CALCULATION OF ENERGY STRESS ON SURGE ARRESTERS IN

275kV TRANSMISSION LINES

R. Bhattarai
*
, H. Griffiths, N. Harid and A. Haddad
High Voltage Energy Systems Group, School of Engineering, Cardiff University, UK
*Email: BhattaraiR@Cardiff.ac.uk

Abstract: Application of line surge arresters is found to be an efficient tool to improve the
lightning performance of transmission lines. Suitable selection of arrester rating and configuration
along the line are crucial for achieving improved reliability of the line. This paper presents a
systematic calculation of energy stress carried out on gapless Zinc-Oxide surge arresters applied to
a 275 kV double circuit shielded transmission line. An operational line was modelled in TFlash,
and simulations were carried out under conditions of lightning strokes hitting the phase conductor
and the shield wire. To determine arrester characteristics, a number of important parameters that
influence energy stress calculations were evaluated. The arrester failure performance analysis was
carried out to estimate the overall reliability of the transmission line. Stroke to phase conductor
was found to be the main source of arrester failure in the line.


1. INTRODUCTION
Faults caused by lightning are the main source of line
outages especially in areas with high ground flash
density, high earth resistivity and poor shielding.
Previous studies [1-3] have shown that lightning
performance can be guaranteed by careful selection and
location of zinc-oxide (ZnO) surge arresters. These line
arresters are exposed to high-magnitude lightning
strikes and have to survive higher energy discharge duty
imposed by the lightning current.
In comparison to the substation arrester, the line arrester
may experience more energy stress. This is because the
incoming surge to a station is limited either by line
insulator flashover or by the discharge to earth through
shield wire. Therefore, adequate selection of a line
arrester also depends upon assessing its energy
absorption capability so that it does not fail under
conditions of lightning striking either the phase
conductor or the shield wire.
In this study, a systematic calculation of energy stress
was carried out on gapless ZnO surge arresters installed
on a 275 kV double circuit shielded transmission line.
The line was modelled in EPRIs TFLASH program
which is designed to examine all arrester options and
their potential benefits to improving line performance
[4]. The simulation was carried out under conditions of
lightning stroke hitting the phase conductor (shielding
failure) and the shield wire (which may lead to
backflashover). In both cases, the energy stress in the
arrester depends on the line and lightning stroke
parameters. A parametric analysis was carried out to
select the appropriate arrester, considering different
parameters that affect its energy calculation. Using these
studies, it is possible to determine an optimum
application of surge arresters, and make a more accurate
selection of arrester rating in terms of protective level
and energy stress. Furthermore, arrester failure
performance analysis was carried out to assess the risk
of failure due to lightning strikes onto the phase
conductor and shield wire of surge arresters installed on
the line. A statistical simulation was performed to
calculate the risk of arrester failure on the line.
2. SIMULATED DATA
2.1. 275kV transmission line
An existing 35 km long, 275 kV double circuit shielded
line was considered in this study. The line with 300 m
span length was assumed to be located on flat terrain
with ground flat density of 0.5 flashes per kilometre
square per year (flashes/km
2
/year). Figure 1 show the
tower structure and conductor geometry in each tower
of the line.
Twin 175 mm
2
Lynx type ACSR conductors were
used in phase conductors and a single Lynx ACSR
conductor was used as shield wire. The diameter of each
conductor is 19.53 mm, and the bundle spacing of 30.48
cm was used for the twin phase conductors.
A 3.31 m long line insulator string composed of 16
individual glass insulator disc producing an overall
critical flashover voltage (CFO) of 1646 kV was used.
2.2. Surge arrester
The following specifications of zinc-oxide surge
arresters were used.
Nominal discharge current : 10 kA
Max. continuous operating voltage (MCOV) : 220 kV
Energy capability : 7.8 kJ/kV of MCOV
Table 1 summarises the arrester V-I curve under an 8/20
impulse.
Table 1: Arrester discharge voltage for 8/20 impulse
current.
I (kA) 3 5 10 15 20 40
V(kV) 581 601 635 666 690 762
ISBN 978-0-620-44584-9
Proceedings of the 16
th
International Symposium on High Voltage Engineering
Copyright c 2009 SAIEE, Innes House, Johannesburg
Pg. 1 Paper G-11
3. TFLASH MODEL
Modelling transmission line components in TFLASH is
facilitated with its existing database of towers, earth
types, insulators, conductors and arresters [5]. Facilities
are provided within the program to allow further input
and modifications of the above components if required.
A brief description of the line model used in this study
is given in the following sections.
3.1. Line model
Each span on the transmission line is represented as a
multiphase untransposed distributed parameter line
section. In order to avoid reflections in the line,
matching impedances were used at both terminations of
the line. Each span section is further divided into
smaller sub sections to enable stroke simulation at a
number of points along the spans. In this model, 8
towers from the middle of the line were used in each
direction with lightning surge striking at a mid tower
position.
A high-voltage transmission tower can be considered as
a network of short transmission lines carrying transient
current from its top to the earth and its reflection back
towards the top [5]. Therefore, in this study, the tower
itself is modelled as a short vertical transmission line
section with constant surge impedance describing the
voltage produced on the tower, per unit current flowing
through it. The steel lattice tower has surge impedance
of 173.1 calculated using Equation (1) [6].

(
(

|
|

\
|
+
=

2 1
avg 1
T
h h
r
tan 5 . 0 cot ln 60 Z (1)
where, h
1
and h
2
are the tower heights from base to
midsection and midsection to the tower top respectively
and r
avg
is the weighted average tower radius given by
Equation (2).

( )
) h h (
h r h h r h r
r
2 1
1 3 2 1 2 2 1
avg
+
+ + +
= (2)
where, r
1
, r
2
and r
3
are the radii at the top, midsection
and base of the tower respectively.
The tower footing resistance plays an important role in
calculating the arrester energy. A non-linear tower
footing resistance model as shown in Figure 2 was used.
The resistance is calculated using Equation (3) [7].

g
lc
i
I
I
1
R
R
+
= (3)
With;
2
lc
g
g
R 2
E
I

= (4)
where, R
lc
is the low-current tower footing resistance,
is the soil resistivity, I is the stroke current through the
tower footing and, E
g
the soil ionisation critical
electrical field (4 kV/cm).
In this study, a low current tower footing resistance of
10 with soil resistivity of 200 m was used in case of
a stroke hitting the phase conductor, and a resistance
value of 80 with soil resistivity of 1600 m was used
for the case of a stroke hitting the shield wire at the
tower top.

3.2. Arrester model
In TFLASH, arresters are modelled as nonlinear
voltage-controlled current sources, and the arrester
current is calculated based on the applied voltage across
the arrester [5]. Therefore, the resulting equivalent
circuit of an arrester installed between a tower structure
and a line conductor is shown in Figure 3.
The tower and line conductor node voltages are given
by Equations (5) and (6) respectively.
30.48 cm
19.24 m
(12.19) m
25.33 m
(18.26) m
31.42 m
(24.37) m
36.88 m
(30.22) m
4.57m
4.26m
4.03m
22.55 m
6.09 m
6.09 m
2.15 m
A1
A2
B1 B2
C1
C2
E
Figure 1: 275kV shielded double circuit transmission
line tower. Values in brackets are mid-span heights.
I
Rlc

Figure 2: Non-linear tower footing resistance model.
ISBN 978-0-620-44584-9
Proceedings of the 16
th
International Symposium on High Voltage Engineering
Copyright c 2009 SAIEE, Innes House, Johannesburg
Pg. 2 Paper G-11
T
A
S T
Z
2
I
I V |

\
|
= (5)
L
A
L
Z
2
I
V |

\
|
= (6)
3.3. Lightning stroke parameter model
A range of different lightning impulse shapes were used
in simulating lightning strikes to the transmission line.
In this study, a double exponential, 4/77.5 impulse
current wave as recommended by CIGRE [6] was used.
CIGRE also recommends specific peak values of
lightning current for the different simulations: a peak
current magnitude of less than 20 kA under shielding
failure scenario and currents above 20 kA for
backflashover scenarios [6]. In order to evaluate the
maximum energy absorbed by surge arresters, unless
otherwise specified, a 20 kA lightning stroke hitting the
phase conductor A
1
was used for the shielding failure
case and a 200 kA stroke hitting the tower-top was used
in the backflashover case.
4. ENERGY STRESS ON SURGE ARRESTERS
Based on the line lightning performance analysis carried
out in a previous study [1], the maximum energy
absorbed by a line surge arrester was calculated using
the product of voltage and current traces computed by a
travelling wave simulation technique [5]. In this
investigation, the insulator flashover was neglected
since the voltage measured across the line insulator
which is protected by the surge arrester was found to be
much lower than the CFO of the insulator string even
under high magnitude lightning strike. Therefore, it is
assumed that the insulator does not flashover when it is
protected by a surge arrester.
Figure 4 shows energy stress distribution in surge
arresters installed in all phases in the case when a
lightning strike hits a phase conductor or a shield wire.
When a low current lightning strike hits a phase
conductor, the energy absorbed by arresters at any tower
is different. As expected, the arrester installed on a
stricken phase absorbs excessively high energy
compared with arresters on other phases. However,
when high current lightning hits the shield wire, it was
shown that any two arresters installed at the same height
absorb equal energy. The arresters installed on the
bottom phases absorb more energy than the arresters
installed on the other four phases above.
In a previous study [1], it was shown that the top phase
conductors are more vulnerable to shielding failure
while those at the bottom phases are more vulnerable to
backflashover. Considering this case and the results
shown in Figure 4, arrester energy requirements can be
such that the top arresters are more likely to experience
direct strikes of lower magnitudes while the lower ones
can be subjected to stresses equivalent to those causing
backflashover.
4.1. Parametric analysis
Appropriate selection of an arrester as a function of its
energy stress depends upon different parameters. These
parameters can be classified as line parameters and
lightning stroke parameters. Parameters such as
arresters in adjacent towers, tower footing resistance
and angle of power frequency voltage are considered as
line parameters whereas stroke peak current magnitude,
front time and tail time are considered as lightning
stroke parameters. Influence of each of these parameters
on arresters energy stress was analysed in this
investigation.
Influence of line parameters
In practice, the energy shared by arresters at a tower is
highly influenced by the presence of arresters at
neighbouring towers. The nature of influence of these
adjacent arresters on the struck arresters energy
stress depends upon the position of the lightning stroke
hitting the line. Figure 5 shows the energy discharged
by an arrester as a function of number of arresters in
neighbouring towers. It is clearly understood that, when
stroke hits the phase conductor, the neighbouring
arresters help in sharing some of the duty and, hence,
this reduces the energy stress on the arrester at the
struck tower. However, the case is different when high
current lightning stroke hits the shield wire or the tower
top. In this case, the energy absorbed by the arresters at
the tower increases with increasing number of arresters
at the adjacent towers. This is explained by the current
passing through the adjacent arresters being of opposite
IA
ZL ZL
ZT ZT
IS
IA/2 IA/2
IA/2
IA/2
VT
VL
Surge
Arrester
Figure 3: Equivalent circuit of an arrester installed
between a line conductor and a tower structure. Z
T
=
Tower surge impedance, Z
L
= Line surge impedance, I
S

= Stroke current, and I
A
= Arrester current
0
10
20
30
40
50
60
70
80
90
A1-C2 B1-B2 C1-A2 A1-C2 B1-B2 C1-A2
Arrester Position in Phase
A
r
r
e
s
t
e
r

E
n
e
r
g
y

[
k
J
]
20 kA
10
Stroke to shield wire Stroke to phase conductor
200 kA
80
Figure 4: Distribution of energy stress in surge arresters
installed at a stricken tower.
ISBN 978-0-620-44584-9
Proceedings of the 16
th
International Symposium on High Voltage Engineering
Copyright c 2009 SAIEE, Innes House, Johannesburg
Pg. 3 Paper G-11
polarity, and hence flows back to the striking point
resulting into the increase of energy absorbed by the
arrester at the struck tower [8].
Figure 6 shows the percentage of energy shared by
arresters at adjacent towers, when lightning strikes a
phase conductor and shield wire. Therefore, it can be
said that when lightning hits the phase conductor, the
nearest arresters to the strike absorb almost 40% of
energy, and the arresters on immediate adjacent towers
absorb 28%. This value decreases to only 19% for the
arresters at far end on the line section considered in this
study. However, when lightning hits the shield wire, the
nearest arresters absorb only 10%, while the arresters at
the far end does not absorb any energy.
The energy discharge duty of surge arresters depends on
tower footing resistance. Figure 7a shows this effect
when lightning strikes of different current magnitude hit
the phase conductor. Figure 7b shows the case when
200 kA lightning strike hits the shield wire. The tower
footing resistance was varied from 10 to 80 . For
stroke to a phase, the energy absorbed by the arrester on
the phase decreases with increasing footing resistance.
For a stroke to shield wire, however, high value of
footing resistance increases the arrester energy
discharge. In this case, arresters installed on the top
phases are more stressed with low tower footing
resistance value whereas bottom phase arresters are
more stressed in the case of high footing resistance
value.
Figure 8 shows the effect of power frequency voltage
angle on arrester energy stress. Significant influence of
this angle is seen on both the case of lightning hitting
the phase conductor and shield wire. When stroke hits
the phase conductor, the maximum energy in struck
arrester is obtained at voltage angle of 0
o
. In case of
stroke hitting the shield wire, the maximum energy is
found at voltage angle of 180
o
. With change in voltage
angle, the energy stress in arrester increases when stroke
hits the shield wire, but the energy discharge in this case
is rather low and is unlikely to exceed the maximum
energy absorption capability of the arrester.
Influence of lightning stroke parameters
Lightning stroke parameters have significant influence
on energy stress in line surge arresters. This can be
considered as one of the key factor for selection of the
arrester. The oscillographic simulation was carried out
to understand the effect of these parameters for both the
case of stroke hitting a phase conductor and the shield
wire.
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
-8 -7 -6 -5 -4 -3 -2 -1 0 1 2 3 4 5 6 7 8
Tower Number
A
r
r
e
s
t
e
r

E
n
e
r
g
y

[
%
]
20 kA
10
200 kA
80
Figure 6: Percentage (with respect to energy absorbed
by arrester at tower hit by lightning) of energy shared by
adjacent arresters at towers along the line.
Figure 5: Arrester energy as a function of adjacent
arresters.
0
100
200
300
400
500
600
700
800
900
1000
1 3 5 7 9 11 13 15 17
Number of Towers with Arresters
A
r
r
e
s
t
e
r

E
n
e
r
g
y

[
k
J
]
0
5
10
15
20
25
30
35
40
45
50
A
r
r
e
s
t
e
r

E
n
e
r
g
y

[
k
J
]
200 kA

10
10 kA
20 kA
10
0
5
10
15
20
25
30
35
0 10 20 30 40 50 60 70 80
Tower Footing Resistance [ ]
A
r
r
e
s
t
e
r

E
n
e
r
g
y

[
k
J
]
Arr A1-C2
Arr B1-B2
Arr C1-A2
200 kA
80
200 kA
80
200 kA
80
b: Stroke to shield wire
Figure 7: Arrester energy dependence on tower footing
resistance
0
20
40
60
80
100
120
140
160
180
0 10 20 30 40 50 60 70 80
Tower Footing Resistance [] ] ] ]
A
r
r
e
s
t
e
r

e
n
e
r
g
y

[
k
J
]
10
20 kA
10 kA
a: Stroke to phase conductor
Figure 8: Arrester energy dependence on power
frequency voltage angle at lightning strike.
0
20
40
60
80
100
120
140
160
180
200
0 60 120 180 240 300 360
Power Frequency Voltage Angle in A1 [] ] ] ]
A
r
r
e
s
t
e
r

E
n
e
r
g
y

[
k
J
]
200 kA
80
20 kA
10
ISBN 978-0-620-44584-9
Proceedings of the 16
th
International Symposium on High Voltage Engineering
Copyright c 2009 SAIEE, Innes House, Johannesburg
Pg. 4 Paper G-11
Figure 9 shows the effect of stroke peak current
magnitude on arrester energy. This effect was examined
for different tower footing resistances. The energy
absorbed by the arrester increases with increasing peak
current magnitude, and this is obtained for all cases of
impact point and tower footing resistance.
The effect of impulse front time for different stroke
peak current magnitudes is shown in Figure 10. In the
case of lightning hitting the phase conductor, the change
in front time does not have any influence on the arrester
energy. However, the arresters are less stressed with
high front time when lightning hits the shield wire. On
the other hand, the stroke current tail time has
significant influence on the energy absorbed by line
arresters (Figure 11). The arrester energy increases with
increasing tail time of the lightning impulse.
5. ARRESTER FAILURE PERFORMANCE
The objective of this study is to estimate the failure rate
of arresters due to excessive energy absorption and to
guarantee that the arresters installed on the line have
sufficient energy capability to withstand lightning
strikes to the phase conductors or to the shield wire. To
determine the arrester failure probability, the integrated
energy for each arrester is used with the failure
probability curve from EPRI report 1000461 [9].
The statistical simulation method was used. To integrate
the energy through the arresters over most of the stroke
duration, the method used in TFLASH software adopts
different simulation time limits for strokes to phase
conductor and shield wire. These time limits are much
longer than the flashover statistics time limit (500 s for
a stroke to phase conductor and 100 s for a stroke to
shield wire). Figure 12 shows an example of a typical
waveform used for energy calculation for a 20 kA
stroke current. In this case, the insulation flashovers
were disabled.
-5
0
5
10
15
20
25
30
35
40
45
0 25 50 75 100 125 150 175 200 225
Stroke Peak Current Magnitude [kA]
A
r
r
e
s
t
e
r

E
n
e
r
g
y

[
k
J
]
50
20
80
Figure 9: Effect of stroke peak current magnitude
a: Stroke to phase conductor
b: Stroke to shield wire
-20
0
20
40
60
80
100
120
140
160
180
0 2 4 6 8 10 12 14 16 18 20
Stroke Peak Current Magnitude [kA]
A
r
r
e
s
t
e
r

E
n
e
r
g
y

[
k
J
]
80
10
40
0
20
40
60
80
100
120
140
0 1 2 3 4 5 6
Front Time [ s]
A
r
r
e
s
t
e
r

E
n
e
r
g
y

[
k
J
]
150 kA
200 kA
250 kA
80
Figure 10: Effect of front time (Tail time = 77.5s)
a: Stroke to phase conductor
b: Stroke to shield wire
0
20
40
60
80
100
120
140
160
180
0 1 2 3 4 5 6
Front Time [ s] ]] ]
A
r
r
e
s
t
e
r

E
n
e
r
g
y

[
k
J
]
5 kA
10
20 kA
10 kA
0
50
100
150
200
250
0 25 50 75 100 125 150 175
Tail Time [ s]
A
r
r
e
s
t
e
r

E
n
e
r
g
y

[
k
J
]
10
5 kA
10 kA
20 kA
0
10
20
30
40
50
60
70
80
0 25 50 75 100 125 150 175
Tail Time [ s]
A
r
r
e
s
t
e
r

E
n
e
r
g
y

[
k
J
]
80
250 kA
200 kA
150 kA
Figure 11: Effect of tail time (Front time

= 4s)
a: Stroke to phase conductor
b: Stroke to shield wire
0
5
10
15
20
25
0 50 100 150 200 250 300 350 400 450 500
Time [ s]
C
u
r
r
e
n
t

[
k
A
]
For 20kA stroke current
Front time = 3.4 s
Tail time = 56.2 s
Figure 12: Equal probability waveform (20kA stroke
current)
ISBN 978-0-620-44584-9
Proceedings of the 16
th
International Symposium on High Voltage Engineering
Copyright c 2009 SAIEE, Innes House, Johannesburg
Pg. 5 Paper G-11
The arrester failure performance was analysed for the
whole line section (35 km) when arresters are installed
at each phase and at every tower. Table 2 gives a
summary of the arrester failures for the 35 km line. It
can be seen that the phase conductors are expected to be
hit by 3.704 strokes per year resulting in an arrester
failure rate of 0.128 failures per year, i.e. one arrester
may fail every 7 to 8 years. Since there are hundreds of
arresters installed on the line, the chances of the same
arrester failing again is very low. The table also shows
the failure rate on each phase. It can be seen that there is
no risk of arrester failure for stroke terminating on the
shield wire or tower top but there is risk associated to
the arresters at the top phase (installed on phases A
1
and
C
2
) with direct strokes terminating on the phase.
Figure 13 shows the arrester failure rate at towers along
the section of the line. The maximum failure rate is very
low; e.g. an arrester at the same tower may fail every
815 years. This of course depends on the appropriate
selection of the arresters.
Table 2: Arrester failure performance of 35km line.
(Values in per year basis)
Direct strikes per year = 3.704
Arrester failure per year = 0.128
Arrester Failure by Phase
Failure From
Phase
Direct
Strikes
Shield
Strikes
Phase
Strikes
All
Strikes
A1 0.269 0.000 0.065 0.065
B1 0.000 0.000 0.000 0.000
C1 0.000 0.000 0.000 0.000
C2 0.269 0.000 0.065 0.065
B2 0.000 0.000 0.000 0.000
A2 0.000 0.000 0.000 0.000



6. CONCLUSIONS
Energy stress analysis of ZnO surge arresters installed
on a 275 kV double circuit transmission line was
investigated. It was found that the energy requirements
on the line arresters were moderate. The energy
absorption studies were carried out for the line and
stroke parameters which are essential in the selection
process of line surge arresters.
Lightning strikes hitting the phase conductors (shielding
failure) were found to be the main source of the risk of
failure. In this case, the surge arresters installed on other
phases on the same tower did not help to share the total
surge energy. The risk of failure when lightning hits the
shield wire or tower was found to be zero.
7. ACKNOWLEDGMENT
The authors wish to thank the Engineering and Physical
Sciences Research Council (EPSRC) for financial
support.
8. REFERENCES
[1] R. Bhattarai, R. Rashedin, S. Venkatesan, A.
Haddad, H. Griffiths and N. Harid, Lightning
performance of 275kV transmission lines, in Proc.
of 43
rd
International Universities Power
Engineering Conference, Padova, Italy, Sept. 2008.
[2] J. A. Tarchini and W. Gimenez, Line surge arrester
selection to improve lightning performance of
transmission lines, in Proc. of IEEE PowerTech
Conference, vol. 2, Bologna, Italy, June 2003.
[3] J. A. Martinez and F. Castro-Aranda, Lightning
flashover rate of an overhead transmission line
protected by surge arresters, IEEE Power
Engineering Society General Meeting, Tampa,
Florida, USA, June 2007.
[4] EPRI, Lightning performance analysis of pacific
power company cascade craft substation, Draft
Report, EPRI, Palo Alto, Exhibit No: WWB-4,
Nov. 2006.
[5] EPRI, Handbook for improving overhead
transmission line lightning performance, EPRI,
Palo Alto, CA:2004. 1002019, Dec. 2004.
[6] CIGRE WG 33-01, Guide to procedures for
estimating the lightning performance of
transmission lines, CIGRE Brochure 63, Oct.
1991.
[7] British Std. BSEN 60071-2, Insulation
coordination, part 2: application guide, 1997.
[8] A. R. Hileman, Insulation coordination for power
systems, Marcel Dekker, ISBN 0-8247-9957-7,
1999.
[9] EPRI, Transmission line surge arrester impulse
energy testing, EPRI, Palo Alto, vol. 1000461.
Figure 13: Arrester failure rate at each tower along a
section of the line.
0
0.0002
0.0004
0.0006
0.0008
0.001
0.0012
0.0014
0.0016
49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65
Tower Number
A
r
r
e
s
t
e
r

F
a
i
l
u
r
e
/
Y
e
a
r
ISBN 978-0-620-44584-9
Proceedings of the 16
th
International Symposium on High Voltage Engineering
Copyright c 2009 SAIEE, Innes House, Johannesburg
Pg. 6 Paper G-11