Simultaneous Faults on Three-Phase Systems

BY EDITH CLARKE*
Associate, A. I. E. E.

Synopsis.-The method of symmetrical components now so

extensively used to determine short-circuit currents and stability
limits during transient conditions for three-phase transmission
systems when a fault involving one or more of the three conductors
occurs at any one point of the system, has been extended to apply to
three-phase systems during simultaneous faults at two or more
points of the system.
A general equivalent circuit is developed to replace, in the positive
phase diagram, two simultaneous faults involving any combination of
the six conductors. An approximate equivalent circuit to be used
with the d-c. calculating table when resistance is neglected is also
given.
Special equivalent circuits are employed to rep,lace two simul-

taneous faults and the lines upon which they occur, when the lines
are unloaded feeders radiating from a common point or lines of
equal impedances bussed at both ends.
The methods and formulas given in this paper were developed in
answer to such questions as the following:
1. lVhich is a more severe shock to a system, a double line-toground fault on one circuit or two single line-to-ground faults
occurring simultaneously on two separate circuits?
2. Do simultaneous double line-to-ground faults which involve
the same phases, a and b, on two circuits, result in more or less
ground current than.faults which involve phases a and b on one circuit and phases b and c on the other?
* * * * *

W HEN double circuit towers carry two three-phase

circuits, disturbances may involve one or more
conductors of each circuit. From published
records' of the number of flashovers on double circuit
towers which have tripped out both circuits, and from
opinions expressed by operating engineers of various
power companies who have been consulted, it seems
reasonable to conclude that in the neighborhood of
20 per cent of the faults on double circuit towers involve
conductors of both circuits. In addition there are
instances where faults in substations have involved
conductors of circuits not on the same towers.
It seems worth while therefore, to have in convenient
form, methods for calculating short-circuit currents,
and of determining the stability limit of a system when
faults occur simultaneously at two separate and distinct
points of the system. The general case will cover
simultaneous faults at any two points of the system,
involving one, two or three conductors at each point,
while short circuits on parallel circuits on the same tower
will be a special case in which the two points of fault are
symmetrical with respect to the system, although they
will not be symmetrical with respect to ground unless

circuit becomes grounded, the voltages to ground on

the three conductors and the currents in the three lines
are no longer balanced. If the three unbalanced voltages to ground, Va, Vb and Vc, at the point of fault,
and three line currents flowing into the fault, Ia, Ib and
I,, are resolved into their symmetrical components,
it will be found that there are enough relations existing
between these six components to allow the positive
component of voltage at the fault, Val, to be expressed
in terms of the positive component of current in the
fault, Ial, and the zero and negative phase impedances,
Z0 and Z2 respectively, viewed from the fault. To do
this requires five equations.
Since there is no generated zero nor negative phase
sequence voltage, and the positive direction for all
component currents is taken towards the fault,
(1)
0 - VaO = IaoZo
= Ia2 Z2
0(2)
At a point of fault there are certain relations between the positive, negative, and zero components of
current which flow into the fault, and also between the
p
n
thef
Tese an s be t components
h
the fault. These relations between the components
of current and voltage, depending upon the type of
fault, provide additional equations connecting the
unknowns. For all types of fault there will be three
independent equations connecting the components of
current and voltage at the point of fault. These
equations are tabulated in Table I, and the manner of
their derivation shown in Appendix B.
From five equations
with
the four
ukow
1it SiXmayunknowns,
bek limnate
and
.ve Va,i and
VaZo, Va2t Ia>o and Ia2 may be eliminated, and

the faults

are o

the

same

phase

phases in bot

therfaults. are on the same phase or phases in both

circuits.
Mr. C. L. Fortescue has shown2 that any system of
three vectors may be replaced by three sets of balanced
components. The fundamental equations expressing
actual currents and voltages in terms of their symmetrical components, and expressing the symmetrical
components of current and voltage in terms of the
actual currents and voltages respectively are given
~~~~~~unknowns
in AppendixinA.Appendix A.
17
oVaiexpressed in terms of Ial thus:
SINGLE FAULT

Vai

K Iai

(3)

When one or more of the conductors of a three-phase where K is a function of ZO and Z2 depending upon the
*Central Station Engg. Dept. General Electric

Sceetd, N. Y'

o ietogon alsK=Z
2
Company,tp ffut
for line-to-line faults K = Z2, and for double- line-to-

I. For references see Bibliography.Zo9

Presented at the Middle Eastern District Meeting of the A. I. E. E.,
Pittsburgh, Pa., March 11j-13, 1931j.
31-60

ground fai'lts K 919

920

CLARKE: SIJULTANEOUS FAULTS ON THREE-PHASE SYSTEMS Transactions A. I. E. E.

Equation (3) will be satisfied, and the positive phase

current and voltage distribution may be determined
if the fault in the positive phase diagram is replaced3
by the shunt impedance K, the value of K being determined by Zo, Z2 and the type of fault.

Two SIMULTANEOUS FAULTS

The method of procedure outlined above for determining the positive phase current and voltage
distribution when a fault involves one or more of the
three conductors at a single point, may be followed for
simultaneous faults at two distinct points.
If the two points of fault are C and D and the conductors at C are a, b and c and those at D are a, f3 and ay
(a and a being conductors of the same phase, as are
b and 3, and c and y), the six components of voltage
and current at C will be Va0, Val, Va2, IaO, Ial and Ia2
and those at D will be Vao, Vall Va2, IaOy Iac and Ia2.
It will be shown that there are ten independent equations connecting these twelve unknowns. It is proposed to eliminate the eight unknowns VaO, VaO, Va2,
Va2, IaO7 Ia,0 Ia2 and Ia2 and to reduce the number of
equations to two, expressing the positive components
of voltage, Vai and Vai, in terms of the positive components of current, ILI and Ia., and the known zero and
negative impedances.
Zero Phase System. With a single fault, the zero
phase current in the fault varies directly as the zero
phase voltage to ground at the fault. When there are
two simultaneous faults, the zero phase current in either
fault depends upon the zero phase voltages at both
points of fault. In order to readily express the two
zero phase voltages in terms of the two zero phase
currents, the zero phase impedance diagram4 will be
simplified, remembering that all points of zero potential
for zero phase voltage may be considered bussed at the
same point, S. Equivalent circuits to replace two
parallel transmission lines with mutual impedance between them in the zero phase system are given in Appendix C. When there are more than two parallel]
lines it may not be possible to accurately represent
them by a simple equivalent circuit in the zero phase
system but an approximate equivalent circuit may
usually be obtained. When the zero phase system
can be represented by an equivalent impedance
diagram, it is always possible to reduce it to an
equivalent Y, connecting the two points of fault,
C and D, and the zero potential point, S. In some
systems this may be done by A - Y or Y - A transformations,5 but in complicated systems it may be
necessary to use an a-c. or d-c. calculating table to
determine the branches of the equivralent Y or A connecting the points C, D, and S.

Determination of Equivalent Y for the Zero Phase

Impedance Network by Means of the Calculating Table.
For an exact determination, the use of an a-c. calculating table such as the M. I. T. Network Analyzer6 is
necessary, in which case the currents and volt!ages

measured as described below would have to be in vector

form, requiring the use of a watt-meter as well as an
ammeter and voltmeter. For many purposes the
resistance in the networks can be neglected and the
impedances considered to consist of reactance alone.
In this latter case the d-c. calculating table may be
used. The general procedure in either case is as
follows:
(a) Set up the zero phase impedance network on the
calculating table. Apply a voltage to ground, Vc, at
point C with point D ungrounded, but all zero potential
points, S, grounded. Measure the total ground current,
Is. Current I, at C is the same as the total ground
current I, at S.
(b) Apply a voltage to ground, V,, at C with point D
grounded but all zero potential points, S, ungrounded.
Measure the current Id at D.
(c) Apply a voltage to ground, Vd, at point D with
point C ungrounded, but all zero potential points, S,
grounded and measure the current I,.
s

Iao+Icoo So
180. IP
v

FIG. 1--EQUIVALENT Y FOR ZERO PHASE SEQUENCE NETWORK

If the branches of the equivalent Y are CO, Doy and SO,

then

v.

CO + SO

(a)

CO + Do

(b)

Do + SO

(c)

Vd

From equations (a), (b) and (c) the branches of the Y

are determined.
Fig. 1 represents the equivalent Y for the zero phase
network connecting C, D and S with its branch impedances, CO, Do and SO. Let the positive direction
for zero phase currents be taken towards the faults C
and D. Then since there is no generated zero phase
voltage, by superposing the voltages due to the two
component currents, the following equations are

obtained:
(5)
0- Va0 =- VaO = IaO (CO + SO) + IaO SO
ac
a
(6)
0-Va0 =-Va0 =la(O So + lao (Do S)
Equations (5) and (6) express the zero components of
voltage at the two points of fault in terms of the two
zero phase currents flowing into the faults and the
branch impedances of the equivalent Y.
Rewriting equations (5) and (6) to, express the cur-

CLARKE: SIMULTANEOUS FAULTS ON THREE-PHASE SYSTEMS

September 1931

921

rents Lo and IaO in terms of Vao and Va0o, or from Fig. 1 Equations (12) and (13) are not independent of equations (10) and (11). There are four unknowns Va2,
directly:
Va2, 1a2 and Ia2 in the negative phase system, and two
50
Do + S0
(7) independent equations connecting them.
VaO zOO + VaO zoo
IaO
Relations between Components of Current and of
Voltage at Two Points of Fault. At each point of fault,
CO + sO
SO
the relations between the components of current flowing
(8)
IaO = VaO z - Vao
ZOO into the fault, and between the components of voltage
(9) to ground at the fault are independent of the rest of
where ZOO = CO Do + CO SO + Do SO
Equations (7) and (8) are not independent of equations the system. Table I gives three equations connecting
(5) and (6). There are four unknowns, VaO, VaO, IaO the components of current or of voltage at a point of
and lao in the zero phase system and two independent fault for various types of fault. When there are two
points of fault, there will be three equations connecting
equations connecting them.
Negative Phase System. In the negative phase sys- components of current or voltage at each point. By
ten, just as in the zero phase system, there is no gene- replacing a, b and c in Table I by a, , and y respectively,
rated voltage, and the positive direction for negative the three equations connecting components of current
phase currents is taken from the neutrals of the ma- or voltage at the second point of fault, D, are obtained.
Simultaneous Equations. Three equations expressing
chines towards the faults. The neutrals of generators
and loads are points of zero potential for the negative relations between the components of current or voltage
phase system. The negative phase network4 of a sys- at each of the two points of fault, two equations contem may be reduced to an equivalent Y or A connecting necting zero phase currents and voltages and two conthe points of fault C and D and the points of zero po- necting negative phase currents and voltages give the
tential, which may be considered bussed at a common ten equations needed to eliminate the eight unknowns
point, S. It is important to note that S, the point of VaO, VaO, IlaO IaO, Va2, Va2, Ia2 and Ia2 so that the
positive components of voltage Val and Vai may be
expressed in terms of the positive components of curs
rent, Ial and I al, and the known zero and negative
phase impedances of the system. Since the ten equaIa2 + iaS2s
tions are linear the two resulting equations can be put
132
12
~in the form:
?
Vaz
D
(15)
C
Vai = k Ial + mIal
(16)
FIG. 2-EQuIVALENT Y FOR NEGATIVE PHASE SEQUENCE
Val = n Ial + 1 Ial
NETWORK

zero potential for zero p'iase voltage is not necessarily

the same point of the actual system as S, the point of
zero potential for negative phase voltage. Neither are
the two P points identical. The same letters are used,
however, to obtain symmetry in form for the zero and
negative phase current and voltage equations.
From Fig. 2 which represents the equivalent Y for the
negative phase network connecting C, D and S, the
two negative phase voltages may be expressed in terms
of the two negative phase currents and the impedances
of the Y, thus:
- Va2 = Ia2 (C2 + S2) + Ia2 S2
(10)
(11)
- Va2 = 1a2 S2 1a2 (D2 + S2)
Rewriting equations (10) and (11) to express the
currents 1a2 and Ia2 in terms of Va2 and Va2, or from
Fig. 2 directly:
52
D2+ s2
Vao2
(12)
Ia2 =Va2
22
22

Ia a2
where

S2

222-- a2

C2 + S2

Z22 = C2 D2 + C2 S2 D2 52

where k, 1, m and n depend upon the branch impedances

of the equivalent Y's replacing the negative and zero
phase networks, and the particular combination of conductors involved in the simultaneous faults. It should
be noted that k, 1, m and n do not involve positive phase
impedances.
Equations giving k, 1, m and n for faults which
involve various combinations of the six conductors are
derived in Appendix D and tabulated in Table II.
Positive Phase System. The positive phase system
differs from the negative and zero phase systems because positive phase sequence voltages are generated
at various points of the positive phase system. In
general there will be as many separate positive phase
generated voltages as there are separate generators or
synchronous motors on the system. When the operating conditions just previous to the occurrence of the
faults are known, these generated voltages are determined in magnitude and in phase. For steady state
voltage or voltage behind
calculations, reactance
is required, while for transient
~~~~~synchronous the excitation

transient reactance
-calculations, the voltage behind
case the
in either from
should be known.
(13) generated
given
theinternal
are determined
voltages However,
(14) operating conditions.

922

CLARKE: SIMULTANEOUS FAULTS ON THREE-PHASE SYSTEMS Transactions A. I. E. E.

TABLE I

Relations between the symmetrical components of current in the fault and between the components of voltage to ground at the
fault for various types of fault.
The operators a and a2 have been defined in equations (13a) and (14a).

Case

Phases
involved

Type of fault

A.

(a) Line to ground

a........ IaO

..................

(b) Line to

Current relations

ground

..................

Ial; Ial = a2 Ia2

Iao = a Ial; Ial = a2 IaO
=

1a2 a2 Ial; Ial = a Ia2

c.c....... b, C ......... Iao = 0, 'aI = - Ial a2
b

(a) Line to line

(b) Line to

a,
gc
..............................
I
a Ial; Ial

be

line

1a2 = -

(e) Line to linen V.....................

C.

(a) Double line to ground.

(b) Double line to ground ........

(c)
D.

. .

Double line to ground ............

(a) Three-phase ...........

(b) Three-phase to ground ..

a,
a.

b,

+
V

= (I

m) I",, +

rn-n

= -

Va0 + a Va2)

(a2

.....................Va2 =

(I al- Ial)

Vao = Val, Va2

Vao

...

Va2

Val

aVal;
Val = =
a

..VaO

l Va2

(a2 Iao + a Ia2)

= a

Val Val
Vat; Val =

a2

Val; Val =
Val = 0, Va2 = 0 V
Val = 0, Vao = 0
= a2

VaO

Va2
VaO

Va2
o

The relations expressed in (19) and (20) are satisfied

by substituting for the faults a Y network having the
branch impedances (k - m), (1 - m) and m connecting
the points C, D and ground as shown in Fig. 3. In
this case the positive phase voltages and currents of
the system may be determined as in any other balanced
load distribution problem: that is, by calculation, or
by means of a calculating table, the equivalent Y re-

placing the two faults in the positive phase network.

C

k-r I-r

E>
Iai

(17)
FIG. 3-EQUIVALENT

!-aVaxi

+Iai4

Y REPLACING

PHASE SEQUENCE NETWORK

(Ial + I a)

.....

Ia2

.........

FAULTS IN POSITIVE PHASE SYSTEM

Equations (15) and (16) may be written
(a+
i)Val
(Ial
+ Il

= - a2

al = - (a
IaO + a2 Ia2)

Ial
'..Iao
b, C .

EQUIVALENT CIRCUITS REPLACING TWO SIMULTANEOUS

rnm-+2 n

.................... Va

Val

FIal= 0
Va2
Va2 a2 Val; Val
(Ia2 = - a2 Ial; Ial = -a Ia2
b, c ......... Ial = - (Iao + Ia2)

At point C there is a positive phase voltage, Val, to

neutral and a positive phase current, Ia1 flowing into
the fault. At point D there is a positive phase voltage,
V,12, to neutral and a positive current, I',, flowing into
the fault, the relations between voltages and currents
being given in equations (15) and (16).

(kfl)Iai+
- it) I,, +
= =
V.,Vai
(k

. .....

, .b.....

a, c .
a, b.

Val

ao

= - (Vao + Va2)

.......

B.

Val

Ial, Ia2 = Ial

a2 Iai; Ial = a Iao

Io
b.0..............................
b.
=

Ia2

(e) Line to ground

Voltage relations

FOR

Two FAULTS
SPECIAL CASE

IN

POSITIVE

OF m

Case II. rm and n unequal.

When m and n are not equal, the positive distribution of currents and voltage due to the faults can not
The last term of (17) and of (18) may also be written be obtained by inserting an equivalent Y or l\ between
n-rn
the points, C, D and ground.
2
A. General Equivalent Circuit. Fig. 4 shows an
circuit which may replace the two faults in
Case I. rnand
n equal.
~the positive ph-ase system. The equivalent circuit
consists of a Y having branch impedances k -n, I-mr
When rn and n are equal (17) and (18) become
2
(19) and
connected between points C, D and F,
Vai = (k -mr) Ial + rn (Ia1 + Icri)
V121 = (lo-rn)1122+rm(112+1121)
(20)2
+

(Ia1 -Ia)

(Iai-Ia;i)

(18)

~~~~~~~~~~equivalent

September 1931

CLARKE: SIMULTANEOUS FAULTS ON THREE-PHASE SYSTEMS

923

TABLE II

Values of k, 1, m and n to be substituted in equations (15) and (16) for simultaneous faults at two points of the system.
The operators a and a2 have been defined in equations (13a) and (14a), and Zoo and Z22 in equations (9) and (14).
For the sake of brevity let,

Zco
Zd,
Zcds

CO + SO + C2 + S2

(CO + C2) (D0 + D2) + (CO C2) (So + S2) + (D0 + D2) (S0 + S2)

Do +So +D2 +S2

Case E. Single Line-to-Ground Faults at Two Points.

(a) Phases a and a.
k = Zcs
n = SO + S2
(b) Phases b and a.
k = Zcs
n = a2 So + a S2
(G) Phases c and a.
k = Zs
n = a SO + a2 S2

m =

m = a SO + a2
1 = Zds

m = S2
1 = D2 + 52
m = a2 S2

1 = D2 + S2

in = a S2

1 = D2 + S2

Case G. Double Line-to-Ground Faults at Two Points.

(a) Phases b, c and ,B, y. *
Z22 (CO + S0) + Z00 (C2 + S2)
k =

Zeds

Z22 So + Zoo S2
Zeds
(b) Phases a, c and /, -y.
Z22 (Co +

So)

Zcds

I =

n =

+ Zoo (C2 +

S2

+ 3 SO

a2 So Z22 + a S2 ZOO
Zcds + 3 So 82
(c) Phases a, b and 3, -.
Z22 (Co + So) + Zoo (C2 +
k =
Zcds + 3 SO S2

S2)

Z22 SO + ZOO S2
Zcds
Z22 (D0 + So) + Zoo (D2 +

Zcds

SO Z22 + a2 S2 Zoo
Zcd, + 3 SO S2

S2)

m =

a2

So Z22 +

Case I. Line-to-Line Fault at C and Single Line-to-Ground Fault at D.

(a) Phases b, c and a.
kc = C2 +8S2
n = -8S2
(b) Phases a, c and a.
k = C2 +8S2
n = -a S2
(a) Phases a, b and az.
kc = C2 +8S2
*See "Equivalent Y vs. Equivalent

A"

page 927.

a S2

ZOO

Zcds + 3 SO S2

Z22 (D0 + So) + Zoo (D2 + S2)

Zcds + 3 So S2

Case H. Three-Phase Faults at Two Points.

k =0
n =0

82

S2)

Z22 (Do + So) + Zoo (D2 + S2)

Zcd, + 3 So S2

a So Z22 + a2 So Zoo
Zcds + 3 SO S2

n = - a2

S2

m = a2 So+ a S2
1 = Zd.S

Case F. Line-to-Line Faults at Two Points.

(a) Phases b, c and 3, -y.*
k = C2 + S2
n = S2
(b) Phases a, c and y,-.
k = C2 + S2
n = a S2
(e) Phases a, b and 3, y.
k = C2 + S2
n = a2 S2

SO + S2

1 = Zds

m =0

=0

m - -82
1 = Zdo
m = -a2
1 = Zds

82

m = -a

82

= Zdo

924

CLARKE: SIMULTANEOUS FAULTS ON THREE-PHASE SYSTEMS Transactions A. I. E. E.

TABLE II-Continued

Case J. Double Line-to-Ground Fault at C and Single Line-to Ground Faults at D.

(a) Phases b, c and a.
(CO + SO) (C2 + S2)
k

SO (C2 + S2) + S2 (Co + SO)

Zcs

n=

Z's

-SO (C2 + S2)

S2

(CO + SO)

(SO-S2)2

Zd

Zcs

ZCS

(b) Phases a, c and a.

(Co + So) (C2

S2)

czs

n=

(C2 + S2)

a2 So (C2

a So (C2 + S2) + a2 S2 (Co + So)

Class K. Three-phase Fault at C and Single Line-to-Ground Fault at D.

(a) Three-phase fault not involving ground.
k =0

SO)

(b) Three-phase fault involving ground.

k =0

a S2 (Co + SO)

S02 + SO S2 + S22
z

Do+SoD

C2 S2

C +S

m =0

n =0

and between F and ground an impedance,

Zds

+ S2) +

m =0

=0

(Co +

1
Z - So2 +SO S2 +S22
1 = Z ds ZcS

+aS2(CO +SO)

n = - ZC8

(e) Phases a, b and a.

(Co + So)
kc =

(C2 + S2) + a2 S2

zcs

a2 So (C2 +S2)

So

m - n

C S
CO + So + D2 + C2 + S2

Do + C

until the current through it to ground is double Ial the

or current entering the fault at C, then the current

n-m
If the
(Io,1 - Ial) will flow in themimpedance adjustable
Equa- (~-Ii ilfo nteipdne 2
I h
m 2 Xparalleled
~
~by an ~
~ voltage,
~ ~V,,. ~~~~n
tions (17) and (18) will be satisfied if current (Ial -Ial) impedance - is used, V must be adjusted until
2

vC

D-rnD

k-n

z11,

Ial+6Io im+n2
F}

II-lal1

Mn 2I,,
2
I

Vs

the current through it is double Ial, the current entering

the fault at D.
B. Approximate Equivalent Circuit for Use on the
D-C. Calculating Table. In short-circuit studies it is
customary to neglect capacitance and resistance, and to
consider all generators operating with no load excitations.

The system

calculating table.
FIG. 4-EXACT GENERAL EQUIVALENT CIRCUIT REPLACING TWO

FAULTS IN POSITIVE PHASE SEQUENCE NETWORK

can

then be represented

on a

d-c.

study of Table II shows that with resistance and

capacitance neglected, k and I have no real components,

but are positive reactive impedances larger in magnitude
m -n
than m and n. When m and n are equal, they also have
is made to flow through the impedance 2 .or if no real components and are positive reactive impedances; but when m and n are unequal they have
n -m real components which are equal and opposite, while
current (Ial - Iai) flows through the impedance 2 2 . their reactive components are equal in magnitude and
of the same sign.
If the voltage, V,, is adjusted in phase and magnitude
The error made by neglecting the real components of

September 1931

CLARKE: SIMULTANEOUS FAULTS ON THREE-PHASE SYSTEMS

925

m and n will ordinarily be less than the error made by

neglecting line resistances. When the real components
m - n
of m and n are neglected,
2
-0, and the general

obtained on the calculating table. Then the division

of positive phase currents between the faulted lines,
and the positive phase voltages at the points of fault
may be determined by calculation.
(1) Equivalent circuit for simultaneous faults on unloaded
equivalent circuit in Fig. 4 becomes a Y connecting
feeders radiating from a common point
and
with
branch
C, D,
G,
impedances (k -n), (I
Let the fault points C and D be on unloaded feeders
mi-n
Fi 5 The branches k - n, and radiating from a common point H, then referring to
and
Fig. 6A, let
Vh = positive phase voltage at H.
C
D
Val = positive phase voltage at C.
Vank-n I-rn
Vali
Ma
Vai = positive phase voltage at D.
ZCh = positive phase impedance between H and C.
Iai + I4l; m2 n
Zdh = positive phase impedance between H and D.
2
Ih = positive phase current at H.
G
Ial = positive phase current at C.

-m)fedrraitnfomacmnpit

FIG. 5-APPROXIMATE EQUIVALENT CIRCUIT REPLACING Two

FAULTS IN POSITIVE PHASE SEQUENCE NETWORK

Ial

positive phase current at D.

Then

while

(21)
(22)
(23)

Ial + Ia1
= I,l Z,h
Vh - Val = Ial Zdh

Ih

Vh - Va1

(1 - m) will be positive reactive impedances and therefore can be represented on the d-c. calculating table,
m

may be either positive or negative. If subtracting (23) from (22)

- Val + Val = Iai Z.h - 1.
laZdh
positive it can also be represented on the d-c. table.
subtracting (16) from (15)
It will be remembered that the branch impedance
Val -(Valfrom(k1- ) I,, - (1- m) I.,
2

connected to ground as are the neutrals of the adding (15) to (22) and (24) to (25)
V, = (Z,h + k) Ial + m Ia,
generators in the positive phase diagram. This im0 = (ZCh + k - n) Ial - (Zdh + 1 -m)
pedance is therefore in series with the generator reactance when there is a single generating source and may
2

(24)

(25)

is

be combined with it. When

is negative and it

is not possible to combine it with any other reactance,

this branch of the Y between P and G may be set to
zero and the distribution of currents obtained, these
currents to be increased by the ratio

vp

(26)
(27)

Ia,

la8

H
A

dh

l4ai+I%i

+n
2
where X, is the equivalent impedance between generator neutrals and P, and is found by dividing generator voltage by total ground current when P is
grounded.
m

FIG. 6A-UNLOADED FEEDERS

WITH

FAULTS

AT

AND

_
1h=Iai+

C. Special Equivalent Circuits. Lines upon which

G
the faults occur, together with the faults may be repre- FIG. 6B-EQUIVALENT IMPEDANCE REPLACING THE Two
sented by a particular equivalent circuit when the
FAULTS AND FEEDERS OF FIG. 6A
faults occur; (1) on unloaded feeders radiating from a
common point and (2) on lines of equal impedances Solving equations (26) and (27) for 'aland lal:
bussed at both ends.
Zd +1-in
dh i-(8
When the equivalent circuit represents the lines upon
lal = V h (Z
i-k
(28
which the faults occur as well as the faults, the sum of
+) (Zdh i- 1)-mn
the currents in the faulted lines, the total ground curc+krent, and the positive phase current and voltage
larl = Vh
Z(2i9k)
(9
distribution for the rest of the system may readily be
(Zch i- k) (Zdh i-+ )-m n

CLARKE: SIMULTANEOUS FAULTS ON THREE-PHASE SYSTEMS Transactions A. I. E. E.

926

adding (28) and (29) and replacing Ial + Ial by Ih

Let Ial = Ix + I4 and I , = Ir + It - Ix - I4.
Let Zcr=a, Zct=b, Zdr=C, Zdt=d and a+b+c+d=S.
Z,h + k + Zdh + 1- (m + n)
=
Vh
-m
(30)
Then
Ih
(Zch + k) (Zdh + 1)
Substituting Vh from (30) in (28) and (29)
Val = Vr aIb
(34)
Ial

Ial

1k

~~Zdh + i-rn

Zch + k + Zdh + i- (m + n)
Z,h + k- n
Ih Zch + k + Zdh + 1-(m + n)

From equation (30)

From
equation (30)
Zdh
Vh _ (Zch k) (Zdh + -)--mn
Ih
Zch + k + Zdh + i - (m + n)
Vh

Vai=Vtbc y(6
I"

alVrc(IrIx)
Vai,,Vt
d(It -Iy)

(31)

(7

Substituting the above values for Ial and I,, in (15)

(32) and (16)

(Ix + Iv) + m (Ir + It)

(k-rm)
V,ll = (n-I) (Ix I+ ) + I (I, + It)

Val

(33)

h= Zh = impedance between point H and

Ih
ground, which is the equivalent circuit sought.
Therefore, two unloaded feeders radiating from H
with two simultaneous faults may be replaced by a
single lumped impedance, Zhk, whose value is given by
equation (33). The total ground current and the
positive phase current and voltage distribution for the
system exclusive of the feeders themselves may then
be obtained. Knowing Ih and Vh, the positive phase
currents in the feeders, Ial and I x, may be obtained
from equations (31) and (32), and the voltages at the
points of fault, Val and V,,a, from equations (22) and
(23).
(2) Equivalent circuit for two lines which have two points
in common or are bussed at both ends and two
simultaneous faults
This case will include simultaneous faults on any two
lines which, on the positive phase diagram, have their
ends terminating at common points. Two feeders
radiating from a common point with impedance loads

(38)

(39)

From the six simultaneous equations (34)-(39) the

four unknowns Val, V,,a, I, and I, will be eliminated
and V, and Vt expressed in terms of Ir and It.

but

IX
cr

OR

Ialal

Zdr

Zdt

FIG. 7ASTWO
LINES

BUSSED

'

ENDS
TWO
AT BOTHWSTH

It
2

M-N

2Ir
-Ir

Vs

replaced by shunt impedances to ground, have the

FAULTED LINES OF FIG. 7A
ground as a common terminal point. Even when the
lines upon which the faults occur do not have two
Substituting (38) and (39) in (34) and (37) respecpoints in common by appropriate A - Y or Y-A te antransposng:
transformations it is often possible to represent them
IrMI
,:
k - m) Iz + (k-r) 4 (40)
by two equivalent lines which do have two points in V, = m Ir + m It + (a + kcommon.
Vt = l Ir + (d + I) It + (n-1) Ix + (n-I-d)Iv (41)
Fig. 7A represents two lines bussed at R and T with Subtracting the sum of equations (35) and (36) from the
faults at C and D, in which Zcr, Zct, Zdr and Zdt repre- sum of (34) and (37); and the sum of (34) and (39)
sent positive phase impedances between C and R, from the sum of (36) and (38):
C and T, D and R, and D and T respectively, and 0 = clr- d t-(a + c)
+ (b + d)r T
(42)
Vai
and
Vai
the
voltage
at
T,
C
and
Vr, Vt,
R,
D.O
I)I+(
)I
Ir and It represent the currents at F and T respectively, 0r m-c-1 .(
)I
(a+cI+k-I+l-m-n) Iz (ki1-m-n) 4v
positive direction being towards the faults and Ial
(43)
and Iai the currents flowing into the faults from C Expressing Ix and 4v from (42) and (43) in terms of
and D.
-Ir and It, and replacing a + b + c + d by S.
(im+ b+ n]j
From Fig. 7Agrudcret[b++
Ir + It = I,al + I,al =4I,= total Frudcret[+[
Im
kn]I
Let Ir divide into Iz and Ir - Iz and It divide into Ir=
)-d(k n)] Ib d
S( + [b (1
(4
k+ -nn
(
)(
)
Iv and It- 4
(4
FIG. 7B-EXACT EQUIVALENT CIRCUIT REPLACING THE Two

September 1931

CLARKE: SIMULTANEOUS FAULTS ON THREE-PHASE SYSTEMS

1
F
[a(l-rm)- c(k-n)] Ir
_ L+[(a + c + d)(l-rm)+d(aA+c + k-n)]IJ
- S(k+l-nm-n)+ (a+c)(b+d)
Substituting (44) and (45) in (40) and (41)

ab(k+A+m+n) +2(a+Ab)(kl-mn) (
4ab +2 (a + b) (k + -m- n)

5())

[S(kI-rmn)-ac(m+n)+Ac(a+bA+d)kJ
L
J
+a(b +c +d)1+ ac (b +d)
Vr
Ir
I+
S (k +
m-n) (a + c) (b + d)
S (k + 1

-m -

n) + (a + c) (b + d)

It (

+adrnA-bcnA-cdk
Vt S (klS -(k +n)1-m-n)
+ (a + c) (b + +ablI
d)
ES (k-rmn) -b d (m + n) + d (aA+ b+ic) kl
( A-b (a - cA
bd (a +c)
I
+
+d) (+(a+A-c)
S(k +1-rn-n)
(b A-d)
A-

ab(k +lr+m +n) +2(a+b) (kl-mn)

4 a b +2 (a +b) (k

lines with the two simultaneous faults, rather than

vt

N Ir + L It

(48)

V= I+M()

(49)

If k, 1, m and n in Figs. 4 and 5 are replaced by

K, L, M and N as defined in equations (46)-(49) and
and Ial and Ia1 by Ir and It the desired circuits are
obtained. The general equivalent circuit is shown in
Fig. 7B.
(b) Special Case. The coefficients of It and Ir in
equations (46) and (47) respectively will be identical
if b c m + a d n = a d m + b c n, that is if m = n or
a d = b c. If either of these conditions is satisfied the
two lines with the two simultaneous faults may be
replaced by an equivalent Y having the branch impedances K - M, L - M and M.
Equivalent circuit for two transrnission lines of equal
irnpedances bussed at both ends and two
sirnultaneous faults
Since Zd. = Zcr = a and Zdt = Zc = b, c and d in
equations (46) and (47) may be replaced by a and b
respectively and the equations rewritten thus
a
Vr=[2 Ir

+ t)

Equations (50) and (51) will be satisfied if the two

lines and the two simultaneous faults are replaced by a

Ybetween the points R, T and G,b Fig. 8B, the branch

a

2 and

impedances of the Y being 2

a

b (k A
+ +m + n) +2 (a + b) (k
4ab

+2(aA+b)

(k

I-rmn)

+I-m--n)

In problems where the division of currents between

the two lines and the voltages at the points of fault are
required they may be determined by calculations from
the voltages and currents at the ends of the lines. If
the positive phase current and voltage distribution has
been determined with the equivalent Y replacing the
Zcr a

just to replace the two faults, if it is found advantageous D

to do so. To obtain this circuit, equations (46) and
(47) maybewritten:
V K
I, M It

+1-rm-n)

(51)

A-

(47)
(a) General Case. A comparison of the coefficient of
It in (46) with that of Ir in (47) shows that they are not
identical, and therefore a simple impedance Y or A
can not replace the lines with the two simultaneous
faults in the general case where m and n are unequal
and there is no fixed relation between the impedances
Zcr, Z,d, Zct and Zd t.
The general equivalent circuit or the approximate
equivalent circuit may be used to replace the two

Ir+I)]

b
Vt= 2 i

S (kl-rn) + bcn+-ad n A cd
+

927

Zrt=b

Zrp

Ir

Zdr=a

Zdt=b

p Zpt 2

pg

4 ab+2

(a+b) (kIt-rn-n)

8A-SPECIAL CASE OF FIG. 7A WHEN LINES HAVE EQUAL

IMPEDANCE AND FAULT POINTS ARE EQUIDISTANT FROM ONE END

FIG.

FIG. 8B-EXACT EQUIVALENT CIRCUIT FOR FIG. 8A

lines and the two simultaneous faults, then Vr, VI, I,

and It will be known. I. and I, may be obtained by
substituting the values of Ir and I, in equations (44)
and (45); Va1 may be obtained from equations (34) or
(35); and Va1 from (36) or (37).
Equivalent Y vs. Equivalent A. In the zero and
negative phase systems, equivalent Y's were used to
replace the network connected between the two fault
points and the points of zero potential, assumed bussed
at S. Equivalent A's might have been used through
this paper. The choice of equivalent Y's and impedances has given, in the majority of cases, equations of
simpler form for the constants k, 1, rn and n than would
have been obtained with equivalent A's and admittances. In the case of double line-to-ground faults,
however, the solution with equivalent A's and admittances is a simple one and the form of the resulting
equations of special interest when the faults are on the
same phases. Let the admittances of the A\ replacing
the equivalent Y for the zero phase system be YOcd,
Y0c. and Yod, and those for the negative system be

928

CLARKE: SIMULTANEOUS FAULTS ON THREE-PHASE SYSTEMS Transactions A. I. E. E.

Y2cd, Y2c8 and Y2d,, the first subscript indicating the

sequence and the second and third the terminal points.
Replacing5 the impedances of the Y's by the admittances of the A's in equations (13d) and (14d) of Appendix D:
Ial = Val (Yocs + Y2cs + YOcd + Y2cd)
- Val (YOcd + Y2cd)
Ial = Val (Ocd + Y2cd)
+ Val (Yod, + Y2ds + YOcd + Y2,d)
These equations are satisfied if the zero and negative
phase networks are connected in parallel between the
points C, D and G of the positive phase system, points
C in the three systems being connected, also points D,
and the points of zero potential of all three systems.
With these connections on the calculating table, the
zero and negative phase currents for phase a as well as
the positive, may be read directly for all parts of the
system.
ANALYTICAL SOLUTION OF POSITIVE PHASE NETWORK
When a calculating table is not available, or when
greater accuracy is required than can be secured from
either the d-c. table or the a-c. network analyzer, the
distribution of positive phase currents and voltages may
be determined by calculation.
Equations may be written by Kirchoff's law expressing the internal voltages of the various machines and
the positive phase voltages, Va1 and Va1, at points C
and D respectively, in terms of the positive phase impedances of the systems and the positive phase currents.
If there are n machines, there will be a minimum of
(n + 2) unknown currents, i. e., the n machines currents
and the two fault currents; and a minimum of (n + 2)
voltage equations, i. e., the n machine internal voltage
equations and the two fault voltage equations. Each
loop in the system introduces an additional unknown
current, but also an additional equation, the voltage
drop around the loop being zero. In these (n + 2)
equations the n machines internal voltages are known,
the (n + 2) currents and the two fault voltages
areko
unknown. With equations (15) and (16), (which are
independent of the positive phase sequence impedances
and the internal voltages of the machine, but depend
upon the tes oltand the zero and negative phase
impedances) there will be (n + 4) equations and (n 4)
unknowns. It is possible, therefore, to solve for the
positive currents in terms of the internal generated
voltages and the positive, negative, and zero phase
impedances of the system. Knowing the positive
currents, the positive voltages for the system may be
obtained.
-

SOLUTION AT POSITIVE PHASE NETWORK BY

SUPERPOSITION
To avoid the additional equations and unknowns due

Positive Phase Sequence Equations* Expressing Positive

Voltages in Terms of Positive Currents and Impedances
E =II Q11+I2 Q12+. In
+I, Ial Qic+Iai Qld
E2 =Il Q21+I2 Q22 . +In Q2n+Ial
l Q2d

Q2c+I,a

En =I1 Qnl+I2 Qn2+....... +In Qnn+Ial Qnc+Ial Qnd

Val=Il Qcl+I2 Qc2....... +In Qcn+Ial Qcc+Ial Qcd
Val =Il

Qdl+I2 Qd2+....... +In Qdn+Ial Qdc+Ial

II where V1 is the internal voltage of

machine I, when no fault currents and no machine currents are flowing except I1.

Q InIn when

are flowing except

no fault nor

machine currents

I,.

Val
I 1 when no machine currents are flowing
and no fault current except

Qcd

In

when Vn is the internal voltage of

machine, or when no fault currents and no machine

currents are flowing except In.
The Q's associated with Iai and jcd in theaboveequations will be negative, due to the arbitrary assumptions
for positive direction of current flow.
When there is a point of zero potential on the positive phase system at which no voltage is generated, such
as the neutral of a shunt impedance load, there will be
a return path for the currents when one current only is
flowing, and the Q's may readily be obtained on the
calculating table. When there is no such point it will
be necessary to ground one machine, No. 2, to provide
a return path for the currents when the only current
flowing is the one under consideration. The voltage
equation E2 = 11 Qai +.1.............................
come indeterminate but may be replaced by the current
equation Ia+ 12. + In = Iai +Iain. The current distribution will first be determined with machine NO. 2 grounded (E2 = 0), and then with all

machines except NO. 2 grounded. By superposing the

two sets of currents the distribution of positive phase

to loops in the positive phase system, the following currents will be determined.

equations may be used:

Qdd

where E1, E2. En are known internal voltages on

machines 1, 2 and n, and I,, I2 and In are the corresponding currents. Val and Vaif1 are the positive phase
voltages at the points of fault C and D respectively,
and Ial and I,, are the corresponding currents flowing
into the faults. Positive direction for machine currents
is taken from the generator neutrals. Positive direction
for faults current, la, and Ialy is taken from the system
into the faults.
Also
V

*These equations were suggested by Mr. R. H. Pa.rk.

September 1931

CLARKE: SIMULTANEOUS FAULTS ON THREE-PHASE SYSTEMS

929

SOLUTION OF NEGATIVE AND ZERO PHASE NETWORKS independent equations connecting components of curWhen the positive phase voltages at the points of rent or of voltage.
Nine equations expressing relations between the
fault and the positive phase currents flowing into the
faults have been determined, four of the eight unknowns components of current or voltage at the three fault
VaO, Va0, Va2, Va2, IaO, IaO, Ia2, and 1a2 may readily be points, three equations connecting zero phase currents
obtained from the relations given in Table I. The and voltages, and three- connecting negative phase
four remaining unknowns, two of which will be in the currents and voltages, provide fifteen equations. The
zero phase and two in the negative phase system, may unknown zero and negative phase currents and voltages
be obtained from equations (5), (6), (10) and (11), or may be eliminated from these equations, the number
being reduced to three, and the three positive phase
from (7), (8), (12) and (13).
In general, the voltages at the two points of fault of voltages to ground at the fault points expressed in
either system will not be in phase with each other nor terms of the three positive phase currents flowing into
with the reference voltage of the positive phase system. the faults and the negative and zero phase impedances
The currents in the network of either system may be of the system.
In the positive phase system there will be one more
determined by superposition, i. e., by adding vectorially
the currents due to the fault voltage at C with point D unknown voltage and one more unknown current than

grounded, and the currents due to the fault voltage at with two simultaneous faults, but there will be two
additional voltage equations, one involving positive
D with C grounded.
When an a-c. calculating table is available the cur- phase currents flowing into the faults and negative and
rent and voltage distribution in the zero phase system zero phase impedances, and the other involving positive
may be obtained by applying voltages to ground, Vao phase line currents and positive phase impedances. A
and Vao, at points C and D respectively of the zero complete determination of the positive phase currents,
phase network; and in a similar manner the negative therefore, is possible, and from these, the currents and
phase current and voltage distribution will be deter- voltages over the entire system may be obtained.
mined by applying voltages Va2 and Va2 at point C
Generalizing from the above, it may be seen that n
simultaneous faults would be solved in a similar manner.
and D of the negative phase system.
SYSTEM CURRENTS AND VOLTAGES
SINGLE-POLE SWITCHING
By use of the fundamental symmetrical component
equations given in Appendix A, the positive, negative,
and zero components of currents and voltages may be advisability of switching out one phase and retaining
combined to give actual currents and voltages over the the two good phases when a ground occurs on one phase
only of a three-phase syst2m. When grounds occur at
entire system.
two separate and distinct points of the system but each
THREE SIMULTANEOUS FAULTS
involves one conductor only, it may be desirable to
The method used for determining currents and volt- investigate the possibility of simultaneous single-pole
ages when simultaneous faults occur at two separate switching.
When a conductor is disconnected from the system
and distinct points of the system may be extended to
or
to
more
simultaneous faults.
by opening the breakers at its terminals, the current
apply three
in
it and in the other two conductors and the three
With three simultaneous faults on the system, three
independent equations may be written expressing the voltages to ground at the terminals of the line are the
three zero phase voltages at the points of fault in terms same as they would be if the conductor were open at
of the three zero phase currents flowing into the faults. one point only, provided there is no ground on the
The equations given above for use in determining conductor and capacitance is negligible. For convethe positive phase currents may be used to ad- nience, when a conductor is open at both ends it will be
vantage in this connection; or if three equivalentY's considered open at one point only, the point having
for the zero phase system are drawn, each preserving any convenient location along the line. This will not
the identity of the points of zero potential considered change the currents in the conductors nor the voltages
bussed at a common point, S, and two of the three fault at the ends of the section under consideration. The
points, the three equations each expressing one zero voltage to ground of the open conductor, however,
phase fault voltage in terms of the three zero phase will depend upon the location of the opening.
currents flowing into the faults may be written by
When one conductor of a three-phase transmission
inspection. line is open the voltages to ground on the two sides of
In like manner the negative phase system will provide the opening in general will be different, and a voltage
three independent equations connecting negative phase will exist across the opening. Fig. 9 represents three
fault voltages and currents. conductors a, b, and c with conductor a open at point C.
At each of the three points of fault there will be three Let ea, eb, eC be the differences in voltage on the two sides

CLARKE: SIMULTANEOUS FAULTS ON THREE-PHASE SYSTEMS Transactions A. I. E. E.

930

of the opening for conductors a, b, and c respectively,

and ia, b and ic the line currents at point C.
(a) conductor a open.
Fault conditions: eb = e, = 0, andia = o
From (la), (10a)-(12a)
ial = - (iaO + ia2)
eal = eaO = ea2
c

ib

'.ebl

(53)

la

lveaxl

(52)

c -ec ic
FIG. 9-THREE-PHASE LINE WITH ONE CONDUCTOR OPEN
SHOWING SERIES VOLTAGES USED IN ANALYSIS

ing in parallel across the opening in the positive phase

impedance diagram; and if two conductors are opened,
by inserting the zero and negative phase series
impedance circuits in series. This is seen by adding
(58) and (59) and substituting (52) and (53) in the sum,
then transposing:
eai =

al

+Z2

zo + z2

al

(60)

and when two conductors are open, by analogy from a

single line-to-ground fault, it follows that
eai = (ZO + Z2) ial
((61)
Single-Pole Switching on Two Circuits Simultaneously.
The relations between the components of line current
and the components of series voltages at a point where
a conductor is open are independent of the rest of the
system. These relations are given in equations (52)(57) for conductors a, b, and c. By substituting a,
3 and y for a, b, and c respectively, the relations existing at a second point of opening are obtained.
In both the zero and negative phase networks there
introduced by the opening.
will be two series voltages
If the zero and negative phase networks are reduced to
their simplest form consisting of three branches each,
as shown in Figs. 1OA and 1OB, the identity of the
lines in which the conductors are open being preserved,
equations connecting components of current and voltage
may be written thus

(b) conductor b open.

Fault conditions: ea =- e, = 0, and ib = 0
From (2a) and (l0a)-(12a)
(54)
ial = - (a iaO + a2 ia2)
(55)
eal = a eaO = a2 ea2
(c) conductor c open.
By analogy from (b)
(56)
ial = - (a2 iaO + a ia2)
(57)
eai = a2 eao = a ea2
It is of interest to note that equations (52)-(57) are
(62)
- eao = iaO (Co + So) + iaO So
identical in form with the equations given in Table I,
(63)
- eao = iaO so + iaO (do + so)
C for double line-to-ground faults. In a similar manner
(64)
- ea2 = ia2 (C2 + S2) + ia2 e2
it may be shown that when two conductors are opened
(65)
- ea2 = ia2 S2 + iaO (d2 + S2)
the resulting equations between components of series
voltages and between components of line currents will
be identical in form with those given in Table I, A for as (5), (6), (10), and (11), and since equations
(52)s
(5) (6), (1) and (o1) asine equations (52)single line-to-ground faults.
Series Voltages and Series Impedances. It should be (57) have the same form as the equations of Table I,
c2 eaa
noted that ea, eao, eai and ea2 are series voltages. Due
CO eao
to the opening, a series voltage, eac, appears in the zero
phase network at point C and a line current, iaO, which
/
d2 e2 A
do e
meets a series impedance zo. Likewise the series voltand
C
network
at
in
the
ea2,
phase
negative
age, appears
S.
S21I
a line current, ia2, which meets the series impedance Z2.
in
Expressed equations:
B
A
2

(58)

FIG. 1OA-ZERO PHASE SEQUENCE SERIES EQUIVALENT CIRCUIT FOR TWO LINES WITH CONDUCTORS OPEN AT TWO POINTS
FIG. 1OB-NEGATIVE PHASE SEQUENCE SERIES EQUIVALENT

=(59)

CIRCUJIT FOR TWO LINES WITH CONDUCTORS OPEN AT Two POINTS

meacsfrtezr
andneatv phas newok repciey viwe fro
the point C.
Single-Pole Switching on One Circuit. It has been
thtteefc
shown8~~
~ ~ ntepoiiepaecret
and voltages of the system of opening a single conductor
may be represented by inserting the zero and negative
phase series impedance circuits viewed from the open-

C for double line-to-ground faults, it follows by analogy

from the work on simultaneous faults that the positive
series voltages eai and eao at points C and D respectively, when one conductor is open at each point may
be expressed in terms of the positive phase line currents
7taO and t al thus:
(66)
eai =k' al m' iQ
(67)
eao = n' zia + 1' tcgI

eao

-ia

Z
=ea2

*0
an-2aetesre
where~~
~

z2

September 1931

CLARKE: SIMULTANEOUS FAULTS ON THREE-PHASE SYSTEMS

931

where k', 1', m' and n' are functions of co, do, soy c2, d2 Vr- Vt
Zrt =
and S2, the branches of the simplified circuits represent= equivalent impedances
(69)
Ial + tal
ing the zero and negative phase networks, and are
(70)
Vr - Vt = Zc ial + eal
given in Table II, G if k', 1', m' and n' are substituted
(71)
Zd
tal

eai
for
for k, 1, m, n and co, do, soy C2, d2 and S2
CO, Do, SO,
Vry-V=
Vt=Z
al+e,71
C2, D2 and S2 respectively.
(67) in (70) and (71), respectively,
and
(66)
ngOpenngsinTwoSubstituting
02,uDivalndtS
V-(Z+k)'*72
Circuits Replacing
Equivalent CirespecuitivReply.
O>pemtngs in TwoV
Vr- Vt (Zc + k) al + m'i,l
(72)
Transmission Lines in Positive Phase System. The
(73)
+
ial
+
(Zd
=
n'
Vt
i
Vr
1')
effect of an opening in a conductor in the positive phase
system is to introduce a series voltage opposing the Solving (72) and (73) for ial and i
Zd + - m'.I
flow of line current. When there are two openings, two
series voltages will be introduced, one at each opening,
ial (Zc + kl) (Zd + 1') - ml n/ (Vr - Vt) (74)
the relations between these series voltages and the line
. k' -n
currents being given by equations (66) and (67).
- n (Vr - V) (75)
Equivalent circuits may replace the two openings just
(Zc
+ k') (Zd + 1')mn
faults,
but
as they may replace the two simultaneous
since these equivalent circuits are to be inserted in Adding (74) and (75) and substituting in (69)
series with the lines in which the conductors are open
(Z k') (Zd + 1') - m' n'
they will depend upon the end connections of the lines.
Z _
(76))
rt Zc + Zd + k' + l'- m'When the lines are bussed at one end, and m' and n'
Equation (76) gives the equivalent impedance, Zrt,
are equal, an equivalent Y having the branch
impedances k' - m', 1' - m' and m' may be inserted by which replaces the two lines bussed at both ends each
opening the lines at the common point and connecting having one conductor open. Fig. 11B.
the branch k' - m' to the line in which the opening
occurs at C, the branch 1'- m' to the line with the
zc c
opening at D and the branch m' to the common point.t
iai e.,
T
R
When m' and n' are not equal the general equivalent
circuit or the approximate equivalent circuit may be 1,4I75- Zd D 'IR=a1I1
R
ZrtaT
inserted in the same manner. These circuits may be
ec,ra
obtained from Figs. 4 and 5 if k, 1, m, n, Ial and I, are
replaced by k', 1', in', n', ial and iali respectively.
HSE
ND FIG.
IGIBPOTVE
FI.1A
SEQUENCE
11A AND
EQNC
.
~~~~~~~FIG.
11B-POSITIVE PHASE
*
end
When the lines are -not
t
bussed at either
en the
LINES
BUSSED
AT BOTH
CIRCUITS
REPLACING
Two
EQUIVALENT
positiven
curents
mayrbe
determined
analyticaly
positive currents may be determined analytically as ENDS, EACH HAVIN(l ONE CONDUCTOR OPEN
under two simultaneous faults, remembering that the
voltages at C and D are series voltages and the clurrents
SHORT-CIRCUIT AND STABILITY PROBLEMS
. .
.
are line currents. In the special case where m' and n'
the curfor
have
been
given
determining
Methods
are equal the openings in the two conductors may be
in
all
parts
of
the
system
during
replaced by a four terminal mesh network similar to rents and voltages
that shown in Fig. 17B, Appendix C. If terminals 1 simultaneous faults when there are any number of
and 2 are connected across the opening at point C and 3 connected machines. For practical problems it is not
and 4 across the opening at point D, the impedances of ordinarily required to know the exact currents and
the six branches may be determined from equations (ic) voltages in all parts of the system
For short-circuit studies when a high degree of preZa, Zb and Zab being replaced by k', 1', and m'
cision is not required, resistance and capacitance are
respectively,
When the lines are bussed at both ends a single neglected, and all generated voltages are assumed equal
impedance may replace the lines with the openings in and in phase.
In stability studies one equivalent machine is often
the positive phase system, Fig 11.
to replace several machines or groups of machines,
used
Equivalent ct'rcuit replacing in the positive phase system so that
41%.-41%
k calcuthe number of machines involved in the
twpo lines bulssed at both ends each having one conductor
open. Let the lines with the openings at C and D, hay- lain is smll Iti eesr oko h eeae
?

Equivalen

in

poitv

phs

imeacsbtentrias.

voltages on all

"ni"n

machines

or

I'ri-nr

ani

equivalent machines in

andZd,resectvel, b busedat oins FandT. etorder to calculate the power which each sends out or
the sum of the currents entering the lines at R andl reevs bu ti o eurdtoko h urnsi
power iS calculated from voltages, angular displaceleaving
thm atatTT bbe I,an
r, and letVr
th volt
nd V~t be the
le V and
leaingthe
In order to determine the
and
VOt-ments,
the
a system during transient conditions
ages at R? and T, respectively, Fig. 11A,
stability7'8 of impedances.
transfer8 impedance between the various machines
Then

Ir=ia + ia

(68) before, during, and after the disturbance must be known,

CLARKE: SIMULTANEOUS FAULTS ON THREE-PHASE SYSTEMS Transactions A. I. E. E.

932

and also the driving point8 impedance of each machine

when resistance is taken into account.
The way in which the analytical methods developed
in the preceding part of the paper may be applied to
actual problems is shown in the two examples given
below.
Problem I. Short-Circuit Study. Simultaneous double
line-to-ground faults occur on two circuits; phases
a and c are involved at point C, and phases b
and c (d and -y) at point D. Find the currents in the
six conductors in "per unit"' neglecting resistance and
assuming no-load excitations on all synchronous
machines.
Fig. 12A gives the simplified positive phase diagram of
the system. Since the currents in the faulted lines are
the only ones required, the rest of the system has been
reduced to an equivalent generator viewed from H,
with transient reactance Xd' = 0.34 and excitation
Eal = 1.00. Negative phase impedances are taken
equal to positive phase impedances, Fig. 12c. The
zero phase impedance between C and the zero potential
point, S, is 0.904 and between D and S is 1.80, and there
is no mutual between the two circuits, Fig. 12B.
Since the zero and negative phase diagrams are
already in simplest form the impedances of the branches
of the Y's may be tabulated:
Do = 1.80
CO = 0.904
SO = 0
D2 = j0.75
C2 = 1j.06
S2 = i 0.34
Substituting these branch impedances in Table II
G (b), k, 1, m and n will be calculated.

Zcds

j 1.964

j 2.55 + j 1.964

0.937 j 0.75 + 1 0.66 -

(0.0732 - j 0.0423)
j 3.098
0.022-j 0.439
0.02
ijl.06
j0.543-(-0.0732-j0.0423)
0.937
I0_
j 3.098
= - 0.022 - j 0.498

Ial =-

0.439

From equations (22) and (23)

Vai = 0.681 - (0.0221 - j 0.439) (j 1.06)
= 0.216 - j 0.023
V
681 ( 0.0221 j 0.498) (j 0.75)

a0.02
=

- 1.410 x j 1.80 - 1.627 x j 1.09

(0

019

VaO

Ea

10.34
H

10.34
0.75

=8D2=1

Co= jO904

0.75

A-POSITIVE PHASE SEQUENCE DIAGRAM FOR PROBLEM I

B-EQUIVALENT Y FOR ZERO PHASE SEQUENCE DIAGRAM
C-EQUIVALENT Y FOR NEGATIVE PHASE SEQUENCE DIAGRAM

3 SO S2 = 0

- 6.54

0.4

= 0.308 + j 0.016
From Table I, C(b) and C(a)
Vao = a2 Vai -0.128 -j 0.1755
0.088 + j 0.1985
Va2 = a Val
= 0.08 + j
= Va2
Va0 = Va2 = Vai = 0.308 +j .016

FIG. 12

j 0.34

- 1.410 x j 0.904 - 1.627 x j 1.40

+ j 2.55 x j 0.34 = - 6.54

k =

1.00

Ea,

0+.30
0.34 + j 0.727 j 1.067
Vh = Ea,-(j 0.34) X (-j 0.937) = 1. - 0.3186 = 0.681
From (31) and (32)

16

j 1.06 X j 0.34 + j 1.06 X j 0.75

+ j 0.34 X j 0.75 = - 1.410
=
0.904
X j 1.80 = - 1.627
Zoo j

Z22

Ih = Ial +Ial =

= 1

0.543

From Fig. 12B and equations

(5) and (6)

j
0.660= - 1 V0 = 0.194-1 0.142

6.540.4

a2 X j 0.34 X (-1.627)
M65=

-6.54

n=

a X

j 0.3

X54

)=

0.0732 -1j0.04231.8

0180= 009+j0.171
From Fig. 12c and equations (12) and (13)
a

0.0732 - j 0.0423 Ia2 =-(-0.088 + j 0.1985) -

1.410

1 0 34
In the positive phase diagram the equivalent circuit
+ (0.308 + 1 0.016) - 1 410 = - 0.150-j1 0.142
for two unloaded feeders radiating from the same point
H, may be used and Zhg calculated from equation (33).
1 j0195
0.34008
(j 1.06 + 1 0.543) (j 0.75 + 1 0.660) 0.007
018)- 1.410
I
.8
Zh j 1.06 + 1 0.543 + 1 0.75 + 1 0.660 + 1 0.846
1

3.S098

= j 0.727

- (0.308 + j 0.016) -

1.410

= 0.032 + 1 0.327

September 1931

CLARKE: SIMULTANEOUS FAULTS ON THREE-PHASE SYSTEMS

933

Substituting the component currents in equations (la) Substituting these values in Table II, E(b).
-(3a), the line currents are determined.
k = j 0.4716
Ia = 0.066 - j 0.723
1 = j 0.4716
Ib = 0
m = -0.1067-j 0.0683
n = 0.1067 - j 0.0683
IC = 0.515 + j 0.298
Ia = 0
Generator
1I, = - 0.729 + j 0.303
I,y = 0.701 + j 0.210
= 0.30
If the approximate equivalent circuit, Fig. 5, is used
2=0.40
AA
in the positive phase diagram, Fig. 13 would result.
s
jO.20

0.751
a

loc C

5
iO.O43
.0423
1j

11

i 0.50

j0.50
i

jO0.20

j 0.20

j 0.20

l0D,

4h4t0.3'1
iOSt
0050

0O.34

j 106

j0.20

Ea=1.00

'hi

i 0.20

:^

j0.60

~~

jO.20

0j.20

jl.40

jl.40

j.60

ID

j00O

0.20

IiIo

1a+

I=x0 6

_2=0.05

FIG. 13-APPROXIMATE POSITIVE PHASE SEQUENCE DIAGRAM

FOR PROBLEM I

Motor
B

- j 0.438,
From Fig. 13, Ih = - 0.935, Ial
s
=
0.497.
These
values
check those obtained
IY
j
s0 O0-0066
by the exact method very closely, and the line currents
calculated by using the approximate equivalent circuit
0.175
would differ very little from those given above. 0175
Problem II. Transfer Impedances (for Stability c CO o D
Study). Find the transfer impedance between the
c
generator G. and the equivalent motor, M, in Fig.14A
with simultaneous line-to-ground faults at points C and
D, conductor b being grounded at C and a at D. Find
also the transfer impedance when conductors b and a
s
are open. Reactances are given in per unit on the
2 j0.13
generator base. Transformers are A-Y, solidly grounded
P
on the high sides. The four transmission lines are on
0
0.16
double-circuit towers, which are not on the same rightof-way, and have light steel ground wires. Power is
D
delivered at two points, B and E, to a system which has
been replaced by an equivalent motor, M, and two
impedances.

0.30

'

10.10
10.40

0.5

jO.195
G

.04

0.06
E

FoG. 14

A-SIMPLIFIED SYSTEM DIAGRAM FOR PROBLEM IIA

The negative phase reactances of the system have

been indicated on Fig. 14A when they differ from the
C-EQUIVALENT Y FOR ZERO PHASE SEQUENCE DIAGRAM
positive phase values. The zero phase diagram has
D-EQUIVALENT Y FOR NEGATIVE PHASE SEQUENCE
been constructed, Fig. 14B, using the average zero
phase line and mutual reactances given in Appendix C. DIAGRAM
(a) Transfer impedance between G and M with

E EQUIVALENT POSITIVE PHASE SEQUENCE DIAGRAM WITH

fault on. The negative and zero diagrams redluced toLIEANFUTSRPCDBYQIVETY

Th tw falswt h ie pnwihte
equivalent Y's are shown in Figs. 14c and 14D, respec-

cu

tively.maberpaebyaeqiaetYFi.8,hvn

Tabulating the branch impedances of these equiva-maberpcdbyneqilntYFg.8,hvg

lent Y's:
the branch impedances
C0 =jO0.175
Do = j0.175
S0 =5j0.0066
Zrp = 2 =j 0.10
C2 =jO0.16
D2 =jO0.16
52 = jO0.13

CLARKE: SIMULTANEOUS FAULTS ON THREE-PHASE SYSTEMS Transactions A. I. E. E.

934

= j 0.40
Zt,= 2 2~~~~~~~~

lines with the two open conductors may be obtained.

Z, and Zd are the positive phase reactances of the two

lines.

Zc== Zd 1.00
-~
=jO.195
j 1.799 x j 1.799 + 0.0119
0.64 + 2 x j 1.00 x j 1.080
0.840
Zrt =
j 3.598 + j 0.217
Inserting this equivalent Y in the positive phase
diagram, Fig. 14E is obtained, from which the transfer
If the impedance Z. is inserted in the positive phase
impedance, Z12, between E1 and E2is readily determined diagram to replace the two lines with two conductors
open, Fig. 15c is obtained, from which the transfer
Z12 = j 1.052
(b) Transfer impedance berween G and M with two impedance, Z12, between E1 and E2may be calculated.
conductors open. Figs. 15A and B show the zero and
Z12 = 0.798
negative phase networks respectively reduced to their
ACKNOWLEDGMENT
simplest series impedance diagrams, the identity of
the lines with the open conductors being preserved.
The author wishes to express her appreciation of the
If
Do, SO, C2, D2 and S2 are replaced by co, do, so, assistance given by Mr. H. W. Bibber in checking the
c2, d2 and respectively, k', 1', m' and n' may be calcu- theory and mathematics of the paper and for his many
lated from Table II, G(b).
helpful suggestions in regard to the arrangement of
%,
ea,
material.
zpSpg

- .16 X j.8066 + 2 X j 1.00 X (-0.2063)

-

jj09

Co2,

jl.60

Appendix

160
E

so Do
10.30

1.90

jO.84

U2

e22

.B
e,x2

/055

D2

0 .06

S2 jl.OO

iU.2875

Fundamental Symmetrical Component Equations:2

Ia = laO + Ial + Ij2
Ib = IaO + a2 Ial + a I2
Ic = IaO + a Ial + a2 Ia2
Va = VaO + Val + Va2

0.4

FG
FIG.

A-ZERO PHASE SEQUENCE SERIES DIAGRAM FOR PROBLEM Ilb

B-NEGATIVE PHASE SEQUENCE SERIES DIAGRAM FOR
PROBLEM Ilb
C-POSITIVE PHASE SEQUENCE DIAGRAM WITH LINES
CONTAINING OPEN CONDUCTORS REPLACED BY SINGLE

(2a)

(3a)
(4a)
(5a)

VaO + a Val + a2 Va2

(6a)

IaO = 3

153

(la)

Vb = VaO + a2 Val + a Va2

VC

E2

(Ia + Ib + Ic)

(7a)

Ial

Ia2

3
1

(Ia + a

3 (Ia +

Il +

a2 Ib

a2 IC)

(8a)

+ a Ic)

(9a)

IMPEDANCE
=

j 1.60

do

3s0 S2 = 3 X1j0.90 XjO0.2875 = - 0.776

+ 3 S0 S2 = -13.71

Zed,

k' = jO.799
1' =1j 0.799
A= 0.0093A-A0.1086

nl'

=0.0093-10.1086

Substituting these values of k', l', in', n' in equation

(76) the value of the impedance, Zr, to replace the two

(lOa)

Va0=

j 1.60 SO = j 0.90
d2 = j 1.00 S2 = j 0.2875
C2 = j 1.00
Zoo = co do + co so + do so = -5.44
Z22 = C2 d2 + C2S2+ d2s2 = - 1.575
Zcd, = j 2.60 X j 2.60 + j 2.60 x j 1.1875
+ j 2.60 X i 1.1875 =-12.935

cO

Va1

1
3 (Va+aV6a2VC)

(la)

Va2 =

(Va + a2 Vb + a Vc)

(12a)

Where

a1
2
a2=-

+j 2
-

.0/10
= l.o /120

1a

(12a)

__
-/ 1.00 /240 = 1.00 /-120

(14a)

September 1931

CLARKE: SIMULTANEOUS FAULTS ON THREE-PHASE SYSTEMS

935

From (7a), (8a) and (9a):

NOTATION
a2 Ib
a Ib
Ib
Ia, Ib and I, are the three line currents at any point
Iao = 3 , Ial = 3 , Ia2 = 3
of the system.
Va, Vb and V, are the three voltages to ground at any
= a2 Ial or a Iao
.IaO
Ial
point of the system.
andIa2 = alaa ora2 Ia2 Ial
Ial Ibl and I,, are the positive phase sequence currents
in the three conductors. By defini- B. ine-to-LineFaults.
Fault on phases a and b, Fig. 16B
tion these currents are equal in magniFault conditions: I, = 0, Ib =-Ia and Va = Vb.
tude and Ial leads Ibl by 120 deg. and
From (7a): IaO = 0
I,, by 240 deg.
Ia (1 - a)
Ia2 Ib2 and Ic2 are the negative phase sequence curIal
=
From
(8a):
3
By
rents in the three conductors.
definition these currents are equal in
Ia (1 - a2)
magnitude and Ia2 leads Ih2 by 120
3
From (9a): 3a2 =
deg. and Ib2 by 240 deg.
Iao, IbO and I,o are the zero phase sequence currents in
- 1*the three conductors. By definition
2I
1a2 1 - a
these currents are equal in magnitude
and in phase.
a
Notation for components of voltage corresponds to that
b
for components of currents.
a

Appendix B

RELATIONS BETWEEN THE COMPONENTS OF CURRENT

AND OF VOLTAGE AT A SINGLE FAULT FOR VARIOUS
a
TYPES OF FAULT
b
'a
Let Ia, Ib and I, be the currents flowing into the fault
C
from lines a, b and c respectively, and Va, Vb and V
b
i>
c
the voltages to ground of phases a, b and c respectively,
C'
all at point of fault, C.
B
When a ground occurs on a three-phase system the
c
FIG. 16
conductors which becolne grounded have zero voltage
to ground and the conductors which are not involved
A--CURRENTS FLOWING INTO SINGLE LINE-TO-GROUND FAULT
have zero current flowing into the fault. When the
fault is between conductors and not to ground, voltages
B-CURRENTS FLOWING INTO LINE-TO-LINE FAULT
to ground on the faulted conductors are equal and the
C-CURRENTS FLOWING INTO DOUBLE LINE-TO- GROUND
sum of the currents flowing into the fault is zero. Thus FAULT
three equations may be written in terms of actual curFrom (4a) and (5a):
rents or voltages, which when used with the fundaVa- Vb = (1 - a2) Val (1- a) Va2 = 0
mental equations (la)-(12a) of Appendix A permit
1- a2
zero and negative phase currents and voltages to be
*a Va2 =- 1a Val = a2 Val or Val = a Va2
expressed in terms of positive phase currents and voltage, respectively.
In these fundamental equations phase a is taken as C. Double Line-to-Ground Fault.
Fault on phases a and c, Fig. 16c
reference phase, and the currents and voltages of phases
b and c are expressed in terms of the symmetrical
Fault conditions: Va = V, = 0 and lb = 0.
F~rom (2a) lb = lao + a2 Ial alIa2 = 0
components of phase a.
I
2I
Negative and zero phase voltages and currents will be*I. . Ial =--(alaO a IaI)
expressed in terms of positive phase voltages and curFrom (lOa), (lla), (12a):
rents respectively for typical fault conditions.
VbV2
a
VaO = Vb
A. ixne-to-ground Fault.
Vbl
a=
Fault on phase b, Fig. 16A3Va
Fault conditions: Ia = Ic = 0 and Vb = 0.
.-. Va0 = a2 Vai or Va1 = a Va0
and Va2 = a Va1or Va1 = a2 Va2
From (5a): Va1 = - (a Vao + a2 Va2)

CLARKE: SIMULTANEOUS FAULTS ON THREE-PHASE SYSTEMS Transactions A. I. E. E.

936

D. Three-Phase Fault.
Za Zb - Zab2
Z13 = Z24
= ilmpedance between points
(a) Conductors not grounded.
Zb
1 and 3 and between
Fault conditions: Va=Vb=Vc and Ia+Ib+Ic=O
points 2 and 4.
From (7a): laO = 0
From (lla): Val = 0
Za Zb - Zab2
Z14 = Z23 = impedance1 andbetween
From (12a): Va2 = 0
Zab
4 and
ab
(b) Conductors grounded.
between points 2
Fault conditions: Va = Vb = Vc =0
and 3.
=
0
From (lOa): Va0
(Ic)
From (Ila): Val = 0
Equations, (1c), were derived by analogy from simiFrom (12a): Va2 = 0
lar equations for the equivalent circuit of a two winding

Codcosgopoints

Appendix C
ZERO PHASE EQUIVALENT CIRCUITS FOR Two PARALLEL
TRANSMISSION LINES
It will be assumed that the conductors of all circuits
are completely transposed, so that the same impedance
is offered to zero phase currents by each of the three
conductors of any circuit.
Line A

la

-7,-12

Ib

z13 _

30_

Z32 =14

(2c)
=

Zb
Za +Zab

4Ze
2
Ab
A2

212

If Za

Line B

transformer given by Mr. George A. Campbell in

"Cisoidal Oscillations," p. 890, Proc. A. I. E. E., 1911.
(b) Two Lines bussed at both ends. If two parallel
lines are bussed at both ends the equivalent circuit of
Fig. 17B reduces to a single impedance branch Ze where
1
Za Zb - Zab2
1
Za Zb 2 Zab
1
1
Ze = 1
+
+
+
Z14
Z23
Z12
Z34

aa

14

lbB

lb

Neglecting Resistance

ie

~ ~ 212 b

C Z24 JZ 13

(3c)

2-,UA1,3

Z=

b-Zab

C 4

FIG.C17
FIG. 17
A-PARALLEL TRANSMISSION LINES HAVING IMPEDANCE Za
Zb AND MUTUAL Zab
B-EQUIVALENT CIRCUIT TO REPLACE THE Two PARALLEL
LINES BETWEEN TERMINALS 1, 2, 3 AND 4
CEU TCIRCUIT To REPLACE Two PARALLEL LINES
BUSSED AT ONE END BUT NOT AT THE OTHER

AND

1. Two Parallel Transmission Lines.

(a) General Case. Any two parallel transmission
lines or portions of parallel transmission lines having
impedances Za and Zb and mutual reactance, Zab,
between them may be replaced at their terminals by
an equivalent circuit consisting of a six-branch network
directly connecting the four terminals.1 If 1 and 2
are the terminals of line A, and 3 and 4 the terminals

Xa + Xab

(c

(c) Two Lines bussed at one end but not at the other.
that points 1 and 3 are together
If the lines are bussed so
points 1, 2 and 4,
Fig. 17B becomes a A connecting
which may
be converted
into awhere
Y, Fig. 17c having
branch
impedances
Z1, and

Z2

Z4

Zab
Za- Zab
(5C)
Z4 = Zb Zab(
This case may be extended to include parallel lines
Z1
Z2

which are bussed through unequal impedances. On the

zero phase diagram of an actual system which includes

two parallel lines it is usually possible to find a point

where the parallel lines are bussed at one end or the
other through impedances. Such a point may be the
ground or any branch point of the system, or it may be
obtained by A-Y or Y-A transformations.
If the impedance from the branch point 5 to the terminals 1 and 3 of the lines A and B are Z,1 and ZB respectively, an equivalent Y may be used between points 5,
of line B, see Fig. 17, then the six impedances of the 2 and 4, having the branch impedances Z5, Z2 and Z4,
where
equivalent circuit are:

212 =

= impedance between pit1

b

and
2.

Z34 = ZaZb-Zb = impedance between points 3

Za

and 4.

22= Za + Za -Z

= Zb + Z4l
Z~~~~~~~4

Zab

These above relations follow directly from (17c)

if Za and Zb are the impedances from the junction point

September 1931

CLARKE: SIMULTANEOUS FAULTS ON THREE-PHASE SYSTEMS

937

and therefore include the external impedances Za and by a2 and adding: adding equations (6) and (11); substituting from Table I, A (b) and A (a):
Zo.
Val = (CO + S0 + C2 + S2) Ial + (a SO + a2 S2) lal
MUTUAL REACTANCE BETWEEN Two PARALLEL
(3d)
TRANSMISSION LINES
Val = (a2 So + a S2) Ial + (Do + SO + D2 + S2) Ilc
From (4c)
Xab = 2X,- Xa
(6c)
(4d)
in
Equation (6c) gives the mutual reactance Xab terms
(e) Phases c and a.
Of Xa, the reactance of one line alone, and Xe the
Multiplying equation (5) by a2 and equation (10)
reactance of the two lines in parallel. Xab may be by a and'adding; adding equations (6) and (11); subexpressed in terms of X1, the positive phase reactance stituting from Table I, A (c) and A (a):
of one line, by substituting for Xa and Xe their values Vatit (Co + 50 + C2 S2) Ial + (a2 S0 + a 52) lcd
aSd)
in terms of X1.
For lines without ground wires or with light steel
Val = (a So + a2 S2) lal + (Do + So + D2 + S2) laO
ground wires.7
(6d)
If Xa = 3.5 Xi and Xe 2.5 Xi then
Case
F.
Faults
at
Two
Points.
Line-to-Line
X
v1.5
Xab
(a) Corresponding phases b, c and 1, -y.
For heavy copper ground wires.
Replacing negative phase currents and voltages in
If Xa = 2 Xi and
then
Xe = 1.5
Xi then
IX=21n
1.5
equations
(10) and (11) by their values in terms of
Xi
Xab= = Xl
Xab
positive phase currents and voltages from Table I,
These approximate values for mutual reactance may B(a):
be used in determining zero phase equivalent circuits
(7d)
Val = (C2 + S2) Ial + S2 Ial
Vai = S2laO (D2 + S2) laO
for two parallel transmission lines when extreme
(8d)
accuracy is not required.
Phases a, c and -y.
(b)
Appendix D
Substituting from Table 1,I, B (b) and B (a) in equaRELATIONS BETWEEN POSITIVE PHASE COMPONENTS tion (10) multiplied by a2, and in equation (11):
OF VOLTAGE AT THE FAULTS AND POSITIVE PHASE
(9d)
Val = (C2 + S2) Ial + a2 S2 IaI
CURRENTs FLOWING INTO THE FAULTS FOR
Vai = a S2 lal + (D2 + S2) Ial
(lOd)
Two SIMULTANEous FAULTS

(c) Phases a, b and 1, -y.

Table I expresses negative and zero phase voltages
Substituting
for Table I, B (c) and B (a) in equation
into
the
fault
of
currents
the
and
at
point fault,
flowing
by a, and in equation (11):
multiplied
in terms of the positive phase voltage at the fault and (10)
the positive phase current into the fault when grounds
Val = (C2 + S2) Ial + a S2 li
(lld)
Vai = a2 52 Ial + (D2 + S2) Ial
(12d)
occur on one or more of the three conductors of a threeakin phase
phae aa ass reerene
reference phse.Case
system,sysem,
phase.
taking
phase phas
G. Double Line-to-Ground Faults at Two Points.
Similar relations for a second point of fault are obtained
(a) Corresponding phases b, c and 13, -y.
by substituting a, 1, y for a, b, and c in Table I. These
Adding equations (7) and (12), and (8) and (13),
two sets of relations from Table I substituted in combinations of equations (5)-(14) allow Val and Val to and substituting from Table I, C (a).
be expressed in terms of Ial and Iai.
So S2
Do + So D2 + S2
Iai ValIZO + Z22 ZVaO Zoo + Z22 )
Case E. Single Line-to-Ground Faults at Two Points.
(a) corresponding phases a and a.
(13d)
and
and
and
and
Adding equations (5)
(10)
(6)
(11)
sO S2
CO + So C2 + S2
vao( Coo + C2S)
substituting from Table I, A (a):
Ia1 =-Val(
22
Val = (CO + SO + C2 + S2) Ial + (SO + S2) Ial (1d)
(14d)
Vai = (So + S2) Ial + (Do + S + D2 + S2) IaT (2d)
and
in
terms
of
and
(b) Phasesb
anda.
Vai
Va0
laO
Iai by
solving equations (13d) and (14d) and making use of
Multiplying equations (5) by a and equation (10) equations (9) and (14):

~~~Expressing

(Co

[Z22 (CO + S0) ZOO (C2 + S2)] Iai + [Z22 So + ZOO S2] Iai
+ C2) (Do D2) + (C0 + C2) (So + S2) + (Do D2) (So S2)

(15d)

Val (C0n

Zoo
S2] laO [Z22 (Do So) Zoo (D2 + S2)] I,ai(6d
~[Z22
+ C2)So(D0
+ D2)
+ (Cn + C2) (Sn + So) + (Dn + D2) (Sn + S2

(6d

Vai
V -

CLARKE: SIMULTANEOUS FAULTS ON THREE-PHASE SYSTEMS Transaetions A. I. E. E.

938

(b) Phases a, c and 3, y.

Multiplying (7) by a and (12) by a2 and adding:
adding (8) and (13): substituting from Table I, C (b)
and C(a):
a SO a2 S2\
{Do +So D2 +S2i
A
Ial =Vai(
Z22

Transposing and simplifying equations (8) and (13)

to express Vao in terms of Vao and I,,o and Va2 in terms
of Va2 and Ia2: then replacing Vao and Va2 by V
from Table I, C (a) and Iao and I,2 by Ial from A (a):

V.,(ZOO SZ22

D+SDZoo

Vao = Va

(17d)
'al Vai(a2 So aS2)

00O

Z22

so

Ial

ZOO

co + so

2___
Va2 = Val C + S Ia, C2 + S

V co AS +o C2 +S2)

ZaO

co

Z22

(28d)

(29d)

(18d)
Solving (17d) and (18d) for Val and Va,, remembering Substituting (28d) and (29d) in (27d), and replacing
that a X a2 = 1 and a + a2 = - 1:
Z0o and Z22 by their values given in (9) and (14):

Vai(C0 [Z22 (C0 SO) ZOO (C2 S2) Ial + (a SO Z22 + a2 S2 ZOO) Ial
(Co + C2) (Do + D2) + (CO + C2) (SO + S2) + (Do + D2) (SO + S2) + 3 SO S2
(a2 SO Z22 + a S2 Z0O) Ial + [Z22 (Do + So) + Zoo (D2 + S2)] Ial
+

V.,

(CO + C2) (Do + D2) + (CO + C2)

(c) Phases a, b, and O3, y.

This case may be solved by analogy from (b), a
replacing a2, and a2 replacing a in equations (19d)
and (20d).
Case H. Three-Phase Faults at Two Points.
From Table I, D.
Val

Va, =O

(21d)
Vai = 0 (21d)

(SO + S2) + (Do + D2) (SO + S2)

= Vai (
cas

+ 3 SO S2

1
S
+ S2
+C-

a
+ al

(20d)

S0 SA2
co + SO + C2 + S2

(30d)

~Transposing (30d)

(22d) V _ (C0 + S0) (C2 + S2)

C0A-50A-C2A-522 al
Case I. Line-to-Line Fault at C and Single Line-toGround Fault at D.
SO (C2 + S2) + S2 (Co + SO)
(a) Phases b, c and a.
(C A CA S( IA , (31d)
C A S A-2 A 2
Substituting from Table I, B (a) and A (a) in equaequations
(28d) and (29d), and replacing
Adding
and
in
tions (10)
(6) plus (11)
Vai from Table I, A (a) and Val
Va2)from
by -(31d)
Va=l C (Sd ) by its+value
Val -Ial (C2 + S2)I-(DlA2

SO)

A S)- S 23d) (Va0

(b) Phases a, c and ae.

Substituting from Table I, B (b) and A (a) in equation (10) multiplied by a2, and in equation (6) plus (11):
(25d)
Val = Ial (C2 + S2)-Ia. a2 S2
=
So
Va. -a S2 Ial + Ia1 (Do + + D2 + S2) (26d)
(c) Phases a, b, and a.
By analogy from (b) replacing a by a2, and a2 by a.
Case J. Double Line-to-Grou,nd Fault at C, Single Lineto-Ground Faullt at D.
(a) Phases b, c and a.
Adding equations (7) and (12) and substituting from
fTable J,C (a):
D2-2
,D0-5
IaiVi~j ZOO A Z22)VaOZ
Z22
(27d)

S!

Vai = _ SO (C2 + S2) + S2 (C0 + SO)

C0A-50A-C2A-52
+

Do + D2 + SO + S2-

+ S )I

(32d)
(b) Phases a, c and a.
Multiplying equation (7) by a, equation (12) by a2;
adding and substituting from Table I, C (b):
/D0 A-SO
D2A-5S2 '\
A- z22 J
laO = Vai (\ z

aC

S0

S2

(33d)
- a2 Va2
- Va0
From equations (8) and (13) and Table I, C (b) and
A (a)

September 1931

CLARKE: SIMULTANEOUS FAULTS ON THREE-PHASE SYSTEMS

-0
Vao = a2 Val C so

~0 ~Zoo
co

Adding (6) and (11), replacing Iao and Ia2 by their

(34d) values given in (44d) and (45d), and substituting from
(34d)

Table I, A (a) and simplifying:

Z2
S2
(3)
+ S- I,C + S, (Dd
Substituting (34d) and (35d) in (33d) and transposing:
(CO + SO) (C2 + S2)
v al- CO + So + C2 + S2 a

V.2

939

Val

= =a
a
C

(36d)

a 50 (C2 + S2) + a2 52 (CO + SO)

C0 + S0 + C2 + S2

Do + Cc + SOlS+ D2 + C)
+ S Iai (46d)

V
=Vi

Bibliography
1. 1927 Lightning Experience on the 132-Kv. Transmission

Lines of the American Gas and Electric Company, by Philip Sporn,

A. I. E. E. TRANS., 1929.
The Problem of Service Security in Large Transmission Systems,

Adding (34d) and (35d), replacing (Va0 + Va2) by by P. Ackerman, A. I. E. E. TRANS., 1930.
- Vai, from Table
L, A (a) and Val by its value from
2. Method of Symmetrical Coordinates Applied to the Solution
(36d):
of Polyphase Networks, by C. L. Fortescue, A. I. E. E. TRANS.,
1918.
a2 SO (C2 + S2) + a S2 (CO + SO)

VaC = -

Ial

CO SO C2 S2

+ (D -I- D2 SO + S2

SO S2 + S22
CO + So + C2 + S2 J I,
S

(c) Phases a, b and

(37d)
+

By analogy from (36d) and (37d) replacing a by a2,

and a2 by a.
Case K. Three-Phase Fault at C and Single Line-toGround Fault at D, .
(a) Three-phase fault not involving ground.
Substituting IaO = 0 from Table I, D (a) in equations
(5) and (6), and Va2 = 0 in equation (10):
- VaO = IaO So
(38d)
- Vao = Io (Do + So)
(39d)
S2
=- a2 C2 + S2

Ia2

(40d)

Substituting (40d) in (11) and simplifying

- Va2

Ia2 D2 +

C2 S

(41d)

C2 + S2

3. Studies of Transmission Stability, by R. D. Evans and

C. F. Wagner, A. I. E. E. TRANS., 1926.
4. "Calculation of Single-Phase Short Circuits by the
Method of Symmetrical Components," by A. P. Mackerras,
General Electric Review, 1926.
5. '"Application of Hyperbolic Functions to Engineering
Problems," App. E by A. E. Kennelly, McGraw Hill Book Co.
6. The M. I. T. Network Analyzer, by H. I. Hazen, 0. R.
1

Schurig, and M. F. Gardner, A. I. E. E. TRANS., 1930.

7. System Stability as a Design Problem, by R. H. Park and
E. H. Bancker, A. I. E. E. TRANS., 1929.

Progress in the Study of System Stability, by I. H. Summers

and8.J. B. McClure, A. I. E. E. TRANS., 1930.
9. The Reactances of Synchronous Machines, by R. H. Park

and B. L. Robertson, A. I. E. E. 1928.

10. "Cisoidal Oscillations," by Geo. A. Campbell, Proe.
A. I E. E., 1911.

D c s

Discussion

M. MacFerran: Miss Clarke's paper is of great practical

value because, as is pointed out in the introduction, simultaneous
faults do occur quite frequently especially when multicircuit
tower lines are in use. For example, on the system of the South-

ern
California
Edison
Co., circuits,
Ltd., there
are many
towerseveral
lines
from two
to twelve
and there
have been
carrying
cases in which trouble on such multicircuit structures has in-

volved two lines going to entirely different parts of the system.

adding .(39d) and (41d) and substituting from Table I, Another very interesting case of simultaneous faulting occurred

A (a)

at a substation which was

Vai

I=I

( Do +

So

D2

C2 S2

C2 + S2

alized,

(42d)

From Table I, D.

Vai = 0
(43d)
(b)0 Tle-pa
s(b)
e-ln
Three-phasefault involving ground.
Substituting Vao = 0, and Va2 = 0 from Table I,

D (b) in (5) and (10) and expressing laO and 1a2 in

terms of Iao and 1a2.

one

section being

being operated with buses section-

on one major station,

on one bus

the other on

another. Trouble originating

spread to the other,
thus producing a condition quite similar to that covered by the
numerical example in the paper.
From the standpoint of one who is faced with the task of
calculating simultaneous faults, the outstanding features of

Miss m,
Clarke's
first, the combinations
derivation of ofthefaults;
values of
and n paper
for all are
conceivable
~~~~~~k,
I,
aJnd
second, the development of the approximate equivalent fault

wye which can be used with a high degree of accuracy even

when m is not equal to n. These two features ma.ke it possible

~~~~~~towork out the complete story of any type of simultaUneous
IaO =-o(44d) ...fault in a quite reasonable length of time, and therefore enable
C0 + So
one to give due attention to such faults in setting relays or
analyzing their performa.nce. This in turn represents an
S2
important step forward in protection engineering.
...H.W. Bibber: While Miss Clarke's paper deals very exhausIa2 - -Ia2 C2 + 2 C+S(45d) tively
with problems involving simultaneous faults as they are
5