Lightning Protection of Compact Transmission Lines

Abstract

Determination of the lightning performance of compact transmission lines is described. The tower top transient voltage, following a lightning stroke is determined using the EMTP(ATP) program. Computer models of towers, ground wires and conductors, and the earthing system are presented. Computer simulations include a sensitivity analysis relative to the waveform of the lightning current, stroke location, earth resistivity, type of tower and terrain profile

Full text

Page 1

INTERNATIONAL CONFERENCE ON POWER SYSTEMS TRANSIENTS
University of Washington, Seattle-WA
June 22-26, 1997
Lightning Protection of Compact Transmission Lines

Miguel O.B.C. Melo
Luiz C.A.Fonseca

Companhia Hidro Elétrica do São Francisco
Rua Delmiro Gouveia, 333
Recife - PE - 50761-901
Brazil

Eduardo Fontana S. R. Naidu

Universidade Federal de Pernambuco Universidade Federal da Paraíba
Depto de Eletrônica e Sistemas Laboratório de AltaTensão
Rua Acadêmico Helio Ramos s/n Av. Aprígio Veloso 882
Recife-PE 50740-530 C. Grande-PB 58109-970
Brazil Brazil



Abstract − Determination of the lightning performance of
compact transmission lines is described. The tower top
transient voltage, following a lightning stroke is
determined using the EMTP(ATP) program. Computer
models of towers, ground wires and conductors, and the
earthing system are presented. Computer simulations
include a sensitivity analysis relative to the waveform of
the lightning current, stroke location, earth resistivity,
type of tower and terrain profile.


Keywords: Lightning Transients, Compact Transmission Lines,
EMTP Modeling.


I. INTRODUCTION
CHESF-Companhia Hidro Elétrica do São Francisco,
responsible for the electric power generation and
transmission in the Northeast of Brazil, intends to expand its
transmission system in that region by installation of a number
of long, 230 kV and 500 kV, transmission lines (TL)[1]. In
this regard, feasibility studies were conducted to determine
costs and economical constraints involved in the project.
One of the consequences of these studies was that, use of
compact transmission lines would provide several benefits,
including, longer time periods between upgrades, reduction
of the shunt and series reactive compensations, as well as
increase of the transmission capability of existing lines.

A new type of compact transmission line, named High Surge
Impedance Transmission Line (HSIL)[2], with a higher level
of compactness, has been recently put into operation in a few
countries. The HSIL is a new concept of transmission line
design because it uses a combination of features such as,
distance reduction among conductors belonging to different
phases and increase of both, the number and relative
distances among sub-conductors of a single phase. In
addition, the HSIL uses asymmetrical bundles instead of
the symmetrical and circular distribution of subconductors,
employed in conventional compact lines. This new geometric
configuration equalizes and optimizes the electric field
distribution around all subconductors. The HSIL optimization
process allows obtaining a substantial reduction of the series
inductance as well as a significant increase of the shunt
capacitance, in turn producing a very high intrinsic
transmission line capability.
Recent development of the HSIL technology has limited
its use to a few countries and therefore, further development
and implementation of HSIL towers brings new challenges to
experts on transmission line studies and design[3]. The
purpose of this work is to analyze the lightning performance
of compact lines, including HSIL lines.

II. LIGHTNING STROKES ON POWER LINES

Studies related to the effects of lightning strokes on
power lines are of fundamental importance during the stage
of tower design, as inadequate choice of the electrical
parameters of the tower may lead to high tripping rates.
These studies determine clearances and shielding angles of
the tower and minimum distances required to bring flashover
probability down to acceptable levels. Because of the
complexity and random nature of the mechanism of lightning
discharges, studies in this field require a Monte Carlo
approach to carry out simulations. This method adequately
represents the randomness of the variables involved and has
been widely used in the literature to solve diverse problems,
e.g., solution of simultaneous equations, diffusion of neutrons
through materials, determination of probabilistic thermal
limits of power lines, to name a few examples. This method,
therefore, is adequate for simulating lightning current
intensity, wavefront characteristics, incidence angle of the
discharge, insulation strength on the tower and location of the
stroke.

Page 2

Two situations may occur, depending on the discharge
location relative to the wires composing the transmission
line. The first one, named a direct stroke, takes place when a
phase conductor is struck directly, in turn producing a
voltage increase in that phase. This may lead to a discharge
between the phase conductor and the tower, if the insulator
strength is exceeded. In this case, the operating voltage
maintains the discharge arc, causing a short circuit and
consequently a line tripout. This is commonly referred to as
a shielding failure. By studying direct strokes, it is possible
to obtain an effective shielding by proper distribution of
ground wires.

The second situation, named an indirect stroke, is
illustrated in Fig.1, and occurs when the discharge hits the
ground wire. Unlike the first situation, it is very difficult to
eliminate tripouts entirely. However, tripout occurrence can
be minimized by proper choice of tower clearances,
optimization of the coupling parameters between conductors
and ground wires, and by improving the tower grounding
project.

For the case of indirect strokes, voltage and current
traveling waves are generated at the discharge site,
propagating along the wires thereafter, until they reach the
adjacent towers. This in turn leads to the production of
reflected waves, with characteristics determined by the
relative values of the surge impedances involved in the
process. These traveling waves induce transient voltages on
the conductors, having shapes determined by the electric
coupling among conductors and ground wires. If the voltage
difference between conductor and tower exceeds the insulator
strength, a discharge occurs. This phenomenon is called
backflashover. It is important to point out that a
backflashover is much more likely to occur on insulators
than a flashover at midspan, because of the smaller distance
between phase conductor and ground at the tower relative to
that at midspan.

The voltage wave at the tower is called tower top
transient voltage and is dependent on the lightning current
and associated waveform, location of the stroke, and
transmission line parameters. When applied to the insulator
string it is given by [4],


( )()

V tk VV
stn
=−+
1

( )
t
( )
1


where, Vs(t) is the potential difference between ends of the
insulator, k is the coupling coefficient between conductor and
ground wire, Vt is the tower top transient maximum voltage
and Vn(t) is the instantaneous conductor voltage.
S
L
l
r
P
T
1


T2
Vs
I
I
CABO PARA-RAIOS
Conductor
CABO DE ATERRAMENTO
CABO CONDUTOR
Ground Wire
CABO CONDUTOR
CABO PARA-RAIOS
Conductor
CABO CONDUTOR

Fig.1. Representation of a lightning stroke on a TL.

performance
electrogeometric model developed by Whitehead [4], that
takes into account the lightning type and formation of a
conducting channel at the tip of the stroke. A diagrammatic
representation of the striking distances relative to the TL
wires is shown in Fig.2. According to this model, the
discharge occurs when the electric charge concentration of
the cloud exceeds the air strength, in turn propagating
through a thermally ionized gas column wich is shaped like a
cone. The cone radius corresponds to the strike distance Sd,
illustrated in Fig.2, and is a function of the lightning current.
The approach used in this work to determine
of compact lines is based on the
Ground Wires
Sd
Sd
Sd
Sd
Sd
Conductor
Conductor

Fig.2.Diagramatic representation of the striking distances relative to
the TL, according to the electrogeometric model.

Ground Wire
Conductor
S
T1

Page 3

III. METHOD OF ANALYSIS

the transmission line system, under lightning, is described.
The model is then applied for determination of the
performance of compact transmission lines, including the
recently developed HSIL system.
In this Section, modeling of the several components of
A. Transmission Line Modeling

according to Fig.3, that shows two circuit diagrams
accounting for discharges occurring at the tower and at
midspan, respectively [4-5].
The elements of the transmission line are modeled
Zgw
Zgw
Zgw
Zgw
Zgw
Zgw
Zt
Zt
Zt
Zt
Zt
Zt
Ztg
Ztg
Ztg
Ztg
Ztg
Ztg
I
Zt
Ztg
Zgw
Zgw
Zgw
Zgw
Zgw
Zgw
Zgw
Zgw
Zt
Zt
Zt
Zt
Zt
Zt
Ztg
Ztg
Ztg
Ztg
Ztg
Ztg
Zgw1
I
Zgw1
Zgw1
Fig.3. Equivalent circuits for a lightning discharge striking the tower
(upper diagram) or midspan(lower diagram).

lightning current I, the equivalent ground wire surge
impedance Zgw, and the ground wire, tower, and tower
grounding, surge impedances Zgw1, Zt and Ztg, respectively. It
is worth noting that Zgw=Zgw1, if the tower has a single ground
wire.
Parameters used to represent the circuits in Fig.3 are the
B.Tower Ground Modeling

model suggested by Bewley[6] is employed. In this model
the parameter Ztg, is made equivalent to a series resistor in
parallel with an RL circuit, as shown in Fig.4. In the
equivalent circuit, when a unit step voltage is applied to the
wire underneath earth, here represented by the counterpoise
wire illustrated in Fig.5, its impedance varies over time,
according to the equation,
To represent the tower grounding surge impedance Ztg, a
( )
t
()
( )
ZRZR
tgdsd
tv
l
=+−
−2 2
where, Rd and Zs are the values of Ztg for t→∞ and for t = 0,
respectively. In Eq.(2), v represents the current wave speed
on the counterpoise wire of length l.
(Zs-Rd)
L = 2l(Zs-Rd)
v
Rd

Fig.4. Tower grounding equivalent circuit.
l
counterpoise wire

Fig.5. Typical earthing configuration of the TL.

expression[6],
Parameter
Rd
is obtained from the following

( )
R
l
l
a
l
s
s
l
s
l
s
l
d=++−


+−



ρ
π
2
22
2912 1071.
0645.0145.
2
2
4
4
lnln.
3

where a is the wire diameter, s is twice the counterpoise
depth and ρ is the soil resistivity.
Calculated values of the equivalent circuit parameters are
listed in Table 1, for typical values of soil resistivity and
4AWG copperweld wire.
Tab 1. Parameters of the tower grounding equivalent circuit.
ρ(Ω×m)
500
1500
(Zs−Rd)[Ω]
3.5
57.1
L [µH]
1.05
59.5
C.Lightning Current Wave

waveshape, as illustrated in Fig.6. Parameter Td is the time
required for the current amplitude to fall back to half of its
maximum value, and Tf is the rise time.
Two linear functions are used to represent the current
I
If
Td
tTf
If /2

Fig.6. Model function describing the lightning current wave.

characterizing the lightning current are typically Td ≈50µsec
and Tf < 5 µsec [4].
Experimental observations indicate that the parameters

Page 4

D.Tower Surge Impedance

the self-supporting 500 kV (with one and two ground wires)
and the cross rope chainette (500kV) and HSIL (230kV).
Figures 7, 8 and 9 represent these tower configurations [5,7].
Three types of compact towers were selected, namely,
29. 0
9. 30.
9. 0
6, 40

Fig.7. Self supporting 500kV(one and two ground wires).

Fig.8. Cross rope chainette 500kV.
Fig.9. Cross rope chainette HSIL 230 kV.

To represent the tower surge impedance, it is necessary to
consider the propagation time of the current wave along the
tower structure for each of the configurations shown in Figs.7
through 9. A simple model of this parameter is given in
Ref.[4], yielding the tower diagram models illustrated in
Figs. 10 and 11, along with the expressions for the surge
impedance.




Zh
r
t=





 −






602
2
1 ln



Fig.10. Model of self supporting tower and corresponding surge
impedance expression.
Z
ZZ
t=
+
12
2

Z
h
r
h
r
h
b
1
609060
=





+

+





−

−
ln
Z
bh
2
609060
=
ln



Fig.11. Model of cross rope chainette 500kV and HSIL 230kV
towers and corresponding impedance expression.
Table 2 lists the calculated values of tower surge
impedances considered in this work.

E .Equivalent Ground Wire Surge Impedance

by[4],
The equivalent ground wire surge impedance is given
( )
Z
h
r
gw
gw
eq
=






60
2
ln , 4
where hgw is the average ground wire height, that is a function
of the terrain profile, and req is the equivalent radius of the
ground wire combination. Both of these parameters are
obtained according to the procedure outlined in Ref.[4].
Calculated values of the ground wire surge impedances,
for the tower types and dimensions shown in Figs. 9 through
11, are listed in Table 2.
Tab 2. Calculated values of the parameters Zt and Zgw for the tower
types analyzed in this work.
Tower type
Self supporting (1 ground wire)
Self supporting (2 ground wires)
Cross rope chainette (500 kV)
Cross rope chainette (HSIL 230kV)
Zt(Ω)
173
176
108
100
Zgw (Ω)
557
345
307
304

Page 5

F.Lightning Impulse Insulation Strength

for the distinct tower air gaps [4]. This information, along
with the parameters calculated earlier, are used to determine
lightning performance and voltage transients as described in
the next Sections.
Figure 12 depicts the lightning impulse withstand curves


Fig.12. Lightning impulse insulation strength.

IV. TOWER TOP TRANSIENT VOLTAGE

distinct tower types investigated in this work, the tower top
transient voltage was studied using the EMTP(ATP)
program, for a fixed value of the peak discharge current.
Fixing the peak current in these simulations allowed to
investigate how the voltage transient could be affected by
changing the remaining four parameters, namely, lightning
rise time, soil resistivity, tower type and stroke location.
Calculations were carried out by setting the peak lightning
current at 10 kA, as approximately 80% of observed
discharges are known to produce peak currents exceeding this
value[4]. In order to obtain the results 48 simulations were
carried out using the EMTP(ATP) program.
Figures 13 through 16 illustrate the behavior of the tower
top transient voltage, obtained by varying a single parameter.
The three plots shown in Fig.13 were obtained by setting
Td=50µsec, and attributing the values of 1µsec, 3µsec and
5µsec, to the parameter Tf. One can notice that the voltage
transients follow basically the lightning current waveform
during risetime with undulations occurring during the
lightning current decay time. It is worth noting the changes
of approximately 40% to 60% in maximum voltage for the
distinct curves.
Prior to determining lightning performance for the

influences both the shape and the maximum value of the
parameter Vs. Increasing the soil resistivity from 500 to 1500
Ω×m produces approximately 50% increase of the maximum
voltage. This shows that the grounding modeling plays an
important role in lightning performance studies. Figures 15
and 16 illustrate the effects of changing the type of tower
Figure 14 shows that the soil resistivity greatly
and location of the stroke. For the curves shown in Fig.15,
the largest change observed in the maximum value of Vs is
approximately 25%, while location of the stroke influences
the time lag for onset of the voltage transient, without
significant variation in the maximum value.

Fig.13. Tower top transient voltage dependence on the lightning
wavefront.

Fig.14. Tower top transient voltage dependence on the soil
resistivity.

Fig.15. Tower top transient voltage dependence on tower type.

Fig.16. Tower top transient voltage dependence on stroke location.

Page 6

V.CALCULATION OF LIGHTNING PERFORMANCE

form of tripouts/(100km×year), was done by use of the
Monte Carlo method. To carry out computer simulations, use
was made of the field data on the yearly incidence of
lightning discharges in the Northeast region of Brazil, as well
as TL geometric characteristics together with the
corresponding data relative to the type of terrain profile along
routes. Values of the remaining parameters were selected
according to their associated probabilities of occurrence. The
lightning current peak amplitude was associated to the
accumulated probability function given by[4],
1
Lightning performance determination, expressed in the
()
( )
P I
(
I
I
f
)
/
≥=
+
1 25
2 5

where If is the peak current and I is the independent variable,
both expressed in kA. The parameter Td was set at 50 µsec,
and the rise time Tf was randomly chosen from the set of
values (1 µsec, 3µsec and 5 µsec). Field data in the region
under study indicated a typical variation in soil resistivity
between 500 and 1500 Ω×m and a gaussian probability
density function was adapted to allow statistical selection of
this parameter. Probabilities associated to the stroke location
along the line as well as the lightning discharge incidence
angle were obtained from typical experimental data reported
in Ref.[4]. For each studied case at least 10000 simulations
runs were conducted.
The studies have taken into account corona effects by
using the model reported in Ref. [4]. According to that
model the cable radius has to be corrected by use of the
expression Rc= k1v2 + k2v + r, where, Rc is the corrected cable
radius, k1 and k2 are coefficients that are dependent on the
cable height, v is the voltage and r is the cable radius .
Simulation results expressed in terms of tripout rates are
listed in Table 3, for the tower configurations studied in this
work, for distinct types of terrain profiles, including the case
of a typical terrain profile in the Northeast region of Brazil.
The value adopted to the keraunic level was 15.
Note that the tripout rate listed in Table 3 show a
decrease as the terrain profile is varied from flat to
mountainous. This occurs because the distance between
conductor and ground wire decrease as the profile deviates
from the flat condition, causing an increase in the value of k.
As can be demonstred from (1), this effect will produce a
decrease in the voltage Vs applied on insulator string

An examination of the values listed in Table 3, indicate
that all rates fall around 3 tripouts/(100km×year), a
maximum limit generally adopted as standard by many
companies of electric power generation and transmission. It
is also worth noting that lightning performance, for the types
of tower investigated in this work, can be greatly influenced
by the terrain profile. Finally, a comparison of the tripout
rates listed in the second and third columns of Table 3,
indicate that use of a single ground wire instead of two, in the
self supporting tower, is technically feasible.
Tab 3. Tripout rates of compact lines [tripouts/(100km×year)].
Profile of
Terrain Supporting
(1 g wire)
Flat 0,8
Semi Hilly 0,7
Hilly 0,7
Mountainous 0.2
Typicala 0,7
a 55% Flat, 20% Semi Hilly, 20% Hilly and 5% Mountain
Self
Self
Supporting
(2 g wire)
0,5
0,3
0.2
0.0
0,4
Cross
Rope
500 kV
0,6
0,4
0,3
0.0
0,5
Cross
Rope
HSIL
3,1
2,7
2,4
1.8
2.9
VI. CONCLUSIONS

according to the following conclusions:
1-The tower top transient voltage is influenced by variations
of the lightning current wavefront, soil resistivity and tower
type, as indicated in Figs.13 through 16.
2-Lightning transient studies indicate that it is necessary to
establish a suitable modeling of the grounding system as well
as to include the soil resistivity variation in the lightning
performance studies.
3-It is very important to take into account the type of terrain
profile in studies concerning lightning performance of
transmission lines.
4-Compact towers are technically feasible due to their good
lightning performance, being recommended for consideration
in expansion planning studies.
5-Use of self supporting towers using a single ground wire is
also technically feasible, and their use could reduce TL
installation costs.
The results obtained in this work can be summarized
VII. REFERENCES
[1] Miguel B. C. Melo, D. Brasil, W. Sato et al, "Viability
Studies of Application of Compact 500 kV Transmission
Lines on the CHESF (Brazil) Systems", Leningrad
Symposium on Compact Overhead Lines, CIGRE, 1991.
[2] G. N. Alexandrov,"Scientific and Engineering Principles
of Creating Compact Lines with Increased Natural Capacity"
Leningrad Symposium on Compact Overhead Lines, 1991.
[3] Oswaldo Regis et al, "Unconventional Lines of High
Natural Power Rating-An Exercise in Prospection in 69 kV
and 138 kV", (in Portuguese), V ERLAC, CIGRE-BR, 1993.
[4] "Transmission Line Reference Book 345 kV and Above",
Second Edition, EPRI, Palo Alto, USA, 1982.
[5] M. Darveniza et al, "Modeling for Lightning Performance
Calculations", IEEE Transactions on PAS, Vol 98,
November/December 1979, pp. 1900-1908.
[6] L. Bewley, "Traveling Waves on Transmission Systems",
John Wiley & Sons, New York, 1951.
[7] Miguel B. C. Melo, "Study of Lightning Protection of
Compact Lines", 22nd International Conference on Lightning
Protection, Budapest, Hungary, 1994.