Internati
o
nal
Journal of Ele
c
trical
and Computer
Engineering
(IJE
CE)
V
o
l.
6, N
o
. 5
,
O
c
tob
e
r
201
6, p
p
. 2
088
~209
5
I
S
SN
: 208
8-8
7
0
8
,
D
O
I
:
10.115
91
/ij
ece.v6
i
5.1
006
6
2
088
Jo
urn
a
l
h
o
me
pa
ge
: h
ttp
://iaesjo
u
r
na
l.com/
o
n
lin
e/ind
e
x.ph
p
/
IJECE
Internal On-line Partial Disc
harge Analysis of 68.75 MVA
Generator Stator Winding Insulation
Walu
yo
1
, Siti
Saod
ah
2
, E
l
t
h
a Hi
d
a
ya
tul
l
a
h
3
1,3
Departm
e
nt
of
El
ectr
i
cal
Engin
eering
,
National
In
stitute of
Tech
nolog
y
(It
enas), Bandung,
Indon
esia
2
Department
of
Energ
y
Conv
ersion, Bandung
St
ate Poly
techn
i
c (
P
olban), B
a
ndun
g, Indonesia
Article Info
A
B
STRAC
T
Article histo
r
y:
Received Feb 4, 2016
Rev
i
sed
May 26
, 20
16
Accepted
Jun 15, 2016
Partial dis
c
harg
e is a phenomenon of
electro
n
ionization occurs due to
concen
trat
ed el
ectri
c
field in a di
fferent
edge pl
an
e. Th
is phenom
enon will be
inves
tiga
t
ed b
y
an elec
tric fi
el
d m
eas
urem
ent in a m
eas
uring point. The
intern
al p
a
rti
a
l
discha
rge
wi
ll gi
ve
t
h
e i
n
sula
tion d
a
mage
effect on
a
generator stator
winding due to
void exis
t
e
nce
.
This m
a
nuscript
presents th
e
m
easurem
ent re
sults of the
on-l
i
ne in
tern
al p
a
rt
ial d
i
scharg
e on
the st
ato
r
winding insulation of 68.75 MV
A generator.
I
t
u
s
ed the r
e
s
i
s
t
ant
tem
p
eratur
e
detector method
and CM2000
TM
. The results
were classifi
ed
into three
conditions b
a
sed
on the voids in
the g
e
ne
r
a
tor st
ator insul
a
tio
n,
t
h
e int
e
rna
l
delamination an
d
the
surface d
i
scharge.
Keyword:
Gene
rato
r
On-lin
e
Partial d
i
sch
a
rg
e
Stator
W
i
nd
ing
insu
latio
n
Copyright ©
201
6 Institut
e
o
f
Ad
vanced
Engin
eer
ing and S
c
i
e
nce.
All rights re
se
rve
d
.
Co
rresp
ond
i
ng
Autho
r
:
W
a
lu
yo,
Depa
rt
m
e
nt
of
El
ect
ri
cal
Engi
neeri
n
g
,
Natio
n
a
l In
stitu
te
of
Techn
o
l
o
g
y
(ITENAS),
Jl.
PHH
.
Mu
st
af
a N
o
. 2
3
Band
ung
, 40
124
I
n
d
o
n
e
sia.
Em
a
il: walu
yo
@iten
a
s.ac.id
1.
INTRODUCTION
Hi
g
h
v
o
l
t
a
ge el
ect
ri
c genera
t
o
rs are esse
nt
i
a
l
co
m
pone
nt
s of rel
i
a
bl
e e
l
ect
ri
c powe
r
gene
rat
i
o
n
syste
m
. They m
u
st operate
continuously and so m
u
st
be very
reliable
because a
faul
t in a hi
gh voltage
gene
rat
o
r co
ul
d com
p
rom
i
se t
h
e powe
r
su
ppl
y
.
The c
o
n
t
i
nuo
us o
p
e
r
at
i
on i
s
cri
t
i
cal
, so t
h
at
t
h
e on-l
i
n
e
di
ag
no
si
s i
s
t
h
e best
pre
v
e
n
t
i
v
e m
easure.
Fu
rt
herm
ore,
th
e
fau
lts are larg
ely cau
sed
b
y
insu
latio
n breakd
o
wns
i
n
t
h
e st
at
o
r
wi
ndi
ng
, s
o
t
h
at
t
h
e com
m
on m
e
t
h
o
d
o
f
on
-l
i
n
e di
ag
no
si
s i
s
part
i
a
l
di
sc
har
g
e (
P
D
)
det
ect
i
on i
n
t
h
e st
at
or
wi
nd
i
ng
[
1
]
.
Gen
e
rator i
n
sulatio
n
system
s,
ho
wev
e
r
p
e
rfect th
eir
i
n
itial state, in
ev
itab
l
y
d
e
g
r
ad
e i
n
serv
ice.
Heat,
t
h
erm
a
l
cy
cli
ng,
bar
f
o
rce
s
,
v
i
brat
i
o
n, m
echani
cal
sh
oc
k, s
h
ri
nka
ge
of
su
pp
o
r
t
st
ruct
ure
s
, s
u
ch
as
wed
g
es
a
n
d
spaces, and t
h
e presence
of
high electric s
t
ress, act a
n
d
interact to impair t
h
e
inte
grity of the
diel
ectric
syste
m
s. At so
me stag
es, p
a
rt
ial d
i
sch
a
rg
es will start
an
d
b
e
g
i
n
to
in
crease, p
r
ov
id
ing
an
add
itio
n
a
l ero
s
ive
ag
ing
facto
r
. Th
is usu
a
lly o
c
cu
rs at abrad
e
d
o
u
t
er sh
ield
ing, in
cav
ities or d
e
lam
i
n
a
tio
n
s
, or in th
e end-tu
rn
structure [2].
Partial disc
harges
(PDs) are
sm
all electric
a
l spar
ks
res
u
l
ting
from
the electrical brea
kdown tha
t
occu
r
whe
n
v
o
i
d
e
x
i
s
t
wi
t
h
i
n
o
n
t
h
e s
u
r
f
a
ce or i
n
ot
he
r
hi
g
h
l
y
n
o
n
-
u
n
i
f
o
r
m
el
ect
ri
c f
i
el
d of
hi
gh
v
o
l
t
a
ge
i
n
sul
a
t
i
o
n
of
st
at
or
wi
n
d
i
n
g
s
i
n
ge
nerat
o
r
s
an
d m
o
t
o
r
s
.
The
s
e P
D
p
u
l
s
es ca
n
occ
u
r
beca
use
o
f
t
h
e
m
a
nufact
uri
n
g
an
d i
n
st
al
l
a
t
i
on
p
r
oces
ses,
t
h
erm
a
l
det
e
ri
orat
i
o
n,
wi
n
d
i
n
g c
ont
am
i
n
at
i
on
or
st
at
o
r
bar
m
o
v
e
m
e
n
t
d
u
r
in
g
o
p
e
ration
.
If
th
e vo
id
with
in
an
org
a
n
i
c so
lid
o
r
liqu
i
d
,
th
e
PD
will d
e
grad
e th
e
org
a
n
i
c
material an
d
may ev
en
tu
ally cau
se th
e failu
re o
f
th
e el
ectrical in
su
latio
n
.
As th
e insu
latio
n
d
e
grad
es, th
e
num
ber an
d m
a
gni
t
u
de
of t
h
e PD p
u
l
s
e wi
l
l
i
n
crease. Al
t
h
o
u
gh t
h
e m
a
gni
t
u
de
of t
h
e
PD p
u
l
s
es ca
n
not
be
d
i
rectly related
to
th
e rem
a
in
i
n
g
life
o
f
th
e
wind
ing
,
th
e
do
ub
ling
PD
pulse
m
a
g
n
itudes
approxim
ately every
Evaluation Warning : The document was created with Spire.PDF for Python.
I
J
ECE
I
S
SN
:
208
8-8
7
0
8
Int
e
r
nal
O
n
-l
i
n
e Part
i
a
l
Di
sc
h
a
rg
e A
nal
ysi
s
of
6
8
.
7
5 MV
A
Gene
rat
o
r
St
at
or Wi
ndi
ng
I
n
s
u
l
a
t
i
o
n (
W
al
uy
o)
2
089
6 m
ont
hs, rat
e
of P
D
p
u
l
s
e a
c
t
i
v
i
t
y
i
n
crease rapi
dl
y
,
or t
h
e PD l
e
vel
s
are hi
g
h
l
y
com
p
are
d
t
o
ot
he
r sim
i
l
a
r
mach
in
es, th
is is an
in
d
i
cator th
at v
i
su
al insp
ection
s
and
/
o
r
o
t
h
e
r testing
m
e
th
od
s are n
eed
to
co
nfi
r
m
th
e
in
su
lation
co
nditio
n
[3
],[4
].
Partial discharges (PDs
) have been rec
o
gni
zed as
harm
ful
agei
ng
pr
oces
s fo
r electrical insulation at
th
e b
e
g
i
n
n
i
ng
o
f
th
e last cen
t
u
r
y
wh
en
th
e H
V
techno
logy w
a
s in
tr
od
u
c
ed
fo
r
th
e gen
e
r
a
tin
g
an
d
tr
ansmissio
n
of electrical powe
r. Pa
rtial discha
rge
s
are
conse
q
ue
nce
of local electrical stress conce
n
trations
in the
insulation or on the surface
of
the insulation. Gene
rally, such disc
harg
e a
ppea
r
s as pulse
having a durat
i
on of
m
u
ch
less th
an 1
s [5
].
Partial dischar
g
e com
e
s fro
m
differe
nt fa
ctors, s
u
c
h
as
therm
a
l, elec
trical
fi
el
d, en
vi
r
onm
ent
a
l
a
m
bient and
mechanical vibr
ation
s
.
Defects id
en
tified
b
y
p
a
rtial d
i
sc
harge m
easurements are base
wedges
and
ba
r
vi
b
r
at
i
o
n
,
sl
ot
di
sc
ha
rges
,
br
o
k
en
c
o
n
d
u
ct
o
r
s, c
o
n
t
am
i
n
at
i
on o
f
dam
a
ge of
en
d
-
wi
ndi
ng
,
di
schar
g
es
bet
w
ee
n
pha
se
s as a
res
u
l
t
o
f
vi
b
r
at
i
o
n
[
6
]
.
More tha
n
ten years, an ext
e
nsiv
e re
searc
h
pr
o
j
ect
was
un
de
rt
ake
n
t
o
devel
op a s
u
p
e
ri
or
part
i
a
l
d
i
sch
a
rg
e test tu
rb
in
e gen
e
rato
rs. Th
e resu
lts were th
e tu
rbin
e g
e
n
e
rator an
alyzer test, wh
ich
was
d
e
sign
ed
t
o
be
per
f
o
r
m
e
d by
n
o
n
-s
peci
al
i
zed ge
ne
rat
o
r
st
at
i
on st
a
ff
wi
t
h
o
u
t
a
ge
n
e
rat
o
r
o
u
t
a
ge
.
The m
a
i
n
t
echni
ca
l
adva
nt
age
of t
h
i
s
t
e
st
was t
h
at
fal
s
e i
ndi
cat
i
ons
of
d
e
terioratin
g
insu
latio
n
were
v
i
rtu
a
lly el
i
m
in
ated
. Th
e
main
d
i
fficu
lty in
p
e
rform
i
n
g
an
on
-lin
e PD test was n
o
t
in
d
e
tectin
g
th
e PD sign
als, bu
t rath
er d
i
stingu
ish
i
ng
t
h
e PD
fr
om
el
ect
ri
cal
noi
se.
The
noi
se
wa
s fo
u
nd t
o
be
very erratic overtim
e, and s
o
me time as
much a
s
1
000
tim
es h
i
gh
er th
an
g
e
n
e
rato
r
PD sign
als. Th
e trend
i
n
resu
lts fro
m
th
e on
-lin
e PD test will g
i
v
e
su
fficien
t
w
a
rn
ing
t
o
p
e
rmit t
i
m
el
y
m
o
d
i
f
i
catio
n of
gen
e
r
a
t
o
r
op
eratio
n
o
r
im
p
l
e
m
en
tatio
n
of
relativ
ely in
ex
pen
s
iv
e
stato
r
wind
ing
m
a
in
ten
a
n
ce. Th
e on
-lin
e PD testin
g
facilitated
ex
ten
d
i
ng
wind
ing
life an
d
red
u
c
ed
o
v
e
rall
m
a
i
n
t
e
nance c
o
st
s
by
av
oi
di
n
g
i
n
ser
v
i
ce
fai
l
u
res a
n
d
resul
t
i
ng
p
r
em
at
ure rewi
nds
[
7
]
.
Th
e statistical an
alysis of th
e
d
a
ta b
a
se
was t
h
e
d
i
stribu
tion
o
f
Q
m
as a
fu
n
c
t
i
on
of
wi
ndi
ng
age
.
T
h
e
PD res
u
l
t
s
i
n
t
h
e dat
a
ba
se fr
om
m
achi
n
es t
h
at
were
fr
om
1 t
o
m
o
re 50 y
ears ol
d sh
ow
e
d
n
o
co
nsi
s
t
e
nt
t
r
end
,
whi
c
h wa
s su
r
p
ri
si
n
g
beca
us
e one
w
oul
d
n
o
rm
al
l
y
assu
m
e
t
h
at
ol
de
r wi
ndi
ng
s w
o
ul
d
be m
o
re det
e
ri
orat
e
d
and t
h
us
have
hi
ghe
r P
D
l
e
vel
s
. The i
n
con
s
i
s
t
e
nt
pat
t
e
rn
OPD
vers
us wi
n
d
i
ng a
g
e m
a
y
i
n
cl
ude t
h
e
obs
er
vat
i
on t
h
at
m
a
nufact
ur
e
r
s of m
achi
n
e
s
have a l
earn
i
ng cu
r
v
e t
o
cl
im
b as t
h
e adopt
ne
w desi
g
n
an
d
man
u
f
act
u
r
i
n
g tech
n
i
q
u
e
s or th
at u
tilities
are con
tinu
o
u
s
ly o
s
cillatin
g
b
e
tween
p
r
o
a
ctiv
e and
breakd
own
m
a
i
n
t
e
nance st
rat
e
gi
es,
depe
n
d
i
n
g o
n
m
a
nagem
e
nt
pol
i
c
i
e
s.
An a
n
al
y
s
i
s
of t
h
e st
at
i
s
t
i
cal di
st
ri
b
u
t
i
on
of
PD
fo
r se
veral
m
a
nu
fact
u
r
ers
wa
s al
so
per
f
o
r
m
e
d. T
h
e ca
use
of t
h
e
di
ffe
re
nces
bet
w
ee
n
m
a
nufact
ure
r
s
was
un
k
n
o
w
n,
b
u
t
i
t
m
i
ght
be
due
t
o
di
ffe
rent
m
a
nu
fact
u
r
i
n
g
pr
ocesses
,
el
ect
ri
c st
ress
desi
g
n
l
e
vel
s
an
d
asse
m
b
ly
m
e
t
hods
[
8
]
.
PD an
d /
o
r E
M
I
m
oni
t
o
ri
ng
of hy
d
r
oge
n
-
c
ool
e
d
ge
nerat
o
rs ha
vi
n
g
t
i
ght
wi
ndi
ng
s, wi
t
h
or
wi
t
h
out
co
n
t
am
in
atio
n
,
do
es
no
t g
e
n
e
rally yield
additio
n
a
l in
form
a
tio
n
o
n
th
e in
su
latio
n cond
itio
n. EM
I ind
i
catio
n
s
of c
o
re
-ed
g
e
and e
n
d
-
wi
n
d
i
ng
di
scha
rge
s
coul
d
not
be
con
f
i
r
m
e
d by
i
n
spect
i
o
n
.
EM
I i
s
effect
i
v
e i
n
i
d
ent
i
f
y
i
n
g
pr
o
b
l
e
m
s
out
si
de t
h
e
wi
n
d
i
n
g, s
u
ch as t
hos
e rel
a
t
i
ng t
o
t
h
e i
s
o
-
p
h
ase
b
u
s
duc
t
com
pone
nt
s
or t
h
e
ex
citer.
Th
e in
t
e
rpretatio
n of
b
o
t
h
PD and
EMI sign
atures
req
u
i
res th
e sk
ill o
f
a trained insu
latio
n exp
e
rt [9
].
An
effecti
v
e on
-lin
e
PD m
o
n
ito
ri
n
g
system
m
u
st
ad
eq
uately filter an
d
th
e im
p
act o
f
no
ise and
di
st
ur
ba
nces i
n
o
r
der t
o
pr
ovi
de
val
i
d
re
sul
t
s
. T
h
e
noi
se an
d
di
st
ur
b
a
nces
fr
om
t
h
e p
o
we
r sy
st
e
m
and
harm
onics could influe
nce t
h
e quality of PD signals.
Absol
u
te hum
i
d
ity should
be
include
d in all
trend
anal
y
s
i
s
, es
pec
i
al
l
y
i
f
surface
PD m
oni
t
o
re
d f
o
r as l
o
n
g
as 3
0
y
ears
w
i
t
h
t
h
e sam
e
m
e
t
hod
o
f
n
o
i
s
e an
d
di
st
ur
ba
nce se
parat
i
o
n.
T
h
e
on
-l
i
n
e
part
i
a
l
di
scha
rge
t
e
st
i
n
g
ha
s
becom
e
a rec
o
gni
ze
d
,
p
r
ove
n t
ool
t
o
hel
p
main
ten
a
n
ce en
g
i
n
eers id
en
tify wh
ich stato
r
wind
ing
s
n
e
ed off lin
e testin
g, in
sp
ection
and
/or
rep
a
i
r
s
[10
]
.
M
a
t
l
a
b Sim
u
l
i
nk
based m
odel
has bee
n
i
n
t
r
o
d
u
ced t
o
generat
e
t
h
e
pul
se an
d d
e
t
ect
it
. One
si
nus
oi
dal
ac
cy
cl
e was t
a
ken an
d di
vi
de
d i
n
t
o
8 segm
ent
s
an
d t
h
e n
u
m
b
er of
pul
s
e
s appea
r
i
n
g f
o
r ea
c
h
segm
ent
s
was fo
u
nd
by
M
a
t
l
ab p
r
o
g
ram
.
The res
u
l
t
s
fo
r PD pulse count
for 5 kV and 10 kV. T
h
e calibrating
ci
rcui
t
was
m
odel
e
d i
n
M
a
t
l
a
b Si
m
u
l
i
nk t
o
creat
e desi
red
out
put
PD
p
u
l
s
es wi
t
h
re
qui
r
e
d t
h
e
cha
r
ge l
e
vel
s
as
5
pC
,
10
pC
,
5
0
pC
,
10
0
pC
and
5
0
0
pC
. T
h
e cal
i
b
rat
o
r
c
i
rcui
t
was
co
n
n
ect
ed ac
r
o
ss t
h
e
ob
ject
a
n
d
out
put
p
u
l
ses were si
milar to
th
e calib
ratin
g pu
lses as requ
ired
. Th
e
ph
ysical m
o
d
e
l o
f
calibrato
r was m
a
d
e
and
out
put
wa
ve
fo
rm
s were
obse
r
ved
o
n
DS
O.
T
h
e
out
put
was
sim
i
l
a
r t
o
t
h
e s
i
m
u
l
a
t
i
on res
u
l
t
s [1
1]
.
It
has
been
pr
esent
e
d t
h
e m
ode
r
n
OL
PD t
e
st
i
ng a
nd m
oni
t
o
ri
ng t
e
c
h
n
o
l
o
gy
. It
wa
s
al
so ex
pl
ai
ne
d
som
e
part
i
a
l
di
scha
rge
sen
s
or
o
p
t
i
o
n
s
,
na
m
e
l
y
hi
gh v
o
l
t
age co
u
p
l
i
n
g
capaci
t
o
r
sen
s
ors
,
hi
g
h
f
r
e
que
ncy
cu
rr
en
t t
r
an
sfor
m
e
r
sen
s
or
s,
Ro
go
wsk
i
co
il sen
s
or
s an
d
tran
sien
t earth vo
ltag
e
sen
s
o
r
s. Th
e sign
ifican
t cost
and
o
p
erat
i
o
na
l
bene
fi
t
s
co
ul
d be
gai
n
ed
f
r
om
com
p
l
e
t
e
po
we
r ge
ne
ra
t
i
on, i
n
d
u
st
ri
al
and
pet
r
oc
he
m
i
cal
in
du
str
y
MV netw
or
k
s
. Th
e
data f
r
o
m
co
n
tin
uou
s CM tech
no
log
y
can
be u
s
ed
to suppo
r
t
CBM sch
e
mes an
d
to
d
i
rect prev
en
tiv
e m
a
in
ten
a
n
ce in
terv
en
ti
on
s to
rep
a
ir
p
l
at o
r
cab
les ahead
of in
su
latio
n
failu
re
from PD
activ
ity [1
2
]
.
It
has
been
pre
s
ent
e
d s
o
m
e
PD det
ect
i
o
n m
e
t
h
o
d
s,
nam
e
ly electrical, chemi
cal, acoustic and
optical
m
e
t
hods
o
f
t
h
e sens
o
r
t
y
pes
.
T
h
e
pr
o
b
l
e
m
s
ass
o
ci
at
ed
w
i
t
h
PD
f
o
r
t
h
e
hi
g
h
v
o
l
t
a
ge
equi
pm
ent
and
hi
g
h
Evaluation Warning : The document was created with Spire.PDF for Python.
I
S
SN
:
2
088
-87
08
I
J
ECE
Vo
l. 6
,
N
o
. 5
,
O
c
tob
e
r
20
16
:
208
8
–
20
95
2
090
v
o
ltag
e
po
wer
syste
m
s h
a
v
e
no
t b
e
en
ign
o
red
and
will
n
e
ver b
e
i
g
no
red
by h
i
g
h
vo
ltag
e
syste
m
d
e
sig
n
e
rs and
m
a
i
n
t
e
nance e
ngi
neers
.
P
D
i
d
ent
i
f
i
cat
i
o
n a
n
d
P
D
m
oni
t
o
r
i
ng m
a
ke ec
on
om
i
c
sense l
o
w a
n
d
m
a
ny
y
ears t
o
com
e
[13]
.
A po
sitiv
e PD
is g
e
n
e
rated
by electro
n
s
i
n
itiated
b
y
co
llisio
n
d
e
tach
m
e
n
t
o
f
n
e
g
a
tiv
e ion
s
in
a
h
i
gh
electric field
reg
i
on
. Thu
s
, the p
o
sitiv
e PD
g
e
n
e
ration
d
e
pen
d
s on
wh
et
her or no
t n
e
g
a
t
i
v
e
io
ns ex
ist aroun
d
t
h
e pa
rt
i
c
l
e
t
i
p
. O
n
t
h
e
ot
he
r
han
d
,
ne
gat
i
v
e PD i
s
gen
e
r
a
t
e
d by
a
n
i
n
i
t
i
a
l
el
ect
ron
de
ri
ve
d f
r
om
t
h
e fi
el
d
e
m
ission from the electrode
surface [
14].
Partial discha
rge m
easurem
e
n
ts on operati
ng m
achines c
a
n be
in
flu
e
n
c
ed
b
y
co
nd
itio
ns su
ch
as hu
m
i
d
ity,
te
m
p
eratu
r
e, term
in
al v
o
ltag
e
an
d lo
ad
[15
]
.
This resea
r
c
h
was to a
n
alysis and classifies
th
e condition
base
d on the a
m
ount of c
h
a
r
ges
per cycle
of
pa
rt
i
a
l
di
sc
har
g
e.
The
pa
r
t
i
a
l
di
schar
g
e t
h
at
m
oni
t
o
re
d
was
occu
rre
d i
n
t
h
e
6
8
.
7
5 M
V
A
ge
nerat
o
r
st
at
or
wind
ing
insu
latio
n
.
2.
R
E
SEARC
H M
ETHOD
The use
d
m
easuri
ng m
e
t
hod i
n
t
h
i
s
researc
h
was On
-l
i
n
e p
a
rt
i
a
l
di
schar
g
e (PD
)
m
easurem
ent
usi
n
g
resistan
ce temp
erat
u
r
e
d
e
tecto
r
(RTD) m
e
th
od
. Th
is m
e
th
od
was to
measu
r
e th
e partial d
i
sch
a
rge th
at
o
ccurred
i
n
th
e g
e
n
e
rat
o
r
stato
r
i
n
su
latio
n
.
Th
e illu
strated d
i
agram
o
f
the m
easu
r
e
m
en
t is shown
i
n
Fi
g
u
re
1
bel
o
w.
Fi
gu
re
1.
Di
a
g
r
a
m
of
part
i
a
l
di
schar
g
e m
easu
r
em
ent
The pa
rt
i
a
l
di
schar
g
e m
easurem
ent
used C
M
200
0
TM
t
h
at
con
n
ect
ed usi
ng a cabl
e
f
r
o
m
C
M
2000
TM
to RTD at t
h
e gene
rator
and
ground.
The m
easur
emen
t d
a
ta sam
p
l
i
n
g
co
llectio
n
was carried
ou
t
au
to
m
a
tical
ly
u
s
ing
a co
m
p
u
t
er an
d PD m
o
nito
r
so
ftwa
r
e
that p
r
ov
id
ed
with
CM200
0.
Diagnosis of
pa
rtia
l d
i
scha
rge
:
The m
easurements
were ca
rried ou
t
u
s
ing
th
ro
ugh
th
e RTD
(Resistan
c
e Tem
p
eratu
r
e Detector) as
a
sens
or
m
o
u
n
t
e
d on
t
h
e ge
nera
t
o
r
st
at
o
r
wi
n
d
i
ngs
.
The
used
m
easuri
ng i
n
st
r
u
m
e
nt
was
Tos
h
i
b
a
C
M
20
00
that
connects the
ca
ble pr
o
b
es
fr
o
m
C
M
200
0 t
o
t
h
e
R
T
D of
ge
nera
t
o
r
st
at
o
r
wi
n
d
i
ngs
.
The
part
i
a
l
di
schar
g
e m
a
gni
t
ude
s w
e
re m
easure
d
i
n
m
i
l
l
i
vol
t
a
ge
(m
V).
The dat
a
col
l
e
ct
i
on was pe
rf
orm
e
d aut
o
m
a
t
i
cal
l
y
usi
ng t
h
e com
put
er and m
oni
t
o
r P
D
soft
ware t
h
at
available
on the CM2000.
2.
1.
The terms
those need to be
c
o
nsidere
d whe
n
tes
t
ings:
PD t
e
st
i
n
gs
sh
oul
d
be
do
ne i
n
a t
h
erm
a
l
l
y
st
abl
e
co
ndi
t
i
o
n
.
Th
is m
ean
t th
at th
e stato
r
win
d
i
ng
tem
p
eratu
r
e
d
i
d
no
t ch
ang
e
du
ring testin
g
s
.
Temp
erat
u
r
e stab
ility was
o
b
t
ain
e
d
at t
h
e o
f
th
e
g
e
n
e
rat
o
r
with
t
h
e same lo
ad
in
a certain
tim
e p
e
rio
d
o
r
un
til th
e te
m
p
eratu
r
e did
no
t
change.
1.
PD
t
e
st
i
n
gs sh
oul
d be do
ne f
a
st
l
y
.
If th
e testin
g
s
are do
n
e
fastly
, th
e p
r
o
b
a
b
ility o
f
te
m
p
erature ch
an
g
e
will b
e
min
i
m
i
zed
.
Th
e tem
p
eratu
r
e
ch
ang
e
will in
fl
u
e
n
c
e th
e
co
mpa
r
ab
ility
and
accu
r
acy
of testin
resu
lts. Th
is case will b
e
co
me be
i
m
p
o
r
tan
t
wh
en
it is carried
o
u
t
t
h
e testings with lo
ad
v
a
riatio
n
,
su
ch
as
fu
ll lo
ad
and
no
load
tests
af
te
r
th
e lo
ad
is
released
.
2.
The l
o
ad and t
h
e tem
p
erature
does
not cha
n
ge m
o
re tha
n
a
fe
w
perce
n
t.
Wh
en
co
llecting
d
a
ta / p
a
rtial d
i
sch
a
rg
e test,
th
e test sh
ou
ld
b
e
p
e
rfo
r
m
e
d
un
d
e
r th
e sam
e
co
nd
itio
ns (lo
a
d
and tem
p
erature).
3.
If t
h
e test conditio
n
s
ch
ang
e
,
(eg, coo
ling
air hu
m
i
d
i
t
y
),
will affect th
e level o
f
p
a
rtial d
i
sch
a
rg
e an
d n
eed
to
b
e
con
s
i
d
ered
in th
e an
alysis of
p
a
rtial d
i
sch
a
rg
e.
4.
Wh
en
co
m
p
ari
n
g
th
e
q
u
an
tity o
f
p
a
rtial d
i
sch
a
rg
e testin
g, it is
k
e
p
t
in
min
d
th
e typ
e
o
f
u
s
ed
insu
lat
i
o
n
.
Th
e
d
i
fferen
t t
y
p
e
s
o
f
i
n
su
latio
n are u
s
ed
,
it will p
r
od
u
c
e d
i
fferen
t
d
i
sch
a
rg
e lev
e
ls.
Evaluation Warning : The document was created with Spire.PDF for Python.
IJECE
ISS
N
:
2088-8708
Int
e
r
nal
O
n
-l
i
n
e Part
i
a
l
Di
sc
h
a
rg
e A
nal
ysi
s
of
6
8
.
7
5 MV
A
Gene
rat
o
r
St
at
or Wi
ndi
ng
I
n
s
u
l
a
t
i
o
n (
W
al
uy
o)
2
091
Th
e resu
lts
o
f
p
a
rtial d
i
sch
a
rg
e testing
o
n
g
e
n
e
ra
tors that are no
t in
syn
c
hrono
u
s
con
d
ition
wit
h
n
e
two
r
k
s
, th
ey
w
ill b
e
d
i
f
f
e
ren
t
f
r
o
m
th
e g
e
n
e
r
a
t
o
r
in
synch
r
on
ou
s conditio
n
w
ith
th
e n
e
two
r
k
.
Th
e p
a
r
tial
d
i
sch
a
rg
e in
g
e
n
e
rat
o
rs th
at are no
t in
syn
c
h
r
o
nou
s con
d
ition
with
th
e n
e
t
w
ork will
b
e
h
i
g
h
e
r.
The first equipment used in t
h
e m
easurem
e
n
t was
RT
D PT 100 (resista
nce tem
p
erature detector).
RTD
o
r
also
called
as
resistan
ce th
erm
o
meters is a tem
p
eratu
r
e senso
r
th
at
u
tilize ch
ang
e
o
f
electrica
l
resistan
ce in
a
certain
m
a
teria
l
to
ch
ang
e
th
e te
m
p
erat
u
r
e.
RTD referred
t
o
a m
easu
r
e o
f
p
o
s
itiv
e tem
p
eratu
r
e
co
efficien
t wh
ich
m
ean
s th
at t
h
e resistan
ce
will in
crease w
ith
th
e te
m
p
eratu
r
e. PT
10
0
is a p
l
atin
u
m
ma
terial
with
cond
itio
n at 0
o
C
t
e
m
p
erat
ure
s
has a
r
e
si
st
ance o
f
1
00
o
h
m
s
. R
T
D i
s
use
d
i
n
gene
rat
o
rs t
o
m
oni
t
o
r
winding tem
p
erature c
h
a
n
ges
due t
o
loa
d
c
u
rrent
vari
ations. Six pieces
of
RTD we
re
placed bet
w
ee
n the top-
b
o
tto
m
co
ils so th
at th
e
ch
ange in
each
co
il co
u
l
d
b
e
d
e
tected
.
Fi
gu
re
2(a
)
sh
ows
t
h
e
l
o
cat
i
o
ns
of
R
T
D
on
t
h
e
gen
e
rat
o
r st
at
or
wi
n
d
i
n
g.
I
t
i
s
i
ndi
cat
e
d
i
n
sy
m
e
t
r
i
cal
am
ong
t
h
em
. Fi
gu
re
2(
b)
s
h
o
w
s t
h
e
posi
t
i
on R
T
D
on
t
h
e ge
ne
rat
o
r
st
at
or
wi
n
d
i
n
g
.
Thi
s
fi
g
u
r
e
i
s
t
h
e
extraction in t
h
e sheet of Fi
gure 2(a). T
h
e st
ator
wi
nding of phase R
(U)
was
placed a
n
d se
nsed by the RTD
pr
o
b
es of
B
2
,
C
2
,
B
1
a
n
d
A
1
. Whi
l
e
,
t
h
e
st
at
or wi
n
d
i
n
g of p
h
ase S
(
V
)
was pl
ace
d
a
n
d
se
nse
d
by
t
h
e
R
T
D
pr
o
b
es
of
C
2
,
A2
, C
1
a
n
d
B
1
. Fi
nal
l
y
,
t
h
e s
t
at
or
wi
n
d
i
n
g
of
p
h
ase
T (
W
)
was
pl
ace
d a
n
d se
nse
d
by
t
h
e R
T
D
pr
obes of
A2,
B2, A1
a
n
d
C1.
(a) RT
D
place
m
e
nt
(b
) R
T
D
po
si
t
i
ons
Fig
u
re
2
.
RTD
in
stallatio
n
and po
sitio
n on
t
h
e g
e
n
e
rat
o
r
stato
r
wi
n
d
i
n
g
Table 1 lists the placem
ent of RTD positions. RTD
A1 was installed
on R and T pha
s
es, RTD
A2
was i
n
stalled
on T
and
S
pha
s
e
s and RT
D B
1
was i
n
sta
lled on
S and
R
phases. For
sam
e
sequ
en
ce, RTD
s
o
f
B2, C
1
a
n
d C2 we
re installed
on R a
n
d T
phases,
T
and
S
phases a
n
d S and R
phases
res
p
ectively.
Table 1. Placement
of
RTDs
Figure 3(a) s
h
ows the typical
RTD with cable. The
eq
ui
pm
ent
was use
d
i
n
t
h
e electric generat
o
r for
part
i
a
l
di
sc
har
g
e
det
ect
i
ons
.
The
seco
n
d
use
d
eq
ui
pm
ent
was
C
M
2
0
0
0
TM
a
n
al
y
zer
of
pa
rt
i
a
l
di
s
c
har
g
e
m
easurem
ent
sy
st
em
, Fi
gure
3(
b)
w
h
i
c
h se
r
v
ed t
o
m
oni
t
o
r
t
h
e part
i
a
l
di
s
c
har
g
e t
h
at
m
i
ght
occu
r i
n
t
h
e st
at
or
g
e
n
e
r
a
tor
.
Fi
nally, th
e su
ppor
t to
o
l
s
w
e
r
e
BN
C co
ax
ial cab
les in
clud
ing
pr
ob
es
w
ith
len
g
t
h
of
5
m
e
ter
s
, as
sho
w
n i
n
Fi
gu
r
e
3
(
c).
RTDs
Place
m
e
nt
Position Slot
Type
Phase
A1 Centr
e
1
M
i
ddle
R,
T
A2 E
x
citer
9
L
ong
T
,
S
B1 Centr
e
17
M
i
ddle
S,
R
B2 E
x
citer
25
L
ong
R,
T
C1 Centr
e
33
M
i
ddle
T
,
S
C2 E
x
citer
41
L
ong
S,
R
Evaluation Warning : The document was created with Spire.PDF for Python.
I
S
SN
:
2
088
-87
08
I
J
ECE
Vo
l. 6
,
N
o
. 5
,
O
c
tob
e
r
20
16
:
208
8
–
20
95
2
092
(a) RTD
with
cable
(b) CM2000
TM
(c) Coax
ial
c
a
bl
e and
BNC prob
e
Fi
gu
re 3.
C
M
2
0
0
0
TM
an
alyzer of
p
a
rtial d
i
sch
a
rg
e m
eas
u
r
emen
t syste
m
an
d aux
iliary cab
l
es
The m
easurem
ent
o
f
part
i
a
l
di
scha
rge
i
n
hi
gh
v
o
l
t
a
ge e
q
u
i
pm
ent
i
s
very
im
port
a
nt
bec
a
use
of
t
h
e
o
b
t
ain
e
d
d
a
ta
an
d th
eir in
terp
retation
can
be d
e
term
in
ed
a reliab
ility o
f
eq
u
i
p
m
en
t caused
b
y
ag
ing
an
d th
e
risk of failure can be
analyze
d
.
T
h
e partial
discharge
te
st
sp
eci
fi
cat
i
ons
de
pen
d
s
o
n
t
h
e t
y
pe
of t
e
st
e
qui
pm
ent
an
d
i
n
su
lation
m
a
terials u
s
ed
in
th
e co
nstru
c
tion
of eq
uip
m
en
t. Th
e ex
isten
ce of
p
a
rtial d
i
sch
a
rg
e in
th
e
i
n
sul
a
t
i
o
n m
a
teri
al
can
be
de
t
e
rm
i
n
ed by
t
h
e m
e
t
hod
o
f
a
n
al
y
s
i
s
of
t
h
e
m
a
gni
t
ude
o
f
t
h
e
part
i
a
l
di
sc
har
g
e.
Thi
s
anal
y
s
i
s
i
s
one
of t
h
em
can be
do
ne b
y
usi
ng C
M
2
0
00 a
n
al
y
zer. It
sho
u
l
d
be n
o
t
e
d t
h
at
t
h
i
s
an
al
y
s
i
s
shoul
d
be care
f
ul and always
be com
p
ared with the data stat
e at an earlier tim
e
. Th
e part
i
a
l
di
schar
g
e dat
a
can
b
e
go
od
i
n
fo
rm
atio
n
fo
r th
e con
d
ition
o
f
the g
e
n
e
rat
o
r stator wi
nd
in
g
and
thu
s
it can
b
e
m
a
d
e
of d
a
ta
trend
s
.
Tabl
e 2. Int
e
rp
ret
a
t
i
on of
Pa
rt
i
a
l
Di
schar
g
e [
16]
No
.
PD Me
asu
r
e
m
en
t r
e
su
lts
In
terp
retatio
n
Re
m
a
rk
s
1
40 m
V
Danger
,
action r
e
qui
r
e
d
Action should be t
a
ken
2
30
m
V
Detailed analy
s
is to be done
Analy
s
is trend,
PD polar
ity
, and
histor
y
to isolate cause of PD
3
20
m
V
PD detected
M
onitor
tim
e
tr
end
4
0
m
V
No pr
oblem
found
No action r
e
quir
e
d
3.
MEAS
U
R
EM
ENT RES
U
L
T
DAT
A
A
N
D
DIS
C
U
SSI
ON
The g
e
ne
rat
o
r
t
echni
cal
dat
a
were
3 p
h
ase
,
50
Hz f
r
e
que
n
c
y
,
11
,8
0
0
V
o
l
t
, 30
0
0
r
p
m
rot
a
t
i
on s
p
eed
,
68
,7
5
0
k
V
A capaci
t
y
, 0.
8 p
o
w
er
fact
o
r
l
a
gg
i
ng, B
i
n
su
latio
n
class and
mica tap
e
with
ep
ox
y resi
n
in
su
lating
mater
i
al.
Thi
s
m
easurem
ent
was per
f
o
r
m
e
d once
a
m
ont
h i
n
a
peri
o
d
o
f
1
2
m
ont
hs,
usi
n
g si
x R
T
D
(Resistance T
h
erm
a
l Detector) se
nsors a
s
the detection
tool. Th
e pu
rpo
s
e o
f
th
is m
easurem
en
t to
d
e
term
in
e
t
h
e m
a
gni
t
ude
of
t
h
e
pa
rt
i
a
l
d
i
schar
g
e t
h
at
o
ccurs
i
n
t
h
e
ge
nerat
o
r
st
at
or
wi
n
d
i
n
g i
n
sul
a
t
i
on.
The
eas
ur
em
ent
r
e
su
lts ar
e show
n in
Tab
l
e
3
.
Tabl
e
3. T
h
e
m
easurem
ent
o
f
part
i
a
l
di
sc
har
g
e
M
onth
RTD
A1 A2
B1
B2
C1
C2
1 25.
85
289.
50
48.
72
58.
49
79.
59
19.
59
2 38.
82
224.
18
28.
28
43.
72
58.
54
20.
32
3 29.
53
394.
51
10.
01
59.
96
58.
83
20.
13
4 40.
60
29.
44
0.
01
50.
01
5 45.
37
55.
19
41.
38
6 40.
29
79.
56
7 289.
63
300.
13
8 287.
67
389.
99
9 0.
01
43.
84
10
83.
28
99.
39
11
2.
04
4.
78
4.
51
15.
00
3.
27
7.
64
12
5.
40
10.
07
5.
98
43.
30
5.
96
49.
80
Evaluation Warning : The document was created with Spire.PDF for Python.
I
J
ECE
I
S
SN
:
208
8-8
7
0
8
Int
e
r
nal
O
n
-l
i
n
e Part
i
a
l
Di
sc
h
a
rg
e A
nal
ysi
s
of
6
8
.
7
5 MV
A
Gene
rat
o
r
St
at
or Wi
ndi
ng
I
n
s
u
l
a
t
i
o
n (
W
al
uy
o)
2
093
Th
e
p
a
rtial disch
a
rg
e vo
ltag
e
s tho
s
e
were m
eas
ured
from
the RTD se
nsor re
presente
d the
con
v
e
r
si
o
n
of t
h
e t
e
m
p
erat
ure
scal
e i
n
t
o
t
h
e
am
ount
of
v
o
l
t
a
ge. T
h
i
s
c
o
ul
d
be d
o
n
e
wi
t
h
t
h
e ai
d
o
f
a
n
a
l
y
zer
t
h
at
co
ul
d
hel
p
anal
y
ze t
h
e
p
a
rt
i
a
l
di
scha
rg
e t
h
at
occu
rs i
n
t
h
e
st
at
o
r
wi
ndi
ng
v
o
l
t
a
ge
gene
rat
o
r.
The
val
u
e
s
t
hose
occ
u
r
r
e
d
du
ri
n
g
t
h
e 12
m
ont
hs
fi
ckl
e
d
.
1.
In t
h
e
fi
rst
se
n
s
or
(A
1
)
, t
h
e
m
a
xim
u
m
vol
t
a
ge o
f
pa
rt
i
a
l
di
scha
rge
was
occu
rre
d i
n
t
h
e seve
nt
h m
ont
h
(2
8
9
.
6
3
m
V
) and
t
h
e m
i
nim
u
m
vol
t
a
ge o
f
p
a
rt
i
a
l
di
scha
rge
was
occ
u
r
r
e
d
at
m
ont
h
o
f
ni
n
e
(
0
.
0
1
m
V
).
2.
In t
h
e sec
o
nd
s
e
ns
or
(A
2
)
, t
h
e
m
a
xim
u
m
vol
tage
of
pa
rt
i
a
l
d
i
schar
g
e
was
o
ccur
r
ed
i
n
t
h
i
r
d m
ont
h
(3
9
4
.
5
1
m
V
) and
t
h
e m
i
nim
u
m
vol
t
a
g
e
o
f
part
i
a
l
di
s
c
har
g
e
was
occ
u
r
r
ed
i
n
t
h
e el
e
v
ent
h
m
ont
h
(
4
.7
8 m
V
).
3.
In t
h
e t
h
i
r
d sen
s
or
(B
1
)
, t
h
e m
a
xi
m
u
m
vol
t
a
ge of
part
i
a
l
di
s
c
har
g
e
was oc
cur
r
ed i
n
t
h
e
fi
rst
m
ont
h (4
8
.
7
2
m
V
) and
t
h
e m
i
nim
u
m
vol
t
a
g
e
o
f
part
i
a
l
di
s
c
har
g
e
was
occ
u
r
r
ed
i
n
t
h
e el
e
v
ent
m
ont
h
(
4
.
5
1
m
V
).
4.
In t
h
e f
o
urt
h
s
e
ns
or
(B
2
)
, t
h
e m
a
xim
u
m
v
o
l
t
a
ge
of
pa
rt
i
a
l
di
scha
rge
w
a
s occ
u
r
r
ed
i
n
t
h
e ei
g
h
t
h
m
ont
h
(3
8
9
.
9
9
m
V
) and
t
h
e m
i
nim
u
m
vol
t
a
ge o
f
p
a
rt
i
a
l
di
scha
rge
was
occ
u
r
r
e
d
i
n
t
h
e
f
o
urt
h
m
ont
hs
(0
.0
1 m
V
).
5.
In t
h
e
fi
ft
h
sen
s
or
(C
1),
t
h
e
m
a
xim
u
m
vol
t
a
ge
of
pa
rt
i
a
l
d
i
schar
g
e
was
o
ccur
r
ed
i
n
t
h
e
fi
rst
m
ont
h
(7
9
.
5
9
m
V
) and
t
h
e m
i
nim
u
m
vol
t
a
g
e
o
f
part
i
a
l
di
s
c
har
g
e
was
occ
u
r
r
ed
i
n
t
h
e el
e
v
ent
h
m
ont
h
(
3
.2
7 m
V
).
6.
In t
h
e si
xt
h se
nso
r
(C
2
)
, t
h
e
m
a
xim
u
m
vol
t
a
ge o
f
part
i
a
l
di
scha
rge
was
occu
rre
d at
t
w
el
ft
h m
ont
h (
4
9
.
8
0
m
V
) and
t
h
e m
i
nim
u
m
vol
t
a
g
e
o
f
part
i
a
l
di
s
c
har
g
e
was
occ
u
r
r
ed
i
n
t
h
e el
e
v
ent
h
m
ont
h
(
7
.6
4 m
V
).
The
t
r
en
d dat
a
o
f
part
i
a
l
di
schar
g
e
m
easure
m
ent
res
u
l
t
s
i
n
t
h
e ge
nerat
o
r
st
at
or wi
n
d
i
n
g appea
r
s
t
h
at
m
o
st o
f
th
e
p
a
rtial d
i
sch
a
rg
e activ
ity ten
t
to
in
crease, esp
eci
ally in
th
e m
o
nth
of sev
e
n
t
h
an
d eigh
th
.
Based on the
direct m
easurem
ents, it was classifi
ed into
three c
o
nditions based
on the am
ount
of
char
ges
per cy
cl
e. Fi
gu
re
13
sho
w
s t
h
e pa
rt
i
a
l
di
scha
rge
oc
curs t
h
o
s
e cau
s
e
d by
i
n
t
e
rnal
voi
d i
n
t
h
e ge
n
e
rat
o
r
stator ins
u
lation. These cases were in
dicated
by the am
ount
of cha
r
ge occ
u
pi
ed on the pos
itive parts of c
y
cles
t
hose
ha
ve al
m
o
st
sam
e
am
ount
o
f
t
hose
o
n
t
h
e
negat
i
v
e
on
es.
A
1
– 24
-04
;
1
3
.39
A
1
– 26
-05
;
1
0
.43
A
1
– 23
-06
;
1
3
.43
A
1
– 09
-10
;
1
3
.15
A
1
– 07
-12
;
1
3
.54
A
1
– 30
-12
;
1
3
.33
A
2
– 29
-01
;
1
0
.40
A
2
– 24
-02
;
1
0
.55
B
1
– 24
-0
2;
11
.0
3
B
1
– 30
-1
2;
14
.0
5
B
2
– 29
-0
1;
10
.5
6
B
2
– 23
-0
6;
13
.5
8
Evaluation Warning : The document was created with Spire.PDF for Python.
I
S
SN
:
2
088
-87
08
I
J
ECE
Vo
l. 6
,
N
o
. 5
,
O
c
tob
e
r
20
16
:
208
8
–
20
95
2
094
B
2
– 06
-0
8;
11
.3
7
B
2
– 28
-0
8;
14
.0
5
C
1
– 24
-0
2;
11
.2
6
C
1
– 24
-0
4;
14
.4
5
C
1
– 26
-0
5;
22
.5
2
C
2
– 29
-0
1;
11
.0
8
Fig
u
re
13
. Th
e m
easu
r
e
m
en
t resu
lts t
h
o
s
e the po
sitiv
e an
d
n
e
g
a
tiv
e cycles h
a
v
e
alm
o
st same ch
arg
e
s
Th
e seco
nd
con
d
ition
,
as shown
i
n
Figu
re
1
4
, ind
i
cated
t
h
e in
tern
al d
e
l
a
m
i
n
a
tio
n
.
In
t
h
e in
tern
al
del
a
m
i
nat
i
on,
t
h
e i
n
sul
a
t
i
o
n
wo
ul
d
be
se
pa
rat
e
d
fr
om
t
h
e co
ppe
r c
o
nd
u
c
t
o
r;
t
h
e
r
ef
o
r
e
i
t
w
oul
d
be
weakl
y
bo
n
d
ed
,
ove
r
h
eat
i
ng, a
n
d s
o
fo
rt
h.
T
h
i
s
cas
e co
ul
d
be
see
n
from
the
m
o
st charge
was
on the
ne
gative cycles
was m
o
re
dom
inant tha
n
on t
h
e
negative cy
cle one
s.
A
1
– 29
-01
;
1
0
.31
A
1
– 07
-12
;
1
3
.54
A
2
– 30
-12
;
1
3
.59
B
1
– 07
-1
2;
14
.2
7
B
2
– 25
-0
3;
11
.1
4
Fi
gu
re
1
4
. T
h
e
m
easurem
ent
resul
t
s
t
hose
t
h
e cha
r
ge
o
n
t
h
e
ne
gat
i
v
e cy
cl
e
s
we
re m
o
re
do
m
i
nant
t
h
a
n
t
h
ose
on
t
h
e
ne
gat
i
v
e
o
n
es
The third condition of m
easure
m
ent indicated th
e surface discharges. T
h
is phe
nom
enon was caus
e
d
b
y
sem
i
co
n
ductiv
e a m
a
teria
l
(p
ai
n
t
)
was
da
m
a
g
e
and
l
o
ose its cap
ab
ility, as sh
own
Fi
g
u
re
15
.
A
1
– 24
-02
;
0
9
.42
A
2
– 07
-12
;
1
4
.01
B
1
– 24
-0
4;
14
.3
1
Evaluation Warning : The document was created with Spire.PDF for Python.
I
J
ECE
I
S
SN
:
208
8-8
7
0
8
Int
e
r
nal
O
n
-l
i
n
e Part
i
a
l
Di
sc
h
a
rg
e A
nal
ysi
s
of
6
8
.
7
5 MV
A
Gene
rat
o
r
St
at
or Wi
ndi
ng
I
n
s
u
l
a
t
i
o
n (
W
al
uy
o)
2
095
C
2
– 07
-1
2;
14
.5
6
Fi
gu
re
1
5
. T
h
i
s
p
h
en
om
enon
was ca
use
d
by
sem
i
con
duct
i
v
e a m
a
t
e
ri
al
(pai
nt
)
4.
CO
NCL
USI
O
N
B
a
sed o
n
t
h
e
m
easurem
ent
for
12 m
ont
hs,
t
h
e pa
rt
i
a
l
di
schar
g
e
occu
rs i
n
t
h
e
gene
rat
o
r u
n
i
t
2 wa
s
al
m
o
st cau
sed
b
y
th
e in
tern
al
vo
id
in th
e
g
e
n
e
rat
o
r i
n
sulati
on. T
h
is case
was indicated
by the c
h
arge
on the
p
o
s
itiv
e cycles alm
o
st sa
m
e
with
o
n
th
e n
e
g
a
tiv
e
on
ce.
Th
e g
r
eatest
en
erg
y
o
f
p
a
rtial
disch
a
rg
e was occu
rred
i
n
t
h
i
r
d m
ont
h
,
an
d f
o
l
l
o
we
d
by
o
n
e i
n
ei
g
h
t
h
m
ont
hs
. T
h
e
s
e cases i
n
di
ca
t
e
d t
h
e
di
sc
har
g
e
occu
rs,
t
h
e
r
efo
r
e
th
e cond
itio
n of th
e stato
r
wi
nd
ing
insu
lation h
a
s ch
ang
e
d fro
m
th
e n
o
rm
al
co
nd
itio
n.
ACKNOWLE
DGE
M
ENTS
We
wou
l
d
lik
e to
exp
r
ess th
e
d
eep
est appreciatio
n
to Indonesia Powe
r,
The
Ge
ot
he
rm
al
Po
wer
Pl
ant
Un
it o
f
Kam
o
jan
g
, wh
ich
h
a
s
sup
ported
t
o
prov
id
e
th
e d
a
ta.
REFERE
NC
ES
[1]
H.
Shin,
et al.
, “Denoising of On-line Par
tial
Discharge Signal from High-Volta
ge Ro
tatin
g
Machines using
Standard Deviation Threshold,”
International Journal of In
novatives
Computing, Information and Contro
l
,
vol/issue:7(A), p
p
. 3761-3769
, 2
011. ISSN 1349-4198.
[2]
M
.
Kurtz and
G. C. S
t
on
e, “
I
n-
S
e
rvice P
a
r
t
i
a
l
Discharge
Testing of Generator
Insulation
,
”
I
E
EE Transaction
on
Ele
c
trica
l
Insula
tion
, vol/issue:
EI-14(2), 1979
.
[3]
G. C. Stone, “
P
artial Disch
a
rg
e Diagnosis an
d Elec
tri
cal
Eq
uipm
ent Insulati
on Condition Assessm
e
nt,”
IEEE
Transaction on
Diele
c
trics and
Ele
c
trica
l
Insula
tion
, vol/issue: 1
2
(5), pp
. 891-90
4, 2005
.
[4]
V. W
a
rren, “
P
artial Disch
a
rge
Testing: A Progress Report,”
St
atistica
l
Eva
l
uat
ion of PD Data, Irish Rotating
Machine
Confer
ence
, San
Anton
i
o,
Tx., USA, Ju
ne, 2007
.
[5]
E. L
e
m
k
e,
et al.
,
“
G
uide for Elec
tric
al Part
ial Dis
c
harge Me
asure
m
ents in Com
p
liance
to IEC 602
70,”
El
e
c
t
r
o
, vol
.
241, Techn
i
cal
Brochure 366
,
WG D1.33, pp.
61-67, 2008
.
[6]
P
D
Tech P
o
wer Engine
ering A
G
, “
P
artial Dis
c
harge Bas
i
cs
of M
onitoring,
” M
i
cam
axx
TM
Plus-Partial Disch
a
rg
e
Basics, pp
. 1-18
.
[7]
H. G. Sedding
,
et al.
, “A New On-Line Partial
Discharg
e
Test f
o
r Turbin
e Generators,”
Cigre, 1
992 Session,
#
0
August – 5
Sep
t
ember
, pp. 1-7, 1
992.
[8]
G. C. Stone and
V. Warren, “Ef
f
ect
of
Manufacturer Winding A
g
e and Insu
latio
n Ty
pe on Stato
r
Winding Partial
Discharge Lev
e
ls,”
From DEIS
,
Sept-Oct 2004
, v
o
l/issue: 2
0
(5), p
p
. 13-17
, 2004
.
[9]
J. K. Nelson
and
J. Stein
,
“A Field
Assessment of PD and EMI M
e
thodolog
y
A
pplied to
Large Util
ity
Gen
e
rators
,”
IEEE Transacti
ons on Dielec
tri
c
s and Ele
c
trica
l
Insulation
,
vol/issue:
17(5),
pp
. 1411–1427, 20
10. ISSN:
1070-
9878.
[10]
V. W
a
rren, “
S
tator W
i
nding Partial Disch
a
rge
Activity
for Marin
e
Ai
r-Cooled Generators,”
FMM
S
2013 of ASNE
Confer
enc
e
, August, 2013.
[11]
A.
K.
Gupta,
et al.
, “Modeling o
f
Calibration C
i
r
c
uit
fo
r P
a
rtia
l Dis
c
harge M
eas
urem
ent,”
The
s
is
, Department of
Electri
cal
Eng
i
n
eering
,
Nationa
l
Institute of
Tech
nolog
y
,
Rourk
e
l
a
, Odisha, p
p
. 13
-42, 2013
.
[12]
L. Renfor
th,
et al.
, “Continuous
, Remote On-line Partial Disch
a
rg
e (OLPD) Monitoring of Co
mplete Medium
Voltage (MV) Network,” pp. 1-1
1
.
[13]
M.
M.
Yaa
c
ob,
et al.
, “Review on Partial D
i
scharg
e Detection Techni
ques
Relat
e
d
to Hig
h
Voltag
e
Power
Equipment using
Differ
e
nt Senso
r
s,”
Pho
t
onic S
e
nsors
, vol/issue:
4(4), pp
. 325-33
7, 2014
.
[14]
U.
Khay
am,
et al.
, “
P
arti
al
Dis
c
harge
and
Cros
s
Interfere
nce P
h
enom
ena
i a
Thr
ee-P
h
a
s
e Cons
tructio
n
,
”
International Jo
urnal on Electric
al Engineering and
Informatics
, vol/issue: 1(1)
,
pp. 85
, 2009
.
[15]
B
.
A.
L
l
oy
d,
et al.
, “Continuo
us On-line Partial Discharg
e M
onitoring of Generator Stator
Windings,”
IEEE
Transaction on
Energy Con
version
, vol/issue: 14
(4), pp
. 1131-11
38. ISSN: 0885-
8969.
[16]
Toshiba, “Instruction
Manual; C
M
2000 Specifications,”
Toshiba International
Co
rporations, Pty
.
,
Ltd.
Evaluation Warning : The document was created with Spire.PDF for Python.