IEEE TRANSACTIONS ON POWER DELIVERY, VOL. 38, NO.

2, APRIL 2023 877

Lightning Efficient Counterpoise Configurations

for Transmission Line Grounding
Leonid Grcev , Life Fellow, IEEE, Blagoja Markovski , and Mirko Todorovski , Senior Member, IEEE

Abstract—Different configurations of counterpoises are often The problem of enlarging grounding’s effective size is one
used to reduce the resistance and impulse impedance of the trans- of the recently researched topics and one of the most intricate,
mission line grounding. However, there are no general rules for e.g., [4], [5], [6], [7], [8], [9], [10]. Recent standards and recom-
an actual design. This paper analyses the performance of several
configurations of counterpoises under typical lightning current mendations by IEEE, CIGRE, and EPRI have covered this topic
impulses. The basis is a parametric analysis derived from time [11], [12], [13], [14], [15], [16], but “there is no consensus yet
responses of 11,386 test cases for each configuration with param- on how to apply present knowledge to the design of the actual
eters in broad ranges computed using a rigorous electromagnetic electrode system” [17].
model based on the method of moments. The contributions of the This paper aims to develop a method for evaluating the op-
paper are: Simple formulas for the resistance, impulse impedance,
and effective length of counterpoise configurations are derived timal counterpoise configuration of overhead transmission line
from and verified by the simulation results; a configuration with grounding for best protection against lightning. The method’s
an 8-leg counterpoise bent parallel to the line is proposed, which basis is simple formulas that are derived and validated using
extends the effective length and is efficient in high resistivity soil; simulation results of the time responses of 11386 test cases for
and the method does not depend on the different lightning current each configuration with parameters in broad ranges. Simulations
impulse waveform representations based on Heidler, CIGRE, and
double-exponential formulas, since the results can be easily con- are performed using a rigorous electromagnetic model [18], [19],
verted. A simple procedure is provided to compare the capabilities [20]. The model computes the grounding electrodes’ current
and limitations of counterpoise configurations. New formulas and distribution by numerically solving the electric field integral
procedures make up a general and straightforward method for equation in the frequency domain by the method of moments.
analyzing the optimal design, considering the cost-effective and best The corresponding Green functions are based on a mathemat-
protection against lightning.
ically exact solution of the electromagnetic field in a layered
Index Terms—Transmission lines, lightning protection, ground- medium, which involves Sommerfeld integrals. Such integrals
ing, modeling. cannot be solved analytically; however, accurate computations
are today possible using dedicated computer software, e.g.,
[48]. The solution is not limited in frequency and includes
I. INTRODUCTION all electromagnetic propagation effects. Required quantities,
HE reliability of power systems depends on the lightning such as GPR, can be straightforwardly computed from known
T performance of transmission lines. Such performance is
improved by reducing the transmission line grounding resistance
currents. Fourier transformation techniques are used to obtain
the time-domain response to lightning currents. Olsen et al. [21]
R and impulse impedance Z [1], [2], [3]. The most effective have established this model as an “exact” solution to this problem
way of lowering R, especially in medium-to-high resistive soil, and a “gold standard” for comparisons. Since this model is
is by long horizontal grounding electrodes (known as “coun- compared with many experimental results, e.g., [22], [23], [24],
terpoises”). However, there is a limit to reducing Z since it [25], the simple formulas are consistent with these experiments.
becomes constant for grounding larger than a limiting size called The original contributions of this paper are:
“effective.” r A complete set of new formulas of resistance, impulse
impedance, and effective length for analyzing the light-
ning efficiency of different counterpoise configurations is
Manuscript received 26 April 2022; revised 13 July 2022; accepted 15 August
2022. Date of publication 22 August 2022; date of current version 24 March derived.
2023. This work was supported in part by the Ss. Cyril and Methodius University r A configuration with an 8-leg counterpoise is proposed,
in Skopje under Project [Link].20-21.10 and in part by the Macedonian which, compared to the usual 4-leg configuration, enlarges
Academy of Sciences and Arts. Paper no. TPWRD-00602-2022. (Corresponding
author: Leonid Grcev.) the effective length and is more efficient in high resistive
Leonid Grcev is with the Macedonian Academy of Sciences and Arts, Skopje soil.
1000, Macedonia (e-mail: lgrcev@[Link]). r It is shown that results got using the mathematical repre-
Blagoja Markovski and Mirko Todorovski are with the Faculty of Electrical
Engineering and Information Technologies, Ss. Cyril and Methodius University sentation of the lightning current impulse with a Heidler
in Skopje, Skopje 1000, Macedonia (e-mail: bmarkovski@[Link]; function can be easily converted to other impulse represen-
mirko@[Link]). tations, such as CIGRE and double-exponential functions.
Color versions of one or more figures in this article are available at r A general and straightforward method for evaluating op-
[Link]
Digital Object Identifier 10.1109/TPWRD.2022.3200579 timal design enables the cost-effective analysis of the

0885-8977 © 2022 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission.
See [Link] for more information.

Authorized licensed use limited to: UNIV ESTADUAL PAULISTA JULIO DE MESQUITA FILHO. Downloaded on October 01,2023 at [Link] UTC from IEEE Xplore. Restrictions apply.
 878 IEEE TRANSACTIONS ON POWER DELIVERY, VOL. 38, NO. 2, APRIL 2023

counterpoise configurations as part of transmission line

grounding for best protection against lightning effects.

II. BASIC ASSUMPTIONS

Despite the complexity of the physical processes of grounding
systems dissipating the typical lightning current into the ground,
simplifying assumptions is necessary to derive engineering
formulas suitable for the early phases of grounding design.
Such simplifications are also related to the “high degree of
uncertainties involved in determining the relevant parameters
and verification tests” [3]. It is, however, imperative that such
assumptions must be conservative concerning safety.
Fig. 1. The lightning current waveforms used in IEEE and IEC standards and
CIGRE recommendations, and definitions of their front times T1 .
A. Uniform Soil Model
This study considers the uniform soil model, characterized in Fig. 1):
by constitutive parameters: resistivity, permittivity, and free
1 t10
space permeability. The apparent soil resistivity is determined i(t) = · 10 · e−t/τ2 . (1)
from measurements [26], [27], while the relative permittivity is η t + τ110
usually assumed, e.g., 10 to 15 (here we assumed value of 10). The parameters η, τ1 , and τ2 are determined for given values
Although the two-layer soil model more accurately represents of the current peak, the front time T1 , and time-to-half-peak-
layered soil, the equivalent uniform model offers conservative value T2 [35].
results [28]. In Section XI, we show that, despite the considerable differ-
The non-linear behavior of ground resistivity with heavy ences between the current impulses waveforms, the results got
currents due to soil ionization might improve the lightning for the Heidler function (1) can be converted for the other two
performance of concentrated grounding systems [14]. However, current waveforms in Fig. 1.
this effect might be negligible on long counterpoises where
the current dissipated per unit length is small [29]. Therefore, C. Quantities Used to Characterize Grounding’s Lightning
this effect is neglected in this study (see further discussion in Performance
Section XII). The second phenomenon that predicts improved
lightning performance is the frequency dependence of the soil The low-frequency, low-current value of grounding resistance
resistivity [12]. However, the knowledge of this complex phe- R is the primary quantity that characterizes grounding systems.
nomenon is incomplete [30], [31], [32], and analysis shows sig- The advantage of using R is that its value is easy to calculate or
nificant differences and inconsistencies “between the available measure [27], [36]. R is linearly proportional to soil resistivity
models/expressions which are based on different sets of experi- ρ, and R/ρ depends on the geometry of the grounding system.
mental data” [31]. Also, the impact of “the nonlinear variations R is also commonly used in lightning performance analysis,
of the soil moisture content” has to be further investigated [32]. especially if the grounding system is smaller than the effective
Therefore, it is a conservative assumption to disregard this effect size. In such a case, the voltage v(t) developed between the point
and assume constant soil resistivity. where the lightning current i(t) is injected into the grounding
system and a distant neutral ground is
v(t) ≈ Ri(t). (2)
B. Lightning Current Pulse Waveform
If the grounding system is larger than its effective size, then
There is no consensus on the mathematical representation during the i(t) initial rise and a period after the current peak Im ,
of current impulse waveforms of first and subsequent return v(t) may be
strokes used in international standards and recommendations
(see Fig. 1). IEEE standard [11] uses a double-exponential v(t) > Ri(t). (3)
function. The second one is the current waveform recommended In such a case, v(t) may exhibit a significant peak Vm . The
by the CIGRE study group [33], and the third is the analytical voltage and current peaks define the impulse impedance Z:
form of the current used in the International Electrotechnical
Commission (IEC) standard [34] based on the Heidler function Vm
Z= . (4)
[35]. EPRI report [15] also used the latter. Fig. 1 illustrates the Im
definitions of the front time T1 for the different impulses used Z represents grounding systems in lightning performance
in [11], [33] and [34]. analysis [37]. Z remains constant for all grounding system sizes
We use the Heidler function [35] from IEC standard [34] to larger than the effective one and Z > R. One goal of the grounding
represent the return-stroke current waveform (designated by IEC system design is to reduce the value of R by enlarging the

Authorized licensed use limited to: UNIV ESTADUAL PAULISTA JULIO DE MESQUITA FILHO. Downloaded on October 01,2023 at [Link] UTC from IEEE Xplore. Restrictions apply.
 GRCEV et al.: LIGHTNING EFFICIENT COUNTERPOISE CONFIGURATIONS FOR TRANSMISSION LINE GROUNDING 879

Fig. 2. A simple model of “double-loop” grounding connected to tower

concrete foundations.

Fig. 4. (a) Measured current and voltage pulses in 15-m long buried wire. (b)
Computed voltage for the same current pulse in wire divided into three 5-m long
segments.

IV. PROBLEM WITH LIGHTNING PERFORMANCE OF

COUNTERPOISES AND THE SOLUTION
Counterpoises are horizontal conductors buried in the earth
at about 1 meter connected to the footings [2]. Counterpoises
with extended lengths are often used to decrease tower ground-
ing resistance in medium-to-high resistivity soils. However,
Fig. 4(a) illustrates a problem with the lightning performance
of counterpoises.
Fig. 4(a) shows recordings of one of the grounding system
experiments for sub-microsecond rise time input current pulses
Fig. 3. Grounding resistivity R for any soil resistivity ρ of three TL concen-
that EdF conducted in Les Renardieres, France, from 1976 to
trated grounding configurations. 1978 [38], [25]. Fig. 4(a) shows measured voltage v(t) at current
injection point to neutral ground as a response to injected current
pulse i(t) with front time T1 ≈ 0.47 μs in 15-m long horizontal
wire buried at 0.6 m in soil with ρ = 79 Ωm and εr = 15 1 . The
performance during the current rise and short period after the
grounding electrodes while keeping them smaller than their current peak is much worse than during the rest of the impulse
effective size for which Z ≈ R. where v(t) ≈ Ri(t). In the initial period, due to the inductive
behavior, v(t) exhibits a peak of about 500 V, while the peak of
Ri(t) is near 200 V.
Bewley [39] has shown that such a performance can be im-
III. TOWER FOOTING CONCENTRATED GROUNDS
proved if a single counterpoise is replaced with multiple shorter
Depending on the construction, the usual tower footing foun- ones. This is illustrated in Fig. 4(b), where the 15-m wire is
dations include tower steel, grillage, and concrete/foundation divided into three 5-m wires. Impulse performance is improved
reinforcing cages, which provide substantial grounding resis- because of a much smaller inductive component of the three
tance [15]. Some constructions have two rings around the foun- shorter wires than the single wire. The multiple wire resistivity
dation (see Fig. 2). If the tower is accessible to the public, is about 10% higher than the single wire because of mutual
a potential control ring is often placed around the four foot- resistance between wire segments, but v(t) ≈ Ri(t) and Z ≈ R,
ings (see configuration C2 in Fig. 3). If the required value which is the desired design outcome.
of R is not achieved, radially connected ground rods are of- The process of the counterpoise reconfiguration for better im-
ten installed. Such elements of the grounding system that are pulse performance in Fig. 4(a) and (b) can be viewed in terms of
within about 15 meters of the tower base are known as effective length. An effective area [40] can be imagined around
“concentrated grounds” [2]. the tower footings as circular areas with centers at the footings
For example, Fig. 3 shows the range of R values for any ρ for and radiuses equal to the effective length ef f (see Fig. 5).
the three grounding configurations (denoted as C1, C2, and C3) Multiple counterpoises connected radially to the footings, all
as a function of the footing distance. Additional electrodes in within the effective area, reduce R while remaining Z ≈ R. If the
configurations C2 and C3 improve the grounding resistance in effective area goes beyond the right of way, the conductors can
relation to C1, but further improvement is possible by connecting
more extended electrodes. However, a suitable configuration of
concentrated grounds usually can achieve a required value of R, 1 Interested reader can find details on the measuring setup and the influence
e.g., R ≤ 10 Ω, in soils with ρ smaller than about 300 Ωm. of the measuring system on the results in [25].

Authorized licensed use limited to: UNIV ESTADUAL PAULISTA JULIO DE MESQUITA FILHO. Downloaded on October 01,2023 at [Link] UTC from IEEE Xplore. Restrictions apply.
 880 IEEE TRANSACTIONS ON POWER DELIVERY, VOL. 38, NO. 2, APRIL 2023

TABLE I
VALUES OF THE GEOMETRIC PARAMETERS OF CONFIGURATIONS IN FIG. 6

Fig. 5. Effective area around TL tower footings.

This paper proposes 8- and 12-leg configurations, GS6 and

GS7 [see Fig. 6(c)], variants of GS3 and GS4, whose coun-
terpoises are not radial but bent parallel to the line to remain
within the right of way. Such a modification of the configurations
allows for longer counterpoises and enlarges the total effective
length. The distances between parallel counterpoises are 20 m or
higher (for GS5) but can be 10 to 15 m (especially for GS7). All
counterpoises in Fig. 6 have equal length  and are connected
to the current injection points, i.e., the tower’s footings. The
length  is measured from the footing to the counterpoise open
end. Ranges of the considered geometric parameters are given
in Table I.
Comparison of the performances of these configurations is
given in Section X.

VI. APPROXIMATION OF QUANTITIES THAT CHARACTERIZE

LIGHTNING PERFORMANCE
Grounding resistance R is expressed as a function of  and
impulse impedance Z is approximated by [40]:

Z = R(),  ≤ ef f (5)

Fig. 6. (a) Counterpoises radially connected to TL tower footings. (b) Single
counterpoises parallel to the line. (c) Multiple counterpoises parallel to the line. Z = R(ef f ),  > ef f (6)

Here, ef f is the effective length. Like , it is measured from

the footing to the counterpoise open end. The definition of ef f
be bent parallel to the line to stay within the right of way; it is is discussed in Section XIII.A.
crucial that they remain within the effective area. ef f is approximated by an empirical formula [44]

V. CONFIGURATIONS OF COUNTERPOISES ef f = D ρT1 (7)
Fig. 6 illustrates the configurations of counterpoises analyzed where ρ is soil resistivity in Ωm, T1 is lightning current pulse
in this paper. 2-leg counterpoise, GS1 in Fig. 6(a), can be placed front time in μs. The constant D (7) is determined by the best fit
parallel to the line and the effective length limits its usability in of (6) to values of Z got by computer simulation with the rigorous
more resistive soils. The others in Fig. 6(a), GS2, GS3, and GS4, electromagnetic model [18], [19], [20] (see Section VIII).
are radial configurations recommended by France’s EdF [41] The interested readers can find more details on this method
and used in many countries [42]. However, they cover the entire of approximating Z and ef f in [40].
effective area (see Fig. 5), which is limited by the right of way
for practical reasons. Therefore, such practical considerations
limit electrodes length that can be used to about 35 m [41], VII. LOW-FREQUENCY GROUNDING RESISTIVITY
constraining their usability. Fig. 7 shows an example of the grounding resistance R of two
Configuration GS5 in Fig. 6(b) is one of the most often used. configurations: C1, comprising four radial counterpoises alone,
It is a variant of GS2 with bent counterpoises parallel to the line and C2, the counterpoises together with foundation grounding,
within the right-of-way. The effective length mainly limits its as a function of the counterpoises length . It is shown that the
efficacy. influence of the local grounding system of the tower footings

Authorized licensed use limited to: UNIV ESTADUAL PAULISTA JULIO DE MESQUITA FILHO. Downloaded on October 01,2023 at [Link] UTC from IEEE Xplore. Restrictions apply.
 GRCEV et al.: LIGHTNING EFFICIENT COUNTERPOISE CONFIGURATIONS FOR TRANSMISSION LINE GROUNDING 881

Fig. 7. Grounding resistance R for any soil resistivity ρ of four radial coun- Fig. 9. Influence of h on R/ρ of configurations in Fig. 6 (simulation results).
terpoises alone (C1) and with footing groundings (C2).

TABLE II
PARAMETER VALUES IN FORMULAS FOR RESISTIVITY (8) AND EFFECTIVE
LENGTH (7) OF COUNTERPOISE CONFIGURATIONS IN FIG. 6

Therefore, it is possible to derive a single formula that approx-

imates R of a particular configuration for all considered distances
between footings and parallel counterpoises (see Table I):
Fig. 8. Influence of b on R/ρ of configurations GS1, GS2, GS3, and GS4 in
Fig. 6 [simulation results and formula (8)].  
R = ρ A + BC (8)

where values of A, B, and C are given in Table II for all

becomes negligible concerning the resistance of longer counter- configurations in Fig. 6. Fig. 8 shows the simulation results and
poises. Therefore, analysis of long counterpoises alone will lead (8). The error of (8) considering simulation results is smaller
to results relevant to the complete tower grounding system. than 10%.
It is well known that the burial depth and the radius of the Figs. 8 and 9 also show that 8-leg counterpoises, GS3, and
counterpoises have a minor influence on R [26]. GS6, significantly reduce R compared to 4-leg counterpoises,
Fig. 8 shows that the distance between the tower’s footings, GS2 and GS5. The decrease is about 30% for any length .
b, influences the R values, especially for configurations with However, the reduction of R by 12-leg counterpoises, GS4 and
multiple counterpoises (GS3 and GS4). The differences of R are GS7, in comparison to GS3 and GS6, is smaller, about 10%.
within 10%. Results in Fig. 8 are for the radial configurations, The following are more accurate formulas of R that, besides
but the conclusions are similar for the others. ρ and , consider values of conductor radius a, depth of burial d,
The simulation results of R in Fig. 9 show minor differences and distance between tower footings b. The method for deriving
between radial and bent counterpoises parallel to the line when the formulas is described in [40].
the distance between parallel counterpoises, h, is greater or equal  
to 10 m. The examples in Fig. 9 are for the distance between ρ 2 2 b b
GS1 : R = ln +ln −1.5428+0.1319 −0.2779 ln
tower footings b = 5 m, but the conclusions are the same for 4π a d  
other distances. (9)

Authorized licensed use limited to: UNIV ESTADUAL PAULISTA JULIO DE MESQUITA FILHO. Downloaded on October 01,2023 at [Link] UTC from IEEE Xplore. Restrictions apply.
 882 IEEE TRANSACTIONS ON POWER DELIVERY, VOL. 38, NO. 2, APRIL 2023

Fig. 10. Simulation results of Z and formula (16) that fits the simulation results Fig. 11. Simulation results of Z and formula (16) that fits the simulation results
for GS1, GS2, GS3, and GS4 configurations (see Fig. 6). for GS5, GS6, and GS7 configurations (see Fig. 6).

 
ρ 2  b b
GS2 : R = ln +ln +0.2647−0.5965 −0.6719 ln
8π a d  
(10)
 
ρ 2  b b
GS3 : R = ln +ln +3.1649+0.3036 −2.3945 ln
16π a d  
(11)
 
ρ 2  b b
GS4 : R = ln +ln +6.6469+0.3767 −3.4 ln
24π a d  
(12)
 
ρ 2  b b
GS5 : R = ln +ln −0.8054+0.4758 −1.3536 ln
8π a d  
(13)
  Fig. 12. Effective lengths of counterpoise configurations in Fig. 6.
ρ 2  b b
GS6 : R = ln +ln +0.1619+0.243 −1.4939ln
12.5π a d  
(14)
  Similar to the results for R in the previous section, Figs. 10 and
ρ 2  b b
GS7 : R = ln +ln +2.0907+2.622 −3.7192ln 11 show that 8-leg counterpoises, GS3 and GS6, considerably
19.9π a d   reduce Z compared to 4-leg configurations, GS2 and GS5, but
(15) 12-leg counterpoises, GS4 and GS7, have a much smaller effect.
Figs. 10 and 11 show results of (16) that uses (8) to approxi-
VIII. IMPULSE IMPEDANCE AND EFFECTIVE LENGTH mate Z. The error of the formula (16) considering the simulation
Applying (5), (6), and (8), Z is expressed as: results for GS5 in Fig. 11 is smaller than 5% in 97% of the test
   C  cases. Errors of other configurations are similar. However, more
Z = ρ A + B D ρT1 ,  > ef f (16) accurate (9)–(15) can be used instead of (8) to approximate Z
considering conductor radius, a, distance between footings, b,
where values of A, B, C, and D are given in Table II. The constant and depth, d.
D is determined by the best fit of (16) to simulation results Fig. 12 shows the effective length of all counterpoise config-
of Z (see Figs. 10 and 11). The simulation results are derived urations in Fig. 6 for any values of ρ and T1 computed by (7)
from time responses of 11386 test cases for each configuration with values of D in the last column of Table II. Configurations
with parameters in wide ranges (ρ from 30 to 2000 Ωm, T1 with multiple counterpoises have smaller inductive components
from 0.2 to 10 μs, and  from 1 to 100 m). Figs. 10 and 11 and correspondingly larger ef f (and larger effective areas,
illustrate a remarkable outcome that simulation results of Z/ρ see Fig. 5). Configurations with the same number of legs bent
as a function of ρT1 converge to a generalized curve, enabling parallel to the line have smaller effective lengths than radial
a single approximating function (16). ones.

Authorized licensed use limited to: UNIV ESTADUAL PAULISTA JULIO DE MESQUITA FILHO. Downloaded on October 01,2023 at [Link] UTC from IEEE Xplore. Restrictions apply.
 GRCEV et al.: LIGHTNING EFFICIENT COUNTERPOISE CONFIGURATIONS FOR TRANSMISSION LINE GROUNDING 883

Fig. 13. Simulation results of Z/R for 11386 test cases with parameters in
ranges ( from 1 to 100 m; ρ from 30 to 2000 Ωm; T1 from 0.2 to 10 μs) for
configuration GS5.

IX. RANGES OF PARAMETERS FOR CONDUCTIVE AND

INDUCTIVE BEHAVIOR
Fig. 13 shows all simulation results of Z/R that are derived
from time responses of 11386 test cases for the configuration
GS5 (see Fig. 6). Results for other configurations are similar.
The results enable approximating ranges of parameters when
grounding systems exhibit dominantly resistive (Z/R ≈ 1) or
inductive behavior (Z/R > 1):
ρT1 / < 1, inductive effects
ρT1 / > 100, resistive effects
1 < ρT1 / < 100 and  < ef f , resistive effects
1 < ρT1 / < 100 and  > ef f , inductive effects

Visacro et al. [45] state that Z/R “is typically lower than 1
for electrode lengths shorter than ef f , due to the effects of
capacitive currents in the soil.” Our computations do not show
any significant impact of the capacitive effect on the values
of Z/R, as shown in Fig. 13. The dominantly capacitive effect
is typical for less extended electrodes in high resistivity soil.
For example, in our computations, the voltage impulse lags
the current, and its front time becomes longer in the case of
dominantly capacitive behavior. Still, the voltage peak value Vm
is not significantly reduced, which results in Z/R ≈ 1.

X. COMPARISON BETWEEN CONFIGURATIONS

Fig. 14. Required counterpoises lengths to get resistivity of 10 Ω, 15 Ω,
The length of counterpoises  for obtaining the required value and 20 Ω, and the effective lengths for typical first and subsequent strokes for
configurations: (a) GS1, (b) GS2, (c) GS5, (d) GS3, (e) GS6, (f) GS4, (g) GS7
of resistance R for a given resistivity of soil ρ can be determined (see Fig. 6).
from (8):
1
R/ρ − A C
= (17) typical first and subsequent stroke, i.e., 2.4 μs and 0.35 μs [43].
B To maintain Z ≈ R, it is necessary that  < ef f .
This formula is used for comparison between configurations. Fig. 14 illustrates the capabilities and limitations of the con-
Fig. 14 shows counterpoises length  necessary to get R values sidered configurations. The following conclusions can be drawn.
of 10 Ω, 15 Ω, and 20 Ω, as a function of the soil resistivity. r The length  of the radial counterpoises, GS2 in Fig. 14(b),
Fig. 14 also shows the effective lengths ef f for values of T1 of GS3 in Fig. 14(d), and GS4 in Fig. 14(f), is limited to

Authorized licensed use limited to: UNIV ESTADUAL PAULISTA JULIO DE MESQUITA FILHO. Downloaded on October 01,2023 at [Link] UTC from IEEE Xplore. Restrictions apply.
 884 IEEE TRANSACTIONS ON POWER DELIVERY, VOL. 38, NO. 2, APRIL 2023

TABLE III
RANGES OF SOIL RESISTIVITY AND THE REQUIRED LENGTHS OF
COUNTERPOISE CONFIGURATIONS FOR R ≤ 10 Ω

about 35 m [41], which constrains their use. The other

configurations within the right of way may use longer
counterpoises, which may be limited by their effective
lengths. Fig. 15. Conversion of the simulation results of Z computed with IEC [34]
r More legs in 2-, 4-, 8-, and 12-leg counterpoises, GS1 in and CIGRE [33] mathematical representations of current waveforms of the first
return stroke (see Fig. 1). Simulations results are for GS5 configuration with b
Fig. 14(a), GS2 (b) and GS5 (c), GS3 (d) and GS6 (e), and = 6 m and h = 20 m.
GS4 (f), extend their use in higher resistive soils. However,
12-leg counterpoise, GS7 in (g), is less effective.
r The efficiency of counterpoises bent parallel to the line
(GS5 and GS6) whose distance between the parallel coun-
terpoises is equal to or larger than 20 m is like the efficiency
of radial ones (GS2 and GS3). Distances between counter-
poises of GS7 are smaller, i.e., 10 and 15 m, and GS7 is
considerably less efficient.
r The area covered by the considered configurations does
not significantly impact the grounding system efficiency.
That can be seen by comparison of the radial and bent
configurations with the same number of legs, which cover
different areas, i.e., 4-legs in Fig. 14(b) and (c), 8-legs in
(d) and (e), and 12-legs in (f) and (g). The conductor length
has the dominant impact on the efficiency.
Table III shows another example of the simple formulas use
Fig. 16. The maximal steepness of the waveforms is similar when the CIGRE
for comparison of the capabilities and limitations of the config- and IEC current waveforms are adjusted to lead to nearly identical Z.
urations. Table III gives lengths of counterpoise configuration
in soil with different resistivity that are necessary to achieve R
values of about 10 Ω. Values for GS2, GS3, and GS4 in Table III Fig. 16 shows that when the CIGRE and IEC current wave-
are adapted from EdF recommendations [41]. Connecting more forms are adjusted to lead to nearly identical results of Z, the
counterpoises to the tower footing enables a larger conductor maximal steepness of the waveforms is similar. This implies that
total length, resulting in lower R values in more resistive soils. the maximal steepness of these two current waveforms is one of
The example in Table III shows that GS6 and GS7 enable the dominant factors that determines Z.
R ≤ 10 Ω in soils with 1600 Ωm and 2100 Ωm. Fig. 17 shows that the double exponential mathematical
representation of the current waveform leads to very sim-
XI. INFLUENCE OF CURRENT WAVEFORM ilar results as the IEC waveform, and no adjustments are
necessary.
Fig. 15 shows simulation results of Z derived from time re-
sponses to two different mathematical representations of current
waveforms of return strokes, i.e., used in IEC standards [34] and XII. DISTRIBUTION OF LIGHTNING CURRENT BETWEEN
recommended by CIGRE [33] (see Section II.B). It is shown that LOCAL GROUNDING AND COUNTERPOISES
results can be easily converted by modifying the impulse front
Fig. 18 shows an example of lightning current distribution
time T1 :
between the local grounding and counterpoises for typical first
(CIGRE) (IEC) and subsequent return stroke current impulses. The considered
T1(30−90%) = 1.82 · T1(10−90%) (18)
grounding system is configuration GS5 (see Fig. 6) with param-
eters shown in Fig. 18. The current dissipated through the local
Note that definitions of the impulse front times are according grounding and counterpoises is insufficient to have a significant
to Fig. 1. soil ionization effect.
Authorized licensed use limited to: UNIV ESTADUAL PAULISTA JULIO DE MESQUITA FILHO. Downloaded on October 01,2023 at [Link] UTC from IEEE Xplore. Restrictions apply.
 GRCEV et al.: LIGHTNING EFFICIENT COUNTERPOISE CONFIGURATIONS FOR TRANSMISSION LINE GROUNDING 885

Fig. 20. (a) Comparison of effective length computed by formulas based on

different definitions (labels are defined in Fig. 19). (b) Comparison between (7)
and formulas in [45].

such definitions on an example of the grounding system GS5

Fig. 17. The double exponential mathematical representation of the current [see Fig. 6(b)].
waveform leads to very similar results as the IEC waveform.
Simulation results in Fig. 19 show that Z ≈ R for coun-
terpoise lengths smaller than a length denoted by ef f (1) .
Z becomes larger than R for counterpoise lengths larger
than ef f (1) , and, at more extensive lengths, converges to
a constant value while R continuously decreases. Follow-
ing quantitative definitions of effective length exist in the
literature.
1) Gupta and Thapar [44], [29]: The length of an electrode
when Z decreases to a value within 3% of the final value
(denoted as ef f (2) in Fig. 19).
2) He et al. [9]: The length of a grounding electrode at which
the derivative of Z with respect to  is smaller than 50
Fig. 18. The lightning current distribution between the local grounding and (denoted as ef f (3) in Fig. 19).
counterpoises for typical first and subsequent stroke current impulse.
3) Grcev [43]: Maximal length of the ground electrode for
which Z ≈ R, and above which Z > R (denoted as ef f (1)
in Fig. 19).
4) Grcev et al. [40]: The length for which R is equal to the Z
final value (denoted as ef f in Fig. 19).
This paper uses the latter definition, denoted as ef f in Fig. 19.
It enables the approximation of Z in (5) and (6) (represented by a
full line in Fig. 19). (5) and (6) show that such an approximation
leads to an outcome that only two quantities, i.e., resistivity R ex-
pressed as a function of , and effective length, ef f , are required
to define the lightning performance of considered grounding
systems.
Fig. 20(a) shows that the different definitions of the effective
length lead to different results (the example is for the four-arm
star configuration, shown in Fig. 20(a), for which corresponding
formulas exist in the literature).
Fig. 19. Simulation results of R and Z (broken lines); approximation of Z
[20] (full line); and different quantitative definitions of the effective length in
Fig. 20(b) shows a comparison of (7) with a formula pub-
literature: ef f (1) [43], eff [40], ef f (2) [29], [44], ef f (3) [9]. lished in [45]. The formulas in [45] differ significantly from
this paper. They are specialized for one counterpoise config-
uration, i.e., GS5, and for two waveforms of the lightning
XIII. COMPARISON WITH PREVIOUS WORK current impulse, one for the first return stroke and the other
for the subsequent stroke. The first one is a waveform with
A. Effective Length
two peaks with T30-90% = 3.83 microseconds, represented
Effective length is defined concerning the change of Z when by seven Heidler functions. The second is represented by
the length of electrodes enlarges. However, there are different two Heidler functions with T30-90% = 0.67 microseconds. In
definitions of effective length in the literature. Fig. 19 illustrates Fig. 20(b), T1 for the first stroke is adapted to 3.83/1.82 =

Authorized licensed use limited to: UNIV ESTADUAL PAULISTA JULIO DE MESQUITA FILHO. Downloaded on October 01,2023 at [Link] UTC from IEEE Xplore. Restrictions apply.
 886 IEEE TRANSACTIONS ON POWER DELIVERY, VOL. 38, NO. 2, APRIL 2023

TABLE IV
COMPUTATION RESULTS (ρ = 1000Ωm AND R = 10Ω)

Fig. 21. Errors of formulas (9)–(15) and Chisholm formula [46] for each
considered counterpoise configuration (see Fig. 6).

2.1 microseconds (since the two-peak waveform has a similar

front to the CIGRE model). Results are similar for the subse-
quent stroke but diverge for the first stroke in high resistivity
soil.

B. Low-Frequency Resistance
Chisholm’s formula of R [46] is applicable for any grounding
system form, from single rods to grids. Fig. 21 shows errors in
this general formula and formulas (9)–(15). The reference values
for evaluation errors are simulations by an accurate computer Fig. 22. Voltages developed at the counterpoise configuration GS6 in case of
model [47]. The error of formulas (9)–(15) is less than 6%, typical first stroke lightning current impulse.
while Chisholm [46] goes up to 40%.

XIV. NUMERICAL EXAMPLE

In this example, we look at the choice of the counterpoise Columns 7 and 8: Voltage peaks for the first and subsequent
configuration with R = 10 Ω in the soil with ρ = 1000 Ωm, return strokes, VmFS and VmSS , are computed by (4).
which requires the smallest total conductor length and develops 8-leg counterpoise GS6 might be chosen for the optimal
a minimal voltage peak. The typical lightning first return stroke’s configuration with  = 44 m since the total length of 352 m
current peak and front time are Im = 30 kA and T1 = 2.4 is the smallest and VmFS = 300 kV is minimal while VmSS <
μs, and of the subsequent return stroke Im = 12 kA and T1 = VmFS . Note that EdF recommends using GS4 in this case [41],
0.35 μs [43]. but it requires a total conductor length of 420 m.
The results of computations by the developed formulas are in Since the formulas are approximate, we check the correctness
Table IV. The following steps lead to these results. of this solution by computing the developed voltage in the case
Columns 1 and 2: For each configuration, length  and total of the first return stroke with the rigorous electromagnetic model
conductor length tot are computed by (17) using values of A, B, [20]. Fig. 22 shows the simulation results. The distance between
and C in Table II and tot = n, where n is the number of legs the tower footings, b, is 5 or 10 m, and between the parallel
in the configuration. counterpoises, h, is 10 or 20 m. Although there are differences of
Columns 3 and 4: Effective lengths for the first and subsequent about 10%, it can be seen that voltages in all cases are consistent
return strokes, ef f F S and ef f SS , are computed by (7) using with the requirement of VmFS = 300 kV.
values of D from Table II. Note that the numerical example results in this Section depend
Columns 5 and 6: Impulse impedances for the first and subse- on input data and do not imply that GS6 is optimal in all cases.
quent return strokes, ZFS and ZSS , are computed by (16) using The trends can be studied in Fig. 14. A comparison can be made
values of A, B, C, and D from Table II. More accurate values by generating results such as in Table IV for given values of R,
can be computed by (9)–(15) with  = ef f F S and ef f SS ρ, and Im and T1 of the first and subsequent return stroke current
computed by (7) and values of D from Table II. (represented by CIGRE, IEC or double-exponential waveforms).

Authorized licensed use limited to: UNIV ESTADUAL PAULISTA JULIO DE MESQUITA FILHO. Downloaded on October 01,2023 at [Link] UTC from IEEE Xplore. Restrictions apply.
 GRCEV et al.: LIGHTNING EFFICIENT COUNTERPOISE CONFIGURATIONS FOR TRANSMISSION LINE GROUNDING 887

XV. CONCLUSION [7] O. Kherif, S. Chiheb, M. Teguar, A. Mekhaldi, and N. Harid, “Investiga-
tion of horizontal ground electrode’s effective length under impulse cur-
1) Radial 4-, 8- and 12-leg counterpoises are limited to rent,” IEEE Trans. Electromagn. Compat., vol. 61, no. 5, pp. 1515–1523,
about 35 m because of practical considerations. 2- and Oct. 2019.
[8] S. Sekioka and T. Funabashi, “A study on effective length for practical
4-leg counterpoises bent parallel to the line have no such design of grounding system in a wind turbine,” in Proc. Int. Conf. Lightning
limitation, but their effective length constrains the use in Protection, Cagliari, Italy, 2010, pp. 1–6.
high resistivity soil. [9] J. He et al., “Effective length of counterpoise wire under lightning current,”
IEEE Trans. Power Del., vol. 20, no. 2, pp. 1585–1591, Apr. 2005.
2) This paper proposes an 8-leg counterpoise bent parallel [10] J. He, R. Zeng, and B. Zhang, Methodology and Technology for Power
to the line. This counterpoise enlarges the effective length System Grounding. New York, NY, USA: Wiley, 2013.
and is efficient in high resistive soil. A 12-leg counter- [11] IEEE Guide for Improving the Lightning Performance of Transmission
Lines, IEEE Standard 1243-1997, Dec. 1997.
poise bent parallel to the line is less effective. [12] W. Chisholm et al., “Impact of soil-parameter frequency dependence on
3) The influence of the mutual resistances between the the response of grounding electrodes and on the lightning performance of
parallel counterpoises is negligible for 20 m or higher electrical systems,” CIGRE, Paris, France, Tech. Brochure 781, 2019.
[13] CIGRE, “Procedures for estimating the lightning performance of trans-
distances. Counterpoises as close as 10 m can also be mission lines – new aspects,” CIGRE, Paris, France, Tech. Brochure 839,
used, leading to about 10% higher peak voltages. 2021.
4) A complete set of new simple formulas of resistance, [14] “Tower grounding and soil ionization report,” EPRI, Palo Alto, CA, Tech.
Rep. 1001908, 2002.
impulse impedance, and effective length of all considered [15] “Handbook for improving overhead transmission line lightning perfor-
counterpoise configurations is derived. mance,” EPRI, Palo Alto, CA, USA, Rep. 1002019., 2002.
5) Remarkably, simulation results of Z/ρ as a function of [16] “Guide for transmission line grounding: A roadmap for design, testing,
and remediation: Part I—Theory book,” EPRI, Palo Alto, CA, USA, Tech.
ρT1 converge to a generalized curve, enabling a single Rep. 1013900., 2007.
approximating function. [17] P. Chowdhuri, Electromagnetic Transients in Power Systems. Hoboken,
6) Ranges of parameters for which behavior is dominantly NJ, USA: Wiley, 1996.
[18] L. Grcev and Z. Haznadar, “A novel technique of numerical modelling
resistive or inductive are determined. of impulse current distribution in grounding systems,” in Proc. Int. Conf.
7) It is shown that impulse impedance results got for the Lightning Protection, Graz, Austria, 1988, pp. 165–169.
different mathematical representations of the lightning [19] L. Grcev and F. Dawalibi, “An electromagnetic model for transients in
grounding systems,” IEEE Trans. Power Del., vol. 5, no. 4, pp. 1773–1781,
current impulses such as Heidler, CIGRE, and double- Oct. 1990.
exponential functions, can be easily mutually converted. [20] B. Markovski, L. Grcev, and V. Arnautovski-Toseva, “Fast and accurate
8) Different definitions of the effective length in the litera- transient analysis of large grounding systems in multilayer soil,” IEEE
Trans. Power Del., vol. 36, no. 2, pp. 598–606, Apr. 2021.
ture lead to different results. The definition used in this [21] R. G. Olsen and M. C. Willis, “A comparison of exact and quasi-static
paper is consistent with a large number of simulation methods for evaluating grounding systems at high frequencies,” IEEE
results. Trans. Power Del., vol. 11, no. 3, pp. 1071–1081, Apr. 1996.
[22] L. Grcev and V. Arnautovski, “Comparison between simulation and mea-
9) New formulas and procedures make up a general and surement of frequency dependent and transient characteristics of power
straightforward method for analyzing the optimal design transmission line grounding,” in Proc. Int. Conf. Lightning Protection,
of counterpoises as part of transmission line grounding, Birmingham, U.K., 1998, vol. 1, pp. 524–529.
[23] L. Grcev, “Computer analysis of transient voltages in large grounding
considering the cost-effective and best protection against systems,” IEEE Trans. Power Del., vol. 11, no. 2, pp. 815–823, Apr. 1996.
lightning. [24] L. Grcev, “Time- and frequency-dependent lightning surge characteris-
10) The underlying simplifications of the formulas are con- tics of grounding electrodes,” IEEE Trans. Power Del., vol. 24, no. 4,
pp. 2186–2196, Oct. 2009.
servative and can be used in the initial grounding design [25] R. G. Olsen and L. Grcev, “Analysis of high-frequency grounds: Com-
phase. If more detailed data on soil characteristics from parison of theory and experiment,” IEEE Trans. Ind. Appl., vol. 51, no. 6,
measurements on the actual site is available, more elab- pp. 4889–4899, Nov./Dec. 2015.
[26] G. F. Tagg, Earth Resistances. London, U.K.: Newnes, 1964.
orate models and dedicated software, e.g., [48], might [27] IEEE Guide for Measuring Earth Resistivity, Ground Impedance, and
allow for more accurate computations. Earth Surface Potentials of a Grounding System, IEEE Standard 81, IEEE,
Manhattan, NY, USA, 2012.
[28] P. Dawalibi and R. D. Southey, “On the equivalence of uniform and
two-layer soils to multilayer soils in the analysis of grounding sys-
REFERENCES tems,” IEE Proc. Gener. Transmiss. Distrib., vol. 143, no. 1, pp. 49–55,
Jan. 1996.
[1] W. W. Lewis, The Protection of Transmission Systems Against Lightning. [29] B. R. Gupta and B. Thapar, “Impulse impedance of grounding grids,” IEEE
New York, NY, USA: Dover, 1965. Trans. Power App. Syst., vol. PAS-99, no. 6, pp. 2357–2362, Nov. 1980.
[2] A. R. Hileman, Insulation Coordination for Power Systems. Boca Raton, [30] L. Grcev, “High-frequency grounding,” in Lightning Protection, V.
FL, USA: CRC, 1999. Cooray, Ed., London, U.K.: The Inst. Eng. Technol., 2009, pp. 503–527.
[3] F. Kiessling, P. Nefzger, J. F. Nolasco, and U. Kaintzyk, Overhead Power [31] D. Cavka, N. Mora, and F. Rachidi, “A comparison of frequency-
Lines. Berlin, Germany: Springer, 2003. dependent soil models: Application to the analysis of grounding sys-
[4] Y. Liu, N. Theethayi, and R. Thottappillil, “Investigating the validity of tems,” IEEE Trans. Electromagn. Compat., vol. 56, no. 1, pp. 177–187,
existing definitions and empirical equations of effective length/area of Feb. 2014.
grounding wire/grid for transient studies,” J. Electrostatics, vol. 65, no. 5/6, [32] M. Nazari, R. Moini, S. Fortin, F. P. Dawalibi, and F. Rachidi, “Impact of
pp. 329–335, May 2007. frequency-dependent soil models on grounding system performance for
[5] S. Miyamoto et al., “Effective length of vertical grounding wires connected direct and indirect lightning strikes,” IEEE Trans. Electromagn. Compat.,
to wind turbine foundation,” J. Int. Council Elect. Eng., vol. 7, no. 1, vol. 63, no. 1, pp. 134–144, Feb. 2021.
pp. 89–95, May 2017. [33] Guide to Procedures for Estimating the Lightning Performance of Trans-
[6] K. Yamamoto, S. Sumi, S. Sekioka, and J. He, “Derivations of effec- mission Lines, Working Group 01 (Lightning) of Study Committee 33,
tive length formula of vertical grounding rods and horizontal ground- CIGRE, Paris, France, Tech. Brochures 63, 1991.
ing electrodes based on physical phenomena of lightning surge prop- [34] Protection Against Lightning - Part 1: General Principles, International
agations,” IEEE Trans. Ind. Appl., vol. 51, no. 6, pp. 4034–4042, Electrotechnical Commission IEC 62305-1, Geneva, Switzerland, Int. Std.
Nov./Dec. 2015. IEC 62305-1, Dec. 2010.
Authorized licensed use limited to: UNIV ESTADUAL PAULISTA JULIO DE MESQUITA FILHO. Downloaded on October 01,2023 at [Link] UTC from IEEE Xplore. Restrictions apply.
 888 IEEE TRANSACTIONS ON POWER DELIVERY, VOL. 38, NO. 2, APRIL 2023

[35] F. Heidler, J. M. Cvetic, and B. V. Stanic, “Calculation of lightning [42] F. M. Gatta, A. Geri, S. Lauria, and M. Maccioni, “Simplified HV tower
current parameters,” IEEE Trans. Power Del., vol. 14, no. 2, pp. 399–404, grounding system model for backflashover simulation,” Elect. Power Syst.
Apr. 1999. Res., vol. 85, pp. 16–23, Apr. 2012.
[36] IEEE Guide for Safety in AC Substation Grounding, IEEE Std. 80-2013, [43] L. Grcev, “Impulse efficiency of ground electrodes,” IEEE Trans. Power
IEEE, Manhattan, NY, USA, Dec. 2013. Del., vol. 24, no. 1, pp. 441–451, Jan. 2009.
[37] S. Visacro, “The use of the impulse impedance as a concise repre- [44] B. R. Gupta and B. Thapar, “Impulse impedance of grounding sys-
sentation of grounding electrodes in lightning protection applications,” tems,” IEEE Trans. Power App. Syst., vol. PAS-99, no. 6, pp. 2357–2362,
IEEE Trans. Electromagn. Compat., vol. 60, no. 5, pp. 1602–1605, Nov. 1980.
Oct. 2018. [45] S. Visacro and F. H. Silveira, “Lightning performance of transmission
[38] R. Fieux, P. Kouteynikoff, and F. Villefranque, “Measurement of the lines: Requirements of tower-footing electrodes consisting of long coun-
impulse response of groundings to lightning currents,” in Proc. Int. Conf. terpoise wires,” IEEE Trans. Power Del., vol. 31, no. 4, pp. 1524–1532,
Lightning Protection, Paper 4.3, pp. K4:40–K4:55, Uppsala, Sweden, Aug. 2016.
1979. [46] W. A. Chisholm, “Evaluation of simple models for the resistance of
[39] L. V. Bewley, Traveling Waves on Transmission Lines. New York, NY, solid and wire-frame electrodes,” IEEE Trans. Ind. Appl., vol. 51, no. 6,
USA: Wiley, 1951. pp. 5123–5129, Nov./Dec. 2015.
[40] L. Grcev, B. Markovski, and M. Todorovski, “Lightning performance of [47] R. P. Nagar, R. Velazquez, M. Loeloeian, D. Mukhedkar, and Y. Gervdis,
multiple horizontal, vertical and inclined grounding electrodes,” IEEE “Review of analytical methods for calculating the performance of large
Trans. Power Del., early access, Dec. 21, 2021, doi: 10.1109/TP- grounding electrodes part 1: Theoretical considerations,” IEEE Trans.
WRD.2021.3137361. Power App. Syst., vol. PAS-104, no. 11, pp. 3123–3133, Nov. 1985.
[41] Guide D’application De La Note H115 – Principes de Mise à La Terre [48] “TRAGSYS-software for high frequency and transient analysis of ground-
Des Ouvrages Du Service du Transport et Des Télécommunications, Paris, ing systems,” version. 2.0 2005. [Online]. Available: [Link]
France: Électricité de, Sep. 1991. com

Authorized licensed use limited to: UNIV ESTADUAL PAULISTA JULIO DE MESQUITA FILHO. Downloaded on October 01,2023 at [Link] UTC from IEEE Xplore. Restrictions apply.