IEEE Transactions on Power Delivery, Vol. 14, No.

3, July 1999 897

A PRACTICAL APPROACH FOR COMPUTATION OF GRIDCURRENT

Hans R. Seedher, Senior Member, IEEE J. K. Arora, Senior Member, IEEE S. K.Soni
Department of Electrical Engbe&ng,- PUIljab
- Engineehg
- College H.P. State Ele&k&y Board
chandigartz India-

A M Whenagmundfitultoccursatasubstation,thepart of the earth fault current depending on the location of fault,

of the fault current that flows between the groundihg system c o d p a t i o n and parameters of ground wires and phase
andthe sumuunding earth is known as grid current. This current conductors, and ground resistance of the station. The problem
dekmines the magnitude of the dangerous voltages within and of determination of grid current has been dealt by several
around the station. It may vary fiom a few percent to almost researchers and a number of analytical methods have been
1WA of the earth fault current. It is necessary to make an reported [l-51.In a survey conducted by IEEE 161, it was,
accurate estimate of grid current for an economical and safe however, found that majority of utilities the world over did not
design of ground electrode. The available analytical methods for appropriately account for current diversion by the alternate
determining the grid current require complete system data, paths to determine the maximum grid current. The maximum
which is voluminous and often unavailable. In this paper a value of the ground fault current or an arbitrary fiaction of it
simple and accurate d y t i c a i method for det- the grid wasbeing used in place ofthe maximum grid current. This may
current is presented. The method requires limited daw A be because of the requirement of an elaborate set of data about
computer program based on the proposed method has been all the transmission lines, ground wires, transformers and
developed and tested. generators of the system for application of analytical methods
for its determination. In such methods, determination of grid
Keywords: grounding, earth fault, substations, grid current, current is an extension of the algorithm used for determining
ground wires, step voltage, touch voltage, transfemd potential fault current. A number of approximate simpWed methods
h v e also been proposed [7-91for evaluation of grid current.
L INTRODUCTION In the method proposed by Thapar and Madan [7,81, the
current diverted by aerial ground wires has been divided into
Gmundii system of a generating station or a substation is two components. One of these components is the current
designed with the prime objective of providing safety to diverted fiom the station through conduction by ground wires of
personnel during an earth fault. The fault current during an the transmission lines,contributing to fault current, which are
earth fault has s e v d alternatepaths for returning to the sources connected to the grounding system of the station. The second
which feed the M t . A part of the current flows between the part is the current diverted by the ground wires because of
growding system and the surrounding earth for returning to the p mutual induction between ground Wires and phase conductors
sources of origin. The remaining current may return through of the respecti\e transmissionh s . The method, however, may
ground wires or may flow through a metallic path consisting of give erroneous results for many situations. In case of two
the conductorsof the grounding system and its cotlnectionto the circuits which are coupled conductively and inductively, it is not
neutrals of the sources of supply. The component of the fbdt appropriate to separately calculate the current in a circuit as two
current that flows between the grounding system and the components, those becauseof conduction and by induction, and
surroundingearth is d e d grid current. Onty this component of then use superposition to obtain the total current as done by
the current is responsible for creating dangerous voltages, Thapar and Madan. Further, in actual implementation of the
within or around the station. Evaluation of grid current is thus method in [7,8], the current diverted by ground Wires of lines
of paramount importance for the design of a grounding system. not contributingto the fault has not been taken into Bccount. It
The grid current may vary from a few percent to almost 100% has been shown by Joy, Meliopoulous and Webb [5] that a
substantial amount of current may be diverted by ground wires
PE-168-PWRD4-09-1998 A paper recommended and approved by of such lines.
the IEEE Substations Committee of the IEEE Power Engineering Garrett, Pate1 and Myers[9] have prepared a number of
Society for publication in the IEEE Transactions on Power Delivery. graphs drawn on logarithmic scale for obtaining ratio of grid
Manuscript submitted April 22. 1998; made available for printing current and earth fauh current. These graphs, obtained by using
October 7,1998.
computer program of [5], however, do not fit into all practical
situations. Interpolation and a p p r o d o n have to be used in
most cases. Graphs only provide a rough estimate of the grid
current.
In this paper a simple but accurate methd is developed
for computation of grid current at a station. Data
requirement of the method is same as that of the method of

0885-8977/99/$10.00 0 1998 IEEE

898

[7,8] and S i to that of 191. The method uses simple remote stationsvis a vis that h m the local sources. The current
equations to solve for the current and at the same time removes supplied to the hult by the local sourcesflows through a metallic
limitations of these two methods [7-93. As in [7,8] as well as path consisting of the conductors of the grounding system and
[9], it is assumedthat the ground fault level for buses Within the Connections to the neutrals of the sources of supply, and does
station is known from the short Circuit studies. A computer not contn'bute to grid current. The fault current supplied
program with the symbolic name PAG @racticalbproach for through a transmission line has two paths for returning to the
computation of m d current), which is based on the proposed
method, has been developed and tested. The data requirement
source. A part of it returns to the source through ground wires
of the program is quite simple. and the rest of it flows between grounding system of the station
and the source through earth. The grid current is thus equal to
ILRELATION BETWEEN FAULT the difference of the current fed to the fiiult through transmission
CURRENT AND GRID CURRENT lines and the current diverted by ground wires. Generally for a
power station forming part of an interconnected system, the
The grid ment Ig at a station can be expressed as a product of maximum value of grid current is obtained when the fault is
four factors [ 101 as inside the station [8,9]. An algorithm for determination of
(1) current division factor, when a single line to ground fault occurs
Ig =If Df Cp Sf inside a station, is developed in the following sections of this
where
If = Symmetrical value (without taking into 8ccount d c paper.
offset) of ground fiiult ament for the fault case resulting IV. MODEL OF GROUND WIRE
in the maximum grid current, A
Dr = Decrement factor to take into 8ccount d c offset A ground wire is connected to the earth at various towers
c, = Corrective projection factor accounting for future through the tower footing resistance. It can be represented by
increase in fault current during the substation life span a ladder network as shown in Figure 1. The impedance zs
is
Sf= Current division factor, fraction of total ground fault Station ZS ZS zs
current that flows between the grounding system and the
surrounding earth
Values of factors Df and Cp are discussed in IEEE Standard
80 [lo]. For design of ground electrode, the value of fault
duration is usually taken 0.5 second or more for which Df may
be assumed as 1.0. The current projection factor Cp takes
into account the increase in the fault current due to future
Figure 1. Ladder network representation of a ground wire
growth Its value can be estimated fiom the available
projections. The value of fault current at a station depends on the selfimpedance of an average span of ground wire with earth
the type of fault and the fault location. 1, in (1) should return. Each shunt branch has an impedance equal to the
correspond to such a M t location and fault type as result in the average tower footing resistance Rt,. The input impedancez,
greatest flow of current between the grounding system of the of the ground wire can be determined as input impedance of the
station and the surrounding earth. From a parametric study in ladder network consisting of number of sections equal to
[5J, it is concluded that number of spans of the line [l 13. If the number of spans is 20 or
i) for a given fault location, the maximum grid current is more, the network of Figure 1 can be considered as an infinite
generated from single line to ground or double line to ground
ladder network for the purpose of determining its input
fault, and
ii) for practical power systems, the grid currents for single line impedance [SI. Input impedance 2, of infinite ladder network
to ground fault and for double line to ground fault are is [1,8]
approximately equal.
For much higher probability of occurrence, only a single line to
ground fault may be considered for computation of maximum
grid current. The selfimpedanceper metre length of the ground wire can be
obtained as
ICL CURRENT DIVISION FACTOR
The current division factor S depends mainly on location Zg=r, + 9.87 x 10-7f +
l
of f d and on overhead ground m e s connected to the station. j28.94 x 10-7f log,,(D,lGMR) (3)
The former determines contrihtion to the fault current from the
 899

where conductors of the transmission line are replaced by an ideal

= self impedanceof the ground wire in ohm/m m ent source of magnitude I, in series with the self impedance
Zg
rc = resistance of the ground wire in ohm/m of the line. Since the impedance in series with an ideal current
f = fkquencyinHz source has no e&t on the rest of the network, its value need
De = equivalentde ofearthreturninm not be known. The SdfimpeaanCe zeof the ground wire can be
= 6 5 8 . 4 d (4) obtained by (2) if the number of spans of the line are 20 or moxq
p = resistivityofearthinoiun-m for a shorter line, it is to be obtained as the input impedance of
GMR= geometric mean radius of ground wire in m the ladder network of Figure 2. The mutual impedance per
The impedance of the series ann zsof the ladder network is mege, zP, betweenthe p&*e he conductors
zs=zgx
Is (5) is [121

where
Is = average span length of the line in m
zp = 9 . 8 7 x 10-7f+
j28.94 x 10-7f log,,(De/GMDsep) (6)
V. MODEL OF TRANSMISSIONLINE

A transmission line with one or more ground wires is to be where

modelledsuch that diversion of currentby ground wires can be GA4Dsep= geometric mean distance between ground wire
computed. To develop such a model a generating station
supplying power to an interconnected system through a step up and phase conductors in m.
transformer and a singie transmission line, shown in Figure 2, is
considered. The current flow when a single line to ground fault
le ZQ
occurs on the HY side of transformer is shown. The

Figure 3. Equivalent circuit for computation of cur-

* f
1 I F
If /M
rent diversionby ground wire in Figure 2

The mutual impedance Zm between phase conductors and

Grounding system ground wire, shown in Figure 3, is assumed to be the same
of the station fraction of Z p as is Ze of Zg . It is obtained as
I g = Ir - Ie zm=Zg,X(ZelZg) (7)
Figure 2. Path for flow of the current supplied to the fault Polarity of the e m f i n d u d in the ground wire due to
by remote source through transmission line mutualcoupling with the phase conductors would be as per the
dot markings shown in Figure 3. By Writing loop equation the
current supplied to the fault by the Iocal source does not current diverted by ground wire, le,is
contribute to the grid med . Only the current 1, supplied to
8 'e =Ir(Zm +',)'('e +Rg) (8)
the fault through transmission line contributes to the grid
and grid current Ig is
current. A part of this current, Ie, is returned to the source
through the ground wire. The grid currend is thus equal to Ig =Ir -Ie (9)
g The equivalent circuit for a station with a number of
1,-Ie.To mmputethe grid current I ,current lediverted by
the ground wire, is to be determined. 6urrent I returns to the transmission lies and feeders, for computing current diversion
g
remote source through soil, and resistance offered to its flow is by ground wires, can be obtainedby extension of the equivalent
equal to the station ground resistance R . circuit of Figure 3. The equivalent circuit for a station with n
An eqkdent circuit of the system of Figure 2 is shown in lines is shown in Figure 4. In the figure, the following notations
Figure 3. By application of substitution theorem [131, phase are used:
900
In matrix form, these equations can be written as
AZ,=B (14)
where A is an n x n square matrix with the elements
a..
!
I
= Rg
J

= R, + Ze, i =J'

Z, is a column vector consisting of currents Iel, le,, . ..,Ienand

B is a column vector withb, in ith row as
b, =RgIr+ Zm,IrI (16)
If ith line is a feeder, current'., is zero. The solution of (14)
gives the currents le,,Ie2,... I diverted by ground wires.
' en
Grid current 1, is obtained from (9).

VIL COMPUTER PROGRAM PAG

Figure 4. Equivalent circuit for a station with n lines
The equations of the previous section can be easily
I = current fed to the fault by ith transmission line
programmed on computer. A computer program with the
rl symbolic m e PAG eractical uproach for computation of
z.,= self impedance of ground wire of ith trammission line @id current) has been developed. With this program grid
current at a generating station or a substation can be computed.
%
' I, = mutual impedance between ith transmission line and its BesidesContributions to earth fault current fiom Merent lines,
ground wire for a line to ground 6 d t at the station, the required data
umsists of the self impedances of the ground wires and mutual
I = current diverted by ground wire of ith transmission line mpedama between the phase conductm and the ground wires
e,
ofthe rqective lines. In case these impedances are not known,
1, = current fed to the fault by all the transmission l
ine these can be computed in the program by spedjing i)
n
frequency, ii) average soil resistivity, iii) resistance per metre
= Irk (10) length of ground wire, iv) geometric mean radius of the ground
k =1
= current diverted by ground wires of all the transmission wire, v) geometric mean distane between the ground wire and
1, n phase conductors, vi) average tower footing resistance, and vii)
tines and feeders = Iek the average line span.
k =1
A feeder connected to the.station is modelled as a transmission Vm.TEST PROBLEMS
line except that its contribution to the fault current is zero.
Computer program PAG has been tested by using it to
VL COMPUTATION OF GRID CURRENT determine grid current for threetest problems. Results for these
test problems as obtained by computer program SMECC [5]
Equationsfix determining current diverted by ground wires which was beloped at EPRI and which requires elaborate data
are obtained from equivalent circuit of Figure 4. For the are availaMein literature[5,9]. The results obtained by PAG for
equivaient Circuit with n ground wires, we can write for the loop the test problems are compared with those obtained with
formed by ith ground wire, station ground r e s i c e and the SMECC and the approximate methods of Thapar and Madan
earth and of Garrett, Pate1 and Myers [9].
The first of the test problems is taken from IEEE Tutorial
Course [9]. Single l i e schematic diagram of the substation of
this problem is shown in Figure 5. The substationsfeeding the
Ot 115 kV lines are representedby equivalent sources. A delta star
transformer suppliesthe 12.47 kV bus to which three feeders are
connected. A single line to ground fault occurs at the 115 kV
bus. By using the sequences impedances obtained from the data
given in [9], the fault currents fed by the lines are found to be
3358 A and 5947 A. By using program PAG, the grid current
 901
r---
-1
115kV Line
0
Equtvalent
I 3
I source I----- Load
Substation
under study

Figure 6.Single line diagram of the system for test problem 2

from a 1 15 kV main line. At the station, a staddelta transformer

Feeders steps down the voltage and feeds a 12 kV feeder. The feeder
consists of a section of overhead line, a cable followed by
Figure 5. Single line diagram of the system for test problem 1 another section of overhead line. To reduce the operating
voltage during the test a mobile deltdstar transformer was
is obtained as 2527 A as compared to the value of 2515 A by connected at the feed end of the 115 kV main line. This
SMECC. transformer was rated at 2 MVA and had voltage ratio of
The second test problems is h m EPRI report [5]. A sipisle 2a4.16kV;theactuai tap used was 18.7/4.16kV. In the EPRI
line diagram for the problem is given in Figure 6. The report the computed value of hult current is 98.2 A and the grid
substation k x h g the 115 kV tmnmision line is represented by current is 62.0A, by the method of this paper the grid current
an equivalent sowe. At the substation under study, the voltage is found to be 62.6 A. The grid current measured [ 5 ) by
is stepped down to 12 kV with a de1Wsta.r transformer. A 12 creating a single line to ground fault at the Texas Valley
kV feeder supplies a three phase load. The grid current would substation was 73.2A, the report ascribes the discrepancy to
bethemaximumforasinglelinetogroundfaultonthe 115kV miations in the assumed values of tower footing resistance and
bus in the station. In the EPRI report [5], the fault and their actual values
the grid current obtained with SMECC program for such a Eault For comparison, the test problems are also solved by the
are 1492 A and 742 A respectively. The value of grid current by approximate methods of Thapar et al [8] and Garrett et al[9]
the method of this paper is 745 A which require Siplified data similar to that required by PAG.
The third problem [5] describes the results of a test camed The value of grid current for the three test problems as obtained
on the system shown in Figure 7.The Texas Valley substation, by the program SMECC,the method of this paper, and by the
where the fault is simulated, is supplied by a tap line methods of [8] and 191 are summarised in Table I

Test transformer
Line opened a t X for

&CY 0%
115 kV Line

115 kV I
i
the test problem

'-1T5711hkV-1
--- To Hommond

Equivalent
source
(Summer vi \le)
tap line

I
c-
I>%
--
I
--A
~ 1 --
12 kV OH
I Feeder
I
-,-
12 kV
Coble
P
12kV
Feeder

Sub - station under

study ( Texas volley)

Figure 7. Single line schematic diagram of the substation for test problem 3

Table I. Comparison of the grid w e n t (A) obtained by different methods

S.No. Program SMECC[5] Program PAG Thapar et al[7-8] Garrett et d[91
Problem 1 2515 2527 4992 2763

Problem 2 742 745 989 438

Problem 3 62.0 62.6 68 43
902
IX CONCLUSIONS 12. C. F. Wagner and R D.Evans, Symmetrical Cornpmnb,
McGraw-Hill Book Company Inc., New York, 1933.
To obtain good estimate of dangerous voltages when a grouad 13. R. E. Scott, Linear Circuits Part I, Addison Weslqr
fault occurs at a station, it is egsential to calculate grid ament. Publishing Company, Inc., Massachusetts, USA, 1960,pp.
algorithm for CalCuIating grid current when ~0-m of -
131 135.
different lines to fault caurent are known hss been develaped.
The algorithm is easy to apply to a p a case. Its data
req- is b i t 4 to fault current contribution of lines -
(M’86 SM’88) was
comected to the station and their configurstion. The data born in Dem G b i Khan, now in
reguireanent ofthealgonhnis similarto that ofthe approlrhnate Pakistan,on F&.3,1946. He received
meth0dsByitsuselimitationsoftheseapprotcimate~~ the B. Sc.Engg., M.Sc. Engg. and
avoided. It is shown that the results obtained by the use oftbe PhD. degrees in El- Engineerhg
algorithm are comparable to those obtained with the analytical in 1968,1971aad 1991, respectively,
method developed at DIU, which requires elaborate system from Pasijab uzkrsity, chsndigarh,
data. The dgorithm is easy to program OR a computer. India.He has been teaching at Punjab
Engineermg College, Cbandigarfi since
x REFERENCES August 1%9 9nd is now an Assistant
P r o h r m the ElectricalEngineering
1 J Endrenvi. “Analvsis of Transmission Tower Potentials Depammt. Hehasbeenaconsultant
During &bnd Fhts”, IEEE Transactions on Power for design of grovnding systems for a number oflarge power
Apparatus and @stems, Vol. PAS-86, No.10, Oct. 1%7, stations.
p ~ 1274
. - 1283. He is a hkmber of The Institutionof E0,gineerS (India) and
2. S. Sebo, “Zero Sequence Current Wbution Along Life Member of Indian Society of Technical Education.
Transmission Lines,” I . Tramxiim on Power
Appardus and System, Vol. PAS-88, No.6,June 1969,
-
pp. 910 919. (M’71-SM’80) WBS
3. R. Varma and D MukWkar, ‘Ground Fault Current born in Amritsar, India on September
D i s t r i b u t i o n m Tawers
~ and Ground Wm,”E E E 28, 1938. He received the B. SC.
T ~ ‘ b m o n P t n v e r A p a t u s a m d S y s t e m s , V dPAS-
. Engg. (Hons.) and M. Sc. Engg.
98, NO.3, May/J~ne1979,p ~ 724 . - 730. degreesin e l d d e n g h x h g h m
4. A Meliopoulos, R Webb, E. Joy and S. Patel, Pan.jabU~,Chancfgar~Indiain
“Computation of Maximum Earth Current in Substation 1960 and 1%7, repedvely, and the
~IEEET~~monPawetAppararusa nd Ph. D. degree firam Montaaa State
m, VOLPAS-102,NO.9,Sept. 1983,Pp.3 131 3 139. - u n i v m , Bozemao, in 1971.
S. Analys~sTechnique for Power SuWiton Groluding In 1960 he jomed Punjab
Systemr,EiW Final Report EL-2682,Vdagnes 1 and 2, Ensineering College, Chandigarh, as
ElectricPowerResearchInstitute,Pal0 Alto, USA, October Lecturer. He retired fromthe same
1982. college in 1996 as Professor of Electrid Engin-. His
6. F.Dawalibi, M.Bouchard and D.M&edkar, “ S w e y on research interests are in electric power distribution, grounding
Power system orounding Wlgn PraCti~’’J m and~ofoptianizationtechniques.H e h a s b e e n o ~ g
T ~ C T ~I h PM Sn Appwral~and S Y H ~ SVol. , PAS- consultancy in grounding and power distriition.
99, NO.4,J u l y / A ~ g ~ -
~ t pp, 13% 1405.
1980, Dr.Arora is a Fellow ofTbe IustimbofE0,giOeers(India)
7. S. K . Madan, Current for Design @Grounding G r a in and Life Member of Indian Society of Technical Education.
Substations, M.E. Thsesis, Panjab UIliversity, Chandigerh,
Ma,1982. S. K. Soa was born on October 10,
8. B. Thapar aad S. K.Madan, “Current for Design of 1948 at Nagrota, Distt. Kangra m
Grounding Systems,” IliEE T-W on Power erstwhile Punjab, India. He obtained
Apparatus a d s y s t e m , Vol. PAS-103,Sept. 1984,pp. his B.Sc. Engg (Elect. Engg.) &om
-
2633 2637. Panjab Umversity in 1970. He joined
9. D. L. Ganett, IEEE T ~ t ~ r i aCCWUS I - Proctiwl Himrtchal Pradesb h4iltipulpose
Applications of ANSI/ IEEE Skmdzrd 80 1986, I - . Projects & Power Department (now
M * w , chapter 3, p ~23. 39,IEEE,N ~ w- YO&. Himachal Pradd State Electricity
10. ANSI 1IEIB Standard 80-1986,LEEE Guidefop. sqfkty in Board)in 1971.At present he is posted
AC Substation Grounding, IEEE,NEW York, 1986. BS an Ekecuthe EIlgineea at Shimla.
11. M.E. Van Valkenburg, Network A m & i s 3rd ed., Pteaaice
Hall of IndiaPvt.Ltd.,New Delhi, 1984.
\