Overvoltages During Power-System

grounded through reactance, so that the

resultant zero-sequence impedance viewed
from the fault is inductive rather than ca-
pacitive, the fundamental-frequency volt-
Faults ages-to-ground on the unfaulted phases at
the fault will not be greater than 1.73 times
normalline-to-neutral voltage. '

EDITH CLARKE S. B. CRARY H. A. PETERSON 3. If a system is grounded through resist-

ance, so that the resultant zero-sequence
MEMBER AlEE MEMBER AlEE ASSOCIATE AlEE impedance viewed from the fault is induc-
tive rather than capacitive, the fundamen-

O
tal-frequency voltages-to-ground of the un-
VERVOLT AGES may be produced scattered and incomplete form. It is the faulted phases at the fault will always be
by lightning, switching surges, purpose of this paper to review and less than twice normal line-to-neutral volt-
faults, both solid and arcing, and the analyze the factors which determine the age.
overspeeding of machines due to loss of maximum voltages which may be ob- 4. In order that the overvoltages during
load. The effect of circuit and machine tained on a system following the occur- faults shall not exceed that which is con-
characteristics on the duration and rence of faults, so that a more rational sidered safe for the operation of grounded-
magnitude of these overvoltages has re- neutral lightning arresters, it is advisable
selection of system protective equipment that careful attention be paid to the method
ceived considerable attention.vr!" Fur- can be made and the best method of sys- of grounding and the number of grounding
thermore, field tests have been made in tem grounding determined. points in the system. Consideration should
order to determine the effectiveness of Results of calculations, and of tests on also be given to the possibility of temporary
different methods of grounding, and the disconnection of these grounding points and
a miniature system are presented. In-
the resulting overvoltages which may occur.
accuracy of the methods used to calculate cluded in these results are the transient
the magnitude of overvoltage during 5. Isolated-neutral systems may be sub-
voltages as well as the" fundamental-
jected to particularly high overvoltages if
system faults. U - 13 frequency voltages that occur on a system the system is large in extent. The over-
It seems to be the proper time to con- during a fault. The information on vol tages at or near the region of resonance
sider and review the overvoltages pro- transient overvoltages is of value in are appreciably reduced by fault and line
duced by the occurrence of system faults, indicating the maximum magnitude of resistance. This may be an important fac-
tor, particularly in low-voltage systems.
in order that full advantage of present-day voltage that may be obtained on a sys- On the other hand, resistance in the fault or
knowledge may be taken by those con- tem during a solid fault, thus showing ground return for a grounded-neutral sys-
cerned with system design and operation. the magnitude of transient voltage which tem may slightly increase the overvoltages
There are essentially three components is inherent in the circuit and which cannot obtained on one of the open phases.
of voltage due to the occurrence of a sys- be reduced by the elimination of excess 6. The transient overvoltages obtainable
tem fault: voltages due to switching or arcing, or on a system which ordinarily might be con-
sidered to be solidly grounded may ap-
1. Fundamental-frequency voltages. by building lightning-proof lines. Fur-
proach 2.73 times normal. Sustained volt-
thermore, as it seems reasonable to as- ages are generally higher for systems
2. Natural-frequency voltages usually of
short duration which are superimposed upon sume that arcing at the fault may sub- grounded through resistance than for sys-
the fundamental-frequency voltages. ject the circuit during the arcing period to tems grounded through corresponding val-
voltages of the same order of magnitude ues of reactance. The natural-frequency
3. Harmonic voltages resulting from un- transient voltages obtained on a system
balanced currents flowing in rotating ma- as those obtained immediately subsequent grounded through resistance are practically
chines in which the reactances in the direct to the initial application of the fault, the negligible except when the resistance is high
and quadrature axes are unequal. curves presented can be used as a guide relative to the positive-sequence reactance.
In general, protective devices, par- to indicate the maximum voltage which 7. The transient overvoltages during faults
ticularly lightning arresters, must be ca- may be expected due to arcing across an are not expected to be of sufficient magni-
pable of withstanding the transient over- insulator. tude to cause breakdown of the major in-
The paper analyzes the effect of those sulation provided it is in good condition,
voltages to which they may be subjected except for the case of an isolated-neutral
for periods of short duration, and then factors which determine the magnitude
system having an appreciable amount of
must be able to seal off the power follow of the overvoltages during faults, which line or cable capacity to ground. The
current associated with the sustained are the most common causes of high sus- overvoltages during faults in a grounded-
tained voltages. A study of these volt- neutral system are not as great in magnitude
voltages. Unnecessary lightning-arrester as those which may be expected from
failures due to excessive sustained over- ages is usually a necessary first step.
lightning or switching surges.
voltages during fault may be prevented after which refinements and the more
unusual cases can be considered. The 8. Ground-fault-neutralizer (Petersen coil)
by properly grounding and operating the systems are subjected to transient and
system. effect of different methods of grounding
fundamental-frequency overvoltages which
It is the fundamental-frequency over- and the influence of fault and arc resist- are, in general, higher than those of a solidly
voltages which largely determine the ance are included. grounded system but lower than the volt-
lightning-arrester rating and the corre-
sponding insulation protective levels. Conclusions Paper number 39-52, recommended by the AlEE
committees on power transmission and distribution.
These in turn are a factor in determining and protective devices, and presented at the AlEE
winter convention, New York, N. Y., January 23-
the allowable apparatus insulation levels. From the results discussed in this paper, 27, 1939. Manuscript submitted November 12,
From a technical standpoint, the funda- the following conclusions can be drawn: 1938; made available for preprinting December 28,
1938.
mental-frequency voltages can be deter- 1. Line-to-ground faults for most systems EDITH CLARKE, S. B. CRARY, and H. A. PETERSON
mined with fair accuracy. can be used as a basis for determining the are all in the engineering division of the central-
Although this phase of the subject is of maximum fundamental-frequency and tran- station department of the General Electric Com-
pany, Schenectady, N. Y.
considerable importance, it has been dis- sient voltages during faults.
1. For all numbered references, see list at end of
cussed in the literature only in somewhat 2. If a system is solidly grounded or paper.

AUGUST 1939, VOL. 58 Clarke, Crary, Peterson-r-Ooerooltages During Faults TRANSACTIONS 377
G~ 6..J
I ,C(
Q:
....
c(
a:
to-
II :J =>
5~ 5~
RO/X\"O
I ...... ~- ~'UJ ~
Jr' 4~~
,
4~~
Ii 1\
-c(
-JIo- ..J ....

v
.,J..J ..,J:.J

~~ <0
l:>
3(X
.3~
~
~ \ If)
z J 0
0
z
-_\\ '91."-
uJ If)
b= ~ R o/ x '
5 ~ UJ
r--- 2.- 2!
-
I1"i"
00 00
t-
~
>- L....-~~~
" ........ 1 I-- 00
-
1\ I---'t-"
00
N ~~
/6
....... ~ ~

1/
-00
'-
-10 -8 -6 -4 -2
XO/XI
0 +2 +4 +6 +8
-
+10 +00
'---
-8 -6 -4 -'2 °
Xo/x,
+'2 +4 +G
-
+10 +00

Figure 1. Fundamental-frequency voltages. the transient overvoltage obtained immedi- Figure 2. Fundamental-frequency voltages
Line-to-ground fault. Maximum voltage-to- ately following the occurrence of a solid Double-line-to-ground fault. Voltage-to-
ground of unfaulted phases at the fault fault. Under this assumption the transient ground of unfaulted phase at the fault
voltages presented in the paper are an
Z, = Z2 = 0 + jX indication of the maximum voltage which Z, = Z2 = 0 + jX 1
Zo = Ro + jX o
1
may b~ obtained due to arcing at the fault."
Z« = R« + jX o
R, = 0 2. Only fundamental and natural-fre- R1 = 0
quency components of voltage are consid-
ered. Harmonics produced by the saliency
effect of rotating apparatus are neglected. machines not equipped with amortisseur
ages which can occur on an isolated-neutral This assumes in effect that Xq" / Xd" = l. windings, is to make the negative-sequence
system. Exceptionally high voltages may occur due impedance less than the positive and to
to saliency effects, particularly for the case increase the losses with fault duration.
of water-wheel generators not equipped with These effects, in general, tend to reduce the
Basis of Study amortisseur windings and disconnected magnitude of overvoltage calculated with
from their load. These overvoltages tend equal positive- and negative-sequence im-
The calculations, and tests on a minia- to be increased if the generator is left con- pedances.
ture system, made to determine the nected with an unbalanced fault to an ap-
preciable amount of transmission line or 5. The magnitude of the positive-sequence
magnitude of overvoltage following the impedance is assumed to be constant for the
cable so as to form a capacitive load. The
occurrence of system faults, were based conditions for this important exception are period in which the overvoltage is being
on the following assumptions and con- usually evident and are discussed fully in determined. As the effective impedance of
siderations : two recent papers.v? a rotating machine changes with time, it is
It is important to realize, however, that a necessary to consider the impedance under
1. The fault is a solid fault; that is, there certain amount of distortion due to this conditions which may result in the greatest
is no arcing. Arcing tends to subject the effect may occur for many system faults and overvoltage. On an isolated system, for
circuit to a more or less continuous transient that this will tend to increase the sustained example, the greatest overvoltage may be
condition. I t has been shown that very voltages above those based only upon a obtained based on impedances correspond-
high voltage may be obtained in a circuit consideration of the fundamental-frequency ing to the synchronous condition rather than
with no losses under assumed conditions of components. the subtransient or transient condition.
arc interruption and restriking.t-l! Evi- Because these impedances vary with time
dence indicates that these mechanisms of 3. Sudden disconnection of load and over- after the fault occurs, it is desirable that re-
arcing do not ordinarily occur at the point of speeding of the system and connected gen- sults be presented in the form of curves so
fault, 12, 13 although such phenomena may be erators is not considered. Loss of load on a that the range of voltage can be easily
associated with switch operation.P It is generator may result in overvoltages due determined.
believed that the maximum voltage which both to the sudden disconnection of the load
and also to the overspeeding. 4 The results 6. The effects of saturation and corona are
may be expected due to arcing can be taken
presented in this paper can, of course, be not included. Both would tend to reduce
to be approximately that corresponding to
modified in order to indicate the increase in
voltage which might be obtained under
these conditions. This effect is most pro- Figure 4. Fundamental-frequency voltages.
Figure 3. Fundamental-frequency voltages. nounced in portions of the system which Line-to-ground fault on phase a. Phase b
Line-to-ground fault. Zero-sequence volt- overspeed due to loss of load on nearby voltage-to-ground at the fault
age at the fault hydro stations.
4. Negative-sequence impedances are as- = Z2 = R1 + jX 1
Z,
Z, = Z2 = 0 + jX 1
sumed equal to positive-sequence imped- Zo = Ro + jX o
Z« = Ro + jX o ances. The effect of rotating loads and R, = 0
R, = 0 machines, except in the case of synchronous R1/X1 = 0.4

I \0
4.
0 4w~
'7w ~C(
UJt!)
V Z4: Ro/x'A j~
3:J~ c(>
Ro/XI:OJ ..Jo
V 3~~.
V/ \
4:>
~ t--
~-'
ere(
b:;::::: ~ r--. ~51\
Zt-
\f}=»
~
2~~
....... ~
~ \\ If)::>
LULU r-- r-...2 K" \
2~~
t-:-F-
...10
~

~
L . - ~ I---
- r2- ~~
00
It-
~z

5
00
- ~~ 1 ~
,.......
~
r--
~ r- ~
~
~-
\[76
~

I..-
-00 -10 -8 -+00 'V -
-4 -2 0 +'2 +4 +8 +\0 -00 -10 -8 -6 -4 +4 +6 +8 +10 +00
Xo/X,

378 TRANSACTIONS Clarke, Crary, Peterson-r-Ouervoltages During Faults ELECTRICAL ENGINEERING

the magnitude of overvoltage. The error in +
o jX 1 and Z« = R o + jXo are the posi- of resonance with a small ratio of R o/ X 1
neglecting these factors, in general, in- tive- and zero-sequence impedances; re- is expected to be unusual, and undoubt-
creases with increase in the magnitude of
overvoltage. In the region of resonance, spectively, viewed from the fault. X« edly will occur only for isolated-neutral
saturation is an important factor in limiting may be positive or negative. See equa- systems or grounded systems following the
the excessive fundamental-frequency over- tions 10 and 11, appendix A. loss of the system ground point.
voltages. Figure 2, similar to figure 1, gives the Figures 4 and 5, similar to figure 1,
7. Overvoltages caused by unusual system maximum fundamental-frequency voltage show the overvoltages which may be ob-
conditions are neglected. For example, on the unfaulted phase at the point of tained on phases band c, respectively, at
nonlinear circuit instability may be pro-
fault following a double-line-to-ground the point of fault for the case of a single-
duced as a result of the opening of fuses or
single-pole disconnect switches, thereby fault. Fault resistance and positive- conductor-to-ground fault on phase a.
causing the magnetizing currents of poten- and negative-sequence resistances are These two curves apply for RI/X1 =
tial or power transformers to flow through neglected. See equation 13, appendix A. 0.4. Other curves similar to these have
line capacitance with resulting overvoltages. Figure 3 gives the zero-sequence volt- been prepared for -early publication by
8. The miniature system used for deter- age at the point of fault for a line-to- E. M. Hunter with different ratios of
mining transient overvoltages was set up to ground fault under the conditions of positive-sequence resistance to reactance.
represent a low-loss system; that is, the
ratio of rl/XI was approximately 0.03. For
figure 1. The zero-sequence voltage at This particular case is selected as being
systems having higher losses, the transient the point of fault for a double-line-to- representative. As will be noted from
overvoltages will be somewhat less because ground fault under the conditions of these curves, the voltages on phase c
of greater decrement factors. figure 2, is one-third the voltage on the are, in general, higher than those on phase
9. Tests on the miniature system for deter- unfaulted phase given by figure 2. See b, although this is not true for all ratios
mination of maximum transient voltage equations 9 and 12, appendix A. of RolXl and XO/X I •
were made by representing the system by For a system which has its neutrals
lumped impedance elements. This will, in
Figure 6, for the line-to-ground fault,
general, lead to higher transient overvolt- solidly grounded or grounded through shows the effect of fault resistance. Re-
ages at the point of fault than will be ob- reactance, the ratio XO/X 1 will ordinarily sistance except in the fault is neglected.
tained in an actual system. This pessi- be positive. That is, the zero-sequence As Rf / X I is increased from zero to RII X 1
mistic result is due to the fact that the num- impedance viewed from the fault is induc- = 1, approximately, the voltage of one of
ber of circuit natural frequencies is less than
that on an actual system, and the resultant
tive rather than capacitive in effect. the unfaulted phases is increased and the
transient voltages are therefore likely to be With neutrals grounded through resist- other decreased, except in the region of
of greater magnitude. ance, XO/X I may be either positive or resonance where both are decreased.
negative. For an isolated-neutral system The curve for RI/X1 = 1 is given in
Discussion of Results XO/X I is negative. Also in a system figure 6. As R IIX 1 is further increased,
which is extensive in number of miles of the voltages decrease. Any increase in.
FUNDAMENTAL-FREQUENCY VOLTAGES connected line or cable compared with fault resistance decreases the overvoltage
the total admittance of the grounded obtained in the region of resonance.
Based on the foregoing assumptions, points, the ratio of XO/X I may be nega- Ro includes the effect of resistance in
figure 1 gives the maximum fundamental- tive. With XO/X 1 negative, the voltages the ground return and in the neutral,
frequency voltage to ground which will obtained correspond to those on the left- while Rlof figure 6 is the resistance in the
occur on either of the unfaulted phases at hand side of the vertical axis. Under this arc or fault. Figures 1, 2, 4, and 5 can
the point of fault following a line-to- condition, it is possible to obtain high be-used to determine the magnitude of the
ground fault. Positive- and negative- overvoltages, particularly for small values fundamental-frequency fault voltages of
sequence resistances are neglected, but of ~/Xl and values of X 0/X, in the region systems with isolated neutral, neutral
zero-sequence resistance is included. of resonance. solidly grounded or grounded through
Fault resistance is neglected. Figure 1 With all resistance neglected, infinite resistance, reactance, or impedance. Z.
has an abscissa of XO/X I , where Zl = fault currents and voltages will occur if
XO/X 1 = -2 for a line-to-ground fault,
and - 0.5 for a double-line-to-ground Figure 6. Fundamental-frequency voltages.
Figure 5. Fundamental-frequency voltages. Line-to-ground fault. Maximum voltage-
fault. See appendix A, equations 9-13.
to-ground of unfaulted phases It fault
Line-to-ground fault on phase a. Phase c Values of X 0/Xl in the neighborhood of
voltage-to-ground at the fault
-2 or -0.5 are considered to be in the Z, = Z2 = 0 + jX 1
Z, = Z2 = R1 + jX 1 resonance region. Operation in the region Z« = 0 + jX o
Zo = Ro + jX o
RI = 0
Rl/X 1 = 0.4
7
l/ ,
I
-{
3;;i,
(X

~
UJ
/ g
Z

Ro/X,=
/
A
\
5.
~UJ
uJ(!)
z<{
--
Rf"/XI&~ ~ ~
::: :::- '"'()
f':
\
1 w
~~
2..J ...
UJ

....J:.J

--
4:i~ \ <0
\
--
~> i-
/ 0.5", :.JO ~
"'"- ~ ~ a
<{> 5
V t:::
0 f--
e- r;::::
"'\
Rf/~;!--~ '--

-- ~
1:,J Z
/
......... t"--
33<l .........
l....e e: \\
I---
~ z~ r---.... ~ ~~
~
~ l.--
~

--
~

K
If)::::> ~
5
~~~
UJUJ
1:Z 00
~
~ ~~ ~ I---- 00
~ c:.
21- \V
I-- 10 ~
I--

~ ~
......"".

fJ.-1o-"'"

-00 -10 -8 -4 -2 0 +2 +4
:-
+10 +00
~

000 -12. -10 -8 -to -4 -2 0 +2 +4 +6 +8 +10 -

+12+00
xo/x, Xo/x,

AUGUST 1939, VOL. 58 Clarke, Crary, Peterson-i-Ooeruoltages During Faults TRANSACTIONS 379
increased, voltages increase more rapidly impedance diagrams for part a. Lower-
~~
F
X for the line-to-ground fault than for the case letters are used to differentiate the
double-line-to-ground fault. With RolXl indicated impedances in figures 7 from

EG:
(a) = 1, higher values for the line- to-ground the resultant effective impedances viewed
fault than for the double-line-to-ground from the fault which are represented by
fault are obtained. From equation 13, capitals.

F
X
f.:CI appendix A, any increase in R I reduces
the voltage on the unfaulted phase for
the double-line-to-ground fault; also,
In figure 7 the positive- and zero-se-
quence fundamental-frequency imped-
ances viewed from the fault are
(b)

E
any increase in R, reduces the voltage
except when X o/Xl lies between -0.2 and

~co
1.0.
Since resistance is always present, it z, = [ro+ 3r n + j(xo + 3x n ) ]xeo
can be concluded that with Z, = Z2, Xco - (x o + 3x n) + j(ro + 3r n)
X the maximum fundamental-frequency
N .
ro +Jxo
s, + jx,
F voltage at the fault for the line-to-ground
(c) For a ground-fault neutralizer, -
fault is as great as, or greater than, that X co

for the double-line-to-ground fault, ex- X o + 3x n , and

Figure 7
cept for values of Xo/X I between -0.2
(a)-One-Iine diagram of power system R ~ (xo + 3x n)x eo
and -1.2, and very little resistance in the o
consisting of synchronous machine, transformer ro + a-,
system. This represents a condition that
bank, and transmission line
is not expected to be met in practice The coil is generally tuned so that ..Y"l)
(b)-Approximate positive-sequence funda- except under the most unusual conditions. has a very small positive value.
mental-frequency impedance diagram Resistance Neglected. With all resist-
(c)-Approximate zero-sequence funda- TRANSIENT VOLTAGES ance neglected in figure 7, transient volt-
mental-frequency impedance diagram Transient voltages here include both ages are expressed in appendix A in
the fundamental-frequency component of terms of the fundamental-frequency im-
voltage and the natural-frequency com- pedances viewed from the fault and the
= &, +
[X« is the zero-sequence imped-
ponents. Curves for transient voltages ratio of the positive to the zero-sequence
ance viewed from the fault and therefore
following faults in terms of system im- capacitive reactance.
includes the effect of neutral impedance.
pedances viewed from the fault cannot be With resistance neglected, the funda-
Comparing corresponding curves in
drawn for the general case, as has been mental-frequency impedances viewed
figures 1 and 2, it is seen that for the same
done for fundamental-frequency voltages. from the fault in figure 7 are:
system and fault location, that is, same
In the actual system, the transient volt-
values of XO/X I and Ro/XI, higher volt-
ages are affected by the number, connec-
ages are obtained for the line-to-ground
tion, and arrangement of the circuits.
than for the double-line-to-ground fault,
To simplify the work, and give an indica- . (xo + 3x n)x co
except for the R o/ Xl = 0 curve when Zo = J = 0 + j X.«
XO/X 1 lies between 1 and 5 and between tion of the maximum transient voltage xeo - (xo + 3x n )
to be expected, a system consisting of a
-0.2 and -1.2, approximately. With With (x, + 3x n ) greater than X co, X 0
synchronous generator, transformer bank,
XO/X I between 1 and 5, and RO/XI = 0, is negative.
the voltages for the line-to-ground fault and transmission line open at the distant
Figures 8 and 9 for line-to-ground and
end was considered. With the fault on
are but slightly less than those for the double-line-to-ground faults, respectively,
double-line-to-ground fault. As R o/ Xl is the line near the transformer terminals
give the maximum transient voltages in
as a first approximation, the open trans-
mission line was replaced by its lumped
Figure 8. Transient voltages with resistance capacitance at the point of fault. Figure
7, part a, gives a one-line diagram of the Figure 9. Transient voltages with resistance
neglected for system shown in figure 7.
system studied; partsbandc,respectively, neglected for system shown in figure ·7.
Line-to-ground fault. Maximum voltage-to-
show the positive- and zero-sequence Double-line-to-ground fault. Maximum
ground of unfaulted phases at the fault
voltage-Io-ground of unfaulted phase at the
-n-f~7
fault
t---t-----I~--- ~~- - -:,' I~ ~ ----rl-----t----i-t--+-+----+--+----+-----J
" 1 T
6
~ -t---t~f---+__+-+--+--l---+----l
6 6 ~-~,--~~

t
/, ,,' IfI
I ....J
I
I ~ <!
~
i i
I ~ --t---t-f--+--+-+---+--+----+---~ I l-
s:::>
lJJ
~~ I 5 ~ - - FUNDAMENTAL FREQUENCY I z ., i
r-T--+-----t---+--+-~.~¥__+--+__--+---J~----l ~ _ VOL TAGE5 I f! FUNDAMENTAL FREQUENCY
"" ~/ J f2 ---- MAXIMUM TRAN51ENT !.II 4'w
w(9
VOLTAGES
XCO"~CI""''' / 4~- VOLTAGES II
Ii I 2<1: - - - - MAXIf\.1UM TRANSIENl
-r ::J~J \JOLTAGt~
- -'xco='lXCI V ~~ " I
rr 3~~,
II

./V ~ 3 ~ --t---t-f-,--+--+-+--+--l---+----l I I T
_....
s --
,,~
I C(
1

- --
0
- -tL--------iI---r-t-l--"-+l.--"'"~.-£..--.t-\.--I----+--Uh-\\--! -t---t-t--+-~x---leo-.--:-;-)(eLI---+-_-._-+-- IX9f~~- V I 2 I I I

=-
I

r- -=-----L
I
-t-+--
--
2~ XCO')(CI
r-----r--+-----t----+--+-~---+-~--!--~-.l\.Ll\ 2 ~ _ ~_ - .:: : Xco: 2XCI L..--- ~ I ~ 2<~O:'[i~ :-
r---r---t---t--t--t---+--+------+----+---1----.J\t- ~ .... - _ - - - - xeo =.'2Xel '-- F ~ I-~ ~"'"'
J.,...::::-I-"
I fjV-
r-T--+-----t---+--+----+--------+----+---+~\~
,~~~~I---
11
-- f------
,J..~I=-t-r--+--+--+----+----+--+--+----.j
- -- -- -- -
II
I '--
I ~- --
-00 -10 -8 -G -4 -2 0 +'2 .. 8 +10 +00
XO/XI

380 l'RANSACTIONS Clarke, Crary, Peterson-Overvoltages During Faults ELECTRICAL ENGINEERING

7
~z \
\
w
t o
6« ~5
\\
1 ~
I 0 ~
>
\ 1\
I ex
, S-J0: u,

,\ ",
4 I
1 - FUNDAMENTAL FREOUENCY
-J4
I
t!
----
,;' I ~- VOLTAGES
z

I
I
4 z-
LU
---- MAXIMUM TRANSIENT
~3 \' -, i"-... r--. Xcl/xl2(
$
-
I I VOLTAGES
I
IV
!I
I
I
Q
z \ 1"0....
r-- r--
u.. \ ~
W
/ I
Xci/X, x ::J
I 10
-
/
---
3~
xco,xs.' :~ \ ..J
\1)2
\
r----
~ r--
----
..J
« ~ 5
- :d-
V 5 UJ
:'1--- xc~~ L ~ 2.5 ; -f--

7 2.5 r--- ~
\ 1\ 20:' .....

-~
\\

~
0
-
- 11._ z -- ~~_I
-
----- -- --
f---
if)
\\ w XCO' XCI XCI/XI :1.5
-~
~:;7
--- I~
t= ~.l)(C'

~
~
-10 -8 -G -4 "' I +4 +G +8 +10
- -10
'00 -8 -f> -4 -2 0 t2 t4 +0 +B + 10
-
+0<'
xo/x,
Figure 11. Natural frequency with line-to-
ground fault for system shown in figure 7,
Figure 10. Transient voltages with resistance unity. The region between 0 and -2
with resistance neglected and Xco = XCI
neglected for system shown in figure 7. Line- corresponds to a very low ratio of XCI/Xl,
to-ground fault. Maximum zero-sequence such as would probably not be encoun-
voltage at the fault tered in a practical system operating nor- 13. As will be noted, exceptionally high
mally. voltages are not obtained until the ca-
pacitive reactance x co becomes small com-
per unit of normal peak line-to-neutral METHOD OF GROUNDING
pared with Xl. This condition corre-
voltage on the unfaulted phases at the Reactance Grounding sponds" to a system having a very large
fault in terms of XO/X1 for the system Figures 8 and 9 can be used to deter- amount of line-charging capacity com-
shown in figure 7. The curves in figures 8 mine transient voltages with resistance pared with the connected generation.
and 9 were calculated from equations 19 neglected when the neutral is solidly Also, it will be noted that transient volt-
or 20 and 23, respectively, of appendix A grounded or grounded through any react- ages do not exceed twice the fundamental-
with (Jo = 90 degrees. Since resistance. is ance, including a ground-fault neutralizer. frequency voltages with either reactance
neglected there will be no decrement, and For a solidly grounded neutral, X n = O. or resistance grounding, for the conditions
therefore the peak value of the funda- For a ground-fault neutralizer with re- assumed.
mental-frequency term indicated in figures sistance neglected, 3x n +
Xo = X co, and The sustained voltages for resistance-
8 and 9 by full-line curves, and the peak Xo = ex>.
grounded systems are generally higher
value of the natural-frequency term are than those for corresponding reactance-
added directly (since they must eventu- Comparison of Resistance grounded systems. This is particularly
ally come in phase) to give the total and Reactance Grounding true if the neutral grounding ohms are
maximum voltage. Peak values of tran- A comparison of the effects of resistance selected to give the same value of short-
sient voltages are shown by dashed lines and reactance grounding on the funda- circuit current, as can be seen f rom an
for Xco = XCI and X co = 2x Cl • mental-frequency fault voltages following analysis of figures 12 and 13. For high
Figure 10, for the same conditions as a line-to-ground fault can be obtained ohmic values of neutral-grounding im-
figure 8, gives the maximum transient from figures 1, 4, and 5 when the positive- pedance, the transient voltages may be
zero-sequence voltage at the fault for a and zero-sequence impedances viewed higher for reactance-grounded than for
line-to-ground fault calculated from equa- from the fault are known. resistance-grounded systems. For low
tion 18, appendix A. The maximum To compare the effects of reactance and ohmic values, the natural-frequency com-
zero-sequence transient voltage at the resistance grounding on transient overvolt- ponent of voltage for resistance-grounded
fault for a double-line-to-ground fault ages following faults, the curves of figure systems decreases and becomes negligible
will be one-third the voltage of the un- 8 for a line-to-ground fault have been re- for values of 3r n/Xl < 5. Accordingly, a
faulted phase given by figure 9. See plotted in figure 12 in terms of the im- reactance-grounded system may not sub-
equation 23, appendix A. pedances indicated in figure 7. Figures ject the protective equipment to as high
With Xo/X I between 0 and -2 in 12 and 13 show the magnitudes of the sustained voltages as a resistance-
figure 8, and between 0 and -0.5 in transient voltages which may be obtained grounded system, but from the standpoint
figure 9, the natural frequency, WnJ is less following a line-to-ground fault when the of natural-frequency overvoltages, . par-
than unity. See equations 21 and 24, neutral is grounded through reactance ticularly those associated with switching
appendix A. Higher transient values and through resistance, respectively, with phenomena, the resistance-grounded neu-
than those plotted in this region could Xo = Xl and X co = XCI· tral may be more desirable.P'!"
therefore have been obtained with (Jo = The voltages in figure 13 and some Figures 14 and 15 show the correspond-
o degrees instead of 90 degrees. points in figure 12 were determined by ing neutral-to-ground voltages for the
The natural frequency, WnJ in per unity using a miniature system having the cases shown in figures 12 and 13, respec-
of fundamental frequency given by (21) characteristics and features described in tively. This information is of value in
for the line-to-ground fault with Xco = XCI appendix B. The full-line curves corre- indicating the voltages which may occur
is plotted in figure 11, with abscissa XO/XI spond to fundamental-frequency voltages from neutral-to-ground under the condi-
and parameter XCI/Xl. From figure 11, for different values of the ratio of 3x n / Xl tion of line-to-ground faults. Excep-
at the resonant point (XO/XI = -2), the or 3r n/Xl, where X nis the grounding react- tionally, high overvoltages are not ob-
natural frequency is unity and for values ance in the case of figure 12 and rn is the tained at the neutral except for systems
of XO/XI between -2 and 0, less than grounding resistance in the case of figure which have a relatively large amount of

AUGUST 1939, VOL. 58 Clarke. Crary, Peterson--Ooervoltages During Faults TRANSACTIONS 381
connected lme in proportion to the posi- 9). The maximum neutral-to-ground neutral voltages. See figures 1,3,4, and 5.
tive-sequence reactance. The curves, in transient voltage at any point in the sys- If low-resistance grounding is used, the
general, have the same shape as those in tem, for example, the neutral of an un- natural-frequency voltages are practi-
figures 12 and 13, as it is essentially the grounded bank, is about normal line-to- cally eliminated and the maximum volt-
shift or rise in neutral voltage which ground voltage due to the occurrence of a ages are essentially, the fundamental-
causes the overvoltage on the open phases. solid fault. See figure 10. frequency voltages which, however, are
Figures 16 and 17 are similar to figures 2. Neutral Grounded Through React- generally higher than the fundamental-
12 and 13, except that X co/ XCI = 2 in ance. When a system is grounded frequency voltages obtained with corre-
figure 16, and XO/XI = 0.5 in figure 17. through reactance less than that of a sponding values of neutral-grounding
Only transient voltages are plotted in ground-fault neutralizer, the zero-se- reactance.
figures 16 and 17. quence impedance viewed from the fault is 4. System Grounded Through Fault
Since the magnitudes of overvoltages inductive rather than capacitive and the Neutralizer. For systems grounded
which are obtained are considerably zero-sequence resistance is relatively through ground-fault neutralizers, with
affected by the method of grounding, this small; accordingly, the fundamental-fre- resistance neglected, X« is infinite; with
discussion is classified accordingly. quency phase-to-ground voltages will not resistance included, R o is very large while
1. Solidly Grounded System. There exceed normal line-to-line voltage, and X« is negative. Based on either assump-
are many degrees of solid grounding and the neutral-to-ground voltage will not tion, the fundamental-frequency voltages
the term means very little as far as the exceed normal line-to-neutral voltage. on 'the unfaulted phases at the fault fol-
overvoltages are concerned, as it depends See figures 1-6, with XO/X1 positive and lowing a line- to-ground fault are essen-
upon the number, location, and kilovolt- RO/X1 < XO/X I. tially line-to-line voltages. See figure 1.
ampere capacity of the grounding points. Following a fault, systems with react- The maximum transient voltages-to-
However, if the grounded-neutral system ance grounds will have maximum tran- ground on the unfaulted phases are less
is considered to be one in which grounded- sient voltages to ground on the unfaulted than 2.73 times normal, and of the neu-
neutral lightning arresters may be used phases not exceeding 2.73 times normal. tral-to-ground less than 1.67 times nor- ·
(line-to-ground sustained voltages not to The voltage to ground of the neutral will mal line-to-neutral voltage. See figures 8
exceed 140 per cent to 150 per cent of not exceed 1,67 times normal line- to- and 10.
normal), it becomes apparent that the neutral voltage. See figures 8-10. Higher voltages may be obtained at
ratio of X o/ Xl should be kept below about- 3. Neutral Grounded Through Resist- points removed from the ground-fault
3 or 4. See figures 1, 2, and 6. Accord- ance. When a system is grounded neutralizers where there is in effect con- ,
ingly, this will ordinarily mean that a through resistance, the zero-sequence centrated an appreciable amount of zero-
large percentage of all transformers or impedance viewed from the fault may be sequence capacitance to ground. This
machines must be solidly grounded. inductive or capacitive, depending upon indicates the advisability of placing the
Systems operating in this classification, the number and location of the grounding ground-fault neutralizer at the centers of
that is, XO/X1 not greater than 3 or 4, points and the amount of connected line the system, and also the desirability of us-
have a maximum transient line-to-ground or cable. With low-resistance grounds, ing, under certain conditions, more than
voltage on the unfaulted phases not ex- X o will ordinarily be positive and the one ground-fault neutralizer.
ceeding 2.0 time's normal (see figures 8 and fundamental-frequency phase-to-ground 5. Isolated Neutral. In the case of an
voltages will, in general, not exceed isolated-neutral system, X 0 is negative
normal line-to-line voltage and the neu- and of the order of magnitude of the
tral-to-ground voltages will not exceed capacitive reactance while Ro/X1 is
Figure 12. Transient and fundamental-Ere-
normal line-to-neutral voltage. With relatively small. From figures 1, 4, and
q~ency voltages with resistance neglected for
high-resistance grounds, X o may be
system shown in figure 7. Line-to-ground
fault. Maximum voltage-to-ground of un- negative. In that case, phase-to-ground
Figure 13. Transient and fundamental-fre-
faulted phases at the fault voltages may be greater than normal line-
quency voltages for system shown in figure 7.
to-line voltages, and neutral-to-ground Line-to-ground fault. Maximum voltage-to-
Xo = Xl XCO = XCI
voltages greater than normal line-to- ground of unfaulted phases at the fault

Wi
Xo = xi Xco = XCI
I I I
~ I- 3X n/X I = UJ I I
~ 2 5 5'2000'2000 ~ " ~lo20=r;:;;OO~~_-+--+-_ _--+-_-+--.l_L.--..L_..L-.....I.....-_ _---L-_-t
06
> 1\ \ \\ \ , ----TRANSiENT ~ frl
l
00
-----TRAN5IENT
FUNDAMENTAL - - FUNDAMENTAL
I'J \ \ ~5 I'
-I
« ~ \
\ FREQUENCY

'
\
FREQUENCY 100\\
~5
::> \\ \.\
\ \
~ II so\ '
\ '\ ~
UJ \ ~ I \ \, \
~4 ~ \ \
~4 \ ..., " ,\.
o
\ "\
\\ ~
, " " ... (; rr: ~
Z
' ..
~"~\ ,~, "

---- ------
l-
\ \ ~,\ ... ... , .... 1/",,- "'~', "" :-' . . _
,,~~~, _--..,::r-- __
I

~3 I : - --
"
~3
\1, .... , - --- 3Xn /Xl= =:rn~=-
-................ _- --- --
.... ....
..J ~\ 'I~ ---- - - - - - - - 00
L ----- ---
-"
-- -- --
20- -.J ~~.... 100-
..J
'\ .......
........ "'~ ...... ~ ....:::::::~ .........
--
2< _ _ I ' - .... - _ ----- 50
'- 5- ...", _ - - - - - _ _ _ _ _ 20-
~2 -::==--. -
- - --
............
~

...... i"--
,~
~ ...... - 00 ~ I ~-~ ~
0:
o
Z
............ ~- --- 1
20
I~
a
o o
2--
I
0
ZI 0-
<1>1 t/)
UJ UJ
~ ~
i=o
. 2 3 4 5 E) "7 6 10 20 30 40 50 GO 80 100 200
-
00
~ O2 3 4 5 ~ 8 10 20 30 40 50 60 eo 100 200
XCO/~I xco/ x,

382 TRANSACTIONS Clarke, Crary, Peterson-r-Ooervoltages During Faults ELECTRICAL ENGINEERING

W
G ~3x.Jx,=
I I
~
s
o
00

,,
- - - - TRANSI ENT
- - FUNDAMENTAL
FREQ.UENCY
3 r n / xl =
00
----TRANSiENT
- - FUNDAMENTAL
>5 ~
I,
FREQUENCY
.J
4:
,
\ \

...a::>4 ~.
\

,
\ \
\

,'\
UJ I 20~
Z \ ~ 50
I
1\ ,
o
t;3
\ \
, \
,
&'
\
20 \
,,
~ ~oo
uJ

,"
f-OO

~ ~ ~
\
Z \ 20 , 1\

\ , ' .....
-- .........-- ...---
J ,

--- -- -- -1- ~ ", "

_--- ----
20 """ ............
;i2 "
- --_
"
~ 3xn/XI~
..... '"""-,

-- --
~
5~ ~
--00 .... 3rn/XI=
20-
~ ~ ......
- -
~ 1___ ~-

--- --
a: ~-

--
5 !-oo -~ 00
'--' t"'-.... ......
o ~~~ I----- f- :5
i " r' 1""--

----- - -- --
·5"",,-

--- -- ... -----

~~ ~- 50
ZI
2 -..r-::- "'---
00_
20 ~-
~ ~~ ~ .00
20-
----- --- --
(/)
UJ 2 -..... 5
~ 1-
I I--
~-
2- 5
1: I ~- I
i=o
2 4 5 (0 8 10 30 40 50 GO 80 100 200
3 4 5 G 1 8 10 20 30 40 50 GO 80 100
XCO/XI
Figure 14. Transient and fundamental-fre-
quency voltages with resistance neglected for those occurring at the point of fault may Figure 15. Transient and fundamental-fre-
system shown in figure 7. Line-to-ground be obtained. This is particularly true at quency voltages for system shown in figure 7.
fault. Maximum voltage-to-ground at the points which are located at a distance Llne-to-greund fault. Maximum voltage-to-
neutral ground at neutral
from the fault and the grounding points,
Xo = Xl Xco = XCI but located where there is, in effect, an Xo = Xl Xco = XC!
appreciable amount of zero-sequence
capacitance to ground. The fundamen-
5, the fundamental-frequency voltages tal zero-sequence voltage at such points tant from the fault than at the point of
obtained may be in excess of normal line- can be readily calculated from the zero- fault has been discussed in a recent paper."
to-line voltage; and in some cases, par- sequence voltage at the fault and the
ticularly when the system is large in ex- zero-sequence network. Figure 18 shows Nomenclature
tent, considerably in excess, so that a the simplified zero-sequence impedance
very undesirable condition is created diagram with the identity retained of the Z, = R1 + +
jX1, Zo = Ro [X« = positive
when faults occur. The fact that such and zero-sequence fundamental-fre-
fault point F and point P, at which
voltages may be obtained makes it highly quency impedances, respectively,
voltage is required. The zero-sequence viewed from the fault point
desirable that under no condition of voltage at P is
operation shall a system lose its grounding R, = fault resistance
points, if by so doing it is liable to be in Vao (at P) = Vao (at F) ~Z Zl(P), Zo(p) = operational expressions for
the region of resonance. Otherwise, ZlJ + Z
the positive and zero-sequence im-
damage to the protective equipment or With Z; capacitive reactance and ZlJ pedances, respectively, viewed from
flashover of major equipment may result. inductive reactance, the zero-sequence the fault point
The minimum voltage rating commonly voltage at P will be higher at P than at F. Lower case letters apply to the circuit shown
used in isolated-neutral arresters is 1.83 If it is appreciably higher, additional in figure 7.
times normalline-to-neutral voltage. calculations are required for determining' Zl = '1 + j X lI Zo = '0 + [x« = positive
both fundamental- and natural-frequency and zero-sequence fundamental-fre-
VOLTAGES DISTANT FROM FAULT
voltages. The case of higher funda- quency impedances, respectively,
The fundamental-frequency and tran- mental-frequency voltages at points dis- from neutral to the fault point
sient voltages given by the curves in this r« + i»« = fundamental-frequency
paper are at the point of fault. Under impedance between neutral and
certain conditions higher voltages than ground
1\ X Cb XCO = positive and zero-sequence funda-
4 mental-frequency capacitive re-
~~
~ !""'iiI actances, respectively, of trans-
1\ r-, 3rn~XllJ-
~ ....
~~
~ ......
~ """"'--- - ""--- 3r~/xl= 00-
mission line
+
f-t--
~~
~
~
5' fJ = (}o t = angle between direct axis of
r--.~
50..- f - - t - r-.....i"-o -~~
<0 reference machine and axis of phase a
~ ""'--
~

20_ f - - t -
5
5 Wn natural frequency in times funda-
I
mental frequency
5 10 50 100
10 50 . 100 XCO/XI
XCO/Xl

Figure 16. Maximum transient unfaulted- Figure 17. Maximum transient unfaulted- Appendix A. Fundamental- and
phase voltage-to-ground for system shown in phase voltage-to-ground for system shown in
figure 7. Line-to-ground fault. Neutral
Natural-Frequency Voltages
figure 7. Line-to-ground fault. Neutral
grounded through resistance grounded through resistance
The instantaneous phase voltages at any
Xo = xi Xco = O.5xcl Xo = O.5xI XCO = XCi point in a balanced three-phase system in

AUGUST 1939, VOL. 58 Clarke. Crary, Peterson-OVervoltages During Faults TRANSACTIONS 383
per unit ot crest voltages to neutral or to Zo - Zl In terms of X« and Xl,
ground at that point are
2Z1 + z, + 3R!
ea =- sin 8 (10) Xo •
stn 8 -
eao= X
+X
+ 20
2 1 0
1 .
= 2" sm 8 cos 8 1.0 ZO - z,
eb
(1) ec = - 2+ J 2 - 2Z 1 + ZO + 3R! (11) 2 (XCI X O - Xl\
1. 0 Xco ) X

(1 + 2::J<2Xl + X
ec = - sin 8 - - cos 8
Double Line-to-Ground Fault (Phases b and c)
2 2 o)
Let e and i with appropriate subscripts
refer to instantaneous values of voltages-to-
ground at the fault and currents flowing into
Zo
(12)
(Sin 00 cos wnt + ± cos 00 sin wnt) (18)
the fault, respectively. In reference 17 are
given the relations for transformation from
3Z o + 6R! (13) with Xco = XClJ equation 18 checks that
three-phase quantities (a, b, c) to sym- given by Fallou.!?
metrical-component quantities for the case
of instantaneous currents and voltages,
Transient Voltages eb = - eo
3
2
+ -02 cos 8 (19)
including the transient as well as the funda-
mental-frequency components. An appli-
To obtain the natural-frequency com- 3 ,,/3
cation of these transformation relations is
given in reference 5, except that a and (3 ponents of equations 2-6 necessitates replac- ec = 2 eo - 2 cos 8 (20)

components are used. However, for the ing ZI{P) and Zo{p) by their operational ex-
case under consideration, Za(p) = Zf3(P) = pressions in terms of system constants. where
ZI(P) = Z2(P), so the application here is This can be done only when the system is
quite similar. U sing the generalized given. For the system represented in fig-
method of symmetrical components, the ure 7,
following equations are obtained.
Line-to-Ground Fault (Phase a)

~
ZI{P) (14) XO(XCI + 2 XCI\ + 2 (XCI - 1\
..
- ~ - ---------
- sin 8 _ X I Xl X co) XI )

XCI +l)
-
+ +
- XO(2
~
'11 - a2 - co - 2Z I(p) Zo(p) 3R!

. Zo(p) sin () Xd) Xl x co

= - ~aoZO(P) = 2Z I(p) +
Zo(p) + 3R! [ro + 3r n + p(xo + 3x n ) ] - (21)
QO
P
(2) Double-Line-to-Ground Fault. With a
0
ro + 3r n + p(xo + 3xn ) + - Xro
double-line-to-ground fault at F on phases b
1 P
ell = - sin () + -
2 2
cos 8 + (IS) and c, the instantaneous zero-sequence and
phase a voltage-to-ground at the fault in per
Zo(p) - ZI(P) . (3) Resistance Neglected. With resistance neg- unit of normalline-to-neutral peak voltage
- - - - - - - - Sin 8
2Z I(p) +
Zo(p) 3Rt + lected in figure 7, ZI(P), Zo(p), ZI, and Z« at the fault are obtained by substituting
(14a) and (15a) in (5) and (6) and solving
become,
1
ec = - sin 8 -
0
- cos 8 +
the operational equations. Expressed in
2 2 terms of the positive- and zero-sequence
ZI(P) (14a) reactances viewed from the fault,
Zo(p) - ZI(P) . (4)
- - - - - - - - sin 8 Xo
2Z I(p) + Zo(p) + 3R! +
Zo(p)
+ 3xn)xco
p(XO
(ISa)
- - - - sin 8
Xl + 2X o
Double Line-to-Ground Fault (Phases b and c) p 2(xo + 3x n ) + Xco
sin 8 jXIXci .
- - - =JX I (16)
XCI - Xl
-Zo(p) sin 8
- ioZo(p)
+ 2Z o(p) + 6R! Z«
• (XO + 3xnJx co
ZI(P)
(5)
= J
X(o - (Xo + 3xn ) = jX o (17)
( sin 80 cos wnt ----;:;;:8 sin wnt )
+ COS (22)

e = -
3Z o(p) 6R! + .
sin 8 (6) Line-to-Ground Fault. With a line-to- (23)
a ZI(P) +
2Z o(p) 6Rf + ground fault at F on phase a in figure 7, the
instantaneous zero-sequence voltage at the where
Line-to-Line Fault (Phases b and c) fault in per unit of normal line-to-neutral
ea = - sin 8 peak voltage at the fault, obtained by
(7) substituting (14a) and (15a) in (2) and
1 solving the operational equations is
eb = ec = - sin 8 (8)
2
Zo
eo = sin () -
Fundamental-Frequency 2Z 1 + z,
Components of Voltage
2(
xo + 3Xn --
Xl)
F Zy
Introducing vector quantities- by replac-
ing - sin 8 by unity, cos 8 by -j, and ZI(P)
and Zo(p) by Zl and Zo, respectively, in
(2)-(6), crest voltages of fundamental-fre-
quency are obtained.
Line-to-Ground Fault (Phase a)

~ cos 0
Zo Figure 18. Simplified zero-sequence imped-
(9) (Sin 00 cos wnt + 0 sin wnt)
ance diagram

384 TRANSACTIONS Clarke, Crary, Peterson-r-Ooerooltages During Faults ELECTRICAL ENGINEERING

Appendix B. Description of Subcommittee on Development and Research-
Edison Electric Institute and Bell Telephone
age are considered as the constants of the
circuit will change in a direction tending to
. Miniature System System, October 25, 1934.
limit the magnitude of voltage. Such
3. SYMMETRICAL COMPONENTS (a book), C. F.
Wagner and R. D. Evans. McGraw-Hill Book
changes in circuit constants may be due to
The miniature equivalent system used in Company, chapter 11. corona, magnetizing current of transform-
making the laboratory tests is shown in ers, etc. The practical range of the curves
4. OVERVOLTAGE ON TRANSMISSION SYSTEMS
figure 19. CAUSED BY DROPPING LOAD, E. J. Burnham. is probably from -1 ratio to the positive
The reactors used in this system had low AlEE TRANSACTIONS, March 1932, volume 51, values.
ratios of resistance to reactance which pages 105-12. A value of XO/Xl of 3 to 4 is suggested for
varied from about 0.01 to 0.04 at 60 cycles 5. OVERVOLTAGES CAUSED BY UNBALANCBD a safe application of a grounded-neutral
depending upon the portion of the total SHORT CIRCUITS-EFFECT OF AMORTISSBUR WIND- lightning arrester. It is believed that in
INGS, Edith Clarke, C. N. Weygandt, and C. Con-
winding being used. The miniature power cordia. ELECTRICAL ENGINEERING, August 1938, drawing this conclusion the effects of fault
volume 57, pages 453-68. resistance have been ignored. When high
6. UNSYMMETRICAL SHORT CIRCUITS ON W ATBR- values of fault resistance are considered the
WHEEL GENERATORS UNDER CAPACITIVB LOADING, ratio of X o/X, must be lowered if the fault
C. F. Wagner. ELECTRICAL ENGINEERING, No- is at the point of application of the arrester.
vember 1937, pages 1385-95. In the work published by Monteith and
7. ARCING GROUNDS AND EFFECT OF NEUTRAL Roman- it was considered improbable to
GROUNDING IMPEDANCE, J. E. Clem. AlEE have a fault at the point of application of the
TRANSACTIONS, July 1930, page 970.
arrester since the arrester should prevent
8. ApPLICATION OF STATION-TYPE LIGHTNING flashover. When the fault is considered at a
ARRESTERS, A. C. Monteith and W. G. Roman.
Electric Journal, March 1938, pages 93-9. point other than where the arrester is ap-
Figure 19. Miniature equivalent system plied, the ratio of 3 to 4 for XO/X l , can be
9. SYSTEM ANALYSIS FOR PETBRSEN COIL ApPLI-
CATION, W. C. Champe and F. Von Voigtlander. selected and the criterion will be entirely
AlEE TRANSACTIONS, volume 57, 1938, pages 663- independent of fault resistance.
72 (December section). (See also discussion, page 412.)
transformer was also of low-loss design so
10. MISB A LA TBRRE DU POINT NBUTRB PAR
that the entire system caused relatively L'INTEMEDIAIRB D'UNE INDUCTANCE, Jean Fallou.
low damping of natural-frequency oscilla- Revue Generale de l' Electricite, May 10, 1930. REFERENCES
tions. This was desirable for purposes of 11. AN INVESTIGATION OF GROUND FAULTS ON A 1. SOME EFFECTS OF UNBALANCED FAUI.TS, R. D.
checking calculations based on no-loss cir- 33-Kv TRANSMISSION SYSTEM, Engineering Report Evans and S. H. Wright. ELECTRICAL ENGINEBR-
cuits. Number 4 of the Joint Subcommittee on Develop- lNG, June 1931, pages 415-20.
ment and Research.
The miniature system was energized from 2. ApPLICATION OF STATION-TYPB LIGHTNING
a three-phase 110-volt 60-cycle voltage 12. EXPERIMENTAL STUDIES OF ARCING FAULTS ARRESTERS, A. C. Monteith and W. G. Roman.
ON A 75-Kv TRANSMISSION SYSTKM, Eaton, Peck, Electric Journal, March 1938, pages 93-8.
source of such capacity that its terminal and Dunham. AlEE TRANSACTIONS, December
voltage remained essentially constant re- 1930, page 1469.
gardless of transient conditions imposed by 13. PETERSEN COIL TESTS ON 140-Kv SYSTEM,
faults in the equivalent circuits. J. R. North and J. R. Eaton. ELECTRICAL ENGI- Edith Clarke, S. B. Crary, and H. A. Peter-
A synchrenous commutator was used to NEERING (AlEE TRANSACTIONS), volume 53, Janu- son: In answer to the discussion by
apply and remove repeatedly the fault in ary 1934, page 63.
Messrs. Evans, Monteith, and Witzke:
synchronism with the system voltage. The 14. VOLTAGES INDUCED BY ARCING GROUNDS, The simple network of figure 19, although
commutator drum was driven by an 1,800- Peters and Slepian. AlEE TRANSACTIONS, volume
42, 1923. page 478. an approximation, gives transient voltages
rpm synchronous motor by means of a ten- at the point of fault not less than the maxi-
to-one gear reduction so that one revolution 15. EFFECT OF RESTRIKING ON RECOVERY VOLT-
AGE, C. Concordia and W. F. Skeats. AlEE
mum transient voltages which may occur
of the drum was completed every 20 cycles. TRANSACTIONS, volume 58, 1939, pages 371-6 (see "Basis of Study," 9), and thus enables
The contacts on the switch were so made (August section). us to draw conclusions in regard to the
that during these 20 cycles the fault would 16. POWER SYSTEM VOLTAGE RECOVERY CHARAC- maximum transient voltages which can oc-
be on 5 cycles and off 15 cycles. The rela- TERISTICS, H. A. Peterson. AlEE TRANSACTIONS, cur at the point of fault for various condi-
tively long period during which the fault volume 58, 1939, pages 405-13 (August section). tions of grounding.
was off afforded ample time for steady-state 17. RELATIONS AMONG TRANSFORMATIONS USED The curves for negative values of Xo/X J
conditions to be reached before the switch- IN ELECTRICAL ENGINEERING PROBLEMS, C. Con- have been found practical in determining
cordia. General Electric Review, July 1938, pages
ing operation was repeated. 323-5. voltages corresponding to high negative
A cathode-ray oscilloscope was used to values of X 0/X 1 beyond the resonance re-
measure the transient voltages. Another gion. As pointed out in the paper, the very
contactor on the synchronous commutator high voltages indicated by the curves would
was used to control the grid on the cathode- not be obtained in the resonance region,
ray tube so that the light beam would appear Discussion since they are critical values and therefore
on the screen for only one cycle out of each greatly influenced by saturation, corona,
20. This made it possible to obtain a clear P. A. Jeanne: See discussion, page 395. and variation of machine reactances with
picture of the transient voltages during the time (see "Basis of Study," 5 and 6). They
period of particular interest. are useful, however, in that they indicate
The contactors on the commutator were R. D. Evans, A. C. Monteith, and R. L. regions where high voltages will be obtained,
mounted on a rack which could be rotated to Witzke (all of Westinghouse Electric and that is, voltages in excess of normal line- to-
vary the instant of application of the fault Manufacturing Company, East Pittsburgh, line voltages.
corresponding to various points on the volt- Pa.) : Concerning the results in the Clarke- The effect of fault resistance was not
age wave. Thus it was possible to deter- Crary-Peterson paper, it may be observed ignored in suggesting a value of XO/X1 be-
mine the maximum voltage that could be that the simple form of network illustrated tween 3 and 4 for safe application of a
reached by selecting the angle of fault appli- in figure 19 for transient analysis, might grounded-neutral arrester. This conclusion
cation which gave the highest transient lead to appreciable error. The use of such is based on the curves of figure 6, which in-
voltage on the oscilloscope. a simple equivalent circuit reduces to an clude the effect of fault resistance. The
extent the usefulness of the results. curves apply to cases in which the arrester
Also, in connection with this paper, the and fault locations are far enough from each
Bibliography authors have carried the curves to negative other so that mutual ground resistance of
value of XO/Xl whereas in the previous pub- the arrester and fault is not encountered,
lished work by Evans and Wright! only and yet not so remote from each other that
1. SOME EFFECTS OF UNBALANCED FAULTS, R. D
Evans and S. H. Wright. ELECTRICAL ENGINEER-
values for positive ratios were considered. voltages to ground at the arrester are ap-
ING, June 1931, pages 415-20. The curves in the negative zone are no doubt preciably different from those indicated.
2. OVERVOLTAGES ON TRANSMISSION LINES DUE
correct from an academic point of view. Figure 6 in the region of X 0/X 1 = 1 to 5
TO GROUND FAULTS AS AFFECTED BY NEUTRAL However, we are inclined to question their checks the curves of figure 4 given by
IMPBDANCE, Engineering Report Number 30, Joint practical use when high values of overvolt- Messrs. Monteith and Roman in reference 8.

AUGUST 1939, VOL. 58 Clarke, Crary, Peterson-r-Ooerooltages During Faults TRANSACTIONS 385