IEEE Transactions on Power Apparatus and Systems, Vol.PAS-98, No.

5 Sept/Oct 1979 1659

PARAMETRIC ANALYSIS OF GROUNDING GRIDS

F. Dawalibi, Member IEEE D. Mukhedkar, Senior Member IEEE

Montel Inc. - Sprecher & Schuh Ecole Polytechnique
Montreal, Quebec, Canada Montreal, Quebec, Canada

ABSTRACT The analysis presented in this paper, concerns

square and rectangular grids having equally spaced con-
The results of an extensive study conducted using ductors. In practice however, unequally-spaced conduc-
a computer program designed to determine grounding per- tors often provide superior grounding performance [1,
formance in two-layer soils [1,2,3,5] are described and 10]. Despite this fact, equal spacing is considered in
discussed. A variety of grounding grid configurations order to:
and two-layer soil conditions are analysed in detail.
The calculated grounding resistances, step and touch a- Decrease the number of design parameters which
potentials are summarized in several charts which must be varied.
could be used conveniently for practical design purpo- b- Provide reference results which can be compared
ses. The results obtained prove that in general, con- with existing published data (usually for uni-
ventional methods of analysis, such as the IEEE 80 me- form soils and equally spaced grid conductors).
thod [4] which uses "Km & Ki" irregularity factors,
fail to predict accurately grounding grid performance. Computer program "MALT" was used to obtain the re-
Also, it is shown that there are certain grid configu- sults of this paper. This program is based on theore-
rations and/or burial depth, which lead to optimum tical studies already published [1,2,3], and is capable
touch or mesh potentials. Finally, it is shown that of handling grounding systems having an arbitrary geo-
certain two-layer conditions will lead to considerably metrical configuration and buried in a two-layer soil.
higher touch (or mesh) potentials than the other compa- Assumption of a two-layer structure is the minimum re-
rable alternatives. quirement for accurate modelling of a grounding grid in
non-uniform soil. A uniform soil equivalent cannot ade-
1.0 INTRODUCTION quately represent the conditions of non-uniform soil,
since the ground resistance is a function of the top
The basic design quantities of a grounding grid and deep soil layers, while touch potential is, for all
are the resistance (or potential rise, which is the practical purposes, a function of the top soil. The
product of the resistance by the grid current), touch adequacy of a two-layer soil structure as an equiva-
and step voltages. In a uniform soil, the resistance lent model for more complex type of soils is discussed
can be calculated with an acceptable accuracy using se- in a companion paper [11].
veral simplifying assumptions [4,6,7]. Touch and step
voltages are not easy to calculate and are usually de- 2.0 CASES ANALYZED
termined by analytical expressions [4], adjusted by ap-
plying correction factors. These factors are also Square (30m x 30m), or rectangular (40m x 20m)
based on experimental results conducted in a uniform grids with identical length of conductors, having one to
soil model [8]. 144 meshes, and few other less conventional grid types,
are analyzed in detail. The basic grid dimensions, men-
Recently, doubts have been raised about the accu- tioned above, are kept constant throughout the study.
racy of the IEEE 80 proposed analytical expressions Table 2.1 further describes the individual grounding
[1,9]. The measurements in [8] have been performed grid arrangements. The following range of design para-
with grid models buried at a fixed depth which does meters is considered:
not correspond to actual depths as used in practice.
Note that an IEEE working group to revise guide IEEE 2.1 conductor radius "r", from 5mm (0.005m) to 250mm
80 is investigating also this particular problem. (0.25m.);
A recent paper [9] have proposed new expressions 2.2 depth of burial "z", from 0.lm to 100m;
for the irregularity correction factors "Kmi" and
"Ksi" which should prove useful for grounding design 2.3 total number of meshes "n", from 1 to 144;
in uniform soils. Unfortunately, when soil is not
uniform (almost all grounding systems are installed in For a base case, the parameters r=0.01m and z=0.5m, are
non-uniform soils), the previous expressions for the used.
correction factors are inadequate and can not be used
to predict step and touch potentials. This is shown 2.4 two basic soil structures:
in detail later in this paper. Also, the grounding a) uniform soil and,
resistance is shown to vary considerably with the de- b) two-layer soil.
gree of heterogeneity of the soil.
(Top layer resistivity Pi kept constant and equal
to 100Q-m);
A reflexion factor "K", defined by the following
expression was used:
K = (P2 P1)/(P2+Pl)
F 79 243-7 A paper recommended and approved by the where: P1i Top layer resistivity
IEEE Substations Committee of the IEEE Power Engi- P2 Bottom layer resistivity
neering Society for presentation at the IEEE PES Winter "K" can vary theoretically between -1 and +1. The
Meeting, New York, NY, February 4-9, 1979. value of K=-l corresponds to P2=0 and K=+l corresponds
Manuscript submitted August 28, 1978; made available to P2=00. As shown in Table 2.1, "K" varies between -0.9
for printing December 14, 1978.
and +0.9.
0018-9510/79/0900-1659$00.75 C91979 IEEE
1660
2.5 top layer height "h" from 0.lm to 300m. 2.6 grid current "I" of 10,OOOA is assumed to flow in
the grid.

Identification GID Dienion Total length, Number of Radius "r" Bur1al depth "e" First layer Refle.xIo
Symbo I RID (ners.) of eduors meshes "n" ( eters) (Ieters) height "h" factor

0.005 0.1 0.1 -0.9

S9ii 30x30 120 1 to toS to
0.25 100 0.5 +0.9

0.1 0.0 -0.9

4 330-30 180 4 0.01 to to 0t
100 300 +0.9

S64 _ S _ 30+30 240 9 0.01 0.5 - 0.0

0.1 0.0 -0.9

S16 ~~~~~~3030 300 16 0.00 to to to
100 300 +0.9

S25 30+30 360 23 0.01 0.5 - 0.0

S36 30+30 420 36 0.01 0.3 - 0.0

064 30.30 540 64 0.01 0.5 0.0

S144 30-30 780 044 0.01 0.3 - 0.0

EiIiIiiil 40-20 120 1 0.01 0.3 0.0

R4 40-20 180 4 0.01 0.5 - 0.0

R9 E 40 20 240 9 0.01 0.5 - 0.0

S1P9 s H 30-30 240 - 0.01 0.5 - 0.0

0016 30 30 S.5., 285 - 0.01 0.S -

0.0

RP9 40a20 240 0.01 0.5 0.0

Table 2.1

3.0 RESULTS presented in Figures 3.1 (one mesh grid), 3.2 (four
mesh grid) and 3.3 (sixteen mesh grid).
The following results were obtained for each case
analyzed:
a- Current density in grid conductors (A/m). (Fi-
gures 3.1, 3.2 and 3.3). 40

K= -0.9
b- Earth surface potentials along various direc- .1.
tions over the grounding grid. (Figures 3.4 to
3.7). 30

E
c- Grounding resistance (or potential rise) of z _Z= 0. 5m
grounding grid. (Figures 3.8, 3.9 and 3.10). Soil ; I ('h=5m
20 C, (02-PI)
K-
d- Step and touch potentials. (Figures 3.11 to
3.17). Grid
10
z
X
(total curre_ L Current
densit
4xlOOOA) 'o
It is not possible to present all the results ob-
profile

tained for each case analyzed. However a number of ty-

pical cases have been selected and are presented in de- LOCATION ON CONDUCTOR IN METERS
I
tail. 5 10 15
.

20 25 30)

3.1 Current Density Curves

The current density function (current per meter of Figure 3.1

conductor flowing into earth) along one or two conduc-
tors of a grounding grid buried in a two-layer soil is One-Mesh Grid Current Density
 1661
In contrast, when the bottom soil layer is more conduc-
tive than the top soil layer (K<O), current density is
80 larger at the center of the grid. This result can be
explained physically by the fact that current will go
downwards into the bottom low resistivity layer when
K<O, and sidewise when the top layer is more conductive
60
that the bottom one.

40
K= -0.9
h= 2m
c Grid potential -*ss. -11 _- X ax i.$
22 rise P'rof iles
20 / ...,M.-I.. S9-2
S9-1I
he~~~~
.r - -,
20 r-
m
S16-2
016-1
!
18 3f~ 9,S25-2
S025-1
16 S16-2
----------------~~ S16-1 g( S25-2 /
Figure 3.2 14

Four-mesh grid current density 12 SS

10 '''- 09-1
8 Si
Grid current IOKA
6 Uniform soil K-O.; p= lOOOQ-m__
4

POTENTIAl. I(liNT X 1N METERS

-18 -14 -10 -6 -2 2 6 10 14 1R8

Figure 3.4
Earth surface potential profiles

11 ro f ilIe s
Figure 3.3
11l
-1. S:st-2
Sixteen-mesh grid current density --
~S16-1
v.
I.-
The current density is shown for
various soil
structures. The current density curves are based on a
20 -I
IlI
-
S64-2
total grid current of lOOOp amperes, where p is the El -.- c
5 Lb_?
Grid potential
total number of conductors in the grid. Therefore, if rise S144-1
the current density is uniform, each grid conductor 16
would carry lOOOA/30m = 33.333 A/m. Figures 3.1 to 3.3 _'_- '4'-''_-.':''.t_'' '_' ''-_'': E S64-2 S144-2
show that the current density function is not uniform
(except for the uniform soil case K = 0.0 and one 12
mesh grid), but varies according to a complex mathema- S36-2 S36144-1 S64-1

tical expression, which is not only function of the Si 36-1

grid arrangement and location of the conductor conside-

red, but also function of the soil structure (h and K). 8
Grid current 1OKA
Uniform soil.p 101o1-m
The almost uniform current density in the case of Z- 0.5m
one mesh grid and uniform soil, is a logical result sin-
4
ce, the grid shape is very close to a circular ring.

Further analyzis of Figures 3.1 to 3.3 show an in- POTENTIAL POINT X IN METERS
teresting result not previously reported in the litera- . I . . I .

ture: -18 -14 -10 -6 -2 2 6 10 14 18

It has been
usually accepted that current densi-
ties larger towards the edge of a grounding grid or
are
electrode, than at the center. This is not always true
as shown in the previous figures. When soil is uniform Figure 3.5
or when the bottom soil layer has a resistivity value
larger than the top layer resistivity (K 2 0), the cur- Earth surface potential profiles
rent density is effectively larger at the grid edges.
1662
3.2 Earth Surface Potentials tivity. This influence is more pronounced with a high
resistivity deep soil than in the alternative. However
The earth surface potentials along various direc- at depths of approximately twice or more grid dimension
tions parallel to the grid conductors, are shown in Fi- the deep soil effect may be neglected without affecting
gures 3.4 and 3.5 for various square grids (1 mesh grid
to 144 mesh grid) and in Figures 3.6 and 3.7 for the 1
and 16 mesh grids buried at various depths. In these
figures the grid potential rise is also shown (as a
straight line). Grid potential O
-Potential
rise -prof i *-
Figure 3.4 and Figure 3.5 show that the effect /- (O. Im) Z= OlIm
Z= O. 25m
of increasing the number of conductors (or meshes) de- ........ (O. 25m) Z= 0.5m
u-: Z= l.Om
creases the grid potential rise (by decreasing the grid c -------- ----(O. 5m)
-
Z= 2.55m
resistance) and increases the minimum earth surface po- 20

tential above the grid. As a result the touch poten- (2.5m)

tials are also decreased. It is however worthwhile no- (5m)
Grid current
ting that the potential rise decrease, the earth surfa- 16 (lOm) Uniform soil.p=lOOMi-,
= lOKA

ce potential increase and the corresponding touch po- (20m) variable grid deEth Z
tential decrease, diminish progressively and are almost (50m) T

negligible when the number of meshes approaches 144.

It can be noticed also that the location of worst touch 12
potential, moves regularly towards the edge of the grid / _ x= = < < _~~~~~~~
,Z2. 5;'
when the mesh number is increased. These results howe-
ver, apply to uniform soils (case of the figures) or Z= 5m
8
two-layer soils with positive reflexion factors K and
equally spaced grid conductors. When K is negative the Z= lOm
highest touch potential location is not easily predic-
table (see Section 3.4) specially if the grid conduc- 4 Z= 29?m
Z= 50m
tors are not equally spaced.
Z= lOOm
Figures 3.6 and 3.7 show the effect of grid depth POTENTIAI. POINT X - (METERS)
. ,. I . .
on both potential rise and earth surface potentials. I

There,an increase of the grid depth results in a simul-

-18 -14 -10 -6 -2 2 6 10 14 18
taneous decrease of the potential rise (or grid resis-
tance) and the earth surface potentials. Consequently
the touch potential will decrease or increase depending
on which of the potential rise or earth surface poten- Figure 3.6
tial, decreases at the highest rate. With a moderate
increase of the depth (say 0.1 to 0.5m), the minimum Earth surface potential profiles
earth surface potential does not vary significantly. In
contrast the potential rise decreases considerably. As
a result the touch potential decreases with a moderate
increase of the grid depth. When the grid is buried at
a large depth (say 50m or up) a further increase of the
depth does not decrease the potential rise significan-
tly (in theory, the resistance of a grounding grid is
reduced by half when it is moved from earth surface to
ll) tvnt i;al
infinite depth-uniform soil). However the earth surfa- EFS
ti -> ro f l e-
ce potential drops sensibly (at infinite depth, earth
-15 *15

surface potential vanishes). The corresponding highest Grid curprelt - lIKA

Uniform soil -n- llOiLi-m
touch potential is now increasing. 201 Variable grid deptll Z
Grid potential rise
The above facts suggest that there is an optimum z
grid depth which leads to the minimum touch potential. (O. 1)
This prediction is confirmed in Section 3.4, Figure 16

3.13.
Finally, although not shown in this paper, the re- 12
sults of the study have indicated that the effect of
grid conductor radius on earth surface potentials, is
negligible (at least, when the radius varies from 005m 15>l00) 250

to 0.25m). However the resistance decreases moderately

10
when the conductor radius is increased (a 25% resistan-
ce decrease when radius is increased from 0.005m to
0.25m).
3.3 -
Grounding Resistance 51)
.1
.._.. _ . _ .. ... ... . t _. . ..._ _ ._ ...... _..

100 POTENTIAL POINT X (METERS)

Figures 3.8 and 3.9 can be used to determine the - ..IN.-.1.4
1.0 2 2 6 1 1

ground resistance of 4 or 16 mesh grids buried in any -18 -14 -lO -6 -2 2 6 10 14 18

two-layer soil structure. The numerical values shown
are valid for a 30 x 30m grid buried at 0.5m from earth
surface (conductor radius = 0.Olm). However other grid
dimensions can be analyzed as explained in Appendix A. Figure 3.7

The figures confirm that ground resistance is Earth surface potential profiles
still considerably influenced by the deep soil resis-
 1663

significantly the accuracy of the results. The same Uniform soil

K= 0.0
figures show clearly that when the grid location chan- 2.0 = 100Q-m
ges from the top layer to the bottom layer (h=z0.5m),
the grid resistance variation is extremely abrupt when 1 .8 (one
~~~~~^_*_< ~~~shown only)
point of curve is

K approaches 1.0 or -1.0.

1.6

100 0 ' _ :
=tr ~~~~~~~~~~~g3S4
S

0.8 _
Depth of grid
' K Z= 0. 5m 0.4 Base case
grid depth
la~~~~~~~ 90 Z= 0 .5m GRID DEPTH Z -IN METERS
*aI,1 d a*I *-
**.

Figure 3.10
0.0 Resistance versus grid depth
-0.5
3.4 Step and Touch Potentials
1 -
3.4.1 Touch Potentials
Figures 3.11 and 3.12, 3.14 and 3.15 give respec-
tively, the maximum touch and step potentials occuring
I 10 100 in a 4 or 16 mesh grid buried in a two-layer soil. The
TOP LAYEIR HEICHT (METERS)
touch and step potentials are expressed in percent of
the corresponding grid potential rise. If an actual
value is desired, Figures 3.8 and 3.9 can be used to
determine the grid potential rise, which is the product
of the resistance by the grid current. The grids di-
Figure 3.8 mensions are the same as in Section 3.3 (base values).
Other grid dimensions can be analysed as explained in
Four - mesh grid resistance Appendix A. The location of maximum touch potentials
occurred always at the corner meshes except for S16
with negative K values in excess of -0.8 and top layer
heights varying from approximately 0.8m to l0.m. In
such cases, the maximum touch potential occurred at the
center meshes close to the grid edges.

Figure 3.11 and Figure 3.12 shosf that the touch

potentials may be almost anywhere between 0% and 100%,
when soil is not uniform. Consequently, the use of
IEEE guide 80 [4] or Koch's results [8] which give the
mesh potentials in percent of grid potential rise (in
uniform soil ) will lead to wrong solutions if the soil
considered is not uniform. Since the "Kmi" factor was
derived from Koch's experiments,the use of "Kmi" should
also be avoided for non uniform soils. Although the
-

touch potential expressed in percent of the potential

rise varies widely with soil structure
, its value in ,

volts fluctuates much less For example, with a one

.

mesh grid type Sl buried in a two-layer soil (pl=100Q-m

and h=2.0m) the touch potential Vt is a function of the
reflexion factor K and varies as shown in Table 3.1:
(grid current = lOkA).

100

TOP LAYER HEIGHT (METERS)

K -0.98 -0.9 -0.8 -0.7 -0.5 -0.3 -0.1 0.0

Figure 3.9 Vt in % 98.9 95.4 90.1 85.4 76.5 67.8 58.3 55.8
Sixteen-mesh grid resistance Vt in kV 8.98 9.19 9.43 9.52 9.97 10.49 10.82 11.41

K 0.98 0.9 0.8 0.7 0.5 0.3 0.1 0.0

Figure 3.10 gives the resistances of various grids Vt in % 3 8.2 10.1 20.3 33.5 42.1 50.4 55.8
buried at various depths in a uniform soil. At an in-
finite depth the grid resistance will reach half the Vt in kV 5.90 8.41 10.71 11.13 12.27 12.02 11.45 11.41
resistance value corresponding to zero depth. The rec-
tangular grids RI and R4 have the same total length of
conductors as their counterparts S1 and S4, but their TABLE 3.1
resistances are slightly larger.
 ~ 0.9
1664
2
The touch potentials in kV vary between 5.9 and CZ Base case
Grid depth
12.3 kV only Table 3.1 draws attention to an impor-
.
-C
11-
H Z= 0.5m
tant conclusion: z
w
1- Uniform soil
0 K= 0.0
10C Cl.
Cl p= 100Q-m
The maximum touch potential obtained with a grid ,
u
buried in a variable two-layer soil, does not occur in 8C
4.
0
the case of the highest or lowest deep layer resistivi- t'
ty (K = -0.98) or when soil is uniform (K=0.0) but,sur- -3

prisingly, this maximum value occurs when the bottom 60 z

w
soil layer has a resistivity value moderately larger 8
A.

than the top layer (K = 0.5 for the case of Table 3.1). 40 U
:D
0

in cn
20
lx
u
LI
WES4 ix
.0
:x
0 GRID DEPTH "Z" IN METERS
100 Grid depth
t K= -0.9 Z= 0.5m
.4 0.01 0.1
c 1.0 10 100
1-
z
80 w
1-
0 la
Cl. U)
Figure 3.13
=M.
u
Touch potential versus grid depth
-D .4
60 a -.c
1-
E- z
w'A.I.-
0 0
w 3.4.2 Step Potentials
40 :R CL.

Figures 3.14 , 3.15 and 3.16 give step potentials

20 in percent of grid potential rise and are similar to
Figures 3.11 , 3.12 and 3.13 for touch potentials. Si-
milar conclusions and discussions can be derived by
detailed analysis of the figures. However, three par-
ticular points deserves some additional comments:
Figure 3.11- Four-mesh grid touch potential
a- First the step potentials values are signifi-
cantly lower that touch potentials. Since the
zI HH
r-T,,
S16
maximum allowable step potentials [4] are al-
0. ready larger than the allowable touch poten-
Grid deptb
ZO 0.5m tials , it may be concluded that , except for
;
some unusual cases , once a grid design is dec-
lared safe for touch or mesh potentials it will
generally be safe for step potentials.

S4
30 X
c
Grid depth
-
Z - 0.5m
-0. 5 >
~~~~~~~~~K=
0.9 _ ----- O-j.:;,
---

SDP LAYER HEIGHT

"l"IN METERS
20

-I 11, I*I .1.

0.1 10 100
15
- s c { -~~~~~~~~~0.
5

Figure 3.12- Sixteen-mesh grid touch potential 10

Figure 3.13 shows the influence of grid burial

depth on touch potentials for various grids (Sl,S4,S16, TOP LAYER HEIGHT "h" IN METFRS
Rl and R4). 0.01 0.1 1.0 1in OS 150

As already mentionned in Section 3.2, the increase

of grid depth causes , first a moderate decrease of the Figure 3.14
touch potential which reaches a minimum value at a fi- Four-mesh grid step potential
nite optimum grid depth ( this optimum depth decreases
with an increase of the meshes ) , before starting to b- As in the case of resistance and touch poten-
rise again. Figure 3.2 suggests that, with large num- tials , there is an abrupt rise or dip of the
ber of meshes, grids buried at depths within 0.5 meters step potential when the top layer height "h"
will be subject to the lowest touch potentials. This equals the grid burial depth "z" (h =z=0.5m).
conclusion however, valid for uniform soils, may not be However , in the case of step potentials, this
true for non-uniform soils. Additional two-layer soil abrupt change (rise or dip) is immediately fol-
studies are required before extending this conclusion. lowed by another moderate change in the reverse
direction. This phenomena, did not occur in the
It is interesting to note that the rectangular resistance or touch potential cases. The au-
grids Rl and R4 have a better touch potential perfor- thors have not yet a satisfactory explanation
mance than their square counterparts S1 and S4. Note of this curious behaviour.
that the actual touch values in volts for Rl and R4 are
obtained using Figures 3.10 and 3.13 . Despite their c- The step potential decreases continuously with
slightly higher resistance , Rl and R4 still have lower the grid burial depth "z" (Figure 3.16) and be-
touch potentials in Volts. comes negligible when z is approximately 1/10
the grid dimensions.
 1665

30
Similar conclusions apply for step potentials. The
above facts are considered by the authors as solid ar-
guments against the use of irregularity grid factor
25 "Kmi" and "Ksi". The authors believe that only direct
analytical solutions are suitable in practice where,
20 irregular grids with unequally spaced conductors are
buried in non-uniform soils.
15
2
10 w

F--,
z
w
5
0
a

0
C4 4. t
0
E-
u
;5
Figure 3.15 3- z
w
Sixteen-mesh grid step potential m
F-
R
I u
u
w 2- :D
c
ce F-
2. VI
tz
I I 0
3:

W~04S Uniform soil

K= 0.0
- S16 p= 100Q-m

Figure 3.17
Touch potential, resistance and Kmi irregularity
correction factor

Base case
grid depth DEPTH "Z" IN METERS
4.0 UNCOMMON GRID FORMS
RID
Z= 0.5m 4
-t ...
. .............
..

0.01 0.1 1.0 10 100

Symetrical grids composed of square or rectangular
Figure 3.16 meshes of uniform size are not always as efficient as
Step potential versus grid depth other grid forms having less frequent cross connections.
Some cross connections , however, are desirable to pro-
3.4.3 Irregularity Factors Kmi and Kms vide multiple paths as further assurance of continuity
in case of mechanical damage to a grid conductor.
Figure 3.17 gives the maximum touch potential "Vt"
in percent of grid potential rise ) as a function of Table 4.1 gives the resistance, touch and step po-
the grid number of meshes "n" (uniform soil). The grid tentials computed ( uniform soil ) for various uncommon
resistance "R" is also shown in this figure. Using this grids (RP9, SP9) and their more conventional equivalent
figure,it is possible to calculate the "Kmi" correction square or rectangular grids (R9 , S9) and (S16). The
factor using the following relation [91: comparison is based on equal total length of buried
conductor "L". See also Table 2.1.
L V % R
K (.1 Grid L Resistance Touch Touch Step Step
100 p ....
=
mi K
m
...(3.1) Type (m) R(Q) Vt(%) Vt(V) Vs(%) Vs(V)
The values obtained using (3.1) are also shown in S9 240 1.6601 28.7 4759 10.3 1715
Figure 3.17 When the mesh number "n" is not large,
.

the authors results are comparable with IEEE 80 equa- R9 240 1.6896 26.8 4524 11.1 1874
tion (17) [41 or Nahman and al [91 equation (21). (See
Figure 3.17). SP9 240 1.6412 25.4 4169 10.9 1784
With higher values of "n" Nahman and IEEE 80 "Kmi" RP9 240 1.6708 23.5 3932 11.5 1925
values are respectively higher and lower than the au- Grid L Resistance Touch Touch
thors values. Step Step
Type (m) R(Q) Vt(%) Vt(V) Vs(%) Vs(V)
Figures 3.11 3.12 3.13 and 3.17 show that the
, , S16 300 1.5933 22.6 3596 12.2 1947
maximum touch potential is not only a function of the
grid number of meshes "n" but is also a complex func-
tion of: SD46 293 1.5874 21.4 3400 12.3 1948

a- The grid burial depth "z" (Figure 3.13) and,

when soil is not uniform (two-layer soil): SDm 285 1.5969 20.8 3316 11.8 1878

b- The top layer height "h". =5D6

SD6n 277 1.6109 22.8 3671 11.4 1838
c- The reflexion factor "K".
Table 4.1
1666
From Table 4.1 it may be concluded that: APPENDIX A
a- Step potentials are consistantly lower than The results of this study are applicable to any
touch potentials. Therefore, the optimum grid grid which is similar to those presented in this paper
with respect to step potentials, is not usually (Reference grids).
of interest for power substation grounding de-
sign. The only restriction is that all grid dimensions,
including conductor radius,must be proportional to the
b- Grids composed of parallel conductors are more reference grid.
effective (with respect to resistance and touch Assume that a grid "a" times larger than the refe-
potential ) than the equivalent grids composed rence grid , is buried at a depth "z" in a two-layer
of meshes. Rectangular grids are more effec- soil having a top layer resistivity "p" and height "h"
tive than square grids with respect to touch and a reflexion factor K. In order to use the paper
potential. However the rectangular grid resis- charts it is necessary to:
tance is slightly higher than the equivalent
square grid (=2%). 1- Select a modified height hm=h/a
2- Select a modified depth Zm=Z/a
c- The diagonal square grid type SD16, despite the
fact that the total length of buried conductor 3- Read, knowing the reflexion factor K, the desi-
is less than square grid SD16 (300m), is still red value Xm.
more efficient with respect to resistance, step - If Xm is a percentage, then the actual value X=Xm
and touch potentials ( when S is selected ade-
quately). Grid SD16 performance can be optimi- - If Xm is a resistance , then the actual value X=
zed by selecting and adequate spacing "S". XmP/pla. Where p1 is the top layer resistivity as-
sumed in the paper (lOOQ-m).
- If Xm is a potential , then the actual value X =
CONCLUSIONS Xm pI/plaIl . where I and I1 are respectively the
actual and chart assumed grid currents.
The analysis of the computer study results has led
to the following conclusions and achievements: REFERENCES
1- F. Dawalibi, D. Mukhedkar, " Optimum design of sub-
a- Effect of the various parameters which influen- station grounding in two layer earth structure",
ce grounding grid performance in uniform and part I, II, & III, IEEE Transactions, Vol. PAS -94,
two - layer soils. Briefly, it was shown that No.2, March/April 1975, pp. 252-272.
soil characteristics ( p, h, K ) and grounding
grid configuration and burial depth influence 2- F. Dawalibi, D. Mukhedkar, "Multi-step analysis of
significantly conductor current density, grid interconnected grounding electrodes" IEEE Transac-
resistance, step and touch potentials. tions, Vol. PAS -95, No.1, January/February 1976,
pp. 113-119.
b- For uniform soils, the authors have obtained an 3- F. Dawalibi, D. Mukhedkar, " Transferred earth po-
irregularity correction factor "Kmi" which di- tentials in Power systems",IEEE Transactions, Vol.
ffers from the ones already published [4,9]es- PAS - 97, No.1, January/February 1978, pp. 90 - 101
pecially when the number of meshes is large.
4- "Guide for safety in alternating current substa-
c- It was shown that: tions grounding, IEEE Standard #80, 1961.
5- F. Dawalibi, D. Mukhedkar, "Resistance calculations
- When soil is not uniform, for a given grid of interconnected grounding electrodes", IEEE Tran-
design , there is a soil structure which sactions, Vol. PAS 96, No.1, January/February,1977.
leads to the highest touch potential.
- Grids composed of parallel conductors are 6- P.G. Laurent , "Les bases generales de la technique
more effective than grids composed of meshes des mises a la terre dans les installations elec-
and that rectangular grids are more effective triques " , Bulletin de la Societe Francaise des
than square grids. Electriciens, 7 iame serie, Tome 1, No.7, july 1951
- When soil is uniform there is an optimum grid pp. 368-402.
burial depth which leads to the minimum touch 7- S.J. Schwartz, " Analytical expression for the re-
potential. sistance of grounding system" , AIEE Transactions,
- Adequate location of grid conductors results Vol. 73, part III-B, 1954, pp. 1011-1016.
in optimized grid performance.
8- W. Koch, " Erdungsmassnahmen Fur Hochstspannungsan-
d- Finally several charts , valid for uniform and lagen mit geerdetem sternpunkt", Elektrotechnische
two - layer soils are provided and may be used Zeitschrift , Vol. . 71 ,February 1950, pp. 89-91.
advantageously for design purposes. 9- J. Nahman , S. Skuletich , "Irregularity correction
factors for mesh and step voltages of grounding
grids " , IEEE Transactions paper F 78 731-2, Los-
Angeles., July 1978.
10- J. Sverak , " Optimized grounding grid design using
variable spacing technique " , IEEE Transactions,
Vol. PAS 95, January/February 1976.
ACKNOWLEDGEMENTS
The authors express their appreciation to the Na-
tional Research Council of Canada for providing the 11- F. Dawalibi, D. Mukhedkar,"Influence of ground rods
necessary financial support of the work. The authors on grounding grids". Companion paper submitted for
also acknowledge the cooperation and facilities offe- presentation at the IEEE PES winter meeting, New-
red by Ecole Polytechnique and Montel Inc. York, 1978.
 1:f_.,t ;
1667
Farid Dawalibi (M'72) was born in Lebanon, in dicate that for a crushed stone surface over a low resistivity soil, in-
1947. He obtained his Bachelors' Degree in Elec- creasing the depth of the crushed stone will increase the step potential
trical Engineering from Ecole Superieure d'In- voltage. Does this indicate a safe limit to the depth of the crushed
genieurs, Universite de Lyon, Beirut, in 1971. In stone?
1972, he obtained his M.S. Degree from Ecole The Authors are to be commended for their work and analysis of
Polytechnique, Montreal University. From 1972 ground grids in two-layer earth conditions.
^ to 1977 Mr. Dawalibi was employed by
Shawinigan Engineering Co. Ltd., where he Manuscript received February 15, 1979.
worked on various major power system projects.
In 1977 he obtained his Ph.D. degree from Ecole
Polytechnique. He then joined Montel-Sprecher P. Kouteynikoff (Electricite de France, Clamart, France): The
& Schuh as manager of technical services department. In 1979 Mr. authors should be complimented for the numerous, original and in-
Dawalibi founded Safe Electrical Systems dm ltd, a consulting firm teresting results they have presented. The present comment deals only
specializing in power system safety and electrical interference. with mesh voltages and irregularity factor K,,. In the authors' paper
Mr. Dawalibi is an IEEE member and a registered professional figure 3.17 gives calculated values of earth resistance, touch voltage and
engineer in the province of Quebec. factor K.,i, as a function of the number of meshes. For K,,,1, the authors
compare their results with those given by the IEEE formula and by
D. Mukhedkar, for photograph and biography please see page 728 of Nahman and Skuletich [1] equation (21). It should be pointed out that
the May/June 1979 issue of this TRANSACTIONS. this equation was obtained through computer-aided calculations and
was said to be a very good approximation by Nahman and Skuletich.
We also carried out computations [2], considering the same grounding
arrangements. For better clarity, only values of Kmi are given, since it is
the most sensitive parameter. s, number of divisions, equals the square
Discussion root of number of meshes, or n - I (using guide 80 definition of n). We
obtained a good agreement with Nahman's results for s ranging from I
C. J. Blattner (Niagara Mohawk Power Corp., Syracuse, NY): The to 10. It should be reminded that Nahman and Skuletich computed K,,,
Authors have provided a valuable analysis of grounding grid perfor- for s ranging only from I to 8 (n from 2 to 9) and that we used their
mance under fault conditions. The effect of the two-layer soil condi- equation (21) to plot the curve.
tions on the grid performance is particularly significant. The Authors
have also presented their data in a practical form for use in the design of I
: :

a ground grid.
The results published in this paper have been used to analyze a ;o
4, i~~~I
substation grounding problem we recently encountered. The substation
site soil condition consisted of a two-meter thick layer of soil over solid
rock. During the substation construction phase, most of the top soil
was removed and, unfortunately, replaced with gravel backfill. As a
result, the initial grid design was no longer valid, and the grid had to be
redesigned. The field test results, obtained during the redesign phase,
-.- NmhmcL n
;, .et o
i1 ,
-

are in close agreement with the predicted values of ground resistance

obtained using the data provided in this paper. _/ T-I---:1
Based on our original grid design, we anticipated a final ground -T--. -'-f-4:- --- ! ;-
grid resistance of 0.7 ohms. Using the data provided in this paper, the
final ground grid resistance was calculated to be 1.4 ohms. Due to the
field construction problem, no actual corresponding test values are
01. a At :. ii

available; however, the difference in calculated values is significant. - - -4+ - -g -i

Field ground grid resistance measurements, made with the gravel _-
__.,_
.t !--
.f. :A,'-<-
+D
backfill in place, resulted in a test value of 6.3 ohms. Calculations in-
dicated that the resistivity of the soil in the yard area was now approx.
600 ohmmeters, or about ten times the original soil resistivity. Using 5
XJ-:-1--- -- ----_i] .f :

this value of resistivity, and the data in this paper, a ground grid
resistance value of 6.7 ohms was calculated. This is in close agreement
with the 6.3 ohm field measurement. 7. .- -. __ --- t-
!o:r .j. .
-4i- ~I-44-
We next planned to add four 20-meter deep ground wells in an at- .~ -... -V. _- ,_/

'a.L lors*~
- -I
-, .
-4-.-
.
--
tempt to obtain the desired ground resistance. Based on the original /
field test data, a resistance of 3.5 ohms per ground well was anticipated.
Based on the Authors' companion paper on ground rods, a resistance -,
value of 17 ohms can be expected. Actual field test measurements of 14
:
-!--1-/
.~~~~~~~~~~~~~~~~~~~~~~~~~~~~
-4-:i-, .- 1.~
.,_ -C.,i IKi
.

ohms were obtained, which is in reasonably close agreement to the value X _

predicted by the Authors' method. IEEE
Calculations, based on the actual ground grid arrangement finally - --I
installed, and the data in this paper, result in a predicted ground
resistance of 2.2 ohms. The actual final ground resistance obtained was -4----
2.0 ohms.
There are three questions raised by this paper that we would ap- ::
i--_- _
-
I
,_
- f: -
.KI K
-
:. .i ,.I. rx
preciate the Authors' comments on: i :T I ~.,
1. The Authors note that the maximum touch potential occurs for
-
-. _j -

i1
:

,
K = 0.5. Is this condition related to the change in current density be-
,
j j I_s"

tween the center conductors and outside conductors as described in Sec- 0 5 10 $ ,,u.vnbro
tion 3.1 ?
2. Figures 3.11 and 3.12 indicate that for h<z, Vt(%) is constant These discrepancies between three sets of computed values are
for all values of K. Figures 3.8 and 3.9 indicate Rg remains constant for believed to be caused by the different simplifications that no method
the same condition. This appears to indicate that the grid potential and can avoid (uniform leakage current distribution along each segment,
the mesh potential voltage remain constant. However, for the same con- etc). Further discussion is required, but it would be too lengthy to be
dition, the step potential Vs(%) is not constant, indicating the step developed here.
potential voltage varies. Can the Authors explain the reason for this? However, even if not perfect, computer calculations are the only
3. Figure 3.15, for values of K = -0.5 and -0.9, indicates that ones suitable for practical use, especially because grids are not uniform-
Vs(%) increases as h increases from h = o to h = z. The ground ly spaced. So we fully agree with the authors' conclusions that use of
resistance remains constant for the same condition. This appears to in- factors Ki,, and K,, should be avoided.
1668
REFERENCES a- When the grid is buried in a uniform soil (bottom layer) at a
small depth (smaller than 0.5m as compared to its dimension, its
[11 J. Nahman, S. Skuletich-Irregularity correction factors for mesh resistance is not altered very much if a thin layer is added on the top of
and step voltages of grounding grids. Paper F78 731-2 presented the soil (top layer). See figure 3.10.
during the IEEE PES Summer Meeting, Los Angeles (1978). b- Touch potential is calculated at the centre of a mesh about 4 to 8
[21 P. Kouteynikoff-Numerical computation of the grounding meters from the closest conductor (point M see figure Cl). The major
resistance of substations and towers. Submitted for presentation at voltage drop due to current flowing in the soil is in the bottom layer.
the IEEE PES Summer Meeting 1979, in Vancouver.
Manuscript received March 5, 1979.
---Nk -. ---_ - -i b $Thin layer
F. Dawalibi and D. Mukhedkar: We would like to thank the discussers
for their interest in the paper and their pertinent comments and ques- Mesh dimension
tions. 7.5 or 15 m Mes
Mr. Kouteynikoff is right when he says that further discussion and (grids S16 or S4) conductor
work are required to determine exactly the reasons for the discrepancies
between three computed curves. We believe however that one important Fig. Cl- Touch Potential
factor is the nature and number of subdivisions of the grid conductors
especially when the grid has a large number of meshes. In this case the
Kmi parameter becomes very sensitive to the number of subdivisions
"n". In our opinion, in order to determine accurately the Kmi
parameter, it is preferable to subdivide the conductors unequally; The
extremity segments being smaller than the centre ones. This techniques
however is suitable only for uniform soils.
Again we fully agree with the discusser that use of Kmi and Ksi fac- I IW II ml I I
III U III Ih I 11 1Thin layer
tors should be avoided, when computer methods are feasible. ~
-^tyzz5 O.5Sm x
We greatly appreciate Mr. Blattner work which compares actual MOL it- uter loop
field results with the predicted values as computed by our methods. We lm conductor
are particularly pleased and encouraged by the results of Mr. Blattner
comparisons and we hope that more engineers will also be encouraged Fig. C2- Step Potential
and will use accurate computer methods to design their grounding
systems.
We will answer Mr. Blattner questions in the sequence they are
raised. The influence of the thin top layer is thus negligible.
[1] It is difficult to explain the reason why maximum touch poten- c- The previous paragraph does not hold for the step potential
tial (in volts) occurs for K = 0.5. Mr. Blattner explanation describes because this potential is calculated based on a vertical voltage drop as
one of the reasons. Another probable reason is due to the fact that the shown by figure C2. The influence of the top layer is now significant.
potential rise or grid resistance is mainly a function of the equivalent [3] There are practical limits to the depth of the crushed stone (5 to
combined earth resistivity (top and bottom soil layers) while the surface 10 cm) which will prevent that the problem of higher touch or step
potential is mainly determined by the top soil. The touch potential voltages (caused by excessive depth) occurs in practice.
(which is the difference between the two previous quantities) is However, figure 3.15 shows that the increase of step potential is
therefore a function of the top and combined soil resistivities and of the small except when the grid is buried in the crushed stone layer. Also
current densities in the conductors. This function exhibits a maximum usually, step potentials are safe or at least, can be controlled easily,
which, for the case described in the paper, corresponds to K = 0.5. when compared to touch potentials. In our opinion, we believe that in
Other cases would probably have a maximum at different K values. practice the safe limit of this depth can not be exceeded.
[2] The reason why the resistance and touch potentials are constant
when h is smaller than Z = 0.5m is explained as follows: Manuscript received May 14, 1979.