The three cables
composing the circuit are arranged to give an equilateral spacing of 8 in.
between centers. Determine the positive-, negative- and zero-sequence
1.mpedances, assuming thc cable sheaths are insulated from ground:
2. What will be the per ccnt rise in voltage at the open end of a 30-mile
section of the three-phase cable described in Fig. 106 when energized by
balanced voltages? Use a method of calculation similar to that discussed
in Sec. 78.
3. Determine the zero-sequence impedance of the three-conductor cable
of Fig. 106 if the sheath is parallelcd by a 4/0 stranded copper conductor
located 8 in. from the center of the cable, assuming no return current in the
earth.
4. A communication circuit located in a lead sheath cable parallels a
power cable for 1 mile at a separation of 100 ft. Determine the voltage
induced in the telephone wires for the case of a conductor-to-ground fault
causing 1,000 amp. to flow in the conductor through the parallel. Assume
that the power and telephone sheaths are grounded through connections of
zero resistance.
5. Assume the three-conductor cable as shown in Fig. 106 with the sheath
solidly grounded. The voltage impressed on this cable has a third harmonic
of 180 cycles which constitutes a zero-sequence. Find the lengths of cable
which will give open-circuit and short-circuit resonances to the third har-
monic voltage. Assume f = 60 and p = 200.
6. Assume a three-phase circuit consisting of th:tee single-conductor cables
with sheath return and a source of negligible impedance. Determine the
analytical expression for the current for a line-to-sheath fault by the method
of symmetrical components and compare thc result with that by the single-
phase method. Show that the analytical expressions obtained by these two
methods are identical.
éHAPTER Xi
POWER SYSTEM VOLTAGES AND CURRENTS UNDER
FAULT CONDITIONS*
It has been rather common practice to base the requirements

of system apparatus on normal load conditions and on balanced
three-phase short-circuits. The effects of unbalanced faults
have been either ignored or considered to a limited extent only.
However, as shown in this chapter, two systems of the same
voltage, which, for particular fault locations have identical
values of three-phase short-circuit current, may have radically
different voltages and currents for unbalanced faults. Thus,
the relations between the sequence impedances havean important
bearing on many applications and may result in higher currents
than for three-phase short-circuits or higher voltages than for
ordinary operation. High currents are important because of
mechanical stresses in rotating machines, transformers, reactors,
circuit-breakers, and bus structures. High voltages are impor-
tant in apparatus or line insulation and in lightning arresters.
Combinations of voltage and current are important in circuit-
breaker and relay applications. The relation between the
sequence impedances or the range of voltages and currents
under unbalanced fault is important in system problems such
as grounding, stability, and inductive coordination with com-
munication circuits.
The value of voltages and currents that obtain under abnormal
conditions on power systems thus is of great importance in
varied applications. In previous chapters the methods of
analysis have been described and boundary conditions for
certain limiting cases have been defined. This chapter gives a
systematic presentation of the ranges of voltages and currents
that may occur on a power system under fault conditions.
92. Assumptions.
In this study, the system will be assumed to be symmetrical
up to the point of fault, or stated otherwise, the system will b~
* Based on a paper of the same title by R. D. Evans and S. H. Wright. (69 )
~22
POWER SYSTEM VOLT AGES AND CURRENTS 223
assumed symmetrical except for the fault itself. The types of
faults which will be considered are as follows:
l. Three-phase (3<P):
.
í
Either line-to-line-to-line (L-L-L), or
l Three lines-to-ground (3L-G).
~t- 2. Single line-to-ground (L-G).
3. Double linc-to-ground (2L-G).
4. Line-to-line (L-L).
Source wifh Symmetrical
~~CE~, ":~:::. 1: : 1:
3 Ph01Se L·L·l or 3 L-G
J CE z, z----'
IL----: Zo
2
Llne -+o- srounol

F
r------------, a 1
~~~--~ ~~4-'\
e
Double line·to- grou11d
F
l CE 1~.-. .- : 0
z,z,_____.z
:
Line-to-tit'le
1: tE
FIG. 111.-Types of faults on three-phase systems.
These types are illustrated schematically in Fig. 111. It will

be noted that fault resistance as indicated by the diagram will be
taken into account. For the sake of simplicity, the different
types of faults are assumed to occur on particular phases, as
indicated. This, of course, introduces no limitation on the
analysis since the system is assumed to be symmetrical up
to the point of fault. Fault currents, and the voltages between
conductors and from conductors to ground at the point of fault,
can be determined by the methods discussed previously. I t
~
~
J+:o.
TABLE XIIL-FAULT CURRENTS*
Type of fault Vector expression, effect of fault Magnitude when fault resistance equals
resistance included t R o= Rx = R2 =O
Eo Eo
3cp la = z +R
1
l a= h = l e = Xr ~
>.:::
~
L-L lb = -jVSEo
h = l e= X 1
v'3E o
+ X2 ~
le= ~I; Z2 + R ~
~
.......
Q
L-G la = 3Er; l _ 3E0 ;:t..
Zo + Zr + Z 2 + 3R a- Xo + Xr + X2 t-t
Q
-V3E - /'5 o
+ RL ) + j (2Zo + z2 + 3RL + 6Ro)l + X oX 2 + X2 2
2L-Gth = 2óv orv 3(Z2 h = l e -_ V3E"VXo2
Ó.M
~
o
-y'3E - /'5
le = 2~v "[V 3(Z2 + RL ) - j (2Zo + z2 + 3RL + 6Ro)] l 0 = 3Eox2
ÓM ~
~
1 0 = h + l e = 3lo ~
3Eo(Z 2 + RL ) ~
= -
Óv
Óv = (Zr + RL )(Z2 + RL ) + (Z¡ + z2 + 2RL )(Zo + RL + 3Ro) óM = XxX2 + Xo(Xr + X2 )

* See Fig. 111 for definiti on of impedances.
t If fault has reactance as well as resistance, replace the latter by the fault impedance.
t It may readily be shown that lb = le for all values of R n. when R L = R1 = R 2 = O.
will be assumed that all t he generated e.m.fs. may be reduced

to a single positive-sequence set of generated voltage* E 11 , and
that the equivalent circuits may be represented by the series
impedances Z1, Z2, and Z 0 for the positive-, negative-, and zero-
sequence networks, respectively.
93. Formulas for Line Currents and Voltages between Line
Conductors and to Ground.
The method of symmetrical components has been applied in
the development of formulas for the short-circuit currents
in t he different phases, and the voltages between the line con-
ductors, and between the line conductors and ground at the
point of fault. The more important of these formulas, which
include the effect of fault resistance, are summarized in Table
XIII for currents and Table X IV for voltages.
94. Basis for Current and Voltage Curves.
The general formulas for currents and voltages referred to
in the preceding section are complicated to such ::m extent t hat
it is difficult to visualize readily the range of voltages and
currents t hat may obtain. This is due to the fact that there are
several types of faults to be considercd, and that for each type
of fault there is a wide range of values which the impedance of
the system elements may have. F urther complication arises
from the fact that in machines t here are two values of X1 which
apply during transient conditions.
Range of the Sequence Impedances. In general, positive-
and negative-sequence impedances of a system are of the same
order of magnitude, whereas the zero-sequence impedance may
vary t hrough a very wide range from being very small to very
large in comparison with the positive-sequence impedanee.
Also, in general, t he positive-sequence resistance R 1 and the
negative-sequence resistance R 2 are small in comparison with
the positive- and negative-sequence reactances. Consequently,
the effect of t hese two resistances on t he magnitude of the fault
voltages a nd currents is relutively small. Are resistance may be
very low, particularly during t he interval following ftashover,
and, of course, may be zero, as in t he case of a metallic connection.
* The method of obtaining a single equivalent generated e.m.f. is dis-

cussed in Sec. 113, Chap. XII.
•
• TABLE XIV.-FAULT VOLTAGES* ~
~
Vector expression, effect of fault Magnitude when fault resistance equals o:>
Type of fault
resistance included t Ro= Rr = R2 =O
3ct> Ea = Eu- R - Ea= O
L-L
2Zz+R 2X2
Ea = E uzr + Z2 + R Ea= Euxl +X2
~R
2
+ f'/3R
2
+ Zz 1::1;¡
~
E Xz
Eb = -Eu Z¡ + Zz + R b = Ec = E ux 1 + Xz ~
tR - f\/3R + Z2 . ~
~
2 2 E 3X2 ~
Ec = -Eu Zr + Z2 + R e = EuX1 + Xz ......
Q
3R ~
L-G
Ea = E 0 Z o + Zr + Zz + 3R Ea= O t:"-1
Eb = -V3Eo[ V3(Zo + R) + j(Zo + 2z2 + 3R)J E =E = yi3E VXo

2+ XoXz + X22 Q
o
+ X1 + Xz
~
2 Zo + Z1 + Z z + 3R b e u Xo
Ec = -V3Eu[ y'3(Zo + R ) - j (Z o + 2Zz + 3R ) l y'33 X o + 2X2
EuX o+ X 1 + X
o
2 Zo +Z1 +Z2 +3R EA = 2
~
'&E0 3 ~
2L-Gt.Ea = l1v (Z2 + RL)(Zo + RL + 2R 0)
Ea=
f1M
EuXoX2
~
Eb = -~Eo[y'3(Zz + RL) (RL + 2Ru) + jRL(2Zo + z2 + 3RL + 6Ru)l Eb = Ec =O
Ec = -~E0 [y'3(Zz + RL )(RL + 2R

1)
jRL(2Zo + Zz + 3RL + 6Ru) J
0) - Ec =Ea
l1 v = (Zr + RL )(Z2 + RL) + (Zr + Zz + 2RL )(Z o + RL + 3R 0) l1M = X1X2 + Xo(X¡ + Xz)
* See Fig. 111 for d efinition of impedances.
t If fault has reactance as well as resistance, replace the latter by the fault impedance.
:t It may readily be shown that Eb = Ec for all values of Rg, when RL = Rr = R2 = O.
POWER SYSTEM VOLTAGES AND CURRENTS 227
In view of this, and the complication resulting from the con-
sideration of the effect of are resistance, it has been deemed
advisable to consider only the case of zero resistance in the fault.
In the case of zero-sequence, however, the resistance Ro may
be small or may be very large as a result of a neutral grounding
resistor. For this reason it is necessary to arrange the curves
to cover a wide range of zero-sequence resistance and reactance.
As a result of the study of the range of values for impedances,
the number of variables has been so reduced that sets of curves
may be plotted with values of zero-sequence impedance as
abscissas, and voltages or currents as ordinates. Static appa-
ratus, as previously pointed out, has identic:;t.l values of positive-
and negative-sequence impedances. This also represents a
kind of mean value for rotating machines, especially for the
reactances effective at the instant of short-circuit. This sug-
gests a set of curves plotted for equal values of positive- and
negative-sequence reactances, supplemented by additional sets
of curves to indica te the probable maximum and mínimum values
of the ratio of negative- to positive-sequence impedance, taking
into account the characteristics of all types of synchronous
machines.
Ratio of Sequence Impedances Used in Curves. The basis
for selecting the values of these ratios will now be considered.
The principal machine reactances that apply to transient condi-
tions are zero-sequence reactance X o, the negative-sequence
reactance X2, and the positive-sequence reactance X1; X1 may
be either the subtransient or transient reactan~e, depending,
respectively, on whether the initial high decrement component of
current is considered or neglected. The ratio of X2 to X1
for commercial machines will normally líe . within the range of
0.5 to 1.5, though it is quite possible with special machines to
somewhat exceed this range. The higher ratio of X2 to X1 is
obtained in machines without dampers, and for this class of
machines the average ratio may be taken as 1.5. Machines
with dampers or their equivalent, such as solid-rotor turbine
generators, always have x2 less than the transient reactance
(X1), and the ratio of X2 to X1 for machines of this class may be
as low as 0.5. It may be well to point out that all machine
reactances vary to sorne extent with saturation; also, that the
value of the negative-sequence reactance of a machine varies
somewhat with external reactance because of distortion resulting
'2.0
\
1.8
1.6 1\.
' ~
\
~ ~1\
~'l1 =0.5 -l--+-t-t-t-t-if----1
14
1\ ~ 2 L - G short-c1rcuit
'~
.¡.-
t:
~ 1.0
1.'2
i'
" "
r-...
t--.
1~ ~
1"'-
"'
J 1 l .1 1 1 1 1
1 1 1 1 1 1 1 '
1
x2 1
1.-
¡-..... ~
t:::,. '2 L-G short-circuit Xt =LO
:;¡
UQ.8 r-...
0.6
0.4 1--1-t-t-+-t--+--+--t l-G short-circu it X7./X :Q_s/i_o-'Ü ·
0.'2
'1\
0 o~~~~~~~~~2~~~~~3~~~~~4~~~~~5~~~~~6~CO~
R01+io Xq/X¡
'
F10. 112.- Curves of fault curre nts vs. systcm reac tances for single and double
line-to-ground faults. Each curve is labeled to indicate the type of fault and the
ratio of X 2/ X 1. All currents are expressed as a ratio to the threc-phase short-
circuit current. For these curves, all r esistances are assumed et¡ual to zm·o.
2.0
1.8
1
" 1!'
1.6 ){ j >Ct: \5
f-~
· cvit ;,..2..r- .11
9 rr-clr
1.4 1--
11
X
11
"l.vG ~n~
..;.?'1 1 Jrt~circu,t X'l Xt ·
. ~;: 1 :\o -1:::..:: - t:=:=l1
~--~ ..........
_'1_
(1¡
m
~ 1.0
o
1.2 - x"' ~
-Y- .f- -~
f.--
í. v-· V
' 1 V
~
.....
......
~ -
~ '1.V G_;;,b....--
~
,.....
¡;-.. . . . .
_.. i-'"'
- :::.C ]_ J. _j_
-:;:::-
'""""
L- G short-circuit ~,. 0.5 1.0' 1.5
1
L-G s hod-c1rcuit X2fX¡=0.5
lJ]~
-
>o.a JI/.- V
ll) ........ ~
0.6 ~1V V ·X2/X1 :Q.5
/
lHV
0.4 'f/1
J
0.'2 1
.1
o '1
o 3 4 5 6 QO 2
R01tio Xo/Xs
FIG. 113.-Curves of fault voltages vs. system reactances for single and double
line-to-ground faults . Each curve is labeled to indicate the type of fault and the
ratio of X2 / X1. The voltages are from line-to-ground and are expressed as a
ratio to the norma lline-to-neutral voltages. For these curves, all resistances are
assumed equal to zero.
from the inequality of direct and quadrature axis subtransient

rcactances. However, except perhaps for special machines or
cases where high accuracy is required for a wide range of external
circuit reactance, it is permissible to use a single value of any
one machine reactance. These considerations have led to the
selection of sets of curves for three ratios of X 2 to . X 1, namely,
0.5, 1.0, and 1.5.
Basis for Comparing Systems. The final consideration in
p 1o t ti n g t.hese curves for
unbalanced conditions was 14
- ---
to find a basis suitable for 1.2 Eca
comparing all systems. For .,10 ....... ~
r- ......., ...:~
~!-"' ....
~
,....... -t;;-
this purpose the following ~ 08 .......
...... -;-..
ref eren ce quantities ha ve ~+
~ lb or le
.............
~06
becn taken :jor line-to-ground t
vo~tages, the normal line-to- JoA
neutral voltage; jor line-to-line 0·2
voltages, the normal line-to- 0
o.5 0.6 0.1 0.8 09 1o 11 1.1 1.3 1.4 15,
Roo+io X1/ X1
line voltage; jor currents, the
Fw. 114.- Curves of fault voltages and
three-phase s h o r t- ci re uit currents vs. system reactances for line-to-line
current; and jor impedances, faults. Line-to-ground and line-to-line
voltages a re exprcssed as ratios to their
t h e positive-sequence react- respective normal values. Current is ex-
ance. Thus all values are prcssed as a ratio to the three-phase short-
circuit curren t. All resistances are assumed
plotted not in units such as equal to zero.
volts, amperes, and ohms,
but simply as ratios or decimal fractions of the reference
quantities. All voltages are computed for the point F (not G) of
the diagrams in Fig. 111.
95. Fault Current and Voltage Curves.

Curves prepared in accordance wit.h the preceding discussion
are shown in Figs. 112 to 116 inclu:,ive. Figures 112 and 113
show the ranges of line currents and line-to-ground voltages
respectively for single n.nd double line-to-ground faults, on
systems for which the three sequence reactances only require
consideration; i.e., fault resistances, Ro, R1, and R2 are taken
· equal to zero. Figure 114 shows the range of line-to-ground
voltages, line-to-line voltages, and line currents that obtain for
line-to-line short-circuits. Figures 115 and 116 show the range
of currents and voltages, respectively, that obtain for single
and double line-to-ground faults on systems for which the zero-
sequence resistance requires consideration; the curves are

plotted with the ratio 1: as abscissa, with families of curves to
2.0 2.4
1.8 \ h:J r-2L-~ Sh~rt!cir~ui~- 2.2
2L-G Short-tircuit
l
A~o
1.6
:C¡I t--
r-1 9 (or 31 0 ) 2.0
lb
3.0 >st X¡ :05.
-~2
~
~
1.4 .! ShOrtTrcu1
L-G . "t 2.8 ~ 1-- ~ 2 :05 1.8
+ 1.2 0.5 la 11 Xt . o'<
e ....... ~2-0.5 t 2.4 + 1.6
<1.1
L 1.0
f'..\ [\ x, ~ '2.0 ~~o
x ,t
e
~ 1.4
O.)_ ~
0.8 t-- ~
l. O 1/.
u
~
;:)
2iJ ¡......._ ~ ~
s t.6
o't
; 1.2 /
l.iO
_,,
~O()
·----
QJ¡ V 1.2 V 1.0
5.0 ¡---.::: ~ ;).._\
0.4
0.2
- f\ 0.8
0.4 ~~
0.8
0.6
'
~-
0.2
0 4
oo 1 2 3 4 5 6·-«> o 1 2 3 4 5 6-a::> ' 0 1 2 3 4 5 6 --ao
R01tio R 0 /x1 R01tio Ro/X 1 R01tio R0 /X1
3.2 2. o r--T""""~.---.---..--r->
2 L-G Short-circ.uit
2.8 l--
t-fg<lor 3 lo)- :-- 1.8
2.4- 1\ 1.6
X2: 1 o
+e 1_2.0 \ XI . -:: 1.4
Ql
1- ~ \.6 ~ ~~o
x, ~1.2
~
1... 1-
1-
.p.~~i---+---1 3
~~~~~~~
;:) 1. 2
u t-,. ~ 3 0.8
1.0 t=
0.8 ~- V
!_"':: ~
0.2 120 r....,
o.6 t-t--t--t--t--f---i~
1-'-0
1 2 3 4 5 6--oo
R01tio R0 /X1
3. 2 2.0 .----.--..--
¡r-1r-1r-1---,
2.8 1.s Er ;x
Xó
1
-1r----1r--r-t-1
2.4 O + 1.ó 0v 21 -G Short-circuif

+
e + 2
e ·
o e 1.4 J\
v ~J\
1
.J b t - t -
Ql
1...
L ~ 1.6 t 1.'2 /.O es:~ -~! =\.5t-
::;¡ ; 1.'2 .~ 1.0
u
nrl-+=F~::'k---1 v o.8 '-' o.8 ~.~~-
~~~,-~ . . . ~~d
-~;:~ :~
0.4 O.6 t-=co
=-t--t---t---t---t--t--t
o5 o4 '--'---J'----1-----l-----l---L--J
O 1 '2 3 4 5 6--oo · O 1 2 3 4 5 6--cc
R01tio R0 / X 1 R01tio R0 jX1
(a) Line curren+ (b) Groun~ current (e) Line curren+
Sin9le line-to-9round f01ult Ooub1e line-to-~round f01ult Double line-to-~rovnd faul+
Fw. 115.- Curves of fault currents vs. system impedances. The legend with
each group of cur ves indicatcs the type of fault, the current plotted, and the ratio
of X2 / X1. The indi vidual curves in each group are for various values of the
ratio of X 0 / X¡. Al! currents are expressed as a ratio to the three-phase short-
circuit current.
cover the range of the ratio of i:' and groups of curves to cover
the three ratios of :i": of 0.5, 1.0, and 1.5.

96. Discussion of Curves.
The ranges of line currents, ground currents, line-to-ground
voltages, and line-to-line voltages that exist on a system undet
unbalanced fault conditions as shown by the curves have been

summarized in Table XV. One of the most important facts
brought out by these curves is that the line-to-ground voltage
may rise to twice the normalline-to-neutral voltage, i.e., toa value
1.8 1.1 1.8

].0 1.6
0.9 1.4 __ Xo
~
r 1.2 Cl) 0.8 (!) 1.2
t ~l. O ~0.7 ~1.0
~
+ 0.8
t to.e ~0.6 0
~- >o.6 > 0.5 >o.6
0 .4 0.4 0.4
0.3 0.2
1 2 3 4 5
R01tío R0jX1
2.0 1.6 1.6
1.8 1.4 1 1 1.4 .{0 ..--+-
1.6 1.2
L.XoLo
-¡-x¡-T 1.2 li1tfo
el) 1.4 ~1.0
lS
-· ~1.0 ~ o.l
lS)
1 1
~1.2 ±0.8 ~0.8 z¡:;:-Xo
~ 1.0
o
>0.6 r-
l~hort-circu'1t
Ebc ~0.6 -lr..
2L·G Short-circuit
o.e 0.4 r-,..-Xz 0.4
-= 1.0 +:¡·E~
X 1
0.2 r-r-~· 1
0
0.2
o
r,¡ 1.01
o 1 2 3 4 5 6oo o 1 2 3 4 5 600
R01ti o Ro/Xt R01ho Ro/XI
2.4 1.5 J. S 00
l ~,.,.
2.2 1.4· - r-:.Xo 1.6 '5.0

¡' x, 1.4 " .o
-rr·
~S
2.0 1.3 ~
?-91
1.6 <111.2 ~ Cl) 1.2
+
~1. 1
0.8
-r
·. .....
ol.o 00 ·2.
~r.o
Xz
~~
>o.9 - (-GShorl'circuit-
Eb~
~l. O
..¡...
"00.8
>o.6
0.2
4r.
a·· y¡
'd. 2l·G '1Shori-circuit
IEct
x2 1
0.1 - ¡ - y¡¡ l. r 0.2
Xi'"l· 5
¡ -¡ -
O.G o 1 2 3 4 5 6 00 oo 1 2 3 4 5 6 00
R01tio RojX1 R01tio R0fXa
(01) Line-to- ~rounol volt01\'je (b) Line·to · line volt01~e (e} Line-to-~rounc.l voltot~e
Sin~le li11e·to·!'Jrounol fotult Sin!')le line · to· 9rounQI f01ult Double li11e -to-~rounol f01ult
Fw. 116.- Curves of fault voltages vs. system impedances. The l egend with
each group of curves indicates the type of fault, the voltage plotted, and the
r atio of X 2/ X 1. The indiv idual curves in each group are for the various values
of the ratio of Xo / Xt. L ine-to-ground and line-to-line voltages are expressed
as ratios to their respective normal values.
somewhat greater than the normal line-to-line voltage- a fact.

that is not generally appreciated. It is well to note that the
ratios of line current and ground current to the three-phase
short-circuit current may be as great as 2.0 and 3.0, respectively;
also, that the ground current of a double line-to-ground fault
TABLE XV. -RANGE OF FAULT CURRENTS AND VOLTAGES
Current ratio* Voltage ratio

Type of
fault
Line Ground L-Gt L-Lt
1 1
L-L o. 69-1.15 • o ••••••• O. 67-1.20 . 0.58-1.04

L-G 0 .00-2. 00 0 .00-2.00 o.58- 1.97 0.67- 1 .20
2L-G 0.69-2 .02 0.00-3.00 o.00- 1. 80§ 0.00- 1 .04§
Based on
O -< Ro
x1-
<
oo
and
2 0
X
Xt =
05 15 O -<X
. to . ; X1 -< 00
* Rcfcrence is the three-phase short-circuit curr cnt. ,

t Referencc is thc normal IinP-to-neutral voltagc.
:j: R efercncc is the normal linc-to-linc voltagc.
~ These voltages are identical; only the Iine-to-ground voltage curves ha ve bccn plott<'d.
may be 50 per cent greater than the maximum value of a single

line-to-ground fault.
Examination of the curves shows that thc maximum values
of fault currents are obtained gencrally with low values of X 0
and X2, and, correspondingly, t.hat the maximum values of
voltages occur with high values of Xo and X2. Howevcr, this
is not the case for zcro-sequence rcsistance Ro, as the maximum
values of currents and voltages are obtained with intermediatc
values of R o. The effect of adding zero-sequencc rcsistance to
a system having no resistance is, first, to incrense the line current
on a double line-to-ground fault; subscquent increase of Ro
reduces line currents as well as ground currents and incrcascs
line-to-ground voltages.
97. Systems Giving Maximum Voltage or Curren.t Ratios.
The powcr-system layouts which give the maximum values
of voltage or current ratios are accurately defined by the sequence
impedance ratios used in plotting the curves. However, it
may be well to describe the limiting conditions in other terms.
The simplest illustration is that of several g0nerators connected
toa bus; a low ratio of Xo to X1 is obtained for one machine or
for several machines in parallel, provided that in every case the
machine neutral is solidly grounded and provided that all
generators have two-thirds pitch windings. For such a case

the zero-sequence reactance may be as low as 1 per cent, and
the ratio of X o to X 1 may readily be less than 0.2. A high ratio
of X oto X 1 may, of course, be obtained by introducing a neutral
reactor. Anothcr method is to ground thc neutral of only one
of sevcral machines operating in parallel; for such a case the
machines should have a full pitch winding and therefore a
relatively high value of X 0 • Considering a fault on the bus,
the effective valu e of the positive-sequcnce reactance is X 1,
n
whcre X 1 is the reactance per machine and n the number of
machines; similarly, the effcctivc value of the negative-sequence
reactance is X 2 ; and, finally, the effective value of the zero-
n
· of course, T,
sequence reac t ance 1s, Xo smce
· only one mac h'me
• has its neutral grounded. Thus, it is clcar that a relatively
high ratio of X 0 to X 1 may be obtained if there are several
machines connectcd toa bus and only one mn,chine has its neutral
grounded. Incidcntally, under thesc conditions the single line-
to-ground fault current may become several times the three-
phase short-circuit current of a single machine, which may
produce unusually high strcsses in the mn,chine whose neutral is
grounded.
Another illustration of a case wherc the ratio i: is low, is
that of a fault on the grounded-star side of delta-star step-up
transformers supplied by waterwheel type generators. The zero-
sequence reactance is merely that of the transformers, while the
positive-sequence reactance is the transformer reactance plus
the reactance of the watcrwheel generators, which is normally
sevcral times that of the transformers. For faults on the trans-
mission line remote from the grounded star-delta transformers,
the zero-scquence reactance may become relatively Iarge and
the ratio of X o to X 1 may be considerably greater than unity.
A high ratio of X 2 to X 1 for synchronous machines alone occurs
for machines without dampers, while a low ratio occurs for
machines with dampers (considering transient reactance).
For system faults this ratio has correspondingly high or low
values, depending upon the type of machine, when the fault
occurs at machine terminals; this ratio approaches unity for
faults remo te from machine buses, i.e., with externa! reactance

in series with the machines.
98. Effect of Method of Grounding. Zero-sequence Impedance
Ratio.
Examination of the curves of Figs. 115 and 116 shows that the
ranges of voltages and currents under unbalanced fault condi-
tions are controlled principally by the ratio of Z 0 to X 1 ; i.e.,
by the method of system grounding and the valucs of the ground-
ing impedances.
The common classification of systcms with respect to grounding
is:
l. Direct-grounded.
2. Grounded through impedance:
a. Resistor.
b. Reactor.
3. Free neutral.*
Systems are usually said to be direct-grounded if the neu trals
of one or more generators or transformers are connected directly
to ground. Systems grounded through impedanccs may approach
direct-grounded systems on one hand and free neutral systems
on the other, depending upon the magnitude of the grounding
impedance. In fact, the zero-sequence impedance of a direct-
grounded system may actually be higher than that of a system
of similar capacity and extent, which is grounded through a
neutral impedance of low value. Obviously, a classification
of systems into grounded or ungrounded without further restric-
tions is unsatisfactory because it takes into account only the
impedance betwcen the ground and the neutral points of appa-
ratus, instead of the total effective zero-sequence impedance
which includes line and apparatus impedance as well as that of
grounding resistors and reactors. A more accurate classification
is suggested by the curves of Figs. 115 and 116. These curves
cover thtl range of system grounding from a direct-grounded
system to a free neutral system with negligible capacitance,
and they suggest that the ratio of Z o to X 1 be used as a criterion,
* For the purpose of this classification, system grounding through neutral
reactance which rcsonatcs with Jine capacitance, such as the Petersen coil,
Bauch transformar, or Jonas dissonance coi!, may be viewed as free neutral
systems, since thc object of these rcactance devices is to nullify line capaci-
tance. In this discussion no consideration is gi.ven to the phenomena of
arcing grounds.
IITI.~illllilllllllll
0.4 f-1-.Xt
~ 0.3
~ ~20
~0.6 05_ l-G Short-circuit ~ 2 L-G Short-cwcu1t
mm
,'1.(7
~ o.z ~~-o.~.5_ilil~jE~
:: = r ~¡¡~
~ 0.5 o
J. Et
0.4
0.3
~1. = 0.5 0.5
0.1
0 .2
o~~~~~~~~~~~~~~
2 3 4 5 6 --ot o 2 · :; 4 5 6 --«
R01tio Ro/Xt Ratio Ro / Xt
1.0
1.0
0,8
cu
2.0
~ 0,1 /. ~o.3
+-
a,.
o. 0
~0.6 L-G Short-circuit 2 L-G Shor+-circuit
> 0 .2 o
·~ D.5'
:> 0.4 Et Et
x2 x,_ -- 1'o
0,3 XI" 1.0
0.1
x,
o:z
0.1
00
o
6 --oc O 2 3 4 5 6 - -oc
R01tio Ro/Xt
0 .6 ~~
5.0
0.5 2.0
0,4 lO.
cu
<S'\ 0.5
cu 0.1
~o.:;
~0. 6 L-G Short-circuit ~ 2 l-G Short-circuit
~ Et 0.2 E1
0.4
0.3
X -- 1, 5
X2
XX~f = 15
.
0.2
0.1
00
o
'Z. 3 4 5 6 --oc O 2 3 4 5 6 --oe.
R01tio Roj X1 R01tio RojXt
Fw. 117.- Curvcs of positive- Fw. 118.-Curves of positive-
sequence voltage at thc faultvs. system sequcnce voltages at the faul t vs.
ÍQ1pedances for single line-to-ground systcm impedances for double line-to-
fault. ground fault.
The l egend with each group of Figs. 117 and 118 indica tes the ratio of X 2/X 1.
The individual curves in each group are for various values of the ratio of X 0 / X1.
Vol tages are expressed as ratios to the normal line-to-neutrai vol taíte-
or a measure of the degree of grounding, low values of this ratio

corresponding to a solidly grounded system and high values
toa free neutral system. Hence, from the standpoint of system
performance, the criterion is not whether there is an impedance
between one or more neutral points and ground, but whether
the ratio of Xo or R o to X1 is high or low. The ratio of Zo to
X 1 is the zero-sequence impedance ratio.
99. Positive-sequence Voltage at the Fault.
Different methods of grounding a system and different values
of the zero-sequence impedance ratio have rather complicated
effects on the positive-sequence voltage uscd in stability analysis.
Since any type of unbalanced fault may be replaced by a three-
phase short-circuit through a symmetrical three-phase imped-
ance, it has been suggested that the positive-scquence component
of voltage E 1 at the fault is a measure of the stability of a system
under fault conditions.czsl.C 35) Accordingly, Figs. 117 and 118
have been plotted to show the range of E1 for single and double
line-to-ground faults, respectively. These curves have been
plotted with the ratios of E1 to normal voltage as ordinates,
and with ratios of Ro to X 1 as abscissas; fami.lies of curves cover
the range of the ratios of Xo to X1, wh»le groups of these families
cover the three values of the ratio of x2 to xl of 0.5, 1.0, and 1.5.
It should be realized that the positive-sequence voltage at
the fault is, for given generator voltages and system impedances,
a measure of the synchronizing power that may be exchanged
between machines past thc point of fault. The drop in positive-
sequence voltage may cause an underexcited and overloaded
synchronous machine to pull out of step, without causing the
system as a whole to lose synchronism. Consequently, the
reduction of the positive-sequence voltage at the fault is also a
measure of the effect of grounding on the tendency of these
machines or of the system as a whole to pull out of step.
Problems
1. A power systern has the following impcclanccs as mcasured from the
110-kv. line:
zl = + j20 per cent}
z2 = + j20 per cent based on 200,000 kva.
Z0 = 40 + jlO per ccnt
Find the largest fault current ancl the highest line-to-ground voltage for (a)
single line-to-grouncl fault; (b) double line-to-grouncl fault; (e) line-to-line
fault; (d) three-phase fault. Make use of the curves of Figs. 114, 115, and
116.
2. Plot a curve for the neutral voltage of power systems for a line-to-
ground fault assuming that thc zcro-scqucnce rcsistancc is all in a grounding
resistor and all thc zcro-scqucnce rcactan ce is in a dclta-star transformer.
Use data from curves of Chap. XI assuming X2 = X1 and Xo = 0.5X¡.
3. Circuit-brcakcr duty is sometimos mcasurcd in kilovolt-amperes
obtaincd by multiplying thc r.m.s. valuc of the current through a breaker
pole bcforc opening and thc r.m.s. valuc of the voltage that appears across
the brcakcr aftor opening. Assurne that tho systcm of Prob . 1 is subjected
to a zoro-resistancc throe-phasc fault to ground. Assuming that thc polos
open individually in the ordor of loast duty oxprossod in kilovolt-ampercs
and that the high froquoncy and wavo distortion may be ignorcd , find the
duty on oach polo.
4. Assumo tho same conditions as in Prob. 3 hut without rosistance, and
determine the circuit-brcakcr duty opcning onc polo at a time as before.
The above systom impcdancc ratios are thosc which obtain at the sending
end of a tmnsmission lino supplicd b y a grounclcd star-dolta transformer.
Explain why ono polo of a brcakor on such a systom may opon appreciably
ahead of tho two othors.
6. A gcncmtor and a stcp-up transformcr connectccl delta on thc low-
voltago sido and groundocl-star on thc transmission-linc sido has impedances
of Zt = +j40 por cont, Z 2 = +j35 per ccnt, Z o = +jlO por cent. The
transmission lino has tho following impodancos for a 10-mile section: zl =
+i5 por cont, Z2 = +i5 por cont, Z o = +jl5 por cont. Assume that it is
desired to limit tho line-to-ground voltagc to 120 pcr cont of normal for a
line-to-ground fault. What is the lcngth of linc that can be protectcd by
grounding the systom at the stop-up transformor station? In this calcula-
tion neglect the receiver-end characteristics.
CHAPTER XII
SIMULTANEOUS FAULTS
Experience has shown that a certain numbm of the faults
on power systems occur concurrently at different points in the
circuits. These "simultaneous" faults may be encountered as
the result of a ground at one location raising the voltage on
a sound phase so that flashover occurs at a second point. They
may also occur on separate circuits which are physically close
together, as, for example, in the case of two circuits on the sáme
structure for which the fl.ashovers are produced by the same
lightning disturbance.
Simultaneous faults are important even though relatively
infrequent because relay systems giving satisfactory operation
íor a fault at a single location may fail to isolate simultaneous
faults. A scheme for obtaining correct relay operation under
these conditions has been proposed by R. M. Smith. * He showed
the applicability of this scheme to a particular case by calcula-
tions of fault currents and voltages using the methods developed
by the authors and presented in this chapter in an amplified form.
G. OberdorferP 4 published the first analysis of this problem in
1930 and this was followed by a more comprehensive treatment
of the same subject by Edith Clarke70 in 1931.
'
This cltGLpter will present in sorne detail a method for calculat-
ing simultaneous faults at two locations, after which the method
of attack for more than two simultaneous faults will be outlined.
The discussion reviews the special case of a fault at a ~ingle loca-
tion with several sources of e.m.f. differing in magnitude or phase
position.
100. Simplest Equivalent N etwork.
Befare taking up the solution of simultaneous faults, it is
desirable to discuss briefl.y certain methods of representing
and simplifying networks. The simplest general form of network
* Suggestions for Avoiding Faulty Action of Ground Relay, Elec. World,
vol. 95, p. 1092, Ma.y 31, 1930.
238
SIMULTANEOUS F'AULTS 239
for a particular problem is determined by the number of terminals
whose identity must be retained. For t he calculation of simul-
.( taneous faults these terminals are of t hree types, namely, (1)
the fault locations, (2) the e.m.f. sources whose magnitude and
phase relation are not identical, and (3) the neutral of the
system, required for cases involving shunt load. Mínimum
number of branches in the simplest general network is obtained
when branches connect all terminals without additional star
points or parallel branches. Thus 10 branches are, in general,
required to represent a network involving two separate sources
Neutral bus
,--- ----------------
,...-"
/
1
1
/
lx = Ea. + E. +
1 Zax Zbz
1
' Zxy
E 11 _E (-1- +
x Z az
-z1bz +-z
1
+ Yxn )Zl/
fu =_Ea + E, +
Zav Z by
Ex _ E(_!_
11
+
Zx y Znu
1 1 )
Z- +z-+ Y)/11
tlx
' b11 xy
(a) (b)
Fw. 119.- Simplest equivalent network for a system with four line terminals
and one neutral terminal. (a) Diagram of connections; (b) equations for termi-
nal currents expressed in terms of ter minal voltages and network constants.
of e.m.f., two fault locations, and shunt loads from line-to-neutral.
Such a network is illustrated in Fig. 119.
101. Determination of the Branches of the Forro Zax·
The series impedance branches of the equivalent network
for any specified number of termina ls will be designated by the
impedance symbol Z with two subscripts to denote the ends
of t he branch; thus Zax represents the impedance of the branch
between the points a and x. The shunt loads may be repre-
sented by t he admittance symbol Y with similar subscripts,
such as Yan•
Determination of Branch Impedances by the Calculating
Board. The various branches of the equivalent network may
conveniently be determined by means of the alternating-current
calculating board. Consider the network of Fig. 120(a). If
the identity of terminals a, b, x, and y is to be retained, t he
240 SYMME1'RICAL COMPONENTS
simplest equivalent network is that shown in (e). The branches

may be obtained by applying voltage to one terminal and short-
circuiting the remaining terminals, then measuring the current
flowing at each terminal as illustrated in (b). The vector
ratio of the source voltage to the current in any terminal gives
.?)1.. ~
(01) Oric;,in01l11etwork
1.
1
1
1
1
1
1
(e:) Simplifieol network (ol) Br0111ches founol from (b)
Fw. 120.-Method of simplifying network p articularly suited to the alternating-

curren t calculating board.
the impedance of the branch between that terminal and the

source, thus
(249)
From the measurements indicated in (b), the network branches

shown by full lines in (d) are obtained. This procedure must
be repeated using other terminals as the source until all of the
branches of the equivalent network are obtained. It will be
noted that checks will be obtained on the branch impedances.
The network of Fig. 120(a) showed only series branches
between the terminals a, b, x, and y . If shunt loads were taken
from this network, a maximum of four additional branches
in the equivalent .:Jircuit from the terminals to neutral would
be required. These shunt branches may be obtained by measur-
ing the current in the source as well as in the remaining terminals.
SIMULTANEOUS FAULTS 241
The value of the branch admittance is determined by the
expression
(250)
The admittance of the remaining branches may be obtained by

using other terminals as the sources.
Determination of Branch Impedances by Analytical Methods.
The simplest equivalent network may also be obtained analyti-
cally in a similar manner or by means of network reduction
met hods. The latter procedurc will be illustrated subsequently
in connection with the determination of the positive-sequence
impedance constants.
102. Circuit Equations for Simplest Equivalent N etwork.
Probably the principal advantage of reducing the network
to its simplest equivalent forro is the ease with which the basic
equations relating voltages and currents at the terminals may
be obtained since they may be written from inspection. For
present purposes it is convenient to assume the positive direction
of current flow as outward at every terminal and to express
these currents in terms of the voltages of the several terminals
and the network constants. Thus for the network of Fig. 119,
the currents drawn from the points x and y may be written as
follows:
¡'
These equations may be rewritten as follows:
1 x= -zEa
ax + Eb
-z
bx + E 11
-
z xy -
E
X
(y
xn+ -z1ax+-z 1
bx + -z
1)
xy
111 = Ea + Eb + Ex _ E 11
Z au Zb 11 Z xu
(y
+ ___!_ + ___!_ + ___!__)
yn Zau Zhu Zxv
(252)
Because of the symmetry of these equations it is readily apparent

that similar expressions may be obtained for additional terminals.
103. Impedance-drop Constants of the Forro Dxx and Dxy•
For certain calculations it is desirable to carry the network

analysis a step farther and to define the set of constants known
as impedance-drop constants* of the form Dxx and Dx11 • Thus,

for a source at a and for current drawn from the network at
the points x and y, the voltages at these points may be written
as
Ex= E a - Dxxl x - Dx¡¡ly
E 11 = Ea- Dxyl x- DIJIJ[IJ }(253)
The determination of the drop constants may be made by

means of the alternating-current calculating board as illustrated
E Nefwork E E Nefwork
y y
___..___--J ~
D::c::c = E : c
I ::c
D::cv = Ev Dxu = E x
l::r; !y
(a) (b)
FIG. 121.-Determination of impedan ce-drop constants by m eans of t he alter-
nating-current calculating board.
in Fig. 121. Impress voltage at x and measure the current

1 x and the voltages E x and E 11 • The drop constants are
(254)
and
D~11 = ~: (255)
By applying voltage at y and measuring thc voltages E 11 and

Ex and the current 1 11 , t he impedance-drop constants obtained are
(256)
D x11 = E x (257)
1 11
The drop constant D u is merely the impedance of the network

between the point x and neut ral, while t he drop constant D x11
* Note that the symbols Z ::c::c and Z ::c11 are not infrequently used instead of
Dxx and D::cv· The notation hcre used has been chosen to avoid confusion
váth the symbol for the impcdance of the branch between the t erminals
denoted by subscripts ás discussed previously.
ia the mutual impedance drop from the neutral to the point y
due to the flow of unit current into the network at the neutral
and out at the point x.
104. Discussion of N etwork Restraints.
For the solution of any short-circuit problem it is convenient
to make use of the restraints imposed by the impedance constants
of the different sequence networks and also by the type of fault.
For the case of two simultaneous faults it becomes necessary to
determine the value of 12 unknown quantities, namely, the
three sequence currents and the three sequence voltages at the
two fault locations. For this purpose 12 independent equations
must be set up. Since the system is assumed to be balanced,
the three sequence rretworks will be independent of each other
and therefore it will be possible to obtain:
l. From a consideration of the positive-sequence network
impedances, two equations relating the positive-sequence
voltages and currents at each fault location.
2. From a consideration of the negative-sequence network
impedances, two equations relating the negative-sequence
voltages and currents at each fault location.
3. From a considerR.tion of the zero-sequence network imped-
ances, two equations relating the zero-sequence voltages and
currents at each fault location.
In this manner six of the requisite equations may be obtained.
At each point of fault, three imaginary leads may be brought
out and the particular conditions for the fault imposed upon
them. Regardless of the characteristics of the fault, the
restraints imposed will always give three equations relating
sequence voltages and currents at that point, and since two
.
1
1
faults are involved six additional equations will thus be obtained .
~ It follows then that this procedure produces a sufficient number
of equations to permit solving of the unknown quantities.
l
i 105. Two Simultaneous Faults with Two Sources of E.M.F.
lt The solution of the case of two simultaneous faults with two
sources of e.m.f. will now be undertaken. The solution will be
illustrated by application to a particular case. For this purpose
consider t he network of Fig. 122(a) which is identical with that
of Fig. 23, Chap. IV, with the point x corresponding to the point
1 F. The two faults are assumed to occur at x at an intermediate
point on one line, and at the point y on the right-hand bus.

Let it be assumed further that the generated voltages of the
machines A and B are identical, each being equal to Ea1, and
that the generated voltages of the machines C and D are identical,
each being equal to Ect. The shunt loads will be neglected for
this particular illustration so that it is necessary to consider
only four terminal points, the voltages of which are Ea1 1 Ec1, Ex,
and E 11 • The sequence components of the currents and the
(j) General equations

lx 1 = Ea1 + Ec1 +. Ev1 _
Zaxl Zrx1 Zxv1
E 1 (_1 +-1 +-1)
:z; Z axl Zcxl Zxl/1
¡ = Ea1 + Er1 + E,1 _
111
Za¡¡l Z cyl Zx111
,.-------------.-2-------------,
1
j425 j 40 :
E,(-1
11
Z ay1
+-1
Z cy1
+-1)
Z xl! !
ÓE.:.a m j 6 jl6 Eca Q
(g) Equations for the spccific
I"'Xt (e) ly, ~ network
j 220 Ex, . y E y,
l x1 = - jO.OJ66Eal -
~;
Resulfof str:.r-dellu
/rr::lnsli:Jrmafton in
in ~lfmmalmg poini m j0.0943E11 1 + jO.lllE.,1
,...--------------------------, [ 11 ¡ = -j0.0045Ea1 -
1
b~
J220 "4{)
~b
:
j0.0250Ec1 - j0.0943Exl +
j0.124E11 1
(e) Equivalcnt network

FIG. 122.-Reduction of positive-sequence network to two equations relating
the posi tive-sequence currents and voltages at the two fault points.
voltages at the two points at which faults are assumed to óccur

will be designated by the addition of the subscript numbers
1, 2, or Oto the subscripts x and y to indicate t he point to which
the quantities refer. Thus the positive-sequence component
of fault current at the point x will be designated as I xl· Similarly
the sequence impedance of the various· branch impedances is
indicated by the addition of the numerals 1, 2, orO, for example,
Zax1·
106. Restraints Imposed by the Sequence N etwork Constants.
It was pointed out in the previous section t hat six of the 12
equations required for the solution of the simultaneous faults
are determined from the restraints imposed by the impedance
of the sequence networks. The determination of these sequence-
impedance constants will now be described.
Positive-sequence Network Constants. The positive-sequence
impedance diagram of the assumed system is shown in Fig.
122(b). The network may be simplificd as illustrated in (e),
(d), and (e), eliminating the point m by means of a star-delta
transformation. * It will be noted t hat Fig. 122(e) is in the form
of the simplest equivalent network of the type illustrated in
Fig. 120, but in this particular case t here are no shunt branches
and the impedance branches Zac and Z cx are infinite.
j22
j/6
y
(j) Drop constants
_ (j76)(j10.5 + j27.1)
(a)
D
xzZ - j(76 + 10.5 + j27.1)
= j25.2
C e"><>' O
j76(j27.1)
AondB ~
.)
(b)
y
'279 X )' D,. 112 = j (76 + 10.5 + j27.1)
. ~~%,~~¡:;,~~/la = jl8.1
(e)
J16 X 30.4
¡------------------------..,
e hmmolmgpomi m
D _ (j27.1 )(j10.5 + j76)
j 279 jJO 1
11112
- j(76 + 10.5 + j27.1)
= j20.6
(g) Equations relating terminal
voltages and currents
E :cz = -j25.2I :cz - jl8.ll11 2
E 11 z = - jl8.ll:cz - j20.6lyz
<a) Bquival ent Carcuit
.f Fw. 123. -Reduction of negative-sequence n etwork to two equations relating the

negative-scquence currents and voltagcs at the two fault points.
The expression for voltages and currents at the various ter-

minals may now be written by comparison with equations (252)
or by inspection with the results given in ítem (f) of Fig. 122.
The drop constants of the positive-sequence network are
not used in the calculation of two simultaneous faults with
two sources of e.m.f. .: these constants are, however, used for the
calculation with a single source as described in Sec. 113.
N egative-sequence N etwork Constants. The negative-..se-
quence network is handled in a manner similar to that of the
positive-sequence except that the impedance-drop constants
are used for all cases regardless of t he number of e.m.f. sources
involved. The negative-sequence network has the same number
* This formula is given in the Appendix.
of branches as the positive-sequence network but in general

includes no sources of generated e.m.f. The negative-sequence
network of the system shown in Fig. 122(a) is represented by
the impedance diagram of Fig. 123(a). By means of successive
steps of network simplification, the network is reduced to
the form shown in Fig. 123(e) . The impedance-drop con-
stants Du2, D 1111 2, and Dz11 2 may readily be computed from (e)
J77
(e) Drop constants
D _ (j46.5)(j42.7 + j35.6)
zxO - j(46.5 + 42.7 + 35.6)
A~j/
70 =j29.2
Y ResuHofstar-dtJ/ta
IITinslórmalion - (j46. 5) (j3 5. 6)
r-:~~ -~--~: _~~-!:~':n_:":::::_r~--. Dxvo
5 9
= j(46.5 + 42.7+ 35.6)
j/70 "45 1 = j13.3
o D _ (j35.6)(j42.7 + j46.5)
'J42.7 yyO - + +
j(46.5 42.7 j35.6)
>< (e) = j25.4
And 611tc~ lite polenfials oF A and O are z::ero (j) Equations relating terminal vol-
tages and currents
Exo = -j29.2J.,o - jl3.3111o
Evo = -j25.4Ivo - j13.3/ xo
FIG. 124.-Reduction of zero-sequence network to two equations relating the

zero-sequence currents and voltages at the two fault points.
with the results shown in (f). The equations for the terminal
voltages E x2 and E 11 2 in terms of the currents I z2 and I 11 2 and the
network constants are
Ez2 = -Du2J z 2 - Dzy2J¡¡2
} (258)
E112 = - D zyd z2 - D""d112
whose numerical values are given in Fig. 123(g). In general, the
real and imaginary parts of Du2, Dx11 2, and D 1111 2 will be positive.
Zero-sequence N etwork Constants. The zero-sequence net-
work is handled in a manner similar to that of the negative-
sequence network. The steps are shown in Fig. 124, resulting
finally in the simplest equivalent network shown in (d). The
zero-sequence voltages Ezo and E 11o may be expressed in terms of
the sequence currents Izo and I 11o and the network constants by
means of the equations
E xO = - D xzol zO - D xyol yO }
E 11o = -D 1111ol11o - Dx_11olxo (
259 )
The impedance-drop constants Dxxo, D1111o, and Dx 11o may be
obtained readily from the network of Fig. 124(d). The numerical
Voltages and currents are expressed in terms of star components.
Line-to-ground faults
lx2 = lxt l xo = l xt
Ext = -Ex2 - Exo + 3Z¡:xlxt Phase a
l x2 = alxt I xo = a 2lxt
Ext = -a 2Ex2 - aExo + 3Z¡xl:z:t Phase b
lx2 = a 2lxt lxo = al:xt

Ext = -aEx2 - a 2Exo + 3Z¡xlxt Phase e
Line-to-line faults
Phases b and e
l xz = -alx1 l :xo = O
Ext = a 2Ex2 + Z¡xlxt Phases a ande
l xz = -a2lx1 lxo = O
Ext = aEx2 + Z¡xlxt Phases a and b
Double line-to-ground faults
:~ Phases b and e
Ol~b l xo + alxz = -a2lxt 2

e '---¡zfx E xl = a 2Ex2 Ex2 = a (Exo - 3Zfz)l:z:O Phases a and e
Ol~b
lxo + a2Jxz = -al xt
e 1zfx' Ext = aEx2 Ex2 = a(E:xo - 3Z¡xlxo) Phases a and b
Three-phase faults
01
e
{jj b J xO = O
E xt = O Ex2 = O Not to ground
:~-=-
"'~ b Exl = O Exz = O Exo = O To ground
Fw. 125.-Equations relating sequence quantities at point x for different kinds

of faults.
248 SYMMET!lléAL COMPONENTS
value of thesé constants substituted in equations gives the

equations of item (f).
107. Restraints Imposed by the Type of Fault.

The restraints imposed upon the sequence quantities at the
point of fault vary with the type of fault and the particular
phase or phases involved. In Fig. 125 a table is given showing
the relations between the sequen e e
a----~-----------
quantities for the most important types
b -----+------.p------ of faults. These relations also include
e----+---,_-~-- the effect of fault impedances. The
1 derivation of these relations is illustrated
2 tx
in Chap. III, Sec. 16 to 19 inclusive.
It will be observed that three equations
are obtained for each type of fault. ·
With the exception of the double
F10. 126.- Device f o r line-to-ground fault, the fault impedance
including equal impedances
in the conductor paths could just as well have been included
:in addition to the common as a part of the positive-, negative-,
impedance Z¡x.
and zero-seq uence networks. For
example, the single line-to-ground fault could have been applied
through the impedance Z 1 x inserted in all three phases, but,
since only one phase carries current, the presence of the two
other impedances is immaterial. Similarly, for a line-to-line
fault, the fault could have been applied through the impedance
Z;x in series with each phase. This device can be used to advan-
tage to include the effect of series resistance in two of the faulted
phases for a double line-to-ground fault. The diagram for such
a case is illustrated in Fig. 126 which shows an impedance
Z' fx in each leg of the imaginary leads coming from the line.
This device is limited, of course, to equal impedances in the two
legs.
Similar equations can be obtained for faults applied at the
point y.
108. Solution by Combination of Restraints Due to N etwork
Constants and Types of Faults.
Having obtained the two equations relating the negative-
sequence currents and voltages from equations (258) and two
relating the zero-sequence currents and voltages from equations
(259) and the six relations from the terminal restraints, it is
possible to eliminate the eight negative- and zero-sequence
unknown quantities, obtaining therefrom two equations relating
the positive-sequence currents and voltages at the fault points.
This has been done with the results as tabulated in Fig. 127
for the most important combinations of faults at the two points.
In order to illustrate how these relations are obtained, the
principal steps for a typical case will be given. Assume a
single line-to-ground fault at x on phase a and a single line-to-
ground fault at y on phase b. In the table of Fig. 125 it is shown
that
J x2 = J xl (260)
J xO = J xl (261)
Exl = -Ex2 - Exo + 3Z¡xf xl (262)
and also that
lv2 = al vt (263)
lvo = a 2lvt (264)
Evt = - a 2E 112 - aE 11o + 3Z¡yf 111 (265)
Substituting equations (258) and (259) in (262) and (265)'
there results
and
E 11 t = a 2Dv112l 11 2 + a 2Dx 2lz2 + aD u0l o + aDxyolxo + 3Zrvlvt
11 11 11
In -turn, substituting the relations from (260), (261), (263),

and (264) and collecting terms
Ext = (Dxx2 + Dxxo + 3Z¡x)I xt + (aDxv2 + a 2Dxvo)l11 t (266)
E 11 t = (a 2Dxv2 + aDxvo)l xt + (D 1111 2 + D 1111o + 3Z¡11 )1 11 t (267)
These equations check with those tabulated for the second
case in Fig. 127.
It will be observed that having given the type of faults and
.the phase involved and also the self and mutual drop constarús from
the negative- and zero-sequence networks, it is merely necessary
to substitute the values of Ex 1 and E 11 t, obtained from Fig. 127,
into the equations for 1 xt and 1111 , obtained jrom the positive-sequence
rrdwork. The solution of th~ two resulting equations gives the
values of 1 xt and ] 111 , from which the other unknown quantities
at tbe point of fault ar~ d~rived ..
109. Simultaneous Faults at More than Two Locations.

The method just described can be extended to the calculation
of simultaneous faults at any number of points. For three
simultaneous faults, 18 quantities comprising the six sequence
currents and voltages at each of the three locations are unknown.
From the positive-sequence network it is possible to obtain three
At po•nh. A+po•n+y Single line-to-ground faults at both points

~:::::t== b~ Ezt = (D u2 + D uo+ 3Z¡z) lzt + (Dzu2 + Dzuo)lut
e .j.Zf~t e ~Zfy Eut = (Dzu2 + Dzuo)lzt + (Duu2 + Duuo + 3Z¡u)l¡¡t
E.ct = (Duz + D uo+ 3Z¡:e)lz! + (aDzu2 + a 2D zuo) lu!

~~ ~~¡Z;=fy E 11 t = (a2Dz112 +
aD.cuo)lzt +
(D11112 Duuo +
3Z¡u)lut +
E x!= (Dn2 + D:uo + 3Z¡z)lz! + (a 2Dxu2 + aDzuo)lut
Ept = (aDzu2 + a 2Dzuo)lzi + (Duu2 + Duuo + 3Z¡II)ly¡
Double line-to-ground faults at both points.
E"! = - (Duzl:r;2 + D" 2l11 2)
11 E 11 t = -a 2 (Dzuzlz2 + D uu2luz)
in which
1
l :r;2 = KN _ LM{[LDzuo- N(Dzzo + 3Z¡x))l :r;t +
[L(Duo + 3Z¡u) - NDzuollud
1
l u2 = KN _ LM((M(Duo + 3Z¡z) - KDx o)l:r;t + 11
[MDzuo - K(Duuo 3Z¡u))lud +

K = Du2 + D uo + 3Z J:r; L = Dxu2 +
D zuo M = L
N = D 11u2 + Duuo + 3Z¡u
E zt = - (Dzz2l:r;2 + Dzu2lu2) Eut = -a(Duuzlu2 + D.cu2ld)
in which
1
I "'2 = KN _ LM([aLDzuo- N(Drxo + 3Z¡x)]lxt +
(a2L(Dwo +
3Z¡11 ) - aNDzuoliutl
1
lu2 = KN _ LM{[M(Duo + 3Z¡x) - aKDxuo)lzt +
[aMDxuo- a 2K(D 11uo 3Z¡u)Iud +
K = Du2 + Duo + '3Z¡z L = D zu2 +
a 2Dzuo M = D zu2 aDzuO +
N = D 1111 2 +
Du110 +
3Z¡u
Ezt ~ -(Du2lz2 + D xuzlu2) Eut = -a 2 (Duu2lu2 + Dzuzlz2)
in which
1
Iz2 =
KN - L
M{[a 2LD.cuo- N(Duo 3Z¡x))lzt+ +
f==s= ~~ [aL(D 11 vo +
3Z¡u) - a 2NDxuollud
+'Z.tx e: --¡rry I 112 = KN ~
LM([M(Duo + 2
3Z¡x) - a KDxuo]l .ct +
[a 2MDx 11o- oK(Dvuo 3Z¡u)lud +
K = Du2 +Duo + 3Z¡x L = Dxu2 + aDxuO M = Dzu2 + a 2 Dxuo
N = Duu2 +
Duuo + 3Z¡u
Fw. 127.-For descriptive legend see opposite page.
equations of the type developed in Fig. 122, which express the

currents I xl, I yt, and I zt in terms of Ex~, Ey1 1 and E zl, the con-
stants of the positive-sequence network, and the generated
e.m.fs. From t.he negative-sequence network the three following
equations can be derived.
SIMULTANEOUS F AULTS 251
Ex2- ·- D xxdx2- Dxy2Jy2- D xzdz2

E112 = -Dxud x2 - D1111d112 - D11zdz2
E z2 = -D x:.2f x2 - Duz2f y2 - D zzdz2
Similar equations relating the zero-sequence constants can also
be set up. In this manner nine of the requisite equations are
determine d.
ee: Line-to-line faults at both points
6 i 6 Ezt = (Du2 + Z¡,.)[z¡ + Dxvzlut
+ (Dvv2 + Z¡ll)l11t
~e
Zfl! E111 = D xv2lx1
ee
6 Ext = (Du2 +
Z¡,.)lxt + aDxv2I 1 11
+
+ Z¡¡¡)lut
i 6 ~=a=
Ztll E¡¡t = a 2 Dz¡¡Úzt (Dl/112
Ezt = (Du2 + Z¡x) l zl + a2Dzu2l¡¡t

Ztx Zty Evt = aD,.llzlxt + (Dvu2 + Z¡v)lut
Line-to-line fault at x and single line-to-ground fault at y

~ E:r:t = (Du2 + Z¡x)lzl - D :r:uzlul
e b
Ztx
g~ Ellt = -Dxuzlz~ + (Dllv2 + Dv¡¡O + 3Z¡v)lvt
~ ~ Ext = (Du2 + Z¡.,)lxt- aDx112l11 1

e: b
Ztx
e:
¡zty Evt = - a 2Dxvzlz¡ + (Dvl/2 Dv¡¡O + 3Z¡¡¡)[¡¡¡
Ezt = (Du2 + Z¡,.)[,.¡ - a 2Dxvzlllt

+
te; ~
+
5 e:
z;;¡ E 11 1 = -aDx11 zlzt (DI/l/2 + Dll¡¡0 + 3Z¡u) l ut
If7.
Double line-to-ground fault at x and single line-to-
ground fault at y
1
Ezt = D
:z:z2
+ D
:tzO
+ az /z {Du2(Duo
+ 3Z¡,.)lxt -
[D,¡¡oDzz2 + D z¡¡2(Duo + 3Z¡,.)][¡¡¡)
E D xuoDu2 + Dxu2(Duo + 3Z¡,.)I +
¡¡l "" - D u2 + D uo + 3Z¡,. :z:l
(Dzl/2 - D z¡¡0)2 ]
[ (DIIv2 +
Dv¡¡O + 3Zlv ) - Dr:z:2 + D:z:zo 3Z!:o I ¡¡! +
1
E:.t = D
:t:t2
D +
z:xO
+ az {Du2(Duo + 3Z¡x) lxt -
/:t
[a 2DzvoDu2 + aD:.¡¡2~Duo + 3Z¡,.) ]lud
aDx¡¡0Du2 + a2D:.v2(Duo + 3Z¡,.) I +
E¡¡l = - D zz2 +Duo + 3Z¡,. zl
[
D 2 +D 11
+ 3z _ 2
D xu2 + Dzv2Dx¡¡0 + D 1
zv ]I111
lll/ ¡¡0 f¡¡ Dzx2 + Duo + 3Z¡., O
1
E xt = D + D + 3 Z {Du2(D uo + 3Z¡.,)lzt -
:t:t2 .c.cO /:t
[aDu2Dz¡¡0 + a2D x112(Duo + 3Z¡,.) llvt}
Eu! = a 2Dz¡¡0Dzz2 + a D xv2(D uo + 3Z¡z) I%! +
Du2 + D u o+ 3Z¡.,
D + D + 3 z _ D 2xv2 + D zu2Dz¡¡0 + D 2z¡¡0 ]l111
[ l/U
2 11
¡¡0 111
D u2 + Duo + 3Z¡x
Fw. 127 (Continued).-Equations relating positive-sequence quantities at fault
locations for simultaneous faults at two points in terma of negative- and zero-
sequence constants. Each pair of equations can be represented in the forro:
E :J:l = ki zl +mi 11 1; E 11 1 = ni zl + li ¡¡1 ·
At each oí the locations, the terminal restraints are imposed,

which, depending upon t he character and phases of the faults,
provide three additional equations of the type shown in Fig. 125
for each location. These restraints supply the nine equations
still required.
110. Simultaneous Faults for the Case of a Single Source of

E.M.F.
In the calculation of system short-circuit cunents asimplifying
assumption frequently used is that all of the generated e.m.fs.
are identical in magnitude and phase. In this case the positive-
sequence network can be handled much like the negative- and zero-
sequence networks. Neglecting for the moment the generated
e.m.fs. the self and mutual drop constants for the positive-sequence
At point x At point y
b~ b~ S!ngle line-to-ground faults at ~oth points
e e E u = (D.<xl + Dxxz +Duo+ 3Z¡x)l :rt +
ZfK Zfy (Dxut + D xvz + Dxuo)lvt
~ ~ Eu = (Dxut + D ..vz + D ..uo) l xt +
a~ a
b
e
Zfx
b
e i zf
Eu = (Dut +Duz+ D uo+ 3Z¡x) l :r.t
(Dvllt + Dnz + Dvuo + 3Ztli)Illt
(Dxvt aDxuz
+
+ +
a 2Dx~~D)I11 t
~ Y Eu = (D:r.vt + a 2 D .. vz + aD:r.uo)Izt +
"'~ oo (Dwt + Dl/Yz + Dv!IO + 3Ztli)I11 t
b b---------
e e Eu = (D:r.xt + D:uz +Du o+ 3Z¡:r.)l:rt +
::-
z
fx 1
_
-
Zf
) Eu = (D:r.vt + aDxvz + a 2 Dx~~D) lxt
(Dxvt
+
+ + aDxi~D)I11 t
a 2 Dxvz
(Dwt + D11vz + Dv!IO + 3Z¡,)I"t
a------- Line-to-line faults at both points

01
b b b Ea = (Dut + Duz + Z¡..,)I:r.t + (D:r.ul + D :r.uz) Iut
e
b
Zfx
e Eu = (D:r.ut + Dxuz)Izt + (Dnt + Duvz + Ztli)IIIt
Q
b
e
b Zfx
01
b
e
b Zfx
0 1 - - - - - - - ~~ Line-to-Iine fault at x and single line-to-gr ound fault e.t y

~:::::¡b::::;;:::: e Eu = (D xx t +Duz+ Ztx)Ixt +
(D:rut - D,uz)lllt
Zfy Eu = +(D:r.ut- D:x;uz)lxt + (Duut + Duuz +
Du!IO + 3Ztli)Ivt
Zfx
Q ---------- 01
~====¡6=::=
b
e
Íz-ry
a------~
Z-r,. 01
~ ====+G=::;=
b
e
Zfy¡
Zfx
FIG. 128.-Equations relating positive-sequence quantities at fault locations
for two simultaneous faults. All gencrated voltages equal.
network can be dete.r mined as indicated in Fig. 123 or 124. The

voltage equations for the positive-sequence network can then be
written in terms of these impedances and the generated e.m.f. Eo
as follows:
E 0 - Ex1 + Dull xl + Dx¡¡llyl
}(269)
E 11 - Eul + Dxud xl + D1111d112
SIMULTANEOUS F'AULTS 253
The values of E xl and E 11 t are also defined in terms of the
restraints due to the type of fault as given in Fig. 127 which
relations are in the form
Ext = k! xl + m/ 11 1
}(270)
Evt = ni xt + li
11 1
After substituting Exl and E 11 1 from equations (270) in equations

(269), the value of 1 xt and 1 vt may be expressed in terms of the
generated voltage E 0 and the system constants, with the following
result:
E 0 = (k + Dxxl)J xl + (m+ D.z: 11 ¡)l11 t
E o = (n + Dx111)1 x1 + (l + D 1111 ¡)l11 t
from which
(l + Dyyl) - (m + Dx11 1) Eo (271)
fxl =
(k + Dxxt)(l + D 1111 t) - (m + Dx11t)(n + Dx 11 1)
(k + Dxxl) - (n + Dxyl) E
/¡;1 = (272)
(k + D:ul)(Z + Dvut) - (m+ Dxvl)(n + Dx11 t) {1
The coefficients of equations (271) and (272), when developed

in terms of the positive-, negative- and zero-sequence quantitief.
result in quite symmetrical forms as may be seen from the table.
of Fig. 128, which has been prepared for the cases which result
in the more simple expressions.
111. Current and Voltage Distribution for Simultaneous Faults.
Having obtained 1 xt and / 11 1 at the fault points, the other
sequence currents at these points may be obtained through
the use of the relations shown in Fig. 125 and the sequence
voltages by use of equations (258) and (259). The current
distribution within the individual networks can be obtained
either analytically by working back through the networks in a
reverse manner, as in obtaining the self and mutual impedances
for the entire network, or by means of a calculating board by
applying the calculated voltages to the fault points.
112. Equivalent Circuits Representing Fault Conditions.
The solution of the equations from Fig. 127 in conjunction
with those from the positive-sequence network will, in general,
offer little difficulty. It is of interest, however, to provide a
complete solution utilizing the calculating board.
The solution of the equations from Fig. 127 may be expressed
in the form of equations (270). For the special case in which
m = n, considerable simplification results. This condition
holds for only the first case of each group in Fig. 127. The
positive-sequence network for these special cases may be repre-
sented by the equivalent circuit
shown in Fig. 129, in which the
mutual impedance is the common
Posifive-seqvence nefwork branch and the fork branches are
equal to (k - m) and (Z - n) for
the x and y points, respectively.
That this represents the equiva-
lent of equations (270) may readily
be verified as follows:
Fw. 129.-Equivalent positive-

sequence diagram for faults at x and
y for the special case in which
E xt = m(lx1 + l 11 t) +
(k- m)I xt = kl xt + ml11 t (
E 11 t = nClxt + lyt) +
! 273
)
m =n. (l- n)l11 ¡ = nlxt + ll 11 t

The cases in which m ~ n, are more involved, and for their
representation use will be made of a device <7 o) described by Miss
Edith Clarke. Equations (270) may be written:
m+n
Exl = (k-n)J x l + -- (l xl
2
m-n
+-
+ I yt) 2
- (1111 - I xt)
Posifive- sequence nefwork
(274)
m+n
E 11 t = (l-m) l 11 t +- - (l :r.l
2
m-n
+I11 t)+--(I 11 t-I .z:t)
2
Figure 130 shows the equivalent
circuit of the positive-sequence
network which will represent this
m-n
condition. The im pe dance
2
is paralleled by a source of poten- Fw. 130.- Equivalent positive-
sequence diagram for faults at x and
tia! which must be varied so that y for the general case. The imped-
the current flowing through the n-m
ance - - may be used in place oí
2
im pedance m ; n is equal to m-
- - n ; w h en t h'18 18
· d one, t h e cur-
2
Cl111 - Ix1). It may develop that rent through the 8hunt source oí
potential 8hould be 21 111 instead of
21:~:1•
the impedance m ; n involves a
negative resistance in which case the impedance n ~ m may be

used instead and the current through it adjusted to (1 xl - I.v 1).
m-n
If is used, the shunt voltage may be varied until the
2
current flowing through it is 21 xl (in both magnitude and phase)
and, when n ~ m is u sed, until the current is 21111 •
Special Cases. Two special cases, dueto Miss Edith Clarke, <7 o>
are worthy of consideration. The solution for the currents and
voltages at the fault points by these methods are about as
complicated as for the ordinary analytical methods. Their
principal value is in providing an equivalent circuit for stability
studies.
---------o--------
Remainder ofposifive- Remainder of posifive-
sequence nefwork Sf!quence nefwork
(Zhx+ k)(Zhy+lhn
Zhy+l-m
(b)
Ext =k lxt +m Iy1
Eyt = t1l')(1 + l lyt
(01)
Fw. 131.- Equivalent positive-sequence network for two simultaneous faults
on two unloaded feeders radiating from a common point. (a) Actual circuit ;
1
(b) equivalent circuit.
a. Equivalent Circuit for Simultaneous Faults on Unloaded

Feeders Radiating from a Common Point. In Fig. 131(a) let
the faults occur at the points x and y on the feeders radiating
from the common point h somewhere in the positive-sequencc
network. The positive-sequence impedances of the branches
between h and the fault points are Zhx and Zhy· The positive-
sequence voltages and currents at x and y are defined in terms
of the ncgative- and zero-sequence network constants by equa-
tions (270). The remainder of the positive~sequence network
is indicated by the box.
N ow, from the figure,
Eh = E :d + Zhx1 xl }(275)
Eh = E 11 1 + Zhy1y1
Substituting (270)
Eh = (k+ Zhx)Ixl + m/11 1 (276)
Eh= nlxl + (l + Zh 11 )111 1 (277)
.S olving
I Zh 11 + l - m E (278)
xl - (Zhx + k)(Zh 11 + Z) - mn h
1 111
= Zhx +k - n Eh (279)
(Zhx + k)(Zh11 + l) - mn
The reciprocals of these coefficients are the equiv~lent imped-

ances to be placed in the x and y branches and are shown in
Fig. 131(b). This circuit may be simplified a little further
by paralleling the two impedances, in which case the single
equivalent impedance reduces to
(Zhx + k)(Zhy + l) - mn
(280)
Zhx + k + Zhy + l - m - n
--------o--------
Remainder of' po.sifive- Remainderof positive-
.sequence network .sequence network
ftl 1
Ztt +ltt
Ztt
2
,"1f
/
1
Exl : k lxl + m lyl ZriZtl (k ti tm tn) + 2 (Zrt + Zt1) (kl- mn)
Eyl =- n lxt + llyt
(A)
, 4 Zr(Ztt + 2(IrJ t Zu) (k+l+mtn)
(b)
Fw. 132.- Equivalent positive-sequence n etwork for two simultaneous faults
on two transmission lines of equal impedances bussed at both ends. (a) Actual
circuit; (b) equivalent circuit.
b. Equivalent Circuit for Simultaneous Faults on Two Trans-

mission Lines of Equal Impedances Bussed at Both Ends.
Another important case that may arise is that of a double
simultaneous fault at a given point along the line of a double-
circuit transmission line. In this case the lines will be bussed
at points on either side of the faults and the positive-sequence
impedance between the fault points and the buses will be equal
for both lines. The equivalent circuit and explanation of the
circuit constants are sufficiently developed in Fig. 132 to be
self-explanatory.
113. Single Fault for the Case of Several Sources of E.M.F.
It has been pointed out in Chap. II that synchronous machines
are represented individually in the positive-sequence network
by an impedance and a source of e.m.f. which in general will
be different in magnitude and phase position for each machine.
For short-circuit calcubtion, as pointed out in Chap. IV, Sec. 24,
it is usually permissible to assume that these e.m.fs. are identical
in magnitude and phase position. The more general case
where this is not permissible will now be considered.
Network solutions with several separate sources of e.m.f.
may readily be obtained on the alternating-current calculating
board, since independent adjustment of magnitude and phase
position of the e.m.f. sources is provided.
The case of a fault ata single location on a system with several
sources of e.m.f. which are not identical in magnitude and phase
position may be considered as a special case of the problem
involving simultaneous faults. Using this method, the solution
may be obtained by following the steps outlined in the previous
section for two simultaneous faults by merely ignoring the
currents at points other than the :first fault location. Thus
equations (252), (258), and (259) reduce to
1 zl ~
= -
Zax
~ + ··· -
+ -Zbx E xl( Y x + -Zax
1 + -1
Zbx
.. )
•.' [Eq. (252a)]
' Ez2 = - Duz! x2 [Eq. (258a)]
Exo = - Duolxo [Eq. (259a)]
for a fault at a single location. These, together with the proper
set of equations from Fig. 127 for the type of fault, provide
enough equations to solve for all the unknown voltages and
currents.
An alternati ve method * may be used which is based on the
reduction of the positive-sequence network with its severa!
sources of e.m.f. to a single equivalent impedance and a single
equivalent source of e.m.f. Considering the simplest equivalent
general network for two sources and one fault point and using the
notation of Fig. 119, the total fault current for a single three-
phase zero impedance short-circuit at the point x is
I
z
= Ea
Zaz
+ Zb:c
Eb (281)
,.. This is b2.~~t:l uvon a. more general method described in Appendix I of
bibliography item ~b.
The relation that the same fault current is to be obtained from

the equivalent network gives the equation
_ E equiv.
I :z:- z
equiv.
(282)
If the equivalent impedance is assumed to have the same value
as exists when the e.m.fs. are identical then
1
= _!_ + _!_ (283)
Z equiv. Z az Zb:z:
The e.m.f. of the equivalent source is then given by
E equiv . _ Ea + Eb (284)
Zequiv. Za:z: Zb:z;
or
E . =
equtv.
ZbxEa
Z
+
+ ZZ axEb (285)
a:z: b:z:
The method using equivalent impedance and equivalent

voltage sources has been outlined for the case of two sources,
bu~ it is readily apparent that additional sources can be taken
care of by adding suitable terms to the right-hand· members of
equations (283) and (284).
Problems
1. Two grounded neutral gcnerators A and B, operating at no load, are
paralleled through a transmission circuit. The reactances on the base of
the rating of machine A are as follows:
Machine A Machine B Transmission Line
X¡ = 0.30 X¡ = 0.15 ;t;¡ = 0.25
X2 = 0.30 X2 = 0.15 X2 = 0.25
Xo = O Xo = O Xo = 0.75
What are the fault currents for a simultaneous line-to-ground fault on phase
a at the t erminal of each machine? Neglect all resistance and the capaci-
tance of the transmission line.
2. Determine the fault currents for short-circuits between phases b and e
on the terminals of machine A and between phases a and b on the terminals
of machinc B of Prob. l.
3. A 5,000-kva., 2,200-volt, 80 per cent power-factor generator is con-
nected to a transmission line whose reactance is 0.2 ohm per phase. R esist-
ance loads of 4.0 ohms are taken off across phases b ande at the terminals,
5.0 ohms across phases b and e at the end of the line, and 5.0 ohms across
phases a and e also at the end of the line. If x 1 of the machine is 1.0 ohm
and X2 is 0.5 ohm per phase and the positive-sequence internal voltage is
1,500 volts to neutral, what are the voltages between the phases at the
machine terminals and at the end of the line?
CHAPTER XIII
DETERMINATION OF SEQUENCE QUANTITIES FROM
PHASE QUANTITIES
Frequently it is desired to determine the value of sequence
quantities when only the ordinary single-phase instrument
readings are available. For three-phase systems involving
zero-sequence, six unknown quantities are required to define the
three vectors. These unknown quantities may be expressed in
terms of thrce amplitudes and three phase angles with respect
to an arbitrary reference, or in terms of six amplitudes, such as
t he t hree line-to-line voltages and the three line-to-neutral
voltages. For three-phase systems in which zero-sequence
does not appear, three amplitudes are sufficient to define the
vectors with respect to each other though an additional relation
is required to relate them to a reference vector. In this chapter
will be presented severa! methods using analytical expressions,
graphical constructions1 or charts by means of which the sequence
quantities may be determined from the phase quantities. The
developments will follow along two general lines, those for the
general case and those for the special case in which the three
vectors form a closed triangle, i.e., the case in which the zero-
sequence quantity is equal to zero.
114. Determination of Vectors from Amplitudes.

Before undertaking the determination of the sequence quanti-
ties, methods will be described for obtaining the vectors from
the amplitudes of the phase quantities. When zero-sequence
is absent the line-to-neutral voltages may be plotted to form a
~. closed triangle so that their phase angles with respect to each
¡
other are readily determined. In case the zero-sequence quanti-
ties are present, it is necessary to obtain six relations such
as given by the six amplitudes of the three line-to-line voltages
and the three line-to-neutral voltages. The phase relation
of the line-to-neutral voltage vectors may be obtained by first
plotting the line-to-line voltages to form a closed triangle as
259
indicated in Fig. 133. The line-to-neutral voltages may then

be plotted from the corresponding apexes of the triangle, and
the common intercept O will define the neutral point of the
system.
In case the zero-sequence quantity is absent and only the
amplitudes of the line-to-line
'"""- voltages are available, the
linc-to-neutral voltages may
be obtained as indicated
in Fig. 134. Form the
triangle from the voltages
EA, Es, and Ec. From any
Ee two apexes draw two lines,
/
01
such as ce and ad, to bisect
the opposite sides. Their
Fw. 133.- Determination of line-to- intercept at O' gives the
neutral vectors from the amplitudes of neutral point of the system.
Ea, Eb, Ec, EA, Es and Ec.
Thus the voltage vectors E a,
Eb, and Ec are the corresponding line-to-neutral voltages.
The validity of this theorem can be proved by showing that
•
the vector sum of E a, Eb, and Ec vanishes in the construction of
Fig. 134, that is e
Ea + Eb + E c = O (~86)
The construct.ion of Fig.
134 gives
Ba = O'a = !(de + ca)
=!(~EA+ Es)
=!EA+ iEB (287)
Similarly,
Eb = iEB + Fw. 134.- Construction showing that the
j-Ec (287a) intersection of the medians is the neutral
and point of t he system of threc vectors com-
Ee = i Ee + .¡EA ( 287b) prising the sidos of the triangle.
Adding equations (287) gives equation (286) and proves the
construction. It shows that Ea, Eb, and Ec are the particular
set of star voltages which produce the delta voltages EA, EB,
and Ec and have no zero-sequence component.
115. Analytical Expressions.
The sequence quantities expressed as vectors can be obtained
from the vector value of the phase quantities with the aid of the
fundamental equations (27), (28), and (29) given on next page.
DETERMINATION OF SEQUENCE QUANTITIES 261
Eao = i(Ea + Eb + Ec)

Eal = !(Ea + aEb + a Ec)
2
Ea2 = HEa + a 2Eb + aEc)

When the three vectors form a closed triangle, i.e., when
Eao = O, the expressions for the positive- and negative-sequences
become, after substituting Ea = - Eb - E e in equations (28)
and (29),
Eal ·= i[(a - l)Eb + (a 2 - l)Ec] (288)
_ _ _l:_(Ebe-iaoo +E etaoo) (289)
y3 e
or
(290)
And
Ea2 = i[(a2 - l)Eb + (a - l)Ec] (291)
= - _l_(Ebe¡aoo + Ece- ;aoo) (292)
V3
eiaoo
= - yi3(Eb + Ece-i 600
) (293)
For this special case other methods are available which do

not require a knowledge of the phase values in vector form.
Kennelly,can Sah,<38> and Norman< 40 > have developed expressions
which give the positive- and negative-sequence components in
terms of the amplitudes of the three unbalanced vectors forming
a closed triangle. These are
(294)
and
(295)
in which
(296)
A- 8 2--
Vi(Ea + Éb + Éc)(Éa + Eb- E c)(Ea-Eb + É c)( -Ea+ Éb +Ec)
(297)
It is interesting to note, as Dr. Kennelly has observed, that
"Am 2 is the mean of the squares of the three unbalanced sides and
As 2 is the square of the equilateral side of an unbalanced system
having the same area as the unbalanced triangle."< 37> Evans and
Heumann have developed expressions (not previously published)
•
that reduce to the above but in addition enable one to obtain

the phase relations of the sequence components. They have
simplified the expressions by letting Ea = 1, F:_b = x, and ~e = y.
Ea Ea
The analytical expressions for the various quantities are given
in Fig. 135.
116. Graphical Constructions.
The methods for determining the sequence components of
three-phase systems by graphical construction may conveniently
be divided into two groups: (a) methods for the determination
1
r = 2 + VJ. 2 3
-
V
1
(1 +X + y)(l +X - y)(l -X+ y)( -1 +X +y)
t =
1
2- v
1
2 3
-v(1 + x + y )(1 + x - y )(1 - x + y) e_ 1 + x + y)
y2 _ x2
S = ..;;__---==--
20
8
82 = tan- 1 - -
t
Eal = v?=2+8 Ea 2
Ea2 = v't 2 + s2 Ea
Fw. 135.-Analytical expressions for positive- and negative-sequence com-
ponents of a set of vectors without a zero-sequence component. (Evans &
Heumann.)
and subsequent elimination of the zero-sequence component, and

(b) methods for the determination of the positive- and negative-
sequence components from vectors not containing a zero-sequence
component.
a. Determination and Elimination of Zero-sequence Compo-
nents. One method for determining the zero-sequence compo-
nent is shown in Fig. 136. It is based on the fundamental
equation (27) and consists in taking one-third of the vector
sum of the three phase vectors. The method is simple and
gives directly the quantity desired.
r
.f
1
Another method, shown in Fig. 137(a), rcquires the determina-

tion of the neutral point of the system with zero-sequence
components eliminated. As described in Sec. 114, the neutral
point of this system of
vectors i.s found at the point
O', the intersection of two
.
¡
~
' median lines. The vector
connecting the points O and
~ O' gi ves directly the zero-
s e que n e e vol t a g e in its
¡ proper phase position with
respect to the line-to-neutral
~
1
vectors E a, Eb, and Ec. b
Ei ther of these methods Fw. 136.-Determination of the zero-
provides the zero-sequence sequence component by taking one-third of
thc sum of the phase quantities.
component from which the
three new vectors indicated in Fig. 137(b) as (Ea- E o), (Eb- E o) ,
and (Ec - Eo) can be ob\iained and which include only the
positive- and ncgative-sequence components.
r e b. Determination of
i
¡. Positive- and N egative-
t
~ sequence Components.
1' Referring to the funda-
~ a mental e q u atio n s (28)
and (29), these may be
rewritten as
1 E al = !(Ea + Ebei 1200 +
(Q)~ Ecei2400 ) (298)
Eaz = !(Ea + Ebei2400 +
E cei1200 ) (299)
E.,.·Eo
(b) From the standpoint of
g r a phi e a 1 constructions,
Fw. 137.- Determination of zero-sequence
component from intcrse')tion of medians.
these signify that to deter-
(a) Determination of E 0 ; (b) closed triangle
after eliminating E o. mine E a l, Ebmust be rotated
through 120 dcg. and E e through 240 deg., both in a counter-
clockwise direction and the resultant vectors added together with
Ea. The resultant vector is equal to 3Eal· Similar relations
hold for Ea2· These constructions are shown in Fig. 138.
After rotating Eb and Ec, one-third of the sum of the three
vectors may be obtained conveniently by constructing a triangle
whose apexes are the termini of the three vectors whose sum is
desired. The intersection of the median lines of the triangle
so formed locates the terminus of the vector E al· While this
construction is not shown, it may be verified by trial in Fig.
120°
'1
1
'
Fw. 138.-Graphical construction for positive- and negative-scqucnce com-

ponents.
138(a). A similar coustruction is applicable to the negative-

sequence system. This construction, applied to positive- or
negative-sequence, corresponds to that described in Fig. 137
•
for zero-sequence .
/
,... - - 1'20°--...... ,
........
/ ''
/
/
' ,---
\ ............
1'20°
'\\
(01)
Fw. 139.-Graphical construction for positive- and ncgative-sequence volt-
ages when the voltages forro a closed triangle (Eo = 0). Vectors rotated 120
deg. (a) Positive-sequence; (b) negative-sequence.
For three-phase three-wire systems or more generally when

the zero-sequence components are equal to zero, a number of
interesting constructio~s are available. Most of these are based
on the fact that for this case the vectors may be drawn as the
sides of a closed triangle. One of these methods is given in
Fig. 139, which, in fact, is the same as that shown in Fig. 138
except that the construction is on the triangle. The method is
self-explanatory and appears as a very direct application of the
e Ea b
(a) Construction to Determine (b) Alternoote Construtt10n
.f5 Eat Giving Ecot Di rectly
(e) Construc+ion Givmg A\1
Three Ph01ses of the Positive·
Sequence Volt01ees
FIG. 140.- Graphical construction for positive-sequence voltages when voltages

form a closed triangle (Ea o = O). Vectors rotated 30 deg. (Genkin.)
(b)
(01)
Fw. 141.- Graphical construction for negative-sequen ce voltages when voltagee
forma closcd triangle (Eao = 0) . Vectors rotated 30 deg. (Genkin.)
(b)
FIG. 142.- Graphical construction for positive- and negative-sequence voltages
when voltages form a closed tria ngle (Eo = 0). Vectors rotated 60 deg. (a)
Construction for Eat; (b) construction for Ea 2·
fundamental theory. The particular construction illustrated

in Fig. 139 is included in the German Electrical Standards. *
In Fig. 140(a) is shown a construction by Genkin,Cl 34) based
on equation (289). Vectors Eb and Ec are each rotated outward
from the triangle through an angle of 30 deg. from their common
apex as a center and the terminals connected together. The
* Regeln für die Bewertung und Prüfung von elektrischen Maschinen,
§16.
resultant vector gives the correct ohase position of Eal, but its
amplitude is equal to V3Ea1· Similar procedures apply to
the other apexes and provide the positive-sequence components
of voltage in the remaining phases. For the negative-sequence
the vectors must be rotated inward, with the lines connecting
the terminals as before. This gives the vector of the proper
1.0
(b)
Fw. 143.- Positive-sequence component of systems of closed t riangles.
phase position but whose magnitude is v3Eaz· This construc-

tion is shown in Fig. 141(a).
A modification of this method has also been given by Genkin
whereby the result gives not only t he correct phase position but
also the correct magnitude. This modification is shown in Fig.
140(b) in which isosceles triangles are constructed upon the sides
of the triangle by drawing lines making angles of 30 deg. with
the sides. The apexes of these triangles are connected together
giving the desired vector. The method lends itself to the
determination of the positive- and negative-sequence vectors
in each of the three phases as illustrated in Figs. 140(c) and
141 (b), respecti vely.
The preceding two methods may be briefly described as based
on (1) the rotation of two vectors through 120 deg., and (2)
on the rotation of two vectors through 30 deg. A third method
is also available which can be similarly described as being based
on the rotation of but one vector through 60 deg. This method
1.0
E~
0.8 E.,.
o o
(01)
1.0
0.99 099
0.98 098
0.97 0.97
0.96 0.96
fb 0.95 0.95 Ec
0.94 0.94 -
Ea 0.93 0.93 Ea
0.92 0.92
0.91 0.91
090 0.90
0.89 0.89
0.88 0.88
0.87 0.87
0.86 0.86
085 o.a~
(b)
Fw. 144.- Negative-scquence component of systems of closed triangles.
is shown in Fig. 142 and is bascd on equations (290) and (293).

Given the thrce vectors formed by the closed triangle abe,
constructing an equilateral triangle externally on any side such
as ab (which is equivalent to rotating that side through 60 deg.),
and connccting the apex d of that triangle to the apex e, opposite
the side used as a base for the equilateral triangle, gives a line
whose length is equal to v3
tÚnes the positive-sequence compo-
nent. Similarly, constructing the triangle internally and con-
necting the apexes, a line is obtained whose length is equal to V3
times the negative-sequence component as shown in (b). The
positions of these lines do not give the correct phase position

of the components. For the positive-sequence it is necessary
to rotate the vector backward 30 deg. as shown in (a), and
for the negative-sequence component forward through 30 deg.
as shown in (b).
Of the three types of construction described, that of Fig. 139
is the most fundamental and direct but involves an operation
u pon two quantities through inconvenient angles; that of Figs.
140 and 141 also involves operation upon two quantities, but
the graphical construction is somewhat easier; but the construc-
tion of Fig. 142 is the simplest when only the magnitude of the
sequence components is desired, as it involves operation on only
one quantity by means of a very simple construction.
117. Charts.
The positive- and negative-sequence components of voltages
or currents on three-phase three-wire systems or more generally
where Eao = O may conveniently be determined by the charts
of Figs. 143 and 144 for positive- and negative-sequence, respec-
tively. These charts are based on the vectors forming a closed
triangle as indicated in the figures. The ratios of the two shorter
sides to the longest side are used as coordinates with the point
so located asto determine the magnitude of the ratio of positive-
sequence voltage to the voltage of the longest side which was
chosen as a reference. Figures 143(b) and 144(b) have been
drawn t6 a larger scale in order to give better accuracy for the
ordinary small unbalances encountered in practice. The nega-
tive-sequence components are dctermined in a similar manner.
If only the amplitude of three voltage vectors is known, then it
is not possible to determine whcther the triangle should be
plotted as shown above the reference line Ea or below it. Con-
sequently, it is not possible to determine whether thc larger
component is the positive- or thc ncgative-sequence component.
This, in general, can be determined from sorne other information.
However, in the usual problem, the positive-sequence quantity
will predominate.
118. Unbalance Factor.
The ratio of the negative- to positive-sequence amplitudes
is commonly called the unbalance factor and is included in the
German Electrical Standards. This factor was proposed by
~!.:"-"' .,._:~1""'1;;~¡:,.-~
'Eb
E,. to 1.1 1.1. 1.3 1.4 1.5 1.6 17 1.8 1.9 2.0 2.1 22 23 2.4 25 2.6 2.7 2.8 2.9 3.0
20r-~-r~~--~-r-.~--~Ar7r~~~~-r~--v-.-rr~--~.--r~-M~T7-r-n~~
1.9 1-+-+-+-+---+--+-1-+-~L.Ah.L+--J,f.-+--.I;~-Y---+
L8t--+--+-l-+--+-+-+--.~'--/- ~
~
17r-t-,_-r-t-4--t-/.0~~~r-~+-~~~~~~~~~~~~h-~~~~~~~~
L61-+---t--t-f--+::-~>4f7L-b.LT---7l~+-l--;kH-/-fr:---b-L4~4-b<~--Ji~f-4l--.J..4ih~4!:~__;
~
~
1.4
1.5 ~
Eb Ec
-~
~
E012 ~
13 8 ......
1.2 E"1
E" a
~
~
~
{:)
0.7 ~
~
~r-+-~~~~~dA~~~~~~44~~~~~~~ 104
u¡u.u 1.00 ~
o.s 1-+--+--t-+- 1\&1
G
~
0.4 1-+--+--+----l- cr->~~Ft-ol-::::l:±-::~~---:AS1'7!T-:-f~ 096t32o·~H <:::!
;:c..
Q3r-+--r~-+--+-~~~~~~~~~~~~ 0.92 ~
~
021-+--+--+~--+--~-h·~ ~~~~~ 0.88 ......
~
0.1 1--1---+----11-+--+-1-+--··:.n·~~-l OM ......
ÜL--...1--'----''--...l.-___L_L__....L-----l...___..:~.:.IL..l:...l..___l
o 01 Q2 0.3 Q4 0.5 0.6 0.7 Éb
~OL-~~~~L--L~-~~__¡_---'
080 OB4 O.Ba 092 Q96 lOO 1.04 1.06 liZ 116 1.20
~
Ea Eb
Ea t-=>
FIG. 145.-Determination of unbalance factor. 0':1
1;,0
Evans and Pierce for measuring the unbalance in voltages and

currents on three-phase three-wire systems. They also devised
means of measuring voltage and current unbalance factors. *
The unbalance factor may conveniently be determined by
means of cha~s, such as those given in Fig. 145. In this figure
the ratios ~b and ~e are used as coordinates and the point thus
Ea Ea
determined gives the unbalance factor by interpolation of a
family of curves. The angle plotted in the diagram gives the
angle by which the vector Ea2 leads the vector Eal· The smaller
chart of Fig. 145 gives the unbalance factor on a larger scale and
is convenient for estimating the arnount of unbalance on systems
that are approximately balanced.
The first work along this line was published by R. Dubusc< 143 >
in 1927 and arnplified by von G. Hauffe< 161> in 1929 and by the
authors in this work.
Problems
l. The line-to-line voltages of a system measure EA = 75, EB = 65, and
Ec = 40. Determine the positive- and negative-sequence components in
the three phases by graphical construction.
2. Using the values of phase voltages from Prob. 1 determine the positive-
and negative-sequence componcnts, thc unbalance factor, and the anglo
between the components by means of the charts in Figs. 143 to 145.
3. Given the voltagcs:
Ea = 25 - j2.()
Eb = - 13 - j2.0
Ec = O + j15
Determine thc three scqucnce
components graphically.
4. Prove the relations of
Fig. 135.
e~====----::::..t.-~-j--j----:~b 5. Dubusc in his report to
the International Congrcss on
e Large Elcctrical Networks in
1927 has given the construc-
tion shown in the accompanying figure to obtain the positive- and negative-
sequence components. Given the voltagc triangle abe, draw the median ad.
On this median construct the 30-deg. right triangle adc. The positive- ancl
negative-scquence voltage components relative to the a phase are then givcn
by ce and eb, respectively. Prove this relation.
*U. S. Patents 1,535,589 and 1,567,582.
.'
CHAPTER XIV
MEASUREMENT OF SEQUENCE VOLTAGES AND

CURRENTS
Sequence components may be measured by means of instru-
.~·
'
ments or meters operating on well-known principies requiring
only spccial externa! networks or in sorne cases special windings.
In all the previous discussions, sequence quantities have been
treated as component vectors of the single-phase vector quantities
measured in the different phases. For the proper understanding
of the prcsent discussion, it is preferable to consider the sequence
quantities as having an existence as real and definite as the single-
phase quantities; for example, positive-sequence voltage is
just as real as a line-to-line voltage and will be measured in a
similar manner by a meter element and an externa! network.
Sequence quantities that can be measured include for each
sequence the same quantities that can be measured in a single-
phase circuit, namely, voltage, current, real power, reactive volt-
amperes, power factor, volt-amperes, etc. In addition, the
sequence quantities may be combined in various ways for meas-
uremcnt. The sequence quantities may be used to operate
indicating, intcgrating, and recording instruments and meters.
In addition, they may be uscd to opcrate relays of the various
kinds, such as protective relays and voltage regulators.
The first proposal for measuring sequence quantities, such
as voltage or current, was the result of a discussion between
Fortescue, Chubb, and Slepian. They contemplated the use
of rotating machines to eliminate one sequence to permit the
measurement of another sequence quantity. The fi.rst measure-
ments, however, were made by R. D. Evans who proposed static
networks to segregate a desired sequence quantity in the presence
of severa! sequence quantities. A large number of different
networks have been worked out by Evans, Allcutt, Fortescue,
and others. The first device actuated by a sequence component,
embodied in a commercial form, was the negative-sequence
overcurrent relay. Subsequently, voltage regulators were built
271
to be operated by positive-sequence voltage. At the present

time, other devices responding to sequence components are
being used for a number of purposes and are being considered
for many other applications.
On a three-phase four-wire system there may exist the three
sequence voltages. Consequently, in order to measure all
of these sequence quantities on a commercial system, it is neces-
sary to have available positive-, negative-, and zero-sequence
voltmeters. On a three-phase three-wire system the zero-
sequence voltage is indeterminate and the corresponding volt-
meter is unnecessary and inapplicable. Similar relations hold
for the sequence currents, and positive-, negative-, and zero-
sequence ammeters are necessary. In general, positive- and
negative-sequence voltmeters are similar but are different from
the zero-sequence voltmeters which are simpler. The sequence
ammeters may likewise be clasRified as positive- and negative-
sequence ammeters and zero-sequence ammeters.
In this chapter on sequence measuring devices, the various
types of sequence voltmeters and ammeters will be studied. In
the following chapter the various combinations of sequence volt-
ages and currents in wattmetric devices will be considered.
119. Measurement of Zero-sequence Voltage.
The zero-sequence voltage, as shown previously, is determined
from the line-to-neutral voltages by means of equation (27):
Eo = !(Ea + Eb + Ec)
=+====~=========~
-+-------r------~-----c
r;u:=r (01)
Fw. 146.- Zero-sequence voltmeters.
Consideration of the above equation indicates clearly that the

zero-sequence voltage may be measured by means of a voltmeter
reading one-third of the sum of the three line-to-neutral voltages.
MEASUREMENT OF SEQUENCE VOLTAGES 273
Several alternative methods for measuring zero-sequence voltage
are shown in Fig. 146. The form most commonly used is shown
in Fig. 146(a); it makes use of three potential transformers and
a standard voltmeter calibrated to read one-third of the voltage
actually impressed across it. Figure 146(b) requires three
accurately balanced impedances with the voltmeter connected
between the star point and neutral or ground. Calibration must
take into account the three parallel impedances f, in series
with the meter across the calibrating voltage. The special
winding arrangement of (e) may also be used. For calibration
the three terminals connected to phases a, b, and e should be
connected together to form one side of the circuit upon which
the calibrating voltage is impressed.
120. M easurement of Positive- and N egative-sequence Voltages.
The positive-sequence voltage as shown previously is deter-
mined from the line-to-neutral voltages by means of equation
(28):
E1 = }(Ea + aEb + a 2E c)
The negative-sequence voltage is determined in a similar manner
by equation (29):
E"' = }(Ea + a 2Eb + aEc)
Equation (28) defines the conditions which must be satisfied by
the network for the meter to measure positive-sequence voltage.
Thus, the meter must respond with equal magnitude to voltages
from phases a, b, and e, individually, but the currents through
the meter must have relative phase angles of O, 120, and 240
deg., respectively, corresponding to 1, a, and a2 , respectively.
The phase shift of zero degrees may be obtained by using a very
large resistance externa! to the meter winding. Phase shifts
of 120 and 240 deg. by externa! impedances alone would involve
negative resistance. This diffi.culty is overcome by reversing
the connections of windings b and e with respect to winding a
as indicated in Fig. 147. The effective value of the meter
current is indicated by
]M = Ea _ Eb _ Ec (300)
Za Zb Zc
With this connection the power-factor angles of the impedances
1
in series with windings a, b, and e, from comparison of equations

(28) and (300), should be O, +60, and -60 deg., respectively,
for the positive-sequence voltmeter.
For measuring the negative-sequence voltage, it may be
shown in a similar manner, from equations (29) and (300), that
the impedances Za, Zb, and Zc should have the values given in
connection with Fig. 147(b).
01 01
b b
e e
~ z¡,; le? Za; 7b~ Zc~
.1 1
:u:.J:JI :....,._,._._
1
\J\I:J :':H.J
'-"""'
~ :\HJI :·:u
""'""""' ...............
1
¡ :f·.J':J
·*
_... ~ ......
~ Q
(01) Pos itive- Sequence
.L/
(b)
j Q~
Neg01tive>- Sequence
General Relations
Zb = Zae+iGo• Z b = Z ae-; 6 o•
Z c = Zae-iGo• Z c = Zae+iGo•
Typical Values
Za = R Za = R
zb = R(0.5 +
j0.866) zb = R(0.5 - j0.866)
Zc = R(0.5 - j0.866) Zc = R (0.5 j0.866) +
Phase rotation a b e
for positive-sequence voltage )-
Fw. 147.-Positive- and negative-sequence voltmeters.
121. Common Forro of Positive- and Negative-sequence Volt-

meters Applicable in Absence of Zero-sequence Voltage.
In general the voltage vectors of a three-phase system possess
three degrees of freedom, and therefore the network branches
must possess a corresponding number, which in the networks of
Fig. 147 were obtained by the three impedances shown. How-
ever, if thc zero-sequcncc voltage wcre eliminated, the meter and
network would require only two degrees of freedom. The
elimination of ze.r o-sequence voltage may, of course, be obtained
by using line-to-line voltages instead of line-to-neutral voltages.
Special windings in the voltmeter are unnecessary as the instru-
ment may be made to read directly by appropriate connection
and adjustment of the externa! network. Thus positive-sequence
voltage may be obtained directly by means of the simplified
network due to C. T. Allcutt, illustrated in Fig. 148. Since

this network is the one most commonly used, an analysis of
its operation will be given. The basic equations of this type
of network may be derived from an application of Kirchhoff's
laws giving the equations
Ea- Ec = ZBJB + M(IA +lB)
Ec- Eb = ZAJA + M(JA +lB)
(01) Positive- Sequence (b) Neg01tive -Sequence
General Relations
Typical Values
ZB = R ZA = R
ZA = R ei60" ZB = Rei6o•
= R (0.50 + j0.866) = R(0.50 + j0.866)
for positive-sequencc voltage
FIG. 148.- Positive- and negative-sequence voltmeters. Applicable in absence

of zero-sequence voltage. The most commonly used connections.
Multiplying the first equation given above by ZA and the sec-

ond by Z B and adding them, and remembering that I A + I B = I M,
there is o btained:
JMZA(Ea - E c) + ZB(Ec - Eb)
= (302)
ZAZB + MZA + MZB
Expressing Ea, Eb, and E c in terms of the sequence voltages,
the meter current IM becomes
IM = (1 - a)(ZA + aZB) E + (1 - a 2 )(ZA + a 2ZB) Ez (303 )
ZAZB + M(ZA + ZB) ZAZB + M(ZA + ZB)
1
The condition for the meter to measure only positive-sequence

voltage is that the coefficient of Ez in equation (303) must equal
zero. Hence, the necessary condition defining the positive-
sequence segregating network is
ZA = -a 2ZB = ZB e+i600 = Z 8 (0.5 j0.866) + (304)
For the negative-sequence segregating network a similar expres-

sion may be obtained by setting the coefficient of E1 in equation
(303) equal to zero which gives:
ZB = -a2ZA = ZA e+i600 = ZA(0.5 + j0.866) (305)
The network most commonly used with the positive- or
negative-sequence voltmeter consists of two branches, one a
resistor and the other an inductive impedance with a power-
factor angle of +60 deg. as indicated in the tables accompanying
Fig. 148.
The quantity measured by the positive-sequence voltmeter
of Fig. 148 is obtained by substituting equation (304) in equation
(303) which gives
3El 3El
¡MI = Zn + M(1 - a) - 7 B + y3Me-i30° (306)
- (306a)
ZA + y3Me+i3oo
The corresponding expression for the negative-sequence volt-
meter is given by
3E2 3E2
(307)
1M2 = ZB + M(1 - a2) - Zn + y3Me+'3oo
3E2e-1600
- --- (307a)
ZA + V3Me-i300
Vector Diagram and Range of Impedance Values. Equation
(306) may be used as a basis for a vector diagram for the positive-
sequence network of Fig. 148. In this diagram, shown in Fig.
149, resistance is plotted as the abscissa and reactance as the
ordinate, plotted upward if inductive and downward if conden-
sive. If the meter impedance M is zero, the current through
the meter willlag behind the positive-sequence voltage of phase
a by the power-factor angle of the impedance Zn, and if Zn
is apure resistance the current will be in phase with that voltage.
If the meter impedance is not zero, then the equivalent impedance
of the meter and network may be obta.ined by adding to the
impedance Zn the impedance M multiplied by V3 and shifted
through 30 deg. in a clockwise direction as indicated in the
diagram. The current through the meter will be equal to three
t\mes the positive-sequence voltage divided by the effective
impedance OD and the phase of the current through the meter
element wíll be shifted back of the positive-sequence voltage
of phase a by the angle that OD makes with the X-axis as
indicated by the position of IM. It is important that the trans-
former connections shown in Fig. 148 be followed carefully
for the above conditions to be fulfilled.
According to equation (304), the only condition which must
be satisfied, so that the meter reading be proportional to the posi-
tive-sequence voltage, is that the power-factor angle of ZA lead
that of Zs by 60 deg. By plotting this relation, the diagram of
Fig. 149 may be used to visualize the phase-angle range of the
impedances ZA and Zs. The limitation of avoiding negativ~
resistance in both quan-
+jX
ti ties fixes the range of
the impedance angles of D
these quantities to the

values
Z A = R ( e- i3oo to e+i9oo)
z s = R ( e-i9oo to e+'3oo)
Furthermore, it is not
feasible to reach the ex-
treme range of ei ther
e+i900 or e- i 900 because of
the fact that all circuits
have sorne resistance. -jX
The most important case FIG. 149.- Diagram for studying positive-
sequence segregating networks of the type shown
is that in which Zs is in Fig. 148.
pure resistance and ZA Phase a ngle of meter current = negative qf
is made up of a resist- phase angle of OD.
. dUCt ance, the M agmtu 3 1
ance an d In . d e o f meter current = . d E (OD)
1mpe ance
combination having a
power-factor angle of +60 deg.
A negative-sequence diagram similar to t hat of Fig. 149
for the positive-sequence can be derived from equation (307).
It is to be noted, however, that the phase-angle shift for the nega-
tive-sequence will be different from that for positive-sequence.
Conversion of Sequence N etworks- Interchange of Leads or
Network Impedances. A positive-sequence voltmeter may be
converted into a negative-sequence voltmeter by the mere
interchange of two leads between the network and the system
whose voltage is being measured. Thus the positive-sequence

voltmeter of Fig. 148(a) becomes a negative-sequence voltmeter
by tapping network leads lz and Z3 to phases b ande, respectively.
Also, a positive-sequence network can be converted into a
negative-sequence network by the interchange of the phase-
shifting impedances, for example, by interchanging the imped-
ances of branches ZA and ZB of the network of Fig. 148. This
interchangeability of the positive- and negative-sequence net-
works is a result which is borne out by analysis and is to be
expected since the mere interchange of two leads on a three-
phase system converts a positive-sequence voltage to a negative-
_sequence voltage and vice versa.
In connection with Fig. 148(a), it should be pointed out that
if leads l1, lz, and la are tapped to phases abe, bca, and cab, respec-
tively, the voltage applied to the meter will be identical in
magnitude, but the phase angle will successively be shifted by
120 deg. in the same sense as the positive-sequence voltages.
In the case of the negative-sequence voltmeter of Fig. 148(b), the
tapping of leads Zt, lz, and la to phases abe, bca, and cab, respec-
tively, cause the meter to read identical in magnitude, but the
phase angle of the voltages applied to the meter will succes-
sively be shifted by 120 deg. in the same direction as the
negative-sequence voltage considering phase rotation abe for
positive-sequence voltage.
122. Second Form of Positive- and Negative-sequence Volt-
meters Applicable in Absence of Zero-sequence.
The positive- and negative-sequence networks that have just
been described are characterized by the fact that the impedance
angles of the two impedances differ by 60 deg. and one transformer
connection is reversed. Another form of network, which may
be used when the zero-sequence component of voltage is zero,
is characterized by the fact that no reversa! in transformer
connection is required but the impedance angles of the two
impedances must differ by 120 deg. A network of the latter
class is indicated in Fig. 150. The network may be made with
the impedances
ZB = R(0.5 - j0.866)
Zc = R(0.5 + j0.866)
for which one branch is a condensive impedance and the other
branch is an inductive impedance. If these conditions are
fulfilled, and the transformer connections of Fig. 150 followed,

the current through the meter will have the same value relative
to E1 as that in Fig. 148.
a Ot
ffi b
e
EL
z,~;;:""~"'~:+,
b
e
Zc
(a) Positive -SeO(uence (b) NeeJOitive- Sequence
General Relation
Zc = aZB ZB = aZc
Typical V al u es
ZB = R(0.500 - j0.866) ZB = R(0.500 + j0.866)
Zc = R(0.500 + j0.866) Zc = R(0.500 - j0.866)
for positive-sequcnce voltage
Fw. 150.-Positive- and negative-sequence voltmeters. Scgregating network
with impedance angles differing by 120° instead of 60° as in Fig. 148.
------------~----- 01
----~-------r-----b
-----r--;---~-----c
(e)
----------------~-01 ------------------~---~
--------~-------+-b ------------~----~---b
~------~---4---+-c
----~----~~----~---<:
(b) (o!)
Fw. 151.-Connections for eliminating transformers in positive-sequence volt-
meters. (a) Special windings with network constants similar to Fig. 148; (b)
Fig. 148 with one transformer omitted; (e) Fig. 148 with both transformers
omitted. (d) Negative-sequence vol tmeters may be obtained by merely inter-
changing leads or network impedances.
123. Reduction and Elimination of Voltage Transformers for

Sequence N etworks.
In sorne cases it will be desirable to reduce the number of
potential transformers required with the positive- or negative-
sequence voltmeters or even to eliminate them completely.

Various schemes for this purpose are outlined in Fig. 151. The
connection of (a) with two windings on the meter obviously
avoids the necessity for potential transformers, since the connec-
tion of the meter winding may be made in such a manner as
to give the same result as with the transformer connections of
Fig. 148. In case the use
G1 of a voltmeter with special
b
e windings is not desired, one
of the potential transform-
ers of Fig. 148 may be
tooooooo [.Q~QQQQQ
omitted as indicated in Fig.
151 (b) . If there is no objec-
tion to the use of a con-
denser as an element in the
network, then the trans-
z~ formers of Fig. 150 may be
eliminated and the connec-
tion reduced to that of Fig.
151(c). A fourth method
involving only one auto-
transformer is shown in Fig.
General Relations 151 (d).
Z'A = -a 2 Z'n
Z"n = -a 2Z"A 124. Simultaneous Meas-
Typical Values urement of Sequence
Z"A = Z'n = R Voltages.
Z'A = Z"n = R(0.500 + j0.866)
Sometimes it is desired to
Fw. 152.- Pos i ti ve- and n egative-
sequence voltmeters for simultaneous meas- measure simultaneously both
urement in meters Vt and V2, respectively. the positive- and negative-
sequence vol t a g es . The
simplest network for this purpose is a combination of (a) and (b) of
Fig. 148, as indicated in Fig. 152. In case it is desired to measure
all three sequence voltages simultaneously, the connection of Fig.
146(a) may be used for t he zero-sequence with auxiliary potential
transformers to give the equivalent of Fig. 152 to obtain the
positivc- and negative-sequence voltages. In case special meter
windings can be used, it is possible to superpose on the connec-
tion of Fig. 146(c) for zero-sequence, Fig. 151 (a) for positive-
sequence and a similar one for negative-sequence, with the branch
impedances ZA and Zn interchanged.
125. Measurement of Zero-sequence Current.
The zero-sequence current, as shown previously, is deter-
mined from the line currents by means of the expression
lo = i(I a + h + l e) (308)
Consideration of the above equation indicates clearly that the
zero-sequence current may be measured by means of an ammeter
reading either (1) one-third of the sum of the line currents or
(2) the neutral or ground current. Severa! alternative methods
for measuring zero-sequence current are shown in Fig. 153. In
~
01 01
'"""""""
hnnnnn.
b b
e e
n
L@
(01}
- - 01 --------~------0
b ------~-+------b
¡--- e ----~~-+------~
n
,,
1 1
:1( )~
~~r-----------01 01
-+~~~T-------b b
-+--r-r-+-~~--c e
-+--~~+-+-4---n n
o
lo
l®- L@ L@
(f)
Fw. 153.-Zero-sequence ammeters.
the schemes of (a), (e), (e), and (f) the sum of the three line
currents are used, while in the schemes of (b) and (d) the neutral
or ground current is used. The most common connections
make use of current transformers but direct measuring devices
are shown in (e) and (d). In connection with the schemes of
(b) and (d), it should perhaps be pointed out that it is essential
that all the zero-sequence current flows through the neutral
wire or ground connection in which the meter is installed, or
that this current bears a known proportionality to the total

current. The scheme used in Fig. 153(a) requires the bussing
of all current transformers. This restriction may be avoided
by the use of separate ammeter windings as in (e), or by the
addition of separate current transformers as in (f). In the
latter case it may be noted that if the current transformers
have a ratio of 3 to 1, the ammeter will read zero-sequence
current directly. With the schemes of (e) or (f) the ordinary
current transformer loads, indicated by L in the figures, may be
connected in delta, for example, without interfering with the
measurement of zero-sequence current.
126. Measurement of Positive- and N egative-sequence Cur-
rents.
The positive-sequence current, as shown previously, may be
determined from the line currents by means of the following
expression
(309)
The negative-sequence current is determined in a similar manner
by means of
(310)
From equation (309) it will be seen that the meter must respond
with equal magnitudes to currents from phases a, b, and e,
individually, but the currents through the meter must have
relative phase angles of O, 120, and 240 deg., respectively,
corresponding to 1, a, and a 2 , respectively. Figure 154 shows a
meter and network suitable for measuring positive-sequence
current according to equation (309). This meter connection
corresponds closely to Fig. 147 for measuring positive-sequence
voltage. The response of the meter of Fig. 154 is given by
(311)
where Za, Zb, and Zc represent the impedance in the path through
the meter windings of phases a, b, and e, respectively; and
Sa, S~>, and Se represent the shunt impedances in parallel
with Za, Zb, and Zc, respectively. In this analysis the mutual
impedance between windings will be ignored because it is
assumed to be small in comparison with the series impedances
Za, Zb, and Zc. Subsequently, other analyses will be given
which definitely take into account the impedance of the meter
windings. It will be noted that equation (311) is in the forro
(312)
\. where Ka, Kb, and Kc are the coefficients of la, lb, and le in
equation (311), respectively. By companson with equation
't
l.
f
\
(DI)
Positive- SequeMee
(b)
Neg01tive -Sequel"'ce
~ --~~-----------a
--r-~~~~------b
~~~-----------~
-+--+-~~---- b
~--~~~--~~-e -+--+-~-4--~~--c
--~~~~~~~-" -+-~+-~~--~-n
1
·,
'
General RPlation
Sb Sa e+; 60o Sb = Sa e_ 160o
Zb + Sb Za + Sa Zb + Sb Za + Sa
Se = Sa e-; 60o Se = So e+i 60o
Ze +Se Za + Sa Ze +Se Za + Sa
Typical Values
for positive-sequence voltage
Fw. 154.- Positive- and negative-sequence ammeters.
.. (309) it will be seen that the meter of Fig. 154 will measure
the positive-sequence current when
in which K is numeric. Hence positive-sequence current is

measured by the meter of Fig. 154 if (1) the meter responds
equally to phases a, b, and e, individually, and (2) the current
1
of winding a is in phase with the line current of phase a, the

current of winding b lags the current of phase b by 60 deg.,
and the current of winding e leads the lin'e current of phase e
by 60 deg.
For measuring negative-sequence current it may be shown
in a similar manner from equations (310) and (312) that the
positive-sequence network may be used, provided the impedance
branches zb and sb are interchanged with those of branches
Z c and Se, or provided that the line currents from phases b
and e are interchanged.
127. Common Forro of Positive- and N egative-sequence Amme-
ters Applicable in Absence of Zero-sequence Current.
The positive- and negative-sequence ammeters of Fig. 154
have been considered, in spite of the fact that they are not the
most practica! schemes, because they use in simple form the
most general relations involved in the segregating networks.
Means for simplifying the meter and network will now be con-
sidered. In general, the current vectors of a three-phase system
possess three degrees of freedom and therefore the network
branches must possess a corresponding number. However, if the
zero-sequence current were eliminated, the network would
require only two degrees of freedom. It is possible to eliminate
the zero-sequence current before it enters the network and thus
to require the connection of the network to only two phases
of the system. The elimination of zero-sequence current is a
subject of considerable importance in sequence measurement
and for this reason it will be discussed in a subsequent section.
For the purposes of the present discussion, it will be assumed
that the zero-sequence current is non-existent as is the case
with the three-phase three-wire system. Another method of
simplifying the meter and network is to eliminate the special
meter windings by obtaining the same result through a con-
nection of the externa! impedances. These considerations
lead to the type of network illustrated in Figs. 155 and 157.
The simplest type of positive- or negative-sequence network
applicable where zero-sequence currents do not exist was pro-
posed by C. T. Allcutt and is illustrated in Fig. 155. This
form of network is frequently considered for commercial devices
and for that reason an analysis of its operation will be given,
taking into account the impedance of the meter element. By
using the notation of Fig. 155, the expression for the meter
current may be written
z zb
IM = - zb + z : + M 1c - zb + Zc + Mh (314)
In terms of sequence currents, I M becomes

¡M = _ (aZc + a2Zb)I 1 + (a 2Zc + aZb)I 2 (315)
Zb + Zc +M
The condition for the meter to measure only positive-sequence
current is that the coeffi.cient of I 2 be zero, which gives
(316)
General R elation
Typical Values
Zo = R Zc = R
Zc = R(0.500 + j0.866) z, = R(0.500 + j0.866)
e
for positive-sequence voltage ';-~
b
Fw. 155.- Positive- and negative-sequence ammeters applicable when zero-
sequence current is not present.
In a similar manner the characteristics of the network for

measuring negative-sequence current may be derived from
e quation (315) by making the coeffi.cient of I 1 equal zero w hich
gives
(317)
In the table associated with Fig. 155 are given not only the
general relations which must be satisfied for the meter to measure
positive- or negative-sequence current but also the typical values
which are commonly used.
The quantity measured by the meter element of the network
of Fig. 155 for positive- and negative-sequence connections is
obtained by substituting in equation (315) the necessary relations
for the segregation of sequence quantities indicated by equations

(316) and (317), which give for the positive-sequence current
I Ml and the negative-sequence current I M2 the following:
1 1
[Ml = - (318)
1+ M e-i30° 1+ __JI_ ei(OM- fJzb-30°)
V3Zb V3Zb
1 1
[M2 = - (319)
1+ M e+i30° 1+ M efCOM-0zb+30°)
V3Zb V3Zb
in which ()M and 8 zb are the impedance angles of the impedances
M and Zb, respectively.
Vector Diagrarn. Equation (318) for positive-sequence cur-
rent may be plotted graphically in the manner illustrated by
Fig. 156. In this case the solution is obtained by plotting ~
pass in g through the point
(1 +jO). Rotate ~ clockwise
30 deg. and divide by The V3.
reciproca! of the distance OS
S
then gives the factor by which
1al must be multiplied to
~o\.¿:,~---~:...._---~~~· obtain the amplitude of the
meter current. The angle AOD
equal to the angle AOS gives the
Fw. 156.- Vcctor diagram for angle of shift of I u1 with respect
positive-sequence ammeter of the type to 1 al, the positive-sequence
shown in Fig. 155(a). Current
through m eter lags Iat by angle AOS current in phase a. A similar
and magnitude is I at divided by length diagram may be made for nega-
of OS.
tive-sequence, but it is to be
noted particularly that the phase shifts are different.
128. Bridge Type of Positive- and N egative-sequence Ammeters.
The bridge type of network for measuring positive- and
negative-sequence currents is due to C. L. Fortescue. This
type of network, as illustrated in Fig. 157, is applicable to three-
phase three-wire systems, and, with special connections for
eliminating zero-sequence current as discussed later, it may
be used on three-phase four-wire systems. This type of network
was developed originally for the simultaneous measurement of
~-
both the positive-sequence current (current in branch M 1)

and the negative-sequence current (current in branch M 2).
In t~e analysis of this type of
network, refer to Fig. 157 for ----------------------------------------------~
the notation. A necessary ------------~~~----b
~~~~~r---~~--c
condition for the simulta-
neous measurement of both
pos i ti ve- and n e g a ti ve-
sequence currents is that the
meter im pedances be equal, so
that letting
~~ :~ Z ~ 2M}(320) Z +
General Relation
M = -a 2 (R +M) =
eiGO(R +M)
The meter current 1MI for
Fw. 157.-Bridge-type positive- and
positive-sequence and 1M2 for negativc-sequence ammeters. The
negative-sequence may be bridge-type network as shown is appli-
cable only when zel"o-sequence current is
written from inspection as not present.
follows:
1
_R+M1 _Z+M1
.Ml = Am b Am e
Expressing lb and le in terms of sequence currents, equations

(321) become
J .M1 = - ~~¡[a 2 (R +M) + a(Z +M)] - ~:[a(R +M) +

a 2 (Z +M)]
1 1 (322)
1M2 = - A~[a 2 (Z +M) + a(R +M)] - A:[a(Z +M)+
a 2 (R +M)]
Examination of "equations (322) shows that if
(Z +M) = -a2 (R +M),
Thus it will be seen that the currents in branches M1 and Mz

of the bridge-type network of Fig. 157 give in correct magnitude
and phase position the positive- and negative-sequence currents,

respectively, of phase a.
It is interesting to observe that if M2 of Fig. 157 be made
equal to 'zero, the network reduces to that of Fig. 155(a).
Vector Diagram. The bridge-type network lends itself
particularly well to exposition by means of vector diagrams
as given in Fig. 158. The line currents of Fig. 157 divide
~mi
R
~m2
(a)
(b) (e)
Fw. 158.- Vector analysis for positive- and negative-sequence brid~Z;e-type
ammeters of Fig. 157. (a) Current division of nctwork of Fig. 157. (b) Meter
currents for line currcnts of positive-sequence. (e) M eter currents for line
currents of negative-scquence.
into components in the network branches M1, M2, R, and Z as

indicated in Fig. 158(a) . Thus the current lb divides into
the components hm1 and l bm2 flowing through the branches Z
and M 1, and R and M 2, respectively. Similar!y the current
l e divides into the components l cm1 and l cm2 flowing through the
branches R and M 1, and Z and M2, respectively. Figure 158(b)
and (e) give vector diagrams for the branch currents correspond-
ing to line currents l a, h, and l e, of positive-sequence and
negative-sequence, respectively. It will be noted in both
diagrams that the components Ibml and Ibm2 are equal in magni-
tude, but they lag by 30 deg. n,nd lead by 30 deg. the resultant
current lb. Also, in bcth diagrams the compónents Icml and
Icm2 are equal in magnitude but len,d by 30 deg. and lag by 30
deg. the resultant currcnt l e. This division results from the

fact that in the network of Fig .. 157 the branches
(Z +M) = (R + M)ei 600 •
In the case of Fig. 158(b) for positive-sequence line currents,
the total current in the branch M 1 is (J ml) which is equal to
- (Ibml+ Icmi), assuming the positive sense as indicated by
the arrows in Fig. 158. It will be observed that the current
I mt is equal in mn,gnitude and in phase with the positive-sequence
current of phase a. The current in branch M2 is zero since the
componcnts Ib m2 and I cm2 exactly cancel each other.
In the case of Fig. 158(c) for negative-scquence line currents,
the analysis is similar to that givcn in connection with Fig.
158(b) for positive-sequence. In the present case 1 ml = O
and I m2 is identical in magnitude and in phase with the negative-
sequence current of phl:).,se a.
In connection with Fig. 158(b) and (e), it should be noted
that the positive- and negative-sequence currents, although
plottcd as though thcy wcre equal and in phase, may actually
h ave any magnitudc or phase position. By superposition it can
be shown that, regardlcss of the unbalance in line currents of
a three-phase three-wire system, the meter branches M1 and M2
will carry only positive- and negative-sequence currents, respec-
tively. The type of analysis just used can conveniently be
applied to check the correctness of any assumed network.
The mathematical analysis of the different schemes has been
given principally because it is more general and more readily
permits dcfining the type of network which must be used to
obtain a segregating network.
129. Elimination of Zero-sequence Current.
Thc elimination of zero-sequence current is a matter of impor-
tance not only in sequence measurements but also for relaying< 60>
and other applications. Consequently, the various connections
for this purpose will be considered by themselvcs and then a few
typical combinations with positive- (or negative-) sequence
ammeters will be described. In Fig. 154 the zero-sequence
current was prevented from actuating the moving element by

the method of connecting the windings into the network, the
element thus béing rendered unresponsive to the zero-sequence
current.
The zero-sequ~mce current may, of course, be eliminated
from load circuits supplied from current transforniers if the
secondaries are connected in delta. as shown in Fig. 159(a).
This connection of current transformers for eliminating zero-
sequence current correspondA to the line-to-line connection of
G1 ()1
1
b ~
b
1 e ~ e
n n
[iiad circuits 1 1 Load circuif6 1
(a) (b)
()1
b
()1
b
e
n
e l
n
(e)
(d)
Fw. 159. -Connections for eliminating zero-sequence current from load
circuits. (a) Delta connection. (b) Cancellation by another phase. (Genkin.)
(e) Cancellation by neutral current. (Goldsborouoh.) (d) Zigzag shunting
transformer. (W. A. L ewis.) Double-secondary-type current transformers
may be used in items (b) and (e).
potential transformers for eliminating zero-sequence voltage.

In both cases they give quantities which are functions of the
positive- and negative-sequence components in the circuit to
which the instrument transformers are connected. For certain
instrument connections line currents are required, and for these
cases the cross connection from another phase, as illustrated in
Fig. 159(b), will of course accomplish the elimination of zero-
sequence current since the line currents have equal values of
zero-sequence current. Where the total neutral current (or a
known proportion of it) is available, the zero-sequence current
from the different phases may be eliminated by the connection
of Fig. 159(c). It is to be noted for this case that the current
transformer for the neutral conductor should have a ratio so as
to give only one-third of the current of the corresponding line
conductors; this difference in ratio is required because the
neutral current contains three times the zero-sequence current
that flows -in an individual phase conductor. Another type of
transformer connection is illustrated in Fig. 159(d) and operates
on the principie of providing no path for zero-sequence current
through the load circuits but including in shunt therewith a
low-irnpedance path for zero-sequence current and a high
impedance path for positive- or negative-sequence currents.
-+----fi'ARf---Ia -=~~==--n ------~~--Q

-+--"t'fR<Ft:r---Ib -+-----+--01 -~~--~~--b
--f'~f'\d--+---- le -+---f-1~~-b --~~--~-1--C
-+--+--+----In
(b)
,.
Fw. 160.-Typical positive- and negative-sequence ammeters for four-wire
service. (a) N egative-sequence ammeter. Zero-sequence current eliminated by
delta connection. (b) Positive-sequence ammeter. Zero-sequence current
eliminated by means of return curront. (e) Positive-sequonce ammeter. Zero-
sequonce current eliminated by cross connection of two phases.
The current transforrner is of the zigzag type with three separate

cores and two equal windings on each. It should be noted that
the main current transformers are connected in star which
is the most frequently occurring case. For t his connection it is
permissible to introduce load circuits both befare and after ·the
shunt connections added to effect the elimination of zero-
sequence components from phase currents.
The use of the connections of Fig. 159 will now be considered
for several typical sequence ammeters. For exarnple, the nega-
tive-sequence network of Fig. 155(b) can be adapted for four-
wire service by using current transformer connections of Fig.
159(a) or (d) and an additional current transformer for reversing
the polarity of one current as shown in Fig. 160(a). The posi-
tive-sequence ammeter of Fig. 155 may also be operated from
the current transformer connections of Fig. 159(b) or (e), a
variation of the latter being illustrated in Fig. 160(b). The
positive- and negative-sequence bridge-type ammeters of Fig.

157 may also be operated from the current transformer connec-
tions of Fig. 159(b) or (e), the former being illustrated in Fig.
160(c).
130. Simultaneous Measurements of Sequence Currents.
In case it is desired to measure simultaneously both the
positive- and ne¿;ative-sequence currents, the bridge-type
network of Fig. 157 may be used for three-wire circuits and the
network of Fig. 160(c) for four-wire circuits, as this connection
requires the minimum number of current transformers and
network branches. If it is desircd to measure the zcro-sequence
current in addition, it is mcrely nccessary to add to sorne positive-
and negative-sequence bridge network the connections necessary
to measure zero-sequence currcnL. This may be done by mcans
of the connections of Fig. 153(!) or by inserting an ammeter in
the star connection of Fig. 159(d).
131. Frequency and Transient Errors.
The departure of frequency from normal will always cause
an error in the response of mcasuring devices for both positive-
and negative-sequence quantitics but may not cause error in
zero-sequencc devices. Considcr first the frequency errors
for the different scquence voltmcters. There will be a srnall
error in the response of thc positive-sequencc voltmeter to
positivc-sequcnce voltage of abnormal frequency. For negative-
sequence voltage of normal frequency the components of current
through the positive-sequence voltrneter just cancel, producing
zero response. Va.riation from n01rnal frcquency in either
direction will prevcnt canccllation, thus producing an error
in response. If the zero-sequence voltage is elirninated from
~he positive-sequencc volLmcter by rneans of instrument trans-
forrner connections, the zcro-sequcnce voltage during departures
from normal frequency cannot affect the meter reading. Similar
remarks apply to negative··sequence. The zero-sequence volt-
meter will not respond to positivc- or negative-sequencc voltages
provided the zero-sequence voltage is obtained by instrument
transformer connections.
The character of the frequency error can conveniently be
analyzed by considering the cffects in connection wi th a typical
instrument, such as t he positive-sequence voltmeter of Fig.
148(a). For ihe purpose of this study the meter impedance

may be ignored and the network assumed to consist of
ZB = R
z. ~ ~(1 + j~!) (324)
where f is the system frequency whose normal value is 60 cycles.

The response of the positive-sequence voltmeter will be obtained
by substituting the above impedance values in equation (303)
which gives for the meter currents the following relations for
positive-sequence voltage:
(325)
and for negative-sequence voltage:
(326)
Examination of equation (325) will show that a 5 per cent

variation in the system frequency of the positive-sequence
voltage will cause the current in the positive-sequencevoltmeter
to be in error by 1.1 per cent. There is also a shift in the phase
of the voltage applied to the voltmeter which amounts to about
one degree lag or lead for system frequencies of 105 and 95 per
cent normal, respectbely.
The positi ve-sequence voltmeter does not respond to negative-
sequence voltage of normal frequency, 60 cycles in this case,
since the components of the meter current just cancel each 0ther.
However, when the frequency is either 5 per cent above or below
normal, the meter will mcasure about 5 per cent of the negative-
sequence voltage on the system. The total error is of course the
combination of the errors dueto positive- and negative-sequence
voltage responses. Thus on a normal system, the negative-
sequence voltage is usually relatively small and the total error
will be only of the order of 1 per cent even for a 5 per cent varia-
tion in frequency from normal. However, in the case of the
negative-sequence vol'tmeter, the frequency errors may be of
importance because of the fact that the negatlve-sequence

voltage is normally zero. Thus a 5 per cent variation in fre-
quency from normal will cause a normally high positive-sequence
voltage to produce a response in the negative-sequence voltmeter
corresponding to about 5 per cent of normal. It should be
clearly understood, however, that variation from normal fre-
quency is not a matter of much importance at the present time
because power-system frequencies are closely regulated in order
to provide time-keeping service.
Another type of error that may arise is due to electrical
transients which include (1) the meter response to the direct-
current components of transients in voltages or currents in the
measured circuits, and (2) the lag in inductive branches of the
segregating network in following changes in the measured circuit
quantities.
The meter response to the direct-current component is of brief
duration and for many purposes it may be ignored. This error
may be minimized, if dcsired, by shunting the meter element by
a low-resistance, high-inductance branch, thus by-passing the
direct-current component but not the alternating-current com-
ponent of the transient current. There is also. the error due to
the time lag of current building up in the inductive branch of a
sequence ammeter network, but fortunately this error is very
small and will reduce to about 2. 7 per cent in 1 cycle and less
than 0.1 per cent in 2 cycles on a 60-cycle systcm. Thus the
error is entirely negligible for ordinary instruments but may
require consideration in the interpretation of oscillographic
records of sequence voltages during the first half cycle. There
is also a similar error in connection with the sequence voltmeters,
but this error is of much less magnitude and is entirely negligible.
The presence of harmonics on a system may constitute another
possible source of error in the response of sequence instruments.
Thts problem of course can be considered as merely one. of an
excessive variation in frequency from normal. Thus by con-
sidering equation (326), it can be shown that a 10 per cent fifth
harmonic, which is of negative-sequence, will cause a positive-
sequence meter response of about 5 per cent of the normal meter
reading. Thus while it appears that the error in meter response
may be quite large, the magnitudes of the harmonics fortunately
are usually small so that the resulting error can be ignored. If
desired, a low-resistance high-inductance branch may be placed
in series with the meter element and thus reduce the effect of
these harmonics to a negligible value.
In concluding this discussion of the errors of the sequence
measuring devices, it should perhaps be pointed out that while
there is a theorctical possibility of these errors, they have not,
as a mattcr of fact, been of much practica! importance so that
it has not been nccessary, except in a few isolated instances, to
consider means for minimizing these errors.
132. Adjustment and Testing of Segregating N etworks.
The adjustment of segregating networks is best made by
determining accurately the impedances including the reactance
and particularly the alternating-current resistance of the network
elements. The procedure is most readily followed by considering
a typical network, such as that of the positive-sequence voltmeter
of Fig. 148(a). The value of the reactance of the branch ZA
should first be obtaincd and then its resistance adjusted to give
the power-factor angle of 60 deg., and finally the resistance
of the branch Zn should be adjusted to have the same impedance
as the branch ZA. These data are sufficient to establish a
segregating nctwork. However, as . a practica! matter, the
impedances of thc various elements are approximately determined
and the final adjustmcnts are made on the combination. A
given voltage is impressed between terminals zl and z2 in parallel
with z3 and then between terminals l3 and z2 in parallel with zl,
and the network adjusted to give equal deflections of the volt-
metcr. The transformer lcads connected to the e phase are
then separated so that the same voltage as before can· be
epplied to both transformcrs in parallal. The meter reading
should then be V3 times the former deflection. The final
check is to connect the voltmeter and network to a balanced
three-phase supply. If the network is correctly adjusted, the
meter will read the line-to-neutral voltage correctly for one phase
rotation and zero voltage if two of the leads are interchanged.
For the adjustment of positive- or negative-sequence current
networks a similar procedure may be followed. For the network
1
r of Fig: 155(a) the steps are almost identical with those given
!t. above for the positive-sequence voltmeter. In the case of the
current networks, it may be necessary to disconnect any current
transformer whose primary sido is open-circuited. So long as
the voltage across the transformer is small so that saturation
·-
does not occur this is unnecessary, but for larger voltages satura-
tion will decrease the exciting impedance so that appreciable
current is by-passed.
The effect of the impedance of the measuring erement of
the voltmeter or ammeter on the adjustment of the segregating
network will now be considered. Examination of the equations
for the various networks for the voltmeters show that they will
respond only to voltages of the desired sequence without placing
any restriction on the meter impedance. Thus the voltages
impressed on the positive-sequence voltmeter elements of Figs.
147 to 155 inclusive, are affected in both magnitude and phase
relation by the meter impedances, but they do not affect the
kind of quantity to which the meter responds, which is positive-
sequence voltage only. This statement also applies to the
sequence ammeters cxcept for the case of thc bridge-type net-
works when used to measure both positive- and negative-sequence
currents. For this case the impedance of one meter is a part
of the segregating network for the other meter. Thus for
a network of t his typc, such as Fig. 1.~7, it is necessary for the
two meter branches to be equal and that the meter impedances
and the other branches bear the relation
Z +M = (R + M)ei 600
• (327)
133. Application of Sequence Voltage and Current Devices.
The sequence voltage and current instruments and rclays are
applicable for a wide variety of uses. Several of these devices
have been made available in commercial form by the Westing-
house Electric and Manufacturing Company who pioneered
in this development. They possess the advantagc of measuring
directly the quantity used in the analysis of power systems under
unbalanced conditions. Sequence instruments required for
particular purposes will usually be less in number and never
in excess of the required number of single-phase instruments
to give t he complete indication of system performance. In
addition to the advantage in the smaller number of devices,
the sequence instruments possess the very real advantage of
giving the data in a more practica! form, which is required
for many applications. "When single-phase instruments are
used, the data for the various uses can be obtained only by making
calculations from the combined readings, as for example by the
methods outlined in the previous chapter.
The application of the sequence instruments will first be
reviewed in a general way and this will be followed by a more
detailed description of three of the more common applications.
Knowlcdge of system voltage is of importance for many applica-
tions, as, for example, the tendency of synchronous motors
to pull out of step on the occurrence of a dip in the system
voltage. For this purpose a positive-sequence voltmeter is
particularly convenient. Hcnce a single meter of the sequence
segregating type will give more satisfactory results than three
instruments of the conventional single-phase type. Another
illustration is the use of the negative-sequence · ammeter to
determine the amount of unbalance arising from the single-phase
applications of certain loads, such as those due to electric fur-
naces. For this purpose a single negative-sequence ammeter
will give more satisfactory results than three single-phase
instruments.
A further advantage of the sequence measuring devices of
the negative- and zero-sequence types arises from the fact
that these instruments usually carry no current when the system
is in normal condition but have a positive indication under
abnormal conditions; consequently, they provide an ideal basis
for relay operation since more sensitive and faster operation
can usually be secured from relays which have to distinguish
betwccn a zcro anda finite value instead of between two differences
in magnitude.
134. Positive-sequence Graphic Voltmeter.
The positive-sequence voltage, as pointed out previously,
is frequcntly required for analysis of power-system operation
particularly in respect to the effect of faults on systems reducing
the system voltage and tending to pull synchronous motors
out of step. For this purpose the positive-sequence recording
voltmeter is being used. The meter element itself is of the
conventional type used for single-phase purposes. The differ-
ence, however, is in the externa! network which consists of the
sequence segregating network of Fig. 148(a). The network,
consisting of merely reactor and resistor elements, is extremely
simple and will not occupy any considerable space. A typical
network built by the Westinghouse Electric and Manufacturing
Company is illustrated in Fig. 161. As a matter of interest a
standard resistor for a recording wattmeter is shown alongside
for comparison. In connection with the positive-sequence

voltmeter it should perhaps be pointed out that two potential
transformers are required when connectcd line-to-line, or three
when connected line-to-neutral. To obtain equivalent data
it is necessary to use the same number of instrument transformers
with sequence-segregating or single-phase devices but the
number of meters will be a mínimum if sequence-segregating
devices are used.
,r., --7'~:.:-.-·:··; :,· ._,• -~~ ~" ~- •:- -- ~~ ··;~~"·-·.: :. ·.

1 .
;
' h j
---·- ---- ~--..J
(b)
Fw. 161.- Relative size of network and standard meter external resistor.
(a) Standard rcsistor for rccording wattmeter; (b) positivc-sequence network
for recording voltmetcr.
135. Positive-sequence Voltage Regulator.

It is a matter of operating experience that the method of
connecting the voltage-regulator operating coil to the machine
terminals through a single potential transformer does not always
result in the proper action of the voltage regulator due to the
fact that the regulator responds to what is happcning on that
particular phase only. The voltage on the unloaded phases of
three-phase machines supplying a single-phase railway load will
be appreciably above the loaded phase. Similar conditions
exist on a power system at times of unbalanced faults . There-
fore, since the throwing of a single-phase load or a single-phase
fault on to one phase may result, under certain conditions, in a
rise rather than a drop in voltage on one of the other phases, it
frequently happens that a regulator will cause the exciter voltage
to build down when it should build up, if the potential trans-
former happens to be connected to one of the unloaded phases.
An example of this rise in voltage is given in Fig. 162 which
shows an oscillogram taken on a 120-kva., 1,100-volt, three-phase,
60-cycle generator. Fullload at unity power factor was thrown

on one phase and the current on the loaded phase, and the
voltage between all three terminals recorded. The voltage
AB actually rises about 6.5 per cent when the single-phase
load is thrown on phase CA. In the oscillogram, voltage AB
stays above normal for a period of about 15 cycles; its fall
~~~~~
Volts AC
VVVV
1 1, \'25 VAB
VAs Vac VeA
100
~ VBc,VcA
1--1-+ =§ 15
e::>
r - ~-~g 50
~,..·-
1--t---:.<11 E25
0...!-
{!:. o
-4 -'2 o '2 4 6 8 10 12 14 16 16 20 22
Time of Lo01ol Appl ic01tion
Fw. 162.- 0scillogram showing rise in voltage on phase AB whcn a resis tance
load is applied to phase AC. The curves at the bottom give the r.m.s. voltages
correctcd for drop in generator speed.
from the maximum vn.lue is due principally to the decrease in

speed of the prime mover in supplying the additional load.
In this case it is obvious that with a voltage regulator connected
across terminals A and B, t he exciter voltage would have been
reduced whereas it was really required to build up since t he
load was demagnetizing the main machine. A positive-sequence
network, however, supplying voltage to the regulator coil would
have caused it to respond correctly. In addition the positive-
sequence voltage regulator gives the same indication irrespective

of the phase to which the unbalanced load or fault is applied,
which is obviously much more satisfactory than the condition
with the single-phase voltage connection. The need for the
positive-sequence voltage regulator was recognized about the
same time as quick-response excitation was introduced. It is
obvious that if excitation is to be changed in a very short
space of time, it is exc€edingly important that the change be
made in the desired direction. For applications where slow-
response exciters are used, the positive-sequence voltage regulator
is not ordinarily used since the circuit-breaker will generally
isolate the fault before there has been much opportunity for the
exciter voltage to change. Of course, if a system were supplying
heavy single-phase loads on different phases, a positive-sequence
voltage regulator would provide the best means of dctcrmining
the proper mean value of machine excitation.
In connection with the application of the positive-sequence
voltage regulator there is one special problem introduced owing
to the fact that with the vibrating type of regulator it may be
desirable alternately to increase and decrease the voltage applied
to the regulator coil in order to introduce anti-hunting features.
This was accomplished in t he regulators operating on a single-
phase voltage by short-circuiting a resistance in series with the
regulator coil. This same principie can be introduced with
the positive-sequence voltage network of the type of Fig. 148(a).
As pointed out before, the meter bra nch of the network including
the meter element and externa! resistance can be varied without
affecting the kind of voltage impressed on the coil, i.e., the
adjustment of the externa! resistance will still leave positive-
sequence voJtage im pressed on t he coil.
136. N egative-sequence Overcurrent Relay.
The negative-sequence current network of Fig. 155(b) or 157
may be used in connection wit h a relay instead of an ammeter.
Such a negative-sequence current relay may be used to provide
protection against single-phase operation and phase reversal.
Its principal application to date has been for protecting induction
motors against single-phase operation, the excessive heating
of which is discussed in Chap. XVII on Induction Motors.
A commercial form of the negative-sequence relay which makes ·
use of the bridge-type network is the type CQ built by the
Westinghouse Electric and Manufacturing Company. Sequence
relays involve one feature which requires consideration, namely,
the provision for changing tap
settings. Sueh eh a n g es in
taps ordinaril'y alter thc
meter impedance, and this,
in turn, will affect the current
fiowing through the meter
element since all of the cur-
rent networks have shunt
paths in parallel with the
meter element. In case net-
works of the type of Fig.
155(b) are used, a change in
the meter impedance affects FIG. 163.- Negative-sequence relay.
the magnitude and phase
angle of the current fiowing through the opcrating coil of the
relay, but it responds only to negativc-sequence current. In the
case of the negative-sequence
bridge-type network of Fig.
157, a change in the meter
impedance alters not only the
response to negative-sequence
current but also permits the
fiow of positive-sequence cur-
rent through the relay element.
Hence it is essential that the
J
j
relay give constant impedance
' for the full range of tap set-
. tings. This is accomplished in
, the type CQ relay, a diagram
1of which is shown in Fig. 163,
] by introducing a dummy coil
in branch M 1 equivalent in
impedance to that of the relay
element M 2 · The connections
of Fig. 163 were also selected so
that, if desired, a relay element
Fw. 164.- Negative-sequence
over- could be substituted for the
currcnt relay.
dummy coil and thus permit
measurement of the positive-sequence current in one branch and
negative-sequence current in another branch of the network. The

constant impedance characteristic of the type CQ* relay is accom-
plished by shifting contacts which vary the amount of secondary
load on the relay in such a manner asto secure constant imped-
ance independent of relay setting. The physical appearance
of the type CQ phase-balance and reverse phase-current relay
is shown in Fig. 164.
137. Polyphase Sequence N etworks and Instruments.

The sequence segregating voltage and current devices that
have been described have been of the single-phase type, and it
------------------------------~----a
--------------------~--------~-----b
--------~----------+---------~-----e
--~-----+----~----+-----~--~-----"
Fw. 165.- Typical polyphase sequenco voltage segregating networks together

with positive-, n egative-, and zero-sequence voltmeters.
seems desirable to describe a few typical pol)Tphase devices.

The polyphase sequence segregating network can readily be made
by combining thrce single-phase networks with the networks
connected to their different phases in cyclic order. The poly-
phase relay may be of the induction-motor or synchronous type;
in either event there would be a distributed polyphase winding.
In general, the connections for three single-phase devices operat-
ing from different phases will require a few additional network
elements in excess of the mínimum required for polyphase
* Further information on the detailed characteristics of the negative-
sequence overcurrent relay are given by J. V. Breisky.( 21 )
service; hence, several typical polyphase sequen ce segregating

voltage and current devices will be described.
A typical connection for measuring polyphase positive- and
negative-sequence voltages is shown in Fig. 165. In connection
with this scheme it should perhaps be pointed out that zero-
sequence current may fiow through both the positive- and
F10. 166.- Simplified polyphase positive-sequence voltmeter.
F10. 167.- Typical polyphase sequence current segregating network together

with positive-, negative-, and zero-sequence ammeters.
negative-sequence polyphase meters. However, since zero-

sequence currents fiow equally in each phase of the distributed
polyphase winding there will be no tendency to operate the
meter. The particular set of connections of Fig. 165 incidentally
::;hows t he meter for measuring zero-sequence voltage. H
only the polyphase positive-sequence voltage is desired, and
line-to-line potential transformer8 are available, then the con-
nections of Fig. 166 may be used. In connection with the

network of Figs. 165 and 166 it may be observed that the positions
of Z and R may be interchanged, or the phase connections of
the potential transformers may be reversed in order to cause a
positive-sequence meter to measure negative-sequence voltage
and vice versa.
R Z
Fw. 168.-Simplified polyphasc ncgativc-sequence amm eter.
----------~------~~~~==~a
(o)
~ ~;
~'
(b)
w, w, M2
Fw. 169.- Polyphase sequencc ammetcr using bridge-typo n etworks. (a)
Schematic connection; (b) meter winding arrangcment.
A typical set of polyphase sequence segregating ammeter

connections is shown in Fig 167. Positive-, negative-, and
zero-s~quence currents are measured by the instruments marked
! 1, [ 2 , and ! 0 , respectively. Here again zero-sequence current
circulates through the polyphase metcrs but produces no resultant
MEASUREMENT OF SEQUENCE VOL'l'AGES 305
effect. Figure 168 shows a simplified form of negative-sequence

ammeter. In this case the zero-sequence current is eliminated
by means of the delta connection of current transformers. The
bridge-type networks of Fig. 157 may be used as a basis of the
polyphase sequence meters of Fig. 169. In this case six terminals
are required for each polyphase meter winding as illustrated in
(b); the connections for the individual phases are shown in the
schematic diagram of (a).
Problems
1. A positive-sequence voltmeter whose meter lmpedance is 500 ohms and
whose resistance branch of the segregating network is 2,000 ohms has 100
per cent voltage applied to the primary of each transformer singly. What
are the meter readings for thc two cases?
2. Prove that the connections of Fig. 151 (d) measure positive-sequence
voltage.
3. In a certain oscillograph test it was desired to measure the positive-
sequcnce current. Only two current transformers and a positive-sequence
bridge-typc network wcre availahlc, but since the system was grounded this
conncction did not eliminate the zero-sequence current. However, a set
of three curren t transformers giving 3/ o for re lay operation constituted part
of the installation and, while these could not be disturbed for cross-connec-
tion purposes, a resistance was inserted in this circuit so that the drop across
it could be used to buck down the zero-sequence drop in the bridge-type
network. What should be the connection and the relation between the
resistance of the oscillograph shunt and the resistance inserted in the relay
circuit?
4. A certain polyphasc device requires for its operation a positive-sequence
set of currents of 5 amp. and only a single-phase source is available. The
impedance of the device per phase is 0.02 + jO.Ol ohm. The different phase
currents are obtained by connecting the three windings in series ar_d shunting
the individual windings by the impedances Za, Zb, and Zc. If Za be equal to
0.025 ohm, what must be the values of Zb and Zc? Advantage may be taken
of the fact that since the windings are insulated from each other, the effect
of a negative resistance may be obtained · by reversing the terminals of
a winding.
5. A positive-sequcnce voltmetcr of the type shown in Fig. 148(a) has a
meter element of 400 + j80 ohms, and R = 5,000 phms. By rewinding the
element with three times the turns of finer wire, the element impedance is
3,600 + j720 ohms. Calculate the value of R to obtain the same sensitivity
6. In Prob. 5, compare the currents in the segregating elements for 50 volts
negative-sequence applied to both the original voltmeter and the rewound
combination.
CHAPTER XV
THE MEASUREMENT OF POWER QUANTITIES ON

POLYPHASE CIRCUITS
This chapter reviews t he measurement of power quantities

on single-phase and polyphase circuits, the latter including
both balanced and unbalanced conditions. Conventional watt-
meter, reactive volt-ampere meter, and power-factor meter
connections for three-phase circuits are examined with respect
to both balanced and unbalanced conditions. It is shown that a
number of the conventional connections with two or more
elements give incorrect indications with unbalanced currents
and voltages though giving correct indications with balanced
currents and voltages. The discussion continues with an analysis
of the sequence power quantities including positive-, negative-,
and zero-sequence wattmeters, reactive volt-ampere meters,
and power-factor meters. The chapter concludes with a dis-
cussion of the flow of the various power quantities on unbalanced
circuits with various kinds of loads.
138. Power Expressed in Terrns of Voltage and Current Vectors.

Power in an alternating-current circuit is at any instant
equal to t he product of the instantaneous values of current and
voltage. In Chap. II, equation (16), it was shown that the
instantaneous value of a sinusoidally varying voltage is
The corresponding expression for the instantaneous value of

current is
i = v2(Jé»t + Jciwt) (328)

2
The instantaneous value of power is then
ei = !(El+ EfEiZwt) +!(El+ EiciZwt) (329)
306
MEASUREMENT OF POWER QUANTITIES 307
The two parts on the right-hand side are conjugate to each
A ,..._ '
other. El and El are conJugate vectors of equal magnitude
independent of time, lyi.ng
in opposite quadrants. Q
/
_, , - - - -- -
-..--...... ......
.......
The do u ble-frequency / ''
rotational vectors E1Ei 2wt /
/
', \
and EÍE- i 2wt are also con- \
1 \
jugate to each other and 1 \
(
have the same absolute
values as EÍ and El. 1 /
'
/ ....... ------
Thus it will be seen that \ /
\ /
1
the expression for instan-
taneous power consists of
a constant term with a
1
sinusoidal term of double J
1
fre:quency superposed. 1
These relations are illus- 1
trated in Fig. 170. \ /
'
I t can be seen that the
average valucs of the
'
' .......... _ .........
./ /
/
/
/
---~'
double-frequency compo- Fw. 170.-Powcr diagram showing double-
nents are zero so that the frequency rotating vectors.
mean value of power is
p = !(EÍ +El) (330)
Now if E = Eéa and 1 = JEifJ then
'
P = ![EJEi(a- {J) + EJE-i(a-fJ)]
= ~i [ {cos (a - {J) +j sin (a - .B) }

+ {cos(a- .B) -jsin(a -.B)}] =Eicos(a-.B) (331)
This is the familiar form of expression that is commonly used.
It will be observed that the real parts of either EÍ or El are
equal to this value and from this point of view, having once
defined P as the real part of El or El, either expression can be
used.
139. Reactive Volt-amperes Expressed in Terms of Voltage and

Current Vectors.
For recurrent phenomena, the reactive volt-amperes consumed
by a load are equal to the product of the voltage and the com-
,
,·
ponent of current lagging 90 deg. behind the voltage. Starting

with the definition of reactive volt-amperes as equal to
Ei sin (a - {3), it is possible to show that the imaginary parts
of the above expressions EI and El are equal in magnitude t.o
the reactive volt-amperes. Since EI and El are conjugates
of each other, their imaginary parts are t he negative of each
other. The choice of the expression to use is dependent upon
the conventional sign associated with inductive or capacitive,
lagging or leading reactive volt-amperes. The authors of the
earlier pape~s presented before the American Institute of Electri-
cal Engineers chose to assume inductive reactive volt-amperes
as positive. The same convention was adopted by C. L. For-
tescue in presenting for the first time a vector expression <5l
for power quantities. Others have chosen the opposite conven-
tion. A search of the literature as to the use of the sign asso-
ciated wit h inductive reactive volt-amperes was recently made
by the authors. This search revealed a preponderance of
the use of the positive sign. for this purpose. Most of the refer-
ences cited used reactive volt-amperes only incidentally
in the development of sorne other main subject. Of even greater
significance than mere numerical superiority is the fact that
of those authors such as Kennelly, Jackson, Fortescue, and
Slepian, whose subject matter dealt primarily with reactive
volt-amperes, all preferred the use of the positive sign. In the
absence of standardization, the authors have followed the
conclusions resulting from this search in considering inductivP-
reactive volt-amperes as positive. With the adoption of this
convention real and reactive volt-amperes will have signs
corresponding to (R + jX)i 2 •
The drop across an inductive reactance X is equal to jXl
and using the expression EI, Q, the reactive volt-amperes,
is equal to (jXI)(l) or jXi 2 • On the other hand, for El, Q
is equal to ( -jXI )(l) or -jXi 2 • From the foregoing it is
evident that for the sign of Q to agree with the convcntion
adopted, El must be chosen to express power and reactive
volt-amperes, so that
P +jQ =El
140. Power Quantities on Single-phase Circuits.
The power quantities on single-phase circuits as used in
this volume consist of the power quantity P; the reactive volt-
ampere quantity Q, which is sometimes referred to as reactive
power; and power factor PF. These power quantities have
physical significance in a ccordance with common usage. Thus
the power P indicates the rate of consuming energy in a circuit
and is commonly measured in watts or kilowatts. The reactive
volt-amperes Q consumed by a load is commonly measured in
volt-amperes or kilovolt-amperes, usually abbreviated "vars"
or "kilovars." Power factor is defined in terms of the power
quantities used in this discussion as follows:
(333)
141. Power Quantities on Polyphase Circuits.

The amount of power which must be supplied to polyphase
circuits is, of course, the sum of the power requirements for
the several phases. Thus for a three-phase system with phases
a, b, ande, the total power
...
(334)
The total reactive volt-amperes Qr are, of course, equal to the
sum of the reactive volt-ampere requirements of the several
phases. Thus for the three-phase system
Qr = Qa + Qb + Qe (335)
It follows, therefore, that
Pr + jQ1' = (Pa + Pb +Pe) + j(Qa + Qb + Qc)
= (Pa + }Qa) + (Pb + }Qb) + (Pe+ jQc)
= Eala + Eblb + EcÍc (336)
On a ba.lanced three-phase system, power factor has the same
significance as on a single-phase system. Similarly, total power
factor under unbalanced conditions will be defined in terms of
Pr and Qr in a manner identical with that for the single-phase
circuit. Thus the power factor for a three-phase system even
under unbalanced conditions * is
PFT =
Pr (Pa + Pb +Pe)
(337)
-v
--¡:P;:::=r=;:=2 =+==:::=
Q==-
r2 - V (P a + p b + p cF + (Qa + Qb + Qe)2
, * A.I.E.¡?. Standards, 1922 ed ., §3243.
These relations for total power, reactive volt-amperes, and

power factor, in terms of the quantities for the individual phases,
may best be clarified by reference
... to Fig. 171.
1
1
1
1
142. Measuring Devices for Single-
phase Power Quantities.
Q Measuring devices for single-
phase power quantities include t he
wattmeter, the reactive volt-ampere
meter, the single-phase power-
factor meter, and are indicated
-::
0- f L - - - - = - - - ' - - - - -----l._ P schematically in Fig. 172Jfor con-
- - - - - - - - - PT ______ _,... ventional indicating instruments.
FIG. 171.-Power quantities on an
The watt-meter shown in Fig.
unbalanced three-phase circuit.
172(a) has a rotating part which
operates against a spring and a potential coil with external
resistance to cause the current through the coil to be in phase
with the voltage of the circuit to which it is connected. The
reactive volt-ampere meter shown in Fig. 172(b) is quite similar
but has a reactance in series with the voltage coil which causes
the current through it to lag 90 deg. behind the voltage of the
"'- . : Vo/fcrge
cod.s
'--+-+--''Vo 1fage '-+- t - "Volfage

L--_____J coi! .___, coi!
(01) (b) (e)

Fw. 172.-Electrodynamic instrumen ts for indicating single-ph asc power
quantities. (a) Wattmeter. (b) Reactive volt-ampere meter. (e) Power-factor
meter (or reactive-factor meter whcn suitably calibrated).
circuit to which it is connected. A conventional form of the

single-phase power-factor meter or reactive-factor meter is
illustrated in (e). Such a meter may be calibrated toread power
factor, reactive factor of the circuit, or the difference in phase
angle between voltage and current. It may be pointed out that
the power-factor and reactive-factor meter differ from the
wattmeter and reactive volt-ampere meter of (a) and (b) in
that the latter operate against a spring whereas the former
take a position depending upon the relative phase of voltage
and current. In general, the reactive-factor meter is used
for determining the excitation of machines because its scale
is not so crowded for the usual high power-factor operating range.
Because of the similarity of the action of the reactive-factor
• 1
and power-factor meters the subsequent discussion will be
restricted to power-factor meters since that is the more com-
monly used factor.
1m
P hase-angle ~~~---0
C o r r e e ti o n . In '"m -t-j Xm Re -jXc
-+--- ------ --- -- E ----- ------- ->-
connection w i t h
the simple circui t
shown in Fig.
172(a) and (b), it
m ay be o bserved
that the current
through the volt-
age coil should
ha ve phase angles
of exactly O deg.
and 90 deg., re-
spectively, from -<-------- --------- -- E -------- ___________ .,..
the phase of the (b)
~---- E ---- ~
vol tage of the cir- ~m Re
cuit to which the o ~~-~----~----~o
rm+jXm lt
instrument is ----+15
con nected. This Rs '+JXs
relation is approxi-
ma tely realized by
the use of a non-
inductive external Ie (e)
Fw. 173. -Mothods for correcting phase angle in poten-
resistor Re and an tia! circuits of wattmeters.
external ~eactance
coil X c. However, due to the fact that the voltage coil will have
sorne reactance, and the reactance coil Xe will have sorne resist-
ance, this solution remains approximate.
Full control of the phase angle of the current in the voltage
coil of wattmeters may be obtained in several ways as shown
in Fig. 173. In this figure, (a) shows a series capacitor whose
impedance is just equal to that of the reactance of the meter ele-
ment. In this manner the effect of meter reactance is annulled.
The scheme shown in Fig. 173(b) utilizes a capacitar in shunt

with all ora part of the external resistance. The vector diagram
is self-explanatory in showing the shift of the various vector
quantities which result in the meter current I m being in phase
with the voltage across the potential circuit. Condensers
cannot conveniently be divided into small divisions. The
method of Fig. 173 (b) i~ therefore of particular valu e in tha t
the small corrections in phase angle may be accomplished by
adjusting the value of the externa! resistance shunted by the
condenser. Figure 173(c) shows a third scheme which consists
of shunting the meter element by a highly inductive shunt.
One method of correcting the phase angle of the potential
coil of a reactive volt-ampere meter is shown in Fig. 174. This
consists of shunting the
meter element with a
condenser. The vector
diagram illustrates the
I
relation between the va-
rious vector quantities.
The phase-angle cor-
rections, the possibility
of which has been
demonstrated in connec-
tion with Figs. 173 and
174, will be assumed in
the subsequent discus-
Em sion, and the shunting
Fw. 174.-Method for correcting phase angle devices will not be shown
in potential circuits of r eactive volt-ampere . th · · •t
meters. 1n e ma1n ClrCUl con-
nections since such pro-
cedure obscures the essential phenomena.
143. Total Power Quantities in Terms of Sequence Qpantities.

It has been pointed out in Sec. 141 that the total power
quantities on a three-phase system may be expressed in terms
of the phase quantities as follows:
Each of the above three expressions may be expanded in

terms of their sequence voltages and currents. For term Eaia,
EJa = (Eo + E1 + Ez)cJo + ]¡ + Iz)

= Eoio + Eoi1 + Eoi2
+ EJo + E1I1 + EJz
+ Ezio + Ezi1 + Ezfz
The two other terms Eblb and EJe can be expanded in a
similar manner. The results of these expansions are shown in
tabular form in Table XVI, the tabulated values representing
v TABI..E XVI.- ExPANSION OF SINGI..E-PHASE PowER ExPRESSION IN TERMS

OF SEQUENCE CoMPONENTS TO Snow TnAT ToTAL PowER Is EQuAL
TO 3(EJ¡ + E 2f 2 + Eof o)
Meter quantity Etft EJ2 i E 1f o 1 E2lt 1 E2f 21 E 2l o 1Eoft 1 Eof21 Eolo
Eafa ... . ......... 1 1 1 1 1 1 1 1 1
Eb!b.............. 1 a a2 a2 1 a a a2 1
Ecfc . . .. .. ... .... . 1 a2 a a 1 a2 a2 a 1
Total. .......... 3 o o o 3 o o o 3
the coefficients of the quantities given at the top of any column.

Adding the coefficients in the respective columns, it will be
noted that the coefficients of six of the products disappear leaving
the total power quantity in terms of the sequence voltages and
currents
Pr + }Qr = 3(Elll + E zfz + Eoio) (338)
Examination of the above equation suggests that the total
power quantities are made up of the positive-, negative-, and
zero-sequence components in each of the three phases. Sequence
power quantities may therefore be de:fined in a manner similar
to that for the quantities of single-phase circuits.
The positive-sequence power quantities per phase are
P1 + jQ1 = E1I1 (339)
The negative-sequence power quantities per phase are
Pz + jQz = Ezfz (340)
The zcro-sequence power quantities per phase are
Po + jQo = E oio (341)
Thus the total power quantities on the three phases may also
be expressed as
Pr + jQT = 3[(Pl + P2 +Po) + j(Q¡ + Qz + Qn)] (342)
144. Power Quantities on Unbalanced Three-phase Circuits

Analyzed by Sequence Components.
There are a number of conventional wattmeter and reactive
volt-ampere meter connections which are in use today. Sorne of
these are correct for balanced or unbalanccd conditions, while
others give correct results under balanced conditions but incorrect
results under unbalanccd conditions. In examining these
connections for measuring power quantities it is convenient to
TABLE XVII.-CoEFFICIENTS OF SEQUENcE PRonuoTs OF VoLTAGE AND

CoNJUGATE OF CuRRENT FOR VARIOus PHASE VoLTAGE AND
CuRRENT CoMBIN ATIONS
Meter Scquencc voltage-currcnt products

Item quan-
tity E¡[¡ E1!2 E Jo E2!1 E2l2 E]o 1 EofJ 1 Euf2 1 Eofo
1 1 1 1 1
1 E ala 1 1 1 1 1 1 1 1 1
2 Eafb a a2 1 a a2 1 a a2 1
3 Eafc a2 a 1 a2 a 1 a2 a 1
4 EJa a2 a2 a2 a a a 1 . 1 1
5 Eb!b 1 a a2 a2 1 a a a2 1
6 Ebfc a 1 a2 1 a2 a a2 a 1
7 EJa a a a a2 a2 a2 1 1 1
8 EJb a2 1 a 1 a a2 a a2 1
9 EJe 1 a2 a a 1 a2 a2 a 1
10 EA!a a - a 2 a -a 2 a - a 2 a 2 -a a 2- a a 2 - a o o o
11 EA!b a 2 -1 1-a a-a 2 1-a2 a-1 a 2-a o o o
12 EA!. 1 - a a 2 - 1 a - a 2 a-1 1 - a 2 a 2- a o o o
13 EB!a l -a 1 - a 1 -a 1 - a 2 1 - a 2 1-a2 o o o
14 EB!b a - a 2 a 2 - 1 1-a a-1 a 2 - a 1-a2 o o o
15 EB!. a 2 -1 a-a 2 1 - a a 2-a a-l 1-a2 o o o
16 Ecfa a 2 - l a 2 -1 a 2 - 1 a-1 a - 1 a-1 o o o
17 Ec!b 1-a a - a 2 a 2 - 1 a 2-a 1 - a 2 a-1 o o o
18 E ef e a-a 2 1 - a a 2 - l 1-a2 a 2- a a-1 o o o
make use of the sequence components. For this purpose the

method, illustrated in Table XVI, has been extended in Table
XVII to give the coefficients of the nin~ products of the sequence
components for the 18 possible products of a line current and a
line-to-neutral or a line-to-line voltage. This table may be
used for determining either the real or reactive component
of each of the nine sequence power products which may be
measured by a meter element with the understanding that the
current through the voltage coil must be in phase with the circuit
voltage in arder to measure the power component or to lag
90 deg. behind the circuit voltage in arder to measure the reactive
volt-ampere component.
It will be noted that for the three-phase four-wire system
the total power quantities may jnclude three sequence com-
ponents, positivo-, negative-, and zero-sequence; consequently,
three meter elements are required to measure the total power.
These measurements may be made either in terms of the three
sequence components in line with equation (338) or in terms 9f
the three phase components, equation (336), or with various
othcr combinations. On the three-phase three-wire system,
the total power quantities will include only two sequence com-
ponents, namely, the positive- and negative-sequence com-
ponents. Thus it is clear that the total power quantities may
be measured by means of two wattmeter-type elements which
t.' may be either of the single-phase or of the sequence types. In
case of symmetrical systems with balanced loads, only the posi-
tive-sequence power quantities exist and these mq,y be measured
by one meter of either the single-phase or positive-sequence
type.
145. Analysis of Meter Connections for Power Quantities on
Three-phase Circuits.
In this discussion the meter reading will be designated by
the symbol PM + jQM, in which PM represents the value obtained
when wattmeters are utilized to make the measurements and
QM represents the value obtained when reactive volt-ampere
meters are utilized to make the measurements. As stated
previously, reactive volt-ampere meters differ from wattmeters
only in that reactance instead of resistance is inserted in series
in the voltage element.
Three-element Wattmeter for Four-wire Service. The con-
ventional connection of wattmeters for measuring total power or
reactive volt-amperes using three meter elements is indicated
in Fig. 175. This meter measures power quantities in accordance
with equation (336). It has previously been shown by means of
Table XVI that this meter connection measures three times the
positive-, negative-, and zero-sequence quantities in accordance
with equation (338). By adjusting the impedance of the
potential circuit so that the current through the potential coil
is in phase with the applied voltage the meter will measure
the total power. Similarly, by adjusting the impedance in the
potential coil circuit so that the current lags 90 deg. behind the
potential applied to this circuit the meter will measure reactive
volt-amperes.
Ic:;l
Gl
--+-
!
lb ic
b i lb
~
b
Jc i
~ r
e -A
:
Ic
e ___ _ _ ___¡
~
" ____________._______
__,~,
~------4------
FIG. 175.- Conventional three-elemcnt Fw. 176.- Conventional two-

wattmeter connection. wattmeter m c thod of measuring
tota l power on a three-phase three-
wire system.
Conventional Two-element Wattmeter for 'J'hree-wire Service.

The conventional connection for measuring total power on three-
phase three-wire systems is indicated in Fig. 176. It will be
noted that the two-meter elements measure Ecla and EAic,
respectively. Referring to Table XVII these quantities are
represented by items (16) and (12) and when the polarities are
adjusted so that the difference of these · items is taken, the
meter measures
PM + jQM = 3(Etlt + E2I2) - 3a 2Etlo - 3aE2Io (343)
On three-phase three-wire systems, lo is, of course, zero. Con-
sequently the meter measures merely
PM + jQM = 3(Etft + E2I2) (344)
Thus the meter connections of Fig. 176 measure the total
power quantities for all positive- and negative-~equence voltage
and current combinations but do not respond correctly in case
zero-sequence current is present.
On three-phase four-wire systems the zero-sequence current
may be present and the last two terms of equation (343) may
MSASUREMENT Of? POWER QÜANTITIES 311
introduce quite a ppreciable errors. H owever, if the zero-
sequence component is eliminated from the phase currents as
explained in t he discussion of Fig. 159(c), then the last two
terms of equation (343) disappear and the meter reads the sum
of the positive- and negative-sequence power. This is a closer
approximation to the true total power in a four-wire circuit
with grounded neutral than the connection of Fig. 176, as the
zero-sequence power quan-
tity is the product of Ea a - -.......-------~~~
and lo which are usually b ---+-------~~~-+-
small in themselves, where-
as the term E1lo of equation e ---+--......---R~~~--+-
(343) involves one large n ~-~......----(i:---~~--+--~
term.
no~
Special Wattmeter Con-
nection for Balanced Volt-
ages and Un balanced
Currents. It is of interest
to consider the wattmcter FIG. 177.- Wattmeter connec tion for
o
mcasuring total power quantitics with bal-
connect ion of Fig. 177 anced voltages and unbalanced currents.
which is sometimes recom-
mended for measuring the total power quantities on a system
with balanced voltages but with unbalanced currents. Exami-
nation of t his connection with t he aid of Table XVII shows t hat
PM + jQM = Ec(fc - lb) + Ea(fa - fe)'
= 3E 1]¡ + 3E2J2 - 3aEol 1 - 3a 2E 0l 2 (345)
Under balanced voltage conditions, E2 and Ea are both equal
to zero and thus equation (345) reduces to
(346)
Referring to equation (338), it will be seen that when E2 and

E o are equal to zero the total power is 3E J 1· A single positive-
/,
sequence wattmeter will measure correctly the total power on a

ctcuit with balanced voltages but unbalanced currents. Thus
for balanced voltages a single positive-sequence· wattmeter
measures the sarLe quantity as the two-element meter of Fig. 177.
On the othcr hand, a single wattmeter whose current elem ~nt
is connected in phase a and whose voltage element is connected
between phase a and neutral will not measure the total power
318 S YMME'i'RiéAL COMPONÉNTS
quantities. This may be verified from item (1) of Table XVII
which states that the meter in this connection measures
which evidently is not the total power quantity.

Power-factor 41Meters for Polyphase Circuits. The ordinary
indicating power-factor meters for three-phase circuits are built
with a single-phase current winding and three voltage windings.
These instruments do not give correct indication under unbal-
anced conditions though they are correct for balanced conditions.
They are used generally for operating purposes as on switchboards
or for measurements on balanced circuits. An indicating power-
factor meter giving correct operation under unbalanced condi-
tions is relatively complicated and for this reason is not used
commercially. However, sorne of the recording power-factor
meters are built with two elements, one measuring total watts
and the other total reactive volt-amperes which are combined to
give a correct indication of power factor under unbalanced
conditions.
146. Cross Connections for Measuring Reactive Volt-amperes.

The measurement of total reactive volt-amperes of polyphase
circuits, as has been pointed out in the previous section, may
be made with the same connections as described for measuring
power, with the exception that the potential coil circuit of each
meter element must be altered so that the current through the
coillags the applied voltage by 90 deg. This method of measur-
ing total reactive volt-amperes seems so straightforward that
further discussion would perhaps seem unnecessary. However,
practically all the schemes that have been used to date have
been based on the use of the ordinary wattmeter with a line
current and with the voltage of another phase in order to obtain
the required 90-deg. displacement. These schemes generally
give correct operation for balanced conditions but give the
difference of the positive- and negative-sequence reactive vct;-
amperes under unbalanced conditions when zero-sequence
current is not present, and a still less significant quantity when
zero-sequence current is present. Most of these schemes were
proposed befare the method of symmetrical components carne
into general use and t hus were proposed befare the significance
of the quantities measured by these meters could be determined.
It is interesting, therefore, to study sorne of these proposals
by the method of symmetrical componcnts.
A common connection proposed for measuring reactive volt-
amperes is shown in Fig. 178.
The three meter elements measure Gil
PM +jQM =
EAla + EBlb + Ecfc (347) b -~.......~~
which from Tablc XVII may be
shown to be equal to
FIG. 178.-Cross con·nection
P M + jQM = acherne using wattmeters of the
jy3[3(Elll - E2I2)] (348) ordinary type, suitable for measuring
reactive volt-amperes on systems
The presence of the j on the with balanced voltages and unbal-
anced currents.
right-hand side indicates that
wattmeters used with this cross connection measure quantities
proportional to the imaginary terms within the brackets. Thus
under balanced voltages the wattmetcrs of Fig. 178 measure
the total reactive volt-amperes of the circuit. In spite of the
use of three meter elements, this connection under unbalanced
conditions fails to measure the true total reactive volt-amperes,
namely, the imaginary component of 3(EJ1 + E2l2 + Eolo).
For balanced conditions, for which this connection measures
reactive volt-amperes correctly, a sin-
gle wattmeter energized, for example,
by EA and l a might just as well have
been used.
Another combination which is sorne-
times considered is that shown in Fig.
179 which will be found to correspond
with Fig. 176, with the exception that
FIG. 179.- Cross-connec- voltages are taken from other phases
tion schcme for mcasuring h h d b 1 d d' ·
reactive volt-amperes in three- w ic ' un er a ance con ltiOns, are
phase systems with balanccd 90 deg. away from those used for the
voltages and currents. conventional two-wattmeter scheme for
three-phase three-wire service. This scheme may be analyzed
with the assistance of Table XVII with the result that
It will be observed that under balanced voltage conditions when

l o = O, wattmeters used in this connection measure a quantity
proportional to the total reactive volt-amperes, irrespective of
the val u e of 1 2· Thus the limitations are similar to those of the
connection shown in Fig. 178.
147. Metersforthe ~easurement of Sequence Power Quantities.
Sequence power quantities may · be measured in the same
manner as phase quantities by impressing a sequence voltage
and a sequence current upon the meter element instead of a
phase . voltage and a phase current. Thus positive-sequence
watts, reactive volt-amperes, and power factor may be measured
by impressing positive-sequence voltage and positive-sequence
current upon the meter elements. In a similar manner negative-
and zero-sequence power quantities may also be measured.
Positive-sequence Wattmeter. The method of measuring
sequence power quantities is conveniently illustrated by the
positive-sequence wattmeter. For obtaining positive-sequence
watts any of the general positive-sequence segregating networks
for voltage or current, as discussed in Chap. XIV, may be used ;
the only restriction is that the impedances be so chosen t hat
the same relation between the sequence currents through t he
meter coils be obtained as in t he conventional connections for
phase quantities. Thus for measuring positive-sequence watts
the current in the potential coil should be in phase with t he
positive-sequence voltage of the system, and the current in the
current coil should be in phase with the positive-sequence
current of the corresponding phase of the system. More gener-
ally, the meter will measure a quantity proportional to positive-
sequence watts provided that the phase displacement of the
cu-rrent through the potential coil with respect to the positive-
sequence system voltage is the same as the displacement of t he
current through the current coil with respect to the positive-
sequence current of the system.
In order to illustrate these general relations assume that it is
desired to use the positive-sequence voltage network of the
common type of Fig. 148 and the bridge-type current network
of the type of Fig. 157. A sequence wattmeter with t hese
segregating networks is shown in Fig. 180. For this meter
the expression for the current through the potential coil is given
in Chap. XIV, equation (306), as
_ 3El
1M 1 -
ZB + y3Mci 300
which may be written as
IM l = ~El (350)
KEiOK
The current through the current coil Iml is from (323)

lml = ]¡ (351)
Let it be further assumed that thc impedance M of the potential
coil is negligible and that the impedance ZB is apure resistance
RB, then
I Ml = 3El (352)
RB
~ ------------------------------------~
b --------------~~+-------~--------+-
e ---+~9+----~~--~------~--~----~
.
'·
~
lmt
'• 1
FIO. 180.-Posit.i vc-sequence wattmeter using bridge-type current network.
The power quantities measured by the system of Fig. 180 may

now be wri.tten as
P M+ J'QM= 3El"'
RB ]¡ (353)
Thus the meter of Fig. 180 gives an indication proportional to

the total positive-sequence power of the three-phase circuit.
Since the resistance term RB is included in the meter constant,
the meter scale will read the total power quantity.
The effect introduced by the impedance of the potential coil
having a finite value may readily be corrected by making the
power-factor angle of the meter coil branch exactly 30 deg.
lagging. Under these conditions it will be seen from equation
(306) that the current through the potential coil is in phase

with the positive-sequence voltage when ZB = RB, a non-induc-
iive resistance. There is also the effect of the potential coil
impedance on the calibration which must be taken into account.
Another combination
using a different segre-
gating network may be
considered in order to
further illustrate the
method of proportioning
Zs the network impedances
to measure the desired
sequence power quanti-
(01) Oi01gram of connections ties: Assume that it is
desired to make a posi-
Potenfia/ coi/
tive-sequence wattmeter
using the positive-se-
Compensating coi! quence voltage network
(b) Maqnitude and phose on9le of Fig. 148 and the posi-
compensation sc:heme for (o)
tive-sequence cur~en t
Fw. 181.-Positive-scqucncc wattmcter using network of Fig. 155, the
current network of Fig. 155.
combination being illus-
trated in Fig. 181. The expression for the current through the
current coil is from equation (318)
lml =
1 + m_ ci3oo (354)
V3Zb
in which m is the meter impedance. For this connection the
impedance of the current coil cannot be ignored. Consequently,
the procedure will be to make a simple current network and
adjust the voltage network to give the desired phase relations.
Thus, if the impedance Zb be a non-inductive resistance Rb,
equation (354) may be written as
1 + m
E-130
. o kéiok (355)
-yi3Rb
In arder that the meter may measure positive-sequence watts,
the phase angles Ox and fh of the right-hand members of equations
(350) and (355) must be equal. The determination of the
proper impedance values for the network and meter branches

of the potential circuits may be carried out conveniently by
the method illustrated in Fig. 149 where the angle DOR of the
voltage network must be equal to the angle fh of the current
· network.
The positive-sequence wattmeter of Fig. 181 is, however,
more conveniently adjusted by using voltage and current-
segregating networks each having a non-inductive resistance
branch. If this is done, it is necessary to adjust the phase
relation of the current through the potential coil by shunting
the meter element with a suitable impedance, so that the current
through the potential coil lags the _¡)ositive-sequence voltage
by the same angle that the current through the current coil
lags the positive-sequence current. This may be accomplished
by replacing the voltage coil branch of the meter of Fig. 181 (a)
by the connections of (b) for which a shunt inductive coil of a
suitable value is adjusted to obtain the proper phase-angle
correction, and the externa! resistance is adjusted for final
scale calibration.
Under balanced voltage conditions, the sequence voltage
segregating network may be eliminated. Thus an ordinary
wattmeter may be used in Fig. 180 provided the current coil
is made a part of a positive-sequence bridge-type current segre-
gating network.
Positive-sequence Watthour Meters. Watthour meters may
be modified by suitable externa! networks to register a sequence
watthour quantity instead of a phase quantity by impressing
sequence voltages and currents of the system instead of phase
voltages and currents. Watthour meters differ from wattmeters
of the dynamometer type in that the current through the poten-
tia! coil of the former lags approximately 90 deg. behind the
impressed voltage, whereas the current is in phase with the
voltage in the dynamometer-type meter. The network con-
nections are shown in Fig. 182(a) and differ from those of Fig.
181 principally by the fact that the voltage network is connected
so that the e phase component rather than the a phase component
of positive-sequence voltage is used in connection with the a
phase component of positive-sequence current. The reason
for this is developed later.
Constructing the impedance diagram of Fig. 156 u pon 1al
in Fig. 182(b), and using a typical phase angle value for the
meter impedance m, the phase position of I ml relative to I al

is obtained. Constructing the voltage imp~dance diagram of
Fig. 149 u pon - Ec1, it may be seen that the current through
the voltage coil I M lags behind -Ecl by the angle Ox. For
the meter to measure true watts, l M1 must lag behind I ml by
approximately 90 deg. It will be seen that this condit ion
is fulfi.lled only when - Ec1 is used as the reference voltage.
The ultimate condition to be fulfi.lled is that the flux set up by
I MI lag I mi by 90 deg. This final adjustment is accomplished
by means of a compensating coil in the magnetic circuit of the
voltage element. An alternate or supplementary means to
~------------~---------- la•
b------~~--~--~------
e~~+-+-~--~--~----~
Die19r.::~m of connections Vector clic;¡gmm
(b)
F10. 182.-Positive-sequence watthour meter .
obtain this adjustment is to shunt the voltage element with a

resistan ce.
148. Flow of Power Quantities Due to Unbalance.

It has been pointed out previously that symmetrical machines
generate no negative- or zero-sequence voltages but only posi-
tive-sequence voltages; consequently, they can supply to the
system only positive-sequence vector power (P 1 + jQ¡) . In a
symmetrical system with symmetrical loads (or short-circuits),
only the positive-sequence power quantities can fiow. However,
in the case of an unbalanced load (or short-circuit), negative-
and zero-sequence power quantities may be produced; thus
in a sense positive-sequence power quantities are supplied
by the generators and the dissymmetry converts part to negative-
and zero-sequence power quantities which are fed into the
system at t he point of fault. This conception coincides with
ME.ASUREMENT OF POWER QUANTITIES 325
the equation~ for the flow of pow~r under unsymmetrical con-
ditions as is readily brought out from the examination of a
particular case.
Consider a single unbalanced load from line-to-neutral
on a system supplied through a grounded generator with positive-,
n e g a t i ve- , and
zero-sequence im-
pedances of Z 11 Z 2,
Zt <>-t--.---- DI
and Zo, respec- Zz o-t--+---- b
tively, and with a Zo o-;--:r--- ~
generated positive-
sequence line-to-
neutral voltage of
Egl· Assume that
this system is sup- Fw. 183.-Diagram for illustrating the fiow of sequence
plying a load on power quantities.
phase a whose impedance is Z L as illustrated in Fig. 183. This
problem was solved in Sec. 21 of Chap. III with the result that
E1 = (Z2 + Zo + 3ZL)~e1
'
~· E2 = - Z2Eg 1
' Zt
Eo = -ZoEg 1
Zt
l1 = l2 = lo = Egl
Zt
where
Zt = Z1 + Z + Zo + 3ZL
2
The total phase quantities at the point of load are readily

obtained from the above sequence quantities with the following
results
E a-- 3ZL
Egl
-
Zt
la= 3Eg1
Zt
The sequence power quantities per phase at the load are
(356)
P, + jQ, = -z{t')' (357)
Po+ jQo = -zo(~.J (358)
The total power quantities may be obtained by combining

the sequence quantities as in equation (342), giving
PT + jQT = 3[(Pl + jQ¡) + (P2 + jQ2) + (Po+ }Qo)] =
9ZL(~, )'
1
(359)
From the single-phase solution the total power quantities equal
PT + ]QT =
• A
Eafa + "' A IE 01 )2
Ebh +EJe = 9Zd-=- (360)
· \Zt
In general, the significance of the negative sign of equations
(357) and (358) is that the flow of P2 + jQ2 and Po + jQo is
in a direction opposite to that of P1 + jQ1. In other words,
an unbalanced load not only draws from a symmetrical system
the total power quantities required for the load but also the
negative- and zero-sequence power quantities which are fed
back into the system at the point of unbalance and which are
consumed in producing negative- and zero-sequence losses
and in absorbing (in the ordinary system) inductive reactive
volt-amperes.
Similar expressions may readily be derived for a line-to-line
fault or single-phase load from line-to-line. In this case the
negative-sequence current in phase a will be the negative of the
positive-sequence current of phase a. However, the negative-
sequence voltage at the fault will have the positive sign so that
power quantities are still fed into the network at the point of
un balance.
Combination of Unbalanced I,oad and Symmetrical Machine
Load. In the combination of an unbalanced load, such as due
to a single-phase load and a balanced load due to rotating
machinery, the unbalanced load produces negative-sequence volt-
age at the point of unbalance and th~s negatiYe-sequence voltage
impressed on the system causes negative-sequence current to flow
not only through the source as described previously but also in all
shunt branches including the symmetrical rotating-machinery
load. The power dueto the flowof the negative-sequence currents
produces losses in the rotating-machinery load and increases the

total amount of power absorbed by t he symmetricalload. Rota-
ting machinery on a system tends to balance the load on a system
and to restare the voltages to their normal values. This action
takes place with the flow of negative-sequence current in the rotat-
ing machinery and with an increased amount of lose anda reduc-
tion in the maximum load capacityof the apparatus. Under sorne
conditions negative-sequence current flowing through machines
will produce exccssive heating in the rotor so that unbalanced
operation may rcquire special consideration as discussed for
induction motors in Chap. XVII.
In a number of respects, as pointed out by Dr. FortescueC17),
positive-sequence power quantities reprcsent more accurately
than total power quantities thc requirements of power supply
systems for handling unbalanced loads; and in addition they
give identically thc same result for balanced conditions.
Problems
1. The average clcmcnt of a wattmcter has a resistance of 100 ohms and
an inductancc of 50 mil-henrys. The externa! rcsistance is 15,400 ohms.
What capacitancc is rcquired across the external rcsistance to correct the
phase angle?
2. The voltagc clcmcnt of a reactive volt-ampere meter has a resistance of
100 ohms and an inductance of 50 mil-henrys. The externa! reactor has
an impedance of 3,200 + j14,600 ohms. What capacitance is required across
the element to correct the phase angle?
3. A three-phase system, whose voltages across lines are 110 volts, is loaded
with the impedances 20 + j3, 5 - j20, and 10 + j20 across lines. D eter-
mine PT + jQT and P1 ·1- jQ¡ and P2 + jQ2.
4. A generator whose synchronous reactance Xd is 1.1 and whose negative-
sequence reactance x 2 is 0.15 is loadcd across two terminals by an impedance
2 + j0.3 (in tcrms of per unit por phase). Thc voltage across these two
terminals is maintainecl at normal. What is the excitation, neglecting
saturation and defining unit excitation as the excitation required to produce
normal voltage at no load?
6. What is the positive- and negative-sequence power at the terminals
ef the machine in Prob. 4?
CHAPTER XVI
/
MULTIPHASE SYSTEMS
The principal application of the method of symmetrical

components is in connection with the solution of the problems
of unbalance on three-phase systems. The method, however,
is applicable to other multipbase systems used for distribution
and conversion. These include the single-phase three-wire
system, the two-phase three- and four-wire systems, and the
four-phase four- and five-wire systems used for distribution,
and also the six- and twelve-phase systems used for synchronous
converters and mercury-arc rectifier applications. A study of
these multiphase systems will bring out certain points in con-
nection with the fundamental theory of the method of sym-
metrical components to a better advantage than is possible by
consideration of the three-phase systems alone.
149. Resolution of Multiphase Systems into Symmetrical

Components.
In arder to explain the application of the method of symrnetri-
cal components to rnultiphase systems, it will be necessary to
review certain fundamental propositions relating to the theory
of numbers and then to illustrate them by a few simple examples.
A system of n general nurnbers may be resolved into n (or more)
sets of component numbers, the number of components in each
set corresponding to the number of original general numbers.
The general numbers may be assumed as the voltage vectors
Ea, Eb, Ec, . . . , En
The component numbers of each set may be designatcd by
adding a number to the letter subscript identifying the original
vectors, as follows:
Eao, Ebo, Eco
Eal, Ebl, Ecl
Ea2, Eb2, Ec2
328
MULTIPHASE SYSTEMS 329
The sum of the component numbers associated with each
general number is equal to each general number itself. Thus
Ea = Eao + Eal + Ea2
Eb - Ebo + Eb1 + Eb2 (361)
Ec = E co + E. cl + E c2
This method of solving a problem involving n general numbers
by resolving them into n 2 component numbers or vectors would
not, upon casual examination, appear to facilitate the solution.
In fact, no advantage is obtained for the most general case
though remarkable simplification results when certain kinds of
symmetry exist in the relations between these general numbers.
In analyzing the problems involving the general case of the
three-phase system, Dr. Fortescue recognized the limitations
that existed when considering the two sets of voltages dependent
upon phase rotation. Consequently, he went back · to a more
fundamental consideraLion based on the theory of numbers
which showed that the maximum simplification was obtained
when advantage was taken of the symmetry that existed in the
relation between these nurnbers. After recognizing this possi-
bility of simplification he undertook the problem of finding the
law governing the formulation of the sets and the components
of each set which would reduce the calculations to the maximum
degree and yet possess sufficient generality to insure a solution
for each particular problem. As a result of this study Dr.
F ortescue discovered the principies governing the method of
resolving the general numbers into sets of component numbers
which constitutes the basis of the method of symmetrical coor-
dinates or symmetrical components.
In the method of symmetrical components, the canonical
form of setting up the sets of numbers is to make all the com-
ponent numbers of equal magnitude in each set. Their phase
position, however, is dependent upon the power of a character-
istic operator. This characteristic operator is one which pro-
360
duces a phase rotation of deg., where n is the number of
n .
phases or the number of the original vectors. The method of
setting up the different sets of vectors is to take the vectors
in any cyclic order, choosing one vector or phase as a reference.
The component vectors of each set may be derived from the
~30 SYMMETRICAL COMPONENTS
vector in the reference phase by rotating it successively by the

characteristic angle of each set. The characteristic angles
for the different sets are taken as increasing multiples of the
characteristic angle of the system. This method of setting
up the sequence components for a multiphase system results in
Ea = ei0 E ao + ei0 Eal + ei0Ea2 + ei0 Ea3 . . .
Eb = ei 0 E ao + e-ieEal + e-i 28Eaz + e-i 38 Eaa · · ·
Ec = ei 0 Eao + e-i 20Eal + e-i 48Eaz + e- i 68 Ea3 (362)
Ea = ei 0Eao + e-i 30Eal + e-i 60Eaz + e-i 90Ea3 · · ·
It will be noted that the characteristic angles have been taken

as negative. This choice was made in order that the generated
voltages of sequence 1, namely, Eal, Ebl, E cl, . . . will have their
maximums occurring in time in the cyclic order according to the
alphabet. In connection with the above set of equations, it is
reallzed that the vectors E ao, Eal, Eaz, . . . may have any
magnitude and any phase position; in other words, these vectors
have the same number of degrees of freedom as possessed by
the original vectors E a, Eb, Ec, . . .
The set of equations (362) may be written in the form of (363)
by using r as a complex number equal to Ei!J.
Ea = E ao + Eal + Eaz + E ,3 + · · ·
Eb = E ao + rn- Eal + rn- E az + rn- 3Eaa +
1 2
Ec = E ao + rn- Eal + rn- 4Eaz + rn- Ea3 +

2 6
Ea = Eao + rn- Eal + rn- Eaz + rn- Ea3 +

3 6 9
The analytica.l expression for the sequence components can

be obtained by solving the n equations of (363) resulting in the
valu es
(364)
The general method of symmetrical components is therefore

applicable to any number of phases. The characteristic features
of the dH'fereut sequence systems for several multiphase systems
are summarized in Table XVIII. As a device for visualizing the
characteristics of the different sequence systems the component
in t he reference phase has in every case been taken as of unit
magnitude and in phase with the reference. This procedure is
used in order to make it easier to remember the relation between
the various components in relation to the component in the
reference phase. Table XVIII covers multiphase systems with
two-, three-, four-, five-, six-, and twelve-phase conductors.
The characteristic angles for the various systems are 180, 120,
90, 72, 60, and 30 deg., respectively. That the characteristic
angle of the two-phase system is 180 deg. instead of 90 deg. may
a t first seem surprising, but comparison with the other systems
will quickly establish the correctness of this statement. The
., . two-phase three-wire system is in reality an irregular system
which will be reviewed in greater detail subsequently. In this
analysis the multiphase systems without neutral wire or ground
connection are considered as special cases of the system with
neutral wire for which the zero-sequence components are zero.
Examination of Table XVIII shows that when the number of
phases n is prime as is the case for the three- and five-phase
systems, the original vectors may be resolved into one set of
zero-sequence vectors and (n - 1) symmetrical sets of vectors
which differ only in their subscript designation. Also it may be
shown that different sets of components may be obtained from
. the same set of original vectors, depending upon the original
choice of the arbitrary order in which the phases were taken.
Consequently, it is possible for certain problems to obtain
considerable simplification by choosing the proper arbitrary
order, which will normally be taken so that the generated e.m.fs.
will always be of sequence l . Further examination of Table
XVIII shows that for the cases in which the number of phases n
is not prime, the system may be resolved into multiphase sys-
tems of lower order, for example, the twelve-phase system may
be resolved into four three-phase systems or into three four-
phase systems. The twelve-phase system may also be viewed
as made up of four twelve-phase vector systems (sequences 1,
5, 7, and 11), two six-phase vectors (sequences 2 and 10), two
four-phase vectors (sequences 3 and 9), two three-phase vectors
(sequences 4 and 8), a two-phase vector (sequence 6), and the
single-phase vector (sequence O). In case the twelve-phase
system is obtained from three-phase transformers, there may
be an appreciable advantage in certain problems in resolving

the twelve-phase system into four three-phase systems. Simi-
larly, if the twelve-phase system were derived from a four-
phase supply, then there may be a corresponding advantage m
TABLE XVIII.-SEQUENCE CoMPONENTS FOR MuLTI-PHASE
SYSTEMS
ltem Number Seque~e Component

Number of Vectors Seq.O Seq.l Seq.2 Seq.o Seo¡.4
2 _Eao Eb¡--Ea•
Ebo
Ec1 Eb2
E~ o
Z 3 ---+Ebo
Eco
)-Ea. )-Eaz
Eb1 Ea
3 4
4 5
Seo¡. JI
Seq.G
--
Eb,
Ea,
E(l 6
Ecó
Ef6 Ee6
Eh6 Eg,
Ej& E¡c,
E¡ 6 E._.
resolving the twelve-phase system into three four-phase systems.

Otherwise, it is generally found desirable to consider the twelve-
phase system as set up in Table XVIII.
150. Two-phase and Four-phase Systems.
The various forms of the two-phase and four-phase systems
will be discussed together since they have the common feature
of generated e.m.fs. 90 electrical degrees a part. The four-phase
.' :five-wire system is the most general case and may be solved as
outlined in item 3 of Table XVIII. The four-phase four-wire
system without neutral connection may be viewed as a special
case of the five-wire system for which the zero-sequence compo-
nent is zero. Frequently the four-phase system without neutral
connection, or the two-phase system as it is commonly called,
f.s operated with conductors and loads symmetrically disposed
Die~gram of connectsons
(01)
la¡
lb
Unbale~nced line currents Resolution into components - sec¡uence 1 a na sec¡uence 3 .
(b) (e)
Fw. 184.- Two-phaso four-wire system.
with respect to ground and with loads taken only between wires
a and e, and b and d, as indicated in Fig. 184. In such a system
the currents are
} (365)
The above set of equations may be analyzed as a four-phase

system with t l:e result that
l o= O
l _la+ jlb l _ l a - jlb
1- 2 3- 2
Thus it will be seen that t he four-phase or two-phase system

, of Fig. 184 may be solved in the manner indicated in ítem 3
of Table XVIII using only the positive-sequence components
(sequence 1) and the negative-sequence component (sequence 3)
as illustrated in (b) and (e) of Fig. 184. For balanced conditions
the voltages and currents will all be of sequence 1, and only

one sequence of generated e.m.fs. will be required for the analysis
of the system.
151. Irregular Systems.
The two-phase three-wire system of Fig. 185(a) may be
:.:egarded as an irregular system because the characteristic angle
¡e i1 q: -I~
-lb
-Ic
Two-phc:~se three-wire
-Ia
-rb
-re
Single-phc:~se
~~
three-wire
(oc) (b)
la1
Ea
(e) (d) (e)

(e) Normal generated c.m.f.'s in a single-phase three-wire system.
(d) Generated e.m.f.'s in a two-phase three-wire system in terms of the
symmctrical componcnts of a single-phase three-wire system.
(e) R esolution of phase currents into their symmetrical components.
(f) Analytical expressions
l ao = Hla +lb)
l at = !(/a- l b)
la = lao + lat
h = 1 ao - 1 al
Fw. 185.-Reprcsentation of a two-phase three-wire system by componentl! of a
single-phase three-wire system.
360
of the system - = 180 deg. is different from the phase angle
n
of the generated e.m.fs. Because of this fact the generated
e.m.fs. cannot be represented by a single sequence of a regular
system such as those that have been described. However,
by building up certain types of irregular sequence systems it is
possible to represent the generat@d e.m.fs. by a single sequence.
The two-phase three-wire system has been selected for analysis
to illustrate the general treatment of irregular systems.
The two-phase three-wire system shown in Fig. 185(a) can
be analyzed in terms of the single-phase three-wire system
illustrated in (b) of the same figure. Each phase of su eh a
system may vary in magnitude and phase position. In the single-
phase three-wire system ordinarily the generated e.m.f. is a
positive-sequence (sequPnce 1) only, as shown in Fig. 185(c);
but for this system to represent the generated voltages of a
two-phase three-wire system, two sequences (sequence 1 ana
sequence O) are required. This is illustrated in Fig. 185(d)
-In
Two-ph<"se three-wire system
(e;¡¡)
Resolution of ph01se currents

Ib¡ "----;~oo-la2 into symmetric01l components
Positiva- sequen ce NegOttive-sequence
components components (c4)
(b) (e)
Fw. 186.-Representation of a two-phasc thrcc-wire system by components of an
irregular sequence system .
in which Ea and Eb represent the generated phase voltages and

the other vectors the sequence components. The resolution
of two general phase currents into their sequence components
is shown in (e) of the same figure. The different sequence
networks do not have coupling between them if the phase wires
are symmetrically transposed with respcct to the neutral wire
or ground. The generator impedances require sorne manipula-
tion to convert from one system to the other.
The two-phase three-wire system may also be handled by an
irregular system of sequence components such as that of Fig.
186. It will be noted that in this irregular system t he two sets
of sequence components are taken, one set comprising components
90 electrical degrees apart in the positive sense and the other
1
set 90 electrical degrees apart in the negative sense. This

'rregular system requires only one set of sequence components
of the type illustrated in Fig. 185 to take care of the normally
generated e.m.fs. of a two-phase three-wire system. The
addition of sequence 2 takes care of unbalance. The resolution
of the two phase currents into the symmetrical components
is shown in Fig. 186(d). It is to be realized, however, that
coupling will be introduced between positive- and negative-
sequence networks in case of an impedance in the neutral con-
nection or in case the impedance of the distribution line requires
consideration.
The several methods which have just been outlined for handling
the two-phase three-wire system have been described for the
purpose of illustrating the types of sequence systems that
may be used for solving special problems. The choice of the
sequence system to be used with the different types of networks
which have been described will be reviewed briefly. For the
general case of the four-phase four- or five-wire system, the
four-phase system as outlined in item 3 of Table XVIII is
preferred. For the ordinary two-phase system of Fig. 184,
the same sequence analysis may be used except that sequence O
and sequence 2 may be omitted. The two-phase three-wire
system without neutral impedance or without distribution lines
of sufficient impedance to require inclusion in the calculations
will be found most convenient by the use of the "irregular
system" described in Fig. 185. If the two-phase three-wire
systcm has a neutral impedance or is used for distribution, · it
will be found more convenient to use the conventional two-phase
system as outlined in Fig. 185. This choice for the two-phase
three-wire system is suggested because of the fact that it is
ordinarily simpler to add generated e.m.fs. in the negative-
sequence network than the alternative of requiring coupling
between the sequence networks though using but a single source.
162. General Treatment of Polyphase Circuits.
All polyphase systems that are symmetrical, viewed from
the standpoint of system constants, may be solved in a manner
similar to that for the three-phase systems which have been
described previously. The constants can be determined sepa-
rately for the individual sequence systems and the constants
of the individual sequence network set up in a network of its
own. For this case there will be no reaction between the different
networks. The tie between the different networks will be the
relations imposed by the particular kind of fault or unbalanced
terminal condition.
Unsymmetrical polyphase systems are more complicated
than symmetrical systems in that mutual impedances exist
between quantities of the different sequences. This problem
will be discussed in Chap. XVIII.
153. Impedances of Symmetrical Polyphase Systems.

The impedances of static equipment, such as transformers
and transmission or distribution lines, are so much like those
for the three-phase systems
as to require no further
\
consideration. Rota ting

equipment introduces a
quite different phenome-
non which warrants further
discussion.
For this purpose consider
a polyphase machine in
which all the windings are 0
distributed, wound so that
a sinusoidal space field is FIG. 187.- Distributed winding whose axis
produced by a current flow- makes an angle {3 witt the X-axis.
ing through each of the
phases. ln Fig. 187 let {3 be the phase position of the axis of a
winding relative to sorne arbitrary reference line which will be
assumed to lie along the horizontal X-axis. Consider first that
this angle is zero and that the current i = V21 cos (wt + a)
flows through this winding. As shown in Chap. II, equation
(16), this current can be expressed as
(367)
This current can be represented as the sum of two conjugate

vectors rotating in opposite directions as shown in Fig. 188.
The imaginary parts thus cancel out, leaving only the real terms
which express the instantaneous value of i. For t = O, the
vatue of the curreut is
- ;-
V 2-
- [Jé:.
2
+ JE-ja]
-
(368)
The sinusoidal current flowing through the winding under

consideration sets up a pulsating field which at any instant has
a sinusoidal space distribution. The axis of this sinusoidal
flux wave, since {J = O, lies along the X-axis. According to
the well-known device due to Ferraris, a pulsating sinusoidal
flux wave can be resolved into two sinusoidal flux waves of
opposite rotation. The magnitude and position of these flux
waves coincide with the position of the two vectors of Fig. 188
representing the instantaneous values of current.
w
Vz _ j(wt+cx)
-/ZIE
.__ z-u
1/2 _ .j(wt+oc)
Fw. 188.- Resolution of a Fw. 189.-Rcsolution of a pulsating

pulsating flux in the X-axis into flux in a winding whose axis makes
two fluxes rotating in opposite angle f3 with the X-axis, into two
directions. components rotating in opposite
directions.
N ow consider a similar winding whose axis lies at an angle

{J from the reference line. If the same current be made to flow
through the winding, two rotating fields will again be set up
which will have the same positions relative to the axis of the
winding. Since, however, this axis is now at an angle {J ahead
of the reference line, the vector position of the resultant m.m.f.
or current field will be given by
0 j é(c,;t+a) +if3 + y2JE-j(wt+a) +if3 (369)

2 2
which consists of two oppositely rotating fields as shown in
Fig. 189.
With this fundamental relation showing the vector position
of the two rotating fields as a function of the phase position
of the current in the winding, and the action of the winding
established, the results may be applied to an n phase machine.
The phase displacement between the windings of such a machine
360
will be deg.; {3 will then have successively the values O,
n
360 · ~360 · n -
1
360. For the zero-sequence set of
n' n ' n
currents the phase currents are all equal to the a phase current.
The field set up may therefore be written:
For the a phase
V2-
21é<wt+a) v2-¡~-j(wt+a)
+2 (370)
For the b phase

v2- ;(wt+a+.!.360) v2J- -i(wt+a-~360) (371)
2/~ n +2 ~ n
For the e phase

v2¡-./(wt+a+~36o)
2 ~
0- -i(wt+a-~360)
+ 2/~ n (372)
For the n phase

v2- i( wt+a+n- 1360) 0¡- -i( wt+a-n- 1 360)
2 /~ n +2 ~ n (373)
The resultant field is the sum of the separate fields due to the
n phases. Adding up the fields of positive rotation and factoring
out v2¡-~i ( wt+a) , t h e followmg

. coeffi ment
. .
remams:
2
(374~
Since these terms are all equal except for their arguments which
360
differ by multiples of , the sum is zero. Therefore, the net
n
field rotating in a positive sense is zero. The coefficient for
the negatively rotating field is the same. It follows that the
air-gap flux in such a machine due to a zero-sequence set of
currents is zero. The reactance to the zero-sequence is merely
the effect of departures from a true sinusoidal wave shape and
leakage effects between phases. This reactance will in general
be low and be affected greatly by chording.
For the positive-sequence set of currents, calling the a phase
the reference phase, a will have successively the values a,
a - ~360 a - ~360 · · ·
1
360. The fields set up
a - n -
n ' n ' ' n
in the different phases are then the same as for the zero-sequence
[equations (370) to (373)] except that the alphas must be given
the val u es indicated. Thus the fields are:
For the a phase
V2-1é(wt+a) + -vl2-1e-i(wd-a)
- (375)
2 2
For the b phase
v/2- i (wt+a-.!.360+.!.360)
21e n n
V2- -j(wt+a-.!.360-.!.360)
+ 21€ n n
or
v/2
_ _1- eJ'(wt + a) + _v/21-€ -j(wt+a-~360)
n (376)
2 2
For the e phase
V2 - i (wt+ a-~360+~360)
2 1€ n n + 2V2-1€ -j (,.t+a-~360-~360)
n n
or
V2 - eJ'(wt + a )
_ _1 + _v/21- € -j (wt+a-~360)
n (377)
2 2
For the n phase
v'2 - €i(wt+a- n-l360+n-
_ 1 n n
l360)
+ v/2- -j(wt+a- n-l360-~!3ao)
- c- 1 € n n
2 :¿
or
V
_ ;2-
- '( + ) + _V_ _
/ñ2
J:JJ-€
·(wt+a--
2(n-1)
- 360
)
__ 1 el wt a
-J n
(378)
2 2
It will be observed that for t he positively rotating components,
the contributions from the individual phases are all equal and
add up directly in space phase to produce a net resultant field.
This field i:u the case of a synchronous machine rota tes in the
same direction and wi th the same speed as the rotor. The
impedances associated with it are of the same nature as
the positive-sequence impedances described in Chap. V. With
regard to the negatively rotating field, after factoring out the
coefficient ~21e - i(w, +a) , the remaining terms, on adding, are
2 +J'~360
e'·o +en
· 360
+ J-
+en + (379)
The summation of n terms separated by the angle ~360 is equal

n
to zero. This may be shown by dividing the terms into
two groups, the one consisting
n
of the first
2 terms and the
other of the remaining terms.
The summation of the terms
in each gro u p is zero as indi-
cated for the first group in
Fig. 190. The negatively
rotating field is therefore zero.
Corresponding expressions Fw. 190.-Diagram showing that i
can be set up for the other equally spaced vectors separated by the
2
sequences. For the second- angle, n-360, add to zero.
sequence in particular, after .
factoring out the common terms, the coefficient for the positively
rotating field is
E;o +E - ; n +E - i~360
360
-j~360
n + E n (380)
and for the negatively rotating field is
E'.o +E - i-ª-360
n +E - i~360
n + (381)
Both of these summations add to zero, so that the second-

sequence air-gap flux in an actual machine is practically zero.
The impedance for this sequence partakes of the nature of the
zero-seq u ence.
Similar considerations apply to the other sequences, except
the (n - 1)th or the negative-sequence. For this sequence
the coeffi.cient for the positively rotating field becomes zero,
but the components of the negatively rotating field are all in
phase and add just as the components of the positively rotating
field added together for the positive-sequence. Therefore,
since only a negatively rotating field is present for the negative-
sequence, the negative-sequence impedance is dependent upon
the impedance of the rotor and will depend, therefore, upon the
leakage of both the stator and rotor.
Summarizing, in symmetrical machines the positive-sequence
current (sequence 1) produces a positively rotating field which,
reacting with the rotor circuits, gives rise to the various kinds
of positive-sequence impedance. The negative-sequence currents
(sequence n - 1) develop a negatively rotating field which
reacts with the rotor circuits and gives rise to the negative-
sequence impedance. All other sequences develop no resultant
rotating fields. The impedances associated with these sequences
are of the nature of armature-leakage impedances and vary with
chording.
164. Harmonics.
Distorted waves of a periodic nature can be resolved into
a Fourier series consisting of a fundamental and higher harmonics.
The result can always be expressed in the form
i = "\1'21 1 cos (wt +a¡) + v/212 cos 2(wt + a2) + V2la cos
3(wt aa) + + ··· (382)
in which 11 12, . . . , represent the r.m.s. values of amplitude
of the respective harmonics. This relation can also be expressed
as
i = real part of lv/21 1é(w,+at) + v/2l 2e2i(w,+az) +
v/21 3e:lj(w¡+aa) + · · · ] (383)
Since the impedance to the fundamental component of current
in sorne cases is influenced to a very considerable extent by its
sequence, it is highly desirable to determine the sequence of
the harmonics. This analysis will be restricted to three-phase
systems, but it will be apparent that it may be extended to
systems of any number of phases.
Assuming that the system under consideration is symmetrical
and of positive-sequence, then the three-phase currents will be
identical except for the phase shift of 240 and 120 deg. In
this case the three phase currents may be expressed analytically
by the following:
ia = real part vf2[1lé(w¡+a¡) + 12é2 (w¡+cu) +
1 3é3(w¡+aal + ]
'tb = real part y'2[1ré(w¡+at+240) + 12é2(w¡+az+240) +
1 3ei3(w¡+aa+240) + ...
'te =real part vf2(1¡é(w¡+at+120) + 12é2(w¡+at+120) +
J 3é 3(w¡+aa+ 120) + . . .]
1
MUL1 IPHASE SYSTEMS 343
or
Ía =real part y'2[Jlé(wt+a¡) + l2é2(wt+a2) +
J 3é3(wt+aa) +
'tb = real part v2[a2Jlé(wt+a¡) + al2é2(wt+a2) + . (385)
J 3é3(wt+aa) + ..]
. 'te = real part vi2[al 1é<w,+a~> + a2J2f.i2(w,+a2) +
' J é 3(wt +a a) + . . . ]
3
Thus it will be scen that the fundamental is of positive-sequence,

the second harmonic of negative-sequence, and the third harmonic
of zero-sequence. This result is summarized in Table XIX.
For the negative-sequence, the three phase currents may be
written
'ta =real part yl2[]¡é(wt+a¡) + 1 2f.j2(wt+a2)
+ J 3f.j3(wt+aa) +
'tb = real part yi2[J é<w,+a¡+lzo> + J 2é 2(w,+a2+Izo>
1
+ J 3é3(wt+aa+l20) + ..
'te = real part yl2[llé(wt+a¡+240) + 12é2(wt+a2+240)
+ laé3(w¡+as+240) +
or
Ía = real part y/2[llé(wt+a¡) + 12é2(wt+a2)
+ J 3és(wt+aa) + . . . ]
'tb = real part yl2[al1é(wt+at) + a2J2f.j2(wt+a2)
+ j f.j3(wt +aa) + . . . ]
3
'te = real part v2[a2J lé<w,+a¡) + a l 2é2(w,+a2>
+ J3é3(w¡+as) + ... ]
The fundamental is thus of negative-sequence, the second
harmonic of positive-sequence, and the third harmonic of zero-
sequence. These results are also summarized in Table XIX.
For the zero-sequence currents the currents in all three phases
are in phase so that the fundamental and all the harmonics
are of zero-sequence.
The fundamental frequency positive- and negative-sequence
impedance of machines rotating at normal speed differ con-
,, siderably. This is due to the fact that the rotating field set
·'
up ·by the positive-sequence component is largely magnetizing,
no currents are induced in the rotor; whereas for the negative-
sequence, currents of double frequency are induced in the rotor
and the impedance is determined largely by the leakage fluxes
of the stator and rotor. For the harmonics, currents are always
\•
T ABLE XIX.-SEQUENCE OF HARMONICS IN THREE-PHASE SYSTEMS
Sequen ce
Harmonic
Positive Negative Zero
1 1
1 1 2 o
2 2 1 o
3 o o o
4 1 2 o
5 2 1 o
6 o o o
7 1 2 o
8 2 1 o
9 o o o
set up in the rotor, so that the difference between the positive-·

and negative-sequence impedances decreases as the order of the
harmonic increases.
Problems
1. A two-phasc system is converted to a three-phase system by means
of a "Scott connection." If p be the ratio of transformation of the trans-
formers used and m the portion of "teaser" transformer used, prove that
in which Eat and Ea2 are the positive- and negativc-sequence voltages to
neutral on the three-phase side and E,z:~ and E., 2 are the corresponding values
to the two-phase side.
2. Devise a metering scheme t0 measurc the zero- and positive-scquence
voltages of a four-phase system.
3. Devise a metering scheme to measure the zero- and positive-sequence
voltages of a six-phase systcm.
4. A three-phase generator whose impedances Zt, Z2, and Zo areknownis
loaded across two terminBJs by mcans of an autotransformer, by which proc-
ess a single-phase three-wire system is produced, the neutral being taken
off midway of the transformer. Determine the positive- and zero-sequence
impedances of the generator for thc single-phase system in terms of Z 1, Z 2,
and Zo.
6. Given la = 20 + j60 and h = 25 - j10, determine (a) the positive-
and zero-sequence components of a single-phase system, and (b) the irregular
components of a two-phase three-wire system.
CHAPTER XVII
INDUCTION MOTORS
An important application of the theory of symmetrical
components is the analysis of induction-motor operation under
unbalanced conditions. For this purpose a symmetrically wound
machine will be considered. It can readily be understood
without discussion that in such a machine there is no reaction
between the different sequence quantities. Being symmetrical,
it is well known that if balanced voltages of a certain sequence
be applied to the stator and rotor, only balanced currents of
the same sequence, regardless of slip, will flow. The operation
of induction motors under three different conditions will be
discussed: (1) application of unbalanced voltages to the stator
of a symmetrical machine, (2) one phase of the stator of a
symmetrical machine open-circuited, and (3) unbalanced imped-
ances in the rotor of an otherwise symmetrical machine. Before
discussing differcnt unbalanced conditions of operation, consider
the characteristics of the motor for positive-sequence voltages
and currents.
155. Characteristics with Application of Positive-sequence
Voltages to the Stator.
Assume a balanced set of positive-sequence voltages applied
to the terminals of an induction motor having a certain shaft
load. The rotor will rotate in the positive sense (same direction
as rotating field) with slip s expressed as a fraction. The char-
acteristics can be best defined in terms of the well-known equiva-
lent diagram. In this development it will be assumed that the
rotor has the same number of turns as the stator or has been
reduced to the equivalent so that the rotor constants and currents
and voltages may be expressed on the same turns basis as the
stator. Let
R11 = resistance per phase of the stator
Xa = leakage reactance per phase of the stator at supply fre-
quency.
345
Rr = resistance per phase of the rotor.

Xr = leakage reactance per phase of the rotor at supply
frequency.
I.t = positive-sequence current in t he stator referred to
phase a.
1 rl = positive-sequence current in the rotor referrcd to phase a.
E . t = positive-sequence line-to-neutral voltage in the stator
referred to phase a.
The application of the voltage E.L to the stator sets up a.
field which rotates with synchronous velocity. This field
cuts the rotor windings and sets up therein a current l r1 of slip
frequency, which produces a field that ro.tates with slip velocity
relative to the rotor. But since the rotor revolves at a speed
(1 - s) times synchronous velocity, the field set up by the
current lrt !otates also with synchronous velocity relative to
the stator.
Designate by E 0 1 the voltage induced in the stator by the
resultant synchronously rotating field produced by 1.1 and lrt·
The applied voltage E.1 must be equal to the sum of this voltage
and the resistance and leakage reactance drops in the stator, or
Esl = E ol + (Rs + jXa)l al (388)
The voltage induced in the rotor must be proportional to
the slip, or sE0 1. It is this voltage which gives rise to the current
lrt of slip frequency. Therefore
sE 0 1 = (Rr + j sXr)l rt (389)
where the leakage reactance Xr is measured at supply frequency.
Dividing equation (389) by s
E,,= e·+ jX},, (390)
The air-gap voltage E 0 1 is produced as the resultant of / 81

and Iri, and if the positive direction of flow of lsl be considered
magnetizing and Irl demagnetizing, then E 0 1 is proportional to
(/. 1 - I ri) . Expressing this proportionality by means of the
magnetizing reactance jX m, there results
Eol = jXm(lsl - l rl) (391)
The three conditions expressed analytically by equations
(388), (390), and (391), are fulfilled by the equivalent circuit
INDUCTION MOTORS 347
for positive-sequence, shown in Fig. 191 (a). It follows, there-
fore, that the solution of this network gives the characteristics
of the induction
motor. It is con-
venient to split
up the resistance
Rrm
- . t o t wo com-
s
ponents Rr and
1- S
--Rn as shown
S
in Fig. 191(b) .
By this means the
power absorbed
by Rr represents
the copper loss,
and, since this is
the only rotor loss
considered, the
power absorbed Fro. 191.- Equivalent circuits of induction motors
1 - S (positive-sequence).
by Rrrepre-
s
sents the shaft power output. Therefore, calling P1 the total
shaft power,
(392)
This is equal to the shaft torque T, times the speed, so that
T 1_- K3Rr
- 1- 2rl (393)
S
in which K is a constant dependent upon the units. When

3Rri 2rt is expressed in watts, and T in pounds at 1 foot,
K= 33,000 33,000p
(7 46) 27r(r. p.m.) synchronoua (7 46)27rf60
0.117p
- (394)
f
where pis the number of pairs of poles andjthe supply frequency.
A usual simplifying assumption is to place the shunt mag-
netizing reactance jX m directly across the terminals as shown in
Fig. 191(c). In the above development no attempt has been

made to include friction, windage, or core losses. For more
detailed discussion any standard work may be consulted.
If externa! impedance is connected in the rotor circuit its
effect can be included by inserting the equivalent resistance
Rs jXs j(Xr+Xext) Rr+Rext
~·~ Represenfs
shaft/oad
F10. 192.- Inclusion of extcrnal rotor resistance (R • .,1) and reactance (Xeze) in
equivalent circuit (positive-sequence).
and reactance (at supply frequency) in the rotor branch of

the equivalent diagram. This is illustrated in Fig. 192.
166. Characteristics with Application of N egative-sequence

Voltages to the Stator.
The application of negative-sequence voltage is in effect
the same as interchanging the leads to two terminals. A field
is set up which rotates at synchro-
2
nous speed in the direction opposite
~~ to that set up by the positive-sequence
!"'- voltages. It follows then that the
" f'.
f'.
1"-
characteristics for the negative-
sequence· may be obtained by merely
selecting the appropriate value of slip
o
" ~ to use in conjunction with the positive-
f"\. sequence equivalent circuit. F igure
2
S 193 shows the relation between t he
Fw. 193.-Relation be- negative-sequence slip s2, and the posi-
tween positive- and negative-
sequence slips.
tive-sequence slip s. For very small
values of s, s2 is very nearly equal to
2 ; and, for the rotor blocked, both s and s2 are equal to unity.
The equation relating the slips is
S2 = 2 - S (395)
'fhe equivalent circuit for the negative-sequence is obtained
by replacing s in the positive-sequence network by s2 from
~quation (395), as is shown in Fig. 194.
The shaft power of negative-sequence is obtained by calculating
1-s 1-s -
the power absorbed in - Rr, namely -3 Rrl 2r2 •
2 -S 2 -S
The negative-sequence torque T2 is
2
T2 = -K3Rr1 r2 (396)
2-s
The negative sign indicates that for small slips the torque is
opposite to that for
positive-seq uence
voltages; hence, for
small slips, the load
supplies power to
the motor through
Fro. 194.- Ncgativc-scquence equivalcnt circuit for an
the shaft. induction motor.
Since Rr + jXr
does not usually exceed 15 per cent and jX m is of the order of
400 per cent, jXm may usually be omittcd. In addition,
1- R
-
2
S
_ sRr may be replaced by - -j,
so that the equivalent circuit may be
_R,.. reduced to that shown in Fig. 195.
2
157. Characteristics with Application
Fw. 195.- Approximatc of Both Positive- and N egative-
negativc-sequcnce equivalent
circuit for an induction motor sequence Voltages to the Stator.
operating at small values of When both positive- and negative-
slip.
sequence voltages are applied simul-
taneously, the currents of the two sequences do not react upon
each other; thus they may be computed separately by the
methods discussed. The total shaft power output may therefore
be written
1 - S - 1 - S -2
P total shaft output = 3 Rrf2rl - 3 Rrf r2 (397)
S 2 -S
and the total shaft torque
T = 3KRr(J2rl - ]2r2 ) (398)
S 2- S
The significance of the power, torque, and the loss relations
will be made clearer by the consideration of the characteristics
of a particular machine.
158. Application of Unbalanced Voltages to the Stator of a

Particular Motor.
Gonsider the case in which both positive- and negative-
sequence voltages are applied to the stator of an induction
Per Cent Rottirlg +
200 tOO O 100 200 300
(a) P ositive-Sequence
2.0
o
1
1
'
1
1
1
1 ~1.0
1~
1
(b) Negative- /
Sequence L
,,... _Torque
/
/
,..,., _,.,./
~----- _____ ....
2.01--:::::::..__ _ _ _ _ _ __
Fw. 196.-Induction-motor characteristics for 100 per cent applied voltage.
motor. In particular, consider a 500-hp. 2,200-volt 25-cycle

motor, the constants of which are
Xr = Xs = 0.9 ohm per phase
Rr = Rs = 0.2 ohm per phase
X m = 40 ohms per phase
The characteristics of this machine for 100 per cent positive-
and negative-sequence applied voltages are shown in Fig. 196(a)
and (b), respectively, in which shaft output and loss are expressed
in per cent of rating, and torque in per cent of the torque at
rated speed. It will be observed that the negative-sequence
characteristics have the same shape as the positive-sequence
r' characteristics except that they are inverted in various ways.
Assume that 100 per cent positive-sequence and 10 per cent
negative-sequence voltages be applied simultaneously, with
the load such that the slip is 1.5 per cent. The torque due to
the positive-sequence voltage is 93 per cent of normal in a
positive sense, a fact which may be verified approximately from
Fig. 196(a). The torque due to the negative-sequence voltage
(since the torque varies as the square of the applied voltage)
is T~~lí X 40 or 0.40 per cent of the normal positive-sequence
torque but in the negative direction. It may be seen, therefore,
that the unbalance of voltages that may occur in service will
generally have a negligible effect upon the torque characteristics.
The loss due to the positive-sequence set of voltages is 2.8
per cent, and that due to the negative-sequence voltage is
Tllí X 155 or 1.55 per cent. This loss is quite appreciable
when compared with the loss upon which the rating of the
machine is based. It follows, therefore, that unbalanced voltages
will cause trouble due to heating if the negative-sequence
voltage is suffi.ciently large.
169. Operation with Open-circuited Stator Phase.
An interesting study of induction machines is the phenomena
that occur when one phase is accidentally opened, as by the
blowing of a fuse in one lead. The treatment is the same whether
the machine is star- or delta-connected; in the latter case it is
necessary merely to replace the delta connection by its equivalent
star. This is a special case of unbalanced applied voltages.
Assuming the a phase to be open, there result from the current
relations
la= O
h = I
le=-[
35:t SYMMETRICAL COMPONENTS
or
1
l,1 = (o +al - a 21) (400)
3
1
la2 = 3(0 + a 21 - al) (401)
The only voltage known is the voltage impressed between ter-

minals b and e, namely EsA· This quantity may be constructed
from the line-to-neutral components in phase a as follows:
E sb = a 2Est + aEs2
Ese = aEst + a Es22
Therefore
EsA = Ese - Esb = (a - a 2)(Ed - Es2) = +jy'3(Est - Ea2)
(402)
P 'Rs jXs jXr
Shaft ouf-pvf
~
1-S
Rs jXs jXr Rr - z-s Rr S
k------------------E$2···-·········--·- · ·--·-~-~=---········--·-··--~
Fm. 197.-Equivalent circuit of an induction motor with one stator phase
open. Conditions to be satisfied are that Ist = -1.2 and Es1 = -E.2 =
E ..A
iv3·
In Fig. 197 the upper half of the circuit represents the positive-
sequence circuit lFig. 191 (b)] redra wn and the lower half the
negative-sequence circuit (Fig. 194). By connecting Q and S,
the condition expressed by equation (401) that ls2 = -1st
is fulfilled. In addition, the voltage across PR is E at - Es2·
But from (402) this voltage is also equal to j~· It follows,
therefore, that if the voltage j~ be applied to the circuit
shown in Fig. 197 the positive- and negative-sequence currents

will be the currents that flow in the respective circuits.
The total shaft output and net torque are then obtained
from equations (397) and (398).
A customary approximation in induction-motor theory is to
transfer the magnetizing current branch jX m from across the
portion of the circuit representing t he rotor to across the source.
A similar approximation will be made in this case; the branch
jXm of the positive-sequence network will be placed across the
source, that is, across the
.:¡~:
terminals P R. Since all
the impedances in the Es1-Es2T.E:2(Rs+Rr) j2(Xs+Xr) 'V\/V\1'
___,..-
negative-sequence net- EsA or ·x
J m
Ir
work but jX m are of the ·1{3 .
order of leakage react- JL~~----------------------~

Fictitious resistance of shaft output =
ances, the magnetizing 1 - S 1 - S) 2 ( 1 - S) 2
branch j X m in this net- ( - s
- - --
2 - s r
R = s(2 - s)
Rr
work may be neglected. Fm. 198.- Simplification o f equivalent net-
work of Fig. 197 by placing magnetizing branch
After rnaking these ap- acrosa the terminals.
proxirnations, and corn-
bining, the equivalent circuit becomes that shown in Fig. 198.
The shaft power and torque beco me:
6(1 - s) 2 - 2
Total shaft power output = s( 2 _ s) Rrlr' (403)
6(1 - s) - 1 2
Net torque = K s( 2 _ s)Rrl r (404)
in which K is defined by equation (394).

From equation (404) it will be observed that for small slips
the torque is positive, which interpreted, means that the motor
will continue to operate if the load is not too great. For s = 1
(standstill), the torque is zero and the machine will not start.
Any of the sequence components of voltage or current, in
either the stator or rotor, may be obtained by solving the circuit
of Fig. 197 which is clearly marked to indicate the various
quantities. Having obtained the sequence components, the
indiv idual phase components may readily be computed. Such
characteristics as t he maximum pull-out torque, loss, or heating
·can thus be determined satisfactorily.
160. Unbalanced Impedances in the Rotor Circuits.

Another important practica! problem involving unbalanced
operation of induction motors is that of unbalanced impedances
in the secondary rotor circuits. For the analysis of this problem,
positive-sequence voltage alone will be assumed applied to the
stator terminals of the machine. Currents of fundamental fre-
quency in the stator will give rise to currents of slip frequency in
the rotor. The unbalance in the rotor will, however, cause these
currents to be unbalanced. These unbalanced currents of slip
frequency in the rotor may be resolved into positive- and negative-
sequence currents. The latter sets up fields which rotate nega-
tively with respect to the rotor at slip frequency. These
currents in turn induce currents in the stator of frequency
equal to fundamental frequency minus twice slip frequency.
The plan of a ttack will be to analyze thc reactions in the m achine
into the two groups; first, those associated with fundamental
frequency in the stator and slip frequcncy of positive-sequence
in the rotor, and second, those associated with slip frequency
of negative-sequence in the rotor and fundamental minus twice
slip frequency in the stator. These two groups of reactions
will then be tied together through the characteristics of the
unbalanced impedances in the rotor.
The diagram in Fig. 199 shows schcmatically the induction-
motor circuits under consideration. With the positive-sequence
-JSI In
voltage Esl applied
to the stator, the
resultant air-gap flux
rotating wi th synchro-
b nous velocity induces
Stafor Rotor in the stator the volt-
FIG. 199.-Schematic dia gram of a n induction age E 0 1 and in the
motor with unbalanced impedances in the rotor rotor the voltage sE 1,
0
circuit.
where s is the slip
expressed as a fraction. The impedance drop in the stator then
provides the relation
Esl = Eol + (Rs + JXs)fsl (405)
In the rotor, the voltage equation is
sE0 1 = JsXrlrl + Rrlr1 + Er1 (406)
where Erl equals the rotor terminal voltage of positive-sequence.
The voltage Erl is the voltage at slip frequency that would be
\
:.
' measured at the tcrminals of the rotor by means of a positive-
sequence measuring device. Dividing equation (406) by s
E ol ·xr I rl
= J + -Rr¡rl + -Eri (407)
S S
The air-gap flux is produced by the resultant of the currents

Isi and Irl· Therefore, considering I sl as positive when mag-
netizing, and Irl as positive when demagnetizing, the propor-
tionality between these currents and the voltage produced by
the air-gap flux may be expressed by
E"1 = jXm(Isl - Irl) (408)
From equations (405), (407), and (408), the equivalent diagram
of Fig. 200(a) can be obtained just as that of Fig. 191(a) was
obtained previously. The approximate diagram is shown in
Fig. 200(b). From this circuit it may be seen that
E;l = E sl - [ R 8 + ~r + j(Xs + Xr) ]Jrl

or
(409)
The development of the equivalent circuit for the negative-
sequence may best be approached by assuming a negative-
sequence voltage Er2 of
frequcncy sf applied to Is1
Rr
the rotor and assuming ,
E;.,
the rotor to be stationary ~s• jXm
T1
'
and the stator rotating
(01)
with speed (1 - s) in the
negative direction. This
gives the same relative
condition as though the
stator were stationary and (b)
the rotor rotating with 1 _ 8( Ert) _
speed (1-s). The ap- Shaftpower=3- s - Rr+realpartlrl I 2rl
plied voltage sets up a flux 1 -S
= 3- - (RrPrl + p o 8 i ti V e-
-
S
which rotates with veloc- sequence power output of rotor)
ity - s, and, since the Fw. 200.-Equivalent positive-sequence dia-
assumed velocity of the grama for an induction motor with voltage
Ert on the rotor. (a) Accurate; (b)
,
. stator is - (1 - s), these Approximate .
velocities will be equal
for s = 0.5. For this slip, no voltages are induced in the stator
and the characteristics correspond to that of an ordinary induc-
tion motor operating at synchronous speed. For s < 0.5

the negative velocity of the stator is greater than the negative
velocity of the flux set up by the negative-sequence voltage
and the machine therefore acts like an induction generator, but
for s > 0.5 the negative velocity of the stator is less than the
negative velocity of the flux set up by the negative-sequence
voltage and the machine acts like an induction motor.
With voltage Er2 applied to the rotor and for the present
considering current flowing out of the rotor as positive (as
indicated in Fig. 199), the voltage equation for the rotor may
be written
(410)
where I r2 = negative-sequence rotor current.
E02 = voltage induced in the rotor due to magnetic flux
associated with the ncgativc-sequence reactions
within the machine.
1-S
- t·2.S Rs
1 -S -
Shaft powcr = -3 _ R.f2.2
1 28
Fw. 201.- Equivalcnt diagram for negativc-sequcnce currents in the rotor of
slip frequency voltagcs of frequency s applied to the rotor of an induction motor
operating at speed (1 - s) and stator short-circuited.
The speed of the flux which produces E 0 2 is, relative to the

rotor, of course, -s and, with respect to the stator, (1 - 2s).
-e
It follows, therefore, that this same flux induces the voltage
~ 28
)E,, in the stator windings. . Assuming that the
system impedance to the flow of the stator current of frequency
(1 - 2s)f is zero, this voltage must be equal to the impedance
drop of the stator at the frequency (1 - 2s)f, so that
-C ~ )E,, ~ R.<-!,,) -
28
j(l - 2s)x.< -!,,) (411)
where ls2 is the stator current associated with the negative-

sequence current in the rotor and, to be consistent with previous
assumptions, the direction of flow is considered positive as
show~ in Fig. 201. The sign used in connection with the react~
ance drop term is nega.tion because at small slips the machine is
operating as an induction generator. Dividing through by the
coefficient of Eu2,
Eo2 = l ~ 2 SRJs2 - jsXsls2 (412)

1- S •
- 1 _ 2sRJs2 - Rsls2 - JSXsls2 (413)
Assuming that positive values of l s2 are magnetizing and

positive values of 1,.2 are demagnetizing, E 0 2 is proportional
to Cis2 - I rz). The same factor of proportionality used in the
r,
positive-sequence set of currents may be used here also, but since
Eu2 is measured at slip frequency, the equation for E 0 2 is
(414)
From equations (410), (413), and (414), the equivalent
circuit for the negative-sequence may be shown to be represented
by that of Fig. 201. The power absorbed by the resistance
1- S
_ sRs represents the shaft power output as the total
1 2
power input from the rotor can represent only copper loss and
shaft power. At small slips the quantity is negative which.
indicates that the shaft power output for the negative-sequence
is opposite in sense to that of the positive-sequence. For
s > 0.5 the contrary is true, and the negative- and positive-
sequcnce powers are additive.
The equivalent circuit for negative-sequence voltages of
frequency sf applied to the rotor terminals can also be derived
from that of the positive-sequence network, by again assuming
that the rotor is stationary and that the stator is rotating with
a speed of - (1 - s) in the positive sense. The synchronous
velocity of the rotor is - s, so that the equivalent slip expressed
-s- [ -(1-s)]
as a fraction of the synchronous velocity is or
-s
1 - 2s
This expression when substituted for s in the positive-
S
sequence equivalent circuit of Fig. 191 gives the term - : ~ ; 8 Rs

for the equivalent resistance which corresponds to the shaft
power. Since all of the reactance terms are given for the normal
frequency J, it follows that in this equivalent diagram the react-
ance terms must be multiplied by s since the applied frequency

in this case is sf.
The voltage and current relations for the negative-sequence
network may be written from Fig. 201 as follows:
·X jsXm[ -RB + jX (1 - 2s)] ]
E r2 = 1 r2[Rr + JS r + -R + j(1 - 2s)(X + Xm)
8
8
8
(415)
Having developed the two sequence networks, the tie between

them and the external impedances may now be considered.
For the unsymmetrical impedances Za, Zb, and Zc, of Fig. 199,
the phase voltages and currents at the terminals may be written
Ea - E e = Z al a - Z el e }
Eb - E a = Zblb- Zala (416)
Ec- Eb = Zcl c- Zoh
The phase voltages and currents may then be expressed in terms
of the sequence voltages and currents with the following results :*
Erl = HZa + zb + Z c)Irl + i(Za + a 2Z + aZc)Ir2 (417)
Er2 = i(Za + aZb + a 2Zc)Irl + i(Za + Zb + Z c)lr2 (418)
The sequence voltages at the rotor terminals may be eliminated
between equations (417) and (418) and equations (409) and (415).
The remaining equations may then be solved for the sequence
currents from which the sequence voltages may be determined.
The analytical expressions are cumbersome; consequently, it is
usually more convenient from this point to carry out the calcula-
tions in numerical form. From the sequence currents and
voltages the complete performance of the motor as to torque,
power loss, phase voltages, and line currents may readily be
calculated for various unbalanced externa! impedances.
161. Operation with Open-circuited Rotor Phase.
The operation of an induction motor with an open-circuited
rotor phase may be analyzed as a special case of the problem
discussed in the preceding section making use of the equivalent
circuits of Figs. 200(b) and 201. Assume that the stator of
the motor is connected to an infinite source, and that phase a
of the rotor is open-circuited and that phase b and phase e are
* It will be noted that the impedance coefficients of equations (417) and
(418) are of symmetrical forro similar to that of the sequence voltages and
currents. The reason for this is given in the discussion of unsyrnrnetrical
impedances in Chap. XVIII.
short-circuited at the rotor terminals in the usual manner.
The circuit condition corresponds to that imposed upon the
imaginary leads brought out from a transmission line for a line-
to-line fault, for which it was found that I rl = -Ir2, and
Erl = Er2· It will be observed from the equivalent circuit of
the positive-scqucnce diagram that the rotor voltage is taken
into consideration by inserting the voltage Eri in the secondary
S
t-s
---R
s(l-2s) s
!
t;
(01)
Rs
jXs -T-2$
(b)
1 •
¡
Shaft poweroutput = ( 6
1
~ 8 Rrf2r1- 311 ~ ; 8 R.l2• 2) +( -311 ~; 8
R.f 2, 2)
Shaft torque = 1 -
6K [ -;R,J2,¡ - (
1
_1
28
-
)R.J2.z J
Loss = 3(R. + 2R,)l + 3R.l .z
2,
1
2
FIG. 202.-Equivalent diagrams for open-circuited rotor phase of an induction

motor.
circuit. If the constants of the equivalent diagram for the

1
positive-sequence currents be multiplied by -, then the same
S
currents will flow when Er 2is impressed as though Er2 is impressed

S
across the unchanged diagram. It follows therefore if this
' network b e connected across Erl m
. t h e positive-sequence d ia-
' s
l
gram, as shown in Fig. 202(a), that the two conditions imposed

by the unbalance will be satisfied. The currents indicated
are the actual currents fiowing in the machine. A word of
caution is necessary in connection with this diagram as the
power loss in the lower portion of this diagram does not represent
the loss in the negative-sequcnce diagram. N o advantage is
therefore attached to retaining the identity of the two networks
except that it enab]es one to determine the value of the terminal
voltage of the rotor. The dia!;ram may be simplified to that
shown in Fig. 202(b).
Times Norme~l Torque
2.0 1.0 + 1.0 2.0 3.0 4.0
0..
~ 0.6
0.7
0.8
0.9
FIG. 203.-Torque of a three-phase induction motor with one rotor phase open-
circuited compared with torquc for normal conditions.
Since the voltage applied to the stator is by assumption wholly

positive-sequence, then the total input is determined by deter-
ming the total power input into the two terminals of the network
of Fig. 202(b). This input is of course merely the sum of the
loss associated with the different resistances indicated. Sub-
tracting from this quantity the actual J2R for the positive-
and negative-sequence currents in the stator and rotor gives
the total shaft power output shown in Fig. 202. The shaft
K
torque in turn is obtained by multiplying the term by _
1 8
with the result also indicated in Fig. 202.
The operation of a three-phase induction niotor with an
open-circuited rotor phase has been investigated by B. G. Lamme
who showed that this motor has a number of unusual char-
acteristics. * These characteristics are brought out by the
consideration of a particular case for which the motor with the
characteristics of those given in Sec. 158 will be assumed. The
motor torque and shaft output are plotted as functions of speed
in Fig. 203 which shows the curves for an open-circuited rotor
r phase and also for the usual balanced rotor winding. It will
be noted that at half speed the motor is capable of developing
t. very little output, in fact, under sorne conditions, the motor,
: while capable of starting with a relatively high torque, is not
1'
.
capable of accelerating beyond a speed approximately half
normal. Furthermore the pull-out characteristic of the motor
in the vicinity of half speed may be considerably more than the
pull-out in the vicinity of synchronous speed.
Problems
1. A 50-hp., thrce-phasc, 60-cyclc, 1,150-r.p.m., 440-volt, squirrel-cage
induction motor was tested at rated voltage with the following results.
No-load ampcres pcr t erminal = 15.6. Lockcd rotor current = 313 amp.
Locked rotor watts = 112 kw. Stator rcsistance between two terminals. =
0.280 ohm. Determine thc constants of the equivalent diagram assuming
the magnctizing branch dircctly across the t erminals.
2. A 500-hp., 2,200-volt, 60-cyclc wound-rotor induction motor has
a pplied to it the unbalanced voltages of 2,200, 2,000, and 1,800 volts. The
magnetizing current per phas;:) is 25 amp., the stator resistance per phase is
0.2 ohm, the rotor resistancc in terms of stator 0.15 ohm, the stator and
rotor rcactancc are cach 0.6 ohm per phase in terms of the stator. Neglect~
ing friction , windagc, and iron loss, plot the curves of torque and loss as a
function of slip.
3. The machine of Prob. 2 is operating at rated torque when a fuse blows
in onc phase of the stator. D etermine the current in the sound phases
assuming thc balanced supply voltage of 2,200 volts is maintained.
4. A 1,000-hp. , two-phase, 2,200-volt, 25-cycle induction motor has the
following constants:
Stator resistance = 0.050 ohm per phase
Rotor resistance = 0.025 ohm per phase
Stator reactance = 0.20 ohm per phase
Rotor reactance = 0.11 ohm per phase
. stator
Turns rat1o - t- = 1.47
ro or
Magnetizing current = 50 amp. per terminal.
* Polyphase Induction Motor with Single-Phase Secondary, Elec. J our.,
September, 1915.
A positive-sequence voltage of 2,000 +jO and a negative-sequence voltage

of O + j300 are applied to the terminals. Determine copper loss and stator
currents. Neglect supply system impedance.
5. A 1,000-hp., 25-cycle, 2,200-volt induction motor has the following
constants referred to the stator:
Stator resistance = 0.07,5 ohm per phase
Stator reactance = 0.30 ohm per phase
Rotor resistance = 0.10 ohm per phase
Rotor reactance 0.35 ohm per phase
Magnetizing branch = 20 ohms per phase
The rotor is comiected Y. With resistors of 0.5, 0.5, and 1.0 ohm in the
rotor circuit, what is the rcsultant specd-torque curve with 2,200 volts
applied to the stator? Neglect the supply system impedance.
6. The motor of Prob. 5 has two rotor phases short-circuited and the
third opcn-circuited. Calculate the speed-torque curve.
7. Determine the analytical expressions for the shaft torque due to the
positive- and negative-sequence currents for an induction motor with one
rotor phase open-circuited. Show from these that the total shaft torque
eq11als that given in Fig. 202.
CHAPTER XVIII
UNSYMMETRICAL SYSTEMS AND PHASE-BALANCERS
In the preceding chapters consideration has been given only

to systems in which the constants are symmetrical, i.e., to
systems whose constants are the same viewed from any phase.
In these systems the voltage drops and current distributions
in any one sequence network are not influenced by the currents
or voltages in the two other sequence networks. This proposi-
tion is not true for unsymmetrical or unbalanced systems; for
them the currents of one sequence may give rise to currents or
voltages of other sequences. The purpose of the present chapter
is twofold: (1) to discuss the treatment of unsymmetrical systems
and (2) to consider the problems of unbalance and apparatus
for its correction.
162. Voltage Drops in Three Unsymmetrical Self lmpedances.
The first case to be considered is the determination of the
voltage drops produced by cur-
ZOI
rents flowing through three unsym-
metrical impedances. As su m e
three impedances Za, Zb, and Zc
connected in star to a neutral wire 1 ~ •
as shown in Fig. 204. Let cur- E" Eb E~
rents l a, lb, and l e flow in the __:.1 _,_~__,_l_ _ _ _ _ _ _ _..J
respective phases. Considering the Fw. 204.- Three unsymme trical

impedances.
individual phascs, the equations
expressing the relations between currents and voltages are
Ea = Zala
Eb = Zblb }(419)
E c = Z cl c
The currents may be resolved into the three symmetrical
components, the equations for which may be written
1 a = lao + 1 al + 1 a2
lb = l ao + a 2l al + alaz
le = lao +alal + a 21 a2
363
Substituting the values oí currents from equations (420) in

equations (419) gives
•
Ea = (Za)I ao + (Za)Ial + (Za)I a2 }
Eb = (Zb)Iao + (a 2Zb)I + (aZb)Ia2
al (421)
Ec = (Zc)lao + (aZc)lal + (a 2Zc)la2
As shown in Chap. II, the voltages Ea, Eb, and Ec, may be
resolved into their symmetrical components [equations (27),
(28), and (29)], as follows:
Eao = !(Ea+ Eb + Ec)
Eal = !(Ea + aEb + a 2Ec)
Ea2 = !(Ea + a 2Eb + aEc)
The various symmetrical components of voltages may now be
expressed in terms of impedance drops by substituting the values
of Ea, Eb, and E c, from equations (421) in equations (27) to (29)
with the following results:
Eao = t(Za + a 2Zb + aZc) I al +HZa+ aZb + a 2Zc)Ia2
+ HZa + zb + Zc)I ao (422)
Ea! = HZa + zb + Zc)I al + HZa + a Zq + aZc)I a2 2
+ i(Za + aZb + a Zc)Iao 2

(423)
Ea2 = HZa + aZb + a Zc)I + i(Za + Zb + Zc)I
2 al a2
+ i(Za + a2Zb + aZc)Iao (424)

Examination of these expressions shows that there are only
three different coefficients involved, which are in reality the
zero-, positive-, and negative-sequence components of the three
impedances Z a, Zb, and Zc, formed in a manner analogous to
that for currents and voltages. Therefore, let
3 o = t(Za + Zb + Zc) (425)
3t = i(Za + aZb + a Zc) 2
(426)
c82 = i(Za + a Zb + aZc)
2
(427)
It is to be noted that the impedance tcrms 3o, c81, and Bz have
a different significance from the Zo, Z1, and Z2 terms used
throughout this volume. The script 3 has been reserved for
this special use. The sequence components of the voltages and
impedance drops may be written
Eao = 3olao + 3zlai + 3da2 (428)
Ea!= 3dao +Bola! +B2Ia2 (429)
Ea2 = 3zlao + 3tfal + 3ola2 (430)
UNSYMMETRICAL SYSTEMS AND PHASE-BALANCERS 365
It will be noted from these equations that each sequence com-

ponent of current in the general case gives rise to voltage drops
of all three sequences.
These expressions may readily be remembered by observing
a relation between subscripts indicating the sequence for voltage,
current, and impedance terms. This relation is that the sum
of the sequence subscripts for impedance and current in imped-
ance-drop terms gives O o.r 3 for zero-sequence voltage, 1 or 4
for positive-sequence voltage, and 2 for negative-sequence
voltage. Thus the im pedance drops :Bol al, :8 d a2, :Bol a2 and
:8da1 are terms for the sequence voltages E a!, E al, E a2, and Eao,
respectively.
163. Discussion of the Impedances :Bo, :81, and :82 of Unsymmetri-
cal Self Impedances.
A clearer conception of the meaning of the impedance terms
;[30 , :81, and :82 of unsymmetrical systems may be obtained by
considering a few special cases. Consider for a moment the
symmetrical system in which
Za = Zb = Zc = Z
Then, from equations (425) to (427),
:Bo = Z
:81 = o
:82 =o
Substituting in equations (428) to (430)
E ao = Ola! + Ola2 + Zlao
E a! = Zlal + Ola2 + Olao
E a2 = Ola! +
Zla2 Olao +
t(431)
These equations show that, for this special case, only voltages
of the same sequences are set up by t he individual sequence
currents. This conclusion checks common experience in this
regard, and the analysis shows that for the symmetrical system,
only the zero-sequence impedance terms are present.
Now consider the case in which the three impedances form
in themselves a positive-sequence impedance set, in which
Za = Z
Zb = a 2 Z ~ (432)
Z c = aZ
This requires in the impedances both positive and negative

reactances and resistanc~s. The physical set up of such imped-
ances offers sorne difficulty, for while negative reactances may
be represented by condensers, sorne special devices such as
vacuum tubes or series commutator machines are necessary
to represent negative resistance. These ideas are merely intro-
duced to clear up the conceptions of unbalanced impedances.
In this case
S o = i[Z + a Z + aZ] = O
2
81 = i[Z + a(a2Z) + a 2(aZ)] = Z ~ (433)

82 = i[Z + a (a 2Z) + a(aZ)]
2
= O
Substituting in equations (428) to (430), one obtains

Eao = Olal + Zla2 + Olao
Ea1 = Ola1 + Ola2 + Zlao
E a2 = Zla1 + Ola2 + Olao
N ow if zero-sequence currents alone are present, then, from these
equations, it is seen that positive-sequence voltages only are
developed. This is to be expected, for since the zero-sequence
components in all phases are equal, the voltages in phases a, b,
and e, are Zlao, a 2Zlao, and aZiao, respectively. These form a
positive-sequence. In this manner positive-sequence voltages
are obtained from zero-sequence currents.
N ow if the positive-sequence currents alone are made to flow
through the three impedances, negative-sequence components
alone are set up. Considering the drops in the phases indi-
vidually, it will be seen that the drop in phase a is ZI; in phase
b, (a 2l 1) (a2Z) = aZl 1; and in phase e, (al 1) (aZ) = a 2 Zl 1•
Thus these voltages form a negative-sequence.
164. General Static Network Containing Unsymmetrical Self
and Mutual Impedances.
The most general case of a static network is that in which all
the self and mutual impedances are of different values. For
the analysis of this case, consider the circuit shown in Fig.
205. It is desired to express the impressed voltages in ·terms
of the symmetrical components of current and the self and
mutual impedances. Let the currents la, h, and l e fiow in the
three phases, respectively, and In in the neutral (or ground)
return. Also, let the self impedances be Zaa, Zbb, Z cc, and Znn·
In addition, let Zab be the mutual impedance between phase&

a and b, Zbc that between phases b and e, and Z ca that between
phases e and a; similarly, ~ Z 01CII
let the mutual impedances r-~---...
between the conductors and rb
"""""+
z<:1b Zcxc zbb Z01n
neutral (or ground) return
be Z an, Zbn, and Z cn· The
equations relating phase-
to-neutral v o 1t a g es and E01 Eb Ec
phase currents and the self --'-_._....__ _-J\
and mutual impedances may Fw. 205.-General static network.
be written from inspection of Fig. 205 as follows :
Ea = (Zaa - Zan)la + (Zab - Zbn)Jb + (Zac - Z cn)lc
+ 3(Znn - Z nn)lo (435)
Eb - (Zab - Z an) Ia + (Zbb - Zbn)Ib + (Zbc - Z cn)Ic
+ 3 (Z nn - Zbn)l o (436)
E c - (Zac - Zan)l a + (Zbc - Z bn) Jb + (Z cc - Z cn)lc
+ 3 (Z nn - Z cn) l o (437)
After substituting sequence voltages and currents for phase
voltages and currents in equations (435) to (437) these equations
may be solved for the sequence voltages. The result ing answer
is rather compbx and in order to simplify the expressions, let
.8aa0 =
Zaa +
Z bb + Zcc
(438)
3
.f Z aa + aZbb + a 2Z cc
i
.8 aal ·-
-
3
(439)
2
Z aa + a Zbb + aZcc
t
l .8aa2 = (440)
1 3
).
~
Zbc + Z ca + Zab
.BbcO = (441)
3
Zbc + aZca + a2 Z ab
.Bbcl = (442)
3
2
Zbc + a Z ca + aZab
.8bc2 = (443)
3
Z an + Z bn + Z cn
.8an0 = (444)
3
Z an. + aZbn + a 2Zcn.
.8an1 = (445)
3
Z an + a Zbn + aZcn
2
- .8an2 = (446)
3
The solution for the sequence voltages may then be expressed as

Eo = L8aa0 + 2(3bc0 - 33ano) + 33nn]lo
+ [3aa2 - 3bc2 - 33an2]11 + L8aa1 - 3bcl - 33anl]l 2 ( 447)
E1 = [3Cia1 - 3bc1 - 33an1]lo + [3aa0 - 3bco]l 1
+ [3aa2 + 23bc2]1 2 (448)
E2 = [3aa2 - 3bc2 - 33an2]1O + L8aa1 + 23bcl]l 1
+ [3aa0 - 3bco]l 2 (449)
It will be noted that only six separate impedance constants
are involved for this general case of the unsymmetrical static
three-phase network. The various constants rnay be grouped
as follows:
Zero-sequence impedance terms, type 3o
3aao + 2(3bco - 33ano) + 3Znn =

Eo
lo (cornponent) (450)
= ~: (component) (451)
E2
= (component) (452)
12
Positive-sequence impedance terms, type 131
3aa1 - 3bcl - 33an1 =

Eo (component) (453)
12
3aal - 3bc1 - 33an1 E1 (
= lo component) (454)
3aal + 23bc1 =
E2
l (component) (455)
1
N egative-sequence impedance terms, type 132
=
Eo (component) (456)
11
= E1 (component) (457)
12
E~
= lo
(cornponent) (458)
In symmetrical static networks the positive- and negative-

sequence impedance terrns of the type 131 and 132 corresponding
to equations (453) to (458) are all zero. Thus the only terms
left are the zero-sequence terms of the type 3o corresponding to
equations (450) to (452). These zero-sequence impedance
terms correspond to Z 0 , Z1, and Z2, in the notation used through-

out this volume. It will be observed that the positive- and
negativc-sequence impedances z 1 and z2 of symmetrical static
networks are always equal to each other as shown by the identity
of equations (451) and (452).
165. Symmetrical Transformations.
In the prcceding section it was shown that for symmetrical
static networks only the zero-sequence impedance terms of the
type Soare present. Also, it was shown that the sequences do
not react u pon each other; in other words, positive-sequence
----
Ic
e
>.
Ior
--
~
1~ 01
e'
b'
- lb ---
Fw. 206.-Interconnected symmetrical booster transformar.
lb
b
currents produce only positive-sequence voltage drops, and

similarly for the other sequences. It is of interest, however,
to considcr the case of static networks involving symmetrical
transformations which is discussed in this section.
A type of transformation that is occasionally encountered
is that shown in Fig. 206 in which both turns ratio and phase
angle are altcred by interconnecting the phases in the same
sequence as the voltages. By this means, both the power and
reactive volt-ampere loading for balanced conditions may be
controlled. Questions, however, arise asto (1) the performance
of such a combination under unbalanced conditions, and (2) the

· introduction of the characteristics of such transformations into
single-line diagrams of the equivalent network.
The transformation may be assumed to involve three single-
phase transformers, each of which possesses a series and a
shunt winding. For the present assume the transformers to be
ideal to the extent that exciting currents and series impedances
are both zero. In short-circuit studies it is common practice to
ignore the exciting current; the transformer impedances will
be considered subsequently. Calling n the turns ratio between
the series and shunt windings, it will be observed from the
lower right-hand leg of Fig. 206, that the voltage from a' to a
is n times the voltage from O to b' or is equal to nEb'. The
voltage of a may then be written as the sum of the voltages
between O anda', anda' anda; or,
Ea = Ea' + nEb'
Similarly,
(459)
and
Ec = Ec' + nEa'
Or
Ea' + nEb' = Ea (460)
Eb' + nEc' = Eb (461)
Ec' + nEa' = Ec (462)
By adding equations (460), (461) and (462) it may be seen
that
(463)
Multiplying equation (460) J:>y 1, (461) by a, and (462) by a 2, and

adding: ~
(464)
Similarly, multiplying (460) by 1, (461) by a2 , and (462) by a,

and adding:
Ea2
1
=
1 ~ naEa2 (465)
For the upper leg, the current through the series winding,
le, produces a demagnetizing e:ffect which to annul requires the
current nl~; in the shunt winding of the same leg. Equatin.t.
UNSYMMETRICAL SYSTEMS AND Pll ASE-BALANCERS 371
the currents about the junction point of this shunt winding gives
the relation
la' = l a + nlc
Similarly,
lb' - l b + nla
and
le' - l e + nlb
Operating on these equations in the same manner as for equations
(460) to (462) there result
1
lao - (1 + n)lao (466)
lal
1
- (1 + na)lal (467)
la2 ' - (1 + na )la2 2
(468)
Thus it is seen that the different sequence networks do not
react upon each other; the equations relating current and voltage
on the two sides involve only terms of the same sequence.
Either side of the transformer may be chosen as a reference
base. If the right-hand side be chosen as the reference voltage,
the currents and voltages on that side will be the actual quantities
but the actual voltages and currents on the left-hand side are
related to the base voltages and currents by the equations
1
Eao (actual) =
1 +1 n
E '
aO (base)
,, Ea/ (actual) - 1 +1 na2E al, (base) (469)
•'
~
1
Ea2 (actual) - 1 +1 naE ,
a2 (base)
~
f and
1
1 a0 (actual) - (1 + n)la o (base) 1
1
lal (actual) - 0 + na)Ial (base> 1
}(470)
(1 + na )1a2 (base) 1
1 2
la2 (actual) -
Equivalent Circuit for Symmetrical Transformations. To
complete the analysis of the symmetrical transformation, it is
necessary to determine the equivalent circuit taking into account
the presence of an impedance on either side of the series winding.
If the positive-sequence impedance Z1'(actual impedance) be inserted
in the left-hand side, and the positive-sequence current 1 a/ (actual)
be made to flow through it, the actual drop across the impedance
is Z /(actual> 1 al' <actual) . In te-:rms of the base voltage from equation
(469) this is equal to (1 + na 2)Zl'cactuaz> I ai'cactuaz>, and in terms

of the base current from equation (470) this equals
(1 + na)(l + na 2)Zl'cactuaz>la/ cbase>·
This voltage must be equal to Z/ ccquiv.) Ial'(base), so that the
equivalent impedance for positive-sequence is
Z1' cequiv.) = (1 + na)(1 + na 2 )ZI'cactual) (471)
The two terms in the parentheses are conjugates of each other,
so that their product is a real number whose magnitude is the
square of the magnitude of either one. The same expression
is obtained for negative-sequence and (1 + n)2 is obtained for
zero-sequence. Thus
Z2' cequiv.) - (1 + na)(1 + na )Zz'cartua z>
2
(472)
and
Zo'cequiv.) = (1 + n) Zo' cactuaz>
2 (473)
To include the effect of transformer impedance, the series
impedance Zt between the series or booster windings and the
shunt winding on any leg should be determincd in terms of the
turns of the series winding as the base. Thus the impedance
will be expressed directly in terms of the right-hand side. The
impedances may thcn be combined in the same manner as any
other externa! impedance of the same value.
Thus the transformer itself may be replaced by a series
impedance Z t in each of the sequence networks; impedances on
right-hand side by their actual values, and the impedances on
the left-hand side by values defined by equations (471), (472),
and (473). To obtain actual currents and voltages in the left-
hand side from thc equivalent voltages and currents, use the
relations defi:Jfed by equations (469) and (470). Voltages and
currents on the right-hand .side are already on their own base.
It is interesting to note the similarity of these rcsults with those
of the special case of the star-delta transformation defined in
Chap. II, Figs. 6 and 7, in which the positive- and negativc-
sequence quantities in passing through the transformer were
rotated in opposite directions.
166. General Case of Unsymmetrical Systems.

Having indicated the general voltage, current, and impedancc
relations that exist in general static networks, it is now prac-
ticable to consider the general case of the unsymmetrical system
involving rotating machines. The basic equations are similar
in form to equations (428) to (430) except that the nine imped-
ance coefficients may be different in value. To take care of this,
the notation for the impedance coefficients Bo, B1, and B2 will
be modified by the addition of a subscript to indicate the current
term with which they are to be used. Thus Bo1 indicates the
zero-sequence impedance term to be used with positive-sequence
current. The general equations for the unsymmetrical system
may be written from comparison with equations (428) to (430)
as follows:
Eao = Boolao + B21la1 + B12la2 (474)
Eal = Btolao + Badal + B2da2 (475)
E a'!. = Bzolao + Bnlal + Bodaz (476)
The determination of the impedance coefficients of equations
(474) to (476) may readily be accomplished by finding the
sequcnce components of voltage produced by the circulation
of unit current of the proper sequence. Thus Bao, B10, and B2o
may be obtained by the circulation of unit zero-sequence currcnt;
the zero-, positive-, and negative-sequence voltages produced
are equal to the desired impedances.
The most general network reprcsented by the equations (474)
to (476) requircs nine separate impedances, i.e., the nine imped-
ances have independent values only when no symmetry exists
in the system. If the impedances have the same value for each
of the three phases, then all the positive- and negative-sequence
impedance terms (Bw, Bu, B12, Bzo, Bz1, Bz2) will be zero. Con-
sequently, the number of the separate impedances varíes with
the system under consideration. The number of impedances
may vary from nine for the most general case down to three
equal impedances for the simple case of three identical series or
shunt impedances.
In ordinary commercial rotating machines the positive- and
negative-sequence impedances are zero. This is due to the fact
that these machines are symmetrical from phase to phase.
Consequently, only three impedances are involved, namely,
Bol, :Soz, and Baowhich correspond to zl, Z z, and Zo, in terms of the
notation used in Chaps. I to XVII.
Static systems not infrequently involve unsymmetrical
i.mpedances, principally unsymmetrical coupling between phases.
/
•
In many cases transpositions are introduced to avoid the effects

of such asymmetry.
167. Independence of the Sequence Components in a Symmetri-

cal General N etwork.
It has been pointed out on several occasions that a great
simplification in the analysis by symmetrical components takes
place if the system is symmetrical except for terminal conditions
such as an unbalanced load or fault at a single location. . This
simplification resides in the fact that in the symmetrical parts
of the network the current of one sequence produces impedance
drops of that sequence only. Stated differently, in a symmetrical
network all the positive- and negative-sequence impedance
terms C810, Bn, B12, B2o, B21, B22) are zero. This leaves only the
zero-sequence terms (Bao, Bo1, Ba2). When these relations are
satisfied, the different sequence networks may be set up inde-
pendently as discussed previously, principally in Chaps. III,
IV, XI, and XII.
The relation which must be satisfied to obtain symmetry is
that for every set of self or mutual impedances, the value for
each phase, considering the phases in cyclic order, must be equal.
Thus on a four-phase system the zero-sequence impedance terms
only are present when
Zaa = Zbb = Z cc = Zdd
Zab = Zbc = Z cd = Zaa
Zac = Zba = Zca = Zdb (477)
Z ad = Zba = Z cb = Zdc
Z an = Zbn = Z cn = Zdn
where the notation corresponds to that used in connection with
Fig. 205, extended to include a fourth phase. It is not necessary,
in general, that the coupling ·between all phases be equal though
this is true for the particular case of the three-phase system.
The condition for the independence of the sequences for the
general case of a symmetrical three-phase network will now be
examined. For the three-phase system the general relations
for symmetry as defined in equations (477) reduce to
Z aa = Zbb = Z cc
Z ab = Zbc = Z ca
Zan = Zbn = Zcn
It can be seen from equations (438) to (446) that the positive-

and negative-sequence impedance terms given by equations
(453) to (458) are all equal to zero. There remain only the
zero-sequence impedance terms which relate voltages and
currents of the same sequence. Consequently, the sequences
are independent.
Independence of the sequences is a basic premise for the
simplified method of symmetrical components which has been
used throughout this volume for the calculation of unsymmetrical
loads or faults on systems which otherwise are symmetrical.
Since independence of the sequ.e nces arises because of symmetry,
only the zero-sequence impedance terms require consideration,
namely terms of the type .8o1, .8o2, and 3 oo. Hence for these
~ cases the first subscript is unnecessary and has been omitted
f in the notation used in all the preceding chapters. To prevent
confusion, the script 3 is used only for impedance coeffi.cients
determined according to equations (425) to (427). In all
analyses of symmetrical systems the general equations for the
unsymmetrical system, equations (474) to (476), reduce to the
following:
E ao = Zolo
Eal = Zd1
Ea2 = Zd2
168. Systems Combining Symmetrical and Unsymmetrical

Phase Impedances.
Special problems distinct from the ordinary short-circuit or

stability problem arise in which a portian of the system is
unsymmetrical. The solution of such problems usually hinges
upon the point in the system where the dissymmetry begins.
To illustrate the handling of such problems, the particular case
of a generator feeding an induction motor load and an unbalanced
impedance load will be considered.
Figure 207 shows diagrammatically the system and the
nomenclature to be used. The generator on the left with
impedances Z 0 o, Z 0 1, and Z 0 2 supplies the motor with impedances
Zml and Zm2 and the shunt load with phase impedances Za, Zb,
and Zc. At the junction point the sequence components of
current must add to zero, giving
J oO = J mO + J LO }
l ol= lml+!Lt (480)
J o2 = J m2 + J L2
In general it will be found simpler to solve as much of the sym-
metrical portion of the network as possible and leave the evalua-
z9o
Zgt lgo lLo
Et
Zg2 Igl Ez Iu
Genercdor
~ Eo 1 L2 Za
z
Zc
¡ ! t lmt
1m2
Zmt
Zmz
Fw. 207.-Diagram illustrating method of combining symmetrical and unsym-
metrical network elements.
tion of the unsymmetrical portion as the final step. Following

this suggestion,
lmo = O
E1 •
Jml = Zml
E2
1m2= -
Zm2
Ea
l oo= - -
Zuo
_ E 0 1 - E1 .
1ol - (482)
Zol
Ez
1 2= - -
0 Zu2
in which E o, E 1, and E2 are the sequen ce components of voltage
at the junction point. Substituting in (480), and solving for
the junction point voltages, gives the results
Eo - -Zuol Lo
The junction point voltage can also be expressed in terms

of the shunt-impedance characteristics, giving
Eo =BolLo+ BdLI + BdL2
E1 = BdLo + BoiLI + BzlL2
E2 = BdLo + BdLI + BoiL2
in which Bo, c81, and :82, are defined by equations (425) to (427).
Equating (483) and (484) and combining terms
(Bo + Zoo)ILo + B2IL1 + BdL2 =O
.ihho + (So+ z,~·:zL)h, + s,h, ~ z,, z_;'zm,E,, (485)
s,JLO + .8tlLI +(.so+ z.:·:';Jh, ~ Ü

These expressions provide a sufficient number of equations
from which it is possible to solve for the three unknown quanti-
., ties, I Lo, I Ll, and I L2·
\
169. Unsymmetrical Series Impedances.

This same method of attack may be applied to other problems
involving conditions that may be reduced to unbalanced ter-
minal restraints. E. L. Harder<so> has analyzed the problem
of inserting three impedances of different values in series relation
in the three phases of an otherwise symmetrical system. For
the present, consideration will be given to only the problem in
which the impedance Za is inserted in series with phase a of an
otherwise symmetrical system. A special case of this problem
in which Za is infinite represents the case of an open-circuited
phase, such as may arise from the mechanical failure of a con-
ductor without grounding or the opening of a single pole of a
breaker. The latter example is important because of the atten-
tion that has been given recently to the use of single-pole breakers
to increase the stabi1ity limit of systems.
Figure 208(a) represents the schematic diagram of the problem.
The two points on either side of the impedance Za will be desig-
nated x and y, respectively. From this diagram it may be seen
that
Eza = E¡¡a +Zafa
.Exb = E11b
Ezc = E¡¡c
378 SYMMET.1UCAL COMPONENTS
From these r·elations

Exo = Evo + !Za1a (487)
Exl = Eyl + !Za1a (488)
E x2 = Ey2 + !Za1a (489)
The line currents and consequently the sequence components,
also, are equal at both x and y when the positive sense is taken
as indicated. In the no-
tation for current the sub-
scri pts i n d i e a t i n g the
position have been omit-
ted. E qua t io ns (487),
(488), and (489) show
that the drops in the three
seq uences must equal
Posiflve- sec¡uence nefwork
!Za(1o + 11 + 12). The
y necessary conditions,
therefore, for an equiva-
lent network are that the
currents flowing into the
respective networks at y
Negcdive -sequence ne fwork
must equal the currents
flowing out at x and that
y
the voltage drops in these
networks between x and y
must all equal
Zero-sec¡uence nefwork !Za(1o + 11 + 12).
These conditions are ful-
y filled by the connection;:;
shown in Fig. 208(b).
The special case for one
open -circuited phase is
represented by letting Za
{b)
Fw. 208.-Equivalent cir cuit for insertion of
be infinite, w h i eh is
impedance Z a in phase a. equivalent to omitting the
Za branch. Since x and
y are any two points, by letting the point y be the neutral point
or ground, the case forZa = oo represents a double line-to-ground
fault. It will be seen that the equivalent network then becomes
the same as that for the double line-to-ground fault developed in
Chap. XII.
E. L. Harder has also shown that the equivalent diagram

for two open-circuited phases is that shown in Fig. 209. The
only difference between this diagram and that for only one
phase open-circuited is that for the latter the negative- and
zero-sequence networks are in parallel, and for the former the
negative- and zero-sequence networks are in series. If the three
phases on the y side of the open-circuit were grounded, then
point y in each of the networks would be connected to their
respective zero potential buses and the network reduces to that
for a single line-to-ground fault discussed previously. How-
ever, if the phase conductors on the y side are merely connected
";l y
1 1
01 -+-o-----o-+-
1 1
b -l---<>
1
o-+-
1
e -i--o o-J--
' 1
Phoses b oncl e open
(a)
Posifive- sec¡uence nefwork Negcrf/ve-seguence network Zero-sequence nefwork
(b)
Fw. 209. -Equivalent circuit for phases b and e open.
together and not grounded, then the y points in the positive-

and negative-sequence networks alone are connected to their
zero potential buses.
170. Power Supply to Single-phase Loads.
In commercial power systems the generating stations, the
substations, and the transmission systems are, with but few
exceptions, designed for balanced load. Individual single-phase
loads on distribution systems are allocated to the different
phases in such a manner that substantial balance is achieved.
There are, however, certain applications such as power supply
to alternating-current railways and electric furnaces where the
single-phase load may be so large in comparison with the total
load that it is necessary to give special consideration to the control
of the unbalanced effects of the load. Two general methods of
power supply are applicable.
l. Separate generation for the single-phase loads.

2. Common generation for both balanced and unbalanced
loads with phase-balancing equipment for the latter.
171. Single-phase Generation.
Single-phase power supply is usually obtained from a separate
generator driven by its own prime mover or by a three-phase
60-cycle motor. The single-phase generator differs from the
ordinary three-phase generator for balanced loads principally
in the use of a heavy damper winding to minimize rotor loss due
to double frequency currents and to prevent local heating,
particularly at the end rings in the case of turbogenerators.
Motor-generator sets of the larger sizes supplying single-phase
loads are provided with spring mounting to absorb the pulsations
in power dueto the single-phase loads. Single-phase generators
are commonly designed on a three-phase basis with the third
phase available for emergency operation or replaced by" dummy"
coils.
172. Phase-balancing of Single-phase Loads.
The flow of power in a single-phase circuit pulsates at a fre-
quency equal to twice that of the alternating-current supply.
Consequently, it is readily apparent
that sorne means of energy storage is
necessary in order to convert a single-
phase load with pulsating power to a
balanced load of constant power. In
order to reduce the periodic variation
in power, it is necessary, in general, to
utilize load from sorne other phase on
the system. Failure to recognize the
significance of the pulsating character
of single-phase loads has led to frequent;
proposals io draw equal currcnts from
Single-phase load the different phases. Typical of these
Fw. 210.- Unsound at- schemes is the one shown in Fig. 210
tempt to balance a single- which has one winding reversed and
phase load.
which draws equal currents in the three
phases. A little consideration will show that additional appa-
ratus capacity is required for handling single-phase load but in
spite of this the power remains single-phase in character. Con-
J,
k
¡
..
i sequently, nothing is gained by the use of the three single-phase

' transformcrs anda single transformer is preferable.
In case the load can be subdivided and distributed among
the different phases, the balance of the system is of course
greatly improved. It is not necessary, however, to balance
the loads by using identical impedances in the different phases.
All that is necessary is that the total pulsating power be balanced.
The fact that loads of different power factor on different phases
can produce balanced power is illustrated in the connections of
- I"
E.:.
b- -6 lb 1,
b
101
le
"' lb
Eb
Ií,
(C1) (b)
Fw. 211.-Diagram illustrating method of balancing singlc-phase resistan ce
(furnacc) load by adding capacitor load on one phase.
Fig. 211 which illustrates a two-phase generator supplying a

transformer whose secondary is connected to form a two-phase
three-wire system. With the a and b phases connected to the
single-phase furnace load as illustrated, the vector diagrams will
be as shown in Fig. 2ll(b). By adding a capacitar connected
across phase b, the total current on the supply lines will be
indicated by the terms I a and h which form a balanced system.
~-------~~----~ ,...,, b
b ,""""//
Three-phase o/<éí. ot''</
loads --........ "
'-...:',
....,,..
Single -phase e
load
(01) (b)
Fw. 212.-Mcthod of using unsymmetrical transformcr tapa to improve balance
on a three-phasc circuit.
Another mcthod of improving the balance of systems by

static means is to alter the transformer ratios from their normal
values. For example, considera three-pl}ase generator supplying
the principal load between phases b and e as illustrated in Fig.
212. Assume that voltage is maintained constant across
phases b and e under all load conditions. Then under light
load conditions the voltage triangle abe will be substantially
balanced, as shown in Fig. 212(b). However, under heavy
382 SYMMETRJCAL COMPONENTS
load conditions the unloaded phases will be of higher value

and the voltage triangle is a'bc. Consequently, it is possible
to choose the transformer ratios so that the system is approxi-
mately balanced under an average load condition, being unbal-
anced in one direction under no load and u:n.balanccd in the
other direction at full load. This may be accomplished by
adjusting the taps on transformer windings to give a"bc under
light load.
Another method of balancing the power drawn by unbalanced
loads is illustrated in Fig. 213 which shows an electric furnace
with two electrodes and one hearth connection. The load may
be represented by the star-connected resistance shown in the
diagram. It will be noted, however, that the resistance in series
with the upper electrode
will be of higher value
than the one in series
with the hearth since the
upper electrodes involve
Fw. 213.-Method of using unsymrnctrical
transforrner voltages to balance the load drawn are drop. By distorting
by an electric furnace with two electrodes and the voltage triangle ap-
one hearth connection.
plied to the furnace,
it is possible to obtain balanced power. This has been dem-
onstrated by means of a theoretical investigation Jescr:bed in
the literature. * Either the dclta-star or the star-delta con·
nections may be used to obtain balanced currents and voltages
on the high-voltage side of the transformer assuming the furnace
load consists of are and hearth resistances of different values.
In connection with static methods for phase-balancing, it is to
be realized that if one load is variable, the other loads must
be correspondingly adjusted if balance is to be maintained.
173. Rotating Balancers.

Rotating balancers tend to balance the voltages and currents
on a power system by periodically absorbing and restoring
energy to the system using in this process the energy stored by
the inertia of rotating parts. Thus rotating machines tend to
provide balancing by inherent action and do not require the
* EvANS, R. D., Electric Fumace as a Central Station Load with Particu·
lar Reference to Phase-balancing Systems, Elec. J our., vol. 17, September,
1920.
adjustable feature characteristic of static balancing systems as

discussed in the previous paragraph.
Rotating balancers are of two general types:
l. Negative-sequence e.m.f. generator.
2. Impedance-type balancer.
Balancers may also be classified as to their connection to the
system which may involve either series or shunt connections or
their combinations. The principal types of balancers will be
/ taken up and discussed separately.
1 174. Negative-sequence E.M.F.-type Phase-balancer.
Probably the earliest proposal for obtaining accurate phase-
balancing is that dueto E. F. W. Alexanderson<l) and illustrated
in Fig. 214. The method is based on the idea of generating a
6enemfor
Unbalancee/
load
Ba/ancecl SynchroruK.Js fype

load auxiliary unif
Main unit
FIG. 214.- Negative-sequonce e.m.f. type of phase balancer. (Alexanderson.)
negative-sequence e.m.f. of the proper magnitude and phase

position to cancel the negative-sequence currents due to the
single-phase loads which flow through the generator and other
symmetrical portions of the system. This negative-sequence
e.m.f. is generated i.n the auxiliary machine shown in the diagram
which machine i.s mounted on the same shaft as the main unit
and is provided with excitation in two axes so that the desired
magnitude and phase relation can be controlled. The auxiliary
machine is in series with the main unit which is of the ordinary
synchronous condenser construction except for the heavy damper
windings provided to take care of negative-sequence current.
The main machine draws balanced positive-sequence power

from the system which the auxiliary generator converts to
negative-sequence powcr and supplies to the system, thus
canceling the pulsating component of load in the generator and
other symmetrical parts of the system.
175. Impedance-type Phase-balancers.
The first proposal to use impedance-type balancers was due
to C. L. Fortescue who suggestcd the series impedance balancer
illustrated in Fig. 215. The avxiliary machine in this case
is of t he induction-motor type and is connected so t hat its phase
6enen:rfor
lnducfion type
auxiliary unif
Main unif
FIG. 215.- Series impcdance type of phase balancer. (Fortescue.)
rotation is opposite to that in which it would normally runas an

induction motor. Consequently, the auxiliary machine offers
very low impedance to positive-sequence currents and very high
impedance to negative-sequence currents, and advantage is
taken of this fact. The negative-sequence currents required by
the unbalanced load must, t.hcrcforc, be supplied by the main
unit, which may be of either the induction or synchronous types.
'l'he rating of the auxiliary machine is determined by the imped-
ance drop due to the negative-sequence current fiowing through
the main unit and the positive-sequence current flowing into
the load.
Power-factor correction may be secured by the main units
of either the Fortescue or Alexanderson type of phase-balancer.
•
Due to the fact that single-phase loads are frequently of low

power factor, the phase-balancing unit would normally be
designed tq give power-factor correction as well.
It might be pointed out that the shunt-type balancer requires
automatic adjustments in the voltage regulator to correct for
the change in the unbalanced condition. Consequently, it
will not be so rapid in its action gcnerally as the inherent type
of phase-balancer making use of the impedance principie, such
as is illustrated in Fig. 215. If the single-phase load can be
segregated from the remainder, then the series machine will
have its current rating determined by the positive-sequence
component of that load. If the single-phase load cannot be
segregated from a considerable amount of balanced load, the
shunt-type balancer may be more
attractive.
The shunt impedance balancer
of Fig. 216 is the simplest scheme
which has been proposed for phase-
balancing. This scheme, due to J.
Slepian, uses a synchronous
machine similar to the main unit
of either of the previously described
balancers and in addition, in series
with each phase, a set of capaci-
tors of such value that the imped-
ance to negative-sequence is made
negligible. This arrangement will,
therefore, prevent n e g a ti ve- Fw. 216.-Shunt impedance bal-
ancer with series capacitor.
sequence current from flowing past (Slepian.)
this shunt machine to the genera-
tor or other balanced machines on the system. The scheme
is also inherent in action. The scheme has not been used
commercially but looks promising. The principal problem
involved is in the protection of the series capacitar units at times
of short-circuit. It has been proposed to take care of this by
connecting the capacitors in the circuit through transformers
which would saturate for loads in excess of the normal rating
of the balancer and thus prevent full balancing action, which
greatly relieves the stresses due to short-circuit currents that
would otherwise flow.
176. Phase-converter.
It is not necessary to balance the entire single-phase load in
all cases. For example, it may be desirable to use a generator
without a damper winding in parallel with several machines
equipped wit h damper windings. The amount of the sin-
gle-phase load t hat can be carried by the combination may be
increased greatly by connecting a negative-sequence auxiliary
machine in series with the generator without a damper winding
as illustrated in Fig. 217. A negativc-sequence machine used
with this con-
Genercdor nection is known
No./
as a phase-con-
verter. The
negative-se-
quence current
will be com-
Unbcdcmce</ pelled by the
loacts
auxiliary ma-
chine to fl.ow
wholly (or par-
tially) twough
Fra. 217.-Use of phasc-converter to minimizo negativo- the machine
sequence current flowing through generator No. 2.
equipped with a
damper winding. Thc positive-sequence current, on the other
hand, may be divided between the machines in the ordinary man-
ner according to the excitation and the governor setting of the
prime mover.
177. Balanced Polyphase Voltages from a Single-phase Source.
Fo"r certain applications it is desirable to obtain balanced
polyphase voltages from single-phase sources; t hus, for control
purposes a single-phase voltage may be applied to the polyphase
sequence segregating network of Fig. 166, Chap. XIV, and a
balanced set of positive-sequence voltages will be obtained from
the terminals which are connected to the polyphase relay.
However, rotating machines will in general be required for
practica! schemes for power purposcs. The principal application
at t he present time has been in connection with alternating-
current railway electrification where a single-phase contact
system is desirable, but polyphase voltages simplify the motor
equipment. This has led to t he phase-converter locomotive,
typical connections of which are shown in Fig. 218. The
phase-converter may be of either the induction or synchronous
type; with the latter, power-factor correction also may be
providcd. The phase-converter is not self-starting, but when
operating at normal speed it will provide polyphase voltages for
starting and running the induction motors. The phase-con-
verter will not provide balanced voltages under all conditions
because of its regulation. To overcome this di.fficulty the motor
side of the transformer winding is provided with several taps
so that tap chang-
ing under load can
be used to secure
bettcr vol tage Confacf
balance. Taps sysfem
~
on either side of
one line terminal
and the mid-point Fw. 218.- Schomatic diagram for phase-converter
lucomotives such as are used on the Norfolk and West-
of the transformer ern and Virginian Rail way electrifications. ·
winding are re-
quired to insure balance during both motoring and regenerating
operations.
Problems
1. Onc form of phasing dcvice consists of two lamps and a reactor con-
nccted in star across the three leads of a three-phase system. The lamps
are connected to the a and e phases and the reactor to the b phase. Assum-
ing the linc-to~ine voltage to be 110 volts and the impedance of each lamp
and reactor to be 10 +jO and O + j10 ohms, respectively, determine the
current through the lamps whcn positive-sequence voltage alone and
negative-scquencc voltage alone are applied to the phasing device.
2. Three resistors of 20, 10, and 5 ohms are connected in star to the a, b,
ande phases, respectively. The junction point of the three resistors is free.
Determine thc voltage bctwecn this junetion point and the neutral of the
system voltagc when a positive-scquencc voltage of 100 volts to neutral is
applied to thc rcsistors.
3. A thrcc-clcctrode furnace is supplicd by a 3,500-kva., 2,200- 200-volt,
threc-phasc, 60-cyclc transformer. The primary is star-connected and the
secondary dclta-conncctcd. At a given instant the characteristics of the
furnacc may be rcplaccd by three star-connected resistances of 0.010, 0.010
and 0.002 ohm in the a, b, and e phases, respectively. If the transformer
reactance is 30 per cent, what are the magnitudes of current in the three
electrodes, neglecting transformer resistance and the impedance of the leads?
4. An interconnected booster transformer of the type shown in Fig. 206
is rated at 20,000 kva., three phase, 110 kv., 60 cycle. The rating is deter-
mined by the voltage of the shunt windings and the current-carrying capac-
ity of the series windings, and on this basis the leakage rcactance is 70 per
cent. Assume the ratio of transformation to be 15 per cent and two three-
phase reactors of 12 por cent and 15 per cent (on ~0,000-kva., 110-kv. base)
are insertcd in a series, one on each side of thc transformer. With the voltage
maintaincd at 110 kv. outsidc of the reactors on the lcft-hand side, compute
the fault current for (a) a thrce-phase faul t on the right-hand side; (b) a
line-to-line fault on the right-hand side.
5. A symmetrical thrce-phase gencrator having tho constants Z 1 = j100
per cent, Z 2 = j30 por cont, and Z o = j5 por cent is loaded between ter-
minals and neutral by the following impedanccs, 100 +jO per ccnt for phase
+
a, O j100 per cent for phase b, and 100 + jlOO per cent for phasc c. The
excitation is 120 pcr cont of the no-load voltage. Determine the line-to-
neutral and line-to-line voltages.
6. Two similar un'grounded gcncrators are synchronized with phase a
open and the positive-sequcnco intcrna l voltagcs 30 deg. out of phase.
Determine the b and e phase currents and the synchronizing powcr in terms
of the ratings of thc machines ancl compare with the case for phasc a closed.
Use Z1 = j0.40 and Z2 = j0.25 in per unit and the interna! voltage of unit
magnitudc.
7. Two similar grounded-ncutral gcnerators are paralleled through the
a phase only. Thc constants of the machines are Z 1 = jl.O, Z 2 = j0.30,
and Zo = j0.05. If the positivc-scquence voltagcs are 30 dcg. out of phase,
what are the voltages between the b phase tcrminals ami between the e phase
terminals? Assume the interna! voltage to be of uni t magnitude.
8. A three-phase electric boiler is supplicd from a three-phase source,
but, due to ccrtain dissymmetry which ariscs from timo to time, a negative-
sequence voltage is developod which reflects in its effect throughout the
entire systcm. It is proposed to place in series with the load an induction
machine capable of limiting the neg'Mive-sequence current so that the
voltage on the supply sido is substantially balanced. It may be assumed
that the amount of the negative-sequencc voltage of 3 pcr cent across the
furnace will not be appreciably changed by a reduction in the negative-
sequence current. Thus a machine of essentially induction-motor con-
struction of 3 per cent of the rating of the boiler is rcquircd. This machine
will be driven ata spccd of 200 per cent slip which corrcsponds to synchro-
nous speed for the ncgative-sequenctJ. Such an arrangcment willlimit the
negative-sequcnce currcnt flowing back into the system to the magnitude of
the magnetizing currcnt. This can be further reduccd by shunting thc
series machine by static condcnsers. If the constants of the series machine
in per unit on its own rating are
Rrotor = 0.01 R .tator = 0.01
Xrotor = 0.15 Xstator = 0.15
Xm = 4.00
what is the rating of the capacitors required to resonate with the magnetizing
current and thus making the impedance to negative-sequence current a
maximum? What 1S the rating of the driving motor in terms of the rating
of the series machine neglecting windage and friction, and iron losses?
(JJVSYMMETR1éAt SYSTÉMS AND PIÍASE-BALANCERS 38g
9. Two single-phase transformers of impedance Z ohms on the primary

turns basis are connected in opcn delta between the a and b and the e and a
phases. With positive-scquencc voltage E A 1 applied to the primary, deter-
mine the expression for the secondary voltagc on the primary turns basis
when positive- and negative-sequcnce currents 1 A1 and 1 A2 on the primary
turns basis are drawn from the sccondary.
10. When positive-scqucnce currcnt alonc is drawn from the open-delta
connection of Prob. 9, dissymmetry is introduced by the un balanced circuit.
.
l
This condition may be corrected by the introduction of a series reactance
¡'' in one of the primary leads. Determine the lead in which this reactance
)
.. should be placed and also its magnitudc in terms of the transformer imped-
ance Z.
r
~
!
•
•
\
BIBLIOGRAPHY
This bibliography is divided into two parts, the first listing
references in the English language and t he second those in foreign
languages.
While the Bibliography is presented without pretensions asto
completeness, the authors believe that it is fairly so with respect
to American publications. Many of the foreign references
were obta.ined from the bibliographic study of symmetrical
components made by Dr. Hague. 09o)
The following is a list of the abbreviations used in the
Bibliography:
Arch. f. Elekt. Archiv für Elektrotechnik (Berlín).
Bull. assoc. ing. élec. Lú~ge
Bulletin de l' Association des 1ngénieurs élec-
triciens sortis de l' 1nstitut électrotechnique
M ontefiore (Li<~ge).
Bell System T ech. J our. Bell System Technical Journal (New York).
Bull. soc. franr;. élecl. Bulletin de la Société franr;aise des ~lectnciens
(París).
Comp?. rend. Comptes rend~ hebdomadaires des Séances de
l' Académie des Sciences (París).
Elec. Rev. The Electrical Review (London) .
Elec. Times The Electrical Times (London).
Elec. World Electrical World (New York).
Elec. Eng. Electrical Engineering (New Y:ork).
Elec. Jour. The Electric J ournal (Píttsburgh).
Elec. L ab. M in. Comms. Re- Researches of the Electrotechnical Laboratory of
searches, Tokyo the Ministry of Communications (Tokyo) .
ETZ Elektrotechnische Zeitschrift (Berlín) .
E .u. M. Elektrotechnik und M aschinenbau (Wíen).
Gen. Elec. R ev. General Electric Review (Schenectady) .
lnd. élec. 1ndustrie électrique (París).
/nst. Civil Eng. Selected Selected Engineering Papers oj the lnstitution oj
Eng. Papers Civil Engineers (London).
J our. A.I.E.E. J ournal of the American 1nstitute oj Electrical
Engineers (New York).
J our. l.E.E. J ournal of the 1nstitution oj Electrical Engineers
(London) .
J our. l.E.E. J apan J ournal oj the I nstitution oj Electrical Engineers
oj Japan (Tokyo).
Jour. Inst. Eng. A~tralia J ournal oj the 1nstit'Ution of Engineers of
Australia.
391
Elettrotec. L' Elettrotecnica. Giornale ed Atti dell' Associa-

zione Elettrotecnica Italiana (Milan).
R.G.E. Revue générale de l' Électricité (Paris).
Tijd. v. Elec. Tijdschrift voor Electrotechniek.
Trans. A.l.E.E. Transactions oj the American Institute oj
Electrical Engineers (New York).
Wiss. Verofftl. Siemens Wissenschaftliche V eroffentlichungen aus dem
Konz. Siemens Konzern (Berlín).
World Power World Power (London).
Z.f. angew. Math. u. Mech. Zeitschrijt für angewande M athematik und M ech-
anik (Berlin).
Part I. English Language References
1913
l. ALEXANDERSON, E. F. W . : Phase-balancer for Single-phase Load on
Polyphase Systems, Gen. Elec. Rev., vol. 16, No. 12, pp. 962- 964,
December.
1915
2. STOKVIS, L. G.: Analysis of Unbalanced Three-phase Systems. Reac-
tions in a Generator Carrying an Unbalanced Load Treated as Equiva-
lent to Two Balanced Loads, Elec. World, vol. 65, pp. 1111-1115,
May l.
1916
3. ALEXANDERSON, E . F. W., and G. H. HILL: Single-phase Power Produc-
tion, Trans. A.l.E.E., vol. 35, pp. 1315- 1327.
4. GILMAN, R. E., and C. L. FoRTESCUE: Single-phase Power Service from
Central Stations, Trans. A.I.E.E., vol. 35, Pt. II, pp. 1329-1347.
1918
5. FoRTESCUE, C. L.: Method of Symmetrical Coordinates Applied to the
Solution of Polyphase Networks, Trans. A.l.E.E., vol. 37, Pt. II,
pp. 1027- 1140.
• 1919
6. WALKER, MILEs: The Supply of Single-phase Power from Three-phase
Systems, Jour. I.E.E., vol. 57, pp. 109- 139.
1920
7. SLEPIAN, J.: Induction Motors on Unbalanced Voltages, Elec. World,
vol. 75, pp. 313- 315, Feb. 7.
8. LYON, W. V.: Unbalanced Three-phase Circuits, Elec. World, vol. 75,
No. 23, pp. 1304-1308, June 5.
9. SLEPIAN, J.: Reactive Power and Magnetic Energy, Trans. A.l.E.E.,
vol. 39, pp. 1115- 1133.
10. FECHHEI ER, C. J.: Fower Factor and Unbalance on a Polyphase
System, Trans. A.I.E.E., vol. 39, Pt. II, pp. 1469-1473.
11. FonTESCuE, C . L.: Polyphase Power Representation by Means of
Symmetrical Coordinates, Trans. A.I.E.E., vol. 39, Pt. II, pp.
1481- 1484.
BIBLIOGRAPHY 393
12. EvANs, R. D.: Measurement of Power Factoron Unbalanced Polyphase
Circuits, Trans. A.I.E.E., vol. 39, Pt. II, pp. 1485-1487.
13. LYON, W. V.: Reactive Power and Unbalanced Circuits, Elec. World,
vol. 75, pp. 1417-1420.
1921
14. TAKATsu, KYOSHI: Measurcment of Power Factor in Polyphase
Unbalanced Circuits, Elec. Lab., Min. Comms., Tokyo, Researches,
No. 90, pp. 1-23, February.
1922
15. FoRTESCUE, C. L.: Transformers and Rectifiers, in F. F. Fowle,
"Standard Handbook for Electrical Engineers," 5th cd., Sec. 6, p.
427, McGraw-Hill Book Company, Inc., New York.
16. LAWRENCE, R. R.: "Principies of Alternating Currents," Chap. 12,
p. 337, McGraw-Hill Book Company, Inc., New York.
1923
17. FoRTESCUE, C. L.: The Measuremcnt of Power in Polyphase Circuits,
Trans. A.I.E .E., vol. 42, pp. 358-371.
18. EvANS, R. D .: New Sequence System of Polyphase Meters, Elec.
World, vol. 81, No. 6, pp. 333-334, Feb. 10.
19. BEKKU, SADATOSHI: Theoretical Researches on Transmission Line
Problcms with Spccial Reference to the Are Suppressing Reactor,
Elec. Lab., Min. Comms., Tokyo, Researches, No. 129, April.
20. HAY, ALFRED: "Alternating Currents," p. 375.
1924
21. BREISKY, J. V.: A Phase-balance Current Relay of the Negative Phase
Sequence Type, Elec. Jour., vol. 21, No. 2, pp. 77-81, February.
22. BEKKU, SADATOSHI: Synchronous Operation of Two Alternators
through Unsymmetrical Impedances, Electrotechnical Laboratory,
Min. Comms., Tokyo, Researches, No. 138, pp. 1-11, February.
23. DuDLEY, A. M.: Induction Motors on Unbalanced Circuits. Vector
Methods of Analysis of Unsymmetrical Systems, Elec. Jour. , vol. 21,
pp. 339-343, Jul)".
1925
24. ScHOENFELD, O. C.: Effect of Unbalanced Voltages on Operation of
Induction Motors, Elec. Jottr. , vol. 22, pp. 30-33, January.
25. EvANS, R. D.: Finding Single-phase Short-circuit Currents on Cal-
culating Boards, Elec. World, vol. 85, pp. 761- 764, April 11.
26. BEKKU, SADATOSHI: Calculation of Short-circuit Ground Currents
on Three-phase Power Nctworks, Using thc Method of Symmetrical
Coordinates, Gen. Elec. Rev., vol. 28, No. 7, pp. 472-478, July.
27. LuNDHOLM, R.: "A Generalized Vector Thcory for Electrical Alter-
nating-current Circuits with lts Application to Certain Practica!
Problems."
1926
28. EvANs, R. D., and C. F. WAGNER: Further Studies of Transmission
Stability, Trans. A.I.E.E., vol. 45, pp. 51- 94.
29. MAcKERRAS, A. P .: Calculation of Singlc-phase Short-circuits by the
Method of Symmetrical Componcnts, Gen. Elec. R ev., vol. 29, No. 4,
pp. 218-231, April; No. 7, pp. 468-481, July.
30. BEKKU, S., M . DoTE, and M. UnusHIBATA: On the Method of Measure-
ment of Zero and Negative Phase Sequence Impedance of the Three-
phase Alternator, Elec. Lab., M in. of Comms., Tokyo, R esearches, No.
170, pp. 1-20, April.
31. BEKKU, S., and K. MAYEKAWA: Calculation of Ground Current Caused
by One Line Groundcd in a Three-phase Transmission N etwork,
Electrolechnical Laboralory, Minislry of Communications, Tokyo,
Researches, vol. 181, pp. 1- 114, September.
32. CARSON, J. R.: Wave Propagation in Overhead Wires with Ground
Return, Bell System T ech. Jour., vol. 5, pp. 539-555, October.
33. HAGUE, BERNARD: Thc Method of Symmetrical Coordinates in the
Theory of Polyphase Circuits, Insl. Civil Eng. (London), Selected
Eng. Papers, No. 40, pp. 1-38.
1927
34. W AGNER, C. F., and A. DovnKov: Impedance of a Rotating Syn-
chronous Machine to Negativo Sequence Voltage, Elec. Jour., vol.
24, No. 3, pp. 117- 121, March.
35. ÜRISCOM, S. B . : Characteristics of Ground Faults on Three-phase
Systems, Elec. Jour. , vol. 24, No. 4, pp. 151-156, April.
36. KARAPETOFF, V.: "Experimental Electrical Engineering," 3d ed.,
vol. 2, Chap. 39, p. 118, John Wiley & Sons, Inc., New York.
1928
37. KENNELLY, A. E.: Computation of the Unbalance Factor of a Thrce-
phase Triangle When Lhe Lengths of the Threc Sides Are Given,
Trans. A.I.E.E., vol. 47, No. 1, pp. 341- 342, January.
38. SAn, A. P.: A Note on the Unbalancing Factor of Three-phase Systems,
Trans. A.I.E.E., vol. 47, No. 1, p. 343, January.
39. WAGNER, C. F., and R. D . EvANS: Symmetrical Components, Elec.
Jour., vol. 25, No. 3, pp. 151-157, March; No. 4, pp. 194-197, Apri.l;
No. 6, pp. 307-311, June ; No. 7, pp. 359-362, July. (Continued in
refcrences 50 and 61.)
40. NoRMAN, H. M .: Mathcmatical Solution of an Unbalanced Three-
phase Voltage System, Jour. A.I.E.E., vol. 47, No. 6, pp. 449-451,
June.
41. PARK, R. H ., and B. L. ROBERTSON: R eactances of Synchronous
Machines, Trans. A.I.E.E ., vol. 47, No. 2, pp. 514-536, April.
42. STUBBINGS, G. W.: Symmetrical Components of Unbalanced Three-
phase Systems, World Power, vol. 10, pp. 512-517.
43. DAHL, O. G. C.: "Electric Circuits," vol. I, Theory and Application,
Chap. 3, p . 68, McGraw-Hill Book Company, Inc., New York.
44. LEwrs, W. W.: "Transmission Line Engineering," 1st ed., Chap. 4,
p. 163, McGraw-Hill Book Company, Inc., New York.
13!13LIOGRAPHY 395
1929
45. PARK, R. H., and E. H. BANCKER: System Stability as a Design
Problem, 'l'rans . A.I.E.E., vol. 48, No. 1, pp. 170-194, January.
46. N.E.L.A .: Calculation of Ground Currents, Serial Rep. Elec. Appa-
ratus Comm., Eng. Sec. Southeastern Division, N.E.L.A., March.
47. DovJIKOV, A.: Application of Automatic Oscillographs, Elec. Jour.,
vol. 26, No. 4, pp. 175-179, April¡ No. 5, pp. 202-206, May.
48. STEW ART, H. R.: How the Positive Phasc Sequence N etwork Works,
Elec. Jour., vol. 26, No. 6, pp. 260-263, June.
49. K u, Y. H.: Transient Analysis of A.C. Machinery, 'l'rans. A.I.E.E.,
vol. 48, No. 3, pp. 707-715, July.
50. WAG NEn, C. F., and R. D. EvANS: Symmetrical Components, Elec.
J our., vol. 26, No. 9, pp. 426- 431, Septcmber; No. 12, pp. 571- 581,
December. (Continuation of refercnce 39 completed in refercnce
61.)
51. CoTo, M.: Various Expressions for Symmetrical Three-phase Circuits
in Tcrm:s of Syrnmctrical Coordinates and a Symbolic Method for
Circular Functions, Elec. Lab., Min. Comms., 'l'okyo, Researches, No.
255, pp. 1-24.
1930
52. SuMMEns, I. H ., and J. B. McCLUHE : Progrcss in the Study of System
Stn.bility, 'l'rans. A.I.E.E., vol. 40, No. 1, pp. 132- 161 , J anuary.
53. JoiNT DEVEL. AND REs. SUB-COMMITTEE: An Investigation of Ground
Faults on 33-kv. Transmission System and the R esulting Voltages
in a Paralleling T elephone System, Eng. Rep. 4, pp. 7- 47, N.E.L.A.
and Bell System, April.
54. MoNSETH, I. E ., and P. H. RoBINSON: Theory and Application of
Rel ay Systcms, Elec. Jour ., vol. 27, pp. 411- 415,473- 480.
55. STIGANT, S. A.: Phasc Sequence on Thrce-phase Systems. Notes on
the Phase Scquence Idea in thc Study of Unbalanced Three-phase
A.C. Systems, Elec. Times, vol. 78, pp. 1021- 1023.
56. GoTo, M.: Undamped Electric Oscillation of an Alternator Connected
to a Transmission Line, Elec. Lab., Min. Comms., 'l'okyo, Researches,
No. 281, pp. 1- 47.
57. BowEN, A. E., and C. L. ÜILKESON : Mutual Impedances of Ground
Return Circuits, Sorne Experimental Studies, 'l'rans. A.I.E.E., vol.
49, No. 4, pp. 1370-1383, October.
1931
58. WAGNER, C. F .: Damper Windings for Waterwheel Generators, 'l'rans.
A.I.E.E., vol. 50, pp. 140-152, March.
59. EvANS, R. D., and S. H. WniGHT: Power System Voltages and Currents
under Fault Conditions-Effcct of Sequence Impedances, A.l.E.E.
Paper, No. 31-59¡ Abstract, Elec. Eng., pp. 415- 420, June.
60. LEWIS, W. A., and L. S. TIPPETT: Fundamental Basis for Distance
Relaying on Three-phase Systems, A.I.E.E. Paper, No. 31-61;
Abstract, Elec. Eng., p . 420, June.
61. WAGNER, C. F., and R. D. EvANS: Symmetrical Components, Elec.

Jour., vol. 28, No. 4, pp. 239--244, April¡ No. 5, pp. 308- 312, May¡
No. 10, pp. 586- 590, October¡ No. 11, pp. 624-630, November.
(Completes references 39 and 50.)
62. MoNSETH, I. T., and P. Il. ROBINSON: Theory and Application of
Relay Systems, Elec. Jour., vol. 28, pp. 375-379, 431- 435.
63. STUBBINGS, G. W.: Unba.lanced Three-phase Circuits, Elec. Rev.
(London), vol. 108, pp. 783- 784.
64. WILSON, W.: Vector Algcbra, World Power, vol. 15, pp. 192- 195.
65. STIGANT, S. A.: Vector Operators in Three-phase Systems, E lec. Times,
vol. 80, pp. 735-736.
66. HAYWARD, A. P .: Measurcment of Symmetrical Components, Elec.
Jour., vol. 28, No. 6, pp. 351-357, June.
67. BoHNER, C. W., and G. W. VAUGHAN: Phase Faults Analyzed on
Direct-current Calculating Board, Elec. World, vol. 97, No. 23, pp.
1068- 1071, June 6.
68. TIPPETT, L. S. C.: Application of Impedance Relays to Earth Fault
Protection, Jour. Inst. Eng. Australia, vol. 3, pp. 233-243.
69. CLEM, J. E.: Reactance of Transmission Lines with Ground Return,
Trans. A.I.E.E., vol. 50, No. 3, pp. 901- 918, September.
70. CLARKE, EDITH: Simultaneous Faults on Three-phase Systems, Trans.
A.I.E.E., vol. 50, No. 3, pp. 919- 941, September.
71. GARZA, E. DE LA: A Note on Symmetrical Components, Elec. Eng., vol.
50, p. 910.
72. KILGORE, L. A.: Calculation of Synchronous Machine Constants,
Reactances and Time Constants Affccting Transient Characteristics,
Trans. A.I.E.E., vol. 50, pp. 1201- 1214.
73. WRIGHT, S. H.: Determination of Synchronous Machinc Constants by
Test, Trans. A.I.E.E., vol. 50, pp. 1331- 1351.
74. DANNATT, C ., and J . W. DALGLEISH: "Electrical Power Transmission
and Interconnection," Chap. 9, Sir Isaac Pitman & Sons, New York.
1932
75. HAHN, W. C., and C. F. W AGNER: Standard Decrement Curves,
Trans . A.I.E.E., vol. 51, pp. 353- 362.
76. WAGNER, C. F., and S. H. WRIGHT: Calculations of Short-circuits on
Power Systems, A.I.E.E. Paper No. 32M- 4¡ Abstract, Elec. Eng.,
vol. 51, No. 2, p . 131, Fcbruary.
77. HAHN, W. C.: D ecrement Curves for Specific Systems, A.I.E.E. Paper,
32M- 5; Abstract, Elec. Eng., vol. 51, p . 131, February.
78. SELS, H. K.: Proposed Definilions of Terms U sed in Power System
Studies, R ept. Subject Comm. on Definitions, A .I.E.E. Paper, 32M-2;
Abstract Elec. Eng., vol. 51, No. 2, p. 106.
79. JoiNT DEVEL. AND REs. SuB-COMMITTEE: Coupling Factors for Ground-
return Circuits- General Considerations and Methods of Calcula-
tions, vol. 2, Eng. Rept. No. 14, pp. 121- 185, N.E.L.A. and Bell
System, April.
80. HARDER, E. L.: Peak Voltagcs on Saturating Reactances in Three-
Phase Circuits, Elec. Jour., vol. 29, pp. 124-129, March.
BIBLIOGRAPHY 397
Part II. Foreign Language References

1912
101. STOKVIS, L. G.: "Dcr Spannungsabfall des synchronen Drehstrom-
generators bci unsymmetrischer Belastung," R. Oldenbourg, Munich.
1914
102. STOKVIS, L. G.: Sur la création des harmoniques 3 dans les alternateurs
par suite des deséquilibrages des phases, Compt. rend., vol. 159, pp.
46-49.
1915
103. STOKVIS, L. G.: En cirkeldiagram voor cheeve belasting von draastroom-
machincs, De lngénieur, Jan. 9.
1917
104. SALDA, C. DELLA: The Thr~e-phase Induction Motor, Fed by Asym-
mctrical Voltage, Elettrotec., vol. 5, p. 294, June 5.
1918
105. M ULLEU, P.: Unsymmctrischc Mehrphasensysteme, E T Z, vol. 39,
pp. 343- 356, Aug. 29¡ pp. 353- 356, Sept. 5.
1921
106. EMDE, F.: Zur Definition der Scheinleistung und der Blindleistung bei
unglcichmassig bclastcten Mehrphasensystemen, E.u.M., vol. 39,
pp. 545- 547.
107. BRUDERLIN, R.: Die Induktionsmaschine als Phasenumformer,
E.u.M., vol. 39, pp. 361- 366.
1922
108. STOKVIS, L. G.: Sur la production des harmoniques 3 dans les machines
a induction a charge deséquilibrée, R.G.E., vol. 12, pp. 619-629,
Oct. 28¡ pp. 661-670, Nov. 4.
109. SToKvis, L. G.: Sur les diagrammes circulaires des systemes triphases
deséquilibrés et la définition de Jeur degré de deséquilibrage, Compt.
rend., vol. 174, pp. 1418- 1420¡ R.G.E., vol. 11, pp. 957-959.
110. SToKvis, L. G.: Sur la définition de la tension, de l'intensité et du
facteur de puissance d'un systeme triphase, R.G.E., vol. 12, pp.
307- 312, Sept. 2.
1923
111. ZACHRISSON, E.: Zur Definition des Begriffes Unsymmetrie eines
Drehstromsystems, E T Z, vol. 44, pp. 897-899, Sept. 27.
112. STOKVIs, L. G.: Essai d'une théorie générale des machines synchronés
a courant alternatif basée sur les champs tournants sinusidaux,
R .G.E., vol. 13, pp. 953- 967, June 9¡ pp. 995- 999, June 16.
113. SToKvis, L. G.: Sur le diagramme de la connection Scott utilisée pour
l'alimentation de deux réseaux monophasés indépendants, Bull.
soc. fran~. élect., vol. 3, pp. 729-750; Over de Scott'sche schakeling
\
als transformatie van draastroom in twee eenphasenstroomen,

Tijd. v. Elect ., May, June, and July, 1925.
114. STOKVIS, L. G., and C. LAVANCHY: Diagramme de la chute de tension
dans un conducteur de ligne triphasée de tension moycnne, cas
général d'un systeme deséquilibré geómetriquement et élcctrique-
ment, R.G.E., vol. 13, pp. 3- 5.
1924
115. THOMALEN, A.: Die Zerlegung unsymmctrischer Drehstromsystcme in
symmetrische, E T Z, vol. 44, pp. 500- 501.
116. SToKvrs, L . G.: Sur la décomposition des systemes triphases ne satis-
faisant pas a la condition que la somme des vecteurs représentatifs
est zéro, Compl. 1·end., vol. 178, pp. 1074-1076, March 24.
117. SToKvrs, L. G.: Sur la décomposition des systcmes triphascs dont la
somme des vecteurs n'est pas nuile, R.G.E., vol. 16, pp. 611-614,
Oct. 18. •
118. SToKvrs, L. G.: Décomposition de l'énergie d'un systeme triphase ou
une éncrgie constante a une énergie pulsatoire, Compt. rend., vol.
179, pp. 595- 597.
119. SToKvrs, L . G.: Sur la transmission de l'énergie par un systeme triphase
deséquilibré pour lequel ni la somme des tensions de phase, ni la
sommc des courants de phase n'est égale a zéro, R .G.E., vol. 16, pp.
979--982, Dec. 20.
120. DUBusc, R.: Relais de deséquilibrage pour circuits triphases, R.G.E.,
vol. 16, pp. 659--670, Oct. 25.
121. ÜENKIN, V.: Théorie du deséquilibre ct son application aux systemes
de protection, R.G.E., vol. 16, pp. 907-916, Dec. 6; pp. 951- 959, Dec.
13.
122. BoTTANI, E.: La moderna matematica dei circuiti trifasi, Elettrotec.
vol. 11, pp. 738-745, 768- 774, 928-937, 966-973, 1013-1023.
1925
123. BEKKU, SADATOSHI: Synchronisierung zweier Drchstrom Generatoren
durch unsymmctrische Impedanzen, E T Z, vol. 46, p. 656, April 30.
124. RASCH, G.: Unsymmetrische Mehrphasensysteme, E T Z, vol. 46, pp.
1446-1447, Sept. 17.
125. DARRIEUS, G.: Protection des grand réseaux, Bull. soc. franr; . élect.,
October.
126. BEKKU, SADATOSHI: Method of Symmetrical Coordinates and Its
Application to the Transmission Line Problem, .four. Inst. Elec.
Eng. Japan, November (in Nipponese).
127. GENKIN, V.: Deséquilibre électrique dans les installations polyphases.
Phasemetre pour circuits triphases deséquilibrés, R.G.E., vol. 18,
pp. 890-892, Nov. 28.
128. GENKIN, V.: Deséquilibre électrique dans les installations polyphases,
Supplément au bull. soc. franr; ., élect. , vol. 5, p. 45.
129. BEKKU, SADATOSHI: Methode der symmetrischen Koordinaten und
allgemeine Theorie der Erdschluss!Oscheinrichtungen, Arch. f.
Elekt., vol. 14, pp. 543- 555.
BIBLIOGRAPHY 399
130. GERSTMANN, B.: Ueber die Dimensionierung von Phasenspaltern,
E .u.M., vol. 43, pp. 653-658.
1926
131. GENKIN, V.: Réseau élcctrique filtrant et autorégulateur pour circuits
triphases, R.G E., vol. 19, pp. 123-129, Jan. 23.
132. GENKIN, V.: f~t.ude des dispositifs électriques applicables aux appareils
de mesure ct aux relais et destinés a modifier un courant dans un
rapport complcxe constant, R.G.E., vol. 20, May l.
133. ÜENKIN, V.: Réccpteurs dyssymétriqucs et procédés d'équilibrage des
circuits triph:.t::;cs a phases inégalement chargées, R.G.E., vol. 20,
pp. 123-136, July 24.
134. GENKIN, V.: Dcséquilibre électrique dans les réseaux, Ind. élec., p.
389, Sept. 10; !J. 419, Sept. 25.
135. F ALLou, J.: Caleul des courants de court-circuit dans les réseaux
triphases, Butl. soc. fran<;. élect., vol. 6, pp. 969- 1002, September.
136. NATALlS, F.: Unsymmetrischc Drehstromsysteme, Wiss. Verofftl.
Siemens Konz., vol. 5, Pt. I , pp. 1- 6.
137. RuoENBERG, REINHOLD: Die Ausbrietung dcr Erdstrome in der
Umgebung von Wechselstromleitungen, Z. f. angew. Math. u. Mech.,
vol. 5, p. 361, October.
138. ÜENKIN, V.: Mesure du facteur de puissance d'un circuit triphase non
équilibre. Phascmetre de la Compagnie des Compteurs, Bull. soc.
fran<;. élect., vol. 6, pp. 1158- 1166.
139. GENKIN, V.: La puissance élcctriquc dans un circuit triphase non
équilibre, R.G.E., vol. 19, pp. 165- 170.
140. GENKIN, V.: La mesure indirecto de l'énergie réactive, Bull. soc. fran<;.
élect., vol. 6, pp. 1148-1157.
141. STOKVIS, L. G.: Ovcr de energie van draaistroomsystemen, Tijd. v.
Elect., July-August.
142. GENKIN, V.: Sur les procédés statiques d'équilibrage d'une charge
monophaséc dans un réseau triphase, Bull. soc. fran<;., élecl., vol. 6,
pp. 841- 853.
1927
143. DUBusc, R.: Los idées modernos en matiere de protection contre les
surintensités, Conférence inlernationale des grand réseaux, vol. 2, p.
469.
144. ÜBERDORFER, G.: Das Rechnen nach dcr Methode der symmetrischen
Koordinaten, E.u.M., vol. 45, pp. 296-301.
145. GENKIN, V.: Coutribution a la t héorie du fonctionnemcnt des appareils
de mesure ou des relais sensibles au deséquilibre électrique, R.G.E.,
vol. 21, pp. 525- 527.
146. HAUFFE, G.: Unsymmetrische Drehstromnetzc, E T Z, vol. 48, pp.
1734.
147. DUBusc, R.: Les idées modernes en matiere de protection contre les
surintensités, R.G.E., vol. 22, pp. 293- 300.
148. FALLOU, J.: Á propos de la mesure de la réact;ance de dispersion des
alternateurs triphases, R.G.E., vol. 22, pp. 543-550.
1928
149. FALLou, J., L. DuPurs, and G. TrNNEL: Note sur le systeme employé
par l'Union d'ltlectricité pour la protection sélective des lignes en
parallele, R.G.E., vol. 24, pp. 29-34, July 7.
150. lLIOvrcr, A.: Dispositifs permettant d'obténir les composantes symétri-
ques des courants des tensions ou des puissances dans un systeme
polyphase sans l'employé des filtres, Bull. soc. fran<;. élect., vol. 8, pp.
839-868, August.
151. FoRTEscuE, C. L., and G. CALABRESE: I../applicazione delle coordinate
simmetriche alla risoluzione delle equazioni algebriche, A tti del
congresso internazionale dei matematici, Bologna, p. 159, Scptember.
152. FocAccrA, B., and G. CERILLO: Determinazione delle constanti delle
linee aeree triphase per l'applicazione del metodo dellc coordinate
simmetriche, Elettrotec., Dec. 15 and 25.
153. ÜBERDORFER, G.: Einige Erdschlussgrundprobleme in symmetrischer
D arstellung, E.u.M., pp. 969-976, Oct. 7.
154. GoRBUNOFF, M. M.: Calculation of Fault Currents and Voltages in
High Voltage Systems under Unsymmetricnl Short-circuit Condi-
tions, Theoretical and Experimental Electricity, vol. 1 (in Russian).
155. BARBAGELATA, A.: Extensione del metodo della quattro letture alla
determinazione delle, eom ponenti simmetriche, di un sistema trifase,
Elettrotec., vol. 15, pp. 482- 490.
156. ÜBERDORFER, G.: Die Spannungsverlagerung in Netzen mit Losch-
transformatoren nach System Bauch, Arch. f. Elekt., vol. 19, pp.
405-423.
157. lLwvrcr, A.: Protection sélective des réseaux, Bull. soc. franr;. élect.,
vol. 8, pp. 1063-1102.
1929
158. DovJrKov, A.: Application of Oscillographs in Electric Power Systems,
American Commerce and Industry, January (in Russian).
159. ÜBERDORFEih G.: Die Leistung in unsymmetrischen Dreiphasen-
systemen, E T Z, vol. 50, pp. 265-267, Feb. 21.
160. GENKIN, V.: Sur la protection sélective des réseaux de distrihution
d'énergie contre les surintensités de courants, R.G.E., pp. 285- 289,
Feb.23;pp.323-329, Mar.2.
161. HAUFFE, G.: Unsymmetrische Drehstromsysteme, E T Z, vol. 50, p.
1446, Oct. 3.
162. ÜRTIZ, JoAQUIN GARCIA: Aplicación de la teoría de las coordenadas
simetricas al studio general de las sistemas trifasicos equilibrados
o no equilibrados, Vease Anales, 'Pomoll, vol. 3, p. 580.
163. ToLVENSKY, V. A.: Voltage Diagram of Three-phase System under
Unsymmetrical Loads." Theoretical and Experimental Eleclricity,
vol. 5 (in Russian).
164. ÜBERDORFER, G.: "Das Rechnen mit symmetrischen Komponenten,"
Verlag B.G. Teubner, Leipzig (72 pages) .
165. BARBAGELATA, A.: La mesure et la tarification de l'énergie daos les
réseaux triphases, R.G.E., vol. 26, pp. 47ü-472.
BIBLIOGRAPHY 401
166. N UTZElBERGER, H.: Drei neue Verfahren der Zerlegung eines unsym-
metrischen Mehrphasensystems in zwei symmetrische, Arch. f.
Elekt., vol. 23, pp. 119-123.
167. ILwvrcr, A.: Les protections sélectives des réseaux basées sur le
deséquilibrage des courants, des tensions et des puissances, R.G.E.,
vol. 26, pp. 517-518.
1 1930
i .
. 168. F ALLOU, J.: Mise a la terre du point neutre par l'intermédiaire d'une
inductance, étudc des régimes transitoires asymétriques succédant a
l'établisscment ct a la rupture du court-circuit, R .G.E., vol. 27, pp.
!:' 739- 751, May 10.
~
t 169. YERNAUX, J .: Á propos de la décomposition des systemes de trois
)
r' vecteurs deséquilibrés, R.G.E., vol. 28, pp. 259--260.

t 170. ZoRN, M.: Bestimmung der Unsymmetrie von Drehstromnetzen,
E T Z , vol. 51, pp. 1233- 1238, Aug. 28.
' 171. ToLWINSKI, W.: Das Problcm der unsymmetrischen Belastung des
Synchrondrchstromgcnerators und des Drehstromtransformators,
Arch. f. Elekt., vol. 23, pp. 497- 521.
172. RrcCIA, A. DELLA: D écomposition et effets des courants polyphasés non
équilibrés, Bull. assoc. ing. élec., Liege, vol. 8, pp. 141- 148.
173. RrMINI, C.: Sistemi trifasi simmetrici es quilibrati, Elettrotec., vol. 17,
pp. 273-278.
174. OsEnoonFER, G.: Der Doppclerdschluss in einer zweifach gespeisten
Einfachleitung im Lichtc der Rechnung mit symmetrischen Kom-
poncntcn, Wiss. Verofftl., S'iemens Konz., vol. 9, Pt. II, pp. 77-87.
175. FROIDEVAUX 1 J.: Étude fonctionnemcnt d'une machine asynchrone a
chargc sur un réseau a charge deséquilibré, R .G.E., vol. 28, pp.
389- 402.
1931
176. BLONDEL, A.: Sur l'étude directe des systemes triphases deséquilibrés
au moyen d 'imp édances et admittances mutuelles de phases dans
les problemes de chutes de tension et de mise en court-circuit,
R.G.E., vol. 29, pp. 3-11, 53- 63, 9ü-100.
177. F ALLOU, J.: Succession de phases des harmoniques dans les branches
d'un réscau équilibré a nombre de phases quelconque, Bull. soc. frant;.
élect., vol. 1, pp. 123- 127, February.
178. HEsSENBERG, K.: Die Berechnung von Symmetriestorungen in Dreh-
·•'
3tromnetzen mit Hilfe von symmetrischen Komponenten und
Ersatzschaltungen, E.u.M., vol. 49, pp. 273-276, Apr. 2; pp. 299-
304, Apr. 19.
179. SToKvrs, L . G.: Over het rekenen met symmtrische componenten, De
Ingénieur, No. 36.
180. FRIEDLANDER, E., and O. ScHMUTZ: Deber Drehfeldscheider zur Auf-
spaltung unsymmetrischer Drehstromsysteme in die symmetrischen
Komponenten , Wiss. Verofftl., Siemens Konz., vol. 10, Pt. I, pp.
24-41.
181. ILIOVICI, A. : Phasemetre a fer tournant pour circuits triphases non
équilibrés, échelle 360°, Bull. soc. jrant;. élect., vol. 1, pp. 978-995.
182. WILD, W.: Der Doppelerdschlusstrom in Drehstromkabeln und seine

Einwirkung auf benachbarte Fernmeldkabel, Wiss. Verofftl.,
Siemens Konz., vol. 10, Pt. I, pp. 51-77.
183. GENKIN, V.: Courants de court-circuit dans les canalisations en cable
armé, Bull. soc. fran~. élect., vol. 1, pp. 886-897.
184. HAUFFE, G.: Ueber Drehfelder bei unsymmetrischen Drehstromsys-
temen, E.u.M., vol. 49, pp. 223-224.
185. GENKIN, V.: Sur la détermination graphique approchée du fonctionne-
ment des systemes triphases deséquilibrés en étoile avec point neutre
isolé, I nd. élec. pp. 281- 283, June 25.
1932
186. HAUFFE, G.: Die Symmetrierung des Drehstromsystems, E.u.M., vol.
50, pp. 85- 87.
187. MARGAND, F . : Relations entre les coefficients fondamentaux des
machines synchrones dans l'état variable et les diverses impédances
apparaissant dans l'étude des r égimes synchronés deséquilibrés,
R.G.E., vol. 31, pp. 241-248.
188. FouRMARIER, P.: Calcul des caractéristiques des moteurs asynchronés
a rotor monophase, R.G.E., vol. 31, pp. 376-380.
189. STOKVIS, L. G.: "La théorie des composantes Symétriques," 4e Section,
Communication No. 4-C-1, Congres international d'électricité,
París.
190. HAGUE, BERNARD: "Méthode des coordonnées symétriques," 4e
section, Rapport 4, Congres international d'électricité, Paris.
APPENDIX
This appendix includes material that has been collected to provide a con-
venient reference for use in calculations by the methods that have been out-
lined in this volume. The material is arranged as follows:
l. Notation
II. General formulas
III. Apparatus constants
IV. Transmission-line constants
V. Cable constants
VI. Equivalent networks
VII. Characteristics of conductors:
Series resistance and reactance
Shunt capacitive reactance
VIII. Exponential functions
I. NOTATION
The principal features of the notation used for voltage, current, imped-
ance, and admittance symbols are givcn below.
Voltage Symbols: E, vector expression for voltage (r.m.s. value); 2, conju-
gate vector expression for voltage (r.m.s. value); E, absolute value of vector
E; e, instantancous voltage.
Current Symbols: I, vector expression for current (r.m.s. value); !, con-
jugate vector expression for currcnt (r.m.s. value); l, absolute value of
vector/; i, insta.ntaneous current.
Impedance Symbols: Z = R + jX, total impedance usually; z = r + jx
for impedance of nctwork element or unit length of transmission line or
cable; Z, absolute value of vector Z.
Admittance Symbols: Y = G + jB, used generally to indicate system con-
stants; y = g + jb, uscd generally to indicate constants per unit length of
transmission line or cable; Y, absolute val u e of vector Y.
Phase Symbols: Subscripts a, b, and e indicate phasc a to neutral, phase b
to neutral, and phase e to neutral, respect.ively, and subscripts A, B, and C
indicate phase betwecn conductors be, pha::se between conductors ca, phase
between conductors ab, respectively.
Sequence Symbols: Subscripts O, 1, and 2 indicate zero-, positive-, and
negative-sequence, rcspectively, in a three-phase system.
L etter Subscripts: These refer to either phase or circuit elements; the exact
significance is to be determined by the context.
403
404 SYMMETRlCAL G'OMPONENTS
11. GENERAL FORMULAS

l. Fundamental Equations:
Ea = E,.o +Ea! + Ea2 lao = -!(la +lb +le)
Eb = Eao + a 2Eat + aEa2 l at = i(la +ah+ a 2le)
Ec = Eao + aEat + a 2Ea2 la2 = Hla + a 2h +ale)
2. Star-delta Current Transforma- 3. Star-delta Voltage Transforma-
tions: tions:
r..
- 1
1
'
1
1
1
1
1
1
1
1
1 1
1
Eel lE 11
1 1
1 1
1
1 1
1
l 1
1
la =lB -le EA = Ec- Eb

h = le- lA En =Ea.- Eo
le = lA -lB Ee = Eb- Ea.
AA ~~,
Positive seqU<lnce
l~z
Negootiw·sequence
E"''
lAo = indeterminate
j
lAt = yalat
1A2 -- _ _j_
y'31 a2
lao =O
lat = -j.y31At
1a2 = +jV3J..u
APPENDIX 405
4. Connections of Sequence Networks for Various Types of Faults:
lone·+o-line f"oou tt
Three·phase f"ooul+
Line-to -ground fault
Double line- +o- <;~rounoi fc:>ult
5. Properties of the Vector a :

1 = é 0 = 1 + jO.O a 4 =a
a = é 120 = - 0.5 + j0.866 a5 = az
a2 = ei 240 = -0.5 - j0.866 -a = eiaoo = 0.5 - j0.866
a3 = é360 = eio = 1 + jO.O -a2 = éoo = 0.5 + .10.866
1 +a+ a2 = O a2 _ a = V3'ei27o = -jy'3
a - a2 V3'ei90 = jy'3 1 - a2 V3'é 30 = -ja-yl3
a2 _ 1 V3'ei210 = jay'3 a - 1 = V3eft5o = -ja2v'3
1 - a ,.. -yl3e"aao = ja 2 y'3
'
i
...
III. APPARATUS CONSTANTS

l. Conversion Formulas: •
2 2
X per 1mi tE kv. X per centE kv. (10)
X ohm• per phau = or
Mva. Kva.
X ohm• per pha•• M va.
Xper unit 2
E kv.
NOTE. These formulas are for thrce-phase circuits and when the line-to-
line voltage is used for E, Mva. and Kva. should express total values, but
when line-to-neutral values are used for E, Mva. and Kva. should express
phase values.
2. Constants of Typical Three-phase Synchronous M achines:
Time
Type of machine xd Xi Xi' x2 X o constants
Turbine generators ................. 110 19 12 12 3 See Fig. 55

Salient-pole motors and generators
(with damper windings) . . ......... 110 33 22 22 6 See Fig. 56
Waterwheel generators (without damp-
er windings) ................... . . 110 35 30 50 7 See Fig. 56
Condensers . . . . . . . . . . . . . . .. . ....... 180 37 25 24 8 See Fig. 56
3. Constants of Typical 'l'ransformers:

Reactance,
Per Cent
Distribu tion .... . . ... ... . ... . . .... . . . . ... . ... ... . . 3
Network . . .. . .................. . ... . ....... . .... . 5
Power
Up to 66 kv ............ . .... . .... ........... .. . 5--7
88 and 110 kv . . ........ . . . .. . ........ ..... .... . &-9
132 and 154 kv ... . ... ........ . ......... . ..... . . 8-10
187 and 220 kv .... ...... .. . . . ..... . ........... . 10-14
The direct-current resistance varíes from 0.35 to 0.50 per cent.
4. Three-winding Transformers:
APPENDIX 407
IV. TRANSMISSION-LINE CONSTANTS
M:
l. Positive- and N egative-sequence 1 mpedance:
T¡ + jxl =Tone conductor+ j0.279loglo G • • •co n duc tor
ohms per mile per phase
r at 60 cycles.
in which D = geometric mean of three separations iu feet.
G.M.R.ccnductor = geometric mean radius of one conductor in feet (see Table
VI).
Ground wires havo negligible effect.
2. Zero-sequence 1 mpedance:
a. Self impedance:
ro + j x 0 = 3 (resistance of all conductors in parallel considered as a group)
+ 0.286 + j0.838 log1o ~~~1. ohms per mile per phase at 60 cycles.
in which G.M.R. = geometric mean radius of conductors in feet.
p = earth resistivity in meter-ohms.
Paveraoe = 100.
b. Mutual impedance:
r,. 0 + jxmo = 0.286 + j0.838logl0 28~VP ohms per mile per phase at 60 cycles
ab
in which dab = geometric mean separation betwccn the two groups of con-
ductors in feet.
NOTE. For very wide separations it may be necessary to use the more
general relations dcveloped in Chap. VII.
c. Ground wire:
+ J0.6?161o9 280 ~
clcab
Equtvcolent corcuit for 60 cycle
The above constants represent the reactances for 60 cycles. The resist-
ance of each lino may be added to each branch.
4. Positive- and N egative-sequence Capacitive Susceptance:

14.64 X I0-6 .
b1 = D mhos per mlle per phase at 60 cycles
log10 -
a
in which D = equivalent spacing in feet.
a = radius of conductor in feet.
5. Z ero-sequence Capacitive Susceptance:
3 883 x w-6
bo = · G.M.D. mhos per mile per phase at 60 cycles
log1o G.M.R.
in which G.M.D. 2 (average height of conductors) in fcet.
=
G.M.R. geometric mean radius of conductora in feet. (In sus-
=
ccptance calculations the G.M.R. of individual conduc-
tors is equal to the physical radius.)
6. Equivalent Circuit of L ong Lines:
équivalenf circuif
? = [(r1 + j x1) or (ro + j xo)] X length in miles

Y = [ (jb 1) or (jbo)] X length in miles
TABLE XX.~CoNSTANTs OF TYPICAL TRANSMISSION LINES
P ositive- a.nd nega.tive-sequenoe
Equiva.-
Reactance per rnile
lent Suscept- Cbarging,
Vt>ltage Conductor
spacing, ance,
feet Per cent k va..
rnicrornhoe per rnile
Ohms ba.sed on
per mile
100,000 kva.
33,000 7 00 A.C.S.R. 0 .80 7 .35 5.69 6.20

66,000 11 00 copper .823 1 .89 5.61 24.5
110,000 15 0000 copper .833 0.69 5 . 16 62.4
154 , 000 19 477 ,000 A.C.S.R. .764 0.32 5.40 128
220,000 24 795,000 A.C.S.R. .767 0.16 5.37 260
Without ground wire or with magnetic ground wire:

X o of single circuit = (2.7 to 3.5)Xt of single circuit
X 0 of double circuit = (3.6 to 6.0)X 1 of double circuit
With non-magnetic ground wire:
X o of single circuit = (1.7 to 2.7)Xt of single circuit
X 0 of double circuit = (2.0 to 4.0)X 1 of double circuit
APPENDIX 409
V. CABLE CONSTANTS
l. Single Conductor:
a. Positive- and negative-sequence resist-
ance.
(1) Conductor (see wire tables such
as X).
(2) Sheath currents (solidly bonded).
AR = increase in resistance due
to shcath currents
Xm2R. h ·¡
= X 2 + R 2 o ms per m1 e
m. a Dimensiona in inchea
per phase
Xm = 0.279 log10
r4
~ r6 ohms per mile per phase at 60
cycles
R. = (
r6
+ 0r4·~~Or5 -- r4
) ohms per mile per phase for lead
at 50°C.
b. Positive- and negative-sequence reactance.
(1) Without sheath currents.
S
X 1 = 0.279 log10 (G M R) ohms per mile at 60
. . conductor
cycles
(2) With sheath currents (solidly bonded).
f S Xm 3
1
X 1 = 0.279 log 10 (G .M.R)conductor -- Xm 2 R. 2 ohms per +
mile at 60 cycles
c. Positive- and negative- and zero-sequence shunt capacitive
reactance.
(See Fig. 108, Chap. X.)
d. Zero-sequence impedance
(1) Solidly bonded sheaths.
Equivalent circuit similar to that shown in Fig. 105;
however, ground circuit may usually be neglected, in
which case Zo = Z 1
(2) Sheaths insulated.
(See Fig. 82 or Appendix IV-2.)
2. Three Conductor:
Positive-, negative- and zero-sequence resistance, reactance, and
shunt capacitive reactance.
(See Table X, Chap. X.) ,
3. Type H:
Positive-, negative-, and zero-sequence resistance and reactance,
same as three-conductor cable.
(See Table X, Chap. X.)
Positive-, ncgative-, and zero-sequence shunt capacitive reactance,
sarne as single-conductor cable with sheath diameter equal to shield
dia.meter.
(See Fig. 108, Chap. X.)
TABLE XXI.-RECOMMENDED THICKNESS OF lNSULATION 1 FOR SINGLE-CONDUCTOR CABLE AND THREE-CONDUCTOR ~
~
H CABLE
TYPE o
A partial list from D. M. Simmons, "Calculation of the Electrical Problems of Underground Cables,"
Elec. Jour., p. 237, May, 1932.
Paper Varnished cambric 4 Rubber7
Size of
Rated conductor Ungrounded Ungrounded Ungrounded Ci:2
Grounded neutral Grounded neutral Grounded neutral 0..:::
voltage2 A.W.G. number neutral neutral neutral
or 1,000 cir. mils
~
64ths Mils 64ths Mils 64ths Mils 64ths Mils 64ths Mils 64ths Mils
~
l--3
1 1 1 ~
1 1 1 .......
1-4/ 0 5 78 5 78 5 78 5 78 5 78 5 78
~
t--;
225-500 6 94 6 94 6 94 6 94 6 94 6 94 (":)
600 525-1,000 109 109 109
7 109 7 109 7 109 7 7 7 C)
~
Over 1,000 8 125 8 125 8 125 8 125 8 125 8 125
C)
1-4/ 0 5 78 5 78 5 78 5 78 6 94 6 94
225-500 6 94 6 94 6 94 6 94 7 109 7 109 ~
1,000
525- 1,000 7 109 7 109 7 109
125
7 109
125
8
9
125
141
8
9
125
141
~
~
Over 1,000 8 125 8 125 8 8
1-4/ 0 6 94 6 94 6 94 6 94 7 109 7 109

225-500 6 94 6 94 6 94 6 94 8 125 8 125
2,000 125
525-1,000 7 109 7 109 7 109 7 109 8 8 125
Over 1,000 8 125 8 125 8 125 8 125 9 141 9 141
7-4 / 0 6 94 6 94 7 109 7 109 8 125 8 125

225-500 6 94 6 94 8 125 8 125 9 141 9 141
3,000 109 125 8 125 9 141 9 141
525-1,000 7 109 7 8
Over 1 ,000 8 125 8 125 9 141 --
9
----- --- -
141
- - - - - - ·-
10 156 ~
10
---
156
14-4/0 7 109 7 109 9 141 10 156 10 156 10 156
5,000 225-1 ,000 7 109 7 109 10 156 u 172 11 172 11 172
Over 1,000 8 125 8 125 10 156 11 172 12 188 12 188
8-1,000 9 141 10 156 11 172 13 203 11 172 14 219

7 ,000
Over 1 ,000 9 141 10 156 11 172 13 203 12 188 15 234
7-1,000 10 156 13 203 14 219 18 281 14 219 18 281

10 ,0003
Over 1,000 10 156 13 203 14 219 18 281 15 234 19 297
4-1 ,000 14 1 219 17 266 19 297 26 406 19 1 297 27 422

15,000
Over 1,000 14 219 17 266 19 297 26 406 20 313 28 438
23,000 2 and larger 19 297 24 375 27 422
34,000 2/ 0 and larger 26 406 33 516 ~
46,000' 4/ 0 and larger 33 516 ~

<:
69,000 See Note 5 48 750 t::1
.......
-- -- ~
t The insulation thicknesses for paper and varnished cambric are I.P.C.E.A. Standards. The rubber walls, though not I.P.C .E.A. Standards, are
believed to be representative of modern practice under average cond.itions.
2 All cables are rated on conductor to conductor basis for three-phase circuits. Al! cables ha ve an operating tolerance of 5 per cent above the rated
voltage except those rated at 15,000 volts and below, which have no operating tolerance. For intermediate voltages take the wall for the next higher
listed voltage.
Paper
a For voltages 10,000 to 35,000 the thicknesses given apply to both single-conductor and three-conductor type H cable.
4 For voltages 40,000 and higher, type H single-conductor cable is recommended.
ti Above 46,000 volts the ratio of the core d.ia.meter to the conductor dia.meter sha.ll not exceed 3 to l.
Varnished Cambric
s For braided or special designa consult I.P.C.E.A. specifications or ma.nufa.cturer.
Rubber
7These thicknesses a.lso apply to multi-conductor cables. For multi-conductor lead-covered cables for voltages 10,000 and higl>er, shield.ing is advis- ~
,_
able. For bra.ided or specia.l designa consult the manufacturer. ,_
TABLE XXII.-THICKNESS OF !NSULATION FOR THREE-CONDUCTOR BELTED CABLES ~
......
R ecommended by the Insulated Power Cable Engineers Association Standards (January, 1932) t,J
A partial list from D. M. Simmons, "Calculation of the Electrical Problems of Underground Cables,"
Elec. Jour., p. 237, May, 1932.
P a p er Varnished cambrict
Size of Conductor Belt Conductor B elt

Rated conduct0r,
voltage* A .W.G. number Ungrounded Ungrounded ~
or 1,000 cir. mils Grounded neutral neutral Grounded neutral
neutral ~
64ths Mils 64ths Mils
64ths Mils 64ths Mils 64ths Mils 64ths Mils ~
~
1 1 1 1
1-4/0 4 63 2 31 2 31 5 78 o o o o ~
1,000 225- 500 5 78 3 47 3 47 6 94 o o o o
525-1,000 5 78 3 47 3 47 6 94 2 31 2 31 ~
Over 1 , 000 5 78 3 47 3 47 7 109 2 31 2 31
......
8-2 6 94 3 47 3 47 5 78 o o o o ~
1-500 6 94 3 47 3 47 6 94 o o o o t:"-4
2,000 525-1,000 6 94 3 47 3 47 6 94 2 31 2 31 ~
Over 1 ,000 6 94 3 47 3 47 7 109 2 31 2 31 o
8-2
1-500
6
6
94
94
3
3
47
47
3
3
47
47
5
6
78
94
2
2
31
31
2
2
31
31 ~
~
3,000 525-1 ,000 6 94 3 47 3 47 6 94 3 47 3 47
Over 1 ,000 6 94 3 47 3 47 7 109 3 47 3 47
8-4/ 0 7 109 4 63 4 63 6 94 4 63 4 63 <:
~
5,000 225-1 , 000 7 109 4 63 4 63 7 109 4 63 4 63
Over 1 ,000 7 109 4 63 4 63 7 109 5 78 5 78
7 ,{)00 8 and larger 8 125 4 63 8 125 7 109 5 78 6 9~
10 , 000t 6 and larger 9 141 4 63 9 141 9 141 6 94 9 141
15,000t 4 and larger 11 172 5 78 11 172 13 203 7 109 13 203
17 , 000 4 and larger .. .. . .. .. .. . .. 14 219 7 109 14 219
*General. All cables for three-phase circuits are rated on conductor to cORductor basis. All cables havean operating tolerance of 5 p er cent except
those rated at 15,000 volts and below which have no operating tolerance.
For intermediate voltage take wall for next higher listed voltage.
t Paper. For voltages 10,000to 15,000 type H cableis recommended where applicable; for voltages 16,000and higher typeH cable only is recommended.
t Varnished cambric. For braided or special designa consult I.P.C.E.A. specifications or manulacturer.
Rubber. For rubber multi-conductor lead-covered cables use thicknesses given in Table XXI.
APPENDIX 413
TABLE XXIII.-THICKNESS OF LEAD SHEATH
Recl)mmended by The Insulated Power Cable Engineers Association
Paper lead power cables
Lead thickness
Core diameter under
lead, mils
64ths Mils
0- i:OO 5 78
401- 1,000 6 94
1 ,001- 1 ,800 7 109
1,801- 2 ,800 8 125
2,801 and over 9 141
Varnished cambric * lead power cables
Lead thickness
Core diameter undcr
lead, mils
64ths Mils
o- 425 3 47
426- 700 4 63
701-1,050 5 78
1,051-1,500 6 94
1,501- 2 ,000 7 109
2,001- 3 ,000 8 125
3,001 and over 9 141
* Usually considered representative for rubber lead power cables also.
VI. EQUIVALENT NETWORKS

l. Star-delta Transformations:
r
3 1.
. 3
Notation for using impedances in formulas

A _ B2Ba
1
- Bt + B2 + Ba
A _ BsBt
2
- Bt + B2 + Ba
_ B1B2
A3 -
Bt +B2 Ba+
B1 =
A 1A2 +
A2Aa
At
+ AaAt =
A 2 +A
a
+ A2Aa
At
B 2 -_ A 1A 2 + AA2A + A aA 3 1 _
-
A
a
+A 1
+ -:¡;-
A aA t
2
B a -_ A 1A 2 + AAa
2A + A aA 3 1 _
-
A
1
+ A + ---x;-
2
AtA 2
3 2
Notation for using admittances in formulas
2. Elimination of Star Point:
Express network in terms of admittances, as
S ~--t'"nT''-\--~'l
The equivalent delta without the junction point is

APPENDIX 415
in which Y12
yl + y2 + y3 + y4 + y6
y2y3
= Y1 + Y2 + Y3 + Y4 + Y6' etc.,
These relations may be generalized to include any number of terminals.
3. Equivalent N etwork for T wo M utually Coupled Circuits:
-I., z..
Ea
21L._
:S2b
Er
-
lo
Ea
4. Equivalent Network for Three Branches Connected inStar and Mutually

Coupled:
NoTE. The signa of the mutual impedances may be ± depending upon

the physical arrangement.
TABLE XXIV.-GENERAL CIRCUIT CONSTANTS FOR DIFFERENT TYPES OF NETWORKS
~
¡-.¿
Equations for general circuit constants in terma of constants of component networks ~
Network Type of network

number
A= B= C= D=
1 1 1
Es ER 1 z o
1 Series lmpedance 1
~
--
2 Shunt admittance fsN

e fRN
1 o y 1 .
V:¡
~
~
3 Transformer
W@n
~RN
1+ZTYT
2 zT (1 +ZTiT) YT 1+ZTYT
2
'"':3
~
......
....¡y¡z Sinh VZY

~
Cosh vlzY y'ZffSinh v'zY t-<
= (1+zJ + = r(1 +z{ +
4 Transmlssion line
N
Z2Y' + ... )
= Z(1+z{+
Z 2f2 + . .. ) Z 2f2 + ... )
Sameas A 8
24 120 120 ~
o
~
5 General network
~ D A B e D t:l;j
~
~
General network and transformer Es ER
6 impedance at receiving end A¡B¡C1 D 1 ~~ cvV', vv At Bt +AtZTB Ct Dt + CtZTB
ZTR
7 General network and transformer

impedance at sending end
Es
---~
ZTs
.A.AA 1A¡B¡C¡D¡ EB
At + CtZrs Bt + DtZTs Ct Dt
General network and transformer E 8 ER

8 impedance at both ends-referred ~ooooo~A 1 B1 C1 D 1 1.~ At + CtZTs Bt +AtZrR+ Ct + CtZTB
to high voltage Z TS Z TR DtZTs + CtZTRZTs Dt
- ---
·:,.~~· x.· -.
General network and transformer

9
impedance at both ends-transformers
having different ratios Ta and Ts Es~
~ z A B CD ~R
E
1 1 1 1
TB 1
Ts(At + CtZTs) TaTs(Bt + AtZTB + CtTBTs
Ts
TB(Dt + CtZTB)
referred to low voltage TS ZTR DtZTs + CtZTBZTs)
--
10 General network anrl shunt
admittance at receiving end
!
ESN
1 A 1 B 1 C 1D 1 1
@= ~RN At +BtYB Bt Ct + DtYB Dt
11 General network and shunt

admittance at sending end fsN@=
1 A¡B¡C¡D¡ 1 rERN
At Bt Ct +AtYs Dt +BtYB
General network and shunt IA¡B¡C¡D¡I Ct+A tYs+

12
admittance at both ends
1sN @= §= , fRN At+BtYB Bt DtYB +BtYBYs Dt +BtYs
~
Es ER
13 Two general networks In series A~B~C~D 1 A 1 B 1 C 1D 1 AtAz+CtBz Bt.Az + DtBz AtCz +CtDt B1C2+ DtD%
~
~
~
Two general networks In series Es z EB A1A2+CtB2 + BtAz+DtBz+ AtCz+ CtD2 + BtCz+DtDz + o........
14 with intermediate impedance ~~ C¡A,Z DtA4 CtC4 DtC4 ~
15 Two general networks In series

with intermediate shunt admittance
lE}A2B 2CzDzi
Es N
0= IA¡B¡C¡D¡'EtRN AtAz + CtB! +
AtBzY
BtAz + DtBz +
BtBzY
AtCz +CtD2+
AtDzY
BtCz + D1D2 +
BtD2Y
Aa(AtA2 + Ca(BtAz +
16 Three general networks In series
Es
--=-jAaB aCaDa HAzBzCzDzHA¡B¡C¡D¡ r ER
+Ba(AtCz +
CtBz) Aa(BtAz + DtBz)
+Ba(BtC2 + D1D2)
Ca(AtAz + CtBz)
+Da(AtC2 + CtDz)
DtBz)
+Da(BtCz +
C1D2) DtDz)
Es AtB1C 1Dt ER Ct + Cz +
AtBz + Ev~2 BtBz (At - Az)(D2 - Dt) B1D2+ D1B2
17 Two general networks in parallel
AzBzCzDz Bt+Bz Bt+Bz Bt+ Bz Bt + Bz
Non:. The exciting current of the receiving end transformers should be added vectorially to the load current, and the exciting current of the senrling end transformers should be
added vectorially to the sending end current. ~
General equations: Es = EBA
that lP all cases AD - BC = l.
+ IBB; EB = EsD - lsB; ls = lBD
·
+
EBC; lB = lsA - EsC. ABa check ln the numerical calculation of the A, B, C, and D constante note ::i
r · TABLE XXV.-NETWORK CoNVERSION FoRMULAs
~
To convert from
To
c:r
Z"'Z
~
..
"ll "ll
~
'"""'
00
ABCD
1
Admittance
1
lmpedance
1
Equivalent r
l Equivalent T o'O o
...¡ S!. >i
+
..... +
.....
Yu Zn
l'l::sl'l
... 1::;¡
~ 9
A .. ABCD constants 1 + ZYB 1+ ZsY •t.:> C> -•
o 11 u
Yu -Zu
~
1 ZuZn - Zut "ll §. "ll l:xl>l t;-, ttt>l ;.....
B•
Yu Zu
z ZR + Zs + YZBZs ;....
tl:l "'~so
p.-
::s ... ::s
p.~ p.
~ ~..
YuYn- YuS 1 ()
1::;¡~ 1
C= Et =AE1 +Bit -Zu YB + Ys + ZYRYs y t:::l
1
V:¡
Is = CE1 + DI1 Yu o..... J:1 tt1> 1.... tt1> 1....
Yu Zu t.:t> t"l "-o:::
D= E1 =DEs- Bis 1 + ZYs + ZBY
1
"'"'3 '""'... ~
[¡ = -CEs+ A l s Yu -Zu ~ ~
~
'O
1 + ZsY
~
A Admittance constants Zn o
Yu =
B YR + z1 ZB + Zs + YZRZs
> ~- ...2!.<"
-1*~~j;;Y;I-
ZuZu- Zut p. (JQ
::: '"-3
3
Yu =
1
B =
- Z u Z uZ u- Zu2 1
z
1
ZR + Zs + YZaZs ...
~- ¡:;·
p.
~
¡:;· ~
11 11
-.::.
~
.......
Yu =
D It = YnEt - YuEt Zu
Ys +z1 . 1 + YZB
ZR + Zs + YZsZs
"'::S
C>
(>"
~ ~.. ~.. ~
B lt = YuEt - YuEt ZuZu - Z 122 ~
"''"
<+
C>
t-t
D Yu lmpedance constants + ZYs
1 1 "'
¡;;- "'"'
1 1
Q
Zu = e1 YR + Ys + ZYRYs
Zs +-y ...... (JQ
C>
-.::. -.::. a
-lz22í;zul- ;;· "'.....
YuYu- Yu2 3 (JQ
- Yu 1 1 'O
...o ~ &, ~
Zu = -e YuYu- Yut YR + Ys + ZYRYs --y ~
p. (JQ
'O ~ ~
""¡j
a
= A Yu Et = -Zult - Znlt 1 + ZYR
+y
1 "'
::S
'O
o o
:¡¡ ~
Zu
e YuY!! - Y u' Et = Z u lt - Zalt YR + Ys + ZYRYs
Zs C>
~ ~
., ... ~
~
~
.....
Zu + Zu
::t>
A-1 Equivalent r YZs o
YR = Yu- Yu
t:z:j
"'o
~
-B- .o C> :¡¡
ZuZu - Z u S z ZR + Zs + YZRZs ¡::
.,
~
¡:;·
e.<" <+c:r
1 ZuZu- ZuS
Z= B
Yu Zu
Ysl#N IYR ZR + Zs + YZRZs oq
...;; n o
Zu + Zu
~
D- 1 YZR o ~ ~
Ys = -B- Yu- Yu
::S
<+ "'
<+
ZuZu- Zut ZB + Zs + YZRZs
"'
¡;;·
"tt
<+
c:r
~
~ ~..
D-1 Y22 - Yu ZYs Equivalent T
-e- + Zu Zs ZR t:z:j
..... ...
+
::S 1 1
.T
ZB = Zu .o
YuYu- Yu 2 YR + Ys + ZYRYs ¡::
~
~ ~
Y= • e YuYu- Y u 2 1
Ys + Ys + ZYRYs <" 1::;¡
:¡¡
o... ~ E..
..."'
Yu -Zu (D u .....
p;
~ ~
A- 1 Yu - Yu ZYR ::S
Zs = -e- Y u Y22 - Y12 2
Zu + Zu YR + Ys + ZYRYs
t"l o...
.......,
. S
"3
'
VII. CHARACTERISTICS OF CONDUCTORS
With the development of the method of symmetrical components, it
became advantageous to prepare a new type of table of the electrical char-
acteristics of transmission and distribution circuit conductora in order to
give the zero-sequence characteristics as well as the ordinary or positive-
sequence characteristics.
Another innovation is the introduction of the characteristics of the
recently developed hollow types of conductora. These tablea give the
characteristics of four types as follows: (1) stranded copper conductora, (2)
aluminum (A.C.S.R.) conductora, (3) Anaconda hollow conductora, and
( 4) General Cable Type HH hollow conductora.
In preparing the new tablea the authors have attempted to simplify
previous tables. They are indebted to W. A. Lewis for the suggestion of
splitting the reactances into components associated with the conductor,
the conductor spacing, and the equivalent depth of earth return. Mter
consideration of loading and ambient temperatures and a discussion with
the engineers of the wire companies, the single value of 50°C. for conductor
temperature was agreed upon as representing average operation andas being
the most satisfactory value for calculations required in power-system studies.
In addition, a column has been added to give the approximate maximum per-
missible current-carrying capacity. The calculations are based on tarnished
conductor surface and air velocity of 2 ft. per second and a conductor tem-
perature of 75°C. with 25°C. ambient, using the method described by Schurig
and Frick (Gen. Elec. Rev., March, 1930).
The conductora included in the tables were selected after discussion with
wire-company engineers. Advantage has been taken of a number of refine-
menta in thc calculation of inductance of stranded conductora which were
recently developed by W. A. Lewis and which will be described by him in
the near future. These tables, within their scope, are intended to supersede
previoua tablea of electrical characteristics of conductora.
The tables were prepared in cooperation with W. A. Lewis of the West-
inghouse Electric and Manufacturing Comp~ny with the assistance of the
engineers of the wire companies, particularly F. R. Da.llye of the Aluminum
Company of America, R. B. Steinmetz of the Anaconda Wire and Cable
Company, and D. M. Simmons of the General Cable Corporation.
Illustration. Determine the positive- and zero-sequence impedances
of the following line at 60 cycles, no ground wires:
Conductor: 795,000 cir. rnils A.C.S.R. 54 aluminum strands.
Effective separation: 26ft. Earth resistivity: 100 rneter-ohms.
From Table XXVI, p. 420: Ta = 0.138 ohms per phase per mile;
Xa = 0.401 ohms per phase per mile.
From p. 421 (or p. 423): Xd = 0.395 ohms per phase per mile.
From p. 420 (or p . 422): r~ = 0.286 ohms per phase per mile;
x~ = 2.89 ohms per phase per mile.
Positive-sequence impedance: Zr = Ta + j(xa + Xd) = 0.138 + j0.796
ohms per phase per mile.
Zero-sequence impedance: zo = Ta + r. + j(xa + x. - 2xd) =
0.424 + j2.50 ohms per phase per mile.
Ata given frequency, changing conductor changes only Ta and xa; changing
spacing changes only Xd; changing earth resistivity changes only x •.
419
¡ '
TABLE XXVI.-CHARACTERISTICS OF ALUMINUM CABLE STEEL REINFORCED
'Q ~~ Ta Xa
Size of
Num-
ber of Co~per
i
Cl)
=g
~~ a~~ = resistance, ohma per phase = reactance at
8-' ~CI)-a pcr mile at 50°C. 25°C. 1 ft., ohma per
~
conductor, wires eQUIVa-
circular
mils or
lent, • cir- c.> a¡
·¡: ......
cular mils {lZ3 ~~~ .§ ~ ~]
)(+- c.>
CISC above 25°C. ambient phaae per mile
A.W.G. ·~ ~ ] or A.W.G. !i~ 8]-
g ;;, ..ts:8·&g
o ~t:e »
c.> D.c. ey25 50 60 25 50 60
eles cycles eyeles cycles cycles cycles
::;¡:¡=
00 ó·S o c.>
<
1,590,000 54 19 1,000,000 1.545 0.0520 1,380 0.0646 0.0656 o. 06751o. 0684 0.150 0.299 0.359
1,510,500 54 19 950,000 1.506 .0507 1,340 . 0680 .0600 .0710 .0720 .151 .302 .362
1,431,000 54 19 900,000 1.465 .0493 1,300 .0718 .0729 .0749 .0760 . 152 .304 .365
1,351,500 54 19 850,000 1.424 .0479 1,250 .0761 .0771 .0792 . 0803 .154 .307 .369
1,272,000 54 19 800,000 l. 382 .0465 1,200 .0808 .0819 .0840 .0851 .155 .310 .372
1,192,500 54 19 750,000 l. 338 .0450 1,160 .0862 .0872 .0894 . 0906 .157 .314 .376
1,113,000 54 19 700,000 l. 293 .0435 1,110 .0924 .0935 .0957 .0969 .159 .317 .380
1,033,500 54 7 650,000 l. 246 .0420 1,060 .0994 .101 . 103 .104 . 160 .321 .385
954,000 54 7 600,000 1.196 .0403 1,010 .108 .109 .112 . 113 . 162 .325 .390
000,000 54 7 566,000 1.162 .0391 966 . 115 . 116 . 118 . 119 . 164 .328 .393
874,500 54 7 550,000 1.146 .0386 949 . 118 .119 .122 .123 .165 .329 .395
795,000 54 7 500,000 l . 093 .0368 897 .129 .131 .136 . 138 . 167 .334 .401
795,000 26 7 500,000 1.108 .0375 901 .129 . 129 .129 .129 .166 .332 .399
795,000 30 19 500,000 1.140 .0393 909 .129 . 129 . 129 . 129 . 164 .327 .393
715,500 54 7 450,000 1.036 .0349 834 .144 .145 . 147 . 148 . 170 .339 .407
715 ,500 26 7 450,000 1.051 .0355 838 .144 .144 .144 . 144 .169 .337 .405
715,500 30 19 450,000 1.081 .0372 845 .144 . 144 .144 .144 . 166 .333 .399
666,600 54 7 419,000 l. 000 .0337 800 .154 . 157 . 159 .160 .172 .343 .412
636,000 54 7 400,000 0.977 .0329 774 .162 . 164 . 168 . 169 . 173 .345 .414
636,000 26 7 400,000 0.990 . 0335 777 .162 . 162 . 162 .162 . 172 .344 .412
636 ,000 30 19 400,000 1.019 .0351 781 . 162 .162 . 162 .162 . 169 .339 .406
605,000 54 7 380,500 0.953 .0321 748 . 170 . 172 . 176 .178 .174 .348 .417
556,500 26 7 350,000 .927 .0313 734 . 185 . 186 .186 .186 . 175 . 350 .420
556,500 30 7 350,000 .953 . 0328 726 .185 .186 .186 .186 .173 .346 .415
500 ,000 30 7 314,500 . 904 .0311 692 .206 .206 .206 . 206 . 175 . 351 .421
477,000 26 7 300,000 .858 .0200 666 .216 .216 .216 .216 .179 .358 .430
477 ,000 30 7 300,000 .883 .0304 671 .216 .216 .216 .216 .177 .353 .424
397,500 26 7 250,000 .783 .0265 591 .259 .259 .259 .259 . 184 .367 .441
397,500 30 7 250,000 .806 .0278 596 .259 .259 .259 .259 .181 .362 .435
336,400 26 7 0000 .721 .0244 530 .306 .306 .306 . 306 . 188 .376 .451
336,400 30 7 0000 .741 .0255 535 .306 .306 .306 .306 . 186 .371 .445
300,000 26 7 188,700 .680 . 0230 493 .342 .342 .342 .342 .191 .382 .458
300,000 30 7 188, 700 .700 . 0241 497 .342 .342 .342 .342 .188 .377 .452
266,800 26 7 000 .642 .0217 457 .385 .385 .385 .385 . 194 . 387 .465
0000 6 1 00 . 563 .00814 340 .485 .514 .567 .592 .242 .484 .581
000 6 1 o .502 .00600 303 .612 .642 .697 . 723 .259 .517 .621
00 6 1 1 .447 .00510 266 .773 .806 .866 .895 .267 .534 .641
o 6 1 2 .398 .00446 233 .974 1.01 l. 08 1.12 .273 .547
.554
.656
.665
1 6 1 3 .355 .00418 199 1.23 1.27 l . 34 l. 38 .277
2 6 1 4 .316 . 00418 179 1.55 1. 59 l. 66 l. 69 .277 .554 .665
3 6 1 5 .281 .00430 157 l. 95 l. 98 2.04 2. 07 .275 .551 .661
4 6 1 6 .250 . 00437 137 2. 47 2.50 2. 54 2.57 .274 .549 .659
5 6 1 7 .223 . 00416 118 3.10 3.12 3.16 3. 18 .279 . 557 .665
6 6 1 8 . 198 . 00394 102 3.92 3.94 3. 97 3.98 .281 .561 . 673
• Based upon copper, 97 per cent; aluminum, 61 pcr cent.

t Based on a conductor temperatura of 75•c. and an ambient of 2s•c.
1 Te* 1
*
Xe, ohms per phase per mile
~
p, meter-ohma ....... All 1 50 tOOt 500 1,000 5,000 10,000
25 cycles............ 0.119 0.923 5
1.04 10
1.10 1.22 1.27 1.39 1.44 1.57 1.62
50 cycles............ .239 1.74 1.98 2.09 2.33 2.U 2.68 2.78 3.03 3.13
60 cycles.... . ....... 286 2.05 2.35 2.47 2.77 2.89 3.19 3.31 3.61 3.73
• Based on Eq. (163).

t This representa an avérage value whlch can be used In the abeence of more definlte informatlon.
APPENDIX 421
TABLE XXVI (Continued).-CHARACTERISTICS OF CoPPER CABLES
Hard drawn 97.3 per cent conductivity
• ~o Ta Xa
Size of conductor
~
..: ·~~~ = resistance, ohms per phaac = reactance at
·¡¡:...
.8 <=l
e"' per mile at so•c. 2s·c.
<ll ~ 1 ft., obms per
.....o ~ El+> .8 t .-2 nee above 2s•c. ambient phaae per mile
Circular A.W.G. .,... :.a

~~
..
.~Ji ~o>CJ
..., - .o* ~[l
<JoS
or .D
'!l"§ "'"'
¡:¡.:::
g-d ~ ~ bO-a D.c. 1 cyclcs
mile B. & S. ~ c:lf ~El.S~
25 1cycles
50 1 cycles 25 1 cycles
60 cycles 50 1 cycles
60
z 8·S
1,000,000 .... 61 1.152 0.0370 1,300 0.0640 0. 0648 o. o672ro. 0685 0.167 0.333 0.400
900,000 .... 61 1.093 .0352 1,220 .0711 . 0718 . 0740 .0752 .169 .339 .406
800,000 .... 61 l. 031 .0332 1,130 .0800 .0806 .0826 .0837 .172 .344 .413
750,000 .... 61 0.998 .0321 1,090 .0853 .0860 .0878 .0888 .174 .348 .417
700,000 .... 61 .964 .0311 1,040 .0914 .0920 .0937 .0947 . 176 .351 .421
600.000 .... 37 .891 .0285 945 .107 .107 .109 .109 .180 .360 .432
500,000 .... 37 .814 .0260 842 .128 .128 .130 .130 .184 .369 .443
400 ,000 .... 19 .725 .0229 730 .160 .160 .161 .162 .191 .382 .458
300,000 .... 19 .628 .0198 607 .213 .214 .214 .215 .198 .397 .476
250,000 .... 19 .574 .0 181 540 .256 .256 .257 .257 .203 .406 .487
211 '600 0000 19 .528 .0167 485 .302 .303 .303 .303 .207 .414 .497
211,600 0000 7 .522 .0158 484 .302 .303 .303 .303 .210 .420 .504
167,806 000 7 .464 .0140 416 .381 .381 .382 .382 .216 .432 .518
133,077 DO 7 .414 .0124 359 .481 .481 .481 .481 .222 .443 .532
105,535 o 7 .368 . 0111 309 .606 .606 .607 .607 .227 .455 .546
83,693 1 7 .328 .0099 266 .765 .765 .765 .765 .233 .467 .560
66,371 2 7 .292 .0088 230 .964 .964 .964 .964 .239 .478 .574
52,635 3 7 .260 .0079 198 1.22 1.22 1.22 1.22 .245 .490 .588
41 '741 4 7 .232 .0070 171 1.53 l. 53 1.53 l . 53 .251 .502 .602
33' 102 5 7 .206 .0062 147 l. 93 l. 93 l. 93 l. 93 .257 .514 .616
26,251 6 7 .184 .0056 127 2.44 2.44 2.44 2.44 .263 .526 .630
• Baaed on a conductor temperature of 75•c. andan ambient of 2s•c.
Xd, ohms per phase per mile
Separation, feet
1011 121 314 1 5 161 718 9
25 cycles................ o ..... o 0. 035 0.056 0.070 0.081 0. 091 0.098 0.105 0.111
10 0.116 0.121 .126 .130 .133 .137 . 140 .143 .146 .149
20 .151 .154 .156 .159 .161 .163 .165 .167 .168 .170
30 .172 . 174 .175 .177 .178 .180 .181 .183 .184 .185
---------------- ----
50 cycles........ . ....... o . .... o .070 .111 .140 .163 .181 . 197 .210 .222
10 .233 .242 .251 .259 .267 .274 .280 .286 . 292 .298
20 .303 .308 .313 .317 .321 .325 .329 .333 .337 .340
30 .344 . 347 .350 .354 .357 .359 . 362 .365 .368 .370
- - - - - - - - - - - - - - - - - - - -- -
60 cycles................ o ..... o .084 . 133 .168 .195 .217 .236 .252 . 267
10 .279 .291 .302 .311 .320 .329 .336 .344 . 351 .357
20 .364 . 369 .375 .380 .386 .391 .395 .400 .404 .409
30 .413 .417 .421 .424 .428 .431 .435 .438 .441 .445
Positive- and negative-sequence impedance:

Z¡ = Z2 = Ta + j(Xa + Xd)
Zero-sequence im pedance:
Zo = (ra +Te) + j(Xa + Xe - 2xd)
TABLE XXVII.-CHARACTERISTICB oF ANACONDA HoLLOW CoPPER

CONDUCTORS
r
¿cj
·~¡i
.:
'&o
';;l ... Ta Xa
.S~ [!3
.... .S
Q)
<::
13
mf=:
e:"'
= resistance, ohms per phase
per mile at 50°C. 25°C.
= reactance at
1 ft., ohms per
z
o
"'"'
::>o
1l..!!!
g·a
·~
....o
.... ~
~:rl
El--"
o2l
·-'-
....
~~
1"'"'~·z
.
Q)Q)>,
o.
r1se above 25°C. ambient pbase per mile
<:: ..8 e*~!3Jl D.c.

~~
00 ·-..e: El:.¡;¡ 25 50 60 25 50 1 60
'¡j ~ $:!CJ
8·5 0á3"' .:¿ El ..S00"'¡;.
Q.
cycles cycles eyeles cycles cycles cycles
A J3"' z ....
163 800,000 22 + 28 1.269 0.0464 1,230 0.0798 0.0800 0.0804 0.0806 0. 155 0.311 0.373
167 800,000 16 + 22 1.185 .0420 1,200 .0798 .0801 .0806 .0810 .160 .321 .385
9e 750,000 18 + 24 1.176 .0422 1,160 .0852 .0854 .0858 .0861 .160 .320 .384
362 750,000 12 + 18 1.092 .0376 1,130 .0852 .0855 .0862 .0867 .166 .332 .398
559 700,000 21 + 27 1.171 .0426 1,120 .0912 .0914 .0918 .0920 .160 .319 .383
361 700,000 16 + 20 1.100 .0388 1,100 .0912 .0915 .0920 .0923 .164 .329 .394
560 650,000 19 + 25 1.106 .0399 1,060 .0983 .0984 .0988 .0990 .163 .326 .391
561 650,000 12 + 18 1.015 .0349 1,030 .0983 .0986 .0992 .0996 .170 .339 .407
360 600,000 22 + 28 1.100 .0402 1,020 .106 .107 . 107 .107 .163 .325 .390
30 600,000 16 + 22 1.027 . 0365 996 .106 .107 .107 .107 .168 .335 .402
563 550,000 24 + 30 1.075 . 0396 968 .116 .116 .116 .117 .163 .327 .392
379 550,000 18+ 24 1.008 .0361 949 .116 .116 .117 .117 .168 .336 .403
4-B 500,000 22 + 28 1. 002 . 0366 903 .128 .128 .128 .128 .167 .334 .401
401 500,000 12 0.995 .0356 901 .128 . 128 .128 .128 .169 .337 .405
405 450,000 26 1.180 .0455 902 .142 .142 .142 .142 .156 .312 .375
426 450,000 15 1.000 . 0367 857 .142 .142 . 142 . 142 .167 .334 .401
51 400,000 19 1.005 .0378 810 .160 .160 . 160 .160 .166 . 331 .397
565 350,000 21 0.968 .0367 749 .182 .183 .183 . 183 .167 .334 .401
378 350,000 12 + 18 .742 .0255 690 .182 . 183 . 183 .183 .185 . 371 .445
178 300,000 12 . 763 .0273 645 .213 . 213 .213 . 213 .182 .364 .437
567 300,000 9 .710 . 0244 631 .213 .213 .213 .213 .188 .376 .451
37 250,000 16 .754 . 0279 587 .255 .256 .256 .256 .181 .362 .434
415 250,000 14 .725 .0264 580 .255 .256 .256 .256 .184 .367 .441
446 0000 21 .755 .0287 540 .302 .302 .302 .302 .180 .359 .431
569 0000 9 . 598 . 0205 504 .302 .302 .302 .302 .196 .393 .471
158 000 16 . 611 .0226 452 .381 .381 .381 .381 . 192 .383 .46o
570 000 12 . 570 .0204 443 .381 .381 '.381 .381 .197 .394 .472
10 00 17 .550 .0205 390 .480 .480 .
.480 .480 . 197 .393 .472
• Based on a conductor temperature of 75°C. and an ambient of 25°C.
1 Te* 1 *
Xe, ohms per phasc per mile
p, meter-ohma .... .... AU 1 5 10 50 lOOt 500 1,000 5,000 10,000

- -
25 cycles.......... .. 0.119
- -
0.923
-- ---- - -
1.04 1.10 1.22 1.27
-- ----
1.39 1.44 1.57 1.62
50 cycles.......... .. .239 1.74 1.98 2.09 2.33 2.44 2.68 2.78 3.03 3.13
60 cycles ........ .. .. .286 2.05 2.35 2.47 2.77 2.89 3.19 ' 3.31 3.61 3.73
• Based on Eq. (163).

t This represente an average value which can be ueed in the absence of more defin.ite lnformation.
APPENDIX 423
TABLE XXVII (Continued).-CHARACTERISTIOS OF GENERAL CABLE
TYPE HH HOLLOW COPPER CONDUCTORS
•o
.!::: • .:S ~ >. Xa
':~ ál ..:
..e
<.> . .!.t:O
><¡:3<0
Ta
.S< .9 s..Jo..., = resistance, ohms per phase = reactance at
.,~
.,
.S d
~ per mile at so·c. 25°C. 1 ft., ohms per
<> ...
...,:lo ~ 8+> .se$
d"'
....o .~ ~ .~J! .§ ~l
rise above 25°C. ambient phase per mile
8] ... "O
.,•
...,g¡
<.>
~ .,~ !3
1l ~...- ~
-
~~ ~
·-..e
2<> :::1 8:.a ... ~ toO<.>
o.
~e.s s D·e· 25 1 cycles
1 cycles 50 1 cyclcs
25 50 1 cycles
60 ey eles cycles 60
J3<> z :Id
o·- ~"' as"'
0'"
1,000,000 12 l . 923 0.142 0.0763 1,580 0.0633 0. 130 0.260 0. 312

950,000 12 1.865 .139 .0739 1,520 . 0666 . 132 .263 .316
900,000 12 1.809 . 136 .0717 1,470 .0703 .133 .266 .320
850,000 12 l. 745 .133 .0691 1,410 .0745 . 135 .270 .324
800,000 12 1.687 . 130 .0668 1,350 .0791 . 137 .273 .328
750 ,000 12 1.622 .126 .0641 1,290 .0844 .139 .277 .333
700,000 10 1.554 . 123 .0614 1,230 . 0904 d d c5 .141 .282 .338
650,000 10 1. 490 .120 .0588 1,170 .0974 ..¿ ..¿ ..¿ .143 .286 .343
... ... ...
.E .E .8
600,000 10 1.418 .116 .0560 1, 110 .106 . 146 .291 .350
550,000 10 1.343 . 113 .0529 1,040 . 115 .,gj .,gj .,gj .149 .297 .357
500,000 9 l . 267 .109 .0498 977 . 127 . 152 .303 . 364
450,000 9 1.188 .105 .0466 009 . 141 J 8
r)g J .155 .310 .372
400,000 8 1.105 .101 .0433 838 .158 .159 .317 .381
350 ,000 7 1.016 .096 .0397 764 . 181 .163 .326 .391
300,000 7 0.920 .001 .0359 686 . 211 .168 .336 .404
250,000 7 .818 .086 .0318 604 .253 .174 .349 .418
0000 7 . 733 .082 .0283 538 .299 .180 .360 .432
000 7 .629 .077 .0241 458 .377 .188 .376 .452
• The conductor diameters for thesc arcas can be incrcascd by as much as 10 per cent and decreased by aa
much as 25 per cent if desired.
t Based on a conductor temperaturc of 75•c. and an ambient of 2s•c.
Xd , ohms pcr phase per mile
Separation, feet
1 2 1 3 1 4 1 5 6 7 8 9
25 cycles........... . .. . . o . .... o 0. 035 0.056 0.070 0. 081 0.091 0. 098 o. 105 0.111
10 0.116 0.121 .126 .130 .133 . 137 .140 .143 .146 . 149
20 .151 .154 . 156 . 159 .161 . 163 .165 .167 .168 .170
30 . 172 .174 . 175 .177 .178 .180 . 181 .183 .184 . 185
- -- - - -- -- - - - - -- -- - - - - -
50 cycles .... . .......... . o . .... o .070 .111 .140 .163 .181 .197 .210 .222
10 .233 .242 .251 .259 .267 .274 .280 .286 .292 . 298
20 .303 .308 .313 .317 .321 .325 .329 . 333 .337 .340
30 .344 .347 .350 .354 .357 .359 .362 .365 .368 . 370
-60 cycles................ o
- - - - - -- -- -- - -
..... o .084 .133 .168 .195 .217
- ------
.236 .252 .267
10 .279 .291 .302 .311 .320 .329 .336 .344 .351 .357
20 .364 .369 .375 .380 .386 .391 .395 .400 .404 . 409
30 .413 .417 .421 .424 .428 .431 .435 .438 .441 .445
Positive- and negative-sequence impedance:

Zt = Z2 = Ta + j(Xa + Xa)
Zero-sequence impedance:
Zo = (ra + Te) + j(Xa + Xe - 2xd)
TABLE XXVIII.-SHUNT CAPACITIVE REAOTANCE

Aluminum cable steel reinlorced Strandc<I copper conductora
o S
Xa' c:i Xa'
ti ::l Shunt capacitivo reactance ti ... Shunt capacitivo roactance
<... ¡::1
·a at 1 ft., megohms per

phase per mile <... "'c.c. at 1 ít., megohms per
phasc per mile
o ::l o o
a-:
Ol
,..en
el~
-
Oj
o
~{!
25 cycles 50 cyclos 60 cyclos
1~
..,en
el~
-
"'o
~{!
..OQ 25 cycles 50 cycles 60 cycles
..0¡::1
'3 '3
"'
,..¡::¡:¡. a~
::l...,
"'...¡::¡:¡. a~
::l...,
6'--' Z"' 6'--' Z"'
1,590,000 54 0.195 0.0977 0.0814 1,000,000 61 0.216 0 . 108 0.0901
1,510 ,000 54 .197 .0986 .0821 900,000 61 .220 . 110 .0916
1,431,000 54 .199 .0996 .0830 800,000 61 .224 .112 .0934
1,351,500 54 .201 .101 .0838 750,000 61 .226 .113 .0943
1,272 ,000 54 .203 .102 .0847 700,000 61 . 229 . 115 .0954
1,192 ,500 54 .206 .103 .0857 600,000 37 .235 .117 . 0977
1,113,000 54 .208 .104 .0867 500,000 37 . 241 .121 .100
1 .033,500 54 .211 .105 .0878 400,000 19 .249 .125 .104
954 ,000 54 .214 .107 .0890 300,000 19 .259 . 130 .108
900,000 54 . 216 . 108 .0898 250,000 19 .266 .133 .111
874,500 54 .217 .108 .0903 0000 19 .272 . 136 . 113
795,000 54 .220 .110 .0917 0000 7 .273 . 136 .113
795,000 26 .219 .110 .C912 000 7 .282 . 141 . 117
795 ,000 30 .217 .109 .0904 00 7 .289 .145 .120
715,500 54 .224 .112 .0932 o 7 .298 .149 . 124
715,500 26 .223 . 111 .0928 1 7 .306 .153 .127
715 ,500 30 .221 .110 .0920 2 7 .314 .157 . 131
666 ,600 54 .226 .113 .0943 ' 3 7 .322 .161 .134
636,000 54 .228 .114 .0950 4 7 . 330 .165 . 138
636 ,000 26 .227 . 114 .0946 5 7 .339 .169 .141
636,000 30 .225 .113 .0937 6 7 .347 .173 . 145
605 ,000 54 .230 .115 .0957 --
556 ,500 26 .232 .116 .0965
556,500 30 .230 .115 .0957 Illustration. Determine the posi-
500.000 30 .234 .117 .0973 tive- and zero-sequence capacitive
477 ,000 26 .237 .119 .0988 rcactances of the line in the illus-
477,000 30 .235 .118 .0980 tration, page 419 .
397,500 26 .244 .122 . 102
Height above ground: 60 ft.
397,500 30 .242 .121 .101 Xa' = 0.0917 megohms per phase
336,400 26 .250 .125 .104 per mile.
336,400 30 .248 .124 . 103
300,000 26 .254 .127 .106 xi = 0.0967 megohms per phase
per mile.
300 ,000 30 . 252 .126 . 105 x.' = 0.426 megohms per phase
266,800 26 .258 . 129 .107
0000 6 . 267 . 133 .111 per mile.
Xt = X a' + xi = 0.1884 megohms
1
000 6 .275 . 138 .115
00 6 . 284 .142 .118 per phase per mile.
X o' = Xa + xo' - 2xi = 0.324
1
o 6 .292 .146 .122
1 6 .300 .150 .125 megohms per phase per
2 6 .308 .154 .128 mil e.
3 6 .317 .158 .132 For other lengths than one mile
4 6 . 325 .162 .135 divide by the length.
5 6 .333 .167 .139
6 6 .342 .171 .142
xe', megohms per phasc per mile

Helght
ground,at?ove
feet 1 10 1 15 1 20 1 25 1 30 1 40 1 50 1 60 l 70 80 90 100
25 ey eles ....
50 cyclcs. . . . .
:¡o.. 64010.7271
0.78810. 836¡o. 87510.93110.98411.02 ,1. 06 1.08 1.11 1.13
320 . 363 . 394 . 418 . 437 . 468 . 492 O. 511 O. 528 0.542 0.555 0.566
60 cyclcs.. . .. .267 .303 .328 .348 .364 .390 .410 .426 .440 .452 . 462 .472
••
APPENDIX 425
T ABLE XXVIII. -SHUNT CAPAOITIVE REACTANCE. -(eontinued)
Anaconda hollow copper conductora General Cable Type HH hollow

copper conductora
o Xa
1
o Xa
1
~ Shunt capacitive reactance ~ Shunt capacitive reactance

..¿
..
o
at 1 ft., mcgohms per
phase p cr mile ..
-<
o
at 1 ft., mcgohms per
phasc per mile
m
:";:) .$
ó 8---:- "8---:-
z ... c:n
o'J~ 25 cyclcs 50 cycles 60 cyclcs
... c:n
oS~ 25 cycles 50 cyclcs 60 cycles
l'l
tlll
-~
'3c.>
• g.
...
.!::~ ~
Cl o"" 6""
·--
163 800,000 0 . 209 0.105 0.0872 1,000,000 0.180 0.0899 0 . 0749
167 800,000 .214 .107 • .0893 950,000 .182 . 0910 . 0758
96 750,000 .215 .107 .0895 900,000 .184 . 0920 . 0767
362 750,000 .220 .110 . 0917 850,000 .187 .0933 .0778
559 700 ,000 .215 .108 .0896 800,000 .189 .0945 .0789
361 700,000 .220 .110 .0915 750 ,000 .192 .0959 .0799
560 650,000 .219 . 110 .0913 700 ,000 .195 .0975 .0812
561 650 ,000 .225 .113 .0938 650,000 .198 .0990 .0825
360 600,000 .220 .110 .0915 600 ,000 .201 .101 .0839
30 600,000 .224 .112 .0935 550,000 .205 .103 .0855
563 550 ,000 .221 .111 .0921 500,000 .209 .105 .0873
379 550,000 .226 .113 .0941 450,000 .214 .107 . 0892
4-B 500,000 .226 .113 .0942 400,000 . 219 .110 . 0913
401 500,000 .227 .113 .0944 350,000 . 225 .113 .0938
405 450,000 .215 .107 .0894 300,000 .232 .116 . 0968
426 450,000 .226 .113 .0943 250 ,000 .241 .120 .100
51 400,000 .226 .113 .0941 0000 .248 .124 . 104
565 350,000 .229 .114 .0953 000 . 259 .130 . 108
378 350,000 .248 .124 .103
178 300,000 .246 .123 .102
667 300 ,000 . 251 .125 .104
37 250 ,000 .245 .122 .102
415 250,000 .249 .125 .104
446 0000 .246 .123 .103
569 0000 .263 .132 .110
158 000 .261 .131 .109
570 000 . 266 .133 .111
10 00 .269 .134 .112
Xd ', megohms per phase per mile
Separation, feet 1 O 1 2 3 4 5 6 7 8 9
25 cycles .... o . . .... o 0 . 0494 0.0784 0 . 0987 0 . 115 0 . 128 0 . 139 0 . 148 0 . 157
10 0.164 0.171 .177 .183 . 188 .193 .197 .202 .206 .210
20 .213 .217 .220 .223 .226 .229 .232 .235 .237 .240
30 .242 .245 .247 .249 .251 .253 .255 .257 .259 .261
50 cycles .... o
- - --- -.0247
.. .... o
--- -- -- -- ---- ---- -
.0391 .0494 . 0573 .0638 .0693 .0740 .0782
10 .0820 .0854 .0885 .0913 .0940 .0964 . 0987 .101 .103 .105
20 .107 .108 .110 .112 .113 .115 .116 .117 .119 .120
30 .121 .122 . 123 .125 .126 . 127 .128 .129 .130 .130
- - - - - - - -- - - - -- - - - -
- - - -o - -.0206
60 cycles .... o .. .. .. .0326 .0411 .0478 .0532 .0577 .0617 .0652
10 .0683 .0711 .0737 .0761 .0783 .0804 .0823 .0841 .0858 .0874
20 .0889 .0903 . 0917 .0930 .0943 .0955 . 0967 . 0978 .0989 . 0999
30 .101 .102 . 103 . 104 . 105 . 106 . 106 . 107 .108 .109
Positive- and negative-sequence shunt capacitive reactance:

X¡' = Xz ' = Xa1 + Xd'
Zero-sequence shunt capacitive reactance:
Xo 1 = Xa 1 + Xe' - 2Xd '
VIII. EXPONENTIAL FUNCTIONS

TABLE XXIX.-ExPONENTIAL FuNCTIONS E-.,
X O 1 0.01 1 0.02 1 0.03 1 0.04 1 0.05 1 0.06 1 0.07 1 0.08 1 0.09
o 1.00000 0.99005 o.98020 0.97045 0.96079 0.95123 o. 94177 0.93239 0.92312 0.91393
0.1 .90484 .89583 .88692 .87810 .86936 .86071 .85214 .84367 .83527 .82696
.2 .81873 .81058 .80252 .79453 .78663 .77880 . 77105 . 76338 .75578 .74826
.3 .74082 .73345 .72615 . 71892 . 71177 .70469 .69768 .69073 . 68386 .67706
.4 .67032 .66365 .65705 . 65051 .64404 .63763 .63128 .62500 .61878 .61263
.5 0.60653 .60050 .59452 .58861 .58275 .57695 .57121 .56553 .55990 .55433
.6 .54881 .54335 .53794 .53259 . 52729 .52205 .51685 . 51171 .50662 .50158
.7 .49659 .49164 .48675 .48191 .47711 .47237 .46767 .46301 .45841 .45384
.8 .44933 .44486 .44043 .43605 .43171 .42741 .42316 .41895 .41478 .41066
.9 .40657 .40252 .39852 .39455 . 39063 • . 38674 .38289 .37908 .37531 . 37158
1.0 .36788 .36422 .36059 .35701 .35345 .34994 .34646 .34301 .33960 .33622
1.1 .33287 .32956 .32628 .32303 . 31982 .31664 .31349 .31037 .30728 .30422
1.2 .30119 .29820 .29523 .29229 .28938 .28650 .28365 .28083 .27804 .27527
1.3 . 27253 . 26982 .26714 .26448 .26185 . 25924 .25666 .25411 .25158 .24908
1.4 . 24660 .24414 . 24171 .23931 .23693 .23457 .23224 .22993 .22764 .22537
1.5 .22313 .22091 . 21871 .21654 .21438 .21225 .21014 . 20805 .20598 . 20393
1.6 .20190 . 19989 . 19790 .19593 . 19398 .19205 .19014 . 18825 .18637 . 18452
1.7 . 18268 . 18087 .17907 .17728 .17552 .17377 . 1720-1 . 17033 .16864 .16696
1.8 .16530 . 16365 .16203 .16041 .15882 . 15724 .15567 . 15412 .15259 . 15107
1.9 .14957 .14808 . 14661 . 14515 .14370 .14227 .14086 .13946 .13807 .13670
2.0 . 13534 .13399 .13266 .13134 .13003 .12873 .12745 .12619 . 12493 . 12369
2.1 . 12246 . 12124 . 12003 . 11884 .11765 . 11648 .11533 . 11418 .11304 . 11192
2.2 .11080 .10970 .10861 . 10753 .10646 .10540 .10435 . 10331 . 10228 . 10127
2.3 . 10026 .09926 .09827 . 09730 .09633 .09537 .09442 .09348 .09255 .09163
2.4 .09072 .08982 .08892 . 08804 .08716 .08629 .08544 .08458 .08374 . 08291
2.5 .08209 . 08127 . 08046 .07966 .07887 .07808 .07730 .07654 .07577 . 07502
2.6 .07427 .07353 .07280 .07208 .07136 .07065 .06995 .06925 .06856 . 06788
2. 7 . 06721 .06654 .06588 . 06522 .06457 .06393 .06329 .06266 .06204 .06142
2.8 . 06081 .0602i .05961 .05901 . 05843 .05784 .05727 .05670 .05613 .05558
2.9 .05502 . 05448 .05393 . 05340 .05287 .05234 .05182 .05130 .05079 .05029
3.0 . 04979 .04929 . 04880 .04832 . 04783 .04736 .04689 .04642 .04596 .04550
3.1 .04505 .04460 . 04416 . 04372 . 04328 .04285 .04243 .04200 .04159 .04117
3.2 .04076 .04036 .03996 .03956 . 03916 .03877 .03839 .03801 . 03763 .03725
3.3 .03688 . 03652 . 03615 . 03579 . 03544 .03508 .03474 .03439 . 03405 .03371
3.4 .03337 .03304 . 03271 .03239 .03206 .03175 . 03143 .03112 . 03081 .03050
3.5 .03020 .02990 . 02960 . 02930 . 02901 .02872 .02844 . 02816 . 02788 .02760
3.6 .02732 .02705 . 02678 .02652 .02625 .02599 .02573 . 02548 .02522 .02497
3. 7 .02472 .02448 . 02423 .02399 .02375 .02352 .02328 .02305 .02282 .02260
3.8 . 02237 .02215 .02193 .02171 .02149 .02128 .02107 .02086 .02065 .02045
3.9 .02024 .02004 . 01984 .01964 .01945 .01925 .01906 .01887 .01869 .01850
4.0 .01832 .01813 . 01795 .01777 . 01760 . 01742 .01725 .01708 . 01691 .01674
4.1 .01657 .01641 . 01624 .01608 .01592 .01576 .01561 .01545 .01530 .01515
4.2 .01500 .01485 . 01470 .01455 . 01441 .01426 . 01412 .01398 . 01384 .01370
4.3 .01357 .01343 . 01330 .01317 . 01304 .01291 .01278 .01265 .01253 .0124o
4.4 .01228 .01216 .01203 .01191 .01180 .01168 . 01156 .01145 . 01133 .01122
4.5 . 01111 .01100 .01089 .01078 .01067 .01057 . 01046 .01036 .01025 . 01015
4. 6 . 01005 .00995 .00985 .00975 . 00966 .00956 . 00947 .00937 . 00928 . 0091 9
4.7 .00910 .00900 .00892 .00883 .00874 .00865 .00857 .00848 .00840 . 00831
4.8 .00823 .00815 .00807 .00799 .00791 .00783 . 00775 .00767 .00760 .00752
4.9 .00745 .00737 .00730 .00723 .00715 .00708 .00701 .00694 .00687 .00681
5.0 .00674
•
INDEX
A Bell Telephone System, 7, 147, 148,
151
a, definition, 14 Blondel, André, 75
properties, 14, 405 Breisky, J. V., 302
Alexanderson, E. F. W., 5, 383, 384
Allcutt, C. T., 271, 275, 284 e
Aluminum Company of America,
419 Cables, general, 198
Ammeter, positive- and negative- dissymmetry in circuit, 198, 203
sequence, 282, 283 independence of sequences in
adjustment and testing, 295 symmetrical systems, 198,
bridge-type, 286 203
common form, 284 sheath bonding, 202
effect of harmonics, 294 submarine, 219
elimination of zero-sequence, typical constants of, 219, 220
289- 291 Cables, single conductor, insulation
frequency error, 292 thickness, 410
simultaneous measurement, 292 positive- and negative-sequence,
transient error, 294 reactance, 207, 220, 409
zero-sequence, 281 resistance, 199- 202, 220, 409
Anaconda Wire and Cable Company, shunt capacitive reactance, 215,
419 220
Applications, 4, 296- 301 sheath thickness, 413
Armature leakage reactance, 78 zero-sequence, impedance, 214,
Arnold, A. H. M., 202 220
Autotransformer, equivalent circuit, resistance, 203, 220
120 shunt capacitive reactance, 215-
conversion of impedances, 121 217, 220
series im pedances, 125 Cables, three-conductor belted, in-
ungrounded, with tertiary, 128 sulation thickness, 412
positive- and negative-sequence,
B reactance, 203, 204-206
Balancers, phase, rotating, 382, 383 resistance, 199-203
negative-sequence e.m.f., 383 shunt capacitive reactance, 215-
power-factor correction with, 217
384 sheath thickness, 413
series-impedance, 384 table of characteristics, 204-206
shunt-impedance, 385 zero-sequence, impedance with
static, 380, 381, 382 sheath and ground return, 211
Bauch transformer, 234 impedance with sheath return
Bekku, S., 7 only, 214
427
Cables, zero-sequence, reactance, Decrement curves, assumptions, 101

207 description, 101
resistance, 203 example of calculation, 84
shunt capacitive reactance, 215- line-to-ground short-circuit, 104
217 line-to-line short-circuit, 103
Cables, three-conductor shielded or mínimum short-circuit current,
type H, 217 104
effect of shields, 218 r.m.s. total current for machines
impedance, 217 without dampers, 104
insulation thickness, 411 standard, 100
!3heath thickness, 413 three-phase, 102
shunt capacitive reactance, 218 unbalanced short-circuits, 100
Calculating board, description, 37 Dissymmetry, cable circuits, 198,
determination of simplest equiva- 203
lent network, 239 transmission circuits, 137-139,
fault calculations, 38, 52, 242, 253, 163, 168-170
(See also Short-circuit calculations) Distribution factor, current, 55, 56,
Campbell, G. A., 140 57
Capacitive reactance, cables, 215 voltage, 62
. long lines, 183 Doherty, R. E., 7
tables, 424, 425 Dovj ikov, A., 7
(See also Susceptance) Dubusc, R., 270
Carson, J. R., 7, 140, 141, 142, 146, Dwight, H. B., 176
148, 164, 208, 209
E
Cascade transformers, 66
Chubb, L. W., 271 Earth resistivity, 141, 146- 150
Circuit constants, ABCD, 182, 416, Earth return current, distribution,
418 148
Clarke, Edith, 238, 254 equivalent depth, 146
Complex numbers, 9 follows line, 150
Conjugate vector, 5 Earth-return-circuit impedance,
Convention of current flow, 21 Carson's formulas, 140
Conversion of currents and voltages curves, 142, .144
due to transformers, 64 parallel conductors, 155
Current and voltage, positive sense simplified formulas, 151, 155
same in all networks, 21, 30, 54, in terms of equivalent depth, 146
59 (See also Zero-sequence imped-
Currents, definitions, ground, 39 ance)
residual, 39 Equivalent networks (or equivalent
under fault conditions, range of, circuits), positive- or nega-
ground, 224, 232 tive-sequence, autotransform-
grounding, effect of, 234 er, 120
line, 224, 225, 228-232 fault calculation, 55, 57, 58
interconnected star-delta trans-
D former, 131, 132
Dallye, F. R., 419 long transmission line, 183
Damper windings, 78, 89, 91 parallel circuits with distributed
short-circuit current for machines inductive and capacitive
without, 104 coupling, 191
•
INDEX 429
Equivalent networks (or equivalent Fault impedance, 223, 248

circuits), positive- or ncga- representation in sequence net-
tive-sequence, simultancous works, 4G-43
fault, 253-256 rcsistance, 225
susceptance, 185 Faults, power system, double line-to-
three-phase transformer, 117,119 ground, 35, 42, 46, 49, 69
three-winding transformer, 114 line-to-line, 35, 43, 47, 50, 69
two-winding transformer, 112 range of curren ts, bearing on
reduction or conversion methods, various applications, 222
elimination of star point, 414 curves, 228, 229, 230
method of sctting up simplest range of sequence imped-
network, 194 ances, 225
star-dclta transformations, 414 systems giving maximum cur-
threc branches connected instar rent ratios, 232
and inductively coupled, 415 range of voltages, bearing on
two inductively coupled cir- various applications, 222
cuits, 415 curves, 22~ 229, 231
zero-sequence, autotransformer, range of sequence imped-
120, 125- 128 ances, 225
cable with sheath, 210 systems g1vmg maximum
interconnected star-delta trans- voltage ratios, 232
former, 131- 133 simultancous, 238
interconnected star transformer, single line-to-ground, 31, 40,
113 44, 48, 57
neutral impedance, 11t three-phase, 36
parallel circuit with earth re- Ferraris, Gallileo, 5, 338
turn coupling, 159, 160 Field leakage, 78
series impedance, 111 Fortescue, C. L., 5, 6, 7, 22, 271, 286,
shunt impedance, 111 308, 327, 329, 384
susceptance, 185 Four-phase systems, 332
three-phase transformer, 117,119 Frick, C. W., 419
three-winding transformer, 117,
119 G
transmission line with ground General Cable Corporation, 419
wires, 165, 166; 171 Generator, fault calculation, bal-
two-winding transformer, 112, anced faults, 37, 74-88, 1ü0-
113 105
(See also Networks, general) unbalanced, 44, 100, 102, 103
Evans, R. D., 7, 222, 261, 262, 270, (See Short-circuit calculations;
271, 382 Decrement curves)
Excitation, determination by Generators, single-phase, 380
A.I.E.E. method, 81 Gencrators, synchronous, imped-
quick response, 300 ances of, negative-sequence
Exciter voltage, variable, 108 reactance, 80, 88, 99
Exponential functions, table, 426 method oí test, 89, 90, 95
negative-sequence resistance, 90
F method of test, 95
Fault calculations (see Short-circuit shaft power associated with, 92,
calculations) 93
Generators, synchronous, positive- Ground wires, (see also Transmis-

sequence reactances, sub- sion lines with ground wires)
transicnt (direct axis), 75, 78, Grounding, effcct of sequence im-
88, 89, 97, 99, 105 pedances, 234, 236
subtransient (quadrature axis), zero-sequence impedance ratio,
75, 79, 88, 89 234, 236
synchronous, 75, 76, 99
transient, 75, 77, 79, 97, 99 H
zero-sequence impedance, 96
method of test, 97, 99 Halperin, H., 202
(See also Multi-phase sys- Harder, E. L., 377, 379
tems) Harmonics, impedances to, 342
Generators, synchronous, interna! sequences of, 344
voltages of, subtransient, 75, 78, Hauffe, G. von, 270
81, 83, 85 Heumann, G. W., 261, 262
synchronous, 75, 81, 83, 84 Historical review, 5
transient, 75, 77, 81, 83, 85, 108
Genkin, V., 265, 266, 290 I
G.M.D. (geometric mean distancc),
reactance calculations, for- Impedance-drop constants, 241
mula, 139 analytical expressions, 246, 251
three-phase circuit, 137, 138 determination, 242
· susceptance calculations, 184 sequcnce constants, 245, 246, 251
G.M.R. (geometric mean radius), Impedances, symmetrical (see specif-
reactance calculations, 137, ic apparatus or circuits)
138, 157 unsymmetrical, general solution,
cables, 207 363
copper-clad conductora, 179 negative-sequence range, 225,
equivalent of interna! reactance, 227, 233
178 positive-sequence range, 225,
steel conductora, 176- 178 227, 233
susceptance calculations, 186 sequence components of, 364,
Genna n Electrical Standards, 265, 368
268 series with one phase, 378, 379
Gilman, R . E., 5 series with three phases, 377
Goldsborough, S. L ., 290 solution in combination with
Graphical constructions, 2, 17, 260, symmetrical impedances, 375
263, 264, 265 zero-sequence range, 225, 227,
Ground currents, distribution deter- 233
mined by zero-soquence net- Independence of sequences, 6
work, 40 cable systems, 198
rotating machines, 23
division with ground wire, 165,
static networks, 22
168, 169, 178
Induction motor, diagram, 91
Ground wires, Carson's formula, 164 operation with open-circuited
current division with earth, 165, rotor phase, 358
168, 169, 178 operation with open-circuited sta-
effect upon positive- and negative- tor phase, 351
sequence impedance, 163 unbalanced rotor impedances, 354
INDEX 431
Induction motor, unbalanced volt- Mackerras, A. P., 7
ages to stator, 345 Magnusson, C. E., 74
equivalent circuit, 347, 348 Mayr, 0 ., 140
heating, 351 Measurement, of currents in power
loss, 350, 351 system sequence networks, 38
rotor resistance, 348 of sequence quantities (see Aro-
shaft power, 347, 349, 351 meter; Voltmeter; Watt-
torque, 347, 349- 351 meter)
Internal reactance of conductora, of voltages in power system
relation to G.M.R., 178 sequence networks, 38
Interna! voltage method, 105 Miller, C. E., 202
accuracy, 106 Multi-phase systems, 328, 336
alternating-current component, impedances of rotating machines,
107 337
assumptions, 106 negative-sequence, 341, 342
direct-current componcnt, 107 other sequences, 341, 342
summary, 108 positive-sequence, 340, 342
Mutual impedance of lines with
J earth return, 155, 158
Jackson, D . C., 308 N

Jonas dissonance coil, 234
National Electric Light Association,
K 7, 147, 148, 151
Negative-sequence, ammeter (see
Kennelly, A. E., 261, 308 Ammeter)
Kilgore, L. A., 7, 75 component, charts, 267
Kirchhoff's laws, first, 26 expression, analytical, 14, -263
second, 27 zcro-sequence is zero, 261
graphical construction, 17, 263,
L 264
zero-sequence is zero, 265
Lamme, B. G., 5, 361 terms of amplitudes, 261, 262
Lawrence, R. R., 91 impedance, definition, 24
Lenahan, B. E ., 279 rotating machines, 23, 24
Lewis, W. A., 136, 290, 419 static networks, 22, 24
Line-to-line voltages, from sequence (See also specific apparatus or
voltages, 63 circuit) ·
Line-to-neutral voltages, from se- networks for power systems, 28
quence voltages, 62 relays (see Relay)
Loads, representation of, 29 segregating networks (see Net-
Locomotive, phase converter, 386 work, s~gregating)
system of vectors, 2, 3, 14, 15, 17
M voltage, measurement of, 39
(See also Negative-sequence
Machine constants, impedance to component)
harmonics, 343 voltmeter (see Voltmeter)
typical values, 99, 406 Nesbit, William, 182
(See also Generators, synchronous) Network, conversion formulas, 417
Network, equivalent (see Equiva- Phase balancing, single-phase load,

lent networks) 380
restraints for simultaneous faults, unsymmetrical transformer taps,
243, 244, 248 381
sequence, connections to repre- Phase converter, 386
sent faults, 31- 36 Phase quantities, determination from
independence, 22, 23, 27 sequence quantities, 16
measurement of sequence volt- numerical example, 18, 59
ages, 38 Pierce, R. T., 270
representation of fault imped- Pollaczek, F., 140
ance, 40-44 Positive-sequence, ammeter (see Am-
of generators, 30 mcter)
of shunt loads, 29 component, charts, 268
segregating, current, 283-296, 303, expressions, analytical, 14, 263
304 zero-scquence is zero, 261
polyphase, 302-304 graphical constructions, 17, 263,
voltage, 273- 280, 292- 296, 302, 264
303 zero-sequence is zero, 265
solution, 1 terms of amplitudes, 261, 262
N etworks, general, analytical expres- impedance, definition, 24
sion, 239, 241 rotating machines, 23, 24
determination of branches, 239 static networks, 22, 24
simplest equivalent form, 238 (See also specific apparatus or
unsymmetrical static, 366 circuit)
sequence components, 368 networks for power systems, 28
series impcdances, 377 regulator, 271, 298
Neutral buses, 29, 30 segregating networks (see Net-
Neutral impedance, sequence imped- work, segregating)
ance of, 111 system of vectors, 2, 14, 16, 17
Nickle, C. A., 7 voltage, measurement of, 38
Norman, H. M., 261 (See also Positive-scqucnce,
Notation, 403 component)
voltmeter (see Voltmeter)
o Power, flow during unbalanced
conditions, 324
Oberdorfer, G., 238 Power factor, polyphase circuit, 309,
Open-circuited phase, one, 378 318
two, 379 single-phase circuit, 309
unbalanced circuit, 310
p Power factor meter, 318
elecirodynamic instrument, 310
Park, R . H., 7, 75, 89 . phasc-angle correction, 311
Per unit notation, 76 Power measurements, conventional
Peters, J. F., 5, 114 two-element for three-wire serv-
Petersen coil, 234 ice, 316
Phase-angle correction of watt- elements required, 315
meters, 311 special connection for balan.ced
Phase balancing, with capacitors, voltages, 317
381 three-element, 315
INDEX 433
Power quantities, analytical expres- Rudenberg; Reinhold, 140, 149, 209,

sion, 306 219
analyzed by sequence-compo-
nents, 314 S
instantaneous, 306 Sah, A. P. T., 261
polyphase circuit, 309, 314 Schurig, O. B., 419
reactive volt-amperes, 307 Sequence, components at point ol
single-phase circuit, 308 fault, 247, 248
unbalanced ci~cuit, 309, 310, 314 impedance terms, 24, 364, 368,
Power supply, single-phase loads, 372, 373
379 independence of, in symmetrical
Power systems, grounding, 234, 236 networks, 22, 374
range of fault currents, 222- 230 measurements, adjustment and
range of fault voltages, 222-231, testing, 295
236 errors, frequency and transient,
rl:inge of sequence ~mpedances, 2!)2-294
225, 233 po1yphase current, 304
Proximity effect in cables, 199 polypbase voltage, 302
Pulsating field, 5 simultaneous current, 292
simultaneous voltage, 280
R (See also Voltmeter; Ammeter;
Reactance, system, for use with Wattmeter)
decrement curves, 102 operator, 6
Reactive-factor meter, 310 power quantities, flow under un-
Reactive volt-ampcre meter, cross- balance, 324-327
connection schcmes, 318 pbase quantities in terms of, 312
electrodynamic instrument, 310 watthour meter, 323
phase-angle correction, 312 wattmeter, 320
Reactive volt-amperes, choice of Series im pedances, sequence im ped-
sign, 308 ance of, 111
measurements, 318 .Sheath currents ·in cables, division
polyphase circuit, 309 with earth, 211
single-phase circuit, 308 single conductor, 200, 207
unbalanced circuit, 310 elimina.tion, 202
vector expression, 307 three-conductor, 203
Relay, elimination of zero-scquence Short-circuit calculations, anulytical
current, 289, 290 discussion, 44, 52
negative-sequence current, 271, assumptions, conventiona1, 39,
300 52, 101
phase balance, 302 calculating boards, 37-39
reverse phase, 302 comparison with test results, 70
type CQ, 300, 301 connection of sequen ce networks,
Relaying, faulty action of ground 30-36, 4Q-43, 405
relays, 238 effect of charging current, 72
simultaneous fault, 238 equivalent impedances and dis-
Robertson, B. L., 7, 75, 89 tribution factors, 55- 57, 62
Rotating machines, independence of example of calculation, 48, 52, 84
sequences in symmetrical, 23 excitation, constant, 74
(See also Generators) variable, 108
Short-circuit calculations, gener- Simultaneous faults, two locations,

ators, three-phase, 74, 102 equivalent circuits, 253
unbalanccd, 100, 103, 104 general method, 238-249
loading condition, with load, 81 single source of e.m.f., 252
no load, 74 two sourccs of e.m.f., 243
outline of calculations, 54 Single-line diagram, 29, 54
sources of e.m.f., severa! sources Six-phase system, 329, 331, 332
and several simultaneous Skin effect, cable circuit, 199
faults, 243, 253 transmission circuit, 137
several sources and single fault, Skinner, M. E., 114
257 Slepian, J ., 5, 271, 308, 385
single source and simultaneous Smith, R. M., 238
faults, 251 Solution of higher order equation, 6
single source and single fault, Specific inductive capacity, S.I.C.,
26, 52 215
systems reducible to single equiva- Stability, positive-sequence voltage
lent generator, 83, 100 at fault a measure of, 236
two or more equivalent generators, Star-delta transformations, conven-
1
105, 243, 250, 256 tion of current flow, 21
interna! voltage method, 105 current and voltagc relations, 19-
(See also Decrement curves; Gen- 21, 64, 65, 404
erators, interna! voltages of; Star-star transformation, 64
Short-circuit currents; Si- Static networks, general, 366
multaneous faults; Faults, independence of sequences m
power system) symmetrical networks, 22,
Short-circuit currents, alternating- 329, 336
current components, 75 Steinmetz, R. B., 419
subtransient, 78, 83, 85, 107 Stokvis, L. G., 5
sustained, 7 5, 85 Symmetrical component systems,
transient, 77, 83, 85, 107 advantages, 3, 4
direct-current component, 80, 83, . combination to forro phase quan-
85, 88, 107 tities, 16
generator, 44, 48 general treatment of polyphase
r.m.s. total, maximum, 80, 85, 86, system, 336
88 multi-phase systems, 328-333
mínimum, 104 reduction into severa! simpler
(See also Short-circuit calcula- systems, 331, 332
tions; Decrement curves) resolution from phase quantities,
Shunt impedances, sequence imped- 16
ance& of, 111 simpli:fi.cation possible with sym-
Shunt load.s, representation in se- metry, 3, 375
queuce uetworks, 29, 30 three-phase systems, 14, 15
Simmons, D. M., 198, 199, 410, 412, unsymmetrical systems, 4
419 Symmetrical components, applica-
Simultaneous ff'-ults, 238 tions, general review, 7, 8 .
relaying, 238 metering, 296-302
three or more locations, 250 canonical forro of setting up, 329
two locations, ~urrent and voltage characteristic angle, 331
distributiou, 253 · characteristic operator, 329, 330
INDEX 435
Symmetrical components, general Transmission lines with ground

treatment of polyphase circuits, wires, impedance (zero-se-
336 quence), comparison of mag-
independence of sequences, 329, netic and non-magnetic
336 ground wires, 178
introduction by Fortescue, v copper-clad ground wires, 179
Symmetrical coordinates, 6 effect of neglecting resistance,
Symmetrical systems, definition of 168, 169
symmetry, 375 general method, 170
Synchronous machines (see Genera- single circuit and one ground
tora, synchronous) wire, 166
single circuit and two ground
T wires, 169
steel ground wires, 176
Time constants, decrement curves, twin circuit and two ground
102 wires, 170
definition, 79 two twin circuits and four
direct-current, 80, 87, 108 ground wires, 171
subtransient short-circuit, 75, 77, typical values, 179, 408
80, 87 impedances (positive- and nega-
transient open-circuit, 79, 83, 98, tive-sequence), 163
102 errors due to dissymmetry, 163,
transient short-circuit, 75, 79, 83, 168, 170
87, 109 typical values, 179, 408
Transformers, conversion of volt- susceptance (positive- and nega-
ages and currents, 64 tive-sequence), 189
equivalent circuit, three-winding, (See also Transmission linea,
114, 406 with~ut ground wires)
two-winding, 112 susceptance (zero-sequence), 189
impedances of, grounding, 113
mutual, 188
interconnected star, 113
Transmission lines (without ground
interconnected star-delta, 131
reactances, typical, 113, 406 wires), capacitive reactance,
representation in sequence net- 424, 425
works, 29, 30 impedance (zero-sequence, correc-
single-phase, 112- 117 tion for height, 148
symmetrical booster, 369 earth resistivity, 149
three-phase, 117- 119 equivalent circuits, 159, 160
three-winding, 114, 117 mutual impedance, 155, 158,
zigzag, 113 407
symmetrical, 369 parallel circuits, 155, 158
unsymmetrical, 381, 382 simplified formulas, 149-155,
Transmission conductora, tables of
407
characteristics, aluminum,
single circuit, 149-155, 407
420, 424
Anaconda hollow, 422, 425 tables, 42ü-423
copper, 421, 424 typical values, lfsO, 408
General Cable Type HH, 423, (See also Earth-return-circuit
425 impedance)
Transmission lines (without ground Vector representation, 12

wires), impedances (posi- absolute value, 11
tive- and negative-sequence), combination of vectors from se-
G.M.D., 137, 138, 139 quence components, 16
G.M.R., 137, 138, 155, 157 conjugate vector, 10
mutual impedance, 139 dircction of rotation, 2
parallel circuits, 139 instantaneous value, 12, 13
single circuit, 136 powcr expression, 306
tables, ,420-423 reactive volt-ampere expression,
typical values, 180, 408 307
unsymmetrical spacing, 137, 139 resolution of vectors into scquence
susceptance (zero-sequence), mu- components, 16, 17
tual, 188 sequcnce of vector, 2
self, 187 Vectors, systems of three-phase, 259
susceptances (positive- and nega- determination from amplitudes,
tive-sequence), G.M.D., 184, 259
186 determination of linc-to-neutral
G.M .R ., 186 from line-to-line quantities,
mutual, 187 259, 260
self susceptive, 183, 189 determination of scqucnce com-
tables, 424, 425 ponents, 13- 16
typical values, 408 (See also particular sequence)
Transmission lines, long, positiv.e- Voltage regulator, positivc-sequence,
and ncgative-sequence con- 271, 298
stants, ABC constants, 182 Voltages balanced, obtained from
equivalent 1r, 183 singlc-phase source, 305, 386
zero-sequence constants, ABC (See also Phase balancing)
constants, 190 under fault conditions, range of,
equivalent 1r, 190 226-232
parallel circuits with distrib- generator, negative-sequence, 33
uted inductive and capacitive positive-sequence, 30, 33
coupling, 191 (See also Generators, synchro-
Twelve-phase system, 329, 331, 332 nous, internal voltages of)
Two-phase systems, four-wire, 332 produced by dissymmetry, 33,
three-wire, 335 324
Voltmeter, positive- and negative-
u sequence, 273
adjustment and testing, 295
Unbalance factor, currents and volt- common forro, 27 4
ages, 268, 269 conversion to other sequence,
Unsymmetrical componcnts, Stokvis 277
proposal, 5 elimination of transformers, 279
Unsymmetrical system, general case, frequency error, 292
372 graphic, 297
lines, 407 harmonics, errors due to, 294
second forro, 278
V
simultaneous measurement, 280
Vector a, definition, 14 transient, errors dueto, 294
properties, 14, 404 zero-sequence, 272
INDEX 437
w Zero-sequence, component, elimina-

tion from phase quantities, 263
Wagner, C. F., 7, 78 current, elimination from phase
Watthour meter, positive-scquence, currents, 289, 290
323 measurement in sequence net-
Wattmeter, cross connection for works, 39
measur~ng reactive volt-am- impedance, basis for single-line
peres, 318 diagram, 29
electro-dynamic instrumenta, 310 definition, 24
phase-angle correction, 311 rotating apparatus, 24
sequence conncction, 320 static networks, 24
threc-phase circuit, 315- 317 (See also specific apparatus or
W estinghouse Elcctric & Manufac- circuits)
turing Company, 297, 301 introduction of, 6
Wollaston, F., 203 networks for power systems, 28
Wright, S. H., 7, 75, 222 segregating networks (see Net-
works, segregating)
z system of vectors, 2, 3, 15, 16
determination by graphical con-
Zero-sequence, ammeter (see Aro- struction, 17
meter) numerical example, 18
component, determination from voltage, measurement in sequence
line-to-neutral quantities, 16, networks, 38
17, 262 voltmeter (see Voltmeter)

Wagner Evan

Uploaded by

Copyright:

Available Formats

Wagner Evan

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Wagner Evan

Uploaded by

Copyright:

Available Formats

t

USED AS A TEXT IN THE DESIGN COURSE

McGRAW-HILL. BOOK COMPANY, INc.

McGRAW-HILL BooK CoMPANY, INc.

All rights reserved. This book, or

to working out, as a part of the text, illustrations of the appli-

lNTRODUCTION, BY c. L. FORTESCUE . . . . xiii

In balanced polyphase problems the phase displacement

1 have had the pleasure of close association with the authors

!·• · INTRODUCTION AND HISTORICAL DEVELOPMENT

Consider for a moment the balanced vectors in Fig. 1 (a)

(01) Posii'ive-Se'fuence (b)

(e) Ne~c:ftive- seqvenc:e (ol)

FIG. 1.-Sequence components and their combinations to form unbalanced phase

tional rotation of vectors in the counter-clockwise direction the

positive-sequence, currents of the negative-sequenc~ only voltage

railway ~-.:;ystcms, particularly from the standpoint of phase-

The method of symmetrical components requires no additional

4. Laws of Complex Numbers.

This is called the polar forro, in which 8 is the argurrumt or

If E8 be expanded into an infinite series,

Separating the real and imaginary components,

The real and imaginary components may be expressed in

from which it follows that the product of two complex numbers,

Division, the inverse of multiplication, can be accomplished

The absolute value of the resultant is equal to the quotient of

5. Vector Representation of Altemating Quantities.

so that, if we let V2Eeia = v2E and V2Ecia = V2E

It will be noted from this equation that the instantaneous

to one of two degrees of freedom. Going a step further, it is

1 = EiO = 1.0 + jO.O

This system of vectors is called the negative-sequence system.

Fro. 4.- Symmetrical vector systems.

and e phases of the positive-sequence voltages be applied to

.. 7. Resolution of Three Vectors into Their Symmetrical Compo-

OrigÍt'IC'I) Zero -sequence

FIG. 5.-Grapbical construction for tbe determination oí sequence components.

Thus these three fundamental equations defining t he sequence

lllustration. Suppose the given voltages are balanced, for example,

The zero- and negative-sequen ce components of voltage in a balanced

= ![(60 +jO) + ( -0.5 + j0.866)( +45 - j75) +

= ![(60 +jO) + (-0.5 - j0.866)(45 - j75) +

By reference to equation (28), it may be seen that

Similarly it may be shown that

Since the delta voltages must by their very nature form a

and rotating machines, the present discussion merely undertakes

The positive-sequence currents thus produce only positive-

10. Sequence lmpedances.

CALCULATION OF UNBALANCED FAULTS

Probably the most important application of symmetrical

a(la2 1 + 1a2 11 + 1 a2 111 ) = O (39)

12. Kirchhoff's Second Law.

tie between these systems will be the terminal conditions at the

point of fault. The impedances to this sequence are radically

neutral is ungrounded, the load impedance does not appear in the

15. Connection of N etworks to Represent Faults.

From equations (27), (28), and (29) of Sec. 6, the symmetrical

lxo = !(lx + 111 + lz) = i(lz) (45)

N ow from equation (48) it can be seen that the currents at

lb' =;~.}a [( -108.6. + j24.3) - (217.1 )] -1,128 + j84.2 =

le' = ;~.}a [(217.1) -· ( -108.6 - j24.3)] = 1,128 + j84.2