I
nte
rna
t
io
na
l J
o
urna
l o
f
E
lect
rica
l a
nd
Co
m
pu
t
er
E
ng
ineering
(
I
J
E
CE
)
Vo
l.
16
,
No
.
5
,
Octo
b
er
20
26
,
p
p
.
2
3
7
9
~
2
3
9
2
I
SS
N:
2088
-
8
7
0
8
,
DOI
: 1
0
.
1
1
5
9
1
/ijece.
v
16
i
5
.
pp
2
3
7
9
-
2
3
9
2
2379
J
o
ur
na
l ho
m
ep
a
g
e
:
h
ttp
:
//ij
ec
e.
ia
esco
r
e.
co
m
Stato
r
int
ertu
rn s
ho
rt
circ
uit
fault i
dentif
i
ca
tion in
d
o
ubly
-
fed
induction
gen
erat
o
r and its ana
ly
sis
using
artificial
n
eura
l
network
Viv
ek
K
us
hwa
ha
1
,
3
,
Sa
nja
y
K
um
a
r
M
a
ury
a
1
,
Arv
ind
K
um
a
r
Ya
da
v
1
1
D
e
p
a
r
t
me
n
t
o
f
El
e
c
t
r
i
c
a
l
En
g
i
n
e
e
r
i
n
g
,
G
LA
U
n
i
v
e
r
si
t
y
,
M
a
t
h
u
r
a
,
I
n
d
i
a
2
D
e
p
a
r
t
me
n
t
o
f
El
e
c
t
r
i
c
a
l
En
g
i
n
e
e
r
i
n
g
,
G
y
a
n
G
a
n
g
a
I
n
st
i
t
u
t
e
o
f
Te
c
h
n
o
l
o
g
y
a
n
d
S
c
i
e
n
c
e
s
,
Ja
b
a
l
p
u
r
,
I
n
d
i
a
Art
icle
I
nfo
AB
S
T
RAC
T
A
r
ticle
his
to
r
y:
R
ec
eiv
ed
Mar
5
,
2
0
2
6
R
ev
is
ed
May
1
,
2
0
2
6
Acc
ep
ted
J
u
l
22
,
2
0
2
6
In
ter
-
tu
r
n
sh
o
r
t
c
ircu
it
(IT
S
C)
issu
e
s
a
re
a
c
o
m
m
o
n
e
lec
tri
c
a
l
fa
il
u
re
m
a
in
ly
c
a
u
se
d
b
y
t
h
e
d
e
teri
o
ra
ti
o
n
o
f
wi
n
d
i
n
g
in
s
u
latio
n
i
n
th
e
m
a
c
h
i
n
e
o
v
e
r
t
ime
.
F
a
il
in
g
to
d
e
tec
t
su
c
h
issu
e
s
e
a
rly
c
a
n
lea
d
to
c
a
tas
tro
p
h
ic
c
o
n
s
e
q
u
e
n
c
e
s.
Th
is
a
rti
c
le
in
v
e
stig
a
tes
th
e
in
tert
u
rn
fa
u
l
t
in
th
e
sta
to
r
wi
n
d
i
n
g
o
f
a
d
o
u
b
ly
-
fe
d
in
d
u
c
ti
o
n
g
e
n
e
ra
to
r
(D
F
IG
)
u
se
d
i
n
win
d
t
u
rb
i
n
e
s.
A
fl
u
x
li
n
k
a
g
e
d
iffere
n
c
e
v
e
c
to
r
(
F
LDV)
m
o
d
e
l
is
in
tr
o
d
u
c
e
d
i
n
t
h
is
stu
d
y
fo
r
fa
u
l
t
d
e
tec
ti
o
n
.
A
d
d
it
i
o
n
a
l
ly
,
a
n
a
rti
ficia
l
n
e
u
ra
l
n
e
two
rk
(AN
N)
m
o
d
e
l
is
p
ro
p
o
se
d
to
c
las
sify
th
e
se
fa
u
lt
s.
S
p
e
c
ifi
c
a
ll
y
,
a
sh
o
rt
-
c
ircu
i
t
fa
u
lt
i
s
in
d
u
c
e
d
in
e
a
c
h
p
h
a
se
o
f
th
e
sta
t
o
r
wi
n
d
i
n
g
,
a
n
d
th
e
fa
u
lt
s
a
re
c
las
sified
b
y
a
ss
ig
n
in
g
d
iffere
n
t
m
a
g
n
it
u
d
e
s to
t
h
e
re
sp
e
c
ti
v
e
p
h
a
se
s.
Th
e
AN
N i
s tr
a
in
e
d
to
id
e
n
ti
fy
wh
ich
p
h
a
se
c
o
n
tai
n
s
a
n
in
tert
u
rn
fa
u
lt
,
wi
th
o
u
tp
u
t
wa
v
e
fo
rm
a
m
p
li
tu
d
e
s
o
f
1
,
2
,
o
r
3
c
o
rre
sp
o
n
d
in
g
t
o
fa
u
lt
s
in
p
h
a
se
s
"
a
,
"
"
b
,
"
a
n
d
"
c
.
"
If
n
o
fa
u
lt
is
p
re
se
n
t,
t
h
e
wa
v
e
fo
rm
m
a
g
n
it
u
d
e
is
d
e
sig
n
a
ted
a
s
"
0
.
"
Th
is
a
p
p
r
o
a
c
h
e
n
a
b
les
e
a
rly
fa
u
lt
d
iag
n
o
sis
b
y
a
n
a
ly
z
in
g
wa
v
e
fo
rm
p
a
tt
e
rn
s
,
th
e
re
b
y
p
re
v
e
n
ti
n
g
o
v
e
rh
e
a
ti
n
g
c
a
u
se
d
b
y
s
h
o
rt
c
ircu
it
s
a
n
d
a
v
o
id
i
n
g
se
v
e
re
,
irrev
e
rsib
le d
a
m
a
g
e
to
th
e
wi
n
d
i
n
g
s.
K
ey
w
o
r
d
s
:
Ar
tific
ial
n
eu
r
al
n
etwo
r
k
Do
u
b
ly
-
f
e
d
in
d
u
ctio
n
g
en
er
at
o
r
I
n
ter
-
tu
r
n
s
h
o
r
t c
ir
c
u
it
f
au
lt
Simu
latio
n
Stato
r
cu
r
r
en
t
T
h
is i
s
a
n
o
p
e
n
a
c
c
e
ss
a
rticle
u
n
d
e
r th
e
CC B
Y
-
SA
li
c
e
n
se
.
C
o
r
r
e
s
p
o
nd
ing
A
uth
o
r
:
Viv
ek
Ku
s
h
wah
a
Dep
ar
tm
en
t o
f
E
lectr
ical
E
n
g
i
n
ee
r
in
g
,
R
esear
ch
Sch
o
lar
,
G
L
A
Un
iv
er
s
ity
1
7
k
m
Sto
n
e
,
NH
-
1
9
,
Ma
th
u
r
a
-
Delh
i Ro
ad
,
C
h
au
m
u
h
an
,
B
h
ar
th
ia,
Uttar
Pra
d
esh
2
8
1
4
0
6
,
I
n
d
ia
E
m
ail:
v
iv
ek
k
u
s
h
wa
h
a@
g
g
its
.
o
r
g
1.
I
NT
RO
D
UCT
I
O
N
T
h
e
escalatin
g
g
lo
b
al
en
er
g
y
d
em
an
d
an
d
r
is
in
g
en
v
ir
o
n
m
e
n
tal
wo
r
r
ies
h
av
e
h
asten
ed
th
e
s
h
if
t
to
r
en
ewa
b
le
en
er
g
y
s
o
u
r
ce
s
.
A
m
o
n
g
th
ese,
win
d
a
n
d
s
o
lar
e
n
er
g
y
h
av
e
em
er
g
ed
as
th
e
m
o
s
t
wid
ely
ad
o
p
ted
o
p
tio
n
s
b
ec
au
s
e
o
f
th
eir
s
u
s
tai
n
ab
ilit
y
an
d
av
ailab
ilit
y
.
W
in
d
en
er
g
y
c
o
n
v
e
r
s
io
n
s
y
s
tem
s
(
W
E
C
S)
ar
e
cr
u
cial
f
o
r
co
n
tem
p
o
r
ar
y
elec
tr
icity
g
en
er
atio
n
,
with
d
o
u
b
ly
f
ed
in
d
u
ctio
n
g
en
e
r
ato
r
s
(
DFI
Gs)
b
ein
g
wid
ely
u
tili
ze
d
b
ec
au
s
e
o
f
th
eir
h
ig
h
ef
f
icien
c
y
,
ab
ilit
y
to
o
p
er
ate
at
v
a
r
iab
l
e
s
p
ee
d
s
,
an
d
in
d
ep
en
d
en
t
m
a
n
ag
em
en
t
o
f
ac
tiv
e
an
d
r
ea
ctiv
e
p
o
we
r
.
E
v
en
wit
h
th
ese
b
en
ef
its
,
DFI
Gs
ar
e
s
u
s
ce
p
tib
le
to
d
if
f
er
en
t
ty
p
es
o
f
elec
tr
ical
f
au
lts
,
with
in
ter
tu
r
n
s
h
o
r
t
cir
c
u
it
(
I
T
SC
)
f
au
lts
in
s
tato
r
win
d
in
g
s
b
ein
g
esp
ec
ially
s
er
io
u
s
.
I
T
SC
f
au
lts
lead
to
a
r
is
e
in
lo
ca
lized
cu
r
r
en
t,
o
v
er
h
ea
tin
g
,
d
eter
io
r
atio
n
o
f
in
s
u
l
atio
n
,
an
d
u
ltima
tely
m
ac
h
in
e
f
ailu
r
e
if
th
ey
ar
e
n
o
t
ca
u
g
h
t
ea
r
ly
o
n
.
Hen
ce
,
t
o
g
u
ar
an
tee
d
e
p
en
d
a
b
le
f
u
n
ct
io
n
in
g
,
lo
wer
u
p
k
ee
p
e
x
p
en
s
e
s
,
an
d
av
er
t
p
o
wer
lo
s
s
,
it is
cr
u
cial
to
d
etec
t I
T
SC
f
au
lts
ac
cu
r
ately
an
d
p
r
o
m
p
tly
.
T
h
e
liter
atu
r
e
h
as
p
u
t
f
o
r
t
h
s
e
v
er
al
m
eth
o
d
s
f
o
r
id
en
tify
i
n
g
an
d
d
iag
n
o
s
in
g
I
T
SC
d
e
f
ec
ts
in
DFI
Gs.
I
n
[
1
]
,
a
ze
r
o
-
s
eq
u
en
ce
v
o
ltag
e
-
b
ased
m
eth
o
d
is
in
tr
o
d
u
ce
d
,
u
s
in
g
MA
T
L
AB
-
b
ased
m
o
d
ellin
g
an
d
s
im
u
latio
n
to
f
ac
ilit
ate
f
au
lt d
etec
tio
n
a
n
d
lo
ca
lizatio
n
.
A
1
0
0
k
W
DFI
G
win
d
tu
r
b
in
e'
s
is
s
u
e
is
d
iag
n
o
s
ed
v
ia
co
r
r
elatio
n
an
aly
s
is
b
etwe
en
v
ib
r
atio
n
s
i
g
n
als
an
d
ex
ter
n
al
m
ag
n
etic
s
ig
n
als
in
[
2
]
.
I
n
[
3
]
,
s
tato
r
an
d
r
o
to
r
f
laws
ar
e
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
0
8
8
-
8
7
0
8
I
n
t J E
lec
&
C
o
m
p
E
n
g
,
Vo
l.
16
,
No
.
5
,
Octo
b
e
r
20
26
:
2
3
7
9
-
2
3
9
2
2380
ex
am
in
ed
u
n
d
er
co
n
tr
o
lled
c
o
n
d
itio
n
s
u
tili
zin
g
a
5
k
W
test
eq
u
ip
m
en
t
f
o
r
e
x
p
er
im
e
n
tal
v
alid
atio
n
.
T
h
er
e
h
as
also
b
ee
n
m
u
c
h
r
esear
ch
d
o
n
e
o
n
ad
v
an
ce
d
s
ig
n
al
p
r
o
ce
s
s
in
g
m
eth
o
d
s
.
I
n
[
4
]
,
t
h
e
n
o
tio
n
o
f
m
a
g
n
etic
f
lu
x
leak
ag
e
(
MFL
)
is
u
s
ed
to
id
e
n
tify
ea
r
ly
d
e
f
ec
ts
.
I
n
[
5
]
,
s
tato
r
in
ter
tu
r
n
d
ef
ec
ts
ar
e
id
en
tifie
d
an
d
th
ei
r
s
ev
er
ity
ass
es
s
ed
u
s
in
g
m
o
to
r
cu
r
r
en
t
s
ig
n
atu
r
e
an
aly
s
is
(
MCS
A)
in
co
n
j
u
n
ctio
n
with
m
u
lti
-
s
ca
le
Par
k
'
s
v
ec
to
r
ap
p
r
o
ac
h
.
Fo
r
a
u
to
m
ated
DFI
G
test
in
g
,
th
e
in
v
er
ted
em
b
ed
d
ed
o
f
f
lin
e
test
is
u
s
ed
in
[
6
]
.
T
h
e
s
u
g
g
ested
tech
n
iq
u
e
d
etec
ts
an
d
ca
teg
o
r
izes
th
e
d
ef
ec
ts
in
th
e
s
tato
r
co
r
e
in
ter
lam
in
ar
in
s
u
latio
n
,
r
o
to
r
win
d
in
g
tu
r
n
in
s
u
latio
n
,
an
d
s
lip
r
in
g
b
r
u
s
h
co
n
tact.
A
m
icr
o
m
ac
h
in
e
an
d
a
s
tan
d
ar
d
litt
le
s
to
r
e
-
b
o
u
g
h
t
wo
u
n
d
r
o
to
r
in
d
u
ctio
n
m
ac
h
i
n
e
(
W
R
I
M)
ar
e
co
m
p
ar
ed
in
[
7
]
.
I
n
th
is
s
tu
d
y
,
r
o
to
r
in
ter
tu
r
n
f
au
lts
in
lar
g
e
-
s
ca
le
an
d
m
icr
o
m
ac
h
in
es
ar
e
co
m
p
ar
ed
.
B
ec
au
s
e
m
icr
o
m
ac
h
in
es
ar
e
p
r
o
p
o
r
tio
n
ate
r
ep
licas
o
f
b
ig
g
er
m
ac
h
in
es,
th
ey
o
f
f
er
an
id
ea
l
r
e
p
r
esen
tatio
n
o
f
wh
at
is
e
x
p
ec
ted
u
n
d
er
th
e
t
r
an
s
ito
r
y
f
u
n
ctio
n
i
n
g
c
o
n
d
itio
n
s
o
f
win
d
tu
r
b
in
es.
T
h
is
wo
u
ld
f
ac
ilit
ate
th
e
d
ev
elo
p
m
en
t
o
f
co
n
d
itio
n
m
o
n
ito
r
in
g
s
o
lu
tio
n
s
s
u
itab
le
f
o
r
th
e
in
d
u
s
tr
y
'
s
r
ea
l
s
y
s
tem
s
.
W
ei
et
a
l.
[
8
]
s
u
g
g
ests
th
e
en
lar
g
ed
Par
k
'
s
v
e
cto
r
m
eth
o
d
,
wh
ic
h
ca
lcu
lates
th
e
m
ath
em
atica
l
r
elatio
n
s
h
ip
b
etwe
en
u
n
b
alan
c
ed
v
o
ltag
e,
l
o
ad
v
a
r
iatio
n
,
an
d
th
e
n
eg
ativ
e
s
eq
u
e
n
ce
cu
r
r
en
t
in
th
e
s
tato
r
u
s
in
g
th
e
Fry
ze
-
B
u
ch
h
o
lz
-
Dp
en
b
r
o
c
k
alg
o
r
ith
m
an
d
t
h
e
n
o
n
-
lin
ea
r
least sq
u
ar
es a
p
p
r
o
ac
h
.
A
m
o
d
el
b
ased
o
n
th
e
f
in
ite
elem
en
t m
eth
o
d
(
FEM
)
is
d
ev
el
o
p
ed
in
[
9
]
t
o
co
m
p
r
eh
e
n
d
th
e
m
ac
h
in
e'
s
b
eh
av
io
r
in
b
o
th
h
ea
lth
y
an
d
I
T
SC
f
au
lt
co
n
d
itio
n
s
i
n
th
e
r
o
to
r
win
d
in
g
.
Attallah
et
a
l.
[
1
0
]
p
r
o
p
o
s
es
d
ee
p
lear
n
in
g
-
b
ased
co
n
v
o
lu
tio
n
al
n
eu
r
al
n
etwo
r
k
s
(
C
NNs)
with
n
o
n
-
in
tr
u
s
iv
e
in
f
r
a
r
ed
im
a
g
in
g
f
o
r
p
r
ec
is
e
d
ef
ec
t
lo
ca
lizatio
n
an
d
class
if
icatio
n
.
I
n
I
T
SC
f
au
lt
d
etec
tio
n
,
m
o
d
el
-
b
ased
a
n
d
o
b
s
er
v
e
r
-
b
ase
d
ap
p
r
o
ac
h
es
h
av
e
d
em
o
n
s
tr
ated
e
n
co
u
r
ag
in
g
o
u
t
co
m
es.
A
n
ew
f
a
u
lt
f
ea
tu
r
e
f
o
r
p
r
ec
is
e
d
iag
n
o
s
is
is
o
f
f
er
ed
b
y
th
e
b
ilater
al
f
lu
x
lin
k
ag
e
d
if
f
e
r
en
ce
v
ec
t
o
r
(
FL
DV)
ap
p
r
o
ac
h
,
wh
ich
was
f
ir
s
t
p
r
esen
ted
in
[
1
1
]
.
I
n
[
1
2
]
,
m
o
d
els
b
ased
o
n
th
e
FEM
ar
e
cr
ea
ted
t
o
ex
a
m
in
e
m
ac
h
in
e
b
eh
av
i
o
r
in
b
o
t
h
b
r
o
k
e
n
an
d
h
ea
lth
y
s
ce
n
ar
io
s
.
Fo
r
DFI
G
win
d
tu
r
b
in
es,
a
f
au
lt
r
id
e
-
th
r
o
u
g
h
tech
n
iq
u
e
is
s
u
g
g
ested
in
[
1
3
]
f
o
r
s
tato
r
in
ter
tu
r
n
s
h
o
r
t
cir
cu
it
f
au
lts
.
T
h
e
s
u
g
g
ested
co
n
ce
p
t
r
ed
u
ce
s
m
ain
ten
an
ce
wh
e
n
an
I
T
SC
f
a
ilu
r
e
o
cc
u
r
s
,
in
cr
ea
s
es
o
p
er
at
io
n
ca
p
ac
ity
,
an
d
p
r
ev
en
ts
p
o
wer
lo
s
s
f
r
o
m
s
h
u
td
o
wn
.
I
n
[
1
4
]
,
a
r
o
b
u
s
t
m
eth
o
d
f
o
r
esti
m
atin
g
r
o
t
o
r
I
T
SC
d
ef
ec
ts
is
d
ev
is
ed
u
s
in
g
an
H
-
in
f
in
ity
o
b
s
er
v
e
r
.
I
n
[
1
5
]
,
an
ex
p
er
im
en
tal
s
y
s
tem
is
cr
ea
ted
to
m
im
ic
th
e
r
ea
l
o
p
er
atio
n
o
f
a
wo
u
n
d
r
o
to
r
in
d
u
ctio
n
m
ac
h
in
e
(
W
R
I
M)
.
T
h
e
s
tato
r
I
T
SC
F
is
d
etec
ted
u
s
in
g
em
p
i
r
ical
m
o
d
e
d
ec
o
m
p
o
s
itio
n
(
E
MD
)
an
d
tim
e
d
o
m
ain
an
al
y
s
is
.
Ad
d
itio
n
ally
,
co
m
p
ar
ed
to
tr
ad
itio
n
al
p
r
o
ce
d
u
r
es,
d
ata
-
d
r
iv
en
ap
p
r
o
ac
h
es
lik
e
lo
n
g
s
h
o
r
t
-
ter
m
m
em
o
r
y
(
L
STM
)
n
etwo
r
k
s
in
[
1
6
]
s
h
o
w
b
etter
d
iag
n
o
s
tic
ac
cu
r
ac
y
a
n
d
q
u
ick
er
r
ea
ctio
n
.
Fo
r
f
au
lt
id
en
tific
atio
n
,
s
tatis
tical
an
aly
s
is
,
an
d
f
r
eq
u
en
c
y
-
d
o
m
ain
tech
n
iq
u
es
ar
e
also
f
r
e
q
u
en
tly
em
p
l
o
y
ed
.
An
ex
p
er
im
en
tal
s
etu
p
h
as
b
e
en
co
n
s
tr
u
cted
in
[
1
7
]
to
id
en
t
if
y
th
e
in
ter
-
tu
r
n
f
a
u
lt
(
ITF
)
in
th
e
r
o
to
r
tu
r
n
s
o
f
a
DFI
G
with
a
m
ag
n
etic
f
lu
x
m
ea
s
u
r
in
g
co
il
a
n
ten
n
a.
I
t
s
u
g
g
ests
th
at,
to
id
en
tify
an
I
T
SC
p
r
o
b
lem
in
DFI
G,
s
ea
r
ch
co
il
v
o
ltag
e
at
v
ar
io
u
s
f
r
eq
u
e
n
cies
is
a
m
o
r
e
ef
f
icien
t
an
d
s
o
p
h
is
ticated
m
eth
o
d
th
a
n
cu
r
r
e
n
t
s
ig
n
atu
r
e
an
aly
s
is
.
Fau
lt
-
r
elate
d
f
ea
tu
r
es
ca
n
b
e
ef
f
ec
tiv
ely
ex
tr
ac
ted
u
s
in
g
m
eth
o
d
s
lik
e
p
o
wer
s
p
e
ctr
al
d
en
s
ity
(
PS
D)
,
tim
e
s
y
n
ch
r
o
n
o
u
s
av
er
a
g
in
g
(
T
SA)
[
1
8
]
,
Kalm
a
n
f
ilter
in
g
[
1
9
]
,
m
ax
im
u
m
p
o
wer
p
o
in
t
tr
ac
k
in
g
(
MPPT)
-
b
ased
d
etec
tio
n
[
2
0
]
,
an
d
T
SA,
as we
ll a
s
C
S
A
tech
n
iq
u
es a
r
e
ap
p
lied
in
[
2
1
]
to
d
etec
t
f
au
l
ts
.
I
n
[
2
2
]
,
an
I
T
SC
f
au
lt
b
etwe
en
r
o
to
r
an
d
s
tato
r
is
h
y
p
o
th
esized
.
Her
e,
th
e
b
est
ac
ti
v
e
p
o
wer
p
r
o
d
u
ctio
n
f
o
r
v
a
r
y
in
g
win
d
s
p
ee
d
s
is
en
s
u
r
ed
b
y
ap
p
ly
in
g
m
a
x
im
u
m
p
o
wer
p
o
i
n
t
tr
ac
in
g
(
MPPT)
m
an
ag
em
en
t.
W
in
d
in
g
f
u
n
ctio
n
a
n
aly
s
is
(
W
FA)
,
d
q
0
tr
an
s
f
o
r
m
atio
n
,
an
d
m
u
lti
-
cir
cu
it
t
h
eo
r
y
ar
e
ex
am
p
les
o
f
ad
v
an
ce
d
m
o
d
ellin
g
tech
n
iq
u
es
th
at
h
a
v
e
b
ee
n
u
s
ed
to
en
h
an
ce
f
au
lt
an
al
y
s
is
an
d
co
m
p
r
eh
en
s
io
n
[
2
3
]
.
Ad
d
itio
n
ally
,
m
ea
s
u
r
es
in
clu
d
in
g
R
MS
cu
r
r
en
t,
to
tal
h
ar
m
o
n
ic
d
is
to
r
tio
n
(
T
HD)
,
s
eq
u
e
n
c
e
co
m
p
o
n
en
ts
,
an
d
p
h
ase
d
if
f
er
e
n
ce
s
ar
e
u
s
ed
in
an
aly
tical
an
d
ex
p
er
im
en
tal
r
e
s
ea
r
ch
to
d
etec
t
I
T
SC
f
au
lts
[
2
4
]
–
[
2
6
]
.
T
h
e
Fas
t
Fo
u
r
ier
T
r
a
n
s
f
o
r
m
is
s
u
g
g
ested
in
[
2
7
]
,
an
d
m
u
lti
-
cir
c
u
it
th
e
o
r
y
is
s
u
g
g
ested
in
[
2
8
]
,
an
d
t
h
e
o
u
tc
o
m
es
o
f
th
e
s
u
g
g
ested
ex
p
er
im
en
tal
m
o
d
el
ar
e
co
n
f
ir
m
e
d
f
o
r
I
T
S
C
f
au
lt
d
etec
tio
n
.
Hig
h
ac
cu
r
ac
y
,
ea
r
ly
-
s
tag
e
id
en
tific
atio
n
,
an
d
r
ea
l
-
tim
e
ap
p
licatio
n
co
n
tin
u
e
to
b
e
h
u
r
d
les
d
esp
ite
th
e
d
e
v
elo
p
m
en
t
o
f
n
u
m
er
o
u
s
ap
p
r
o
ac
h
es.
Fo
r
I
T
SC
f
au
lt
d
etec
tio
n
in
DFI
G
-
b
ased
win
d
en
er
g
y
s
y
s
tem
s
,
m
o
r
e
ef
f
ec
tiv
e
an
d
d
ep
en
d
ab
le
d
iag
n
o
s
tic
m
eth
o
d
s
ar
e
th
er
ef
o
r
e
r
eq
u
ir
ed
.
I
T
SC
f
au
lt
d
etec
tio
n
in
DFI
G
-
b
ased
win
d
en
e
r
g
y
s
y
s
tem
s
h
as
b
ee
n
ex
ten
s
iv
ely
s
tu
d
ie
d
,
b
u
t
th
er
e
ar
e
s
till
a
f
ew
im
p
o
r
tan
t
is
s
u
es
th
at
n
ee
d
to
b
e
ad
d
r
ess
ed
.
Fo
r
ex
am
p
le,
m
an
y
o
f
th
e
c
u
r
r
e
n
t
m
eth
o
d
s
r
ely
o
n
in
tr
u
s
iv
e
m
ea
s
u
r
em
en
t
tech
n
i
q
u
es,
ex
p
e
n
s
iv
e
s
en
s
o
r
s
,
o
r
c
o
m
p
lex
m
at
h
em
atica
l
m
o
d
elli
n
g
,
wh
ich
r
estricts
th
eir
p
r
ac
tical
im
p
lem
en
tatio
n
in
r
ea
l
-
tim
e
ap
p
licatio
n
s
.
Sig
n
al
p
r
o
ce
s
s
in
g
ap
p
r
o
ac
h
es
lik
e
FF
T
,
C
NN,
SVM,
an
d
MCS
A
f
r
eq
u
en
tly
f
ail
to
d
etec
t
in
cip
ien
t
f
au
lts
u
n
d
er
v
ar
y
in
g
lo
ad
a
n
d
n
o
is
e
co
n
d
i
tio
n
s
,
an
d
th
e
y
ar
e
co
s
tly
to
o
,
wh
ile
m
o
d
el
-
b
ase
d
an
d
o
b
s
er
v
er
-
b
ased
tech
n
i
q
u
es
ar
e
s
en
s
itiv
e
to
p
ar
am
eter
v
ar
iatio
n
s
an
d
m
o
d
ellin
g
er
r
o
r
s
.
Mo
r
e
o
v
er
,
ev
en
wh
ile
d
ata
-
d
r
iv
en
m
eth
o
d
s
lik
e
d
ee
p
lear
n
in
g
h
av
e
d
em
o
n
s
tr
ated
en
co
u
r
a
g
in
g
o
u
tco
m
es,
th
ey
f
r
eq
u
e
n
tly
n
ee
d
b
ig
d
atasets
,
a
lo
t
o
f
p
r
o
c
ess
in
g
p
o
wer
,
an
d
ar
e
d
if
f
icu
lt
to
u
n
d
er
s
tan
d
.
As
a
r
esu
lt,
a
r
elia
b
le,
p
r
ec
is
e,
a
n
d
co
m
p
u
tatio
n
a
lly
ef
f
ec
tiv
e
tec
h
n
iq
u
e
f
o
r
ea
r
l
y
-
s
tag
e
I
T
SC
f
au
lt
d
etec
tio
n
u
n
d
er
v
ar
i
o
u
s
o
p
er
a
tin
g
s
itu
atio
n
s
is
r
e
q
u
ir
ed
.
T
h
is
p
ap
er
u
s
es
MA
T
L
AB
/Si
m
u
lin
k
m
o
d
ellin
g
an
d
in
tellig
en
t
d
ata
an
aly
s
is
tech
n
iq
u
es
to
o
f
f
er
a
n
ef
f
icien
t
I
T
SC
f
au
lt
d
etec
tio
n
an
d
an
al
y
s
is
f
r
am
ewo
r
k
f
o
r
DFI
G
to
o
v
er
co
m
e
th
ese
is
s
u
es.
T
h
is
p
ap
er
h
as
a
d
etailed
MA
T
L
AB
/S
im
u
lin
k
m
o
d
el
o
f
a
DFI
G
s
y
s
tem
,
wh
ich
ac
cu
r
ately
r
ep
r
esen
ts
t
h
e
d
y
n
am
ics
o
f
s
tato
r
f
au
lts
u
n
d
er
b
o
th
h
ea
lth
y
an
d
I
T
SC
f
au
lt
s
tates.
Stato
r
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J E
lec
&
C
o
m
p
E
n
g
I
SS
N:
2088
-
8
7
0
8
S
ta
to
r
in
tert
u
r
n
s
h
o
r
t c
ir
cu
it
fa
u
lt id
en
tifi
ca
tio
n
in
d
o
u
b
ly
-
f
ed
in
d
u
ctio
n
g
e
n
era
to
r
…
(
V
iv
ek
K
u
s
h
w
a
h
a
)
2381
cu
r
r
en
t,
v
o
ltag
e,
f
lu
x
lin
k
a
g
e
d
if
f
er
en
ce
(
FLD)
,
an
d
FLDV
ar
e
ex
tr
ac
ted
an
d
an
aly
ze
d
f
o
r
f
au
lt
id
en
tific
atio
n
.
T
o
class
if
y
h
ea
lth
y
a
n
d
f
au
lty
co
n
d
itio
n
s
with
im
p
r
o
v
e
d
ac
c
u
r
ac
y
a
n
d
r
eliab
ilit
y
,
an
in
telli
g
en
t
f
au
lt d
etec
tio
n
ap
p
r
o
ac
h
(
e.
g
.
,
ar
tific
ial
n
eu
r
al
n
etwo
r
k
(
ANN)
)
is
p
r
o
p
o
s
ed
.
E
f
f
ec
tiv
en
ess
o
f
th
e
p
r
o
p
o
s
ed
m
eth
o
d
is
an
aly
ze
d
an
d
co
m
p
ar
ed
with
o
th
er
co
n
v
en
tio
n
al
tech
n
i
q
u
es
in
ter
m
s
o
f
ac
cu
r
ac
y
,
r
o
b
u
s
tn
ess
,
an
d
co
m
p
u
tatio
n
al
ef
f
icien
c
y
t
o
d
e
tect
in
cip
ien
t
f
au
lt
in
ea
r
l
y
s
ta
g
e,
wh
ich
r
e
d
u
ce
s
th
e
r
is
k
o
f
s
ev
er
e
d
am
a
g
e
a
n
d
im
p
r
o
v
es sy
s
tem
r
eliab
ilit
y
.
2.
DF
I
G
SI
M
UL
AT
I
O
N
M
O
D
E
L
As
w
e
k
n
o
w
,
to
i
n
cr
ea
s
e
th
e
l
if
e
o
f
a
m
ac
h
i
n
e
,
o
n
e
s
h
o
u
l
d
av
o
i
d
t
h
e
f
a
u
lts
an
d
a
ls
o
n
e
e
d
s
t
o
d
ete
ct
an
d
r
em
o
v
e
t
h
e
f
a
u
lt
as
ea
r
l
y
as
p
o
s
s
ib
le
.
F
ig
u
r
e
1
s
h
o
ws
a
s
i
m
u
la
ti
o
n
m
o
d
el
o
f
DF
I
G
wit
h
a
wi
n
d
t
u
r
b
in
e
d
e
v
e
lo
p
ed
,
an
d
I
T
F
is
cr
ea
t
e
d
i
n
a
ll t
h
r
e
e
p
h
ases
o
f
s
t
at
o
r
wi
n
d
i
n
g
.
S
h
o
r
t c
ir
c
u
its
ar
e
c
r
ea
t
ed
in
p
h
as
es a
,
b
,
a
n
d
c
s
im
u
l
ta
n
e
o
u
s
l
y
f
o
r
d
i
f
f
er
en
t
t
im
e
p
er
io
d
s
.
A
f
te
r
t
h
at,
a
n
NN
m
o
d
el
is
p
r
o
p
o
s
ed
a
n
d
tr
ain
e
d
t
o
i
d
e
n
ti
f
y
t
h
e
f
a
u
lts
i
n
d
if
f
e
r
e
n
t
p
h
ases
.
As
t
h
e
m
o
d
e
l is
t
r
a
in
ed
,
it
m
a
y
d
e
t
er
m
i
n
e
t
h
e
p
h
ase
i
n
w
h
i
c
h
a
s
h
o
r
t
ci
r
c
u
it
o
cc
u
r
s
.
Fig
u
r
e
1
.
Sch
em
atic
d
iag
r
am
o
f
s
im
u
latio
n
m
o
d
el
3.
I
NT
E
R
T
URN
SH
O
R
T
CIR
CUIT
F
A
UL
T
An
I
T
SC
d
ef
ec
t
is
ch
ar
ac
ter
i
ze
d
b
y
th
e
s
h
o
r
t
-
cir
c
u
itin
g
o
f
2
o
r
m
o
r
e
tu
r
n
s
in
th
e
s
tato
r
o
r
r
o
to
r
win
d
in
g
.
B
ec
au
s
e
o
f
th
is
f
au
lt
,
s
u
d
d
en
ly
,
t
h
e
cu
r
r
en
t
in
cr
ea
s
es
in
th
e
p
h
ase
wh
er
e
th
e
f
a
u
lt
o
cc
u
r
r
e
d
,
wh
ich
in
cr
ea
s
es th
e
tem
p
er
atu
r
e
in
th
e
win
d
in
g
.
Du
e
to
t
h
e
co
n
tin
u
o
u
s
h
ea
tin
g
in
th
e
win
d
in
g
f
o
r
a
lo
n
g
tim
e,
it m
ay
d
am
ag
e
th
e
wh
o
le
win
d
in
g
.
T
h
er
ef
o
r
e
,
to
s
av
e
th
e
m
ac
h
in
e
f
r
o
m
s
ev
er
e
d
am
a
g
e,
it is
r
eq
u
ir
ed
to
d
iag
n
o
s
e
th
e
f
au
lt
at
an
ea
r
ly
s
tag
e
an
d
tak
e
n
ec
ess
ar
y
ac
tio
n
.
T
h
e
n
e
u
r
al
n
etwo
r
k
to
o
l
in
MA
T
L
AB
Simu
lin
k
is
ap
p
lied
to
d
iag
n
o
s
e
th
e
in
te
r
tu
r
n
f
au
lt in
th
e
s
tato
r
win
d
in
g
.
3
.
1
.
M
a
t
hema
t
ica
l
m
o
dellin
g
o
f
I
T
F
I
n
th
is
ar
ticle,
f
au
lts
ar
e
cr
ea
ted
in
all
th
r
ee
p
h
ases
f
o
r
d
if
f
er
en
t
tim
e
p
e
r
io
d
s
.
I
n
Fi
g
u
r
e
2
,
t
h
e
in
ter
tu
r
n
f
a
u
lt
in
p
h
ase
‘
a’
is
cr
ea
ted
b
y
s
h
o
r
t
-
ci
r
cu
itin
g
th
e
p
h
ase
win
d
in
g
.
Similar
ly
,
an
I
T
F
ca
n
b
e
cr
ea
ted
in
ev
er
y
p
h
ase.
Fig
u
r
e
2
.
Sh
o
r
t c
ir
cu
it d
ia
g
r
a
m
o
f
s
tato
r
win
d
i
n
g
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
0
8
8
-
8
7
0
8
I
n
t J E
lec
&
C
o
m
p
E
n
g
,
Vo
l.
16
,
No
.
5
,
Octo
b
e
r
20
26
:
2
3
7
9
-
2
3
9
2
2382
T
u
r
n
s
in
s
tato
r
w
in
d
i
n
g
,
N
s
=
4
8
,
T
u
r
n
s
in
r
o
to
r
w
in
d
in
g
,
N
r
=
3
2
T
u
r
n
s
r
atio
,
=
=
48
32
=
1
.
5
.
Nu
m
b
er
o
f
tu
r
n
s
s
h
o
r
t
-
cir
cu
ite
d
in
th
e
s
tato
r
win
d
in
g
,
T
h
e
n
u
m
b
er
o
f
tu
r
n
s
r
em
ain
s
h
ea
lth
y
in
th
e
s
tato
r
win
d
in
g
,
=
−
T
h
e
m
ag
n
etic
p
er
m
ea
n
ce
=
T
o
tal
I
n
d
u
ctan
ce
o
f
s
tato
r
win
d
in
g
,
=
2
I
n
d
u
ctan
ce
o
f
s
h
o
r
t
-
cir
cu
ited
tu
r
n
s
,
=
2
=
(
)
2
I
n
d
u
ctan
ce
o
f
r
em
ain
in
g
h
ea
lt
h
y
tu
r
n
s
,
=
(
−
)
2
=
(
−
)
2
Mu
tu
al
in
d
u
ctan
ce
am
o
n
g
s
h
o
r
ted
an
d
t
h
e
r
em
ain
i
n
g
win
d
i
n
g
in
th
e
s
am
e
p
h
ase.
=
(
−
)
=
(
−
)
2
Per
p
h
ase
in
d
u
cta
n
ce
m
atr
ix
is
=
[
]
R
esi
s
tan
ce
p
er
tu
r
n
=
R
esi
s
tan
ce
o
f
s
h
o
r
t
-
cir
c
u
ited
t
u
r
n
s
,
=
⋅
=
(
1
)
R
esi
s
tan
ce
to
r
em
ain
in
g
h
ea
lth
y
tu
r
n
s
=
(
−
)
=
−
(
2
)
,
,
an
d
ѱ
T
h
e
v
o
ltag
e,
cu
r
r
e
n
t,
an
d
f
lu
x
lin
k
a
g
e
m
atr
ices
ar
e
s
h
o
wn
in
(
3
)
an
d
(
4
)
.
W
h
er
e,
s
u
b
s
cr
ip
t
,
,
an
d
r
ep
r
esen
t
s
tato
r
,
r
o
to
r
,
an
d
f
au
lt,
r
esp
ec
tiv
ely
.
,
,
an
d
r
ep
r
esen
t
s
elf
-
i
n
d
u
ctan
ce
,
r
esis
tan
ce
,
an
d
m
u
tu
al
in
d
u
cta
n
ce
.
[
ѱФ
ѱ
ѱ
]
=
[
0
0
]
[
]
(
3
)
[
]
=
[
0
0
0
0
0
0
]
[
]
+
[
ѱ
ѱ
ѱ
]
(
4
)
=
[
µ
2
(
−
)
−
µ
2
−
µ
2
−
µ
2
(
−
)
−
µ
2
−
µ
2
−
µ
2
(
−
)
]
(
5
)
=
[
ζ
2
(
−
)
−
ζ
2
−
ζ
2
−
ζ
2
(
−
)
−
ζ
2
−
ζ
2
−
ζ
2
(
−
)
]
(
6
)
an
d
ar
e
th
e
s
elf
-
in
d
u
cta
n
ce
m
atr
ices
r
esp
ec
tiv
ely
f
o
r
th
e
s
tato
r
an
d
r
o
to
r
win
d
in
g
,
w
h
e
r
e
µ
an
d
ζ
ar
e
s
h
o
r
t
cir
cu
it
c
o
ef
f
ici
en
ts
,
r
esp
ec
tiv
ely
f
o
r
s
tato
r
an
d
r
o
t
o
r
win
d
in
g
s
.
µ
=
No
.
o
f
f
au
lt
tu
r
n
s
in
s
tato
r
win
d
in
g
/To
tal
No
.
s
tato
r
win
d
in
g
ζ
=
No
.
o
f
f
a
u
lt
tu
r
n
s
i
n
r
o
to
r
win
d
i
n
g
/To
tal
N
o
.
r
o
to
r
win
d
in
g
.
I
f
t
h
er
e
is
n
o
f
au
lt,
t
h
en
th
e
v
alu
e
o
f
th
e
s
h
o
r
t c
ir
cu
it c
o
ef
f
icien
t w
ill b
e
‘
0
’
.
=
[
µ
ζ
Ѳ
µ
c
os
(
Ѳ
+
120
0
)
µ
c
os
(
Ѳ
−
120
0
)
ζ
c
os
(
Ѳ
−
120
0
)
Ѳ
c
os
(
Ѳ
+
120
0
)
ζ
c
os
(
Ѳ
+
120
0
)
c
os
(
Ѳ
−
120
0
)
Ѳ
]
(
7
)
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J E
lec
&
C
o
m
p
E
n
g
I
SS
N:
2088
-
8
7
0
8
S
ta
to
r
in
tert
u
r
n
s
h
o
r
t c
ir
cu
it
fa
u
lt id
en
tifi
ca
tio
n
in
d
o
u
b
ly
-
f
ed
in
d
u
ctio
n
g
e
n
era
to
r
…
(
V
iv
ek
K
u
s
h
w
a
h
a
)
2383
=
Mu
tu
al
in
d
u
ctan
ce
m
atr
ix
b
etwe
en
s
tato
r
an
d
r
o
to
r
.
3
.
2
.
I
dentif
ica
t
io
n
m
et
ho
d ba
s
ed
o
n F
L
DV
Flu
x
lin
k
ag
es
in
DFI
G
ca
n
b
e
s
h
o
wn
as
in
t
h
e
f
o
llo
win
g
eq
u
atio
n
.
I
n
(
18
)
an
d
(
19
)
r
e
p
r
esen
t
th
e
cu
r
r
en
t a
n
d
v
o
ltag
e
m
o
d
el
to
g
et
th
e
f
lu
x
lin
k
ag
es f
o
r
DFI
G.
[
ѱ
]
=
[
ѱ
ѱ
]
=
[
]
[
]
(
8
)
[
ѱ
]
=
[
ѱ
ѱ
]
=
[
]
−
[
0
0
]
[
]
(
9
)
T
h
e
DFI
G
is
a
f
f
ec
ted
b
y
th
e
s
lip
r
atio
;
th
e
r
o
to
r
cu
r
r
en
t
f
au
lt
f
ea
tu
r
es
ar
e
d
im
in
is
h
ed
,
an
d
it
is
ch
allen
g
in
g
to
p
r
ec
is
ely
i
d
en
tify
th
e
r
o
to
r
s
id
e
I
T
F,
w
h
ich
ar
e
th
e
lim
itatio
n
s
o
f
th
e
ex
is
tin
g
d
iag
n
o
s
is
ap
p
r
o
ac
h
as
a
ch
ar
ac
ter
is
tic
q
u
an
tity
.
T
h
er
e
f
o
r
e,
a
n
ew
a
p
p
r
o
ac
h
is
p
r
o
p
o
s
ed
,
ch
ar
ac
t
er
ized
b
y
th
e
f
lu
x
lin
k
ag
e
an
d
th
e
FLDV
.
Du
e
t
o
th
e
f
l
u
ctu
atio
n
i
n
th
e
im
p
ed
an
ce
o
f
th
e
s
tato
r
an
d
r
o
to
r
win
d
in
g
b
e
f
o
r
e
an
d
af
ter
th
e
I
T
F,
th
e
o
b
s
er
v
ed
f
au
lt v
o
ltag
e
an
d
c
u
r
r
e
n
t a
r
e
in
co
r
p
o
r
ated
i
n
to
th
e
v
o
ltag
e
an
d
f
lu
x
eq
u
atio
n
.
T
h
u
s
,
th
e
f
lu
x
lin
k
ag
e
o
b
tain
ed
is
ѱ
=
ѱ
−
ѱ
(
1
0
)
ѱ
=
[
]
[
]
−
∫
(
[
]
−
[
0
0
]
[
]
)
(
1
1
)
T
h
e
eq
u
atio
n
r
ep
r
esen
ts
a
f
a
u
lt
ch
ar
ac
ter
is
tic
q
u
a
n
tity
.
ѱ
an
d
ѱ
ar
e
cu
r
r
en
t
a
n
d
v
o
ltag
e
m
o
d
els,
r
esp
ec
tiv
ely
,
an
d
th
eir
d
if
f
er
en
ce
r
ep
r
esen
ts
th
e
FLD.
£
=
ѱ
+
ѱ
+
ѱ
(
1
2
)
£
=
ѱ
+
ѱ
+
ѱ
(
1
3
)
£
s
an
d
£
ar
e
t
h
e
r
esp
ec
tiv
e
FLD
V
in
th
e
s
tato
r
a
n
d
r
o
to
r
win
d
in
g
.
T
h
e
s
tato
r
a
n
d
r
o
to
r
FLDV
s
r
o
tate
at
d
if
f
er
en
t f
r
eq
u
en
cies
b
ec
au
s
e
o
f
DFI
G'
s
asy
n
ch
r
o
n
o
u
s
m
o
to
r
f
ea
tu
r
es.
T
h
e
s
lip
r
atio
is
m
u
ltip
lied
b
y
th
e
r
o
to
r
FLDV
f
r
eq
u
en
cy
,
wh
ile
th
e
FLDV
f
r
eq
u
en
c
y
o
f
t
h
e
s
tato
r
is
a
p
o
wer
f
r
eq
u
en
cy
.
I
t
is
p
o
s
s
ib
le
to
d
eter
m
in
e
i
f
an
I
T
F h
ap
p
en
s
o
n
th
e
r
o
t
o
r
o
r
th
e
s
tato
r
u
s
in
g
th
is
f
r
e
q
u
en
c
y
.
3
.
3
.
F
L
DV
t
heo
ry
Ass
u
m
e
th
at
th
e
p
h
ase
‘
a’
o
f
th
e
s
tato
r
win
d
in
g
is
s
h
o
r
t
-
cir
cu
ited
an
d
its
s
h
o
r
t
cir
cu
it
f
ac
to
r
is
‘
µ’
.
(
1
1
)
r
ep
r
esen
ts
th
e
FLDV
ca
lcu
latio
n
f
o
r
m
u
la
u
n
d
er
f
au
lt c
o
n
d
itio
n
s
.
No
w,
ѱ
=
[
ѱ
ѱ
ѱ
]
=
ѱ
−
ѱ
.
=
[
]
[
]
−
∫
(
−
[
0
]
[
]
)
(
1
4
)
Su
f
f
ix
s
h
o
ws
f
au
lty
q
u
an
titi
es
with
DFI
G
p
ar
am
eter
s
i
n
(
1
4
)
.
C
alcu
latio
n
o
f
th
e
s
h
o
r
t
cir
cu
it
ad
d
itio
n
al
f
lu
x
lin
k
ag
e
in
th
e
f
au
lt
m
o
d
el
is
n
o
t
r
eq
u
ir
ed
,
as
th
er
e
is
n
o
ad
d
itio
n
al
ter
m
f
o
r
th
e
f
au
lt
in
th
e
n
o
r
m
al
m
ac
h
in
e
m
o
d
el.
T
h
e
v
alu
es
o
f
f
lu
x
lin
k
ag
e
o
b
tain
e
d
f
r
o
m
th
e
v
o
ltag
e
an
d
f
lu
x
lin
k
ag
e
eq
u
atio
n
s
b
y
a
ty
p
ical
DFI
G
ar
e
th
e
s
am
e.
ѱ
=
ѱ
−
ѱ
=
(
ѱ
−
ѱ
ℎ
)
−
(
ѱ
−
ѱ
ℎ
)
(
1
5
)
I
n
(
1
5
)
,
th
e
f
au
lty
a
n
d
h
ea
lth
y
DFI
G’
s
cu
r
r
en
t e
q
u
atio
n
r
ep
r
e
s
en
ts
th
e
ch
an
g
es to
th
e
f
lu
x
d
if
f
er
en
ce
.
T
h
is
allo
ws FLD
V
to
b
e
ca
lcu
lated
s
ep
ar
ately
.
ѱ
=
(
ѱ
−
ѱ
ℎ
)
ѱ
=
(
−
)
(
µ
2
−
2µ
)
+
µ
(
1
−
µ
)
+
µ
2
(
+
)
−
µ
[
Ѳ
+
(
Ѳ
+
120
0
)
+
(
Ѳ
−
120
0
)
]
(
1
6
)
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
0
8
8
-
8
7
0
8
I
n
t J E
lec
&
C
o
m
p
E
n
g
,
Vo
l.
16
,
No
.
5
,
Octo
b
e
r
20
26
:
2
3
7
9
-
2
3
9
2
2384
T
h
e
3
-
p
h
ase
u
n
b
alan
ce
d
r
o
to
r
cu
r
r
en
t c
a
n
b
e
n
eg
lecte
d
f
o
r
th
e
s
tato
r
p
h
ase
‘
a’
ter
m
I
T
F; th
er
ef
o
r
e,
th
e
last
ter
m
in
(
1
5
)
ca
n
b
e
ig
n
o
r
ed
.
ѱ
=
(
−
)
(
µ
2
−
2µ
)
+
µ
(
1
−
µ
)
+
µ
2
(
+
)
(
1
7
)
Similar
ly
,
th
e
f
lu
x
lin
k
ag
e
o
f
t
h
e
h
ea
lth
y
p
h
ase
o
f
s
tato
r
win
d
in
g
‘
b
’
an
d
‘
c’
is
g
iv
en
b
y
(
1
8
)
.
ѱ
=
ѱ
=
µ
2
(
+
)
(
1
8
)
ѱ
=
[
ѱ
ѱ
ѱ
]
=
µ
(
−
)
[
Ѳ
(
Ѳ
+
120
0
)
(
Ѳ
−
120
0
)
]
(
1
9
)
In
(
1
9
)
s
h
o
ws
th
e
f
lu
x
lin
k
ag
e
o
f
th
e
h
ea
lth
y
win
d
in
g
o
f
e
ac
h
p
h
ase
in
th
e
r
o
to
r
.
T
h
e
F
L
D
in
th
e
v
o
ltag
e
eq
u
atio
n
o
b
tain
ed
b
y
ѱ
=
ѱ
−
ѱ
=
µ
(
−
)
(
2
0
)
Ass
u
m
e
−
=
(
+
Ѳ
0
)
ѱ
=
−
µ
(
+
Ѳ
0
)
(
21)
B
y
u
s
in
g
(
1
7
)
,
(
1
8
)
,
(
1
9
)
an
d
(
2
1
)
,
s
ix
-
p
h
ase
FLDs
ca
n
b
e
f
o
u
n
d
;
also
(
1
1
)
an
d
(
1
2
)
g
iv
es FL
DV.
B
y
th
e
eq
u
atio
n
s
,
FLDs
o
f
3
-
p
h
ases
o
f
r
o
to
r
win
d
in
g
ar
e
s
y
m
m
etr
ical,
s
o
th
eir
s
u
m
(
£
)
is
ze
r
o
.
C
o
n
ce
r
n
in
g
th
e
s
tato
r
win
d
in
g
s
,
th
e
am
p
litu
d
e
o
f
f
au
lt
p
h
ase
‘
a’
d
if
f
er
s
s
ig
n
if
ican
tly
f
r
o
m
th
at
o
f
th
e
o
th
e
r
p
h
ases
.
Sin
ce
t
h
e
v
ec
to
r
s
u
m
o
f
an
d
is
d
ir
ec
te
d
o
p
p
o
s
ite
to
th
at
o
f
,
f
o
llo
ws
th
e
s
am
e
d
ir
ec
tio
n
as
.
Ph
as
e
a’
s
FLD
an
g
le
tr
ac
k
s
its
cu
r
r
en
t.
I
n
th
e
ca
s
e
o
f
a
n
I
T
F
in
th
e
s
t
ato
r
,
th
e
f
au
lt
cu
r
r
en
t
t
y
p
icall
y
ex
h
i
b
its
a
h
ig
h
e
r
am
p
litu
d
e
th
an
th
at
o
f
th
e
s
tato
r
cu
r
r
e
n
t.
C
o
n
s
eq
u
en
tly
,
th
e
FLD
in
th
e
o
th
er
two
p
h
ases
will
b
e
o
p
p
o
s
ite
to
th
at
o
f
p
h
ase
a.
No
w,
FLDV
o
f
th
e
s
tato
r
win
d
in
g
is
o
b
tain
e
d
as:
£
=
µ
(
µ
−
2
)
+
µ
(
+
Ѳ
0
)
+
µ
[
(
µ
−
5
2
)
+
(
3
2
−
µ
)
+
1
2
(
)
]
(
2
2
)
£
=
µ
(
µ
−
2
)
+
µ
(
+
Ѳ
0
)
+
µ
[
(
µ
−
5
2
)
+
(
3
2
−
µ
)
+
1
2
(
+
)
]
(
2
3
)
4.
Art
if
icia
l neura
l net
wo
rk
(
A
NN)
An
ANN
is
b
ased
o
n
th
e
c
o
m
p
o
s
itio
n
an
d
ca
p
ab
ilit
ies
o
f
th
e
h
u
m
an
b
r
ain
,
an
aly
ze
s
d
ata,
an
d
d
r
aws
co
n
clu
s
io
n
s
b
y
u
s
in
g
in
ter
c
o
n
n
ec
ted
n
o
d
es,
o
r
ar
tific
ial
n
eu
r
o
n
s
.
Neu
r
al
n
etwo
r
k
s
ar
e
h
elp
f
u
l
f
o
r
im
a
g
e
id
en
tific
atio
n
,
n
atu
r
al
la
n
g
u
a
g
e
p
r
o
ce
s
s
in
g
,
an
d
p
r
e
d
ictiv
e
a
n
aly
tics
ap
p
licatio
n
s
b
ec
a
u
s
e
o
f
th
eir
ca
p
ac
ity
to
lear
n
f
r
o
m
d
ata
a
n
d
s
p
o
t
p
at
ter
n
s
.
T
r
ain
in
g
d
ata
is
ess
en
tial
f
o
r
n
e
u
r
al
n
etwo
r
k
s
to
le
ar
n
an
d
g
r
a
d
u
ally
in
cr
ea
s
e
th
eir
ac
cu
r
ac
y
.
A
s
tan
d
ar
d
f
ee
d
-
f
o
r
war
d
ANN,
o
f
ten
ca
lled
a
m
u
ltil
ay
er
p
er
ce
p
tr
o
n
(
ML
P)
,
is
co
m
p
o
s
ed
o
f
an
i
n
p
u
t la
y
er
,
o
n
e
o
r
m
o
r
e
h
id
d
en
lay
er
s
,
an
d
an
o
u
tp
u
t la
y
er
,
as sh
o
w
n
in
F
ig
u
r
e
3
.
E
ac
h
lay
e
r
c
o
n
s
is
ts
o
f
in
ter
c
o
n
n
ec
ted
n
o
d
es
(
n
eu
r
o
n
s
)
t
h
at
p
r
o
ce
s
s
in
f
o
r
m
atio
n
u
s
in
g
n
o
n
-
lin
ea
r
ac
tiv
atio
n
f
u
n
ctio
n
s
,
en
a
b
lin
g
th
e
n
etwo
r
k
to
lear
n
co
m
p
l
ex
p
atter
n
s
f
r
o
m
d
ata.
T
h
e
lear
n
in
g
p
r
o
ce
s
s
is
g
o
v
er
n
ed
b
y
a
tr
ain
in
g
a
lg
o
r
ith
m
,
with
b
ac
k
p
r
o
p
ag
atio
n
b
ein
g
th
e
m
o
s
t
f
u
n
d
am
en
tal
m
eth
o
d
.
B
ac
k
p
r
o
p
ag
atio
n
ca
lcu
lates
th
e
g
r
ad
ien
t
o
f
a
lo
s
s
f
u
n
ctio
n
(
cr
o
s
s
-
en
tr
o
p
y
)
with
r
esp
ec
t
to
th
e
n
etwo
r
k
’
s
weig
h
ts
,
allo
win
g
th
e
m
o
d
el
t
o
iter
ativ
ely
ad
ju
s
t
its
p
ar
am
e
ter
s
v
ia
an
o
p
tim
izatio
n
alg
o
r
ith
m
lik
e
s
to
ch
asti
c
g
r
ad
ien
t
d
escen
t
(
SGD)
o
r
A
d
am
to
m
in
im
ize
p
r
ed
ictio
n
e
r
r
o
r
.
T
h
is
ca
p
a
b
ilit
y
to
lear
n
in
tr
icate
,
n
o
n
-
lin
ea
r
d
ec
is
io
n
b
o
u
n
d
ar
ies with
o
u
t e
x
p
licit f
ea
tu
r
e
en
g
in
ee
r
in
g
m
a
k
es AN
Ns
h
ig
h
ly
ef
f
ec
tiv
e
f
o
r
class
if
icat
io
n
task
s
s
u
ch
as th
e
f
au
lt d
iag
n
o
s
is
p
r
e
s
en
ted
in
th
is
wo
r
k
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J E
lec
&
C
o
m
p
E
n
g
I
SS
N:
2088
-
8
7
0
8
S
ta
to
r
in
tert
u
r
n
s
h
o
r
t c
ir
cu
it
fa
u
lt id
en
tifi
ca
tio
n
in
d
o
u
b
ly
-
f
ed
in
d
u
ctio
n
g
e
n
era
to
r
…
(
V
iv
ek
K
u
s
h
w
a
h
a
)
2385
Fig
u
r
e
3
.
Neu
r
al
n
etwo
r
k
m
o
d
el
5.
RE
SU
L
T
S AN
D
D
I
SCU
SS
I
O
N
Simu
latio
n
r
esu
lts
f
r
o
m
MA
T
L
AB
Simu
lin
k
ar
e
s
h
o
wn
in
t
h
e
f
ig
u
r
es
g
i
v
en
b
elo
w.
T
h
e
s
im
u
latio
n
m
o
d
el
s
h
o
w
n
in
Fig
u
r
e
1
,
h
as
a
DFI
G
win
d
tu
r
b
in
e
o
f
1
MV
A,
4
0
0
V,
5
0
Hz
co
n
n
ec
ted
wit
h
a
n
eu
r
al
n
etwo
r
k
m
o
d
el.
I
n
th
e
p
r
o
p
o
s
ed
m
o
d
el,
I
T
F
is
cr
ea
ted
in
p
h
ase
‘
a
,
’
p
h
ase
‘
b
,
’
an
d
p
h
ase
‘
c’
s
im
u
ltan
eo
u
s
ly
.
Fig
u
r
e
4
s
h
o
ws th
e
wav
ef
o
r
m
s
o
f
h
ea
lth
y
DFI
G
f
o
r
p
.
u
.
v
alu
e
o
f
3
-
p
h
ase
v
o
ltag
e,
cu
r
r
en
t,
ac
tiv
e
a
n
d
r
ea
ctiv
e
p
o
wer
in
f
ig
u
r
es
4
(
a)
to
4
(
d
)
r
esp
ec
ti
v
ely
.
T
h
e
wav
e
f
o
r
m
f
o
r
ea
c
h
q
u
an
tity
(
v
o
ltag
e,
c
u
r
r
e
n
t,
ac
tiv
e
an
d
r
ea
ctiv
e
p
o
wer
)
h
as a
g
en
er
al
c
h
ar
ac
ter
is
tic
with
o
u
t a
n
y
d
is
tu
r
b
an
ce
s
,
as th
e
h
ea
lth
y
m
ac
h
in
e
s
h
o
ws
in
th
e
f
ig
u
r
e.
(
a)
(
b
)
(
c)
(
d
)
Fig
u
r
e
4
.
T
h
e
wav
ef
o
r
m
o
f
3
-
p
h
ase
v
o
ltag
e
(
a)
an
d
cu
r
r
en
t (
b
)
,
ac
tiv
e
(
c
)
an
d
r
ea
ctiv
e
(
d
)
p
o
wer
f
o
r
DFI
G
u
n
d
er
n
o
r
m
al
co
n
d
itio
n
s
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
0
8
8
-
8
7
0
8
I
n
t J E
lec
&
C
o
m
p
E
n
g
,
Vo
l.
16
,
No
.
5
,
Octo
b
e
r
20
26
:
2
3
7
9
-
2
3
9
2
2386
No
w,
co
n
s
id
er
th
e
in
ter
-
tu
r
n
f
au
lt
in
t
h
e
s
tato
r
win
d
in
g
o
n
ly
,
an
d
th
e
r
o
to
r
win
d
in
g
is
h
e
alth
y
.
W
e
g
et
d
if
f
e
r
en
t
r
esu
lts
,
wh
ich
ar
e
s
h
o
wn
in
F
ig
u
r
e
5
.
First,
a
f
au
lt
in
v
o
lv
i
n
g
5
s
h
o
r
ted
tu
r
n
s
in
p
h
ase
‘
a’
o
f
th
e
s
tato
r
win
d
in
g
is
in
tr
o
d
u
ce
d
f
r
o
m
0
.
5
to
0
.
7
s
ec
.
As s
h
o
wn
in
Fig
u
r
e
5
(
a)
,
th
e
FLD
is
ze
r
o
p
r
io
r
to
0
.
5
s
ec
an
d
in
cr
ea
s
es
u
p
o
n
f
au
lt
in
ce
p
tio
n
.
Similar
ly
,
th
e
m
ag
n
itu
d
e
o
f
FLDV
in
cr
ea
s
es,
as
d
ep
ic
ted
in
Fig
u
r
e
5
(
b
)
.
C
o
n
s
eq
u
en
tly
,
th
e
p
h
ase
‘
a’
cu
r
r
en
t
s
ig
n
atu
r
e
ex
h
ib
its
an
in
cr
ea
s
e
in
m
ag
n
itu
d
e
d
u
e
to
th
e
r
ed
u
ce
d
p
h
ase
im
p
ed
an
ce
,
as
illu
s
tr
ated
in
Fig
u
r
e
5
(
c)
(
r
e
d
cu
r
v
e)
.
C
o
n
v
e
r
s
ely
,
th
e
p
h
ase
‘
a’
v
o
ltag
e
am
p
litu
d
e
is
r
ed
u
ce
d
,
as sh
o
wn
in
Fig
u
r
e
5
(
d
)
(
r
ed
c
u
r
v
e)
.
Similar
ly
,
an
i
n
ter
-
tu
r
n
f
au
lt
(
I
T
F)
is
cr
ea
te
d
in
p
h
ase
‘
b
’
f
r
o
m
1
.
0
to
1
.
3
s
ec
.
Fig
u
r
e
5
(
e)
s
h
o
ws
th
e
in
cr
ea
s
ed
cu
r
r
en
t
am
p
litu
d
e
in
p
h
ase
‘
b
’
(
y
ello
w
cu
r
v
e)
,
wh
ile
Fig
u
r
e
5
(
f
)
s
h
o
ws
th
e
co
r
r
esp
o
n
d
in
g
d
ec
r
ea
s
e
in
v
o
ltag
e
am
p
litu
d
e
f
o
r
p
h
ase
‘
b
’
(
y
ello
w
cu
r
v
e)
.
Fin
ally
,
an
I
T
F
is
estab
li
s
h
ed
in
p
h
ase
‘
c’
f
r
o
m
1
.
5
t
o
1
.
7
s
ec
.
As
s
h
o
wn
in
Fig
u
r
e
5
(
g
)
,
th
e
cu
r
r
en
t
a
m
p
li
tu
d
e
in
p
h
ase
‘
c’
in
cr
ea
s
es
(
b
lu
e
cu
r
v
e)
,
an
d
th
e
co
r
r
esp
o
n
d
in
g
v
o
ltag
e
r
esp
o
n
s
e
in
Fig
u
r
e
5
(
h
)
d
em
o
n
s
tr
at
es
s
h
o
r
ter
b
lu
e
v
o
ltag
e
p
ea
k
s
co
m
p
ar
ed
to
th
e
h
ea
lth
y
p
h
ases
.
(
a)
(
b
)
(
c)
(
d
)
(
e)
(f)
(
g
)
(
h
)
Fig
u
r
e
5
.
FLD,
FLDV
,
cu
r
r
en
t
an
d
v
o
ltag
e
r
esp
o
n
s
es in
th
e
s
tato
r
win
d
in
g
o
f
DFI
G,
wh
e
n
5
–
tu
r
n
s
o
f
th
e
s
tato
r
win
d
in
g
ar
e
s
h
o
r
ted
(
a)
r
ep
r
esen
tin
g
th
e
FLD,
(
b
)
FL
DV
u
n
d
er
f
a
u
lt,
(
c)
s
h
o
w
th
e
c
u
r
r
en
t
,
(
d
)
v
o
ltag
e
r
esp
o
n
s
es wh
en
5
t
u
r
n
s
o
f
p
h
ase
‘
a’
a
r
e
s
h
o
r
te
d
i
n
th
e
s
tato
r
win
d
in
g
,
(
e
)
s
h
o
wi
n
g
cu
r
r
en
t,
(
f
)
v
o
ltag
e
r
esp
o
n
s
es wh
en
5
t
u
r
n
s
o
f
p
h
ase
‘
b
’
ar
e
s
h
o
r
ted
i
n
th
e
s
tato
r
win
d
in
g
,
(
g
)
s
h
o
w
th
e
cu
r
r
e
n
t,
(
h
)
v
o
ltag
e
r
esp
o
n
s
es wh
en
5
t
u
r
n
s
o
f
p
h
ase
‘
c’
a
r
e
s
h
o
r
te
d
i
n
th
e
s
tato
r
win
d
in
g
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J E
lec
&
C
o
m
p
E
n
g
I
SS
N:
2088
-
8
7
0
8
S
ta
to
r
in
tert
u
r
n
s
h
o
r
t c
ir
cu
it
fa
u
lt id
en
tifi
ca
tio
n
in
d
o
u
b
ly
-
f
ed
in
d
u
ctio
n
g
e
n
era
to
r
…
(
V
iv
ek
K
u
s
h
w
a
h
a
)
2387
As
th
e
cu
r
r
en
t
in
cr
ea
s
es
in
th
e
win
d
in
g
,
th
e
tem
p
er
atu
r
e
o
f
th
e
win
d
in
g
also
in
cr
ea
s
es,
d
u
e
to
wh
ich
th
e
in
s
u
latio
n
m
ay
f
ail,
r
esu
ltin
g
in
p
er
m
an
e
n
t
d
am
ag
e
to
th
e
win
d
in
g
.
So
,
it
is
n
ec
es
s
ar
y
to
ea
r
ly
id
en
tify
th
e
in
ter
-
tu
r
n
f
au
lt
to
s
av
e
th
e
m
a
ch
in
e
f
r
o
m
d
am
ag
e
an
d
to
g
et
s
tab
le
an
d
r
eliab
le
p
o
wer
in
th
e
s
y
s
tem
.
T
h
u
s
,
th
e
n
eu
r
al
n
etwo
r
k
to
o
l
is
ap
p
lied
in
MA
T
L
AB
f
o
r
class
if
icatio
n
an
d
id
en
tific
atio
n
o
f
f
au
lts
.
Firstl
y
,
th
e
m
o
d
el
is
tr
ain
ed
f
o
r
th
e
f
au
lts
o
cc
u
r
r
e
d
in
all
t
h
e
p
h
ases
at
d
if
f
er
e
n
t
tim
es,
an
d
ea
ch
p
h
ase
is
a
s
s
ig
n
ed
a
d
if
f
e
r
en
t
m
ag
n
itu
d
e.
Af
ter
th
at,
th
is
tr
ai
n
ed
m
o
d
el
o
f
NN
is
co
n
n
ec
ted
to
th
e
s
im
u
latio
n
m
o
d
el
o
f
DFI
G,
wh
ich
f
u
r
th
e
r
id
en
tifie
s
th
e
f
au
lt in
win
d
i
n
g
s
as p
er
th
e
tr
ain
in
g
.
I
n
MA
T
L
AB
’
s
n
eu
r
al
n
etwo
r
k
to
o
lb
o
x
,
th
e
“b
est” p
er
f
o
r
m
a
n
ce
r
an
g
e
f
o
r
a
r
e
g
r
ess
io
n
cu
r
v
e
d
ep
en
d
s
o
n
th
e
s
p
ec
if
ic
p
r
o
b
lem
an
d
d
ata,
b
u
t
g
e
n
er
ally
,
a
lo
w
m
ea
n
s
q
u
ar
ed
e
r
r
o
r
(
MSE
)
a
n
d
a
h
ig
h
R
-
v
alu
e
(
co
r
r
elatio
n
co
e
f
f
icien
t)
in
d
ica
te
g
o
o
d
p
er
f
o
r
m
an
ce
.
T
h
e
r
an
g
e
o
f
p
er
f
o
r
m
an
ce
f
o
r
a
r
eg
r
es
s
io
n
cu
r
v
e
d
ep
en
d
s
o
n
th
e
p
r
o
b
lem
an
d
d
ata,
b
u
t
a
lo
w
v
alu
e
o
f
MSE
,
i.e
.
,
clo
s
e
to
‘
0
’
an
d
a
h
ig
h
R
-
v
alu
e
(
c
o
r
r
elatio
n
co
ef
f
icien
t)
,
i.e
.
,
clo
s
e
to
1
,
d
e
s
ig
n
ates
g
o
o
d
p
er
f
o
r
m
an
ce
.
A
n
ex
ce
s
s
iv
e
R
-
v
alu
e
m
ea
n
s
th
at
th
e
p
r
e
d
icted
an
d
r
ea
l
v
alu
es
ar
e
s
tr
o
n
g
ly
co
n
n
e
cted
alo
n
g
a
s
tr
aig
h
t
lin
e,
wh
ich
m
ea
n
s
th
at
th
e
r
eg
r
ess
io
n
is
ac
cu
r
ate.
Fig
u
r
e
6
in
d
icate
s
th
e
b
est
v
alid
atio
n
p
er
f
o
r
m
a
n
ce
is
0
.
1
0
3
9
3
at
e
p
o
ch
2
3
(
i.e
.
,
2
3
iter
atio
n
s
)
,
wh
i
ch
m
ea
n
s
th
e
v
alu
e
o
f
MSE
is
cl
o
s
e
to
‘
0
,
’
i.e
.
,
0
.
1
0
3
9
3
.
T
h
u
s
,
th
e
t
r
ain
in
g
p
er
f
o
r
m
an
ce
cu
r
v
e
d
esig
n
ates
g
o
o
d
p
er
f
o
r
m
an
ce
o
f
th
e
n
eu
r
al
n
etwo
r
k
.
Fig
u
r
e
7
s
h
o
ws
th
e
r
eg
r
ess
io
n
cu
r
v
e
o
f
th
e
n
e
u
r
al
n
etwo
r
k
f
o
r
tr
ain
i
n
g
,
v
alid
atio
n
,
an
d
test
in
g
.
Fo
r
tr
ain
in
g
,
th
e
v
alu
e
o
f
R
=
0
.
9
6
9
6
7
,
f
o
r
v
alid
atio
n
,
R
=
0
.
9
5
3
7
4
,
a
n
d
f
o
r
test
,
R
=
0
.
9
6
9
5
7
,
a
n
d
th
e
o
v
er
all
v
al
u
e
o
f
R
=
0
.
9
6
7
1
5
.
T
h
er
ef
o
r
e,
th
e
r
eg
r
ess
io
n
cu
r
v
e
is
also
s
tip
u
latin
g
g
o
o
d
p
e
r
f
o
r
m
an
ce
as
R
h
as
a
v
alu
e
v
er
y
clo
s
e
to
1
.
Fig
u
r
e
6
.
T
r
ain
in
g
p
er
f
o
r
m
an
c
e
cu
r
v
e
o
f
n
e
u
r
al
n
etwo
r
k
b
y
MA
T
L
AB
Fig
u
r
e
7
.
R
eg
r
ess
io
n
cu
r
v
e
o
f
n
eu
r
al
n
etwo
r
k
b
y
MA
T
L
AB
f
o
r
tr
ain
in
g
,
v
alid
atio
n
an
d
test
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
0
8
8
-
8
7
0
8
I
n
t J E
lec
&
C
o
m
p
E
n
g
,
Vo
l.
16
,
No
.
5
,
Octo
b
e
r
20
26
:
2
3
7
9
-
2
3
9
2
2388
Her
e,
a
f
ee
d
-
f
o
r
war
d
ANN
is
d
ev
elo
p
e
d
f
o
r
th
e
m
u
lti
-
class
class
if
icatio
n
o
f
th
e
o
p
er
atio
n
al
s
tate
o
f
a
s
y
s
tem
in
to
o
n
e
o
f
f
o
u
r
ca
teg
o
r
ies:
Hea
lth
y
,
p
h
ase
‘
a’
f
a
u
lt,
p
h
ase
‘
b
’
f
au
lt,
an
d
p
h
as
e
‘
c’
f
au
lt.
Fig
u
r
e
8
s
h
o
ws
a
co
n
f
u
s
io
n
m
atr
ix
with
a
to
tal
d
ataset
co
m
p
r
is
ed
o
f
1
0
,
0
0
1
s
am
p
les,
with
a
n
o
tab
le
class
im
b
alan
ce
wh
er
e
th
e
h
ea
lth
y
class
co
n
t
ain
ed
1
3
0
1
in
s
tan
ce
s
,
an
d
ea
ch
o
f
th
e
t
h
r
ee
f
au
lt
class
es,
p
h
ase
‘
a
’
an
d
‘
c
’
co
n
tain
ed
2
0
0
in
s
tan
ce
s
,
an
d
p
h
ase
‘
b
’
h
ad
3
0
0
in
s
tan
ce
s
.
T
o
m
itig
ate
th
e
b
ias
in
tr
o
d
u
ce
d
b
y
th
is
im
b
alan
ce
,
th
e
s
y
n
th
etic
m
in
o
r
ity
o
v
er
-
s
am
p
lin
g
tech
n
iq
u
e
(
SMOT
E
)
was
ap
p
lied
to
th
e
tr
ain
i
n
g
p
ar
titi
o
n
to
g
en
er
ate
s
y
n
th
etic
s
am
p
les
f
o
r
th
e
m
in
o
r
ity
class
es.
T
h
e
m
o
d
el
ar
ch
itectu
r
e
co
n
s
is
ted
o
f
an
in
p
u
t
lay
er
,
two
h
id
d
en
lay
er
s
with
6
4
an
d
3
2
n
e
u
r
o
n
s
r
esp
ec
tiv
ely
em
p
lo
y
in
g
th
e
r
ec
tifie
d
lin
ea
r
u
n
it
(
R
eL
U)
a
ctiv
atio
n
f
u
n
ctio
n
,
an
d
an
o
u
t
p
u
t la
y
er
with
f
o
u
r
n
eu
r
o
n
s
u
s
in
g
th
e
s
o
f
tm
ax
f
u
n
ctio
n
to
o
u
tp
u
t c
lass
p
r
o
b
a
b
ilit
ies.
T
h
e
m
o
d
el
was
co
m
p
iled
with
th
e
Ad
am
o
p
t
im
izer
an
d
ca
te
g
o
r
ical
cr
o
s
s
-
en
tr
o
p
y
lo
s
s
f
u
n
ctio
n
an
d
w
as
tr
ain
ed
f
o
r
1
0
0
ep
o
ch
s
with
a
b
atc
h
s
ize
o
f
3
2
.
T
h
e
m
o
d
el’
s
p
e
r
f
o
r
m
an
ce
was
ev
alu
ated
o
n
a
h
o
ld
-
o
u
t
t
est
s
et.
T
h
e
r
esu
ltin
g
co
n
f
u
s
io
n
m
atr
ix
d
em
o
n
s
tr
at
ed
ex
ce
p
tio
n
al
class
if
icatio
n
ca
p
ab
ilit
y
,
with
th
e
m
o
d
el
c
o
r
r
ec
tly
class
if
y
in
g
1
2
9
6
o
u
t
o
f
1
3
0
1
h
ea
lth
y
i
n
s
tan
ce
s
,
1
9
9
o
u
t
o
f
2
0
0
in
s
tan
c
es
f
o
r
p
h
ase
‘
a’
an
d
‘
c’
f
au
lt
an
d
2
9
9
o
u
t
o
f
3
0
0
in
s
tan
ce
s
f
o
r
p
h
ase
‘
b
’
f
a
u
lt,
a
ch
iev
in
g
a
f
in
al
o
v
e
r
all
class
if
icatio
n
ac
cu
r
ac
y
o
f
9
9
.
6
0
%.
Fig
u
r
e
8
.
C
o
n
f
u
s
io
n
m
atr
i
x
Pre
v
io
u
s
ly
,
a
u
th
o
r
s
a
n
aly
ze
d
t
h
e
I
T
SC
f
au
lt
(
lik
e
2
o
r
5
tu
r
n
s
s
h
o
r
ted
)
f
o
r
a
p
ar
ticu
lar
p
h
as
e
o
f
s
tato
r
win
d
in
g
.
B
u
t,
in
th
is
s
tu
d
y
,
t
h
e
f
au
lts
ar
e
cr
ea
ted
in
all
th
e
3
p
h
ases
,
an
d
th
e
y
ar
e
clas
s
if
ied
b
y
d
if
f
e
r
en
t
am
p
litu
d
es.
As
p
er
th
e
tr
ain
i
n
g
,
Fig
u
r
e
9
s
h
o
ws
th
at
a
f
a
u
lt
in
p
h
ase
‘
a’
wh
ich
is
o
cc
u
r
r
ed
b
etwe
en
tim
e
0
.
5
to
0
.
7
s
ec
,
h
a
v
in
g
a
m
p
litu
d
e
1
; a
f
a
u
lt o
cc
u
r
r
ed
in
p
h
ase
‘
b
’
b
etwe
en
tim
e
1
.
1
to
1
.
4
s
ec
,
h
av
in
g
am
p
litu
d
e
2
;
an
d
a
f
au
lt
o
cc
u
r
r
e
d
in
p
h
ase
‘
c’
b
etwe
en
tim
e
1
.
5
to
1
.
7
s
ec
,
h
av
in
g
am
p
litu
d
e
3
.
T
h
is
m
ea
n
s
we
h
av
e
tr
ain
ed
th
e
n
e
u
r
al
n
etwo
r
k
m
o
d
el
d
u
r
i
n
g
a
f
a
u
lt
wh
en
th
e
wav
ef
o
r
m
h
as
an
am
p
litu
d
e
eq
u
al
to
1
,
w
h
ich
m
ea
n
s
th
e
in
ter
-
tu
r
n
f
a
u
lt
o
cc
u
r
r
ed
in
p
h
ase
‘
a
’
.
Similar
ly
,
as
th
e
wav
ef
o
r
m
h
as
its
v
alu
e
as
2
,
t
h
en
t
h
e
s
h
o
r
t
cir
cu
it
will
b
e
in
p
h
ase
‘
b
,
’
an
d
f
o
r
th
e
am
p
litu
d
e
as
3
,
th
e
f
au
lt
will
b
e
in
p
h
ase
‘
c
,
’
o
th
er
wis
e,
th
e
am
p
litu
d
e
will b
e
ze
r
o
,
wh
ich
m
ea
n
s
wh
en
th
er
e
is
n
o
f
au
lt,
th
e
wa
v
ef
o
r
m
h
as z
er
o
m
ag
n
itu
d
e.
Fig
u
r
e
9
.
I
n
p
u
t
p
r
o
v
id
ed
t
o
th
e
n
eu
r
al
n
etwo
r
k
to
tr
ain
th
e
m
o
d
el
f
o
r
id
en
tify
i
n
g
tu
r
n
f
a
u
lt o
cc
u
r
r
e
d
in
p
h
ase
‘
a
,
’
‘
b
,
’
an
d
‘
c
’
Evaluation Warning : The document was created with Spire.PDF for Python.