Indonesian J
ournal of Ele
c
trical Engin
eering and
Computer Sci
e
nce
Vol. 2, No. 1,
April 201
6, pp. 69 ~ 78
DOI: 10.115
9
1
/ijeecs.v2.i1.pp69
-78
69
Re
cei
v
ed Fe
brua
ry 11, 20
16; Re
vised
Ma
rch 1, 201
6; Acce
pted
March 15, 20
16
Induction Motors Stator Fault Analysis based on
Artificial Intelligence
Huss
ein Tah
a
Huss
ein*, Mohamed
Ammar, Mohamed Mousta
fa Ha
ssa
n
Dept. of Electri
c
al Po
w
e
r and
Machi
nes, F
a
cult
y
of Engin
e
e
r
ing, Ca
iro Un
i
v
ersit
y
, Giza, E
g
ypt
*Corres
p
o
ndi
n
g
author, e-ma
i
l
: h.adel
20
11@
hotmai
l
.com
A
b
st
r
a
ct
T
h
is article pr
e
s
ents a metho
d
for fault dete
c
tion an
d di
ag
nosis of stator inter-turn sh
ort circuiti
n
three
ph
ase
i
nducti
on
mach
ines. T
h
e tec
h
niq
ue
is
base
d
o
n
th
e stato
r
curre
nt a
nd
mo
de
lli
ng
in
t
h
e
dqfra
me
usin
g
an Ada
p
tive
Neur
o-F
u
zz
y
a
r
tificial i
n
tell
ige
n
ce a
ppro
a
ch.
T
he deve
l
o
p
ed fau
l
t ana
ly
sis
meth
od
is ill
ustrated us
ing MA
T
L
AB simulati
o
n
s. T
he obt
a
i
n
ed res
u
lts are
pro
m
isi
n
g
base
d
on th
e new
fau
l
t
detectio
n
ap
pr
oach.
Ke
y
w
ords
: F
ault dia
g
n
o
sis, inducti
on
motor
,
turn-to-turn
stator fault, dq
mode
lli
ng, Ne
uro
-
F
u
zz
y
,
ANF
I
S
Copy
right
©
2016 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
The indu
ctio
n machi
ne is commo
nly used in all the
industri
e
s. It is utilized in
90% of
electri
c
al m
o
tor ap
plicatio
n
s
[1]. The me
rits of
the ind
u
ction m
a
chi
ne are its lo
w pri
c
e, ea
se
of
control and
reliability. Investigat
ing induction machi
nes faults is
crucial
to mi
nimize downt
i
me
and the cost
of damage
s [1, 2].
Theind
uctio
n
ma
chine
fa
ults a
r
e
cl
assified
as wi
n
d
ing fa
ults, u
nbala
n
ced
st
ator
and
rotor,
bro
k
e
n
roto
r b
a
rs,
Ecce
ntri
city and
bea
ring
faults. Th
e f
a
ilure
du
e to
stato
r
windi
ng
brea
kd
own i
s
about
30
~4
0% of total i
ndu
ction F
a
u
l
ts [3]. The
p
r
edi
ction
s
of
stator fa
ults
will
save the
high
maintena
nce
co
st [3, 4].There a
r
e
a
lot of approa
che
s
to dia
gno
se
the stator tu
rns
fault. Some
method
s a
r
e
based
on th
e
stato
r
curr
e
n
t
s an
d fa
st F
ourie
r t
r
an
sfo
r
ms (F
FT) while
other metho
d
s
u
s
e
the
torque
profile
a
nalysi
s
a
nd
v
i
bration
st
udy [4, 5]. Rece
nt re
se
arch
work
investigate
d
t
he u
s
e
of i
n
te
lligent
cont
rol
,
Fuzzy
lo
gic (FL
)
,
Neural Network (NN),
com
b
inatio
n
o
f
FL an
d
NN
and
adaptive
co
ntrol i
n
fa
ult analy
s
is [
6
-8]. Thi
s
art
i
cle i
s
o
r
g
ani
zed
a
s
follo
ws:
Modellin
g of the three p
h
a
s
e ind
u
ctio
n motor fo
r bot
h the healthy
and faulty ca
se
s is presen
ted
in section II. An overview
of the Adapti
v
e Neur
o-Fuzzy Inference
System
(ANF
IS) is discussed
in section III. The proposed faul
t analysis techni
que i
s
invest
igated in section IV through
MATLAB si
mulation
s of
indu
ction m
a
chi
n
e
s
with
inter-tu
r
n
st
ator fault
s
. The results
and
con
c
lu
sio
n
are discu
s
sed i
n
se
ction V.
2
Modelling of A Three Pha
se Inductio
n
Motor
A dq fra
m
e i
s
use
d
to
red
u
c
e th
e compl
e
xity of differential eq
uatio
ns. Th
e o
r
igin
al stato
r
and
rotor fra
m
es
of refe
rence a
r
e tra
n
sformed to
a co
mmon
frame that
rot
a
tes
with a
r
b
i
trary
angul
ar velo
city [9].
3 Health
y
Case
The three p
h
a
se
s of a
he
althy motor a
r
e sym
m
etri
cal. Thus, all t
he ph
ases
h
a
ve the
same n
u
mb
er of turns [8-1
2]. The rotor i
s
bala
n
ced st
ar co
nne
ction
cage rotor
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN: 25
02-4
752
IJEECS
Vol.
2, No. 1, April 2016 : 69 – 78
70
Figure 1. 3ph
stator wi
ndin
g
The voltage e
quation
s
of the motor can b
e
written a
s
b
e
low:
V
s
abc
= r
s
abc
i
s
ab
c
+
p
λ
s
abc
,
0 = r
r
abc
i
r
abc
+
p
λ
r
abc
(1)
λ
s
abc
= [
λ
s
a
λ
s
b
λ
s
c
]=
L[i
a
i
b
i
c
]
Whe
r
e
P
=d/dt
Conve
r
ting to
dq stationa
ry frame
X
dq0
=K
X
abc
=
2
3
10
.
5
0
.
5
33
0
22
0.
5
0
.
5
0.
5
[ X
abc
] (2)
The voltage e
quation
s
of st
ator and
rotor arede
rived a
s
:
v
s
dq0
= r
s
dq0
i
s
dq
0
+
p
λ
s
dq0
,
0 = r
r
dq0
i
r
dq0
−
ω
r
01
0
10
0
00
0
λ
r
dq0
+
p
λ
r
dq0
(3)
The stato
r
re
sista
n
ces in t
he dq fram
e depe
nd on th
e stator resi
st
ance value
s
for ea
ch ph
ase.
R
s
dq0
=
11
12
13
21
22
23
31
32
33
s
ss
s
ss
s
ss
rr
r
rr
r
rr
r
(4)
The Roto
r re
sistan
ce
R
r
dq0
=r
r
I
3x
3
(5)
The Moto
r flux equation
Evaluation Warning : The document was created with Spire.PDF for Python.
IJEECS
ISSN:
2502-4
752
Inductio
n
Motors Stato
r
Fa
ult Analysi
s b
a
se
d on Artificial Intelligen
ce
(Hussei
n
Taha Hussei
n)
71
s
qd
0
r
qd
0
λ
λ
=
ss
r
s
qd
0
q
d
0
qd
0
rs
rr
r
qd
0
q
d
0
qd
0
LL
I
LL
I
(6)
Based
on
a
balan
ce
d Y
3ph in
du
ctio
n moto
r, the
neutral current ha
s
ze
ro
value I
s
0
=I
r
0
=0.
Acco
rdi
ng to
balan
ce
d st
ator
con
d
ition, t
he turns
of e
a
ch
ph
ase a
r
e eq
ual
(Na
= Nb
= Nc
=
Ns).
The su
pply voltage in the
dq frame i
s
:
V
q
s
=
2
/3[V
a
s
-0.5(V
b
s
+V
c
s
)],
V
d
s
=1/
3
(-V
b
s
+V
c
s
)
V
0
s
=
1
/3 (V
a
s
+ V
b
s
+ V
c
s
) (7)
The flux equa
tions are:
p
λ
s
q
=V
q
s
- r
s
11
i
s
q
- r
s
12
i
s
d
,
p
λ
s
d
=V
d
s
- r
s
21
i
s
q
- r
s
22
i
s
d
,
p
λ
r
q
=-
r
r
r
i
r
q
+ w
λ
r
d
,
p
λ
r
d
=-
r
r
r
i
r
d
- w
λ
r
q
(8)
The develo
p
e
d
torque a
nd
spe
ed are given by:
T
d
= (3/2
)(P)
(
λ
s
d
i
s
q
-
λ
s
q
i
s
d
) (9)
Whe
r
e P is n
u
mbe
r
of pair poles
P
w
m
=P
/
(
2
J
) (
T
d
-T
L
-Td
amp
) (
1
0)
4
Inter-Turn F
a
ult Ca
se
Und
e
r “
a
” p
h
a
se inte
r-tu
r
n
fault, the motor pa
ramet
e
rs
(stato
r re
sista
n
c
e
, ind
u
ctan
ce
and th
e m
u
tu
al ind
u
cta
n
ce
between
all
pha
se
s a
n
d
the
faulty pha
se) cha
nge
a
s
sho
w
n
in Fi
gure
(1).
X(fault %)=N
a2
(fault turns)/
N
a
(he
a
lthy phase turn
s)
(11
)
r
sh
=X
r
a) health
y
=r
a)f
,
L
a1a1)f
=(1-X
)
2
L
asas)healt
h
y
=L'
asas
,
L
a2a2)f
=X
2
L
m)h
ealth
y
=L
shsh
,
L
a1a2)f
=(1-X
)
X
L
m)health
y
=
L
assh
,
L
asr
=(1
-
X
)
L
m)h
ealth
y
, L
m)heal
th
y
=L'
asr
+L
shar
(12
)
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN: 25
02-4
752
IJEECS
Vol.
2, No. 1, April 2016 : 69 – 78
72
Figure 2. 3ph
Induction mo
tor in dq fram
e with
turn fa
ult in q phase
represent ph
ase a fault
The flux equa
tion in the dq frame after ta
king the
sho
r
ted turn
s in co
nsid
eratio
n is:
sh
λ
q
s
λ
q
s
λ
d
r
λ
q
r
λ
q
=
sh
ssh
sh
r
LL
0
L
0
qq
q
ss
h
s
s
r
LL
0
L
0
qq
q
ss
r
00
L
0
L
dq
sh
r
s
r
r
LL
0
L
0
qq
q
sr
r
00
L
0
L
dd
sh
I
q
s
I
q
s
I
d
r
I
q
r
I
q
(13
)
The stato
r
re
sista
n
ce is gi
ven by:
sh
r
q
s
r
q
s
r
d
=
2
sh
r0
0
q
3
ss
0r
r
11
1
2
ss
0r
r
21
2
2
(14
)
The flux linka
ge derive
d
fro
m
equation
(3
) is:
p
λ
sh
q
=V
q
sh
– r
sh
i
sh
q
,
p
λ
s
q
=V
q
s
- V
q
sh
- r
s
11
i
s
q
- r
s
12
i
s
d
,
p
λ
s
d
=V
d
s
- r
s
21
i
s
q
- r
s
22
i
s
d
,
p
λ
r
q
=-
r
r
r
i
r
q
+ w
r
λ
r
d
,
p
λ
r
d
=-
r
r
r
i
r
d
- w
r
λ
r
q
(15
)
The eq
uation
s
(10-14) sho
w
the in
du
ction moto
r dq
modelin
g with
fault con
d
itio
ns a
nd the
effect
of fault severi
ty on
the motor parameters
.
Evaluation Warning : The document was created with Spire.PDF for Python.
IJEECS
ISSN:
2502-4
752
Inductio
n
Motors Stato
r
Fa
ult Analysi
s b
a
se
d on Artificial Intelligen
ce
(Hussei
n
Taha Hussei
n)
73
5
Adap
tiv
e
Neuro-Fu
zz
y
Inferen
ce Sy
stem
The Adaptive N
e
ur
o-
Fuzzy infer
e
nc
e
sys
tem (
A
N
F
I
S
)
depends
on two main s
y
s
t
ems
fuzzy logi
c (F
L) and
artifici
al neural net
works (A
NN).The Fu
zzy lo
gic a
c
ts a
s
the huma
n
lo
gic
thinkin
g
and
theneu
ral net
work a
c
ts a
s
human b
r
ain
[13]. Both th
e FL and ANN increa
se t
he
system
effici
ency a
nd d
e
c
re
ase the
mathemat
i
c
al
equatio
ns compa
r
ed
to other dete
c
tion
method
s [14]
. The syste
m
is wid
e
ly used for m
any appli
c
ation
s
of system
s m
odelling,
cont
rol
system
s an
d
forecastin
g
predi
ction
s
[
15]. The AN
FIS con
s
ist
s
of IF-then rules, trai
ning
and
learni
ng alg
o
rithms [13].
For the Fuzzy inference
system, considera
sy
stem
with two input
s (X,Y) and one output
(Z).Th
e fuzzy rules b
a
sed
on 1
st
ord
e
r S
ugen
o type [16] are:
Rule
1: IF X is A1 and Y is B1 Then f1=p
1
X+
q
1
Y+
r
1
,
Rule
2: IF X is A2 and Y is B2 Then f2=p
2
X+
q
2
Y+
r
2
,
A
i
,B
i
are the
Fuzzy set, f
i
is
the system outputs withi
n
the spe
c
ified fuzzy rule
s
and the
p
i
,q
i
and r
i
are
the desig
n p
a
ram
e
ters ba
sed o
n
the ANFIS training
[7], [17-20]
Figure 3. ANFIS structu
r
e
for tw
o i/p wit
h
three mmf
and on
e O/P
The a
daptive
neu
ro-fu
z
zy inferen
c
e
(A
NFIS) n
e
two
r
k con
s
ist
s
of
five layers.
A norm
a
lizati
on
layer more than the neu
ro
-fuzzy net
work [21-34].
La
y
e
r 1
is th
e fuzzifi
cation
layer adaptiv
e node
s with
bell memb
ership functio
n
with equation o
f
:
µA
i
(X) =
1
1|
|
^
2
Xm
A
bA
б
A
,
µB
i
(Y)=
1
1|
|
^
2
Ym
B
bB
б
B
(16
)
Whe
r
e the m
A
, mB,
б
A,
б
B, bA and b
B
are the bell fun
c
tion pa
ram
e
ters
= 1, 2, 3 [20]
MF1,i= µA
i
(X)
& MF1,i= µB
i
(Y), for i=1,2,
3
(17
)
The A
i
and B
i
are the ling
u
i
s
tic varia
b
le o
f
X and Y
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN: 25
02-4
752
IJEECS
Vol.
2, No. 1, April 2016 : 69 – 78
74
La
y
e
r 2
is th
e rule
s layer
whe
r
e its out
put is co
nsi
d
e
r
ed a
s
fire st
rength of ea
ch
node
W
i
=µA
i
(X)*µB
i
(Y), i=1, 2, 3
(18)
La
y
e
r 3
is th
e norm
a
lization layer an
d its out
put is th
e norm
a
lized
fire stre
ngth
i
W
i
=
(1
2
9
)
Wi
WW
W
(19
)
La
y
e
r 4
is th
e co
nsequ
en
t layer where
each no
de i
s
an
ada
ptive nod
e an
d i
t
s output i
s
t
h
e
prod
uct of the
con
s
eq
uent
polynomial of
fu
zzy rul
e
s a
nd normali
ze
d firing streng
th
_
_
W
i
f
i
=
_
_
W
i
(p
i
X+q
i
Y+r
i
),i=
1, 2, 3…9
(20)
La
y
e
r 5
is th
e deffu
zificati
on laye
r
whi
c
h ha
s
only o
n
e
no
de
(outp
u
t nod
e)
and
its outp
u
t is the
overall ANFI
S output, su
mmation of the layer 4 outp
u
t
ƒ=
__
__
ff
f
12
12
_
i
1
_
i
WW
W
i
(21
)
6
Simulation and Res
u
lts
The
develop
e
d
fault
analysi
s
te
chni
que
is in
vestig
ated
throug
h MAT
L
AB sim
u
latio
n
s. A
n
indu
ction mot
o
r
with inte
r-t
urn
stator fa
ults is
mod
e
ll
ed in SIM
U
LI
NK ba
se
d on
the eq
uation
s
presented in section II. Figure (4) illustrate
s the fault analysi
s
system procedure.
Figure 4. The
fault analysis system for in
ductio
n
motor with dq mod
e
ling
The
qd
cu
rre
n
t indi
cate
be
tter re
sol
u
tio
n
for fault
det
ection. It i
n
crease a
s
the f
ault pe
rcenta
g
e
increa
se
s as
per Fig
u
re
(5)
Evaluation Warning : The document was created with Spire.PDF for Python.
IJEECS
ISSN:
2502-4
752
Inductio
n
Motors Stato
r
Fa
ult Analysi
s b
a
se
d on Artificial Intelligen
ce
(Hussei
n
Taha Hussei
n)
75
Figure 5. The
d CurrentVS Fault Percent
age
s (X %) At Different Lo
ading
The fault d
e
tection te
ch
ni
que u
s
e
s
a
n
ANFIS network to e
s
tima
te the inter turn fa
ult
percenta
ge.
Traini
ng a
nd t
e
sting
data a
r
e gen
erated f
r
om the
SIMULINK i
ndu
cti
o
n moto
r mo
del.
The moto
r lo
ading
con
d
ition wa
s vari
e
d
to simulate
no-loa
d
, 25
%, 50%, 75%, full-load and
110% lo
adin
g
.
The inte
r tu
rn fault pe
rcen
tage
wa
s vari
ed to
spa
n
th
e ra
nge
of 0
~
16% with
ste
p
s
of 0.005. The
total points a
r
e 199.
The A
N
FIS n
e
twork
wa
s trained
with
66
% of t
he total
data a
nd
ch
ecked/te
sted
with the
remai
n
ing
34
%. The de
sig
n
is
ba
sed
o
n
thre
einp
ut fuzzy mem
b
ership
fun
c
tion
s. It wa
s
notice
d
that the learni
ng pha
se
wa
s co
mpleted i
n
t
he first 120
Epoch
s
out o
f
300 iteration
s
.
Figure (6
) views the A
N
FI
S erro
r for th
e di
fferent lo
ading
ca
se
s and fault pe
rcenta
g
e
s
estimating.
Figure 6. % Erro
r VS fault percenta
g
e
s
at different loading
The fault p
e
rcenta
ge at th
e 25% lo
adin
g
ca
se
gives the high
est
error
as th
e
ANFIS
netwo
rk
wa
s
not traine
d wi
th this fault d
a
ta. T
he maxi
mum % error
is 6.83% at 2
5
% loadin
g
a
n
d
16 % fault. The result
s illustrate t
he ANFIS accuracy
for fault detecti
on even for cases that
were
not inclu
ded i
n
the training
data.
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN: 25
02-4
752
IJEECS
Vol.
2, No. 1, April 2016 : 69 – 78
76
The network
initial configu
r
e ratio
n
ha
s an
effect on
the perform
ance and a
c
curacy of
the fault dia
gno
sis
syste
m
. Table (1) sho
w
s
the
errors for
a
two inp
u
t and three in
put
membe
r
ship
function
s.The
thre
e in
put
membe
r
ship
function
ha
s lower erro
r for te
sting
and
che
c
king d
a
t
a
.
Table 1. The
error comp
ari
s
on for th
etwo and thre
e membe
r
ship
Error
T
w
o MMF
Three MM
F
Training data
9*10-4
6*10-4
Testing (75
%
)
6*10-3
4*10-3
Checking (25%
)
9*10-3
3*10-2
The fault severity wasvari
ed from 0%
till 16%. Ho
wever,an actual fault
will be limited to 10%
fault only at 110% loadin
g
based on the
indu
ct
ion mot
o
r over lo
ad p
r
otectio
n
setti
ng I
o.l
=1.
5
*I
rated
7 Conclu
sion
This p
ape
r shows the faul
t diagno
si
s of inter turn fau
l
t of induction
motor b
a
sed
on a
n
artificial n
e
tworksyst
e
m u
s
ing ofthe
st
ator dq
cu
rrents. The
d
q
stator curre
n
ts give bett
e
r
resolution
for inter-turn
fa
ult diagn
osi
s
.
The A
N
FIS
netwo
rk dete
c
ts th
e inte
r t
u
rn
stato
r
fa
ults
with hi
gh
accura
cy even
for lo
w fa
ult p
e
rcentag
es.
The ave
r
a
ge
ANFIS erro
r i
s
1%
amo
n
g
all
the data trai
n
i
ng, testing
a
nd che
cki
ng.
The ANFI
S
initial stru
ctu
r
e ha
s an eff
e
ct on th
e fa
u
l
t
detectio
n
syst
em accu
ra
cy.
APPENDIX
The moto
r pa
ramete
rs a
r
e
given in Tabl
e 2.
Table 2. Moto
r Param
e
ters
parame
ter
Value
po
w
e
r
2hp~1.5k
w
no. poles
4
R
s
4.05 ohm
L
ls
0.014H
R
r
2.6 ohm
L
lr
0.014H
L
m
0.5387H
I
ra
te
d
2.81A
Referen
ces
[1]
Kripak
aran P,
A Naraina a
nd SN Dee
pa.
Conditi
on mo
nitori
ng
in
in
d
u
ction motor by
para
m
eter
estimatio
n
techni
que
. In: Sprin
ger 2
014,
third Internati
ona
l Co
nferen
ce on Soft C
o
mputi
ng f
o
r
Probl
em Solvi
n
g; India: 20
14: 87-9
8
.
[2]
Drif, M’ham
ed,
and
AJ Car
d
o
s
o.
Stator fault
dia
g
n
o
stics in
squi
rre
l ca
ge t
h
ree-
phas
e i
n
d
u
ction
motor
drives
usi
ng th
e inst
antan
eo
u
s
active
an
d re
active
pow
er si
gnatur
e a
n
a
l
ys
es
.
IEEE trans
actions
on
industrial infor
m
atics.
2
014; 1
0
: 1384-
13
60.
[3]
Bind
u S an
d Vino
d V. T
homas.
Diagn
oses
of intern
al fault
s
of th
ree ph
a
s
e squirr
el ca
g
e
ind
u
ctio
n
m
o
tor—A
review
. In: IEEE 2014 Inter
nationa
l Conference on Adv
anc
es
in Energy Conv
ersion
T
e
chnolog
ies (
I
CAECT
)
. 2014
: 48-54.
[4] Siddiqui,
Khad
i
mMoin, Ku
ld
ee
p Sa
ha
y
and
V
K
Giri.
He
alth
mo
nitori
ng
an
d
fault d
i
a
gnos
is
in
ind
u
ctio
n
m
o
to
r-a
re
vi
e
w
.IJAR
EEIE In
te
rn
a
t
i
o
n
a
l
Jo
u
r
na
l
o
f
Ad
va
n
c
e
d
R
e
sea
r
ch
in
El
e
c
tri
c
a
l
, El
e
c
tro
n
i
cs and
Instrume
ntatio
n Engi
ne
erin
g
. 201
4; 3: 6549-
656
5.
[5]
Karmakar S, Chattop
a
d
h
y
a
y
S, Mitra M & Sengu
pta.
T
u
rn-to-turn fault
diag
nos
is of an in
ducti
o
n
motor
by th
e a
nalysis
of tran
sient
a
nd ste
a
d
y state stator
current
.
Innov
ative S
y
stems
Desig
n
a
n
d
Engi
neer
in
g. 2014; 5: 65-
74.
Evaluation Warning : The document was created with Spire.PDF for Python.
IJEECS
ISSN:
2502-4
752
Inductio
n
Motors Stato
r
Fa
ult Analysi
s b
a
se
d on Artificial Intelligen
ce
(Hussei
n
Taha Hussei
n)
77
[6]
A Medou
ed, A
Lebar
ou
d, A Laifa & D S
a
yad.
Class
ificati
on of in
ducti
on
mac
h
in
e fault
s
using ti
me
freque
ncy re
p
r
esentati
on
a
nd p
a
rticle s
w
arm opti
m
i
z
ation
.
Jo
urn
a
l
of Electrica
l
Engi
neer
in
g
T
e
chno
logy
. 2
014; 15: 7
42-7
49.
[7]
W
Li, H Z
hao,
X Y
ang
& W
Den
g
.
Mod
u
lar
i
z
e
d fa
ult d
i
ag
nosis
mode
l of
ind
u
ction
mot
o
r bas
ed
o
n
radi
al bas
is fun
c
tion ne
ural
ne
tw
ork
.
Journal of Process Me
chan
ical En
gi
n
eeri
ng.
20
14; 0
:
1-8.
[8]
MF
D’Ange
lo, RM Palh
ares, LB Cosme, LA
Agui
ar, F
S
F
onseca & W
M
Camin
has. F
ault
detection i
n
d
y
nam
ic s
y
ste
m
s b
y
a F
u
zz
y/
Ba
yesi
an n
e
t
w
ork formulati
on
.
ELSEVIER
.
2014; 21: 6
47-6
53.
[9]
PC Krause, O Wasy
nczuk & SD Sudhoff.
Analysis
of elect
r
ic mac
h
in
ery
and dr
ive systems
. 2
nd
ed:
John W
i
l
e
y & Sons
. 201
3.
[10]
Anna Ph
ilo An
ton
y
a
nd R S
ankar
an. Simu
latio
n
of perfor
m
ance of a ca
ge in
ductio
n
motor drive
n
spoo
ler driv
e
w
i
th spe
ed a
nd
current fee
dba
ck using fi
eld-
o
r
iente
d
contro
l.
IJAREEIE.
2014; 3: 77
97-
780
6.
[11]
T
o
shiji Kato,
K
aoru
Ino
ue,
an
d Ke
isuk
e Yos
h
id
a. Di
ag
nosi
s
of stator-
w
i
n
din
g
-t
urn f
aults
of i
n
d
u
cti
o
n
motor b
y
dir
e
ct detecti
on
of
n
egativ
e se
qu
e
n
ce c
u
rrents.
Electrical
En
gi
neer
ing
in
Ja
p
a
n
. 201
4;
1
86:
75-8
4
.
[12]
MK Ebrahim
i
a
nd M Ehsa
ni.
A gen
eral
appr
oach for curr
e
n
t-base
d
con
d
i
t
ion mon
i
tori
ng
of inducti
o
n
motors.
Journa
l of Dyna
mic S
ystem
s, Meas
u
r
ement, and C
ontrol
. 20
14; 1
36: 1-26.
[13]
Sidd
iqu
e
, N
a
z
m
ula
nd
Hoj
j
at
Ade
li.
Co
mp
u
t
ationa
l i
n
tel
lig
ence: sy
ner
gi
es of f
u
zz
y
l
ogic,
ne
ura
l
netw
o
rks and e
v
oluti
onary co
mp
utin
g:
John
W
ile
y
& So
ns. 201
3.
[14]
AA Bohari, WM Utomo, ZA
Haro
n, NM Zin, SY Sim and
RM Ariff.
Vector contro
l of
in
ductio
n
mot
o
r
usin
g n
eur
al
n
e
tw
ork
. In:
the 8th Intern
atio
n
a
l C
onfer
enc
e
on
Rob
o
tic, Vi
sion, S
i
gn
al
Pr
ocessi
ng
&
Po
w
e
r Ap
plic
ations; Sin
g
a
por
e: Spring
er. 20
14: 501-
50
6.
[15]
Che
n
, Che
ng-
Hun
g
an
d She
ng-Ye
n Yan
g
. Neur
al fuzz
y
i
n
ferenc
e s
y
st
e
m
s
w
i
th k
n
o
w
l
edg
e-bas
ed
cultura
l
differe
ntial ev
oluti
on
f
o
r non
lin
ear s
ystem control.
ELSEVIER
.
201
4; 270: 15
4-17
1.
[16]
Sidd
iqu
e
, Naz
m
ul. Neur
o-F
u
zz
y
Co
ntrol.
Intelli
ge
nt Contro
l
:
Springer Inter
natio
nal P
ubl
is
hin
g
. 201
4.
[17]
Che
n
, Se
ng-C
h
i, Di
nh-K
h
a
Le
and
Va
n-S
u
m N
g
u
y
e
n
.
Adaptiv
e netw
o
rk-bas
ed
fu
zz
y
infer
enc
e
system
(ANFIS) controller for an active
mag
netic bearing s
ystem
with unbalance
m
a
ss
. In:
Spring
er
,
AET
A 2013 Recent Adv
ance
s
in Electric
al
Engi
neer
in
g a
nd Re
late
d Scienc
es, Berli
n
Heid
el
ber
g
.
201
4: 433-
443.
[18]
P Subb
ara
j
an
d B Kan
n
a
p
ira
n
. F
ault d
e
tection
and
di
ag
n
o
sis of p
n
e
u
m
a
tic valv
e usi
n
g Ada
p
tiv
e
Neur
o-F
u
zz
y
I
n
ferenc
e S
y
st
e
m
appro
a
ch.
ELSEVIER Applied Soft Com
puting.
201
4; 19: 362-
371.
[19] Sidd
iqu
e
,
Naz
m
ul
. Fu
zz
y
Co
ntrol.
In: Springer Internatio
n
a
l Pub
lish
i
ng, Intelli
ge
nt Contr
o
l. 201
4: 95-
135.
[20]
Bo
yaci
ogl
u, M
e
lek Ac
ar a
n
d
Der
y
a Avc
i
. An
ada
ptive
net
w
o
rk-bas
ed fuzz
y
infere
nce s
y
s
t
em (ANF
IS)
for the predicti
on of stock market return:
the case of the Istanb
ul stock e
xchan
ge.
ELSE
VIER exper
t
system
s
w
i
th a
pplic
atio
ns
. 20
10; 37: 79
08-7
912.
[21]
PM Meng
ha
l
and
A Ja
ya
L
a
xm
i.
Ne
ural
Netw
ork
bas
e
d
dyn
a
m
ic per
forma
n
ce of
i
n
ductio
n
motor
drives.
In: Spr
i
ng
er, Proce
e
d
i
ngs
of the T
h
ird Int
e
rnati
o
nal
Co
nferenc
e on
Soft C
o
mputin
g fo
r
Probl
em Solvi
n
g. India. 20
14.
[22]
AA Hass
an, M
R
Sa
ye
d
an
d
MA Moustafa
Hassa
n.
Po
w
e
r s
y
stem
q
ual
ity
improv
eme
n
t usi
ng fl
e
x
ib
l
e
ac transmiss
io
n s
y
stems b
a
s
ed
on
ad
apt
ive
ne
uro-fuzz
y
in
ference
s
y
ste
m
.
WSEAS Tr
ansactions
on
Power Systems
. 2013; 8: 1-1
3
.
[23]
J Guan, D Shi, JM Z
u
rada
& Levitan. A
nal
yzi
ng mass
ive dat
a s
e
ts: an ad
aptive
fuzz
y
n
eura
l
appr
oach for
pred
iction,
w
i
t
h
a real estat
e
illustrati
on.
Journ
a
l of Organ
i
z
a
t
io
na
l Computi
ng an
d
Electron
ic Co
mmerc
e
.
201
4; 2
4
: 94-11
2.
[24]
F
F
aghan
i, M Abzari, S F
a
thi
& SAH Mona
j
e
mi.
Designing
a stock trading system
us
ing
Artificial Ner
o
Fu
z
z
y
infer
e
n
c
e syste
m
s a
nd tec
hnic
a
l
ana
lysis
appr
oach
.
Inter
nati
ona
l Jo
urna
l
of Acad
e
m
i
c
Rese
arch in Ac
counti
ng, F
i
na
nce an
d Man
a
g
e
m
e
n
t Scienc
es.
2014; 4: 76
-84.
[25]
Sharma An
ura
g
, Jha Man
o
j a
nd MF Qureshi
.
Gove
rning
Co
ntrol an
d E
x
cit
a
tion C
ontro
l for Stabilit
y o
f
Po
w
e
r S
y
stem
Based on ANFIS.
IJIRSET
.
2014; 3:13
847-
13
855.
[26]
T
a
mer S. Kamel, MA Moustafa Hassan. Ad
aptiv
e N
euro
F
u
zz
y
infer
enc
e s
y
stem (AN
F
IS) for fault
classificati
on i
n
the transmissi
on li
nes.
OJEEE
. 2010; 2: 255
1-25
55.
[27]
G Banu a
nd S
Suja. F
a
u
l
t lo
cation tec
hni
q
ue us
ing GA-A
NF
IS for UHV line
. Arc
h
ives
of Electrica
l
Engi
neer
in
g Jo
urna
l
. 201
4; 63
: 247-26
2.
[28]
Ali, Moham
ed
M. Isma
il and Moham
ed A. Moustafa Hass
an. Spee
d sen
s
orless fiel
d–
or
iente
d
contro
l
of a si
x–
phas
e
saturated m
o
del
of ind
u
ctio
n motors
driv
e
w
i
t
h
on
lin
e stator resista
n
ce
estimatio
n
usin
g ANF
I
S.
Internati
o
n
a
l Jo
urna
l of Model
li
ng, Identific
at
io
n and C
ontro
l.
201
2; 17: 334-
347.
[29]
T
a
mer S Kamel, Mohamed
A. Moustafa
Ha
ssa
n a
nd
Ahda
b El
–Mor
shed
y. A
d
van
c
ed
distan
c
e
protectio
n
tech
niq
ue
base
d
o
n
multi
p
le
clas
sified
A
N
F
I
S consi
deri
ng
different l
o
a
d
in
g c
ond
itions
f
o
r
lon
g
transmiss
i
on l
i
nes
in EP
S
. Internatio
na
l Jour
nal
of Mode
lli
ng, Ide
n
ti
fication
an
d C
ontrol.
20
12
;
16: 108-
12
1.
[30]
EM Abd El-Gaw
a
d
, MM Hassan, MAM Hallo
uda a
nd
O Abu
l
-Hag
ga
g. Stochastic mod
e
li
n
g
compar
e
d
w
i
t
h
artific
i
al
in
tellig
enc
e b
a
se
d a
ppro
a
ch f
o
r short term
w
i
n
d
sp
eed
forec
a
sting.
Journ
a
l
of
Americ
a
n
Scienc
e
. 201
1; 7.
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN: 25
02-4
752
IJEECS
Vol.
2, No. 1, April 2016 : 69 – 78
78
[31]
Aziz,
Abd
e
l, Moustafa Has
s
an and
EA Z
ahab.
A
p
p
lic
ations
of ANF
I
S in
hig
h
i
m
ped
anc
e fau
l
t
s
detectio
n
an
d c
l
assificati
on i
n
distrib
u
tion
net
w
o
rks.
In:
IEEE International
S
y
mposium Diagnostics for
Electric Mach
in
es, Po
w
e
r Elec
tronics & Drive
s
(SDEMPED). 2011.
[32]
Ali, Mohamed M Ismail and MA
Hassan. Parameter
id
e
n
tificatio
n
usi
n
g ANF
I
S for magn
etical
l
y
saturated induction
motor
.
Internati
ona
l Jo
urnal of Syste
m
Dyna
mics A
p
p
l
icatio
ns (IJSDA)
. 2012;
1
:
28-4
3
.
[33]
Abde
l Aziz,
MA Hass
an
a
nd EA
El-Z
a
h
ab. An
artific
i
al
intel
lig
enc
e
bas
ed
ap
pro
a
ch for
h
i
gh
impe
danc
e fa
ults a
nal
ys
is i
n
distri
buti
on
net
w
o
rks.
Inte
rnatio
nal
Jo
urnal
of Syste
m
Dyna
mi
c
s
Appl
icatio
ns (IJSDA).
2012; 1:
44-59.
[34]
HAT
Hussein, ME Ammar, MAM Hassan.
A
N
F
I
S based th
ree ph
ase
ind
u
ction
mot
o
rs stator turn
s
fault ana
lysis
. In: W
C
IS 2014, T
a
shkent Uzbe
kistan. 201
4; 1: 43-51.
Evaluation Warning : The document was created with Spire.PDF for Python.