Electric Power Systems Research 131 (2016) 178–186

Contents lists available at ScienceDirect

Electric Power Systems Research

journal homepage: www.elsevier.com/locate/epsr

Quick calculation of the grounding resistance of a typical 110 kV

transmission line tower grounding system
J. Trifunovic ∗ , M. Kostic
Faculty of Electrical Engineering, University of Belgrade, Bulevar kralja Aleksandra 73, 11000 Belgrade, Serbia

a r t i c l e i n f o a b s t r a c t

Article history: A general practical method, based on the Finite Element Method modelling and a numerical analysis of
Received 22 September 2014 the obtained results, was applied for the calculation of the grounding resistance of a complex ground-
Received in revised form 5 October 2015 ing system frequently used for typical 110 kV transmission line towers. An approximate formula for a
Accepted 10 October 2015
practical evaluation of such a grounding system resistance, valid for the ranges of input parameters used
in the Serbian transmission power system, is proposed. A new, more accurate formula for calculating
Keywords:
the square loop grounding resistance is also given. The proposed method for generating approximate
Complex grounding systems
formulas intended for fast grounding resistance calculations is general and, therefore, can be applied to
Grounding loop resistance
Practical method
any type of complex grounding systems.
Finite Element Method © 2015 Elsevier B.V. All rights reserved.
Transmission line tower
Approximate formula

1. Introduction electrodes (e.g. in karst and sandy terrains), the contact resistance
becomes a dominant component of the total grounding resistance.
In order to meet the electrical safety standards, grounding In such cases, the aforementioned methods and formulas, valid
resistance of a grounding system must be sufficiently low, which for the calculation of the theoretical grounding resistance, are not
is especially hard to reach in troubled environments [1–3]. Its applicable. The values of the square loop grounding resistance
recommended or maximum permitted value varies from 10  presented in [12], obtained experimentally during a period of 30
for lightning protection [4] to 0.1  for sites where protective months, differed from the theoretical ones from 238% to 1354%,
devices must operate very quickly [5]. In order to ensure that although the soil resistivity variations were modest. It was con-
an adequate grounding system is designed, grounding resistance cluded in [14] that the seasonal variations in grounding resistance
should be estimated in an early design phase [6]. Therefore, due to soil moisture variations, considered in [15,16], can to the
simple methods and the designer oriented formulas for the cal- greatest extent be attributed to the variations of the effective con-
culation of the theoretical grounding resistance of various types tact surface between the electrodes and soil, which cause great
of grounding systems are given in standards [7,8] and handbooks variations of the contact resistance and, consequently, the total
[9–11]. Using them, the designers can relatively easily and accu- grounding resistance.
rately calculate theoretical grounding resistances of differently The authors’ efforts are focused on developing a general method
shaped electrodes (e.g., long buried wires, circular and rectangu- for deriving a set of simple formulas for estimating the total ground-
lar loops, grids, vertical rods and metal plates, as well as complex ing resistance of complex grounding systems buried in troubled
grounding systems containing several elements of the same or dif- soils (characterized by imperfect contact), which will be presented
ferent type) buried in uniform or non-uniform (usually two-layer) in the next paper [17]. Those formulas are based on coefficients
soil. reflecting the mutual (proximity) effect existing between the
It was noticed in [12], and confirmed by the 3D Finite Element grounding system elements [18,19], because it was noticed that the
Method (FEM) modelling in [13,14], that in cases where the ground- proximity effects influence the theoretical grounding resistances of
ing system is buried in soils which form poor contact with the the grounding system elements, and not their contact resistances
(it was reported in [14] that imperfect contact does not influence
the distribution of the current outside a narrow region around
the electrode, and, therefore, does not significantly affect the
∗ Corresponding author. Tel.: +381 11 3218374; fax: +381 11 3370144. grounding resistances of other “distant” elements of the grounding
E-mail address: jovan.trifunovic@etf.rs (J. Trifunovic). system).

http://dx.doi.org/10.1016/j.epsr.2015.10.014
0378-7796/© 2015 Elsevier B.V. All rights reserved.
 J. Trifunovic, M. Kostic / Electric Power Systems Research 131 (2016) 178–186 179

Fig. 2. The grounding system of a typical 110 kV transmission line tower used in the
Serbian transmission power system.

Fig. 2. The upper loop (of dimensions L1 × L1 ) is placed at a depth h1 ,

while each of the four identical lower loops (of dimensions L2 × L2 )
is placed around the concrete foundation of its tower footing, at a
depth h2 . Note that the described grounding system is also used in
other Balkan countries.
A large number of grounding systems such as the one shown
in Fig. 2 have been analysed using 3D FEM, which was followed
by a numerical analysis of the obtained results. The final step was
the derivation of an approximate formula for a practical assess-
ment of such a theoretical grounding system resistance, containing
coefficients which take into account the proximity effect. A new
formula for calculating the theoretical grounding resistance of a
square loop was also derived.
The accuracy of the derived formulas was compared with the
accuracy of the well-known formulas given in standards [7,8] and
handbooks [9–11] in a way that the results obtained using all of
those formulas were compared to those obtained by the 3D FEM
modelling as referent.

2. The ranges of input parameters of a typical 110 kV

transmission line tower used in the Serbian transmission
power system

The size of the grounding system presented in Fig. 2, relevant

for the calculation of its grounding resistance, is conditioned by
the structure parameters of the transmission line tower, and espe-
cially by the dimensions and disposition of its footings (the lower
loop size depends on the footing base dimensions, while the upper
loop size depends on the span between the footings). The footing
dimensions, as well as their span, are conditioned by the transmis-
sion line tower height (H) and soil bearing capacity (). Therefore,
in order to obtain the ranges of relevant structure parameters of the
transmission line towers used in practice, the design documenta-
Fig. 1. A typical 110 kV transmission line tower used in the Serbian transmission tion of 72 different towers used in the Serbian transmission power
power system. (For interpretation of the references to colour in this figure legend,
the reader is referred to the web version of this article.)
system has been examined, with the results presented in Table 1

Table 1
The first step in the development of a method intended for the The ranges of relevant structure parameters of a typical 110 kV transmission line
calculation of the grounding system resistance in case of a troubled tower used in the Serbian transmission power system.
soil is the derivation of simple approximate formulas for the cal-
Parameter Range
culation of the theoretical grounding system resistance based on
coefficients reflecting the proximity effect. The method for gener- H 12–30 m
 100–300 kPa
ating such formulas will be presented taking the grounding system L1 5–10 m
of a typical 110 kV transmission line tower used in the Serbian r 0.20–0.44
transmission power system (Fig. 1) as an example. The grounding h1 0.7 m (fixed value)
system, which usually consists of five electrically connected square h2 2 m (fixed value)
p 0.044–0.088 m
loops (usually made of zinc-protected steel strips), is shown in
180 J. Trifunovic, M. Kostic / Electric Power Systems Research 131 (2016) 178–186

(r = L2 /L1 and p is the perimeter of the strip cross-section, usually

varying from 20 mm × 2 mm to 40 mm × 4 mm).

3. Review of the previous calculation methods

The following basic formulas for the calculation of the loop

grounding resistance (R) in a homogeneous soil are given in [7,8]:

 
R= (1)
4 A
where  is the soil resistivity, and A is the area occupied by the loop,

  
R= + (2)
4 A LT
where LT is the total length of the buried conductors, and
Fig. 3. Grounding resistances of several FEM modelled grounding loops
   ( = 100  m, p = 0.068 m) depending on the loop size and depth.
1 1 1
R= +√ 1+  , (3)
LT 20A 1+h 20/A
difficulties characterizing the experimental approach, computa-
where h is the depth of the loop. tional techniques became more popular among the researchers and
Eq. (1) relates to the grounding resistance of a circular metal the validation of new formulas only by the FEM simulation has
plate installed at the ground surface (zero depth). The second term broadly been accepted [20,23].
in Eq. (2) recognizes the fact that the resistance of any grid is higher In order to better understand the influence of various param-
than that of a solid metallic plate having the same area. It is stated eters of the grounding system shown in Fig. 2 on its grounding
in [7] that Eqs. (1) and (2) can be used with reasonable accuracy resistance, their values were varied (within the ranges presented
for grid depths less than 0.25 m. For loop depths in the range of in Table 1) by the 3D FEM modelling described in Refs. [13,14]
0.25–2.5 m, Eq. (3), which takes into account the effect of the loop (the detailed information regarding the FEM modelling of the con-
depth, is recommended. sidered grounding system is given in Appendix B). Afterwards, a
Formulas which take into consideration the electrode’s cross- numerical analysis of the obtained grounding resistances was per-
section are given in [9,10] (Eqs. (4) and (5), respectively): formed.

 8LT2
R= ln (4) 5. A new formula for calculating the square loop grounding
2LT bh resistance
where b is the width of the electrode, and
A large number of square grounding loops, the parameters of
 BL2
R= ln T (5) which belong to the ranges given in Table 1, were modelled apply-
2LT de h ing 3D FEM. As an illustration of the obtained data, the grounding
where de is the diameter of the equivalent electrode’s cross-section, resistances for some of those loops are presented in Fig. 3. In order
which can be calculated as de = 2b/ (or de = b/2, as suggested in to determine a formula characterized with both simplicity and sat-
some literature), while B is a factor dependent upon the grounding isfactory accuracy (suitable for a quick calculation of a square loop
system structure (B = 5.53 for a square loop). grounding resistance), the fitting of several functions R(,LT ,h,p)
Appendix A contains an explanation of the two methods given was done using both the method of least squares and the itera-
in [11], one analytical (Method A.1) and the other semi-numerical tive calculation method. Analysing the results obtained by 3D FEM,
(Method A.2). Since both methods are intended for calculating the it was concluded that logarithmic dependence should be included
grounding resistance of complex grounding systems, they can be in function R(,LT ,h,p) (as in Eqs. (4), (5) and (A.1)). The idea was to
applied to the grounding system presented in Fig. 2. In both meth- keep the simplicity of Eqs. (4) and (5), increase the accuracy by tak-
ods the square loop grounding resistance is calculated using Eq. ing into account the contact surface between the loop’s electrodes
(A.1), which is similar to Eq. (2), but far more complex, as well as and the surrounding soil (feature of Eq. (A.1)) and adopt only one
accurate. The advanced feature of Eq. (A.1) compared to Eqs. (1)–(5) constant in the fraction within the logarithm. This constant was
is that it takes into account the contact surface between the loop’s determined varying its value by Microsoft Excel Solver [30] (the
electrodes and the surrounding soil. adopted constant corresponds to the minimum value of the sum of
squares of differences between the results obtained by 3D FEM and
4. Application of the Finite Element Method (FEM) the R(,LT ,h,p) function values). Finally, the following formula was
derived:
Many studies have analysed behaviour of different grounding 8.85LT2

systems, buried in various soil structures, using both analytical R= ln . (6)
2LT ph
formulas and FEM simulations (grounding loops were analysed
in [20], grids in [21,22], rods surrounded by a backfill material in It can be observed that the recommended Eq. (6) is similar to
[23], multi-layer soil structures in [24], local soil nonuniformities Eqs. (4) and (5), the main difference being the use of the perime-
in [25] and interference phenomena between separate grounding ter of the electrode’s cross-section (p) instead of the width of the
systems in [26]). Furthermore, it was shown in [13,14,27–29] that electrode or its equivalent diameter. It was established through a
the grounding resistances obtained by the measurements at the large number of the FEM models that for the same value of p the
steady-state conditions are similar to those obtained by the FEM ratio between the sides of the electrode’s rectangular cross-section
simulations. Therefore, both FEM and experimental techniques can practically does not influence the loop grounding resistance. The
be used for the validation of new formulas for the calculation of rel- possible explanation of this finding is that the loop grounding resis-
evant grounding system parameters. Due to the cost and numerous tance is predominantly influenced by the contact surface between
 J. Trifunovic, M. Kostic / Electric Power Systems Research 131 (2016) 178–186 181

Table 2 (again, using both the method of least squares and the iterative
The relative errors obtained by using Eqs. (1)–(6) and (A.1) compared to the FEM
calculation method through Microsoft Excel Solver). A very sim-
results.
ple summation function (L1 , r), containing four summands, was
ıR (%) |ıR|av (%) selected (the first and the second summands take into account the
Eq. (1) [−18.36, 56.44] 17.35 influence of the length of lower and upper loops, respectively, the
Eq. (2) [27.71, 144.71] 75.88 third takes into account the influence of relative mutual distances
Eq. (3) [−10.56, 75.15] 26.84 between the four lower loops, and the fourth the influence of the
Eq. (4) [−8.40, −3.83] 5.21
fixed parameters (h1 and h2 ), as well as a parameter p, which does
Eq. (5) [2.36, 11.56] 7.18
Eq. (A.1) [−3.90, 8.30] 3.97 not significantly affect values of the  coefficients). The most precise
Eq. (6) [−3.50, 1.39] 1.35 approximation was achieved by the following equation:

the loop’s electrodes and the surrounding soil, which depends on (L1 , r) = C1 · L1 · r + C2 · L1 + C3 · r + C4 , (9)
the perimeter and not the ratio between the width and height of
the electrode’s cross-section.
The values of the grounding resistance of a large number of where C1 , C2 , C3 and C4 are the constants, the values of which were
square grounding loops, the parameters of which belong to the varied until the minimum value of the sum of squares of differences
ranges given in Table 1, were calculated using Eqs. (1)–(5), (A.1) between the results obtained by 3D FEM and Eq. (9) was achieved.
and (6) and compared to the values obtained by the 3D FEM calcu- For the considered case and the ranges of input parameters given
lations. The ranges of relative errors (ıR(%)), as well as the averages in Table 1, the following values of the constants C1 , C2 , C3 and C4
of their absolute values (|ıR|av (%)), are given in Table 2. According were determined:
to Table 2, the results obtained using Eqs. (1)–(3) are considerably
less accurate than the results obtained by Eqs. (4), (5) and (A.1),
while Eq. (6) is characterized by the highest accuracy. Since Eq. (6) - (C1 , C2 , C3 , C4 ) = (0.005185, −0.005094, −0.1656, 0.7299) for
produces results deviating less than 4% from those obtained by the UL–4LL , and
FEM calculations, it can be recommended for a quick analysis or - (C1 , C2 , C3 , C4 ) = (−0.005264, 0.004609, −0.7925, 0.8793) for 4LL .
design of square grounding loops, the parameters of which belong
to the ranges given in Table 1.
Note that C1 –C4 do not depend on the uniform soil resistivity.
6. An approximate formula for a quick calculation of the For any combination of the input parameters belonging to the
grounding resistance of a typical 110 kV transmission line ranges given in Table 1 and for any value of the uniform soil resis-
tower tivity, the grounding resistance of the grounding system sketched
in Fig. 2 can be calculated combining Eqs. (6)–(9). The values of
A simple method for determining the theoretical grounding RUL and RLL can be calculated by Eq. (6), the  (UL–4LL and 4LL )
resistance of a complex grounding system is presented in [18,19]. coefficients by incorporating the corresponding constants (C1 , C2 ,
The method, based on the principles of superposition and reci- C3 , C4 ) into Eq. (9), R4LL by Eq. (8), and, finally, R by Eq. (7). If, for
procity, results in a simple formula for the calculation of the any reason, the upper loop is not a part of the grounding system, R
grounding system resistance. Applying this method to the ground- is equal to R4LL .
ing system shown in Fig. 2 and considering the upper loop as the The values of the grounding resistance of a large number of the
first and all of the four lower loops as the second part of the ground- considered grounding systems were calculated using Eqs. (6)–(9),
ing system, the following formulas apply: as well as the analytical Method A.1 (Eqs. (A.3) and (A.2) for calcu-
1
 1 1 lating R and R4LL , respectively) and the semi-numerical Method A.2.
= UL−4LL · + , (7) Then, they were compared to the grounding resistances obtained by
R RUL R4LL
1
 1 1 1 1 4
the FEM calculations. The ranges of relative errors (ıR(%)), as well as
= 4LL · + + + = 4LL · , (8) the averages of their absolute values (|ıR|av (%)), are given in Table 3.
R4LL RLL RLL RLL RLL RLL As an illustration, the grounding resistances of some of the con-
where R represents the grounding resistance of the whole ground- sidered grounding systems, calculated using 3D FEM, Eqs. (6)–(9),
ing system, RUL is the grounding resistance of the upper grounding Method A.1 and Method A.2, as well as their relative errors, are
loop, R4LL is the grounding resistance of all four lower loops, UL–4LL given in Table 4. Also as an illustration, as well as to show the influ-
is the coefficient reflecting the mutual (proximity) effect existing ence of parameter r on the grounding resistance, the data obtained
between the upper and the four lower loops, RLL is the ground- by the 3D FEM modelling for some of the considered grounding
ing resistance of a solitary lower loop, and 4LL is the coefficient systems is presented in Fig. 4.
reflecting the proximity effect among the four lower loops. In order to evaluate the upper limit of the possible errors when
Note that a usual (conservative) approach for the determination calculating the grounding resistance R, caused by neglecting the
of the R values was applied, since the positive effect of the connect- positive effect of the connecting strips and the foundation ground-
ing strips and the foundation grounding systems existing in the ing systems, which was not taken into account in the FEM modelling
tower footings was not taken into account. and when deriving new formulas presented in the paper, the
The values of R, RUL , R4LL and RLL for a large number of ground- grounding resistance of the whole grounding system presented in
ing systems such as the one shown in Fig. 2, parameters of which Fig. 2 was also calculated using Eq. (A.4). Comparing these results to
belong to the ranges given in Table 1, were obtained by the 3D the values of the grounding resistance R obtained by FEM (some of
FEM calculations. By incorporating those values into Eqs. (7) and them are given in Table 4), it was obtained that the relative errors
(8), the  (UL–4LL and 4LL ) coefficients were calculated for each ıR were in the range of [−16.18, −1.78], while the average of their
considered grounding system. It was noticed that the obtained absolute values amounted to |ıR|av = 8.15%. This confirms that there
values of the  coefficients are mainly dependent of L1 and r. There- is no need for FEM modelling of the connecting strips and founda-
fore, it was convenient to determine the function (L1 , r), which tion grounding systems, for which the necessary data is difficult to
was realised by fitting several functions through the obtained data obtain.
182 J. Trifunovic, M. Kostic / Electric Power Systems Research 131 (2016) 178–186

Table 3
The relative errors obtained by using the methods from [11] and Eqs. (6)–(9) compared to the FEM results.

ıR (%) |ıR|av (%) ıR4LL (%) |ıR4LL |av (%)

Method A.1, Eqs. (A.3) and (A.2) [0.55, 8.19] 4.81 [−4.85, 4.23] 2.39
Method A.2 [3.56, 23.09] 13.67 [−0.78, 8.83] 3.10
Eqs. (6)–(9) [-2.24, 0.38] 1.22 [−2.25, 1.08] 0.71

Table 4
Comparison of the results obtained by FEM, the methods from [11] and Eqs. (6)–(9) for several grounding systems of the type shown in Fig. 2 (h1 = 0.7 m, h2 = 2 m and
p = 0.068 m).

 ( m) 50 50 100 100 200 200

L1 (m) 5 7.5 10 5 7.5 10
r 0.20 0.28 0.36 0.44 0.36 0.44

R () FEM 3.42 2.38 3.65 6.13 9.17 7.19

R () Eqs. (6)–(9) 3.38 2.33 3.60 6.16 9.02 7.12
␦R (%) Eqs. (6)–(9) −1.19 −2.24 −1.59 0.38 −1.65 −0.86
R () Method A.1, Eq. (A.3) 3.44 2.50 3.95 6.35 9.76 7.77
␦R (%) Method A.1, Eq. (A.3) 0.55 4.83 8.18 3.57 6.44 8.19
R () Method A.2 3.81 2.63 4.086 7.55 10.59 8.35
␦R (%) Method A.2 11.63 10.48 11.79 23.09 15.52 16.14
R4LL () FEM 4.79 2.90 4.22 7.04 10.63 8.12
R 4LL () Eqs. (6)–(9) 4.68 2.93 4.23 7.11 10.72 8.10
␦R4LL (%) Eqs. (6)–(9) −2.25 0.99 0.33 1.08 0.79 −0.25
R4LL () Method A.1, Eq. (A.2) 4.55 2.94 4.39 6.91 10.89 8.38
␦R4LL (%) Method A.1, Eq. (A.2) −4.85 1.19 4.23 −1.80 2.41 3.29
R4LL () Method A.2 4.75 2.97 4.31 7.66 11.01 8.38
ıR4LL (%) Method A.2 −0.78 2.28 2.25 8.83 3.56 3.24

7. Discussion Another advantage of Eqs. (6)–(9), compared to the other two

methods, is that these equations are easier to apply.
7.1. Comparison of methods for calculating the grounding system
resistance 7.2. Analysis of the obtained results

According to Table 3, all three methods (the one proposed in According to [31], the grounding resistance of the overall
this paper and Methods A.1 and A.2) are generally characterized by grounding system of a 110 kV transmission line tower should be
good accuracy. Considering the fact that Eqs. (6)–(9) from the pro- less than 10 . However, if this request cannot be achieved by
posed method produce results deviating less than 3% from those standard grounding system structures and economically justified
obtained by the FEM calculations, these equations can be recom- means, it is suggested that the safety of such a high-voltage trans-
mended for a quick analysis or design of grounding systems shown mission line should be increased using other techniques (enhanced
in Fig. 2. However, these formulas possess the following drawback: insulators, surge arresters etc.). If not, a higher number of line out-
their exceptionally high accuracy is only valid for the ranges of input ages should be allowed on that transmission line. In case that during
parameters given in Table 1 (the other two methods do not have the operation period some transmission towers become particu-
such a limitation). Fortunately, the majority of the 110 kV trans- larly exposed to failures due to lightning strikes or other types of
mission line tower grounding systems possess parameters which overvoltage, their grounding system should be improved or insu-
belong to the ranges given in Table 1. lation enhanced, or both measures should be applied. However,
quantitative measures for the grounding system improvement or
insulation enhancement are not specified in [31].
Applying Eqs. (6)–(9) on the grounding system such as the one
shown in Fig. 2, it can be concluded that its grounding resistance
will be less than 10  for any combination of input parameters
that belong to the ranges given in Table 1 only if it is installed in
the soil characterized by  ≤ 140  m. For soils characterized by
 > 140  m the request R ≤ 10  cannot be achieved for all com-
binations of input parameters that belong to the ranges given in
Table 1 (when grounding resistivity increases, higher values of
parameters L1 and r are required for achieving R ≤ 10 ). If the
grounding system is buried in the soil characterized by  ≥ 270  m,
the request R ≤ 10  cannot be achieved by any combination of
input parameters that belong to the ranges presented in Table 1.
However, Eqs. (6)–(9) are useful for the design of the transmission
line safety measures even if the soil is characterized by  ≥ 270  m,
because the calculated tower grounding resistance is an essential
input parameter needed for the evaluation of the average number
of faults which occur in the transmission line [32,33]. In addition,
proper transmission line insulation dimensioning is based on stud-
Fig. 4. Grounding resistances of several FEM modelled grounding systems ies requested to accurately (or as accurately as possible) express
( = 100  m, p = 0.068 m) depending on L1 and r. the type and magnitude of the expected overvoltages, which
 J. Trifunovic, M. Kostic / Electric Power Systems Research 131 (2016) 178–186 183

cannot be done without knowing the transmission line tower foot- soil parameters [37–40]). Ref. [11] offers a procedure intended for
ing resistance [34]. The grounding resistance value is also essential practitioners, which uses the LCLF resistance as an input parameter
for a research which should show how surge arresters can, in for calculating the transient impedance of an arbitrary conducting
an economically justified way, be installed on transmission lines object. The same procedure can be applied to the LCLF resistances
[35,36]. obtained by the formulas presented in this paper.

7.3. Possible application of the derived formulas outside the

ranges given in Table 1 8. Conclusions

If the calculated grounding resistance is higher (but not con- A general practical method, involving the Finite Element Method
siderably higher) than the requested one (which can occur in soils modelling and a numerical analysis of the obtained results, was
characterized by  ≥ 140  m), the designer can decide to increase applied to a frequent practical case of a grounding system of a
the size of the upper loop (for installation reasons, the size and typical 110 kV transmission line tower used in the Serbian trans-
position of the lower loops remain the same). In order to provide mission power system (Fig. 2). It is based on the calculation of the
the designer with means useful for a quick grounding resistance grounding resistances of its elements (square loops in this case),
calculation even in such cases, the following grounding system which can be done using simple approximate formulas, as well
was analysed using 3D FEM:  = 200  m, p = 0.068 m, h1 = 0.7 m, as the coefficients taking into account the proximity effects. An
h2 = 2 m, L1 = 5 m and L2 = 1 m, resulting in R = 13.67 . Analysing the approximate formula for a practical assessment of such a ground-
results obtained when the length L1 was increased 3 and 5 times, ing system resistance, valid for the specified ranges of the relevant
it was concluded that Eq. (6) still produces results deviating less input parameters, is proposed. A more accurate formula for cal-
than 4% from those obtained by the FEM calculations. However, it culating the square loop grounding resistance, valid for the same
was also observed that the UL–4LL coefficient cannot accurately be ranges of the input parameters, is also derived. These formulas can
calculated using Eq. (9) and the corresponding constants (C1 , C2 , be recommended for a quick analysis or design of grounding sys-
C3 , C4 ) valid for the ranges of the input parameters presented in tems such as the one shown in Fig. 2. Since the proposed method
Table 1. In such cases, the information that the UL–4LL coefficient for deriving approximate formulas for a quick calculation of a com-
increased 1.194 and 1.275 times when the length L1 increased 3 and plex grounding system resistance is general, it can also be applied
5 times, respectively, could be of help to the designer for a quick to any other type of grounding system.
estimation of the grounding resistance of such an untypical (cus-
tom made) grounding system (for L1  = 3L1 and L1  = 5L1 , R = 6.53 
Acknowledgements
and R = 4.31 , both lower than the requested 10 ).
First, using Eq. (9) the designer should calculate the UL–4LL
This research was partially supported by the Ministry of Edu-
coefficient related to the case corresponding to Fig. 2 (without an
cation, Science and Technological Development of the Republic of
increase of L1 ). Then, using the linear interpolation, a more accu-
Serbia (project TR 36018). The authors would like to thank Mr.
rate value of UL–4LL can be computed: if the length L1 is increased
Ivan Milanov, design engineer from Elektroistok Design Office, Bel-
k times, UL–4LL should increase (0.097k + 0.903) times for 1 ≤ k ≤ 3,
grade, for providing the design documentation of typical 110 kV
and (0.0405k + 1.0725) times for k > 3. Since this procedure is not
transmission line towers used in the Serbian transmission power
very accurate, there is a possibility that the UL–4LL coefficient value
system.
calculated in the proposed manner becomes greater than unity. In
such cases, UL–4LL = 1 should be adopted.
Nevertheless, for the design of such untypical grounding sys- Appendix A. The methods from [11] for calculating the
tems it is recommended to use a FEM based professional software, grounding resistance of the grounding system presented in
because it not only more accurately calculates the grounding resis- Fig. 2
tance, but also provides a possibility to check the touch and step
voltages on the surface of the area of interest. Two methods given in [11], one analytical and the other semi-
numerical, can be applied for calculating the grounding resistance
7.4. Limitations of the presented method and the necessity for of the grounding system presented in Fig. 2.
further research

Not only that the derived formulas are useful for calculating the A.1. The analytical method
grounding resistance of a typical 110 kV transmission line tower
grounding system, but, using the presented method, similar for- According to this method, any complex grounding system can
mulas can be derived for any grounding system buried in a uniform be approximated by an electrode that extends vertically to the soil
soil. surface and envelops all parts of the grounding system, additionally
Influence of spatial variation in the soil resistivity (especially taking into account the fact that the grounding resistance of any
in case of two-layer soil stratification, which is most commonly wire structure is higher than that of the solid electrode occupying
adopted in practice) also deserves to be investigated. The deriva- the same volume. Applying equations 11.8–11.13 and Table 11.1
tion of appropriate formulas which take into account this influence from Ref. [11], the grounding resistance of many differently shaped
could also be performed by the 3D FEM modelling and a numerical electrodes can be calculated.
analysis of the obtained results. For a square loop with the side length L and the perimeter of
The derived formulas presented in this paper relate to the low the strip cross-section p, buried in a uniform soil at depth h, the
current and low frequency (LCLF) resistances under uniform soil following equation applies:
conditions. These resistances can be associated with the perfor-
   
mance of the overhead transmission lines only from the safety point
 2 17 L2 + 2h2 1
 L + 4h
of view. In case of a lightning strike to the line, the tower footing R= ln + ln .
2 L2 + 2h2 2L (L + 4h) 4L 8p
resistance is frequency dependent and can significantly be different
than its LCLF value (due to soil ionization and frequency dependent (A.1)
184 J. Trifunovic, M. Kostic / Electric Power Systems Research 131 (2016) 178–186

If the grounding system consists of only 4 lower grounding loops For the whole grounding system presented in Fig. 2, the [R]5 × 5
presented in Fig. 2, the following equation applies: matrix, denoting the upper loop by 5, must be determined. For i,
   j = (1, 4) all resistances Rij have the same values as in the previ-
 2 17 L12 + 2h22 ous case. The self-resistance R55 can be calculated incorporating
R= ln
2 L12 + 2h22 2L1 (L1 + 4h2 ) the input parameters (, L1 , h1 and p) into Eq. (A.1). All of the
mutual-resistances Ri5 (i = (1, 4)) have an equal value, which can
1
 L + 4h
1 2 be calculated by the following equation:
+ ln . (A.2)
16rL1 32pr
√ √
2 2
For the whole grounding system presented in Fig. 2, the follow- Ri5 = = . (A.9)
2 (L1 − L2 )  2L1 (1 − r) 
ing equation can be used:
  
 2 17 L12 + 2h22
R= ln
2 L12 + 2h22 2L1 (L1 + 4h2 ) Appendix B. Application of 3D FEM

1
 L + 4h The grounding system presented in Fig. 2 was modelled apply-
1 2
+ ln . (A.3) ing 3D FEM similarly as in Refs. [21,24,25]. The 3D models of the
4L1 (1 + 4r) 8p (1 + 4r)
analysed grounding system were very similar to those presented
in Refs. [13,14], where a lot of details about the 3D FEM applica-
If the whole grounding system presented in Fig. 2, including tion on such grounding systems and presentations of the calculated
the foundation grounding systems existing in the tower footings, results were given. In Ref. [14] 3D FEM modelling was explained
is approximated by a single equivalent solid square electrode, step by step, because Ref. [14] was aimed to serve as a theoretical
enveloping all elements of the grounding system, the following background for our subsequent papers, including this one.
equation applies: All 3D FEM calculations applied in our research were per-
   formed using the COMSOL Multiphysics 3.5a software package,
 2 17 L12 + 2h22 installed on a PC equipped with a 2.53 GHz dual core processor
R= ln . (A.4)
2 L12 + 2h22 2L1 (L1 + 4h2 ) and 3.50 GB of RAM. The dimensions of the initial cube model were
40 m × 40 m × 40 m. Each side of this volume, except the upper
side representing the boundary between the soil and air, was sur-
A.2. The semi-numerical method
rounded by an additional 4 m thick wall (a subdomain) made of
boundary infinite elements [41]. The electric potential of the outer
The second method given in [11] is based on the development
surfaces of these walls was set on 0 V (distant ground). The pre-
of the [R]n × n matrix (n is the number of the grounding system
sumed uniform soil was incorporated in the model by the value of
elements), containing the self- and mutual-resistance values (Rij ,
its electrical conductivity. The z axis shown in Fig. 2 represents the
i, j = (1, n)) for each solid, conducting element. Note that the [R]
vertical axis of symmetry of the whole volume (the upper ground-
matrix is symmetric (Rij = Rji , i, j = (1, n)).
ing loop is located in the middle of the horizontal cross-section of
The following matrix equation applies:
the volume at the depth h1 ). All five loops were modelled as sub-
[R]−1 · [V ] = [I] , (A.5) domains with their realistic dimensions and electrical properties.
The electric potential of all upper surfaces of the electrodes was set
where [V]n × 1 is the vector of equal components (Vi = V, i = (1, n)), on 200 V in all cases.
and [I]n × 1 is the vector of currents Ij (j = (1, n)) dissipating from each Due to the symmetry, the final model used in all of the per-
element. formed calculations comprised only one eighth of the initial model
The total current I (dissipating from the whole grounding sys- (Fig. A1). The finite-element mesh in the model volume was made
tem) can be calculated as by four-node solid elements (tetrahedrons). The number of tetra-
n hedral finite elements used in this analysis ranged from 180,000
I= Ij . (A.6) to 230,000, providing high accuracy of the results. Very small ele-
j=1 ments were used close to the electrode (lengths of edges of these
tetrahedrons belonged to the range of 2–10 mm, depending on the
The grounding system resistance R can simply be calculated by size of electrodes used in the models). Their size increased with the
dividing the electric potential applied on the electrodes (V = 200 V distance from the electrode.
in the considered case) by the total current I. The applied 3D FEM software calculates the potential of all
If a grounding system consisting of only 4 lower grounding finite-element nodes of the model. Fig. A1, serving as an illustra-
loops presented in Fig. 2 is considered, the [R]4 × 4 matrix, denot- tion, displays the potential distribution in the soil obtained for the
ing each loop by 1 to 4 clockwise, can be determined and the model corresponding to the grounding system characterized by
above procedure for calculating the grounding resistance R = R4LL  = 100  m, L1 = 7.5 m, r = 0.36, h1 = 0.7 m, h2 = 2 m and p = 0.068 m.
can be conducted. According to Ref. [11], all self-resistances Rii Once the potentials at all points of the model volume are
(i = (1, 4)), having an equal value, can be calculated incorporat- determined, the potential distribution over any plane or along an
ing the input parameters (, L2 = r·L1 , h2 and p) into Eq. (A.1). The arbitrary line within the model volume can be obtained. The poten-
mutual-resistances can be calculated by the following two equa- tial distributions over several planes selected within models similar
tions: to the considered one are presented in [14]. Given also as an illustra-
  tion, the diagram shown in Fig. A2 contains curves representing the
R12 = R14 = R23 = R34 = = , (A.7)
2 (L1 − L2 )  2L1 (1 − r)  potential distributions along the two lines belonging to the ground
and surface (z = 0) determined by y = 0 and y = x (the coordinate system
√ √ is given in Fig. 2), related to the model presented in Fig. A1.
2 2 Considering that the applied FEM software can also calculate
R13 = R24 = = . (A.8)
4 (L1 − L2 )  4L1 (1 − r)  the current density vector (J) at an arbitrary location in the model
 J. Trifunovic, M. Kostic / Electric Power Systems Research 131 (2016) 178–186 185

between the electric potential applied on the upper surfaces of

electrodes (200 V) and the total current.

References

[1] N. Mohamad Nor, S. Abdullah, R. Rajab, Z. Othman, Comparison between utility

sub-station and imitative earthing systems when subjected under lightning
response, Int. J. Electr. Power 43 (2012) 156–161, http://dx.doi.org/10.1016/j.
ijepes.2012.04.042.
[2] S.C. Lim, C. Gomes, M.Z.A. Ab Kadir, Electrical earthing in troubled environment,
Int. J. Electr. Power 47 (2013) 117–128, http://dx.doi.org/10.1016/j.ijepes.2012.
10.058.
[3] Y. Khan, F.R. Pazheri, N.H. Malik, A.A. Al-Arainy, M.I. Qureshi, Novel approach of
estimating grounding pit optimum dimensions in high resistivity soils, Electr.
Power Syst. Res. 92 (2012) 145–154, http://dx.doi.org/10.1016/j.epsr.2012.06.
003.
[4] A. Piantini, J.M. Janiszewski, The use of shield wires for reducing induced vol-
tages from lightning electromagnetic fields, Electr. Power Syst. Res. 94 (2013)
46–53, http://dx.doi.org/10.1016/j.epsr.2012.04.012.
[5] G. Eduful, J.E. Cole, P.Y. Okyere, Optimum mix of ground electrodes and con-
ductive backfills to achieve a low ground resistance, in: Proceedings of the
2nd International Conference on Adaptive Science & Technology, 2009, pp.
140–145, http://dx.doi.org/10.1109/ICASTECH.2009.5409734.
[6] N. Mohamad Nor, R. Rajab, Z. Othman, Validation of the earth resistance
formulae using computational and experimental methods for gas insulated
sub-station (GIS), Int. J. Electr. Power 43 (2012) 290–294, http://dx.doi.org/10.
1016/j.ijepes.2012.04.056.
[7] ANSI/IEEE, Guide for safety in AC substation grounding, in: ANSI/IEEE Std. 80,
ANSI/IEEE, 1986, http://dx.doi.org/10.1109/IEEESTD.1986.81070.
[8] ANSI/IEEE, Guide for safety in AC substation grounding, in: ANSI/IEEE Std. 80,
ANSI/IEEE, 2000, http://dx.doi.org/10.1109/IEEESTD.2000.91902.
[9] A.T. Golovan, P.G. Grudinski, G.N. Petrov, A.M. Fedoseev, M.G. Cilikin, I.V. Antik,
Electrical Engineering Handbook Energy, Moscow-Leningrad, Moscow, 1964
(in Russian).
[10] European Copper Institute and Copper Development Association, Power Qual-
ity Application Guide, Earthing & EMC, Earthing Systems–Basic Constructional
Aspects, European Copper Institute and Copper Development Association,
Brussels–London, 2001.
[11] W.A. Chisholm, Transmission system transients: grounding, in: L.L. Grigsby
(Ed.), The Electric Power Engineering Handbook: Power Systems, third ed., CRC
Press, Boca Raton, FL, 2012, pp. 11-1–11-19.
[12] M.B. Kostic, Z.R. Radakovic, N.S. Radovanovic, M.R. Tomasevic-Canovic,
Improvement of electrical properties of grounding loops by using bentonite
and waste drilling mud, IEE Proc. C 146 (1999) 1–6, http://dx.doi.org/10.1049/
ip-gtd:19990057.
Fig. A1. The potential distribution in the soil (one eighth of the model). (For inter- [13] J. Trifunović, The algorithm for determination of necessary characteristics of
pretation of the references to colour in this figure legend, the reader is referred to backfill materials used for grounding resistances of grounding loops reduction,
the web version of this article.) J. Electr. Eng. 63 (2012) 373–379, http://dx.doi.org/10.2478/v10187-012-0055-
1.
[14] J. Trifunovic, M. Kostic, Analysis of influence of imperfect contact between
grounding electrodes and surrounding soil on electrical properties of ground-
ing loops, Electr. Eng. 96 (2014) 255–265, http://dx.doi.org/10.1007/s00202-
013-0291-9.
[15] V.L. Coelho, A. Piantini, H.A.D. Almaguer, R.A. Coelho, W.C. Boaventura, J.O.S.
Paulino, The influence of seasonal soil moisture on the behavior of soil resis-
tivity and power distribution grounding systems, Electr. Power Syst. Res. 118
(2015) 76–82, http://dx.doi.org/10.1016/j.epsr.2014.07.027.
[16] F.E. Asimakopoulou, G.J. Tsekouras, I.F. Gonos, I.A. Stathopulos, Estimation of
seasonal variation of ground resistance using Artificial Neural Networks, Electr.
Power Syst. Res. 94 (2013) 113–121, http://dx.doi.org/10.1016/j.epsr.2012.07.
018.
[17] J. Trifunovic, M. Kostic, An algorithm for estimating the grounding resistance
of complex grounding systems including contact resistance, accepted for pub-
lication, in: IEEE Trans. Ind. Appl., 2015, http://dx.doi.org/10.1109/TIA.2015.
2429644.
[18] M. Kostic, M. Jovanovic, A new approach of determining grounding resistance
of one class of complex grounding systems, Ser.: Power Eng. 120 (1987) 1–11
(Publications of Faculty of Electrical Engineering, University of Belgrade).
[19] M.B. Kostic, Analysis of foundation grounding systems with external loops
and rods, IEE Proc. C 140 (1993) 73–76, http://dx.doi.org/10.1049/ip-c.1993.
Fig. A2. The potential distributions along the two lines situated at the ground sur-
0010.
face. [20] E. Mombello, O. Trad, J. Rivera, A. Andreoni, Two-layer soil model for power
station grounding system calculation considering multilayer soil stratifica-
tion, Electr. Power Syst. Res. 37 (1996) 67–78, http://dx.doi.org/10.1016/0378-
volume, the total current, I (dissipating from the grounding system), 7796(96)01042-5.
[21] J. Nahman, I. Paunovic, Mesh voltages at earthing grids buried in multi-layer
can easily be computed by the following equation:
soil, Electr. Power Syst. Res. 80 (2010) 556–561, http://dx.doi.org/10.1016/j.
 epsr.2009.10.017.

I= 
J · dS, (B.1)
[22] F. Freschi, M. Mitolo, M. Tartaglia, An effective semi-analytical method for sim-
ulating grounding grids, IEEE Trans. Ind. Appl. 49 (2013) 256–263, http://dx.
S doi.org/10.1109/TIA.2012.2229688.
[23] L.H. Chen, J.F. Chen, T.J. Liang, W.I. Wang, A research on used quantity of ground
where S is a surface enclosing all of the grounding system elec- resistance reduction agent for ground systems, Eur. Trans. Electr. Power 20
trodes. The grounding system resistance represents the ratio (2010) 408–421, http://dx.doi.org/10.1002/etep.322.
186 J. Trifunovic, M. Kostic / Electric Power Systems Research 131 (2016) 178–186

[24] J. Nahman, I. Paunovic, Resistance to earth of earthing grids buried in multi- [33] I.F. Gonos, L. Ekonomou, F.V. Topalis, I.A. Stathopulos, Probability of back-
layer soil, Electr. Eng. 88 (2006) 281–287, http://dx.doi.org/10.1007/s00202- flashover in transmission lines due to lightning strokes using Monte-Carlo
004-0282-y. simulation, Int. J. Electr. Power 25 (2003) 107–111, http://dx.doi.org/10.1016/
[25] J. Nahman, I. Paunovic, Effects of the local soil nonuniformity upon perform- S0142-0615(02)00025-X.
ances of ground grids, IEEE Trans. Power Delivery 22 (2007) 2180–2184, http:// [34] M.Z.A. Ab Kadir, I. Cotton, Application of the insulator coordination gap models
dx.doi.org/10.1109/TPWRD.2007.905284. and effect of line design to backflashover studies, Int. J. Electr. Power 32 (2010)
[26] F. Freschi, M. Mitolo, M. Tartaglia, Interferences phenomena between separate 443–449, http://dx.doi.org/10.1016/j.ijepes.2009.09.020.
grounding systems, IEEE Trans. Ind. Appl. 50 (2014) 2853–2860, http://dx.doi. [35] A.H.A. Bakar, D.N.A. Talib, H. Mokhlis, H.A. Illias, Lightning back flashover double
org/10.1109/IAS.2013.6682577. circuit tripping pattern of 132 kV lines in Malaysia, Int. J. Electr. Power 45 (2013)
[27] N. Mohamad Nor, M. Trlep, S. Abdullah, R. Rajab, Investigations of earthing 235–241, http://dx.doi.org/10.1016/j.ijepes.2012.08.048.
systems under steady-state and transients with FEM and experimental work, [36] M.S. Savic, Estimation of the surge arrester outage rate caused by lightning
Int. J. Electr. Power 44 (2013) 758–763, http://dx.doi.org/10.1016/j.ijepes.2012. overvoltages, IEEE Trans. Power Delivery 20 (2005) 116–122, http://dx.doi.org/
08.031. 10.1109/TPWRD.2004.835435.
[28] J.A. Güemes, F.E. Hernando, Method for calculating the ground resistance of [37] A.B. Lima, J.O.S. Paulino, W.C. Boaventura, I.J.S. Lopes, M.F. Guimarães, W.A.
grounding grids using FEM, IEEE Trans. Power Delivery 19 (2004) 595–600, Chisholm, B. Jamali, F. Bologna, Transient grounding impedance and transient
http://dx.doi.org/10.1109/TPWRD.2004.824761. resistivity measurements using a very short current lead, Electr. Power Syst.
[29] J.A. Güemes-Alonso, F.E. Hernando-Fernández, F. Rodríguez-Bona, J.M. Ruiz- Res. 118 (2015) 69–75, http://dx.doi.org/10.1016/j.epsr.2014.05.014.
Moll, A practical approach for determining the ground resistance of grounding [38] S. Mousa, N. Harid, H. Griffiths, A. Haddad, Experimental investigation of high
grids, IEEE Trans. Power Delivery 21 (2006) 1261–1266, http://dx.doi.org/10. frequency and transient performance of earth rod systems, Electr. Power Syst.
1109/TPWRD.2006.874121. Res. 113 (2014) 196–203, http://dx.doi.org/10.1016/j.epsr.2014.03.024.
[30] D. Fylstra, L. Lasdon, J. Watson, A. Waren, Design and use of the Microsoft Excel [39] J.H. Choi, B.H. Lee, An analysis of conventional grounding impedance based on
Solver, Interfaces 28 (1998) 29–55, http://dx.doi.org/10.1287/inte.28.5.29. the impulsive current distribution of a horizontal electrode, Electr. Power Syst.
[31] The Electric Power Industry of Serbia, Grounding of overhead line tow- Res. 85 (2012) 30–37, http://dx.doi.org/10.1016/j.epsr.2011.07.005.
ers 1 kV, 10 kV, 20 kV, 35 kV and 110 kV, in: Technical Recommendation [40] A. Ametani, T. Chikaraa, H. Morii, T. Kubo, Impedance characteristics of ground-
No. 9, The Electric Power Industry of Serbia, Belgrade, 2010 (in Serbian) ing electrodes on earth surface, Electr. Power Syst. Res. 85 (2012) 38–43, http://
http://www.eps.rs/SiteCollectionDocuments/TP%209.pdf. dx.doi.org/10.1016/j.epsr.2011.07.006.
[32] M.S. Savic, Sensitivity analysis of lightning performance calculations for trans- [41] O.C. Zienkiewicz, C. Emson, P. Bettess, A novel boundary infinite element, Int.
mission lines and substations, IEE Proc. C 132 (1985) 217–223, http://dx.doi. J. Numer. Methods Eng. 19 (1983) 393–404, http://dx.doi.org/10.1002/nme.
org/10.1049/ip-c:19850037. 1620190307.