Annex B (informative) Sample calculations

This annex illustrates the application of equations, tables, and graphs for designing a substation grounding
system. The specific objectives are as follows:

a) To show the application of principal equations of this guide for several refinements of the design
concept toward a satisfactory final design solution.
b) To illustrate the typical differences to be expected between results obtained using the simplified
calculations of this guide and the more rigorous computer solutions.
c) To illustrate such design conditions for which the use of simplified calculations of this guide
would not be appropriate for a safe design, as some of the equations may only be used with
caution.

In view of these objectives, the following series of examples (B.1 through B.4) neither represents, nor is
intended to be, the best or most efficient way to design a grounding system.

A computer-based grounding program described in EPRI TR-100622 [B64] was used to model the grids in
these examples.

For the series of examples (B.1 through B.4), the design data are as follows:
 Fault duration tf = 0.5 s
 Positive sequence equivalent system impedance Z1 = 4.0 + j10.0 Ω (115 kV side)

 Zero sequence equivalent system impedance Z0 = 10.0 + j40.0 Ω (115 kV side)

 Current division factor Sf = 0.6

 Line-to-line voltage at worst-fault location = 115 000 V

 Soil resistivity ρ = 400 Ω-m
 Crushed-rock resistivity (wet) ρs = 2500 Ω-m

 Thickness of crushed-rock surfacing hs = 0.102 m (4 in)

 Depth of grid burial h = 0.5 m

 Available grounding area A = 63 m × 84 m
 Transformer impedance, (Z1 and Z0) = 0.034 + j1.014 Ω (13 kV) (Z = 9% at 15 MVA, 115/13 kV)
The crushed-rock resistivity is assumed to be a conservative estimate based on actual measurements of
typical rock samples. The equivalent system fault impedances and current division factor Sf are determined
for the worst-fault type and location, including any conceivable system additions over the next 25 years.
Thus, no additional safety factor for system growth is added. In addition, it is assumed that the substation
will not be cleared by circuit breakers with an automatic reclosing scheme. Thus, the fault duration and
shock duration are equal.

B.1 Square grid without ground rods—Example 1

Using the step-by-step procedure as described in 16.4 and illustrated in Figure 32, the following design
evaluations can be made.

Step 1: Field data. Although the substation ground grid is to be located within a rectangle of 63 m ×84 m
(5292 m2), for the initial design assessment it may be expedient to assume a square 70 m × 70 m grid with
no ground rods. Consequently, the area occupied by such a grid is A = 4900 m2. An average soil resistivity
of 400 Ω-m is assumed, based on soil resistivity measurements.

Step 2: Conductor size. Ignoring the station resistance, the symmetrical ground fault current If ≈ 3I0, is
computed using Equation (72)

E
I= (B.1)
0 3× Rf + (R1 + R2 + R0) + j(X1 + X 2 + X0)

For the 115 kV bus fault

(3)(115 000 / 3 )
3I0 = 3( 0) + ( 4.0 + 4.0 +10.0) + j (10.0 +10.0 + 40.0)
and, hence

|3I | = 3180 A, and the X/R ratio = 3.33 0

For the 13 kV bus fault, the 115 kV equivalent fault impedances must be transferred to the 13 kV side of
the transformer. It should be noted that, due to the delta-wye connection of the transformer, only the
positive sequence 115 kV fault impedance is transferred. Thus

 13 
Z1 =   [4.0 + j10.0]+ 0.034 + j1.014 = 0.085+ j1.142
115

Z0 = 0.034+ j1.014

(3)
13 000 3

 
)
3I = 3(0) +(0.085+ 0.085+ 0.034)+ j(1.142+1.142+1.014
0

and,
hence

|3I | = 6814 A, and the X/R ratio is 16.2 0

The 13 kV bus fault value of 6814 A should be used to size the grounding conductor.

Using Table 10 for fault duration of 0.5 s, the decrement factor Df is approximately 1.0; thus, the rms
asymmetrical fault current is also 6814 A. This current magnitude will be used to determine the minimum
diameter of ground conductors.

Assuming the use of copper wire and an ambient temperature of 40 °C, Equation (47) and Table 2 are used
to obtain the required conductor cross-sectional area. For 0.5 s and a melting temperature of 1084 °C for
hard-drawn copper, the required cross-sectional area in circular mils is
 K t
Akcmil = I × f c (B.2)

Akcmil = 6.814×7.06 0.5 = 34.02 kcmil

34.02 kcmil = 17.2 mm2

Because A =πd2 /4 , the conductor diameter is approximately 4.7 mm, or 0.0047 m if it is solid
2

mm

conductor.

Based on this computation, a copper wire as small as size No. 4 AWG could be used, but due to the
mechanical strength and ruggedness requirements, a larger 2/0 AWG stranded conductor with diameter d =
0.0105 m (0.414 in) is usually preferred as a minimum.

Consequently, at this stage, the designer may opt to check if, alternately, the use of a less conductive (30%)
copper-clad steel wire and the imposition of a more conservative maximum temperature limit of 700 °C
will still permit the use of a conductor with diameter d = 0.01 m.

Using Equation (46) and Table 1 gives

197.4
Akcmil = I (B.3)
TCAP   Ko + Tm 
  ln 
 tcarρr   Ko + Ta 

2
197.4
Akcmil = 6.814 = 59.81 kcmils or 30.24 mm
 3.85   245+ 700 
   
  (0.5)(0.00378)(5.862)  ln  245+ 40 

In this case, dmin = 6.2 mm, or 0.0062 m solid conductor, which is less than d = 0.01 m desired. Hence, a
30% copper-clad steel wire of approximately 2/0 AWG size is a viable alternative for grid wires, even if a
conservative maximum temperature limit of 700 °C is imposed.

Step 3: Touch and step criteria. For a 0.102 m (4 in) layer of surface layer material, with a wet resistivity
of 2500 Ω-m, and for an earth with resistivity of 400 Ω-m, the reflection factor K is computed using
Equation (21)

ρ−ρs
K= (B.4) ρ+ρs

K == −0.72
Figure 11 indicates for K = –0.72 the resistivity of the surface layer material is to be derated by a reduction
factor Cs ≈ 0.74. The reduction factor Cs can also be approximated using Equation (27)

 ρ
009. 1− 
 ρs 
Cs =1−
2hs + 009. (B.5)

 400 
009.1− 
 2500
Cs =1− = 0.74
2(0102.) + 009.

Assuming that for the particular station the location of grounded facilities within the fenced property 1 is
such that the person’s weight can be expected to be at least 70 kg, Equation (30) and Equation (33) may be
used to compute the tolerable step and touch voltages, respectively, as follows:

Estep70 = (1000 + 6Csρs )0157. ts /(B.6)

Estep70 = [(1000 + 6(0.74)2500)]0.157 0.5 = 2686.6

/ ts

Etouch70 = (1000 +15. Csρs ) 0157.(B.7)

Etouch70 =[(1000 +1.5(0.74)2500)]0.157 0.5 = 838.2

Step 4: Initial design. Assume a preliminary layout of 70 m × 70 m grid with equally spaced conductors,
as shown in Figure B.1, with spacing D = 7 m, grid burial depth h = 0.5 m, and no ground rods. The total
length of buried conductor, L , is 2 × 11 × 70 m = 1540 m.
T

1 That is, not accessible to the general public.

 Figure B.1—Square grid without ground rods
Step 5: Determination of grid resistance. Using Equation (57) for L = 1540 m, and grid area A = 4900
m2, the resistance is

1 + 1  1  (B.8)

Rg = ρ 1+  20A 

1+ h 20 / A
LT
 1   1540

R = 400 1 + 1+
1
 = 2.78 Ω
g
  20×4
9 00 
1+ 0.5 20 4900 

Step 6: Maximum grid current IG. Per the procedure and definitions of 15.1, the maximum grid current IG
is determined by combining Equation (68) and Equation (69). Referring to Step 2, for Df = 1.0, and the
given current division factor Sf = 0.6,

Ig
S f= (B.9)
3× Io

and

IG = D f × Ig (B.10)

Though the 13 kV bus fault value of 6814 A is greater than the 115 kV bus fault value of 3180 A, it is
recalled from Clause 15 that the wye-grounded 13 kV transformer winding is a “local” source of fault
current and does not contribute to the GPR. Thus, the maximum grid current is based on 3180 A.

IG = Df ×S f ×3×I0 (B.11)
IG = (1)(0.6)(3180) =1908 A

Step 7: GPR. Now it is necessary to compare the product of IG and Rg, or GPR, to the tolerable touch
voltage, Etouch70

GPR = IG ×Rg (B.12)

GPR=1908×2.78 = 5304 V

which far exceeds 838 V, determined in Step 3 as the safe value of Etouch70. Therefore, further design
evaluations are necessary.

Step 8: Mesh voltage. Using Equation (86), Equation (87), Equation (88), Km is computed

1   D2 (D + 2×h) 2
h  Kii  8 

π
Km = 2× ×ln16×h×d + 8× D×d− 4×d  + Kh ×lnπ(2×n

−1) (B.13)
where

Kii = (B.14)

(2×n)n

Kii = = 0.57
(2×11)11
and

h
Kh = 1+ (B.15)
h0

05
.
Kh 1+ == 1225.
10
.

1  72 (7 + 2×0.5) 2
0.5  0.57  8 
π
Km = 2 ln16×0.5×0.01 + 8×7×0.01 − 4×0.01 +1.225

lnπ(2×11−1) = 0.89
The factor Ki is computed using Equation (89) through Equation (94)

Ki = 0.644+ 0.148×n (B.16)

where

n=n×n×n×n (B.17) a b c d

2×LC (B.18)
n=
a Lp na=

=11

nb = 1 for square grid nc

= 1 for square grid nd = 1

for square grid and

therefore

n=11×1×1×1=11

Ki = 0.644+0.148×11= 2.272

Finally, Em is computed using Equation (85) and Equation (95)

ρ × IG × K m × K i
Em = (B.19)
LC + LR

Em = = 1002.1 V

Step 9: Em versus Etouch. The mesh voltage is higher than the tolerable touch voltage (that is, 1002.1 V
versus 838.2 V). The grid design must be modified.

For comparison, the EPRI TR-100622 [B64] computer program resulted in 2.67 Ω and 984.3 V for the grid
resistance and touch voltage, respectively, for this example.
B.2 Square grid with ground rods—Example 2

In the previous example, B.1, Step 10 of the design procedure has not been reached due to the failure to
meet the criterion of Step 9. Generally, there are two approaches to modifying the grid design to meet the
tolerable touch voltage requirements

a) Reduce the GPR to a value below the tolerable touch voltage or to a value low enough to result in
a value of Em below the tolerable touch voltage.

b) Reduce the available ground fault current.

Usually reduction of the available ground fault current is difficult or impractical to achieve, so the grid is
modified by changing any or all of the following: grid conductor spacing, total conductor length, grid
depth, addition of ground rods, etc. In this example, the preliminary design will be modified to include 20
ground rods, each 7.5 m (24.6 ft) long, around the perimeter of the grid, as shown in Figure B.2.

Figure B.2—Square grid with twenty 7.5 m rods

Step 5. Using Equation (57) for LT = 1540 + 20 × 7.5 = 1690 m, and A = 4900 m2 yields the following
value of grid resistance Rg:

1 1  1 
ρ
Rg =  + 1+  (B.20)

LT 20A  1+ h 20/ A 

 1 1  1 

Rg = 400 + 1+  = 2.75 Ω

1690 20×4900  1+ 0.5 20/4900 

Steps 6 and 7. The revised GPR is (1908)(2.75) = 5247 V, which is still much greater than 838.2 V.

Step 8. Using Equation (80) and Equation (82), K is computed

 m

1   D2 (D + 2×h) 2
h  Kii  8 
π
Km = 2× ln16×h×d + 8× D×d− 4×d  + Kh ×lnπ(2×n

−1) (B.21)

where

Kii = 1.0 with rods and

h
Kh = 1+ (B.22)
h0

05
.
1+
Kh 10
. == 1225.

1  72 (7 + 2×0.5) 2
0.5  1.0  8 
π
Km = 2 ln16×0.5×0.01 + 8×7×0.01 − 4×0.01 +

1.225lnπ(2×11−1) = 0.77

This time, Em is computed using Equation (85) and Equation (96)

ρ× I × Km × Ki
G
(B.23)

  Lr 
1.55 +1.22 2
L C + × L R
  L x + L 2y 
Em =   

400×1908×0.77×2.272
  7.5 
1.55 +1.22 2
1540 +  150
  70 + 702 
Em =    = 747.4 V

Because the step voltage has not been calculated yet, Equation (94) and Equation (97), Equation (98), and
Equation (99) are used to compute Ki, Es, LS, and Ks, respectively. Note that the value for Ki is still 2.272
(same as for mesh voltage).
 1 (B.24)  1
)
1
Ks = 2×h + D + h + D 1−0.5  ( 1
n −2 
π

1 1 1 1 

)
(
Ks =π2×0.5 + 7 + 0.5 + 7 1−0.511−2  = 0.406
Then

ρ×I ×K ×K
0.75
Es = Gs i
(B.25)
×LC
+
0.85
×LR

Es = = 548.9 V

Step 9: Em versus Etouch. Now the calculated corner mesh voltage is lower than the tolerable touch voltage
(747.4 V versus 838.2 V), and we are ready to proceed to Step 10.

Step 10: Es versus Estep. The computed Es is well below the tolerable step voltage determined in Step 3 of
Example 1. That is, 548.9 V is much less than 2686.6 V.

Step 11: Modify design. Not necessary for this example.

Step 12: Detailed design. A safe design has been obtained. At this point, all equipment pigtails, additional
ground rods for surge arresters, etc., should be added to complete the grid design details.

For comparison, the computer program of EPRI TR-100622 [B64] resulted in 2.52 Ω, 756.2 V, and 459.1
V for the grid resistance, touch voltage, and step voltage, respectively, for this example.

B.3 Rectangular grid with ground rods—Example 3

In this example the preliminary grid design will be reconciled in terms of the actual shape of the grounding
area as an alternative design. Realizing that the full grounding area is only about 8% larger than that used in
the previous calculations, most of the conclusions from Example 2 can be used for arriving at a suitable
final design solution.

Choosing, again, spacing D = 7 m, for a rectangular 63 m × 84 m grid, the grid wire pattern is 10 × 13, and
the grid conductor combined length is 13 × 63 m + 10 × 84 m = 1659 m. Assume the use of 38 ground
rods, each 10 m long, as shown in Figure B.3.
 Figure B.3—Rectangular grid with thirty-eight 10 m ground rods

Step 5. Again, using Equation (57), but for L = 1659 m + (38)(10 m) = 2039 m and A = 63 m × 84 m =
T

2
5292 m , gives

1 1  1 
Rg = ρ + 1+  (B.26)
L T 20A  1+ h 20/ A 

 1 1  1 
Rg = 400 + 1+  = 2.62 Ω
2039 20×5292  1+ 0.5 20/5292 

Steps 6 and 7. Using IG = 1908 A as before, and Rg = 2.62 Ω, the GPR = (1908)(2.62) = 4998.96 V, which
is much greater than 838.2 V.

Step 8. For the particular design arrangement shown in Figure B.3, the equations of 16.5.1 can again be
used to estimate the corner mesh voltage. However, because the grid is rectangular, the value of n to be
used in the mesh voltage computation will be different, based on the factors determined using Equation
(89) through Equation (93).

n = na ×nb ×nc ×nd (B.27)

2×LC (B.28) na=

Lp

na = = 11.29
 Lp
nb =
4× A (B.29)

294
nb =
4 × 5292
= 1.005

n = 1 for rectangular grid c

n = 1 for rectangular grid
d

n =11.29×1.005×1×1=11.35
Now K is computed using Equation (86) and Equation (88) m

1   D2 (D + 2×h) 2
h  Kii  8 

π
Km = 2× ×ln16×h×d + 8× D×d− 4×d  + K ×lnπ(2×n

−1) (B.30)
h

where

Kii =1 for a grid with ground rods

05
.
Kh 1+ == 1225.
10
.

1  72 (7 + 2×0.5) 2
0.5  1.0  8 
π
Km = 2 ln16×0.5×0.01 + 8×7×0.01 − 4×0.01 +1.225

lnπ(2×11.35−1) = 0.77
Equation (94) is used to compute K i
Ki = 0.644+ 0.148×n (B.31)

Ki = 0.644+ 0.148×11.35 = 2.324

Finally, E is computed using Equation (85) and Equation (96)

m

ρ × IG × K m × K i
  Lr 
1.55 +1.22 2 2
L C + L R
  L x +L y 
  

400×1908×0.77 ×2.324
Em = = 595.8 V
  10 
1.55 +1.22 2
1659 +  380
  63 +842  E
   m = (B.32)
Step 9. This calculated mesh voltage is well below the E limit of 838.2, but uses 119 m of additional touch70
conductor and 230 m of additional ground rods, as compared with the previous example. Thus, the mesh
spacing could be increased, the number and/or length of ground rods could be reduced, or both to achieve
the same margin of safety as Example 2.

The remaining steps are the same as demonstrated in Example 2 and will not be repeated here.

For comparison, the computer program of EPRI TR-100622 [B64] resulted in 2.28 Ω, 519.4 V, and 349.7
V for the grid resistance, touch voltage, and step voltage, respectively, for this example.

B.4 L-shaped grid with ground rods—Example 4

In this example the design of Example 2 is modified to illustrate the use of the equations for an L-shaped
grid with ground rods. The total area and mesh spacing are the same as that of Example 2, and the ground
rods are located only around the perimeter of the grid, as shown in Figure B.4. All other parameters are the
same as Example 2, except the number of rods (24). Thus, Steps 1 through 4 are the same as Example 2,
and this example begins with Step 5. Step 5. Using Equation (57) for L = 1575 m + (24)(7.5 m) = 1755 m
and A = 4900 m2, gives T

1 1  1 
ρ
Rg =  + 1+  (B.33)

LT 20A  1+ h 20/ A 

1 1  1 
Rg = 400 + 1+  = 2.74 Ω
1755 20×4900  1+ 0.5 20/ 4900 

Steps 6 and 7. The revised GPR is (1908)(2.74) = 5228 V, which is much greater than the tolerable touch
voltage of 838.2 V.

Step 8. Using Equation (89) through Equation (93), and Equation (86) and Equation (94), n, Km, and Ki are
computed
 Figure B.4—L-shaped grid with twenty-four 7.5 m ground rods

n = na ×nb ×nc ×nd (B.34)

2× LC
na = (B.35)
LP

na = =9

LP
nb =(B.36)
4× A

350
nb ==1.12
4 × 4900
0.7×A

L x × L y L x ×L y
=
nc   (B.37)
 A 

70×105
nc =  4900 =1.21
  nd =1 for L-shaped grid

n=(9)(1.12)(1.21)

(1)=12.2
Now K is computed using Equation (86) and Equation (88) m

Kii =1

Kh == 1225.

  2 2 
1 D (D + 2×h) h Kii  8 
(B.38)
π
Km = ×ln + −  + h ×ln (2×n −1) 2× 
16×h×d 8× D×d 4×d  K  π

1  72 (7 + 2(0.5)) 2
0.5  1.0  8 
Km = ln + −  + ln (2(12.2)−1) = 0.76
2π16(0.5)0.01 8(7)0.01 4(0.01) 1.225 π

Equation (88) is used to compute Ki

Ki = 0.644+ 0.148×n (B.39)

Ki = 0.644+ 0.148×12.2 = 2.45

Finally, E is computed using Equation (86) and Equation (96) m

ρ× IG × Km × Ki
Em = (B.40)
  Lr 
1.55 +1.22 2
L C + L R
 
 L x +L y
2 

  

(400)(1908)(0.76)(2.45)
Em == 761.1 V
  7.5 
 1575+ 1.55+1.22 2 2 180

  70 +105 

Equation (97), Equation (98), and Equation (99) are used to compute E , L and K , respectively. It should ss
s be noted that the value for K is still 2.45 (same as for mesh voltage).
i

11 1 1

Ks = π2×h + D + h + D 1−0.5n−2 ( ) (B.41)

 1 1 1 1 

Ks =  () + 7 + 0.5 + 7 (1− 0.512.2−2 ) = 0.41

π2 0.5
Then

ρ×I ×Ks ×Ki

G

Es = (B.42)
0.75×LC + 0.85×LR

Es == 574.6 V

Step 9. Note that this is close to the results of Example 2, and is lower than the tolerable E limit of touch70
838.2 V. Proceed to Step 10. Step 10. The computed E is well below the tolerable step voltage determined in
Step 3 of Example 1. That s
is, 574.6 V is much less than 2686.6 V. Step

11. Not required for this example.

Step 12. A safe design has been obtained and final details can now be added to the design.

For comparison, a computer program of EPRI TR-100622 [B64] gives results of 2.34 Ω, 742.9 V, and 441.8
V for the grid resistance, touch voltage, and step voltage, respectively, for this example.

B.5 Equally spaced grid with ground rods in two-layer soil—Exhibit 1

Using the computer program of EPRI TR-100622 [B64], an equally spaced grid in two-layer soil was
modeled.

As shown in Figure B.5, the 61 m × 61 m (200 ft × 200 ft) grid consisted of four meshes per side, and had
nine ground rods, each 9.2 m (30 ft) long. The diameter of ground rods was 0.0127 m (0.5 in). The diameter
of the grid wire was 0.01 m diameter, buried 0.5 m below the earth’s surface. The depth of the upper layer
300 Ω-m soil was 4.6 m (15 ft); the lower soil had resistivity of 100 Ω-m.

The computer-calculated values of resistance, corner mesh voltage, and maximum step voltage, are as
follows:

R = 1.353 Ω E = 49.66% of GPR E = 18.33% of GPR g

m s

As can be determined from Figure B.6, the mesh voltage coordinates were X = –75.00 ft, and Y= –75.00 ft,
that is, near the center of the corner mesh. The maximum step voltage (not shown) was calculated outside the
grid, between the grid corner (X, Y = –100 ft) and the point at X, Y = –102.12 ft, that is, approximately over
1 m distance in a diagonal direction away from the grid corner.