Fast Transients Analysis of Non-uniform Multiconductor

Transmission Lines Excited by Incident Field

Alejandro R. Chvez B, Pablo Moreno V., Member, IEEE, Jos L. Naredo, Senior Member, IEEE.


Abstract --In this work the development of a model for the
simulation of fast transients on non-uniform multiconductor
transmission lines excited by incident fields is presented. The
proposed transmission line model is based on the method of
characteristics of partial differential equations and is able to
include resistive losses and wave speed variations. The results
obtained with the model are compared with those obtained with
the EMTP.
Index Terms --Fast Transients, Method of Characteristics, Non-
uniform Multiconductor Transmission Lines, Electromagnetic
fields.
I. INTRODUCTION
HE most popular technique for analyzing electromagnetic
transients on transmission lines is the method of
Bergeron [1]. This method is very efficient for the case of
uniform transmission lines, this is, lines whose electrical
parameters do not depend on the distance. The main reason
for this is that for integration of the telegraphers equations
subdivision of the line is not necessary. However, for a
number of practical cases, such as when the electrical
parameters vary with the distance, when corona effect
appears or when the transmission line is excited by external
electromagnetic fields, a knowledge of the voltage and
currents on interior points is required [5]. In those cases a
discretization of the transmission line length is necessary.
Although subdivision of the line in a number of sections for
EMTP simulation is possible, this procedure is cumbersome
and a great deal of experience is needed in order to define the
optimal number of line sections. Moreover oscillation
problems have been reported when simulating nonlinear
problems using Bergerons method [2].
The method of characteristics, one of several solution
methods that discretize both time and distance, has been used
successfully in calculating transients on lines with nonlinear
effects due to corona. It has been reported that this method
does not present the numerical oscillations that are very
common in finite difference methods [3,7,8,12].
In this work the problem of electromagnetic transients
on non-uniform multiconductor transmission lines excited by
incident fields is analyzed. In particular non-uniformities due
to conductors catenary are addressed.

______________________
The authors are with the Center for Research and Advanced Studies
(CINVESTAV-IPN), Prol. Lpez Mateos Sur 590, Guadalajara, Jalisco, C.P:
45090, Mxico (e-mail: pmoreno@gdl.cinvestav.mx).
In addition equivalent circuits for the transmission
line ends that allow introducing the method of characteristics
into general purpose computer programs are presented.
Finally results obtained with the proposed method are
compared with those obtained with the EMTP.

II. METHOD OF CHARACTERISTICS
The electromagnetic behavior of multiconductor
overhead transmission lines, with frequency independent
electrical parameters is described by means of telegraphers
equations:

0

A
I
V
B
I
V
A
I
V
=
(
(

+
(

+
(

E
B
t x t
(1)
with


LR 0
0 0
B ,
0 L
C 0
A
(

=
(
(

1
1
1
(2a,b)
where V and I are the voltage and current vectors at any
point x, respectively; L, R and C are the inductance,
resistance and capacitance matrices in per unit length [4].
Also
] ... ... [
t
i
0
i
h
inc
z i
dy

= B
B

(3a)
t
i
] ... ... [
i
h
inc
y i
dy

=
0
E C
E
(3b)
Expression (1) represents a system with 2n first order partial
differential equations (PDEs) with x and t as independent
variables.
By definition system (1) represents a hyperbolic
system if matrix A has only real eigenvalues and a complete
set of linearly independent eigenvectors [9,11]. If T
V
and T
I

are the matrices that diagonalize the LC and CL products,
respectively, then
=
V
-1
V
T LC T (4a)
=

I
1
I
CLT T (4b)
Note that the columns of T
V
are the voltage propagation
modes, the columns of T
I
are the current propagation modes
and the diagonal elements of are the squared inverses of
the corresponding velocities.
The modal matrices T
V
and T
I
are related by means
of the following expression [10]:
T
0-7803-7967-5/03/$17.00 2003 IEEE
Paper accepted for presentation at 2003 IEEE Bologna Power Tech Conference, June 23th-26th, Bologna, Italy


-1
I
T
V
T T = (5)
In modal analysis it can be defined two additional
matrices: the modal inductance, L, and the modal
capacitance, C. These matrices are given by:
I
T L T L'
1
V

=
(6a)
V
1
I
CT T C'

= (6b)
Consider now the following pair of 2nx2n matrices:
(
(

|
.
|

\
|
=
1 -
I W
1 -
V
-1
I W
-1
V
L
T Z T
T Z T
E
2
1
(7a)
(

|
.
|

\
|
=
W I W I
V V
R
Y T Y T
T T
E
2
1

(7b)
where
1
W
'

= C L' Z
and
1
W
'

= L C' Y
(7c)
Note that Z
W
and Y
W
have dimensions of impedance and
admittance, respectively.
Matrices E
L
and E
R
diagonalize matrix A, this is
(

0
0
R L
AE E (7d)
where
1 1 1
'

= = C' L (7e)
Expressions (7a), (7b), (7c), (7d) and (7e) provide a
complete set of eigenvalues and eigenvectors for A.
Expressions (7d) and (7e), together with the fact that the
elements of are real and positive imply that the eigenvalues
of A are real.
Expression (1) can be cast in a more convenient way
by multiplying it by E
L
:
0 Z RI T I T Z
V T I I Z V V
W =

+ |
.
|

\
|

|
.
|

\
|

|
.
|

\
|

+ |
.
|

\
|

B
m m
1
V
1 -
I
1
V m m m m
t t x
x
-
x t x t
W
W
E


(8a)
and
0 Z RI T I T Z
V T I I Z V V
V
V
=

|
.
|

\
|

|
.
|

\
|

+ |
.
|

\
|

|
.
|

\
|

B
m m
1 1 -
I
1
m m m m
t t x
x x t x t
W W
W
E


(8b)
where
V T V
-1
V m
= and I T I
-1
I m
= (8c,d)
Being V
m
and I
m
the modal vectors of voltage and current.
Note that the analysis is based on the real matrix LC that
varies along the transmission line.
Expressions (8a) and (8b) define 2n equation pairs of
the form:
0
t t
Z I R I T
x
Z
V T
x x
I
t
I
Z
x
V
t
V
mj j mj j Wj
n
1 k
mk
"
jk j k
n
1 k
1
Ijk j Wj
n
1 k
k
-1
Vjk j
mj
j
mj
Wj
mj
j
mj
=

|
|
.
|

\
|

|
|
.
|

\
|

|
|
.
|

\
|

|
|
.
|

\
|

= =

=
B
m
m
E

(9)
where subscript j indicates the j-th element in the case of
column vectors V
m
and I
m
, and, the j-th diagonal element for
the diagonal matrices and Z
W
;
' '
jk
R , is the element in the j-
th row and the k-th column of the matrix :
I
-1
V
T R T ' R' = (10)
Consider now some of the curves on the x-t plane
defined by the following expression:
j
dt
dx
=
(11a)
Along curves (11a), (9) becomes:

=
=

= +
+
n
1 k
m m W mk
"
jk
n
1 k
k
1
Ijk
W
n
1 k
k
1
jk V
mj
W
mj
0
dt
d
dt
dx

dt
d
dt
dx
Z I R
dt
dx
I T
dx
d
dt
dx
Z V T
dx
d
dt
dx
dt
dI
Z
dt
dV
B E
(11b)
and

=
=

= + +
+
n
1 k
m m W mk
"
jk
n
1 k
k
1
Ijk
W
n
1 k
k
1
jk V
mj
W
mj
0
dt
d
dt
dx

dt
d
dt
dx
Z I R
dt
dx
I T
dx
d
dt
dx
Z V T
dx
d
dt
dx
dt
dI
Z
dt
dV
B E
(11c)
where (11b) corresponds to the positive sign of (11a) and
(11c) to the negative sign. The curves over which the PDEs
become Ordinary Differential Equations (ODEs) are known
as characteristics.
Equations (11b) and (11c), can be approximated by
means of finite differences

=
=

=
+ +
n
1 k
mj
j
mj
j
j W mk
"
jk
j
n
1 k
k
j
1
Ijk j
j W
n
1 k
k
j
1
Vjk j mj
j W
mj
0
t

t
x

t
x
Z I R
t
x
I
x
T
t
x
Z V
x
T
t
x
t
I
Z
t
V
B E
(12a)
and

=
=

= + +
+
n
1 k
mj
j
mj
j
j W mk
"
jk
j
n
1 k
k
j
1
Ijk j
j W
n
1 k
k
j
1
Vjk j mj
j W
mj
0
t

t
x

t
x
Z I R
t
x
I
x
T
t
x
Z V
x
T
t
x
t
I
Z
t
V
B E
(12b)
Substituting the increments by the voltage and current
values shown in Fig. 1, and using averages for the required
voltages and currents at the evaluation point, equations (12a)
and (12b) can be written as follows:
n
D
n
D
n
D
n
D
1 n
E m
1 n
E
1 n
E
t t
D E D E
t t
E D E D
1 n
E

1 1

1
2 2
}
"
] [ { ] [
} 2 2
]
"
] [ [ { ]
1
[

B E
B E
m m
W1
m
1 W1
m
m
W1
m
1 W1
x x
Z
I xR
I
T
I
T U Z V
V
T
V
T U
x x
Z
I xR
I
T
I
T U Z V
V
T
V
T U
m

+ +
=

+ + +

+
+
+ +
+

+
(13a)
and
n
F
n
F
n
F
n
F
1 n
E m
1 n
E
1 n
E
t t
F E F E
t t
E F E F
1 n
E

1 1

1
2 2
}
"
] [ { ] [
} 2 2
]
"
] [ [ { ]
1 -
[

B E
B E
m m
W2
m
2 2 W
m
m
W2
m
2 2 W
x x
Z
I xR
I
T
I
T U Z V
V
T
V
T U
x x
Z
I xR
I
T
I
T U Z V
V
T
V
T U
m

+ +
=

+
+ + +

+ +
+

+
(13b)
where U is the identity matrix and
( ) 2 /
D E W W W1
Z Z Z + = (14a)
( ) 2 /
F E W W W2
Z Z Z + = (14b)
2 / ) (
' ' ' '
D E
R R R
' '
1
+ = (15a)
2 / ) (
' ' ' '
F E
R R R
' '
2
+ = (15b)
x = diag(x
1
,x
2
,.......,x
n
) (16a)
x
j
=
j
t (16b)
the subscript n indicates any time point and n+1
indicates the next time point. The subscripts mD, mE and mF
indicate modal vectors at the corresponding points, as shown
in Fig. 1. Combining equations (13a) and (13b) expressions
for the voltage vector
1 n
E
+
m
V and the current vector
1 n
E
+
m
I can
be obtained.
In the case of a lossy non-uniform transmission line,
the characteristics are curved, therefore if a uniform mesh is
retained the departure points of the characteristics that
intersect at any apex of the mesh are unknown. In order to
find such departure points an iterative process is used at the
beginning of the calculation and then in the time marching
algorithm interpolations are used.
t
x
D
E
F
D
1
D
2
D
n
F
1
F
2
F
n
x
n
x
2
x
1
t = T
t = T + t

1
2

n
x

Fig. 1. Characteristics in the x-t plane.

III.-BOUNDARY CONDITIONS
At the sending end of the line there exist only n
characteristic curves, which correspond to the lower sign of
equation (11a), as shown in Fig. 1. Transforming (13b) to the
phase domain and rearranging it can be written
hist1 1
I V Y I -
1 n
E
1 n
E
+ +
= (17a)
where
)
s4 s3 s2 s1
(
E
v v v v T Y I
V hist1 1
+ + =
(17b)
1 1
E E F F
1
)
E F
(
E
} ] ) ( [ {
' ' 1
+ + + =

I 2 I I W2 V V V V
T xR T T U Z T T T T Y
1
(18)
n
F
n
F
E
F E F m V
V T T U T T T
V
V V V
v ] [ ] [
s1
-1 1
+ + =

(19)
n
F F
E F E F
m
2
I
I
I xR T T U Z T T T v
W2 V V V
} {
1 "
] [ ] [
s2
+ + =

(20)
t / ] [ 2 ) (

s3 m m
1
x Z T T T v
W2 V V V
+ =
+ n
F
1 n
E F E F
E E

(21)
t / ) (- ] [ 2
m m
1
s4
x T T T v
V V V
n 1 n
F F E F
B B
+ + =
+
(22)
At the receiving end there exist only n characteristics
too, which correspond to the upper sign of equation (11a),
transforming the equation (13a) to the phase domain and
rearranging :
hist2 2
I V Y I + =
+ + 1 1 n
E
n
E
(23a)
where:
( )
d4 d3 d2 d1 E
v v v v T
V 2 hist2
Y I + + + = (23b)
1 ' ' 1
} ] ) ( [ ) ( {
1 1
+ + + =
E E
D
D D E E I I
I
V V V 2
T xR T T U Z T T T T Y
1 W1 V
(24)
n
D D
E
D E D m
V T T U T T T v
V
V
V V V
] [ ] [
1 1
d1

+ + =

(25)
n
D D E D E D m 1 W1
I xR T T U Z T T T v
I I V V V
] ] [ [ ] [
' ' 1
d2
+ + =

(26)
t / ) ( ] [ 2

d3 m m
-1
x Z T T T v
W1 V V V
=
+
+
n
E
1 n
D D E D
E E

(27)
t / ) ( ] [ 2

m m
-1
4 d
x T T T v
V V V
=
+
+
n
E
1 n
D D E D
B B
(28)
1 1 + +
=
n
E
n
I I
l
(29)

Equations (17a) and (23a) represent Norton equivalent
circuits for the transmission line ends as shown in Fig. 2.



V
E
n+1
V
E
n+1
I
E
n+1
I
E
n+1
I
hist1
Y
1
+
-

V
E
n +1
V
E
n +1
I
hist 2
+
-
+
-
I
l
n +1
I
l
n +1
Y
2

(a) (b)
Fig.2. Norton equivalent circuits: (a) transmission line sending end
and (b) receiving end.

IV. NUMERICAL EXAMPLES
As an example consider a three-phase transmission line
that is excited by an Electromagnetic plane wave; the data of
the line are the following: length 325m, conductors radius
r=2.54cm, the maximum height of the line is 30m and the
height at middle span is 15m., the ground resistivity is 100
m. The geometry of the line is shown in Fig.3. At the
conductors end A loads of 400 are connected and at end B
the conductors are open ended. The electric and a magnetic
fields only vary with respect to time as shown in Fig. 5 and
Fig.6.
Figures 7 and 8 show the voltage waveforms obtained
at both ends of the 1st and 2nd phases. For both methods the
line was subdivided in 25 sections of length x =13m. To
obtain the voltage waveforms using EMTP, the incident field
effects are included by inserting equivalent current sources
between line subsections as is shown in Fig. 4. In the EMTP
simulations current i
B
flowing through R
eq
reproduces the
induced longitudinal distributed voltage source due to B
z
and
current i
E
takes into account the effects of E
y
.


(a)
z = 325 m.
RL
R0
RL R0
RL R0
d =3m
h=30m
h=15m
A B
1
2
3
B
z
E
y
B
z
E
y

(b)
Fig. 3. Transmission line geometry, (a) Transversal plane ,
(b) Longitudinal plane.

Fig.4. Inclusion of distributed source in the EMTP
0 1 2 3 4 5 6 7 8
x 10
-6
0
0.5
1
1.5
2
2.5
x 10
-6
TIME
M
A
G
N
I
T
U
D
E

Fig.5 Incident Electric Field.

0 1 2 3 4 5 6 7 8
x 10
-6
0
2
4
6
8
x 10
-15
TIME
M
A
G
N
I
T
U
D
E

Fig.6. Incident Magnetic Field.
0 1 2 3 4 5 6 7 8
x 10
-6
-5
0
5
10
15
20
x 10
4
TIME
M
A
G
N
I
T
U
D
E
EMTP
EMTP
CARACT
CARACT
B
A

Fig.7. Voltages response at the ends of the 1st phase.


0 1 2 3 4 5 6 7 8
x 10
-6
-0.5
0
0.5
1
1.5
2
2.5
x 10
5
TIME
M
A
G
N
I
T
U
D
E
EMTP
EMTP
CARACT
CARACT
B
A

Fig. 8. Voltages response at the ends of the 2nd phase.
As can be seen from Figs.7-8 both methods employed for the
simulations provide almost the same results. Here it is
important to note the difficulty involved when using the
EMTP. With EMTP the transmission line has to be divided in
several segments and each of them included as a
transmission line by itself. For a simulation the user has to
decide about the number of segments based on his expertise.
When a new simulation with a different number of segments
is necessary, the corresponding electrical parameters have to
be calculated again and a new data file (or network) has to be
created.
V. CONCLUSIONS
In this work the method of Characteristics has been
applied to the analysis of electromagnetic transients in non-
uniform multiconductor transmission lines illuminated by
incident electromagnetic fields. The proposed method do not
present numerical oscillations which are very common in
finite difference procedures. The results obtained with the
method of characteristics showed good agreement with those
from the EMTP. Results from the EMTP do not present
numerical oscillations, however, this simulation method is
very cumbersome and the accuracy of the results depends on
the expertise of the user.

VI. REFERENCES
[1]. H. W. Dommel, Digital Computer Solution of
Electromagnetic Transients in Single and Multiphase
Networks, IEEE Trans. on Power App. And Syst., pp.388-
398, April, 1969.
[2].- S. Carnerio, J. R. Mart, Evaluation of corona and Line
Models in Electromagnetic Transients Simulations, IEEE
Trans. on PWRD, Vol. 6, No. 1, Jan. 1991.
[3].- Amner Israel Ramrez Vzquez Transitorios
Electromagnticos en Lneas de Transmisin considerando
Parmetros Dependientes de la Frecuencia y Efecto Corona,
Master of Science in Electrical Engineering thesis, CUCEI-
University of Guadalajara, August 1998.
[4].- R. H. Galloway, W. B. Shorrocks and L. M. Wedepohl,
Calculation of electrical parameters for short and long
polyphase transmission lines, Proc. IEE, vol. III, No. 12,
December 1964.
[5].- Jos Alberto Gutierrez Robles, Desarrollo de un
Mtodo para el Anlisis de Lneas de Transmisin no
Uniformes y su Aplicacin al Estudio del Problema del
Flameo Inverso, Master of Science in Electrical
Engineering thesis, CUCEI-University of Guadalajara, June
1998.
[6].- J. L. Naredo, A. C. Soudack, J. R. Mart, "Simulation of
transients on transmission lines with corona via the method
of characteristics", IEE Proc. C, Gen., Transm. and Distrib.,
Vol. 142, No. 1, pp. 81-87, Jan., 1995.
[7].- Allen Taflove, Computacional Electromagnetics, The
finite-Difference Time-Domain Method. Artech House,
Inc., 1995.
[8]. K. S.Yee, Numerical Solution of Initial Boundary Value
Problems involving Maxwells Equations in Isotropic
Media, IEEE Trans. Antennas and Propagation, vol. 14,
1966, pp. 302-307.

[9].-R. L. Courant and D. Hilbert, Methods of Mathematical
Physics, Vol. II, Partial Differential Equations, New
Wiley-Interscience, 1962.

[10].-L. M. Wedepohl, Application of Matrix Methods to
the solution of travelling wave phenomena in polyphase
systems, Proc. IEE, Vol. 110, No.12, pp. 2200-2212,
December 1963.

[11].- J. L. Naredo, The Effect of Corona on Wave
Propagation on Transmission lines Ph.D. Thesis,
Department of Electrical Engineering Faculty of Applied
Science, The University of British Columbia, 1992.

[12].- J. Carlos Gutierrez V. Modelado de lneas de
transmisin monofsicas con parmetros dependientes de la
distancia y condiciones arbitrarias en los extremos para el
anlisis de transitorios electromagnticos., Master of
Science in Electrical Engineering thesis, CINVESTAV-IPN,
Guadalajara Unit, October, 2000.

Alejandro R. Chvez B., was born in Tepic, Mexico, on 1975. He
received the Electrical Engineering Degree from the Instituto
Tecnologico de Tepic, Tepic, Mexico in 1998. He is currently
working toward the M.S.E.E. degree at the Centro de Investigacion
y de Estudios Avanzados del Instituto Politecnico Nacional,
Guadalajara, Jalisco, Mexico. His research interests include
multiconductor transmission lines modeling for transient studies



Pablo Moreno V., was born in Cuernavaca, Mexico, on 1961. He
received the Electrical and Mechanical Engineering degree from the
Universidad Nacional Autonoma de Mexico, Mexico, in 1985, the
M.S.E.E. from the Instituto Tecnologico y de Estudios Superiores
de Monterrey, Monterrey, Mexico, in 1989 and the Ph.D. degree in
Electrical Engineering from Washington State University, WA,
USA in 1997.
In February 1998, he joined the Faculty of the Centro de
Investigacion y de Estudios Avanzados del Instituto Politecnico
Nacional, Guadalajara, Jalisco, Mxico. His research interests
involve electromagnetic transients analysis in electrical and
electronic systems and electromagnetic compatibility.

J. Luis Naredo V. , received the Electrical and Mechanical
Engineering degree from Universidad Anahuac, Mexico, in 1983
and the M. A. Sc. and Ph. D. degrees from the University of British
Columbia, in 1987 and 1992, respectively.
In May 1997, he joined the Faculty of the Centro de Investigacion
y de Estudios Avanzados del Instituto Politecnico Nacional,
Guadalajara, Jalisco, Mexico. His research interests involve
electromagnetic transients analysis in transmission lines,
telecommunications in power systems and digital protections..