Journal of Electrostatics 72 (2014) 381e386

Contents lists available at ScienceDirect

Journal of Electrostatics
journal homepage: www.elsevier.com/locate/elstat

Transient lightning responses of grounding systems buried in

horizontal multilayered earth with a hybrid method
Zhong-Xin Li*, Ke-Li Gao, Yu Yin, Dong Ge
The China Electric Power Research Institute, No 15, East Road, Xiaoying, Qinghe, Haidian, Beijing, China

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

Article history: Combined with the Fast Fourier Transform (FFT), a mathematical hybrid method for accurately
Received 2 April 2013 computing the lightning response from grounding systems buried in multilayered earth model has been
Received in revised form developed in this paper. In the method, electrical circuit consists of “T” typical of basic elements. To
18 June 2014
accelerate calculations of the method, quasi-static complex image method and closed form of Green's
Accepted 27 June 2014
function and analytical formula for mutual induction and impedance coefficients were introduced into
Available online 14 July 2014
this method. With the inverse FFT, the method can be used for studying performances of transient
lightning response from the grounding systems.
Keywords:
Closed form of Green's function
© 2014 Elsevier B.V. All rights reserved.
Grounding systems
Nodal analysis
Horizontal multilayered earth model
Quasi-static complex image method

1. Introduction the grounding systems buried in infinite half homogenous earth

Recently, the computerized analysis about the hybrid method
basing on quasi-static electromagnetic field theory has been
developed [1e10]. The hybrid method can be used to study fre-
quency behavior and lightning response of grounding systems.
The hybrid method has been developed out of conventional
nodal analysis method [11]. The papers [1e6] were developed to
analyze the behavior of grounding systems in frequency domain.
And the methods in the papers [8,9] were used to study the
lightning response of grounding systems. Among these papers, the
papers [1e3] discussed the behavior of grounding systems in the
frequency domain buried in only half infinite homogenous earth.
Although Refs. [12,13] has developed the hybrid method basing on
high frequency electromagnetic field theory, only half infinite ho-
mogenous earth model is considered. These confines in the fre-
quency domain promoted a novel mathematical model to be
developed in [4e7], with quasi-static complex image method
(QSCIM) introduced, which can be used to fast calculate the cur-
rents distribution in the grounding systems buried in horizontal
multilayered earth model in the frequency domain. The hybrid
method can be combined with FFT, and the transient response from
the grounding systems can be achieved. The lightning response of
* Corresponding author. Tel.: þ86 013693134345; fax: þ8601082812256.
E-mail addresses: zxli@fudan.edu.cn, lizx@epri.sgcc.com.cn (Z.-X. Li).
has been developed in [8], and the lightning response of the
grounding systems buried in horizontal multilayered earth has
been discussed in [9].
Basic elements of the electrical circuit used by the hybrid
method are “p” typical of basic elements (see (a) in Fig. 1) for all
these papers [1e9]. However, the typical of basic elements can be
further displaced with “T” typical of basic elements (see (b) in
Fig. 1), which own better physical meanings. The hybrid method
with “T” typical of basic elements has been developed in [10]. The
hybrid method is basing on direct current electrical field theory and
the behavior of the grounding systems buried in half infinite ho-
mogenous earth. The hybrid method can be further developed
basing on quasi-static electromagnetic field theory to study the
behavior of the grounding systems buried in horizontal multilay-
ered earth in the frequency domain, and analyze the lightning
response of the grounding systems buried in horizontal multilay-
ered earth through the inverse FFT. It should be pointed out that the
method is similar with the PEEC method, however, the PEEC
method is developed to calculate the transient response to micro-
wave integrated circuits in air space; Recently, the PEEC method has
been developed to calculate the transient lightning response to the
grounding systems [14], however, only uniform earth model has
been considered.
In this paper, basing on previous work in [1e10], combined with
the FFT, an accurate mathematical hybrid method with “T” typical
of basic elements is developed for calculating the harmonic wave

http://dx.doi.org/10.1016/j.elstat.2014.06.009
0304-3886/© 2014 Elsevier B.V. All rights reserved.
382 Z.-X. Li et al. / Journal of Electrostatics 72 (2014) 381e386

So we have
2 3
  t t  
F
¼ 4 Apb Y bb Abp Apb Y bb Abb 5 Vp
(2)
t t
Is Abb Ybb Abp A Y A Vb
bb bb bb

where ½F is a Np  1 vector of external currents sources; ½Y bb  is the

Nb  Nb branch admittance sub-matrix of the matrix ½Y b ; ½Is  is a
Nb  1 vector of leakage currents; Vp is a Np  1 vector of SEP of the
end points; Vb is a Nb  1 vector of SEP of the middle points; The
Fig. 1. “T” typical of basic element. sub-matrices ½App , ½Apb  and ½Abb  are, respectively, Np  Np, Np  Nb
and Nb  Nb matrices.
For interactions of mutual impedance among these discrete
currents from lightning currents distribution along the grounding conductors, only Nb pieces of conductors should be considered,
systems buried in multilayered earth model. The earth is consid- which means each conductor do not require to be separated into
ered as horizontal multilayered earth model. To accelerate the two parts due to its middle point, so we have
calculation, QSCIM and closed form of Green's function were
introduced, and the mutual inductive and conductive coefficients



Vb ¼ Zs Is (3)
have analytical formulas so as to avoid the numerical integral.
where ½Zs  is a Nb  Nb mutual impedance matrix, which give a
2. Mathematical model of equivalent circuits with “T” typical matrix relationship between SEP of middle point of conductors ½Vb 
of basic elements of the grounding systems in frequency and leakage currents ½Is  through the rapid Galerkin's Moment
domain method [16].
Combining Eqs. (2) and (3), we have
The transient problem is first solved by a formulation in the 2 3
frequency domain. The time-domain response is then obtained by   t t  
Apb Y bb Abp Apb Y bb Abb
applying a suitable Fourier inversion technique.
F
¼4 5 Vp (4)
0 t t 1
A Y A Abb Ybb Abb þ Zs Vb
First, a grounding system is divided in Nb pieces of segments bb bb bp
that can be studied as an elemental unit, where the discrete
At last we have
grounding system has Np end nodes and Nb middle nodes.
The grounding network is energized by injection of single fre-



quency currents at one or more nodes. In general, we assume that a F ¼ Y b Vp (5)
sinusoidal current source of value Fi is connected at ith node. Scalar
electric potential (SEP) Vj of jth (i,j ¼ 1,2,…,Np) node on the
 h t i
grounding network referring to the infinite remote earth as zero Y b ¼ Apb Y bb Abp
h t h
(6)
SEP is defined. Meanwhile, we define a branch current Ibk , a branch t 1 i1 t i
voltage Ubk and a leakage current Isk on kth (k ¼ 1,2,…,Nb) segment.  Apb Ybb Abb Abb Y bb Abb þ Zs Abb Y bb Abp
With the above considerations and “T” typical of basic elements
(see (b) in Fig. 1) is used, the obtained electric circuits may be The vector of nodal SEP ½Vp  may be calculated through solving
studied using the conventional nodal analysis method [11]. For the Eq. (5). The SEP of middle point ½Vb  can be calculated by
interactions of mutual induction among these discrete conductors, h

 t 1 i1 h t i

2Nb pieces of conductors must be considered, which means each Vb ¼  Abb Y bb Abb þ Zs Abb Ybb Abp Vp (7)
conductor has been separated into two parts due to its middle
point. Then we have The leakage currents ½Is , and the branch currents ½Ib  can also be



 calculated [4e7].
Ib ¼ Yb Ud (1) Once branch and leakage currents are known, the electromag-
netic field quantities like as SEP 4, A, E and B at any point can be
where ½Yb  is a 2Nb  2Nb admittance matrix who includes mutual calculated, which can be referred in papers [4e7].
induction and internal impedance of conductors; ½Ud  is 2Nb  1 The study of the grounding systems performance in the fre-
vector of potential difference between end and middle points of quency domain has been reduced to the computation of ½Zs  and
conductors; ½Ib  is 2Nb  1 vector of axis branch currents of ½Y bb  matrices. From Refs. [4e9], we know that each segment is
conductors. modeled as a lumped resistance and self-inductance. Interactions
Considering an incidence matrix ½A, the incidence matrix is of mutual inductance or impedance between branch segments’
(Nb þ Np)  2Nb matrix and used to relate to branches and nodes branch or leakage currents are also included in the method. For
(including middle and end points), whose elements can be referred matrix ½Y bb  case, the diagonal elements consist of self impedance
and self induction, other elements belong to mutual induction
in [4e7]. Then we know ½A½Ib  ¼ 0 and ½Ud  ¼ ½At ½Vn , here ½Vn  is
between branch currents along a pairs of conductor segments. The
(Nb þ Np)  1 vector of all SEP of nodes including end and middle
" # formula for self impedance and self induction can be referred to
App Apb Refs [4e9]. For matrix ½Zs  case, whose element Zi,j is the mutual
points. Meanwhile, we know ½Vn  ¼ ½Vp =Vb  and ½A ¼ t
Apb Abb impedance coefficient between leakage currents along a pair of
  segments, including conductive and capacitive effects of the earth.
Y bb Y bb It must be pointed out that formula of mutual impedance co-
and ½Yb  ¼
Y bb Y bb efficients is almost same as the one in Refs. [4e9], only different
 Z.-X. Li et al. / Journal of Electrostatics 72 (2014) 381e386 383

parts are length of the conductor segments li ¼ 0.5Li or lj ¼ 0.5Lj, induction or impedance coefficient can be fast analytical calculated,
which can been seen below formula. which is just same as the one for infinite conductive medium case.
Z Z
  dti dtj 4. Verification of our model
Zi;j ¼ G4 r j ; r i (8)
li lj
li lj
In this part, our model has been verified through comparison
Here li ¼ 0.5Li, lj ¼ 0.5Lj (seen in Fig. 1). with experimental data from other published papers. It must be
From Refs. [4e9], we know that for infinite homogeneous con- pointed out that the impulse impedance, an essential parameter in
ductivity medium, formula for both mutual induction and imped- grounding systems design, is defined as the following expression
ance coefficients can be solved analytically. [15]: Zc ¼ U/I.
Note the medium surrounding the point current source here
was considered as homogeneous and infinite medium. However, in 4.1. Verification of our method
any practical case, the earth is represented as horizontal multilay-
ered earth model. The QSCIM can be used to dealt with the hori- The first case, from [17], which gives out some measurement
zontal multilayered earth model, this will be discussed next. results about grounding impedances, the grids is used for mea-
surement consisted of 4  4 mesh with size of 100  100 m. The
3. The closed form of the Green's function of a point source in earth is considered as a multilayered earth model. The earth model
horizontal multilayered earth model and the QSCIM is given as, (a) r1 ¼ 120 Um, r2 ¼ 25 Um, and thickness of the upper
earth is 5 m; (b) r1 ¼ 25 Um, r2 ¼ 120 Um, r3 ¼ 10 Um, thickness of
The another important task of simulations of the grounding the first and second layer earth is 5 m and 100 m; (c) r1 ¼ 25 Um,
systems is calculating the element Zi,j of matrix ½Zs  and the element r2 ¼ 120 Um, r3 ¼ 10 Um, thickness of first and second layer earth is
Mi,j of matrix ½Zbb . When the earth model is considered as the 5 m and 9 m, respectively. The radius of the grids conductors is
horizontal multilayered conductivity media, influences from in- 0.5 cm, and buried at 0.5 m depth in the earth. Here, the conduc-
terfaces will be considered, which will lead to infinite integral tivity of copper conductors is sCu ¼ 5.8  107 S/m, and the
about Bessel function associated with Green's function of a point permittivity of earth mode is set ε1 ¼ 5. The results can be seen in
source (a scalar monopole or a vector dipole). However, the Table 1. From Table 1, we know the impedances from two different
element Zi,j of matrix ½Zs  and the element Mi,j of matrix ½Zbb , which methods are consistent.
includes the Green's function, can also be fast calculated by using The second case from Ref. [15], a typical grounding system
the QSCIM. consists of 2  2 mesh with size of 10  10 m, which was made of
The closed form of a point source in the horizontal multilayered round copper conductors 50 mm2 cross section. The grounding grid
earth model has been developed based on the QSCIM, the proce- was buried with 0.5 m depth in two-layer horizontal earth, whose
dure for the closed form of the Green's function of a point source in ratio of resistivity for the upper and the lower soil layers is r1/
arbitrary horizontal multilayered earth model can be refereed in r2 ¼ 50/20 Um/Um, the upper layer thickness being H ¼ 0.6 m. The
[4e6]. Here, we just directly give out scalar Green's function inject lightning current parameter was set as T1 ¼ 15 ms, T2 ¼ 90 ms
G411 ðr; r 0 Þ for three-layer earth model. and Im ¼ 8.2 A, the feed point is at the corner of the grid. For our
model, the permittivities of the two layer earth model were set as
  1 1 k k k k k k ε1 ¼ 10ε0 and ε2 ¼ 10ε0. The transient SEP can be seen in Fig. 2,
G411 r; r 0 ¼ ½ þ ð  01 þ 12  01 12  01 12
4ps1 R0 R1 R2 R3 R4 which was ultimately agreed with the measured curve in Fig. 7(d)
k01 k23 k01 k12 k23 k23 k01 k23 in Ref. [15]; Meanwhile, the impulse grounding impedance was
  þ  Þ
R5 R6 R7 R8 1.60 U given by the Ref. [15], and it is (1.70,2.44  104)U for our
XNc (9) model “T” case, and (1.70,2.45  104)U, from our model “p” case.
k k k k k k
þ an ð  01 þ 12  01 12  01 12 From Fig. 2 we know that the curves of transient SEP at injection
n¼1
R n1 R n2 R n3 Rn4 point from the two methods are identical.
k01 k23 k01 k12 k23 k23 k01 k23
  þ  Þ
Rn5 Rn6 Rn7 Rn8 5. Simulation result and analysis

where the origin of the coordinate system has been moved to the A typical grounding system can be seen in Fig. 3, which is buried
surface of the earth, the source point is at (0,z') and the field point is in 0.6 m below ground surface. The earth is modeled as a three-
at (r,z), so R0 ¼ [r2 þ (z  z0 )2]½, in the same way, we have layer conductive earth model, whose conductivity and permit-
R1 ¼ ½r2 þ z21 1=2 , R2 ¼ ½r2 þ z22 1=2 , R3 ¼ ½r2 þ z23 1=2 , R4 ¼ ½r2 þ tivity is s1 ¼ 1001 S/m, s2 ¼ 6001 S/m, s3 ¼ 2001 S/m, ε1 ¼ 10ε0,
ε2 ¼ 10ε0, ε3 ¼ 10ε0, respectively. The first layer height is 5.6 m, and
z24 1=2 , R5 ¼ ½r2 þ z25 1=2 , R6 ¼ ½r2 þ z26 1=2 , R7 ¼ ½r2 þ z27 1=2 ,
the second layer height is 4.8 m. The material of the grounding
R8 ¼ ½r2 þ z8  , meanwhile, Rn1 ¼ ½r þ zn1  , Rn2 ¼ ½r2 þ z2n2 1=2 ,
2 1=2 2 2 1=2
system conductor is Cu with the conductivity sCu ¼ 5.8  107 S/m.
Rn3 ¼ ½r þ z2n3 1=2 ,
2 Rn4 ¼ ½r2 þ z2n4 1=2 , Rn5 ¼ ½r2 þ z2n5 1=2 , The conductor radiuses are 7 mm. Three kinds of external excited
Rn6 ¼ ½r2 þ z2n6 1=2 , Rn7 ¼ ½r2 þ z2n7 1=2 , Rn8 ¼ ½r2 þ z2n8 1=2 , here, lightning currents with different parameter T1 are considered to
inject at the corner of the grounding system, which is described by
zijði¼18Þ ¼ signa $2h1 þ signb $2h2 þ signc $z þ signd $z0 and
znijði¼18Þ ¼ signa $2h1 þ signb $2h2 þ signc $z þ signd $z0  bn , and
signa ¼ 0 for i ¼ 1,6, signa ¼ 1 for others; signb ¼ 0 for i ¼ 1  4,
Table 1
signb ¼ 1 for others; signc ¼ 1 for i ¼ 1,3,5,6, signc ¼ 1 for others; Comparison with published measurement results: frequency is 80 Hz.
signd ¼ 1 for i ¼ 1,4,6,8, signd ¼ 1 for others. And a pair of quasi-
“T” case “p” case
static complex image's coefficients are an and bn, Nc is the Model Mea.[18]

maximum numbers of the pair of the an and bn. a 0.276 3

(0.230,2.6  10 ) (0.230,2.6  103)
Once the closed form of Green's function of a point source in the b 0.300 (0.279,3.3  103) (0.279,3.3  103)
c 0.130 (0.110,2.9  103) (0.110,2.9  103)
horizontal multilayered earth model has been obtained, the mutual
384 Z.-X. Li et al. / Journal of Electrostatics 72 (2014) 381e386
Fig. 2. Transient SEP at injection point.
Fig. 3. Typical grounding system.
a double-exponential function as I(t) ¼ I0  (eat  ebt)A, whose
parameters are given in Table 2, and the lightning currents can be
seen in Fig. 4.
To calculate the impulse impedance, maximum frequency must
be decided firstly. The maximum frequency can be decided by the
parameter T1 of the lightning current, which means that the shorter
of T1 is, the greater of the maximum frequency will be. From paper
[18], the Nyquist frequency and the number of samples is
Fig. 4. Different incident currents.
Fig. 5. The transient SEP curves for T1 ¼ 12 ms case.
determined by the accuracy required for the representation of the
time domain response at early times of the current excitation. The
Nyquist frequency range is given by Fm 1/2△t. If the time step △t is
specified in advance, the Nyquist frequency yields the maximum
frequency which has to be accounted for in the frequency spectrum.
In order to illustrate the relationship, different maximum fre-
quencies will be used in the example to calculate the impulse
impedance.
The calculated grounding impulse impedances due to different
maximum frequencies are shown in Table 3 below.
From Table 3, we can see that for cases 1, 2 and 3, the impulse
impedances can be guessed to be about 2.13 U, 2.11 U and 2.08 U,
and errors from maximum frequencies less than 1.50 MHz,
0.80 MHz and 0.43 MHz will be discarded, respectively. Which
means that for the three cases, maximum frequencies used for the
calculation must be chosen to be greater than 1.50 MHz, 0.80 MHz,
and 0.43 MHz, respectively.
The curves of a transient SEP due to different maximum fre-
quencies at the injection point for three cases are given in Figs. 5e7.
From these figures, we can see that maximum value of the transient
Fig. 6. The transient SEP curves for T1 ¼ 20 ms case.
 Z.-X. Li et al. / Journal of Electrostatics 72 (2014) 381e386 385

Fig. 8. Amplifying the transient SEP curves for T1 ¼ 12 ms case.

Fig. 7. The transient SEP curves for T1 ¼ 28 ms case.

Table 2
Parameters of three kinds of incident lightning currents.

Case 1 Case 2 Case 3

I0 ¼ 5.334661 I0 ¼ 6.105605 I0 ¼ 7.771485

a ¼ 8.2780505  103 a ¼ 9.6435975  103 a ¼ 1.2069957  102
b ¼ 0.3102623 b ¼ 0.1452526 b ¼ 7.9305030  102
T1 ¼ 12 ms, T2 ¼ 99 ms T1 ¼ 20 ms, T2 ¼ 99 ms T1 ¼ 28 ms, T2 ¼ 99 ms

SEP at injection point is varied responding to the maximum fre-

quencies, to further better illuminate the phenomena, a locally
magnified figure to show the wave crest of the curves in Figs. 8e10.
From Figs. 8e10, we can see that maximum frequencies greater
than 1.50 MHz, 0.70 MHz, and 0.43 MHz will give reasonable re-
sults. In this way, we can conclude that maximum frequencies used
for the calculation must be chosen according to parameter T1 of
lightning currents.

6. Conclusion Fig. 9. Amplifying the transient SEP curves for T1 ¼ 20 ms case.

Combined with the FFT, a mathematical model with “T” typical

of basic element for accurately computing the lightning response
from substations' grounding system buried in multilayered earth
model has been developed in this paper. To accelerate the calcu-
lation, the QSCIM and analytical formula for mutual induction and
impedance were introduced into this model. How to decide
maximum frequency in the model has been discussed. Two con-
clusions be achieved: (1) The mathematical model with “T” and “p”
typical of basic elements can obtained consistent numerical results;
(2) The maximum frequency used in the mathematical model is
decided by the parameter T1 of lightning current.

Table 3
Impulse impedance due to different max frequency Fm(MHz)

Case 1 Case 2 Case 3

Fm Impedance (U) Fm Impedance (U) Fm Impedance (U)

1.83 (2.14,3.07  108) 1.10 (2.11,1.50  107) 0.79 (2.09,1.28  107)

1.67 (2.13,6.95  108) 1.00 (2.11,1.25  107) 0.71 (2.08,9.31  108)
1.50 (2.12,4.93  108) 0.90 (2.11,1.55  107) 0.64 (2.07,1.33  107)
1.33 (2.11,7.06  107) 0.80 (2.09,3.24  107) 0.57 (2.08,2.50  107)
1.17 (2.09,1.88  107) 0.70 (2.08,2.49  108) 0.50 (2.08,9.65  108)
1.00 (2.57,3.43  107) 0.60 (2.07,1.89  108) 0.43 (2.07,8.33  108)
0.83 (2.70,2.48  108) 0.50 (2.08,8.50  109) 0.36 (2.07,1.46  107)
0.67 (2.86,9.48  108) 0.40 (2.09,1.38  108) 0.29 (2.07,1.01  107)
0.50 (2.56,1.65  107) 0.30 (2.06,1.96  107) 0.21 (2.07,2.27  107)
0.33 (2.75,6.33  108) 0.20 (2.08,1.53  107) 0.14 (2.08,6.01  108)
0.17 (2.59,3.78  108) 0.10 (2.16,2.10  108) 0.07 (2.12,3.22  108)
386 Z.-X. Li et al. / Journal of Electrostatics 72 (2014) 381e386
Fig. 10. Amplifying the transient SEP curves for T1 ¼ 28 ms case.
Acknowledgment
This work was supported by the National Natural Science
Foundation of China under Grant 51177153 and the Science and
Technology Projects of State Grid Corporation of China under
Contract Number: GY172011000JD.
References
[1] F. Dawalibi, Electromagnetic fields generated by overhead and buried short
conductors, part II-ground networks, IEEE Trans. Power Deliv. 1 (4) (Oct. 1986)
112e119.
[2] F. Dawalibi, R.D. Southy, Analysis of electrical interference from power lines to
gas pipelines, part I, computation methods, IEEE Trans. Power Deliv. 4 (3)
(1989) 1840e1846.
[3] L. Huang, D.G. Kasten, Model of ground grid and metallic conductor currents
in high voltage a.c. substations for the computation of electromagnetic fields,
Electr. Power Syst. Res. 59 (2001) 31e37.
[4] Z.X. Li, W.J. Chen, J.B. Fan, J.Y. Lu, A novel mathematical modeling of grounding
system buried in multilayer earth, IEEE Trans. Power Deliv. 21 (3) (2006)
1267e1272.
[5] Z.X. Li, W.J. Chen, Numerical simulation grounding system buried within
horizontal multilayer earth in frequency domain, Commun. Numer. Methods
Eng. 23 (1) (2007) 11e27.
[6] Z.X. Li, J.B. Fan, Numerical calculation of grounding system in low frequency
domain based on the boundary element method, Int. J. Numer. Methods Eng.
73 (2008) 685e705.
[7] Z.X. Li, G.F. Li, J.B. Fan, C.X. Zhang, Numerical calculation of grounding system
buried in vertical earth model in low frequency domain based on the
boundary element method, Eur. Trans. Electr. Power 19 (8) (2009)
1177e1190.
[8] A.F. Otero, J. Cidras, J.L. Alamo, Frequency-dependent grounding system
calculation by means of a conventional nodal analysis technology, IEEE Trans.
Power Deliv. 14 (July3 1999) 873e877.
[9] Z.X. Li, G.F. Li, J.B. Fan, Y. Yin, A novel mathematical model for the lightning
response of a grounding system buried in multilayered earth based on the
quasi-static complex image method, Eur. Trans. Electr. Power 32 (1) (2012)
21e30.
[10] J.S. Yuan, H.N. Yang, L.P. Zhang, X. Cui, X.S. Ma, Simulation of substation
grounding grids with unequal-potential, IEEE Trans. Magn. 36 (4) (July 2000)
1468e1472.
[11] J. Choma, Electrical Networks-Theory and Analysis, 1985. New York.
[12] S. Visacro, A. Soares Jr., HEM: a model for simulation of lightning-related
engineering problem, IEEE Trans. Power Deliv. 20 (2) (Apr. 2005)
1206e1209.
[13] Z.X. Li*, Y. Yin, C.X. Zhang, L.C. Zhang, A mathematical model for the transient
lightning response from grounding systems, Prog. Electromagn. Res. B 57
(2014) 47e61.
[14] P. Yutthagowith, A. Ametani, N. Nagaoka, Y. Baba, Application of the partial
element equivalent circuit method to analysis of transient potential rises in
grounding systems, IEEE Trans. Electromagn. Compat. 53 (3) (2011) 33e41.
[15] Z. Stojkovic, J.M. Nahman, D. Salamon, B. Bukorovic, Sensitivity analysis of
experimentally determined grounding grid impulse characteristic, IEEE Trans.
Power Deliv. 13 (4) (Oct. 1998) 1136e1143.
[16] P.L. Zhang, J.S. Yuan, Z.X. Li, The complex image method and its application in
numerical simulation of substation grounding grids, Commun. Numer.
Methods Eng. 15 (11) (1999) 835e839.
[17] J.X. Ma, F.P. Dawalibi, Influence of inductive coupling between leads on
ground impedance measurements using the fall-of-potential method, IEEE
Trans. Power Deliv. 16 (4) (Oct. 2001) 739e743.
[18] L. Grcev, F. Dawalibi, An electromagnetic model for transients in grounding
systems, IEEE Trans. Power Deliv. 5 (4) (Oct. 1990) 1773e1781.