TELKOM
NIKA Indonesia
n
Journal of
Electrical En
gineering
Vol. 12, No. 8, August 201
4, pp. 5814 ~ 5826
DOI: 10.115
9
1
/telkomni
ka.
v
12i8.603
3
5814
Re
cei
v
ed Ma
rch 3
0
, 2014;
Re
vised Ma
y 10, 2014; Accepted Ma
y 26
, 2014
Reliability Analysis of Surge Arrester Location Effect in
High Voltage Substations
Se
y
e
d Ahmad Hoss
eini*, Mohammad
Mirzaie, Tag
h
i Barfo
r
osh
i
Dep
a
rtment of Electrical E
ngi
neer
ing, Ba
bol
Univers
i
t
y
of
T
e
chn
o
lo
g
y
,
Ba
bol,
Iran
*Corres
p
o
n
id
n
g
author, e-ma
i
l
: ahmad
hoss
e
i
n
i1
11@
gmai
l.com
A
b
st
r
a
ct
High v
o
ltag
e s
ubstatio
n
s as
pivota
l sectio
n
s
of
pow
er netw
o
rk play an i
m
p
o
rtant rol
e
i
n
pow
er
netw
o
rk becau
se of supplyi
ng
electrical e
ner
gy for cons
u
m
ers. In recent years w
i
th incre
m
e
n
t of electri
c
ity
consumption and complexity
of power system
, the substa
tion stability and protection
against various faults
or overv
o
lta
g
e
s
has
bee
n a
big c
onc
ern. S
u
rge
arrest
ers
as pr
otective
devic
e h
a
s the
respo
n
si
bility
of
attenuati
ng
ov
ervolta
ge d
ue
to light
nin
g
or
sw
itch
ing occ
u
rring
in trans
miss
ion
or dis
t
ributio
n syste
m
.
Absenc
e or fail
ure of arrester
s inside or n
e
a
r
substa
tions c
an le
ad to loa
d
interrupti
on an
d great outa
g
e
in
the syste
m
. Therefor
e their
prese
n
ce
is i
ndis
pens
ab
le. De
mo
nstratin
g
the surg
e ar
resters pres
en
ce
necess
i
ty an
d
eval
uatin
g th
eir
failur
e
i
m
pact
duri
ng
nor
ma
l
oper
ating
con
d
i
tion
an
d ov
erv
o
ltag
e i
n
differ
ent
scenar
ios of ar
rester locati
on
and l
i
g
h
tnin
g s
u
rge p
o
siti
on i
n
view
poi
nt of EMT
P
simul
a
tion a
nd re
li
abi
li
ty
techni
qu
e is the purp
o
se of th
is pap
er.
Ke
y
w
ords
: sur
ge arrester, E
M
TP-RV, high
voltag
e subs
tat
i
on, rel
i
ab
ility i
ndic
e
s, mi
ni
ma
l cut set
Copy
right
©
2014 In
stitu
t
e o
f
Ad
van
ced
En
g
i
n
eerin
g and
Scien
ce. All
rig
h
t
s reser
ve
d
.
1. Introduc
tion
High voltag
e
sub
s
tation
s a
s
a
key pa
rt o
f
powe
r
sy
ste
m
s conn
ect g
eneration
sou
r
ce
s to
distrib
u
tion n
e
tworks. Fast
growth of el
ectri
c
al en
erg
y
demand wit
h
the develo
p
ment of po
wer
system
s makes these substation
s m
o
re important. Therefore,
substation reliability
is an
importa
nt co
nce
r
n fro
m
b
o
th utility and cu
stome
r
points of vie
w
[1]. Subst
a
tions a
r
e al
ways
subj
ect to va
riou
s ove
r
vol
t
ages such a
s
swit
ching
a
nd lightni
ng t
hat may
cau
s
e
su
bstatio
n
s
failure. The substatio
n
failure can lead
to t
he outage of so man
y
loads in the powe
r
syst
em,
whi
c
h can be
catast
rop
h
ic
to the system
[2].
The most dange
rou
s
o
v
ervoltage i
s
due to lightni
ng
surge
whi
c
h
its frequ
en
cy rang
e is f
r
o
m
10
kHz
to
3MHz. If lightning
strikes the line
s
or
the
towers adja
c
ent to a
sub
s
tation, cau
s
e
s
ba
ck-f
lashov
er,
the
overv
o
ltage wave
prop
agate
s
at
the
sub
s
tation
an
d can
dama
g
e
sub
s
tation
equipm
ent
su
ch
as tra
n
sfo
r
mer d
r
amati
c
ally. Betwe
e
n
5% to 10%
o
f
the lightnin
g
-
ca
used fa
ults a
r
e th
oug
ht to re
sult i
n
p
e
rma
nent
da
mage to
po
wer
system e
quip
m
ent [3].
Surge
arre
st
ers a
r
e b
e
in
g appli
ed to
system
to mitigate
the overvoltage. As
novel
prote
c
tion
de
vices ag
ainst
overvoltag
es with
out
sta
n
d
ing
ope
ratio
nal p
e
rfo
r
ma
nce, m
e
tal ox
ide
surge arrest
e
r
s (MOA
)
were
develo
ped 20
yea
r
s
ago
and a
r
e
no
w being
extensively applied i
n
power
syste
m
s to lower t
he insulation
level and e
n
sure lo
ng time
reliability of the high volta
ge
equipm
ent [4]. Their nonlin
ear resi
stan
ce cha
r
a
c
teri
st
ic prote
c
ts e
q
u
ipment ag
ai
nst overvolta
ges
whi
c
h
may e
x
ceed
ba
si
c l
i
ghtning
impu
lse i
n
sula
tion
level (BIL). Arre
sters ca
n
be
l
o
cated on
sele
cted
poin
t
s of the
net
work to
obtai
n the
re
q
u
ire
d
control
of
overvoltage
s
[5], particularly
insid
e
and n
e
a
r the sub
s
ta
tions. Thu
s
, surge
arre
st
ers incre
a
se the relia
bility of the sub
s
tatio
n
s
durin
g the occurrin
g of overvoltage
s
esp
e
cially du
e to lightning
surg
e.
Ideal surg
e a
rre
sters
sho
u
l
d be o
pen
ci
rcuit
duri
ng n
o
rmal
ope
rati
ng conditio
n
of the
system
an
d
sho
r
t
circuit
durin
g th
e ov
ervoltage
s. B
u
t the i
n
sulat
i
on b
ehavio
r
of MOA
may be
subj
ect to ch
ange
s due to
different deg
radatio
n me
chani
sms that
can be
categ
o
rized a
s
imp
a
ct
from co
ntinu
ous o
p
e
r
atin
g voltage stress, humidif
i
c
ation
cau
s
e
d
by deficien
c
y in stru
ctu
r
e and
poor
seali
ng,
impulse
current stre
ss of
insid
e
par
tial
discha
rge, effects of atmo
sph
e
ri
c su
rfa
c
e
contami
natio
n, etc [4]. Th
ese fa
ctors
can in
cre
a
se
the lea
k
a
ge current of
arre
sters an
d in
cl
ine
them to be sh
ort circuit esp
e
cially in the end of
their life. Short circui
t
behavior of surge arre
ste
r
s
in norm
a
l condition of syst
em reduces t
he reliabilit
y of substations. Also
open circuit behavior of
surge a
r
reste
r
s
duri
ng ov
ervoltage
re
d
u
ce
s the
reli
ability greatly
. In gene
ral,
mode
rn
su
rge
Evaluation Warning : The document was created with Spire.PDF for Python.
TELKOM
NIKA
ISSN:
2302-4
046
Reliability Analysis of Su
rge Arrester Location Effect i
n
High
Voltage… (Seyed A
h
m
a
d Hossei
ni)
5815
arreste
r
s utili
zing
ZnO
blo
c
ks a
r
e ve
ry
reliabl
e ap
pa
ratus with
a l
o
w failu
re
rat
e
. A publi
s
he
d
numbe
r for di
stributio
n arresters i
s
0.1
%
failures
p
e
r
year, whil
e for high volta
ge arre
sters the
estimated
fail
ure
rate
is ev
en lo
we
r [6].
It can
be
con
c
lud
ed th
at f
a
ilure
of
arre
sters
ca
n aff
e
ct
the substation reliability and
can
cause severe harm
to
substation
reli
ability or
great
outage
in
the netwo
rk.
In addition, location
of su
rge arreste
r
s a
ffects the e
quipme
n
t terminal voltage
and
substation rel
i
ability. They
are often i
n
st
alled i
n
the place where the inco
mi
ng li
ne
connects to
the sub
s
tatio
n
and on tra
n
s
form
er feed
ers.
For the investigation of su
rge arresters infl
uence on the HV substa
tions reliability, firs
t
it will be seen from the sim
u
lation analysis point
of view by Electro Magnetic T
r
ansi
e
nt Progra
m
(EMTP-RV).
Then it will b
e
analyze
d
by
mathemati
c
al
method
s giv
en for
cal
c
ula
t
ing of reliabili
ty.
Deh
a
k
su
bst
a
tion which i
s
a
230/6
3
kV sub
s
tation
in Ne
ka, Iran
, is sele
cted
as a
case st
udy.
Asse
ssme
nts prove,
ab
se
nce
or fa
ilure
of arre
ster can imp
o
se a
seri
ou
s p
r
obl
em to
sub
s
tat
i
on
and lea
d
to long time load
unavailability.
2. Sy
stem Modeling
Deh
a
k, a 23
0
/
63 kV and o
ne bre
a
ker a
nd half su
bst
a
tion in Ne
ka,
Iran, is co
nsi
dere
d
as
ca
se
study. It has eight fe
eders
co
nsi
s
t
s
of
six
line f
e
ede
r
s and
two tran
sform
e
r fee
ders. T
he
singl
e line dia
g
ram of this
substatio
n
is shown in Figu
re 1.
Figure 1. Deh
a
k Sub
s
tation
Single Line
Diag
ram
In ord
e
r to
obtain a
c
tual
re
sults f
r
o
m
simul
a
tion
of the sub
s
tation, pa
rti
c
ula
r
ly in
transi
ent co
n
d
ition, prop
er models of substati
o
n
eq
uipment an
d other comp
o
nents
sho
u
ld
be
applie
d
to si
mulation.
In this se
ction variou
s
el
em
ents of
the con
s
id
ere
d
system
with
their
model
s are in
trodu
ced.
Owin
g to i
n
h
e
rent
di
stribu
ted natu
r
e
of
tran
smi
ssi
on
line
s
, fre
q
ue
ncy d
epe
nde
n
t (F
D)
line is u
s
e
d
to model t
r
an
smissio
n
line
s
conn
ec
te
d
to the su
bsta
tion. Electri
c
a
l
para
m
eters
of
these
line
s
t
hat mu
st be
inse
rted i
n
to
the line
dat
a
of FD line
s
are
sh
own in
Table
1. In t
h
is
table, R
0
, L
0
and C
0
are
zero
seque
nce pa
ram
e
ters and
R
1
, L
1
a
nd
C
1
a
r
e
po
sitive sequ
en
ce
para
m
eters.
Table 1. Elect
r
ical Pa
ram
e
ters of Lin
e
s
Data
Neka
Sari
Kordku
y
Kaghazsazi
R
0
(
Ω
/km)
0.3184
0.0892
0.4467
0.3255
L
0
(mH/km)
3.8276
2.3679
2.7315
3.2446
C
0
(
μ
F/km
0.00473
0.00827
0.005348
0.00607
R
1
(
Ω
/km) 0.0612
0.032
0.1058
0.0529
L
1
(mH/km
)
1.294
0.9261
2.0094
0.9844
C
1
(
μ
F/km
)
0.00888
0.0127
0.008775
0.0115
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN: 23
02-4
046
TELKOM
NI
KA
Vol. 12, No. 8, August 2014: 581
4 –
5826
5816
The choice o
f
tower mod
e
l
in analyzin
g
light
ning surge in the po
wer
system i
s
very
importa
nt. The steel to
wers are usually rep
r
e
s
ent
e
d
as
a single condu
ctor distributed
pa
ram
e
te
r
line termi
nat
ed by a resistan
ce rep
r
ese
n
ting the
tower fo
oting impe
dan
ce [7]. Con
s
tant
para
m
eter (CP) line is use
d
to simulate
single cond
u
c
tor of towe
r model. 160
Ω
and sp
eed
o
f
light are dete
r
mine
d for towers surg
e impeda
nce an
d wave veloci
ty as the principal pa
ram
e
ters
of CP line.
Diele
c
tri
c
strength of
line
insul
a
tors
u
n
der
lightnin
g
con
d
ition
s
d
e
pend
on
the
impulse
wave
sha
pe,
magnitud
e
a
n
d
pol
arity. A li
ghtning
-im
pul
se volta
ge,
wi
th a ma
gnitud
e
that ex
cee
d
s
the critical
flash
o
ver
(CF
O
), may
still
not la
st
long
enou
gh to
carry
stre
ame
r
s all th
e
way
to
compl
e
te insulation brea
kdown [8]. In
orde
r to mod
e
l insulato
r string, flashov
er switch a
s
an
element in E
M
TP-RV i
s
u
s
ed.
Lightnin
g
surge
waveform
ha
s
con
s
id
e
r
able
influe
nce on
the
ove
r
voltage
indu
cing
on
sub
s
tation e
q
u
ipment. Th
e
lightning
surge is m
odele
d
by a cu
rre
n
t and a p
a
rallel re
si
stan
ce
(Cig
re
model
). The resi
stan
ce valu
e is ta
ken to
be
40
0
Ω
, whi
c
h
wa
s de
rived
by Bewley [9]. T
h
e
stand
ard i
m
p
u
lse
wavefo
rm (1.2/50
)
with the pea
k
value of 150
kA is u
s
e
d
for mo
deling
of
lightning surg
e.
The a
pprop
ri
ate mod
e
ling
of a
surge
arreste
r
i
s
si
gnifica
nt for i
n
sul
a
tion
coo
r
dinatio
n
studie
s
in order to extract reliable e
s
timations. Fo
r this rea
s
on,
several fre
q
uent-d
epe
nd
ent
model
s of
Z
n
O
su
rge
arreste
r
s, u
s
in
g the
phy
si
cal an
d the
electri
c
al
dat
a given
by t
he
manufa
c
turer, have bee
n p
r
opo
se
d, in a
way that
the model simul
a
tion
re
sults correspon
d
to the
actual b
ehavi
o
r of the arre
ster [
10]. In the relate
d si
mulation, Zn
O su
rge a
r
re
ster i
s
rep
r
e
s
ented
by IEEE recommended
model
whi
c
h is
dem
onstrated in Figure
2. T
h
is model
contai
ns
non-
linear
re
si
sta
n
ce
s in t
w
o l
o
catio
n
s
whi
c
h are sepa
ra
t
ed by an
RL
filter. This
RL
filter provid
e
s
impeda
nce th
at differs b
e
twee
n the cases of fast
surges a
nd
slo
w
surge
s
[11]. Table 2
sho
w
s
IEEE model parameters and other
char
acteri
stics
of arresters used
in Dehak
substation. In this
table, V
r
an
d
I
d
are
arre
ste
r
rated volta
g
e
an
d di
scha
r
ge cu
rrent resp
ectively. Beside
s of surge
arreste
r
mod
e
l, the v-i cha
r
acte
ri
stic of surge a
rre
ste
r
s al
so spe
c
ifies the beh
avior of them. Their
cha
r
a
c
teri
stics sh
ould b
e
n
online
a
r such
that su
rge a
r
reste
r
current
rise
s as ove
r
voltage occu
rs
and ab
ate to very small val
ue wh
en overvoltage de
ca
ys. The arrester lead len
g
th at the top and
at the
bottom
mu
st be
con
s
ide
r
ed
to
accou
n
t for the
effects of a
d
d
i
tional voltag
e ri
se
a
c
ross
the
lead in
du
ctan
ce. A lum
ped
element
rep
r
ese
n
tation
with an i
ndu
ctan
ce
of 1
μ
H/m
will be suffici
ent
[7].
Figure 2.
IEEE Model of Surge Arre
sters
Table 2. IEEE Model Parameters and
Characteristi
c
s of the Arrest
ers
Parameter Value
R
1
(
Ω
)
143.325
L
1
(
μ
H)
33.075
C (nF
)
0.04535
R
0
(
Ω
)
220.5
L
0
(
μ
H)
0.441
V
r
(kV)
198
I
d
(kA)
10
Und
e
r lig
htning ove
r
vo
ltage, su
bst
a
tion eq
uip
m
ent su
ch
as
circuit
brea
ke
r,
discon
ne
ctor,
powe
r
tra
n
s
form
ers an
d instru
ment
transfo
rme
r
s can be
a
pproxim
ated
with
cap
a
cito
rs. E
quivalent cap
a
citan
c
e
s
of the su
bstatio
n
equipme
n
t are sho
w
n in T
able 3.
Evaluation Warning : The document was created with Spire.PDF for Python.
TELKOM
NIKA
ISSN:
2302-4
046
Reliability Analysis of Surge A
rrester Location Effect i
n
High
Voltage… (Seyed A
h
m
ad Hossei
ni)
5817
Table 3. Equi
valent Cap
a
ci
tance of Sub
s
tation Equipm
ent
Equipment
Equivalent
capacitance (pF
)
Power tr
ansform
er
2000
Curre
nt transfo
r
m
er (C
T)
200
Capacitive voltag
e transforme
r
(
C
VT)
5000
Disconnector 150
Circuit breaker
500
Busba
r
and
i
n
terconn
ectio
n
s
between
the
sub
s
tation
equi
pment
should
be
mo
deled
by
distrib
u
ted lin
es or lu
mpe
d
inducta
nce. A model
of lumped pa
ram
e
ter indu
ctan
ce, its unit value
being 1
μ
H/m
,
is u
s
ually t
a
ke
n for
se
ctions, the l
e
ngths
of whi
c
h d
o
n
o
t e
x
ceed
15m [
12],
otherwise di
stributed lin
e
is u
s
ed. Deh
a
k
sub
s
tatio
n
ha
s a lon
g
bu
sba
r
mo
re than
100
m,
therefo
r
e CP
model line
sh
ould be u
s
e
d
to represent that in simulati
on.
3. Simulation Resul
t
s
By means
of data an
d el
ements
mod
e
ls give
n in
previou
s
se
ction, De
ha
k substatio
n
wa
s simul
a
te
d by EMTP-RV software wi
th the applic
a
t
ion of lightning su
rge to
many points
su
ch
as in
side the
sub
s
tation, subs
tatio
n
entrance and lin
e
phase.
Simulation shows ove
r
vo
ltage wave
shape a
nd its amplitude
b
a
si
cally dep
e
nds
on
locatio
n
of lig
htning
strike
and po
sition
of sur
ge a
r
re
sters. Since tran
sform
e
rs
on tran
sfo
r
m
e
r
feeders
and
cap
a
cito
r volt
age tran
sformers (CVT
)
on the
entra
nce,
are
mo
re impo
rtant t
han
other
equi
pm
ent, their volt
age
s d
u
rin
g
applying li
ght
ning
su
rge
h
a
ve be
en m
e
asu
r
ed
an
d t
heir
values a
r
e
sh
own in
Tabl
e
4 for differe
n
t
cases
of lightning
stro
ke
locatio
n
and
surge a
r
rest
ers
positio
n. Sim
u
lation
sh
ows if li
ghtning
stri
kes th
e
p
o
int which i
s
farthe
r th
an
250
m
from
the
sub
s
tation, o
v
ervoltage
wave attenuat
es d
r
a
s
tically and it
s am
pl
itude de
crea
ses g
r
e
a
tly un
der
the tran
sfo
r
m
e
r BIL
(85
0
kV) an
d the
CVT BIL
(10
50kV
)
a
nd th
erefo
r
e
ca
n
not ha
rm the
m
.
Acco
rdi
ng to this table, the
pl
aces of surge arre
sters are:
a)
Absen
c
e of a
rre
ster
(A)
b)
Presen
ce of arreste
r
in su
bstation e
n
tra
n
ce (E
)
c)
Presen
ce of arreste
r
in tra
n
sformer fee
der (T)
d)
Presen
ce of arreste
r
in su
bstati
on e
n
tra
n
ce a
nd tran
sformer fe
ede
r (E+T)
Table 4. Maxi
mum Indu
ced
Voltage due
to Lightning
Ligh
tni
ng s
t
rok
e
locati
o
n
Maximu
m in
du
c
e
d
v
o
l
t
age
(kV)
CVT
Transform
er
A E
T
E+T
A E
T
E+T
Inside the substation
2823.20
854.59
2640.10
818.11
4331.1
3454.40
999.83
814.22
Substation entra
nce
3071.5
1040.60
3071
1038.30
4101.3
1273.10
921.16
575.65
50 m from subst
a
tion
1028.1
522.50
735.65
518.91
1179.20
808.96
506.03
454.28
100 m from sub
s
tation
890.25
492.62
742.49
483.44
1216.6
819.09
498.59
443.94
150 m from subs
tation
809.62
481.17
775.66
484.41
1283
836.51
492.72
444.14
200 m from subs
tation
711.20
466.75
695.40
465.18
1145
845.90
499.47
457.55
250 m from subs
tation
596.78
462.55
596.14
462.55
936.61
800.44
506.02
480.60
300
m
from
subs
tation
569.61
442.43
569.32
442.46
824.05
782.70
532.66
528.73
Shaded valu
e
s
indi
cate to the ca
se
s in which ove
r
voltage exceed
s
the BIL.
4. Reliabilit
y
Assessment Method
In previou
s
section
s
, role of surg
e a
rre
ste
r
s
and their lo
cation
s impa
ct were
investigate
d
.
No
w, reliabilit
y techni
que
s are u
s
e
d
to
analy
z
e
su
rge a
r
reste
r
s
effects f
r
om
the
viewpoi
nt of reliability. The equatio
ns
whi
c
h are nee
de
d for cal
c
ulati
ng relia
bility are given in [1
3]
in terms of failure rate (
λ
), repair time (r) and u
navai
lability
(U) for serie
s
and p
a
rallel
syste
m
s
that may cont
ain one, two
or more co
m
pone
nts.
For sub
s
tations
reliabilit
y evaluation, several techniqu
es h
a
ve been
sug
geste
d in
different re
se
arche
s
su
ch
as Markov modelin
g [14
], minimal cut set [15] and Monte Ca
rlo
simulatio
n
[1
6]. Cut set
and e
s
p
e
ci
al
ly minimal
cut set te
chni
que
can
be
an a
p
p
r
op
ri
ate
approa
ch for cal
c
ulating
reliability indices. Cut set i
s
a di
sconn
e
c
ting
set that brea
ks a pa
th
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN: 23
02-4
046
TELKOM
NI
KA
Vol. 12, No. 8, August 2014: 581
4 –
5826
5818
betwe
en a lo
ad nod
e an
d
a so
urce n
o
d
e
and mi
nima
l cut set i
s
a
cut set whi
c
h
has n
o
sub
s
ets
[17]. A minimal cut set is the
s
m
alles
t
set of comp
on
ents th
at can
ca
use a
sy
stem to fail
if they
fail [18].
In this techni
que f
a
ilure mo
de
s of the
system
should b
e
found. The failu
re mode
s of the
system
are i
dentified
eith
er vi
s
ually
or usi
ng th
e mi
nimal
cut
set
metho
d
[13].
Some
of the
s
e
method
s are introdu
ce
d by [1] and [17].
On the ba
si
s
of cut set techniqu
e, by tra
c
i
ng th
e path
starting f
r
om
sou
r
ce to loa
d
point,
failure m
ode
s can
be fo
un
d. For
simpli
city of finding failure
mod
e
s,
the sub
s
tatio
n
co
nfigu
r
atio
n
can b
e
displa
yed by graph
whi
c
h in
clude
s nod
es a
s
e
a
ch n
ode rep
r
esent an ele
m
ent.
Since
su
rg
e
arre
sters
should
act
co
mple
tely diffe
rent d
u
rin
g
n
o
rmal
ope
rati
ng an
d
overvoltage
condition
s, the
r
eby
reliabilit
y evaluati
on
has to b
e
pe
rformed
for
ea
ch
of the
s
e t
w
o
con
d
ition
s
se
parately. An
i
deal
su
rge
arrest
e
r
sh
ould
be
open
ci
rcuit duri
ng
normal op
erating
con
d
ition
an
d short
ci
rcu
i
t durin
g ove
r
voltage
s.
Po
ssi
ble
rea
s
o
n
s fo
r failu
re
of an
arre
ster
inclu
de the fo
llowing [6]:
a)
Overloa
d
ing
of the active
element
s by energy or current.
b) Moisture
ingress.
c)
Partial fla
s
ho
ver of o
ne
or
se
veral u
n
its
in a m
u
ltiunit
arreste
r
cau
s
ed by
extern
al pollutio
n
,
bird
s or hi
gh
overvoltage
s.
d)
Therm
a
l insta
b
ility due to the effect of heavy external
pollution.
e)
High temp
ora
r
y overvoltag
es.
f)
Dama
ge of some blo
c
ks i
n
one o
r
several unit
s
due t
o
ene
rgy and
current di
sch
a
rge
s
whi
c
h
lead
s to power freq
uen
cy overloa
d
of
the remai
n
ing p
a
rt of the arre
sters.
g)
Mech
ani
cal o
v
erloadi
ng which le
ad
s to an ele
c
trical failure.
Arre
sters ten
d
to be
sho
r
t
circuit
whe
n
they
enter f
a
tigue p
h
a
s
e
of their life.
Failure
durin
g impul
se passa
ge m
a
y occur d
u
e
to overloadi
ng or q
uality probl
em
s su
ch as
signifi
ca
nt
inhomo
gen
eities in the Zn
O blocks, po
or ele
c
trod
e adhe
sio
n
to the material, i
n
suffici
ent su
rface
ins
u
lation, etc [6].
Before evaluating
of substation reliability
fo
r the
cases of
differ
ent
surge
arrester
locatio
n
s, the
param
eters use
d
in this a
nalysi
s
are int
r
odu
ce
d in Ta
ble 5.
Table 5. Parameters used i
n
the Reliability Analysis
Parameter
Descrip
tion
λ
sc
short circuit failure rate of a
rreste
r
s dur
ing normal c
ondition of s
y
ste
m
(fail per
year)
r
sc
repair time of a
r
r
e
sters due to sho
r
t circuit failure (hour)
λ
oc
open circuit failure rate of a
rreste
r
s during overvoltage (fail per
year
)
λ
1
,
λ
2
busbars failure r
a
te (fail per
year
)
r
1
, r
2
busbars repair ti
me (hou
r)
λ
3
,…,
λ
14
circuit breakers failure rate (f
ail per
y
e
ar)
r
3
,…,
r
14
circuit breakers repair time (hou
r)
λ
15
,
λ
16
po
w
e
r t
r
ansform
ers failure rate
(f
ail per
y
e
a
r
)
r
15
, r
16
po
w
e
r t
r
ansform
ers repair time (
h
our)
t
ov
overvoltage
dura
t
ion
λ
L1
Annual number o
f
lightning stroke to inside
the substation due to shield w
i
re
failure
λ
L2
Annual number o
f
lightning stroke to subs
tation entr
ance due to shield w
i
r
e
failure
λ
L3
Annual number o
f
lightning stroke to points bet
w
e
e
n
50 m to 250 m
from substation d
ue to shield w
i
re
failur
e
P
L1
Probabilit
y
of an
nual lightning stroke to inside the substation due to
shield w
i
re failure
P
L2
Probabilit
y
of an
nual lightning stroke to substation entrance due t
o
shield w
i
re failure
P
L3
Probabilit
y
of an
nual lightning stroke to points betw
e
en 50 m to
25
0 m from substation due to shield w
i
re
failur
e
4.1. Modeling of Substati
on Reliabilit
y
due to Surge Arrester Location
In this part, variou
s cases
of surg
e arre
st
er lo
cation
s with their graph
s and eq
uation
s
are given.
a. Absen
c
e of
arre
ste
r
(A)
1) No
rmal o
p
e
rating
con
d
ition
It is worth po
inting out tha
t
annual light
ni
ng st
ro
ke a
nd probability
of annual lig
htning
stro
ke a
r
e correspon
ding a
nd have sam
e
values.
The graph of
this con
d
itio
n is sh
own in
Figure 3.T
h
e
minimal cut
sets o
b
taine
d
are as
follows
:
Evaluation Warning : The document was created with Spire.PDF for Python.
TELKOM
NIKA
ISSN:
2302-4
046
Reliability Analysis of Surge A
rrester Location Effect i
n
High
Voltage… (Seyed A
h
m
ad Hossei
ni)
5819
First o
r
de
r: 15
Secon
d
order: 1-2, 2-3, 1-5
,
3-5, 3-4, 1-4
The ave
r
a
ge
failure
rate,
a
v
erage
un
ava
ilability and
a
v
erage
outa
g
e
time a
r
e
cal
c
ulate
d
usin
g "(1)", "(2)" and "(3)" resp
ectively.
Square and
circle a
r
e
symbols of ci
rcuit
brea
ke
r an
d tran
sform
e
r re
spe
c
tively.
Figure
3
. Gra
ph of norm
a
l operating con
d
ition in
the case of ab
se
n
c
e of arre
ster
11
5
1
2
1
2
2
3
2
3
1
5
1
5
()
(
)
(
)
A
rr
r
r
rr
35
3
5
3
4
3
4
1
4
1
4
()
(
)
(
)
rr
rr
r
r
(
1
)
11
5
1
5
1
2
1
22
3
2
3
1
5
1
5
A
Ur
r
r
r
r
r
r
3
5
3
5
3
434
1
4
1
4
rr
r
r
r
r
(2)
11
1
/
A
AA
rU
(3)
2
)
O
v
e
r
vo
ltage
c
o
nd
itio
n
The graph of
this co
ndition
is sh
own in Figure 4.
Figure 4. Gra
ph of Overvol
t
age in the Case of Ab
sen
c
e of Arreste
r
Acco
rdi
ng to simulatio
n
ou
tcome
s
demo
n
strate
d in Table 4, striki
n
g
the lightnin
g
surge
to the poi
nts
insid
e
the
su
bstation,
sub
s
tation e
n
tra
n
ce
and
bet
wee
n
50
m t
o
250
m fro
m
the
sub
s
tation in
the case of
abse
n
ce of arre
ste
r
ca
n damag
e the equi
pmen
t like CVT a
n
d
transfo
rme
r
a
nd lead to
substatio
n
failure to
supply
load. Loa
d point reli
ability indice
s of thi
s
ca
se are cal
c
ulated u
s
ing "
(
4)", "(5
)
" and
"(6)".
21
2
3
A
LL
L
(4)
22
1
5
AA
Ur
(5)
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN: 23
02-4
046
TELKOM
NI
KA
Vol. 12, No. 8, August 2014: 581
4 –
5826
5820
22
2
/
A
AA
rU
(6)
b. Prese
n
ce of arre
ste
r
in sub
s
tation e
n
t
rance (E)
1) No
rmal o
p
e
rating
con
d
ition
Figure 5
de
monst
r
ate
s
t
he g
r
a
ph
of
this
con
d
ition
.
As
well
as bu
sba
r
, b
r
e
a
ke
r
and
transfo
rme
r
that were co
n
s
ide
r
ed in p
e
rvious case,
in this ca
se surge arre
ster failure is involv
ed
in reliability estimation. The mi
nimal cut sets are as follows:
First o
r
de
r: 15
Secon
d
order: 1-2, 2-3, 1-5
,
3-5, 3-4, 1-4
T
h
ir
d
or
d
e
r
:
17
-
1
8-
19
The load p
o
in
t reliability indice
s are give
n by "(7)", "(8
)
" and "(9
)
".
11
5
1
2
1
2
2
3
2
3
1
5
1
5
()
(
)
(
)
E
rr
r
r
r
r
32
35
3
5
3
4
3
4
1
4
1
4
()
(
)
(
)
(
3
)
s
cs
c
rr
r
r
r
r
r
(7)
11
5
1
5
1
2
1
22
3
2
3
1
5
1
5
3
5
3
5
E
Ur
r
r
r
r
r
r
r
r
33
34
3
4
1
4
1
4
s
cs
c
rr
r
r
r
(8)
12
2
/
E
EE
rU
(9)
2
)
O
v
e
r
vo
ltage
c
o
nd
itio
n
In order to calculate th
e load poi
nt reliability indices
of th
is
condition, all the possi
bilities
of sub
s
tation
failure h
a
ve
to be con
s
id
ered
by refe
rral to si
mulat
i
on re
sult
s gi
ven in Tabl
e
4
.
Shaded
valu
es in
that tabl
e indi
cate to t
he case
s in
whi
c
h ove
r
vol
t
age exceed
s the BIL, so
that
such values can be
causer of su
bstation failure. Fail
ure rate of
each possibilit
y can be seen in
Table
6 in
whi
c
h S
C
a
nd O
C
stan
d for
sh
ort
circuit a
nd o
pen
circuit
status of a
r
re
ster
r
e
spec
tively.
Triangle is
symbol of surge arrester
Figure 5. Gra
ph of Normal
Operating Co
ndition
in the
Ca
se of Pre
s
ence of Arre
ster in
Substation E
n
tran
ce
Figure 6. Gra
ph of overvoltage in the ca
se of
prese
n
ce of arre
ste
r
in sub
s
tation
entran
c
e
Evaluation Warning : The document was created with Spire.PDF for Python.
TELKOM
NIKA
ISSN:
2302-4
046
Reliability Analysis of Surge A
rrester Location Effect i
n
High
Voltage… (Seyed A
h
m
ad Hossei
ni)
5821
Table 6.
Possibilities of Overvoltage
Co
ndition in the
Ca
se of Pre
s
ence of Arre
ster in Substati
on
Entranc
e
Ligh
tni
ng s
t
rok
e
locati
o
n
Entranc
e
arrest
er
stat
us
Failure
rate
Inside the
substation
Substation
entrance
Between 50 m t
o
150 m from
substation
*
SC
or
O
C
λ
L1
*
SC
or
O
C
λ
L2
*
SC
0
*
O
C
P
L3
λ
oc
SC and O
C
stand for short
circuit and o
pen ci
rc
uit status of arre
ster re
sp
ectivel
y
.
Therefore,
load point
reliability indices
of th
is case
are calcul
ated
using "(10)", "(11)"
and "(1
2
)".
21
2
3
E
LL
L
o
c
p
(10)
22
1
5
EE
Ur
(11)
22
2
/
E
EE
rU
(12)
c. Presen
ce o
f
arre
ster in transfo
rme
r
feeder
(T)
1) No
rmal o
p
e
rating
con
d
ition
With the hel
p of Figure
7 minim
a
l
cut se
ts and load point reli
ability indices
can be
obtaine
d.
First o
r
de
r: 15, 17
Secon
d
order: 1-2, 2-3, 1-5
,
3-5, 3-4, 1-4
11
5
1
2
1
2
2
3
2
3
1
5
1
5
()
(
)
(
)
Ts
c
rr
r
r
rr
35
3
5
3
4
3
4
14
1
4
()
()
(
)
rr
rr
r
r
(13
)
1
1
51
5
1
21
2
2
3
2
3
1
5
1
5
Ts
c
s
c
Ur
r
r
r
r
r
r
r
35
3
5
3
4
3
4
1
4
1
4
rr
r
r
r
r
(14)
11
1
/
TT
T
rU
(15)
Figure 7. Gra
ph of Normal
Operating Co
ndition
in the
Ca
se of Pre
s
ence of Arre
ster in
Tran
sfo
r
mer Feede
r
2
)
O
v
e
r
vo
ltage
c
o
nd
itio
n
Same a
s
pre
v
ious
ca
se
a
ll the
po
ssibi
lities h
a
ve to
be
verified
and th
en
loa
d
poi
nt
reliability indi
ces can be estimated.
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN: 23
02-4
046
TELKOM
NI
KA
Vol. 12, No. 8, August 2014: 581
4 –
5826
5822
21
2
3
TL
L
L
o
c
p
(16)
22
1
5
TT
Ur
(17)
22
2
/
TT
T
rU
(18)
Figure 8. Gra
ph of Overvol
t
age in the Case of
Pre
s
e
n
ce of Arre
ster in Transfo
rmer Fe
ede
r
Figure 9.
Gra
ph of Normal
Operating Co
ndition in
the
Ca
se of Pre
s
ence of Arre
ster in
Substation E
n
tran
ce an
d Tran
sfo
r
mer
Feede
r
Table 7.
Possibilitie
s of Overvoltage Condition in th
e Ca
se of Pre
s
en
ce of Arre
ster in
Tran
sfo
r
mer Feede
r
Ligh
tni
ng s
t
rok
e
locati
o
n
Transf
ormer arrester
stat
us
Failure
rate
Inside the
substation
Substation
entrance
Between 50 m t
o
150 m from
substation
*
SC
or
O
C
λ
L1
*
SC
or
O
C
λ
L2
*
SC
0
*
O
C
P
L3
λ
oc
d. Prese
n
ce of arre
ste
r
in entran
c
e a
nd
transfo
rme
r
feede
r (E+T)
1) No
rmal o
p
e
rating
con
d
ition
Cut set
s
and i
ndexe
s
of this ca
se a
r
e a
s
follows:
First o
r
de
r: 15, 20
Secon
d
order: 1-2, 2-3, 1-5
,
3-5, 3-4, 1-4
T
h
ir
d
or
d
e
r
:
17
-
1
8-
19
()
1
1
5
1
2
1
2
2
3
2
3
()
(
)
ET
s
c
rr
r
r
15
1
5
3
5
3
5
3
4
3
4
()
(
)
(
)
rr
r
r
r
r
14
1
4
33
)
()
(
3
s
cs
c
rr
r
(19)
Evaluation Warning : The document was created with Spire.PDF for Python.
TELKOM
NIKA
ISSN:
2302-4
046
Reliability Analysis of Surge A
rrester Location Effect i
n
High
Voltage… (Seyed A
h
m
ad Hossei
ni)
5823
()
1
1
5
1
5
1
2
1
22
3
2
3
E
T
sc
sc
Ur
r
r
r
r
r
33
15
1
5
3
5
3
5
3
4
3
4
1
4
1
4
s
cs
c
r
r
rr
rr
r
r
r
(
2
0
)
()
1
(
)
1
()
1
/
E
T
ET
ET
rU
(21)
Figure 10.
Graph of Overv
o
ltage in the
Ca
se of Pre
s
enc
e of Arre
ster in Substati
on Entran
ce
and
Tran
sfo
r
mer Feede
r
2
)
O
v
e
r
vo
ltage
c
o
nd
itio
n
Owing to
presence of more t
han one arre
ster in thi
s
case, the number of possibilities
that have to be con
s
id
ere
d
become mo
re
, that are sho
w
n in Tabl
e 8
.
Owin
g to ve
ry
sm
all valu
e o
f
overvoltag
e
duratio
n
(aro
und fe
w
microse
c
o
n
d
s
), th
e term
s
including t
ov
can
be ignored, i.e. probability of entra
nce and transformer
arresters failure
simultan
eou
sl
y is practi
cal
l
y impossi
ble
.
So
load point reliability
indice
s ca
n
be written as
follows
:
()
2
1
2
2(
)
ET
L
L
o
c
pp
(22)
()
2
(
)
2
1
5
ET
ET
Ur
(23)
()
2
(
)
2
(
)
2
/
ET
ET
ET
rU
(24)
Finally after
cal
c
ulating the
average fai
l
ure rate and aver
age unavailability of normal
operating
an
d overvolta
g
e
con
d
ition
s
separately, total value
of e
a
ch
ca
se
ave
r
age
failure rate
and ave
r
a
g
e
unavaila
bility will be
su
m of conditi
o
n
s one
s
of e
a
ch
ca
se. Consequ
ently
total
averag
e re
pai
r time is divisi
on of total averag
e unavail
ability on total average failu
re rate.
Table 8.
Possibilities of Overvoltage
Co
ndition in t
he
Ca
se of Pre
s
ence of Arre
ster in
Tran
sfo
r
mer Feede
r
Ligh
tni
ng s
t
rok
e
locati
o
n
A
r
res
t
er st
atu
s
Failure ra
te
Inside the
substation
Substation
entrance
Between 50 m
to 150 m from
substation
Entrance
arres
t
er
Transform
er
feeder a
rrester
*
SC
SC
0
*
O
C
SC
P
L1
λ
OC
*
SC
O
C
P
L1
λ
OC
*
OC
OC
2p
L1
λ
oc
(
λ
oc
t
ov
)
*
SC
SC
0
*
O
C
SC
P
L2
λ
OC
*
SC
O
C
P
L2
λ
OC
*
OC
OC
2P
L2
λ
oc
(
λ
oc
t
ov
)
*
SC
SC
0
*
O
C
SC
0
*
SC
O
C
0
*
OC
OC
2P
L3
λ
oc
(
λ
oc
t
ov
)
Evaluation Warning : The document was created with Spire.PDF for Python.