Int
ern
at
i
onal
Journ
al of
P
ower E
le
ctr
on
i
cs a
n
d
Drive
S
ystem
(I
J
PE
D
S
)
Vo
l.
11
,
No.
3
,
Septem
be
r
2020
, pp.
1241
~
1248
IS
S
N:
20
88
-
8694
,
DOI: 10
.11
591/
ij
peds
.
v
1
1
.i
3
.
pp
1241
-
1248
1241
Journ
al h
om
e
page
:
http:
//
ij
pe
ds
.i
aescore.c
om
Phase t
o phase fa
ult detec
tion of
3
-
phase in
ductio
n motor
using
FRA te
chniq
ue
A.
A. Al
awady
1
, M
.
F. M.
Yous
of
2
,
N
.
Azis
3
,
M
. A. T
alib
4
1,
2
Facul
ty
of Electrical a
nd
Elec
t
ronic
Engi
ne
eri
n
g,
Univer
si
ti T
u
n
Hus
sein
Onn
Mala
y
sia
,
Ma
la
y
sia
1
Coll
ege of
Tec
hnic
a
l
Eng
ineeri
ng,
Th
e
Isl
am
i
c Universit
y,
Naj
a
f,
Ir
aq
3
Facul
ty
of Engi
nee
ring
,
Univ
ersit
i
Putr
a
Ma
la
ysi
a,
Ma
la
ysi
a
4
Te
nag
a
Nasion
al
B
erh
ad
Research,
K
awa
san
In
stit
usi Pe
ny
el
idikan,
Ma
la
ysi
a
Art
ic
le
In
f
o
ABSTR
A
CT
Art
ic
le
history:
Re
cei
ved
Dec
2
7
, 201
9
Re
vised
Feb
1
2
, 2
0
20
Accepte
d
M
ar
5
, 2
0
20
The
purpose
for prepa
ring
thi
s p
a
per
is t
o
innov
ate
a
new
m
et
hod
to
detec
ti
on
and
di
agnosis
th
e
thr
ee
-
ph
ase
in
duct
ion
mot
or
e
le
c
tri
c
al
failure
s
,
espe
ciall
y
the
f
ai
lur
es
th
at
occ
ur
in
St
at
or
winding.
St
at
or
winding
bre
akd
own
la
rge
ly
cont
ributes
to
i
nduct
ion
mo
tor
fa
il
ure
s
.
To
u
nder
stand
intern
al
fau
l
t
in
induc
ti
on
mo
to
rs
winding,
fo
ur
c
ase
s
studies
of
diff
ere
n
t
thre
e
-
ph
ase
induc
ti
on
mo
tor
s
(TPIM)
were
a
nal
ysed
accordi
n
g
to
two
condi
t
i
ons:
normal
winding
condi
t
i
on
and
windings
shorted
be
tween
two
phase
s.
In
t
his
pap
er,
the
m
ea
sure
me
nt
of
fre
quen
cy
r
esponse
ana
lysis
(
FR
A)
on
the
stator
windin
g
with
int
er
-
ph
as
e
short
is
pr
ese
nte
d
.
Additi
on
al
ly
,
Freque
ncy
Response
Analysis
(FRA
)
int
erp
r
et
a
ti
on
t
echnique
class
ify
a
nd
quan
ti
fy
th
e
f
aul
t
is
a
lso
proposed.
For
i
nte
rp
re
ta
t
ion
of
th
e
FR
A,
a
st
at
isti
cal
indi
c
ator,
whi
ch
is
NCEPR
I
al
gori
t
hm
is use
d
for c
ompa
rison
the m
ea
sured respons
e
s.
Ke
yw
or
d
s
:
Fr
e
qu
e
nc
y
re
spon
s
e a
nalysis
Ind
uction m
otor
NCEPRI
alg
or
i
thm
Ph
ase
to p
hase
fau
lt
Stat
or
windin
g fa
ult
This
is an
open
acc
ess arti
cl
e
un
der
the
CC
BY
-
SA
l
ic
ense
.
Corres
pond
in
g
Aut
h
or
:
A.
A. Ala
wady
,
Dep
a
rtme
nt of
Ele
ct
rical
an
d
Ele
ct
ro
nic
Eng
ineerin
g,
Un
i
ver
sit
i
T
un
Hu
s
sei
n O
nn
M
al
aysia
(
UT
HM),
86400, Pa
rit R
aj
a, Bat
u
Pa
hat
, J
oh
or,
M
al
at
sia
.
Emai
l:
ah
me
d.a.al
awady
@gmai
l.com
1.
INTROD
U
CTION
Healt
h
of
a
ny
inducti
on
m
ot
or
is
c
riti
cal
in
vie
w
of
a
vaila
bili
ty
asset
m
anag
e
ment
an
d
cost.
E
very
machine
ha
s
it
s
own
est
imat
ed
li
fe.
This
e
sti
mate
is
mad
e
by
c
on
si
der
i
ng
va
rio
us
fac
tors
s
uc
h
as
m
at
erial
us
e
d,
operati
ng
c
onditi
on,
e
nv
i
ronme
ntal
conditi
ons
et
c.
I
nductio
n
mot
or
s
fail
ure
me
chan
is
m
a
re
diff
e
re
nt
dep
e
nd
on d
i
fferent
par
t,
the
mo
st i
m
porta
nt p
a
rt is stat
or
wind
in
g [1].
Accor
ding
to
study
t
hat
pres
ented
by
Ele
ct
ric
P
ow
e
r
Re
s
earch
I
ns
ti
tute
(EPR
I),
s
how
that
(
36%)
of
mo
t
or
fail
ure
is
beca
us
e
of
sta
t
or
windi
ng
fa
ults
[
2,
3].
M
ot
or
sta
tor
windin
gs
fail
ur
e
or
brea
kdown
is
the
m
os
t
c
ommo
n
reas
on
s
of
TP
I
M
fail
ur
e
[
4]
.
Ca
ses
of
i
nductio
n
mo
t
or
fail
ur
e
am
ount
re
porte
d
bec
ause
of
this cat
eg
ory r
ang
e
d nea
rly fr
om
30% t
o 40% [
5].
Seve
ral
reas
ons
r
esult
to
armatu
re
or
sta
tor
windin
gs
in
su
la
ti
on
fail
ur
e
,
am
ong
these
a
re
[6]:
(a)
S
hort
ci
rcu
it
or
sta
rting
stre
sses,
(b)
Lo
os
e
braci
ng
f
or
en
d
windin
g,
(c
)
High
sta
tor
windin
g
te
mp
erat
ur
es
,
(
d)
Leaka
ge
i
n
coo
li
ng
sy
ste
ms,
(e)
Ele
ct
rical
discha
rges.
From
ot
her
side,
i
nsula
ti
on
fail
ure
al
way
s
relat
ed
TPIM
wi
nd
i
ng
s
fail
ure
a
nd
it
is
co
untt
ing
the
mai
n
reas
on
of
windin
g
fail
ur
e
[5].
As
evi
denced
by
c
red
i
ble
surve
y
[
7
-
10]
,
sta
tor
wi
ndin
g
br
ea
kdow
n
is
a
la
rg
e
gr
oup
of
in
duct
io
n
mo
to
r
fail
ur
es
wh
ic
h
undetect
ed
fa
ul
t
is
co
ns
ide
re
d
a
s
t
he
root
c
ause
f
or
this
c
at
egory
of
the
fau
lt
s.
A
ny
fai
lure
in
windin
gs
sho
ul
d
be
detect
ed
and
diag
nosis
in
inci
pient
le
vel
to
a
vo
i
d
a
ny
cat
ast
r
ophi
c
fail
ure
it
ma
y
ca
us
e
unwa
nted d
owntime o
f pro
duct
ion
.
Evaluation Warning : The document was created with Spire.PDF for Python.
IS
S
N
:
2088
-
8
694
In
t J
P
ow
Ele
c
&
Dr
i
S
ys
t
,
V
ol
.
1
1
, N
o.
3
,
Se
ptembe
r
2020
:
12
41
–
12
48
1242
To
un
der
sta
nd
the
ef
fect
of
sh
ort
-
ci
rc
uit
fa
ult
bet
ween
w
ind
in
g
phases
in
th
ree
-
phase
in
du
ct
io
n
mo
to
r
(
TPIM
),
this
pap
e
r
us
e
d
a
ne
w
met
hod
w
hich
is
f
re
quenc
y
r
esp
ons
e
analysis
(
FR
A)
te
ch
ni
qu
e
.
This
is
achieve
d
by
a
nalysi
ng
fou
r
inducti
on
m
oto
rs
with
eac
h
has
a
diff
e
re
nt
powe
r
rati
ng
an
d
tw
o
dif
fer
e
nt
conditi
ons:
(
no
rmal
windin
g
conditi
ons
a
nd
short
-
ci
rc
uit
be
tween
phases
.
The
f
reque
nc
y
respo
ns
es
(FR
)
of
norma
l
a
nd
fa
ulty
wi
nd
i
ng
s
will
be
analys
ed
by
usi
ng
th
e
North
C
hin
a
Ele
ct
ric
Po
we
r
Re
searc
h
I
nst
it
ute
(N
CEPR
I)
al
gorith
m.
The
im
portance
of
t
hi
s
stu
dy
is
main
ly
to
val
uate
th
e
abili
ty
of
FR
A
t
o
detect
the
fa
ulty
windin
g
in
TP
I
M
.
2.
LIT
ERATUR
E OR
BIB
LIOGR
AP
HY S
URVEY
S
hort
ci
rc
uit
(
SC)
windin
g
f
aults
mig
ht
de
velo
p
to
m
or
e
com
plica
te
d
ty
pes
of
sta
tor
w
ind
in
g
fa
ults
[11]
.
T
ypic
al
ly,
SC
sta
t
or
windin
gs
occ
ur
:
(a)
betwe
en
i
nter
nal
tu
rn
s
in
same
ph
a
se
(turn
-
to
-
tur
n),
(b)
betw
een
t
wo
c
oils
in
sa
me
phase
(c
oil
-
to
-
c
oil),
(c)
be
tween
two
tu
rn
s
in
dif
f
re
nt
phases
,
(
d)
be
tween
ph
a
se
to
ne
utra
l
po
i
nt,
(e
)
between
phase
t
o
gro
und
a
nd
(f)
op
e
n
ci
tc
uit
c
oi
l
[1
2]
.
Fig
ur
e
1
s
how
the
co
mmon
fau
lt
s in
in
duct
ion
mo
t
or
s
.
Figure
1
.
I
nduc
ti
on
m
otors
c
ommo
n
fa
ults a
nd Fail
ures
[12
].
The
t
yp
e
s
of
windin
g
fau
lt
s
that
was
cl
ari
f
ie
d
in
the
prev
iou
s
pa
ra
gr
a
ph
is
na
me
d
by
a
“st
at
or
tu
rn
fau
lt
s”
[
13].
T
he
res
pons
i
ble
for
sta
tor
windin
gs
fail
ures
is
the
stresses,
wh
ic
h
are
ma
y
be
cl
assi
fied
into
el
ect
rical
,
ther
ma
l,
mec
ha
nic
al
,
an
d
en
vir
on
ment.
T
he
factor
that
most
in
flue
nces
on
for
the
dete
rior
at
i
on
of
the
ins
ulati
on
in
m
os
t
of
the
cases
a
mon
g
them
is
the
t
he
rmal
stre
ss
[14].
A
s
hort
-
ci
r
cuit
(S
C
)
t
urns
fa
ult
causes
a
flo
w
of
la
rg
e
ci
rcu
la
ti
ng
c
urren
t.
T
hu
s
,
this
c
urre
nt
ge
nerat
es
ex
cessi
ve
the
rma
l
stress
in
t
he
point
of
sh
ort
ed
w
i
nd
i
ngs t
hu
s
, th
at
m
ay
ca
us
e
for m
achin fai
lu
re [1
5].
In
[
16
],
Stat
ist
ic
al
process
c
on
t
ro
l
te
c
hn
i
que
(S
PCT
)
wa
s
us
e
d
t
o
diag
no
sis
an
d
dete
ct
windin
g
s
fau
lt
s
for
TPIM.
A
ne
w
m
odel
was
est
im
a
te
d
to
detect
s
hort
-
ci
rc
uit
(
S
C)
wi
nd
i
ngs
f
aul
ts
in
ti
me
domain
.
An
aut
hor
in
a
rtic
le
[
17
]
de
ve
lop
e
d
a
fu
zz
y
fau
lt
detect
or
(F
F
D)
by
us
i
ng
Co
ncor
dia
patte
rn
s
to
dia
gnos
is
op
e
n
-
ci
rcu
it
(
OC)
fa
ults
an
d
sta
tor
unbala
nc
e.
A
rtic
le
[
18
],
Discu
ssed
th
e
pe
rforma
nce
of
the
TP
IM
and
it
s
vu
l
ner
a
bili
ty by
m
odel
li
ng fo
r dete
ct
ion
of i
nductio
n
m
ot
or
windin
g fa
ult and it
s locati
on
.
Re
fer
e
nces
[
19
]
a
nd
[
20
]
re
ported
that
the
y
us
e
d
ad
va
nc
ed
sig
nal
-
pr
oc
essing
te
ch
niques
(
AS
PT
)
and
sta
to
r
-
windin
gs
-
c
urre
nts
f
or
sta
to
r
wind
i
ng
fa
ult
de
te
ct
ion
of
the
IM
.
T
he
a
ut
hor
s
in
[
21
]
a
nalyse
d
the
de
viati
on
in
fr
e
quenc
y
r
esp
on
se
(F
R)
by
us
in
g
arti
fi
ci
al
-
neural
-
network
(AN
N)
for
fa
ults
detect
ion.
The
el
ect
rical
fau
lt
f
or
in
duct
ion
m
oto
r
s
fiel
d
-
windin
gs
was
detect
io
n
f
or
fir
st
ap
proac
h
th
rou
gh
FR
A
te
chn
iq
ue wit
h st
at
ic
ex
ci
ta
ti
on
was rep
ort
ed i
n
[
22
]
.
Fr
e
qu
e
nc
y
Re
s
pons
e
An
al
ys
i
s
FRA
is
an
e
le
ct
rical
no
n
-
i
ntr
us
ive
a
nd
powe
rf
ul
use
d
t
echn
i
qu
e
t
o
diag
nosis
of
m
any
c
ha
ng
es
,
f
aults,
or
ab
nor
mal
conditi
ons
in
the
trans
f
ormer
wi
nd
i
ng
[
23
].
FR
A
res
ponse
is
off
-
li
ne
te
c
hn
i
qu
e
for
fail
ur
e
diag
nosis an
d
f
aults detec
ti
on
[
24
]
. F
RA is c
on
si
der
e
d
as
one of the c
omp
arati
ve
te
sts
wh
ic
h
are
require
d
a
refe
ren
ce
data
(R
D)
(at
healt
hy
op
e
rati
on
co
nd
it
ion
)
to
co
mpa
re
with
it
.
Th
eref
or
e
,
the
dif
fer
e
nces
betwee
n
RD
and
t
he
ne
w
meas
ured
res
ponse
s
re
fer
to
there
fa
ult
[
25
,
26
].
Eac
h
f
au
lt
hav
e
fin
ger
-
pri
nt
or
cha
racteri
sti
c
fr
e
qu
e
nc
y
res
ponse
.
By
ob
se
rv
at
io
n
of
dif
f
eren
t
res
pons
e
s
of
data
mea
su
re
d,
mu
lt
iple
t
yp
es
of
fa
ult
a
re
det
ect
ed
a
nd
al
s
o
fau
lt
se
ve
rity
i
s
ide
ntifie
d
[
27
]
.
T
he
FR
A
r
esp
on
se
to
ol
c
an
get
h
ol
d
of
t
he
volt
age
tra
nsfer
functi
on
for
windin
gs
fr
e
qu
e
nc
y
betwe
en
(
20
Hz
to
2
MHz
)
[
28,
29]
.
The
imp
ort
ant
measu
reme
nts
connecti
ons
be
tween
the
FR
A
i
ns
tr
um
e
nt
dev
ic
e
a
nd
othe
r
e
quipme
nt
s
uch
a
s
Evaluation Warning : The document was created with Spire.PDF for Python.
In
t J
P
ow Elec
& Dri S
ys
t
IS
S
N:
20
88
-
8
694
Ph
as
e to
ph
as
e
fau
lt
detect
ion of
3
-
phas
e
in
duct
ion m
otor
usi
ng FRA tec
hniq
ue
(
A.
A.
Alawa
dy
)
1243
trans
forme
r
a
r
e
f
our
fa
ults
cases
wh
ic
h
a
re
e
nd
-
to
-
e
nd
sh
ort
ci
rcu
it
c
onnecti
on,
int
er
-
windin
g
in
duct
ive
connecti
on, i
nt
er
-
windin
g
ca
pa
ci
ti
ve
co
nn
ect
ion
a
nd
op
e
n
ci
rcu
it
c
onnecti
on
[
26
].
3.
METHO
DOL
OGY FO
R
F
AU
LT
A
ND
MEASU
REM
ENT
3.1.
Ex
peri
ment
al un
it
s
ch
o
ose
Four
unit
s
of
three
-
phase
i
nduction
m
otors
wer
e
c
hose
n
to
c
onduct
this
study,
these
unit
s
ha
ve
a
diff
e
re
nt
rati
ng
powe
r
w
hich
are
1,
2,
3
a
nd
5.5
horse
po
w
er
re
sp
ect
ivel
y.
T
he
s
pecifica
ti
on
s
for
t
hese
te
ste
d
un
it
s
a
re
sho
w
n
i
n
Ta
ble
1.
T
his
pap
e
r
f
ocus
on
t
he
W
ye
(
Y)
co
nnect
io
n
config
ur
at
io
n
betwee
n
phase
s
U
,
V
and
W
wh
ic
h
is
widely
us
e
d
for
TP
IM.
T
he
rfor
e
,
phases
for
al
l
exp
e
ri
mental
unit
s
of
TPIM
windi
ng
s
a
re
connecte
d as
Wy
e
(Y)
w
i
nd
i
ng con
fig
ur
at
io
n.
Table
1.
S
peci
f
ic
at
ion
s
of
t
hr
e
e
phase i
nduction m
o
tor
un
it
s
Moto
r
Moto
r
1
Specific
at
io
n
s
Moto
r
2
Specific
at
io
n
Moto
r
3
Specific
at
io
n
s
Moto
r
4
Specific
at
io
n
Manu
facturer
JILAN
G
JILAN
G
JILAN
G
JILAN
G
Mod
el
1
1
0
RK
-
3
DS
Y9
0
S
-
2
Y9
0
L
-
2
Y1
1
2
M
-
2
Ph
ases
3
-
Ph
ind
u
ctio
n
m
o
to
r
3
-
Ph
ind
u
ctio
n
m
o
to
r
3
-
Ph
in
d
u
ctio
n
m
o
to
r
3
-
Ph
ind
u
ctio
n
m
o
to
r
Po
wer
0
.75
KW / 1HP
1
.5 KW
/
2
HP
2
.2 KW
/
3
HP
4
KW / 5.5
HP
Rated
Voltag
e
4
1
5
V /
5
0
Hz
4
1
5
V /
5
0
Hz
4
1
5
V /
5
0
Hz
4
1
5
V /
5
0
Hz
RPM
1
5
0
0
r
p
m
2
8
4
0
r
p
m
2
8
4
0
r
p
m
2
8
8
0
r
p
m
3.2
.
Cre
ati
on
of
f
au
lt
The
a
rtific
ia
l
fau
lt
s
in
windin
gs
wh
ic
h
are
pe
rformed
e
xpe
rimental
ly
i
n
la
borato
ry,
will
us
e
t
o
stu
dy
the
diff
e
re
nces
in
wi
nd
i
ng
s
r
esp
on
se
that
may
be
occur
du
e
to
this
fa
ul
t.
I
n
this
stu
dy,
the
arti
fici
al
fa
ult
create
d
bet
wee
n
t
wo
phases
of
th
ree
-
phase
i
nduct
io
n
mo
t
or. P
has
e
-
to
-
ph
a
se
(PP)
fa
ult o
ccur
w
he
n
si
ngle
-
tur
n
in
phase
U
s
ho
rted
with
si
ngle
-
tur
n
in
phase
V
or
betwe
en
two
c
oils
from
ph
ase
s.
T
his
pa
per
fo
c
us
on
ph
a
se
-
to
-
ph
ase
tu
rns
(PPT)
fa
ults,
f
our
unit
s
of
TP
IM
wer
e
us
e
d
to
cre
at
e
sho
rt
ci
rc
uit
fau
lt
b
et
wee
n
ph
a
ses
windin
gs
to
ge
t
PPT
fa
ult.
I
n
this
stu
dy,
the
sh
ort
ci
rc
uit
fa
ult
was
a
ppli
ed
in
the
la
borato
ry
betwee
n
ph
ase
U
and phase
V w
ind
in
gs o
nly.
T
he
P
PT
fau
lt
s
proce
dures
for o
ne
cas
e
of T
PIM a
re sh
own
i
n
Fi
gure
2
.
(a)
(b)
(c)
Figure
2
.
PPT
f
aults
c
reati
on
sta
ges
on
TP
I
M
, (a
)
N
ormal
windin
g,
(b)
Re
m
ovin
g some
con
du
ct
ors to
create t
he fault
, (c)
Co
mpl
et
ed
windi
ng
with P
PT
fau
lt
.
3.
3
.
F
RA
m
ea
suremen
ts
se
t
up
Ph
ase
-
to
-
phase
tu
rn
s
(P
PT
)
fa
ult
co
mm
on
l
y
occurs
to
TP
I
M
.
Ge
ne
rall
y,
at
an
y
short
-
ci
rcu
it
(S
C)
in
the
windin
gs
s
uffer
from
ins
ulati
on
fail
ur
e
because
of
t
he
thermal
stress
wh
ic
h
occ
urs
du
e
t
o
t
he
incr
ease
of
current
in
the
windin
gs
.
In
th
is
pa
pe
r,
FRA
meas
ureme
nts
are
pe
rformed
on
phases
U
,
V
a
nd
W
t
o
di
agnos
e
the
fa
ult
in
TP
IM
windin
gs.
The
mea
surem
ents
are
pe
rformed
befo
re
an
d
afte
r
the
fau
l
t
has
create
d.
F
igure
3
sh
ows
co
nnect
ion
betwee
n
t
he
eq
uip
me
nts
wh
ic
h
a
re
us
ed
to
perform
the
FRA
meas
ur
e
ments.
I
n
t
his
study,
the
wi
nd
i
ngs
ha
ve
wy
e
(
Y)
c
onnecti
on
c
onf
igurat
ion.
With
wy
e
co
nnect
ion
,
the
fr
e
que
ncy
res
ponse
s
can
be
measu
red
betw
een
t
wo
phase
te
rmin
al
s
(U
-
V
,
V
-
W,
an
d
W
-
U)
only
,
with
a
ssu
mi
ng
that
t
he
ne
utral
po
i
nt
(N)
Evaluation Warning : The document was created with Spire.PDF for Python.
IS
S
N
:
2088
-
8
694
In
t J
P
ow
Ele
c
&
Dr
i
S
ys
t
,
V
ol
.
1
1
, N
o.
3
,
Se
ptembe
r
2020
:
12
41
–
12
48
1244
is
hi
dd
e
n.
T
he
FRA
meas
ur
e
ments
are
m
a
de
on
f
our
unit
s
of
TP
IM
wh
ic
h
a
re
li
ste
d
i
n
Table
1.
T
he
windin
g
fr
e
qu
e
nc
y respon
s
es a
re lat
er
analyze
d
a
nd i
nter
pr
et
ed
.
(a)
(b)
(c)
Figure
3
.
Eq
ui
pm
e
nts
c
onnec
ti
on
to
meas
ur
ed
FR
A
for
mot
or
windin
gs
(a
)
Co
mputer
, (b
)
FR
AN
E
O 8
00
equ
i
pm
e
nt, (c)
Thr
ee
-
ph
ase
in
du
ct
io
n mot
ot
4.
FRA RE
SU
L
TS A
ND AN
A
LYSIS
In
t
his
sect
io
n,
the
FRA
res
pons
e
res
ults
for
the
PPT
f
ault
create
d
pr
act
ic
al
ly
in
th
e
la
borato
ry
on
T
PIM
sta
tor
windin
gs
a
re
presente
d
a
nd
discu
ssed
.
Fo
r
eac
h
case
,
three
ph
a
ses
fr
e
qu
e
nc
y
res
pons
e
measu
red
via
t
wo
ste
ps
,
the
f
irst
one
duri
ng
healt
hy
wi
nd
i
ng
s
co
ndit
ion
and
t
he
sec
ond
at
PPT
fa
ult
in
the
TPIM
windin
gs.
Fig
ur
e
4
,
Fi
gure
5
an
d
Fi
gure
6
s
hows
t
he
fr
e
qu
e
nc
y
re
sp
onses
(F
R
)
curves
of
t
he
t
est
ed
fou
r
unit
s.
As
sh
ow
n,
the
FR
curve
wa
s
dif
fer
e
nt
from
m
achine
to
oth
e
r,
this
is
li
kely
dep
e
ndin
g
on
four
TPIM
un
it
s
po
wer rati
ng.
(a)
(b)
(c)
(d)
Figure
4
.
FRA
M
eas
ur
e
ment
f
or
te
rmin
al
U
-
V
at
normal c
onditi
on a
nd
PP
T
fa
ult i
n p
has
e
s
U
an
d V.
(a)
FRA
f
or
1HP
T
PIM,
(b)
FRA
for 2H
P
TPIM, (c
)
FR
A
f
or
3HP
TP
I
M
a
nd (d
)
FR
A
f
or
5.5HP T
PIM.
Evaluation Warning : The document was created with Spire.PDF for Python.
In
t J
P
ow Elec
& Dri S
ys
t
IS
S
N:
20
88
-
8
694
Ph
as
e to
ph
as
e
fau
lt
detect
ion of
3
-
phas
e
in
duct
ion m
otor
usi
ng FRA tec
hniq
ue
(
A.
A.
Alawa
dy
)
1245
(
a
)
(b)
(
c
)
(
d
)
Figure
5
.
FRA
M
eas
ur
e
ment
f
or
te
rmin
al
V
-
W
at
normal c
onditi
on and
P
PT
fa
ult i
n p
ha
se
s
U
an
d V.
(a)
FRA
f
or 1HP
T
PIM,
(b)
FRA
for 2H
P
TPIM, (c
)
FR
A
f
or
3HP
TP
I
M
a
nd (d
)
FR
A
f
or
5.5HP T
PIM.
(a)
(b)
(c)
(d)
Figure
6
.
FRA
M
eas
ur
e
ment
f
or
te
rmin
al
W
-
U
at
normal c
onditi
on a
nd
PP
T
fa
ult i
n p
has
e
s
U
an
d V.
(a)
FRA
f
or 1HP
T
PIM,
(b)
FRA
for 2H
P
TPIM, (c
)
FR
A
f
or
3HP
TP
I
M
a
nd (d
)
FR
A
f
or
5.5HP T
PIM.
Evaluation Warning : The document was created with Spire.PDF for Python.
IS
S
N
:
2088
-
8
694
In
t J
P
ow
Ele
c
&
Dr
i
S
ys
t
,
V
ol
.
1
1
, N
o.
3
,
Se
ptembe
r
2020
:
12
41
–
12
48
1246
The
main
ef
fe
ct
of
PP
T
f
ault
was
by
c
reats
a
s
hort
ci
rc
uit
betwe
en
phas
es
wi
nd
i
ng
s
le
ad
to
re
duce
the
num
ber
o
f
tur
ns
le
ss
than
the
normal
c
onditi
on
.
T
hus,
t
ha
t
will
red
uce
t
he
wi
nd
i
ng
s
le
ng
t
h
of
the
c
oils.
So
that,
the
t
otal
resist
ance
of
t
he
c
oils
t
hat
m
ake
up
the
phase
will
decre
ase
.
Also,
t
he
fl
ux
in
the
c
oi
ls
will
reduce.
Creat
ion
of
the
PP
T
fau
lt
will
le
ad
to
re
duced
mag
netiz
ing
in
du
ct
a
nce
of
phase
U
a
nd
phase
V
,
because
of
th
e
sho
rted
tur
ns
bet
ween
ph
a
ses
U
-
V
a
nd
the
tu
r
ns
rem
ov
e
d
f
rom
ph
ase
U
a
nd
ph
ase
V
.
Th
us
,
cause
s
a
n
i
ncr
ease
the
mag
nitud
e
of
the
res
pons
e
an
d
t
his
c
on
si
dered
def
ect
i
ve
re
sp
onse
.
And
s
ur
el
y,
the
co
mp
a
rison
of
fr
e
que
ncy
res
pons
es
(F
R
)
in
case
of
no
rmal
c
onditi
on
an
d
at
P
PT
fa
ult
in
windin
gs
was
cl
early v
isi
ble
in
Fi
gure
4
,
Fi
gure
5
a
nd Fig
ur
e
6
,
a
nd
the
re
spon
ses
s
how
s
the
case
s of
e
ff
ect
acco
rd
i
ng
to
the
connecti
ons
of
measu
reme
nts
cond
ucted, w
hi
ch
a
re th
e
FRA
meas
ur
e
d for t
erminals
(
U
-
V, V
-
W
and
W
-
U).
Fo
r
m
ore
cl
ari
f
ic
at
ion
,
t
he
ma
gn
it
ude
of
de
fe
ct
ive
res
ponse
betwee
n
t
he
re
fer
e
nc
e
FR
a
nd
t
he
FR
at
PPT
fau
lt
c
ondi
ti
on
f
or
te
rmi
na
ls
U
-
V
we
re
ve
ry
evi
den
ts
as
show
n
i
n
Fig
ure
4,
beca
us
e
of
the
PPT
fa
ult
was
create
d
in
t
he
ph
a
ses
U
a
nd
V
windin
gs
to
gethe
r
si
nce
t
he
be
ginnin
g
f
or
al
l
TP
IMs.
On
the
ot
her
hand,
the
healt
hy
phase
W
from
t
he
oth
e
r
side
.T
he
reas
on
f
or
t
he
se
de
fecti
ve
r
esp
on
se
was
due
to
t
he
ma
gn
et
iz
ing
inducta
nce
li
m
it
at
ion
in
the
ph
a
ses
windin
gs
of
T
PIM,
wh
ic
h
is
di
rec
tl
y
af
fected
w
hen
a
ny
fa
ult
occur
i
n
mo
to
r win
ding
s.
5.
ANALY
SIS
U
SIN
G
NCEP
R
I A
LG
ORI
T
H
M
NCEPRI
al
gor
it
hm
was
pro
pose
d
in
[
30
]
,
[
31
]
a
nd
has
be
en
use
d
as
one
of
sta
ti
sti
cal
methods
for
analysi
ng
t
he
FRA
res
ults
.
I
n
this
stu
dy,
th
e
NCEPRI
al
gorith
m
is
us
e
d
for
inter
pr
et
at
ion
a
nd
validat
i
on
t
he
respo
ns
e.
This
performe
d
by
c
ompu
ti
ng
t
he
e
ff
ect
ive
de
viati
on (
E
D
) or
a
ss
essment
facto
r (E
12
) b
y usin
g
(1).
12
=
1
∑
(
1
−
2
)
2
=
1
(1)
Wh
e
re
N
is
th
e
num
ber
of
point
i
n
a
f
requ
ency
re
spo
ns
e,
1
an
d
2
are
t
he
t
ran
s
f
er
f
unct
io
n
values
of
the freq
ue
ncy r
esp
on
se
of
nor
mal win
ding
a
nd f
a
ulty
windi
ng
res
pecti
vel
y
at
f
re
quenc
y
i
.
NCEPRI
al
go
rithm
was
util
iz
ed
f
or
disti
nguis
h
the
diff
e
ren
ces
between
f
re
qu
e
nc
y
res
pons
es
at
normal
c
on
diti
on
a
nd
with
P
PT
fau
lt
ca
se
f
or
f
our
un
i
ts
wh
ic
h
a
re
c
le
arly
visible
i
n
Fi
gure
4,
Figure
5
and
Fi
gure
6.
I
n
t
his
st
udy,
the
f
reque
ncy
res
pons
e
(
FR
)
c
urves
di
vide
d
i
nto
f
our
r
egio
ns
a
cco
rd
i
ng
to
the r
es
ponse
w
ave
-
s
ha
pe.
As
sh
ow
n
in
Ta
ble
2
.
Table
2
.
A
sses
sment
facto
r
(E
12
)
of
PPT
fa
ult
fr
e
quenc
y r
esp
on
ses
by
u
s
ing
NCEPR
I
al
gorithm
Frequ
en
cy
rang
e
E12
in d
B for 1HP
m
o
to
r
E12
in d
B for 2HP
m
o
to
r
E12
in d
B for 3HP
m
o
to
r
E12
in d
B for 5.5
HP m
o
to
r
U
-
V
V
-
W
W
-
U
U
-
V
V
-
W
W
-
U
U
-
V
V
-
W
W
-
U
U
-
V
V
-
W
W
-
U
2
0
Hz
-
2
k
Hz
4
.31
0
7
0
.84
7
9
0
.77
0
7
2
.68
0
7
0
.57
4
3
0
.63
4
1
4
.94
9
4
1
.12
3
5
0
.88
8
6
4
.00
9
3
1
.21
5
3
0
.53
4
2
2
k
Hz
-
2
0
k
Hz
7
.56
6
0
1
.31
1
9
1
.28
3
4
4
.35
1
2
0
.87
6
7
0
.95
3
5
1
0
.62
5
6
1
.90
9
2
1
.48
3
9
7
.58
3
4
2
.01
3
0
0
.87
0
6
2
0
k
Hz
-
2
0
0
k
Hz
7
.03
6
7
1
.46
2
1
1
.32
1
1
4
.73
0
2
0
.96
9
0
0
.91
7
1
1
1
.48
0
9
1
.80
7
8
1
.59
9
7
8
.75
5
2
2
.28
7
8
1
.12
0
5
2
0
0
k
Hz
-
2
MHz
1
.57
3
3
0
.83
7
9
0
.15
9
5
0
.45
1
7
0
.13
4
1
0
.24
0
7
0
.99
0
8
0
.29
5
2
0
.34
9
3
1
.10
7
9
0
.84
3
1
1
.12
0
7
The
asse
ssme
nt
factor
(E
12)
values
was
cal
culat
ed
f
or
al
l
cases
in
this
study
by
us
i
ng
the
e
qu
at
io
n
(1).
Table
2
shows
the
val
ues
of
E
12
for
t
he
f
our
unit
s
of
TPIM.
As
cl
ea
rly,
the
E1
2
va
lues
f
or
P
PT
fa
ult
are
high
f
or
t
he
f
reque
ncy
reg
i
on
(20
Hz
-
2kH
z),
it
range
d
betwee
n
(
2.6
807
t
o
4.9
494)
wh
e
n
it
cal
culat
ed
for
te
r
minal
U
-
V.
A
nd
f
or
te
r
minals
V
-
W
a
nd
W
-
U
the
E
12 r
an
ge
d betwee
n
(
0.5
342
t
o 1.215
3). As E
12
vales
sh
owe
d,
t
he
e
f
fecti
ve
of
fau
lt
was
more
e
vid
ent
at
U
-
V
ca
lc
ulati
on
.
This
du
e
to
t
he
im
pa
ct
of
PPT
fa
ul
t
was
occure
d
directl
y
on
t
his
phas
e
.
T
heref
or
e
,
t
his
reg
i
on
is
c
on
si
der
e
d
as
s
erio
us
fail
ure.
Fo
r
fr
e
quenc
y
reg
i
on
s
(2kHz
-
20kHz)
and
(
20kHz
-
20
0kHz
),
t
he
E1
2
value
s
for
te
rmin
al
s
U
-
V
,
V
-
W
a
nd
W
-
U
was
range
d
be
twee
n
(4.35
12
to
11.
4809)
,
(0.87
67
to
2.2
87
8)
a
nd
(0.87
06
to
1.599
7)
res
pecti
vely
as
sho
wn
in
Ta
ble
2.
T
ha
t
mean
it
was
hi
gh
e
r
t
han
the f
irst
va
lues.
S
o
th
at
,
this
re
gion
is
c
onside
red
as
se
r
iou
s d
ist
ort
ed, t
his
due
to
t
he
eff
ect
s
of
PPT
fa
ult
in
mo
t
or
windin
gs
.
For
the
f
re
qu
e
nc
y
re
gi
on
(20
0kHz
-
2MH
z),
the
rec
orde
d
values
we
re
lowe
r
than
t
he values
in
the
r
est
re
gio
ns. T
his
mean
s that the
ef
fect o
f
PP
T fa
ult i
s
r
el
at
ive in
this
f
re
quenc
y re
gio
n.
Evaluation Warning : The document was created with Spire.PDF for Python.
In
t J
P
ow Elec
& Dri S
ys
t
IS
S
N:
20
88
-
8
694
Ph
as
e to
ph
as
e
fau
lt
detect
ion of
3
-
phas
e
in
duct
ion m
otor
usi
ng FRA tec
hniq
ue
(
A.
A.
Alawa
dy
)
1247
6.
CONCL
US
I
O
N
Ph
ase
to
phase
tur
ns
(
PPT
)
f
ault
in
TP
IM
windin
g
was
presents
in
this
pap
e
r.
T
he
c
ompa
rison
of
fr
e
qu
e
nc
y
re
spon
s
es
(
FR)
c
onfirme
d
that
th
e
PPT
fau
lt
in
T
PIM
wi
nd
i
ng
gi
ves
some
of
va
riat
ion
in
the
FRA
respo
ns
e.
T
he
va
riat
ion
s
i
n
FRA
si
gn
at
ures
f
or
wi
nd
i
ng
s
was
inter
pr
et
e
d
a
nd
a
na
lyzed
in
sect
ion
4.
A
fi
nd
i
ng
of
t
hi
s
paper
,
it
has
been
pr
ov
e
n
t
ha
t
the
de
fect
th
at
occurre
d
du
e
to
P
PT
fau
lt
can
be
diag
nos
e
an
d
detect
ed,
t
hro
ugh
the
var
ia
ti
on
in
the
FR
A
re
spo
ns
e
in
phases
wh
e
n
measu
rin
g
t
he
fr
e
quenc
y
re
sp
onse
.
And
al
s
o,
this
pap
e
r
pr
ese
nt
ed
the
validat
ion
f
or
FR
A
at
the
PPT
fa
ult
case
by
usi
ng
one
of
a
ce
rtifie
d
sta
ti
sti
cal
ind
ic
at
or
s
,
w
hich
is
NCEPRI
al
gorithm
.
W
he
r
e,
the
cal
culat
ion
of
t
he
ass
essment
facto
r
(E12
)
sh
ows
t
he
highest
dist
or
ti
on
ind
ic
at
or
of
f
req
ue
ncy
re
gio
ns
that
has
l
ow
an
d
me
dium
f
reque
ncies
range.
And
in
high
frequ
e
ncies
ra
ng
e,
th
e
disto
rtio
n
le
vel
will
be
relat
ively
i
ns
ign
i
ficant.
T
he
method
propos
ed
i
n
this
pa
per
ha
d
a
us
ef
ul
re
su
lt
,
it
can
be
c
ons
idere
d
as
a
ne
w
met
hod
to
de
te
ct
ion
an
d
dia
gnos
is
th
e
PP
T
fau
lt
in three
-
ph
a
se
inducti
on m
otor.
And
al
s
o,
t
he
appli
cat
ion
s
of d
e
vel
op
e
d m
et
hod
ca
n be
use
d for
diag
nos
is and
detect
ion t
he o
ther
t
yp
es
of
fa
ults in T
PIM.
REFERE
NCE
S
[1]
Za
mudi
o
-
Ramir
ez
,
Isra
el,
Roq
ue
Alfr
edo
Os
ornio
-
Rios,
Mi
guel
Trejo
-
H
er
nande
z
,
Rene
de
Jesus
Ro
mero
-
Tronc
oso,
and
J
ose
Alfonso
Ant
onino
-
Daviu,
"S
ma
rt
-
sensors
to
esti
mate
insulati
on
health
in
ind
uct
ion
mo
tors
v
i
a
ana
lysis of
stray
fl
ux,
"
Ene
rgies
,
vol.
12
,
no
.
9
,
pp
.
1658
,
2019
.
[2]
Akar
Me
hme
t
,
Ankaya
Ilya
s
C.
,
''
Broken
rotor
bar
fau
l
t
detec
tion
in
inve
r
te
r
-
f
e
d
squirre
l
c
age
i
nduct
ion
mo
tors
using sta
tor
c
urr
ent
ana
lysis
and fuzz
y
logi
c
,
''
Tur
k
J
E
le
c
Eng
&
Comp Sc
i
,
v
ol
.
2
0,
no
.
1,
pp
.
107
7
–
1089,
2012
.
[3]
Pezz
an
i
C,
Don
olo
P.
,
Boss
io
G
.
,
Donolo
M.
,
G
uzm
an
A.
,
Zo
ch
oll
S.
E.,
"
Det
ecting
broke
n
roto
r
bar
s
with
ze
ro
-
sett
ing
pro
tecti
on
,
"
Industrial
&
Comm
erc
ial
P
ower
Syste
ms
Technical
Con
fe
r
enc
e
(I&
CPS)
,
IEE
E
/IA
S
48
th.
,
pp.
1
–
12
,
2012
.
[4]
Qi,
Xiang.
,
"P
ra
ct
i
ca
l
C
irc
ui
t
de
sign
to
protect
mot
or's
phase
fa
il
ure
op
era
t
ion
,
"
2010
Asia
-
Pa
ci
f
ic
Confe
r
enc
e
o
n
Powe
r E
le
c
troni
cs
and
Design
,
p
p.
104
–
107
,
201
0
.
[5]
G.
B.
Kl
im
an
,
W.
J.
Prem
erl
an
i
,
R
.
A.
Koegl
,
a
nd
D.
Hoewel
er,
"
A
n
ew
appr
oa
ch
to
on
-
lin
e
tur
n
fau
lt
de
te
c
ti
on
in
ac
mot
ors,"
Proc
.
IE
EE Industry Appli
ca
ti
ons So
c
.
Annua
l
M
ee
t
in
g
Conf
.
,
San
Die
go,
CA
,
pp
.
687
–
693
,
1996
.
[6]
P.
J.
Ta
vn
er
a
nd
J.
Pen
ma
n
,
''El
e
ct
ri
ca
l
t
ec
h
nique
s
for
mon
it
oring
,
''
in
Co
ndit
ion
Monit
or
ing
of
Elec
trical
Mac
hine
s
,
L
et
ch
worth,
Her
tfordshire
,
U.K.: Re
se
arc
h
S
tud
ie
s Pre
ss
Lt
d
.
,
vol
.
1
,
p
p.
93
–
132
,
1987
.
[7]
P.
Zh
ang,
Yi
Du
,
T.
G.
Hab
etler
,
Bin
Lu
,
"
A
sur
vey
of
condi
t
ion
mon
it
oring
and
prote
c
ti
on
m
et
h
ods
for
me
d
ium
-
volt
ag
e
indu
ct
io
n
mot
ors
,
"
IEEE
Tr
ans.
Ind
.
Appl
,
vol
.
47
,
no
.
1
,
p
p.
34
-
46
,
2011
.
[8]
Benbouz
id
,
M.
El
Ha
chemi
,
"A
rev
i
ew
of
indu
ct
ion
mot
ors
signat
ure
an
al
ysis
a
s
a
me
d
ium
for
fau
l
ts
de
te
c
ti
on
,
"
IEE
E
transacti
o
ns on
industrial
el
e
ct
ronics
,
vol
.
47,
no
.
5
,
pp
.
98
4
–
993
,
2000
.
[9]
Siddique
,
Arfa
t
,
G.
S.
Yada
va
,
and
Bhim
Sing
h,
"A
rev
ie
w
of
stat
or
fau
l
t
mo
nit
oring
techniq
ues
of
indu
ct
ion
mot
ors
,
"
I
EE
E
tr
ansacti
ons on en
ergy
con
ve
rs
ion
,
vol.
20,
no.
1
,
p
p.
106
–
114
,
200
5.
[10]
Singh,
Arvind
,
Bevon
Grant
,
R
onal
d
DeFour
,
Chandra
bhan
Sh
arm
a
,
and
San
jay
Baha
doorsing
h
,
"A
r
evi
ew
of
induc
ti
on
mo
tor faul
t
modeling
,
"
El
e
ct
ric
Pow
er
S
yste
ms
Re
sear
ch
,
vol
.
133
,
pp
.
19
1
–
197
,
2016
.
[11]
M.
Arkan
,
D.
K.
Pe
rovic,
and
P.
Uns
worth,
"
Onl
ine
stat
or
f
ault
d
ia
gnosis
in
indu
ct
ion
mo
tors,
"
P
roc.
Inst.
E
lect.
Eng.
,
E
le
c
t. P
ow
er
Appl
.
,
vol
.
14
8,
no
.
6
,
pp
.
537
–
547,
2001
.
[12]
A.
H.
Bonn
et
a
nd
G.
C
.
Souku
p,
"
Cause
and
a
nal
ysis
of
st
at
or
and
rotor
failu
res
in
three
-
pha
se
squirre
l
-
ca
g
e
induc
ti
on
mo
tors
,"
IE
EE
Tr
ansac
ti
ons on
Industry
Application
,
vol
.
28
,
no
.
4
,
pp
.
9
21
–
937,
1992
.
[13]
J.
Scot
ti
l
e,
J.
L.
Ko
hle
r,
"
An
on
-
l
ine
method
to
d
e
te
c
t
incipie
n
t
fail
ure
o
f
turn
insul
at
ion
in
r
andom
-
wound
mot
ors
,"
IEE
E
Tr
ans.
on
Ene
rgy
Con
ve
rs
i
on
,
vol
.
8
,
no
.
4
,
pp.
762
–
768
,
19
93.
[14]
P.
J.
T
avne
r
a
nd
J.
Penma
n
,
''
T
em
per
at
ur
e
moni
tori
ng
,
''
in
Condit
ion
Mo
nit
oring
o
f
Ele
ct
rical
Ma
chi
ne
s
,
Le
t
chwort
h,
Her
tfordshire
,
U.K.
:
Resea
r
ch
Stud
ies
Press
Lt
d
.
,
vol
.
1,
pp.
228
–
239,
1987
.
[15]
S.
B.
L
ee,
R.
M.
Ta
l
lam
and
T.
G.
Habe
t
le
r
,
"
A
robust,
on
-
li
ne
turn
-
fau
lt
det
e
ct
ion
techni
que
for
indu
ct
i
on
ma
ch
ine
s
b
ase
d
on
moni
tor
ing
the
sequ
e
nc
e
c
ompone
nt
im
p
e
danc
e
m
at
rix
,"
IEE
E
Tr
ans.
P
ower
Elec
tronics
,
vol.
18
,
no
.
3
,
pp
.
865
–
872
,
2003
.
[16]
C.
Xianr
ong
,
V
.
Cocque
mpot,
an
d
C.
Chr
istophe,
"
A
mode
l
of
async
hronous
mac
hine
s
for
st
at
or
f
aul
t
detec
t
ion
an
d
isola
ti
on
,
"
I
EE
E
Tr
ans.
Ind.
E
le
c
t
ron
.
,
vol
.
50
,
no
.
3,
pp.
578
–
584,
2003.
[17]
F.
Zi
dan
i,
M
.
E
.
H.
Benbouzid,
D
.
Diallo,
and
M.
S.
Nait
-
Said
,
"
In
duct
ion
mot
or
st
at
or
fau
lt
s
di
agn
osis
by
a
cur
r
ent
Concordi
a
patter
n
-
base
d
fu
zz
y
d
e
ci
sion syste
m,
"
I
EE
E
Tr
ans.
Ener
gy
Conve
rs
.
,
vo
l.
18
,
no
.
4
,
pp
.
4
69
–
475,
2003
.
[18]
D.
C.
Pat
el,
M.
C.
Chandork
ar,
"
Modeli
ng
and
ana
lysis
of
stator
int
er
turn
f
ault
locati
on
eff
ec
t
s
on
induc
t
io
n
ma
ch
ine
s
,
"
I
EEE
Tr
ans.
Ind. Elec
troni
cs
,
vo
l. 6
1,
pp
.
4552
–
456
4,
2014
.
[19]
M.
E
.
H
.
B
enbo
uzi
d,
M.
Vieira,
C.
The
ys,
"
Indu
ct
ion
mot
o
rs
’
fa
ult
s
de
tecti
on
an
d
local
iz
a
ti
on
using
sta
tor
cur
ren
t
adva
nc
ed
sign
al
proc
essing
t
ec
hn
ique
s
,"
I
EE
E
Tr
ans.
Powe
r
Elec
t
ron
.
,
vol
.
14
,
no
.
1,
pp.
14
–
22,
19
99.
[20]
Tushar
G.
Vilh
e
kar
,
Maka
r
and
S
.
Ba
llal,
and
H.
M.
Suryawanshi
,
"
Appli
ca
t
ion
of
double
par
ks
ve
c
tor
appr
oa
ch
fo
r
det
e
ct
ion
of
in
terturn
fau
lt
in
in
duct
ion
mot
or
,"
Inte
rnational
Co
nfe
renc
e
on
Con
dit
ion
Assess
ment
Techni
qu
es
in
El
e
ct
rica
l
S
ystem
s (CA
TC
ON),
pp.
173
–
178
,
20
15.
[21]
Gandhi,
Ke
ta
n
R.
,
and
Ket
an
P.
Badguj
a
r,
"A
rt
i
fic
i
al
n
eur
al
ne
t
work
base
d
id
en
ti
ficat
ion
of
d
eviati
on
in
fr
eque
n
cy
respons
e
of
po
wer
tr
ansforme
r
windings,
"
201
4
Annual
Int
ernati
onal
Confe
re
nce
on
Eme
rgin
g
Re
s
earc
h
Are
as:
Magne
tics,
Ma
c
hine
s and
Dr
ives
(AI
CE
RA
/
iCM
MD)
,
pp.
1
–
8.
I
EE
E
,
2014
.
Evaluation Warning : The document was created with Spire.PDF for Python.
IS
S
N
:
2088
-
8
694
In
t J
P
ow
Ele
c
&
Dr
i
S
ys
t
,
V
ol
.
1
1
, N
o.
3
,
Se
ptembe
r
2020
:
12
41
–
12
48
1248
[22]
F.R.
Bl
nqu
ez,
C
.
A.
Pl
at
ero
,
E.
R
ebol
lo
,
F.
Bl
zqu
ez
,
"
Fiel
d
-
wind
i
ng
fau
l
t
d
etec
t
io
n
in
synchronou
s
ma
ch
ine
s
wi
th
stat
ic
ex
ci
t
at
ion
through
fr
equency
response
a
nal
ysis
,"
Int
ernati
onal
Journal
of
E
lectric
a
l
Powe
r
&
Ene
r
gy
Syste
ms
,
vol
.
73
,
pp
229
–
239
,
20
15
.
[23]
J.
C.
G.
Arispe
and
E.
E.
Mo
mbe
ll
o
,
"
Det
ec
t
i
on
of
fai
lur
es
withi
n
tr
ansformers
by
FR
A
usin
g
mul
ti
r
esolut
io
n
dec
omposition,
"
IEE
E
Tr
ans.
Po
wer
Deliv
.
,
vo
l.
29,
no
.
3
,
pp
.
11
27
–
1137,
2014
.
[24]
Vilhe
kar
,
Tusha
r
Gulabr
ao,
Mak
ara
nd
Sudhakar
Bal
lal,
a
nd
Bhi
mra
o
S
.
Umr
e,
"
Ap
pli
cation
o
f
S
wee
p
Frequ
enc
y
Response
Analy
sis
for
the
d
et
e
ction
of
winding
f
aul
ts
in
inductio
n
mot
or
,
"
IECO
N
2016
-
42nd
An
nual
Confe
ren
ce
of
th
e
I
EE
E
Indu
strial
Elec
tronics
Soci
e
ty
,
pp
.
14
58
–
1463,
2016
.
[25]
"
IEE
E
guid
e
for
the
app
li
c
ation
and
interpre
t
atio
n
of
fre
quen
cy
r
esponse
ana
lysis
for
oil
-
i
mm
ers
e
d
tra
nsforme
rs
,"
IEE
E
power
and
ene
rgy
soci
et
y
,
IEE
E
Std
C57.
1
49
-
2012
,
pp
.
1
–
72,
2013
.
[26]
P.
Piche
r,
"
Me
c
hani
c
al
cond
it
io
n
assess
me
nt
of
tra
nsformer
wi
ndings
using
fre
quenc
y
response
ana
lysis
(F
RA
),
"
Cigre,
Ev
alua
ti
o
n
,
vol
.
A2
.
26,
pp
.
30
–
34
,
2008
.
[27]
Behj
at
V
.
,
Vah
edi
A
.
,
Seta
y
eshme
hr
A
.
,
Borsi
H
.
,
Gock
enbach
E
.
,
"
Sw
ee
p
fre
quenc
y
respo
nse
analysis
fo
r
dia
gnosis
of
lo
w
l
ev
el
short
c
irc
uit
f
aults
on
th
e
windings
of
power
tr
ansf
orme
rs:
An
exp
eri
m
ent
a
l
study
,"
Inte
rnational
Jo
urnal
of
Elec
tri
c
al
Pow
er
and
En
ergy
Syst
ems
,
vo
l.
42
,
no
.
1
,
pp
.
7
8
-
90,
2012
.
[28]
S.
Al
-
Amer
i
,
A
.
A.
Al
awa
dy,
M.
F.
M.
Yous
of,
H.
Ahmad
,
Ali
.
A.
Salem,
and
M.
A.
Ta
l
ib,
''Freque
ncy
respons
e
ana
lysis
for
tra
n
sforme
r
t
ap
ch
a
nger
damage
de
te
c
ti
on
,
''
Inte
rna
ti
onal
Journal
o
f
Powe
r
El
e
ct
ro
nic
s
and
Dr
iv
e
Syste
ms
(IJ
PE
D
S)
,
v
ol
.
11
,
no
.
1
,
pp
.
350
–
358
,
2
020
.
[29]
N.
F.
M.
Yasid,
A.
A.
Alawa
dy
,
M.
F.
M.
Yous
of,
S.
Al
-
Am
er
i
and
M.
S.
Ka
ma
rudin
,
''Int
erp
ret
a
ti
on
of
sw
eep
fre
quenc
y
respo
nse
analysis
trac
es
on
inter
-
turn
s
hort
ci
r
cuit
fau
lt
,
''
Inte
rnationa
l
J
ournal
of
Pow
er
El
e
ct
roni
cs
and
Dr
iv
e
Syst
ems
(I
JP
EDS)
,
vol
.
11
,
no.
1,
pp.
309
–
3
16,
2020
.
[30]
OM
ICRON
,
"
FR
an
al
yz
er
user
ma
nua
l
-
SF
RA
ana
ly
ze
r
for
po
wer
tr
ansforme
r
winding
di
agno
sis
,
"
O
MICR
O
N
el
ect
ro
nics
,
20
06.
[31]
NCEPR
I,
"
Appli
ca
t
ion
guid
el
in
e for
tra
nsform
er winding
distor
tion te
st
technolo
gy
,
"
1999.
BIOGR
AP
H
I
ES
OF
A
UTH
ORS
Ahmed
All
awy
Alawa
dy.
He
r
e
ce
iv
ed
h
is
ba
ch
el
or
d
egr
e
e
(B
.
Eng.
)
in
Elec
tr
i
ca
l
Engi
n
ee
r
ing
from
Univ
ersit
y
of
Kufa
,
Ir
aq
in
2007.
He
r
ecei
v
ed
ma
ster
d
egr
e
e
(M.
Eng.
)
in
Elec
tr
ic
a
l
power
sys
te
m
from
(S
HU
ATS),
India
in
2012.
He
is
a
lectur
er
at
Coll
age
o
f
Te
chn
ica
l
E
ng
ineeri
ng,
The
Isla
mi
c
Uni
ver
sity,
Ir
aq
sinc
e
2007.
Curr
ent
l
y
he
is
pursuing
his
Ph.D.
Degre
e
in
Univer
si
ti
Tun
Hus
sein
On
n
Mala
ysi
a
(UT
HM
).
His
rese
ar
ch
is
on
mo
tors
f
aul
ts
d
et
e
ct
ion
u
sing
Freque
ncy
Response
Analy
sis (FRA
)
te
chnique.
Mohd
Fairouz
Mohd
Yous
of
.
He
obt
ai
ned
Di
p.
Eng,
B.
Eng
and
M.
Eng
f
rom
Unive
rsiti
Te
knologi
Mal
a
ysia.
H
e
com
pl
eted
his
Ph.D.
stu
dy
in
2015
fro
m
The
Unive
rsity
of
Quee
nsl
and,
Aus
tra
li
a
and
cu
rre
ntl
y
a
l
ec
tur
er
at
th
e
D
epa
rt
ment
of
Elec
tri
c
al
Pow
er
Engi
n
ee
ri
ng,
U
niv
ersit
i
Tun
Hus
sein
Onn
Mal
aysia
(UT
HM
).
He
was a
visit
ing
rese
arc
h
er
for
a
y
ea
r
wi
t
h
TNB
Rese
arc
h
in
2018.
His
main
rese
arc
h
is
co
ndit
ion
-
bas
ed
m
onit
oring
and
assess
me
nt
of
pow
er
tr
ansforme
r
and
rot
at
ing
m
ac
hin
e.
He
is
ac
t
ive
ly
invol
v
e
d
with
pro
je
c
ts
fr
om
loc
a
l
an
d
int
e
rna
t
iona
l
industri
es.
Norhafi
z
A
zi
s.
He
re
ce
iv
ed
B
.
Eng
degr
ee i
n
Elec
tr
ic
a
l and Elec
troni
c
Engi
n
ee
r
i
ng
(2007)
fro
m
Univer
siti
Putra M
al
aysia a
nd
PhD
degr
e
e
in
E
le
c
tri
c
al
Pow
er
Eng
ine
er
ing
(2012)
f
rom
Th
e
Univer
sity
of
Manc
hest
er in
UK
.
Curre
nt
ly, he i
s
an
associ
ate
prof
essor a
t the
Dep
art
m
ent
of
El
e
ct
ri
ca
l
and
E
l
ec
tron
ic
Engi
n
eering,
Univ
ersit
i
Putra
Mal
aysia, M
al
aysia.
His r
e
se
arc
h
int
er
ests
ar
e
in
-
s
erv
ice agei
ng
of transforme
r
insu
la
ti
on
,
condition monit
oring
,
asset
ma
nag
em
en
t, high voltage
engi
n
ee
ring
and
alter
n
at
iv
e
insul
at
ion
ma
t
eri
a
ls f
or
h
ig
h
voltage
equi
pm
ent
.
Mohd
Aiza
m
Talib.
He
re
ce
iv
ed
his
Bac
h
el
or
in
El
e
ct
r
ic
a
l
Eng
i
nee
ring
from
th
e
Univer
sity
of
Portsmouth,
U
K
in
1997,
Ma
ster
degr
e
e
in
El
e
ct
ri
ca
l
Engi
n
ee
ring
fro
m
Un
ive
rsiti
Te
n
aga
Nasiona
l
(UN
IT
EN),
Mal
aysia
i
n
2001,
and
Ph.
D.
degr
e
e
fro
m
Univer
siti
Te
kno
logi
Mal
aysia
in
2017.
Upon
gra
d
uat
ion
in
1997,
h
e
worked
with
A
BB
Tra
nsm
ission
and
Distributio
n
Sdn
Bhd
as
a
Design
Eng
ineer.
Since
1998
,
h
e
h
as
be
en
em
p
l
oyed
by
TNB
R
ese
arc
h
Sdn
Bhd
as
a
Rese
arc
h
Engi
ne
er
and
no
w
is
a
te
chn
ical
expe
rt
on
tra
nsf
orme
r.
His
r
ese
a
rch
int
er
ests
a
re
in
tr
ansforme
r
condi
ti
on
mon
itoring,
insula
t
ion dia
gnostic a
nd
d
ie
l
ec
tr
ic m
e
asure
me
nts.
Evaluation Warning : The document was created with Spire.PDF for Python.