See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/3058886

Calculating Mutual Ground Impedances Between Overhead and Buried Cables

Article  in  IEEE Transactions on Electromagnetic Compatibility · March 2008

DOI: 10.1109/TEMC.2007.915286 · Source: IEEE Xplore

CITATIONS READS
21 1,059

1 author:

Felipe Uribe
University of Guadalajara
44 PUBLICATIONS   502 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Transient profiles using the Numerical Laplace Transform: Definition and Applications View project

All content following this page was uploaded by Felipe Uribe on 01 June 2016.

The user has requested enhancement of the downloaded file.

198 IEEE TRANSACTIONS ON ELECTROMAGNETIC COMPATIBILITY, VOL. 50, NO. 1, FEBRUARY 2008

Calculating Mutual Ground Impedances Between

Overhead and Buried Cables

F. A. Uribe

Abstract—Aerial and buried electric conductors often share the same

right of way. When a phase-to-ground fault occurs, high transient-induced
overvoltages appear from short-circuit currents to the ground. The mutual
ground-loop impedances between both systems are given by the Pollaczek
coupling integral that does not possess an analytic closed-form solution,
and its integrand is highly oscillatory to perform a direct numerical inte-
gration. An efficient and accurate algorithmic evaluation of the Pollaczek
coupling integral is presented in this paper. The obtained result is used in
the numerical assessment of approximated formulas previously issued by
Lucca, Wedepohl, and Comité Consultatif International Téléphonique et
Télégraphique.
Index Terms—Ground return, inductive coupling, numerical integra-
tion, Pollaczek coupling integral. Fig. 1. Transmission system comprising an overhead conductor and a buried
power cable on an imperfectly conducting homogeneous ground.

NOMENCLATURE
d Distance between conductors. ductors [4]. The solution is numerically efficient, guarantees conver-
h1 Height of the aerial conductor. gence, and the local error is bounded [4]. In this paper, an extended ver-
h2 Depth of the underground cable. sion of this algorithmic solution is presented for the case of evaluating
x Horizontal distance between conductors. Pollaczek coupling integral for calculating mutual impedances between
ξ Cable depth “h2 ” normalized by the skin effect layer thickness buried and overhead conductors. Comparisons between the proposed
“|p|.” algorithm and the direct numerical integration with the Gauss–Lobatto
µ0 Magnetic permeability of vacuum. quadrature rule [5] are also provided.
η Horizontal distance “x” normalized by the cable depth “h2 .” Finally, the developed algorithm is further applied to assess the
ω Angular frequency. accuracy ranges of the approximate formulas previously proposed by
σ Soil conductivity. Lucca, Wedepohl, and Comité Consultatif International Téléphonique
ζ Height of the aerial conductor “h1” normalized by the cable et Télégraphique (CCITT) [2], [6].
depth “h2 .”
II. COUPLING INTEGRAL BETWEEN OVERHEAD
I. INTRODUCTION AND BURIED TRANSMISSION SYSTEMS

Overhead lines and buried transmission systems often share the same Fig. 1 shows a hybrid transmission system comprising an overhead
right of way, and also with another supply services such as oil, gas, water conductor and a buried power cable, with common earth-return. Assum-
pipes, communication lines, and electrified railway systems. During ing a quasi-TEM mode of propagation, the mutual ground impedances
both transient and steady-state operation conditions, coupled power are given by [1]
transmission conductors induce voltages and currents on each other.
Moreover, the aerial conductors could be close to a metallic fence, and  +∞


jωµ0 exp [−h1 |β|] exp[−h2 β 2 + 1/p2 ]
then, the induced-transient voltage poses a safety hazard. To analyze ZG (ω) =

voltage induction levels, it is required to calculate mutual (between 2π −∞ |β| + β 2 + 1/p2
buried and overhead conductors) ground inductances as functions of
× · · · × exp (jβx) dβ (1a)
frequency [1]–[4].
In 1926, Pollaczek postulated mathematical expressions in the form where p is the complex depth skin effect thickness given by
of infinite integrals for calculating electromagnetic fields inside an im-
perfectly conducting ground due to a buried thin filament of current [3]. 1
In 1927, Pollaczek presented a similar formulation for calculating the p= √ . (1b)
jωµ0 σ
mutual ground loop impedance between overhead and buried conduc-
tors [1]. The problem with Pollaczek coupling integral is that it does not Introducing the variable change β = u/|p| and the following identity
have an analytic closed-form solution and is highly irregular oscillatory, into (1a), we have
making its direct numerical integration difficult.

j
Recently, the author proposed an algorithmic solution to the u2 + j + u =
. (1c)
Pollaczek integral to evaluate mutual impedances between buried con- [ u + j − u]
2

Thus,
Manuscript received September 11, 2006; revised March 2, 2007 and   
September 21, 2007. ωµ0
+∞
h1
The author is with the State University of Guadalajara, CUCEI, CP. 44430, ZG = [ u2 + j − u] exp − u
Jalisco, Mexico (e-mail: [email protected]). π 0 |p|
   
h2
2
Color versions of one or more of the figures in this paper are available online
at http://ieeexplore.ieee.org. u
× · · · × exp − u + j cos x du. (2a)
Digital Object Identifier 10.1109/TEMC.2007.915286 |p| |p|

0018-9375/$25.00 © 2008 IEEE

IEEE TRANSACTIONS ON ELECTROMAGNETIC COMPATIBILITY, VOL. 50, NO. 1, FEBRUARY 2008 199

Consider now the algebraic solution for the first radical factor in (2a) A. Truncation Criterion
as [4]
Integrating the pure damping exponential function on the right-hand
u2 + j = F (u) + j G (u) (2b) side of (2e) yields
where  ∞

√ exp (−ξ (ζ u + F (u))) du
u2 + u4 + 1
F (u) = √ (2c) 0
2  u ma x

√ ∼ ··· =
= ∼ exp (−ξ (ζ u + F (u))) du + εr . (4a)
−u2 + u4 + 1 0
G (u) = √ . (2d)
2 On applying F (u) → u, for u > 1 [4] on the right-hand side of (4a),
Applying (2b) in (2a), we obtain we obtain the following relative error:
 +∞ εr = exp (−ξ (ζ + 1) uma x ) ξ (ζ + 1) . (4b)
ωµ0
ZG = [F (u) − u + jG (u)] exp [−ξζu]
π 0 For an arbitrary fixed value of εr , the corresponding uma x is
× · · · × exp [−ξ (F (u) + jG (u))] cos (ξζηu) du
 λe
+∞ uma x = (4c)
ωµ0 ξ (ζ + 1)
= [F (u) − u + jG(u)] exp [−ξ (ζu + F (u))]
π 0
where
× · · · × exp [−jξG(u)] cos (ξηu) du λe = −log (εr ξ (ζ + 1)) . (4d)
ωµ0 A satisfactory relative error εr using (4c) in (2e) can be found by
= J (ξ, η) (2e)
π refining λe . A value of λe = 12 has been determined empirically for
where ξ = h2 /|p/|, η = x/h2 , and ζ = h1 /h2 are normalized param- this paper’sapplications.
eters related to the physical configuration of the system in Fig. 1, while
functions F (u) and G(u) are independent of the physical geometry
of the problem. Another advantage of representation (2e) is that the B. Integrand Oscillations Identification Procedure
integration range is positive defined [0, ∞]. The second step is to include the contribution of the modified regular
Pollaczek coupling integral (2e) is now represented in four factors. and irregular oscillating factors of the integrand in (2e) in the new
The first two are of the damping type, while the second two are mere truncated range [0 − uma x ].
oscillatory. The first factor depends on functions F (u) − u and G(u), Basically, regular and irregular oscillations depend on the magnitude
which are decreasing monotonic. The second factor is a pure damping of parameters η and ξ, respectively.
exponential function. The third and the fourth factors are irregular and On one hand, regular oscillations are introduced to (2e) through the
regular oscillatory functions, respectively. As their function arguments cosine term. This term does not oscillate if
are identical to the case of analyzing Pollaczek integral for buried
π
conductors, the identification procedure for the zero crossings of the ξηu < . (4e)
2
integrand and the truncation criterion for the integral can be applied
through the set of numerical recipes proposed in [4]. Substitution of (4e) into (4c) yields
Basically, the main differences between both Pollaczek integrals
(buried and aerial-buried systems) are summarized in the following η λe π
< . (4f)
relationship that is valid for ζ = 0: (ζ + 1) 2

exp (−ξF (u)) = exp (−ξ (ζ u + F (u))) . (3a) Thus,

π (ζ + 1)
η< . (4g)
The right–hand side of (3a) means that the Pollaczek coupling inte- 2 λε
gral is a particular case of the Pollaczek integral for calculating ground Thus, from (4e), it can be noticed that each of the regular zero
impedances of buried cables. In the case of ζ = 0, the following addi- crossings of the integrand can be identified according to (4c) as
tional damping factor is introduced to the integrand: π
uk < (2k − 1) , for k = 1, . . . , Kma x (4h)
exp (−ξ (ζu)) . (3b) 2ξη
where k is the corresponding oscillation and Kma x is the total number
Thus, the additional damping factor given by (3b) depends on the
of oscillations in the truncated range [0, uma x ].
parameter product ξζ. The more general application case considered
On the other hand, the irregular oscillations are introduced to (2e)
in this paper occurs when ζ = 0.1. Greater values from this limit are
by the complex exponential factor when
considered as damped subcases.
π
ξG (u) = . (4i)
III. POLLACZEK COUPLING INTEGRAL SOLUTION 2
The algorithmic solution for calculating ground-return impedances From (2d), the highest value of the function G(u)
√ (decreasing mono-
for buried conductors proposed in [4] is extended in this paper to the tonically) occurs when u = 0, where G(0) = 1/ 2. Then, the irregular
case of analyzing ground-return impedances between overhead and oscillations will start for [4]
buried conductors through the Pollaczek coupling integral as described π√
in the following. ξ> 2 (4j)
2
200 IEEE TRANSACTIONS ON ELECTROMAGNETIC COMPATIBILITY, VOL. 50, NO. 1, FEBRUARY 2008

TABLE I
APPLICATION RANGES FOR PHYSICAL VARIABLES

TABLE II
RANGES FOR NORMALIZED DIMENSIONLESS PARAMETERS

TABLE III
COMPUTATIONAL TIME AS A FUNCTION OF THE SAMPLES AND ROOT MEAN
SQUARE ERROR ON CALCULATING POLLACZEK ALGORITHMIC SOLUTION

and the total number of oscillations in the range [0, uma x ] is

√
π 2
Kma x = . (4k)
2ξ

Thus, the zero crossings of this factor are given by inverting the
nonlinear function (2d)

 

−1 kπ ξ 4 − (k 4 π 4 /4)
uk = G = , for k = 1, . . . , Kma x .
2ξ πξk
(4l)
The total zero crossings of (2e) in the new truncated range [0, uma x ]
are thus identified by sorting regular and irregular oscillations due to
(4h) and (4l). Fig. 2. Pollaczek induction integral solutions for ranges inside Table II.
(a) Curves of
e{J (ξ, η)}. (b) Curves of m{J (ξ, η)}.

C. Numerical Solution of Pollaczek Coupling Integral

The dimensionless parameters ξ, η, and ζ depend on the physical
quantities h1 , h2 , ω, ε, µ, σ, and x. A broad application range ar-
bitrarily determined for these variables is proposed in Table I. Most IV. NORMALIZED COUPLING IMPEDANCES
of the practical engineering application cases lie within these variable The normalized coupling impedance concept is now introduced as
ranges when calculating ground-return impedances between overhead [4]
and buried conductors of coupled systems. Subsequently, these physi- π 
cal variable ranges are used to establish the parameter ranges of ξ, η, Z G ∆ ZG µ0 . (5a)
= ω
and ζ given in Table II.
The truncation criterion and the integrand oscillation procedure pro- Using (5a), the Pollaczek coupling integral is redefined as
posed in this paper have been implemented to solve Pollaczeck coupling
integral 103 × inside the ranges given in Table II, with 100 samples  +∞
for ξ, 10 samples for η, and ζ = 0.1 (value corresponding to the most Z G = J (ξ, η) = [F (u) − u + jG(u)] exp[−ξ(ζu + F (u))]
general case). 0
Fig. 2(a) and (b) depicts the numerical solution for J (ξ, η) in (2e),
× · · · × exp[−jξG(u)] cos(ξηu)du. (5b)
obtained with the Pollaczek algorithm proposed here for the ranges
given in Table II.
Table III shows the magnitude of the rmse rro r and CPU time as a Now, the approximation of (5b) does not depend on physical vari-
function of the sampling rate “N ” required for this task on a Pentium ables, allowing the possibility to handle broad application ranges for
4M at 1.6 GMHz with 1.0 GB of RAM running MATLAB V. 6.5. the dimensionless parameters ξ, η, and ζ.
IEEE TRANSACTIONS ON ELECTROMAGNETIC COMPATIBILITY, VOL. 50, NO. 1, FEBRUARY 2008 201

V. CLOSED-FORM APPROXIMATIONS AND ADAPTIVE NUMERICAL

INTEGRATION FOR CALCULATING Zg
The difficult task of solving Pollaczek integrals has motivated the
search for new closed-form approximations.
In practice, power engineering analysts use approximate formulas
though their accuracy ranges are not determined.
These are preferred instead of using complicated numerical algo-
rithms that sometimes are very cumbersome and require a long com-
puting processing time. Now, the broad range solutions depicted in
Fig. 2 of the normalized Pollaczek coupling integral in (4b) are used to
assess the following ground-return impedance approximations.

A. Lucca Formulation
Lucca proposed the following two-step formula [2]. The first step
is based on the theory of images, following the work by Wait and
Spies [2]. The second step consists of the suppression of the oscil-
latory exponential factor of the integrand. The latter is very similar
to the idea proposed by Ametani for calculating underground cable
earth impedances by using the Carson’s integral [4]. Thus, the formula
proposed by Lucca is [2]
    
jωµ0 R1 2 2y y 2 − 3x2
ZG −L = ln − 3 6
(6a)
2π R1 2 3γ R1 2



where R1 2 = x2 + (y1 − y2 )2 , ke2 = −jωµ0 σ, R1 2 = y 2 + x2 ,
γ = jke , and y = y1 − y2 + 2/γ.
Transforming (6a) into the dimensionless parameters given in
Table II, and normalized coupling impedances in (5a), we get

   ! 
 λ λ2
− 
λ +η
2 2 2 3η 2
Z G −L = j ln  + √ !
 η 2 + (ζ − 1)2 3 jξ 3 η 2 + λ2 
3

√ % (6b)
where λ = ζ − 1 + 2 j ξ.
Expression (6b) is a function of ξ, η, and ζ only.

B. CCITT Recommended Formula Fig. 3. Gauss–Lobatto quadrature of the Pollaczek induction integral for
ranges inside Table II. (a) Curves of
e{J (ξ, η)}. (b) Curves of m{J (ξ, η)}.
The CCITT formula [2] (a specialized paper concerning telecom-
munication lines protection) recommended the following approximate
formula for calculating mutual ground impedances between overhead following terms can be considered; thus [7]
   
and buried lines: jωµ0 γd 4 (h1 + h2 )
&   ' ZG −W = − log + 0.5 − (6e)
2π 2p 3p
jωµ0 1.851 2jke (y1 + y2 )
ZG −C = ln + . (6c)
2π jke R1 2 3 where γ is the Euler’s constant.
Due to the simplicity of (6e), this expression has been preferred
It is convenient to represent (6c) into the dimensionless variables over different formulas for analyzing induced voltages and currents be-
given in Table II. According to the normalized impedance concept tween overhead and buried power lines [6]. The equivalent normalized
stated in (5a) parameter version of (6e) as a function of ξ, η, and ζ is
  * 
 ( √ ) √ 

1.851 · j 2j j ξ(1 + ζ) Z G −W = j − log γ jξ η 2 + (ζ − 1)2 2
Z G −C = j ln
− . (6d)
ξ η 2 + (ζ − 1)2 3

2

+ · · · + 0.5 − jξ (ζ + 1) . (6f)
C. Wedepohl Formula 3
In 1973, Wedepohl and Wilcox reported a series solution to ap-
proximate Pollaczek integral for calculating ground-return impedances
D. Method of Adaptive Gauss–Lobatto Quadrature Rule
between buried cables [7]. However, this series solution resulted very
cumbersome, and at certain ranges, converged slowly [4], [6]. Subse- The direct numerical integration of Pollaczek coupling integral (5b)
quently, these authors recognized that, up to quite frequencies, only the has also been implemented here for the validation of Pollaczek method
202 IEEE TRANSACTIONS ON ELECTROMAGNETIC COMPATIBILITY, VOL. 50, NO. 1, FEBRUARY 2008

TABLE IV
COMPUTATIONAL TIME AS A FUNCTION OF THE TOLERANCE AND ROOT MEAN
SQUARE ERROR ON CALCULATING POLLACZEK THROUGH GAUSS–LOBATTO

Fig. 5. Relative error on approximations Z G -L , Z G -C , and Z G -W [2], [7].

(a) Curves of
e{δ}. (b) Curves of m{δ}.

VI. NUMERICAL ASSESSMENT OF APPROXIMATE Zg

In order to assess the ZG approximate formulations mentioned be-
fore, consider the first application example reported in [2, Sec. III]. The
Fig. 4. Ground impedances along horizontal distances for the application conductivity of the soil is σ = 0.01 S/m, the height of the aerial conduc-
example reported in [2]. (a) Curves of
e{Z G }. (b) Curves of m{Z G }. tor is h1 = 15 m, the depth of the buried conductor is h2 = 1 m, and the
horizontal distance between conductors is linearly sampled 100 times
for values inside 0 < x ≤ 2000 m range. The frequencies considered
for this test are f1 = 50 Hz, f2 = 500 Hz, and f3 = 5000 Hz.
(using “quadl” routine in MATLAB > V.6.5), through the adaptive Fig. 4(a) and (b) depicts the real and imaginary curves of ZG for
Gauss–Lobatto quadrature rule [5]. each frequency test using the Pollaczek algorithmic solution and the
For this implementation in the new truncated range [0, uma x ], the approximate ZG formulations [2], [5], [7]. Notice in Fig. 4(a) that the
criterion (4c) and (4d) with a fixed absolute error tolerance of 1 × 10−6 set of curves are in good agreement. However, the real component of
has been chosen. both, the Wedepohl and the CCITT formulas presented values that tend
The broad range solutions depicted in Fig. 2(a) and (b) have been to |4.84 × 10−5 |, |46.37 × 10−5 |, and |399.31 × 10−5 | for f1 , f2 , and
reproduced in Fig. 3(a) and (b) through the direct numerical integration f3 , respectively. In Fig. 4(b), the Wedepohl and CCITT formulations
applying the Gauss–Lobatto quadrature rule. presented valid results at the low-frequency range, only.
In addition, Table IV shows the behavior of the magnitude on The Gauss–Lobatto quadrature rule again presents numerical dis-
rmse rro r and the CPU time, both functions of the error tolerance continuities in the high-frequency range. Thus, the latter is dis-
mentioned before. There are some numerical discontinuities present in carded for the numerical assessment of the Pollaczek algorithmic
both real and imaginary components of Fig. 3. methodology.
IEEE TRANSACTIONS ON ELECTROMAGNETIC COMPATIBILITY, VOL. 50, NO. 1, FEBRUARY 2008 203

It can be seen from Fig. 4(b) that the Lucca formula presents an Complexity Analysis of a Heuristic Method
excellent performance during the entire range. for Fixed-Frequency Assignment Including
In addition, the relative errors on approximate formulations for ZG Adjacent Channel Interference
have been evaluated according to [2]
Raúl Chávez-Santiago, Member, IEEE, Eli Gigi,
  and Vladimir Lyandres, Senior Member, IEEE
ZG -a p rox
δ= 1− × 100 (6g)
ZG -P o lla c z e ck
Abstract—In this paper, we extend an existing heuristic algorithm for
fixed-frequency assignment in cellular communication networks in order
where ZG -a p rox is the approximate formulation and ZG -P o lla c z e k is the to account for the effects of adjacent channel interference (ACI). Our algo-
Pollaczek algorithmic solution (taken as a reference for this calcula- rithm utilizes estimations of carrier-to-interference ratio instead of binary
channel separation constraints. The extended algorithm avoids overopti-
tion). mistic frequency assignments that do not guarantee interference-free oper-
Fig. 4(a) and (b) shows relative errors “δ” on real and imaginary ation environments. Its cost is the increased computational complexity. We
components of ZG -L , ZG -C , and ZG -W . Notice from both figures that estimate the upper bound of the algorithm’s complexity when ACI is taken
the Lucca formula again presents the best performance at the entire into account. Simulations demonstrate that despite the theoretical high in-
crease in complexity, the algorithm is suitable even for a quasi-dynamic
range, with a maximum relative error on the imaginary component of frequency assignment scheme in practical situations. The algorithm en-
50-Hz curve of 45.65% at x = 1878 m. However, Pollaczek formula- hances the spectrum utilization in cellular networks while reducing harm-
tions have been derived considering a homogeneous ground; in such a ful interference.
case, the validity of curves shown in Fig. 5 are up to sample x = 606.06 Index Terms—Algorithms complexity, cellular communications, fre-
m in which the relative error on Lucca formula is δ = 2.39%. quency assignment, heuristics, interference.

I. INTRODUCTION
VII. CONCLUSION
The optimization of a frequency assignment plan for a cellular com-
The Pollaczek coupling integral for calculating mutual impedances
munication network is an NP-complete combinatorial problem [1].
between overhead and buried conductors, both with earth-return, has
There is no mathematical proof of whether an NP-complete problem is
been solved in this paper for a broad range of applications. The applied
solvable in polynomial time. Therefore, heuristic methods are applied
methodology is based on a extension of a previously developed accurate
extensively to solving problems of this kind.
algorithm for solving Pollaczek integral for ground-return impedances
A frequency assignment problem (FAP) is commonly modeled us-
of buried conductors [4].
ing the graph theory proposed in [2]. Many heuristics for graph col-
Finally, through the concept of normalized impedances and di-
oring have been applied to this problem. Sequential (greedy) algo-
mensionless parameters ξ, η, and ζ, it has been possible to assess
rithms [3]–[7] are among the most popular ones due to their relative
three well-known closed-form approximations proposed by Lucca,
simplicity, robustness, good running-time, and surprisingly, good so-
Wedepohl, and the CCITT [2].
lutions in many cases. These algorithms are generally used to solve
The Matlab code routines for evaluating Pollaczek coupling integral
the minimum span problem, i.e., find an assignment that minimizes the
can be freely obtained from the author’s Web page or via e-mail.
difference between the lowest and the highest used frequency channels.
Most sequential algorithms model the interference using binary channel
separation constraints assuming a regular hexagonal network layout [5].
ACKNOWLEDGMENT
However, it has been demonstrated in [8] that the algorithms with bi-
The author is deeply indebted for many years of work and friendship nary constraints do not always produce optimal solutions. In networks
with Prof. J. L. Naredo, A. Ramirez, E. Barocio, P. Moreno, V. Ortiz, with irregular layouts, the use of nonbinary constraints such as maxi-
and R. Castellanos. mizing the carrier-to-interference ratio (CIR) over the covered region
is more effective [9]. Based on a similar idea, a sequential algorithm
that utilizes estimations of CIR was presented in [10]. Unfortunately,
REFERENCES this algorithm takes into account only cochannel interference (CCI)
and neglects the effects of adjacent channel interference (ACI), which
[1] F. Pollaczek, “Über die induktionswirkungen einer wechselstromeinfach-
leitung,” Elect. Nachr. Tech., vol. 4, no. 1, pp. 18–30, 1927. Manuscript received September 25, 2005; revised July 23, 2006 and
[2] G. Lucca, “Mutual impedance between an overhead and a buried line with February 26, 2007. The work of R. Chávez-Santiago was supported by the
earth-return,” in Proc. 1994 Inst. Electr. Eng., 9th Int. Conf. Electromagn. National Council of Science and Technology (Conacyt-Mexico) under Gradu-
Compat., pp. 80–86. ate Scholarship 130737.
[3] F. Pollaczek, “Über das Feld einer unendlich langen wechsel stromdurch- R. Chávez-Santiago was with the Department of Electrical and Computer En-
flossenen Einfachleitung,” Elect. Nachr. Tech., vol. 3, no. 9, pp. 339–360, gineering, Ben-Gurion University of the Negev, Beer-Sheva 84105, Israel. He
1926. is now with the Laboratory of Research in Informatics, University of Paris-Sud
[4] F. A. Uribe, J. L. Naredo, P. Moreno, and L. Guardado, “Algorithmic XI, 91405 Orsay Cedex, France (e-mail: [email protected]).
evaluation of underground cable earth impedances,” IEEE Trans. Power E. Gigi was with the Department of Electrical and Computer Engineering,
Del., vol. 19, no. 1, pp. 316–322, Jan. 2004. Ben-Gurion University of the Negev, Beer-Sheva 84105, Isreal. He is now with
[5] W. Gander and W. Gautschi, “Adaptive quadrature—Revisited,” BIT, the Dr. Yoram Tsivion Patent Office, Caesarea Industrial Park 38900, Israel
vol. 40, pp. 84–101, 2000. (e-mail: [email protected]).
[6] H. W. Dommel and J. Sawada, “The calculation of induced voltages and V. Lyandres is with the Department of Electrical and Computer Engineering,
currents on pipe-lines adjacent to ac power lines,” Dept. Elect. Eng., Univ. Ben-Gurion University of the Negev, Beer-Sheva 84105, Israel (e-mail: lyan-
British Columbia Rep. B. C. Hydro Power Authority, August 16, 1976. [email protected]).
[7] L. M. Wedepohl and D. J. Wilcox, “Transient analysis of underground Color versions of one or more of the figures in this paper are available online
power-transmission systems,” Proc. Inst. Electr. Eng., vol. 120, no. 2, at http://ieeexplore.ieee.org.
pp. 253–260, Feb. 1973. Digital Object Identifier 10.1109/TEMC.2007.911937

0018-9375/$25.00 © 2008 IEEE

View publication stats