Vous êtes sur la page 1sur 5

Lightning Performance of 275 kV Transmission

Lines

R. Bhattarai
1
, R. Rashedin, S. Venkatesan, A. Haddad, H. Griffiths, N. Harid
High Voltage Energy Systems Group, School of Engineering, Cardiff University, United Kingdom
1
BhattaraiR@Cardiff.ac.uk


AbstractThis paper presents a comparative lightning
performance study conducted on a 275 kV double circuit
shielded transmission line using two software programs, TFlash
and Sigma-Slp. The line performance was investigated by using
both a single stroke and a statistical performance analysis and
considering cases of shielding failure and backflashover. A
sensitivity analysis was carried out to determine the relationship
between the flashover rate and the parameters influencing it. To
improve the lightning performance of the line, metal oxide surge
arresters were introduced using different phase and line
locations. Optimised arrester arrangements are proposed.
Index TermsLightning Overvoltages, Statistical Analysis,
Electrogeometric modelling, Sensitivity Analysis, Surge Arrester

I. INTRODUCTION
Lightning is a major cause of overhead line faults.
Between 5% to 10% of the lightning-caused faults are thought
to result in permanent damage to power system equipment
[1]. Therefore, the analysis of lightning performance is
fundamental when designing new lines and for uprating
existing lines to higher voltages.
Lightning is a natural phenomenon with random behaviour,
and hence a complete study of the lightning performance of
an overhead line should also include a statistical approach [2].
Computer software packages have been developed for the
evaluation of lightning performance of overhead lines. In this
study, two widely-available programs, viz. TFlash and Sigma-
Slp have been used in a comparative study of the lightning
performance of a generic 275 kV double circuit line.
Both software packages make use of the travelling wave
method for the computation of electromagnetic transients
along the line [3, 4]. TFlash employs a Stroke Incidence
Table (SIT) together with an Electrogeometric Model (EGM)
or the EPRI stroke attraction model while Sigma-Slp uses a
Monte Carlo statistical method in combination with the EGM
to determine strike points on the line. Both programs are
capable of modelling the application of line surge arresters.
In this paper, single and statistical stroke analyses were
made with different amplitudes of the injected stroke current
in order to estimate the flashover performance of the studied
transmission line. In this investigation, both shielding failure
and backflashover were considered. Detailed sensitivity
analysis studies were carried out to determine the relationship
between the flashover rate and the parameters influencing it;
such as tower footing resistance, ground flash density and
front time of the lightning impulse. The overvoltage
magnitude and impulse shape on the struck phase conductor
is computed for both the shielding failure and backflashover
cases.
To improve the lightning performance of the line, the
application of metal oxide surge arrester was studied using
different arrester configurations and locations. The computed
results indicate that arresters installed on the top phase of
each circuit give the most significant improvement in
lightning performance when combined with low tower
footing resistance. Optimised locations of surge arresters
were then derived for practical applications.

II. SIMULATED LINE DATA
A 35km long, 275kV double circuit line with 300m span
length was selected in this study. The height of the steel-
lattice towers of the line is assumed to be 36.88m. The surge
impedance of the tower was calculated to be 173.1 using (1)
[2, 5].

+
=

2 1
1
tan 5 . 0 cot ln 60
h h
r
Z
avg
T
(1)
where, r
avg
is the weighted average tower radius given by (2).

) (
) (
2 1
1 3 2 1 2 2 1
h h
h r h h r h r
r
avg
+
+ + +
= (2)
where,
r
1
, r
2
and r
3
are the radii at the top, midsection and base of the
tower respectively, and
h
1
and h
2
are the tower heights from base to midsection and
midsection to the tower top respectively.
The line is assumed to be located on flat terrain with a
ground flat density of 0.5 flashes per kilometre square per
year (fl/km
2
/yr). It is also assumed that there is no nearby
object present to cause an induced voltage flashover on the
line.
The phase conductors were twin 175mm
2
Lynx type
ACSR conductors with a bundle spacing of 30.48cm and a
single Lynx ACSR conductor was used for earthwire. The
individual conductors have a 19.53mm diameter. The phase
and earth conductors were assumed to have a 7.05m and
6.66m mid-span sag respectively. The Tower structure and
conductor geometry are shown in Fig. 1.

Authorized licensed use limited to: UNIVERSIDADE FEDERAL DE MINAS GERAIS. Downloaded on March 4, 2009 at 11:15 from IEEE Xplore. Restrictions apply.






























The line insulator strings are composed of 16 individual
glass insulator discs having 170mm spacing and each disc has
a creepage distance of 540mm which is equivalent to a total
string length of 3.31m (including the length of the fittings at
top and bottom of the string). The calculated critical
flashover voltage (CFO) of the insulator string is 1646kV.
To study the lightning performance of the line with surge
arresters, metal oxide arresters with a nominal discharge
current of 10kA, a Maximum Continuous Operating Voltage
(MCOV) of 220kV and an energy capability of 7.8kJ/kV were
used. The voltage-current characteristic of the arrester used
in this work is shown in Fig. 2.













III. LINE MODELLING TECHNIQUES
The details of the transmission line modelling in the
software are as follows:
The line consists of 116 spans with each span represented
as a multiphase untransposed distributed parameter line
section. To avoid reflections on the line, a sufficiently long
section is added to each side of the studied line. At line ends,
Sigma-Slp connects coupling matrices while TFlash adds
matching impedances in order to avoid reflections. Each
simulated span section is further divided into shorter sections
to enable stroke simulation at a number of points along the
span.
In TFlash, the tower is modelled as a simple transmission
line with constant surge impedance, and terminated with a
footing resistance. In Sigma-Slp, the tower is modelled by a
propagation element model represented by the tower surge
impedance and its propagation length. The propagation
length is equal to the height of the tower.
A non-linear footing resistance, as derived by Weck [2] and
given in (3), is used in both programs
g
T
I
I
R
R
+
=
1
0
(3)
With;
2
0
2 R
E
I
g
g

= (4)
where, R
0
is the low-current tower footing resistance, is the
soil resistivity, I is the stroke current through the tower
footing and, E
0
the soil ionisation threshold [400kV/m].
In TFlash, the Disruptive Effect (DE) method [6] is used as
the default insulator flashover model whereas Sigma-Slp, uses
a leader progression method. The DE method defines the
disruptive index by
[ ] dt A t V DE
B

= ) ( (5)
where, V(t) is instantaneous value of the impulse voltage, and
A and B are constants; A represents the minimum voltage
below which breakdown cannot occur and B is a coefficient
indicating that the breakdown process is not linear. When the
disruptive index, DE, reaches a critical value, breakdown
would occur.
The leader progression method [2] is represented by

d
t u
l
l
e E
l d
t u
d V
) (
0015 . 0
0
) (
170

= (6)
where, V
l
is the leader velocity, d the gap distance, l
l
the
leader length, u(t) the applied voltage and, E
0
the voltage
gradient (520kV/m).
The power frequency voltage may influence the insulator
flashover. This influence is taken into account by calculating
the voltage across the insulators through a 360
0
phase cycle,
and the flashover rate is determined using its average.
The effect of corona coupling in the line is considered in
TFlash while it is ignored in Sigma-Slp.

Fig. 1. 275 kV double circuit shielded transmission line tower showing
its conductor coordinates. Values in parenthesis are midspan heights.
500
550
600
650
700
750
800
0 10 20 30 40
Arrester Cur rent (kA)
D
i
s
c
h
a
r
g
e

V
o
l
t
a
g
e

(
k
V
)
Fig. 2. Surge arrester voltage-curve.
Authorized licensed use limited to: UNIVERSIDADE FEDERAL DE MINAS GERAIS. Downloaded on March 4, 2009 at 11:15 from IEEE Xplore. Restrictions apply.
-2000
-1000
0
1000
2000
3000
4000
5000
6000
0 2 4 6 8 10
Time (s)
O
v
e
r
v
o
l
t
a
g
e


(
k
V
)
TFlash
Sigma-Slp

(a) Shielding failure flashover
-1000
0
1000
2000
3000
4000
5000
6000
0 5 10 15 20 25 30 35 40 45 50
Time (s)
O
v
e
r
v
o
l
t
a
g
e

(
k
V
)
TFlash
Sigma-Slp

(b) Backflashover

Fig. 3. Overvoltage magnitude and shape on phase conductor A1
computed using TFlash & Sigma-Slp.
The EGM is used to determine the strike point on the line.
In this investigation, more than 20000 stokes are used for the
electromagnetic simulations. The conductor and earth
striking distance used in the software are given by [2, 7]
65 . 0
10 I R
c
= (7)
( ) [ ]
65 . 0
43 ln 7 . 1 6 . 3 I h R
e
+ = for h < 40 m (8)
65 . 0
5 . 5 I R
e
= for h > 40 m (9)
where, R
c
is the striking distance to a line conductor, R
e
is the
striking distance to earth, I is the lightning impulse current
magnitude and, h is the height of the tower.

IV. SINGLE STROKE ANALYSIS
For the single stroke studies, the overvoltage magnitude
and impulse shape on the phase conductor during shielding
failure and backflashover are computed. A 32kA, 2/75
impulse current is applied to phase conductor A
1
at the tower
position to simulate a shielding failure flashover on the line.
A low current tower footing resistance of 10 and 200 m
soil resistivity were assumed. Fig. 3a shows typical impulse
voltage on the phase conductor. As can be seen in the figure,
similar voltage impulse magnitudes were obtained with the
two models. However, close examination of the results
reveals that the TFlash model predicts a slightly higher
overvoltage magnitude and a faster initial rise time compared
with the Sigma-Slp model. For the backflashover studies, a
200kA lightning impulse was injected on to the shield wire at
a tower position. A low current tower footing resistance of
80 and a soil resistivity of 1600m were adopted, which
allow backflashover to be initiated. Fig. 3b shows a typical
impulse shape of overvoltage computed on the phase
conductor during backflashover. Here again, we can observe
that the results obtained with TFlash model indicate slightly
higher magnitudes of overvoltage and faster rise times
compared with the Sigma-Slp model.

V. STATISTICAL ANALYSIS
Statistical flashover analysis of the studied line is carried
out to assess the risk of flashover considering the random
behaviour of lightning. The flashover rate (number of flashes
per 100km per year) is used as the basic parameter in the
sensitivity analysis, and the performance of the line is
analysed with different arrester configurations.
The programs used in this investigation (Sigma-Slp and
TFlash) use slightly different approaches for simulating
random behaviour of lightning. In Sigma-Slp, random
lightning strokes are generated with magnitudes between 1.2
and 161.1kA and with rise times in the range between 1.2 and
4.38 s. A fixed half time of 75 s is assumed. The impulse
shape varies randomly, with a sample size of 2000. In
TFlash, however, the stroke current range can be selected
from 1kA to 300kA, and the range is divided up to 512
current bins. In order to match the two models, as closely as
possible, 32 stroke current beans and a current range from
2.5kA to 160kA were selected.
In the TFlash model, the selected lightning current impulse
shape was used for the entire statistical calculation without
any variation. Again, for close matching of the two models,
simulations were carried out with a 2/75 lightning impulse.
A. Sensitivity Analysis
Sensitivity analysis studies were carried out to determine
the relationship between the flashover rate and the parameters
influencing it, such as tower footing resistance, ground flash
density (GFD) and front time of the lightning impulse.
(i) Effect of footing resistance
It was found that the shielding failure flashover rate
(SFFR) is not affected by footing resistance and a constant
SFFR of 0.7 and 0.63 fl/100km/yr was obtained in TFlash and
Sigma-Slp respectively. Fig. 4a illustrates the effect of
footing resistance on the backflashover rate. As can be seen
in the figure, the backflashover rate (BFR) evaluated by
Sigma-Slp, for a given value of footing resistance is higher
than that computed by TFlash.
(ii) Effect of ground flash density (GFD)
The effect of GFD on SFFR is shown in Fig. 4b. As can be
seen, the SFFR calculated using the Sigma-Slp model is
lower than the SFFR computed using the TFlash model.
The difference may be attributed to the different stroke
statistics patterns of the two models. In the TFlash model,
45% to 50% of the current bins are in the low-current range
Authorized licensed use limited to: UNIVERSIDADE FEDERAL DE MINAS GERAIS. Downloaded on March 4, 2009 at 11:15 from IEEE Xplore. Restrictions apply.
which is more likely to result in shielding failures. On the
other hand, in the Sigma-Slp model, only 10% to 15% of the
total number of strokes are likely to hit the phase conductor as
a result of shielding failure. Therefore, the majority of
strokes in the Sigma-Slp simulations hit the shield wire or the
tower top resulting in higher backflashover rates.
(iii) Effect of impulse shape
The computed effect of lightning impulse shape on the line
flashover rate is shown in Fig. 4c. Only results from the
TFlash model were shown since it is possible to vary the
impulse front time in this case. From Fig. 4c, it can be seen
that, for fast front times (< 2s), the BFR increases as the
front time decreases. The simulations show that SFFR is not
influenced by variations in the front time; a constant value of
0.7 fl/100km/yr is obtained.










































B. Flashover Performance of the Line with Surge Arresters
The objective of this study was to estimate the
improvement in flashover rate by implementing surge
arresters on the line. Different arrester configurations and
locations were analysed and compared to assess
improvements in lightning performance of the line. In these
studies, low current tower footing resistance value was varied
from 10 to 80 keeping the ratio of soil resistivity to
footing resistance constant (/R
0
= 20). An initial study was
carried out with arresters positioned at every tower and on
every phase conductor. This resulted in a zero flashover rate,
but, practically the configuration would be too expensive.
It is well known [8] that, in the low current range, the
lightning strikes hit only the top phase conductors of the two
circuits during shielding failure. This suggests that the
majority of shielding failures for this tower design would
occur on the top phases. Table I shows typical results of an
electrogeometric model (EGM) study obtained with Sigma-
Slp and Fig. 5 shows an example of graphical display of
lightning strokes hitting the line for different current
magnitudes, which were obtained with TFlash. Based on
these studies, four practical different arrester configurations
were studied, and are shown in Table II. In the table, the
black spots on the phases represent the application of an
arrester on that phase.
As can be seen in the table, the Sigma-Slp modelling shows
that the application of arresters substantially improves the
flashover rate. By placing arresters on the top phases only, a
zero SFFR is obtained. However, this arrangement only
suppresses backflashover under low footing resistance
conditions. On the other hand, with arresters installed on the
bottom phases only, a zero backflashover rate is obtained at
expense of shielding failure. When arresters are installed on
the top and bottom phases, both shielding and backflashover
failure can be suppressed. Therefore, to improve the
lightning performance, it is recommended to install arresters
only in the top phases at towers with low footing resistance
and in the top and bottom phases at towers with high footing
resistance.
TABLE I
EGM REPORT IN SIGMA-SLP
G
F
D

M
e
d
.

s
t
r
o
k
e

c
u
r
r
e
n
t

(
k
A
)

T
o
t
a
l

s
t
r
o
k
e

T
o

e
a
r
t
h

T
o

t
o
w
e
r

T
o

s
h
i
e
l
d

w
i
r
e

T
o

p
h
a
s
e

A
1

T
o

p
h
a
s
e

B
1

T
o

p
h
a
s
e

C
1

T
o

p
h
a
s
e

C
2

T
o

p
h
a
s
e

B
2

T
o

P
h
a
s
e

A
2

0
.
5

3
1
.
5

2
0
0
0
0

1
4
0
5
4

3
3
1

4
8
4
0

2
8
0

0

0

2
9
8

0

0






-0. 05
0
0. 05
0. 1
0. 15
0. 2
0. 25
0. 3
0. 35
0. 4
0 10 20 30 40 50 60 70 80 90
Low Curre nt Footi ng Re si sta nce (ohm)
B
F
R

(
f
l
/
1
0
0
k
m
/
y
r
)
TFl as h
Sigma-Slp

(a) Flashover rate vs. low current footing resistance
(GFD = 0.5 fl/km
2
/yr, /R0 = 20)
0
0. 2
0. 4
0. 6
0. 8
1
1. 2
1. 4
1. 6
0 0. 2 0. 4 0. 6 0. 8 1 1. 2
GFD (fl / km
2
/ yr)
S
F
F
R

(
f
l
/
1
0
0
k
m
/
y
r
)
TFlash
Si gma-Slp

(b) Flashover rate vs. GFD (R0 = 10 , = 200 m)
-0. 1
0
0. 1
0. 2
0. 3
0. 4
0. 5
0. 6
0. 7
0. 5 1 1. 5 2 2. 5 3 3. 5 4
Front Ti m e (s)
B
F
R

(
f
l
/
1
0
0
k
m
/
y
r
)

(c) Flashover rate vs. lightning impulse front time in TFlash
(GFD = 0.5 fl/km
2
/yr, R0 = 10 , = 200 m)

Fig. 4. Sensitivity analysis

Fig. 5. Stroke view at different current magnitude in TFlash
Authorized licensed use limited to: UNIVERSIDADE FEDERAL DE MINAS GERAIS. Downloaded on March 4, 2009 at 11:15 from IEEE Xplore. Restrictions apply.
Table III shows the results obtained with the TFlash model,
which agree with the previous findings. The SFFR obtained
in this case is slightly higher while the BFR lower.

TABLE II
FLASHOVER RATE (SIGMA-SLP MODEL)
DIFFERENT ARRESTER CONFIGURATIONS
Footing
Resistan
ce
()

Shielding Failure Flashover Rate (fl/100km/yr)
10 0.63 0 0.63 0.63 0
20 0.63 0 0.63 0.63 0
30 0.63 0 0.63 0.63 0
40 0.63 0 0.63 0.63 0
50 0.63 0 0.63 0.63 0
60 0.63 0 0.63 0.63 0
70 0.63 0 0.63 0.63 0
80 0.63 0 0.63 0.63 0
Backflashover Rate (fl/100km/yr)
10 0 0 0 0 0
20 0 0 0 0 0
30 0 0 0 0 0
40 0.03 0 0 0 0
50 0.08 0.01 0 0 0
60 0.14 0.03 0.01 0 0
70 0.26 0.07 0.02 0 0
80 0.35 0.11 0.03 0 0
Total Flashover Rate (fl/100km/yr)
10 0.63 0 0.63 0.63 0
20 0.63 0 0.63 0.63 0
30 0.63 0 0.63 0.63 0
40 0.67 0 0.63 0.63 0
50 0.72 0.01 0.63 0.63 0
60 0.78 0.03 0.65 0.63 0
70 0.90 0.07 0.66 0.63 0
80 0.99 0.11 0.67 0.63 0
mark indicates surge arrester in the phase

TABLE III
FLASHOVER RATE (TFLASH MODEL)
DIFFERENT ARRESTER CONFIGURATIONS
Footing
Resistan
ce
()

Shielding Failure Flashover Rate (fl/100km/yr)
10 0.70 0 0.70 0.70 0
20 0.70 0 0.70 0.70 0
30 0.70 0 0.70 0.70 0
40 0.70 0 0.70 0.70 0
50 0.70 0 0.70 0.70 0
60 0.70 0 0.70 0.70 0
70 0.70 0 0.70 0.70 0
80 0.70 0 0.70 0.70 0
Backflashover Rate (fl/100km/yr)
10 0 0 0 0 0
20 0 0 0 0 0
30 0 0 0 0 0
40 0.01 0 0 0 0
50 0.04 0 0 0 0
60 0.09 0.02 0 0 0
70 0.15 0.03 0 0 0
80 0.25 0.05 0.02 0 0
Total Flashover Rate (fl/100km/yr)
10 0.70 0 0.70 0.70 0
20 0.70 0 0.70 0.70 0
30 0.70 0 0.70 0.70 0
40 0.71 0 0.70 0.70 0
50 0.74 0 0.70 0.70 0
60 0.79 0.02 0.70 0.70 0
70 0.85 0.03 0.70 0.70 0
80 0.95 0.05 0.72 0.70 0
mark indicates surge arrester in the phase


VI. CONCLUSION
Two similar models of transmission line were studied using
Sigma-Slp and TFlash software packages. Satisfactory
agreement is given by the two models.
It was shown that the lightning performance of overhead
transmission lines can be improved by applying surge
arresters. Arresters on the top phases improve shielding
failure flashover rate, and when applied to the bottom phases
they allow improvement of backflashover rate. Adequate
selection of the arrester configuration in the line can
significantly improve lightning performance and may reduce
the financial burden.


REFERENCES
[1] T. Short, "Application of the IEEE Guide for Improving the Lightning
Performance of Electric Power Overhead Distribution Lines (Std. 1410-
1997)," in IEEE T&D Conference, 1999.
[2] CIGRE Working Group 33-01, "Guide to Procedures for Estimating the
Lightning Performance of Transmission Lines," CIGRE Brochure 63,
1991.
[3] T. E. McDermott and V. J. Longo, "Advanced computational methods
in lightning performance the EPRI Lightning Protection Design
Workstation," Power Engineering Society Winter Meeting, 2000. IEEE,
2000.
[4] Y. A. Wahab, Z. Z. Abidin, and S. Sadovic, "Line surge arrester
application on the quadruple circuit transmission line," Power Tech
Conference Proceedings, 2003 IEEE Bologna, 2003.
[5] W. A. Chisholm, Y. L. Chow, and K. D. Srivastava, "Travel Time of
Transmission Towers," IEEE Transactions on Power Apparatus and
Systems, vol. PAS-104, pp. 2922-2928, 1985.
[6] U. Savadamuthu, K. Udayakumar, and V. Jayashankar, "Modified
disruptive effect method as a measure of insulation strength for non-
standard lightning waveforms," Power Delivery, IEEE Transactions on,
vol. 17, pp. 510-515, 2002.
[7] "IEEE guide for improving the lightning performance of transmission
lines," IEEE Std 1243-1997, 1997.
[8] S. Sadovic, R. Joulie, S. Tartier, and E. Brocard, "Use of line surge
arresters for the improvement of the lightning performance of 63 kV
and 90 kV shielded and unshielded transmission lines," Power
Delivery, IEEE Transactions on, vol. 12, pp. 1232-1240, 1997.

















Authorized licensed use limited to: UNIVERSIDADE FEDERAL DE MINAS GERAIS. Downloaded on March 4, 2009 at 11:15 from IEEE Xplore. Restrictions apply.