In
te
r
n
ation
a
l Jou
rn
al
o
f Po
we
r
Elec
tron
ic
s an
d
D
r
ive S
y
stem
(IJ
PED
S
)
Vo
l
.
1
0
, No
.
1
, M
a
r
c
h
2
01
9
,
pp
.
9
3
~
103
IS
S
N
: 2088-
86
94,
D
O
I
:
10.11
59
1
/ij
ped
s
.
v10
.
i
1.pp
9
3
-1
03
93
Jou
rn
a
l
h
o
me
pa
ge
:
ht
tp:
//i
a
e
score
.
com
/
j
o
u
r
na
l
s
/
i
n
d
e
x
.
p
hp/IJ
PED
S
Determi
n
ing the operational sta
tus of a th
ree phase induction
motor using a pre
d
icti
ve data mining mode
l
A
d
er
i
b
igbe
I
srael
A
d
e
k
itan,
Ad
eyin
ka A
dewa
le
, A
l
a
sh
i
r
i
Olait
a
n
D
e
part
men
t
o
f Electri
cal
a
n
d
I
n
f
orm
a
ti
on
En
g
i
n
e
e
ri
ng,
C
ov
enant
Univers
i
ty,
Ni
ge
ri
a
Art
i
cl
e In
fo
ABSTRACT
A
r
tic
le hist
o
r
y
:
R
e
c
e
i
v
e
d
Sep
1
8
,
2
018
Re
vise
d N
ov
20,
201
8
Ac
ce
p
t
ed
No
v
2
4
,
2
018
Th
e
o
p
erati
o
n
a
l
p
e
rf
orm
a
nce
of
a
t
h
r
ee-phas
e
i
ndu
cti
o
n
m
o
t
o
r
is
i
m
p
aired
by
un
bal
a
nced
v
o
ltag
e
s
upp
ly
d
u
e
t
o
t
h
e
g
e
ner
a
t
i
on
o
f
neg
a
ti
ve
s
eq
ue
n
ce
curren
t
s
,
a
n
d
n
eg
ati
v
e
sequ
ence
to
rqu
e
w
hich
i
ncreas
e
m
o
to
r
l
o
s
ses
and
a
l
so
trig
ge
r
to
rqu
e
p
ulsa
tion
s
.
In
t
h
i
s
stud
y,
d
a
t
a
m
i
n
i
ng
a
pp
ro
a
c
h
wa
s
appli
e
d
in
dev
e
lo
pin
g
a
pred
i
cti
v
e
mo
del
u
s
in
g the
histo
r
ical
,
sim
u
l
a
ted
o
p
e
ratio
nal
da
t
a
o
f
a
m
o
t
o
r
f
o
r
c
l
a
s
s
i
f
y
i
n
g
s
a
m
p
l
e
m
o
t
o
r
d
a
t
a
u
n
d
e
r
t
h
e
a
p
p
r
o
p
r
i
at
e
ty
pe
o
f
vo
lt
age
s
u
p
p
l
y
i
.e.
b
a
lan
ced
(
BV
)
and
un
bal
a
nce v
o
l
t
age su
pply
(U
B
= 1%
t
o
5%
).
A
d
at
aset
c
on
tai
n
i
ng
th
e
val
u
es
o
f
a
t
h
ree-ph
ase
i
n
d
u
ct
ion
m
otor’s
perf
o
r
man
ce
pa
ram
e
ter
valu
es
w
as
a
n
a
ly
sed
us
ing
KN
IME
(K
on
stan
z
In
for
m
a
t
ion
M
i
ne
r)
a
na
ly
tic
s
p
l
a
tfor
m.
T
h
r
e
e
pre
d
ic
tiv
e
m
od
e
l
s
;
t
h
e
N
a
ï
v
e
Bayes,
D
e
c
i
s
i
o
n
Tree
and
the
Probabilisti
c
Neural
N
etwork
(
P
NN)
P
redict
ors
were
d
epl
o
y
e
d
fo
r
com
p
arat
iv
e
anal
ys
is.
Th
e
d
a
taset
was
d
i
v
i
ded
i
nto
tw
o;
70
%
f
o
r
m
o
d
e
l
t
r
aining
a
nd
l
e
a
rn
in
g,
a
nd
30%
f
or
p
erf
o
rm
an
ce
e
v
alu
a
ti
on
.
Th
e
t
h
ree
pred
ict
o
rs
h
ad
accu
raci
es
o
f
98.649%
,
10
0
%
a
n
d
9
8
.
64
9
%
r
e
s
p
e
c
t
i
v
e
l
y
,
a
n
d
t
h
i
s
c
o
n
f
i
r
m
s
t
h
e
s
u
i
t
a
b
i
l
i
t
y
o
f
d
a
t
a
m
i
n
i
n
g
m
e
t
h
ods
for
pred
ict
i
ve
e
valu
ati
o
n
o
f
a
t
h
r
ee-p
h
as
e
in
du
ctio
n
mo
to
r’s
p
e
rf
or
mance
us
ing
m
achi
n
e l
earn
i
ng.
K
eyw
ord
s
:
Mac
h
i
n
e
lear
n
i
ng
Mo
tor
per
f
orm
a
nce
cha
r
ac
t
e
ris
tics
Ne
g
a
t
i
v
e a
n
d
po
si
tiv
e seq
u
e
n
ce
com
p
o
n
e
n
t
P
o
w
e
r qua
lit
y
Three
pha
se
i
n
duc
t
i
o
n
m
otor
V
o
l
t
a
g
e
un
ba
l
a
nce
Co
pyri
gh
t © 2
019 In
stit
u
t
e
of Advanced
En
gi
neeri
n
g
an
d
S
c
ien
ce.
All
rights
res
e
rv
ed.
Corres
pon
d
i
n
g
Au
th
or:
Ader
i
b
i
gbe
I
srae
l Adek
it
an,
D
e
pa
rtme
nt
o
f
El
e
c
t
rica
l
and
Inform
ati
on E
n
gi
neer
in
g,
Cove
na
nt U
n
i
v
e
rsit
y,
O
t
a,
O
gun S
t
at
e,
N
iger
i
a
.
Em
ail:
ader
ibi
gbe.
a
de
k
i
t
a
n
@
c
o
v
e
na
nt
u
n
ive
r
sity.
e
du.n
g
1.
I
N
TR
OD
U
C
TI
O
N
Thr
ee
pha
se
i
n
duc
t
i
o
n
m
o
t
ors
(TPI
M
)
h
a
v
e
fou
n
d
a
p
p
l
i
ca
t
i
ons
i
n
vari
ou
s
c
o
mm
er
cial
a
n
d
i
ndus
tria
l
opera
tio
ns
[
1]
due
t
o
i
t
s
l
o
w
c
o
s
t
,
l
o
w
m
a
in
tena
nc
e
r
e
qu
irem
ent
,
rug
g
ed
d
es
ign
a
n
d
no
n-c
o
mpl
e
x
c
o
nst
r
u
c
ti
on
.
Th
e
i
m
po
rt
a
n
ce
o
f
a
T
P
I
M
was
a
l
so
e
mp
h
a
si
ze
d
by
[
2]
t
hat
pr
op
ose
d
a
d
ri
ve
s
yste
m
for
con
v
er
ti
ng
s
i
n
g
le
p
h
a
s
e
t
o
t
h
r
ee
pha
se
f
or
p
ow
e
r
in
g
ind
u
c
t
io
n
m
ot
ors
i
n
r
ura
l
a
rea
s
w
here
only
si
n
g
l
e
pha
se
sup
p
l
y
i
s
a
v
ai
l
a
ble.
A
t
hre
e
-pha
se
i
nd
uc
ti
o
n
m
otor
i
s
a
p
o
ly-p
h
a
s
e
e
q
u
i
pm
ent
w
h
ic
h
requ
ire
s
a
t
hr
ee
pha
s
e
sup
p
l
y
t
o
ru
n.
T
hree
p
hase
s
upp
ly
s
ys
tem
s
a
r
e
t
he
or
etic
al
l
y
d
es
i
gne
d
t
o
h
ave
a
ba
l
a
nce
d
a
n
d
e
qua
l
vo
lta
g
e
ma
gni
tude
p
er
pha
se
,
bu
t
d
u
e
to
o
pe
rat
i
ona
l
rea
liti
es
s
uch
a
s
u
n
r
e
l
i
a
bl
e
po
wer
su
ppl
y
[3],
l
i
n
e
di
stu
r
b
a
n
c
e
s
,
motor
w
i
n
d
i
n
g
f
a
c
tors,
the
ra
tio
o
f
thr
ee
pha
se
t
o
si
n
g
le
p
h
a
se
l
oa
ds
[
4],
transfor
me
r
fau
l
t
s
,
l
i
ne
t
ra
nsp
o
s
i
t
i
on
i
s
su
es
,
un
equ
a
l
t
r
a
n
sf
o
r
me
r
t
a
p
set
t
i
n
gs,
h
eav
y
c
o
mmerc
i
a
l
l
o
a
d
s
a
n
d
s
o
f
ort
h
;
t
h
e
vo
l
t
a
g
e
ma
g
n
it
ude
o
f
ea
c
h
pha
se
o
f
a
t
h
ree
phase
s
u
p
p
l
y
a
re
unequa
l
s
o
me
ti
m
e
s,
a
nd
t
he
l
i
n
e
t
o
l
i
n
e
pha
se
s
h
i
ft
m
a
y
a
lso
de
v
i
a
t
e
fr
o
m
t
h
e
no
rmal
1
20
.
T
h
is
a
bn
o
r
m
a
l
sup
p
ly
c
on
d
i
t
i
on
is
r
efe
rre
d
to
a
s
vo
l
t
a
g
e
un
b
a
lanc
e
[4],
[
5].
V
o
lta
g
e
un
ba
lanc
e e
x
i
s
ts i
n m
o
st su
p
p
l
y
ne
tw
or
ks a
n
d
it i
s
qu
i
t
e
sev
e
r
e
i
n weak power system
s
[
6]
.
The
perf
orm
a
nce
o
f
a
T
P
I
M
is
i
mpa
i
r
e
d
w
h
en
ope
ra
ti
n
g
u
nder
u
n
b
a
l
a
nce
vo
l
t
ag
e
con
d
i
t
i
ons.
V
o
l
t
a
g
e un
ba
la
nce
st
imu
l
ates incre
ase
d
m
ot
o
r
l
osse
s
w
h
ic
h resul
t
s
i
n
i
n
cre
a
s
e
d
h
eat
g
e
n
erat
io
n
t
h
a
t
m
a
y
l
ea
d
to
e
arl
y
m
o
t
or
f
ai
l
u
re
[
7],
[8].
V
ol
t
a
ge
u
n
b
a
l
a
n
ce
re
duce
s
m
oto
r
ef
fic
i
e
n
cy
t
here
b
y
i
ncr
easi
ng
e
n
er
gy
cos
t
f
or
the
user
[
9],
a
nd
by
im
p
lic
a
t
i
o
n,
t
he
r
e
d
uce
d
e
ffic
ie
nc
y
i
n
cre
a
ses
the
sys
t
em
l
oa
d
o
n
t
h
e
pow
e
r
p
la
nt
w
hic
h
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SSN: 2088-
8694
I
nt
J
P
ow
Elec
& Dr
i
S
y
st, Vol. 10,
N
o.
1, Mar
c
h 2
0
1
9
:
93
–
103
94
un
nece
ssar
i
l
y
d
e
p
l
e
t
e
s
t
h
e
e
n
e
r
gy
rese
rve
of
t
he
pow
er
p
la
nt
[
1
0].
V
o
lta
g
e
u
nba
l
a
nce
c
r
eate
s
a
m
ag
ne
tic
f
lux
t
h
at
opp
ose
s
t
h
e
m
a
i
n
f
l
ux
,
a
n
d
th
i
s
c
au
s
e
s
po
wer
an
d
to
rque
o
s
ci
l
l
at
ion
a
t
t
w
i
c
e
t
he
f
re
que
nc
y
o
f
t
he
s
up
p
l
y.
Conse
que
n
tly,
t
h
e
o
p
p
o
si
ng
fl
ux
lea
d
s
to
t
he
g
ene
r
a
t
i
o
n
o
f
n
ega
t
i
v
e
s
e
qu
en
c
e
c
u
r
re
nt
s
t
h
at
t
r
i
gg
er
i
n
c
rea
s
e
d
motor
los
s
e
s
[
11]
,
and
heat
p
ro
d
u
ct
i
on
w
h
ic
h
m
a
y
resu
lt
i
n
loca
l
ho
t
spo
t
s
i
n
t
he
s
ta
tor
w
i
n
d
in
gs
[
7],
[12],
[13].
U
s
in
g
F
o
rte
s
c
u
e
The
o
re
m
a
n
u
nba
l
a
nce
v
o
l
t
a
g
e
ca
n
be
r
esol
ved
in
to
t
hree
s
ym
m
e
trica
l
s
eq
ue
nc
e
c
o
mp
on
en
ts,
t
h
es
e
are
-
t
h
e
zero
se
qu
e
n
ce
,
the
p
o
s
iti
v
e
s
e
q
u
e
n
c
e
an
d the n
e
gat
i
ve
s
e
q
u
e
n
ce com
p
o
n
e
n
t
s
[
14]
.
G
i
ve
n li
ne
v
o
l
t
a
ge
s
V
a
,
V
b
a
n
d
V
c
, these
ca
n
b
e
transfor
me
d int
o
seq
ue
nc
e
com
p
o
n
e
n
ts
a
s show
n
i
n
F
ig
ure
1.
120
o
12
0
o
V
C
+
V
A
+
V
B
+
120
o
120
o
V
B
-
V
A
-
V
C
-
V
AO
V
BO
V
CO
(a
)
(
b
)
(
c
)
Figure
1.
(
a) Ze
r
o seque
nce
(b)
Positive
se
q
u
enc
e
a
n
d
(
c) N
egat
ive
se
que
nc
e vo
l
t
ag
e
co
m
pone
n
t
s
B
y
d
es
i
gn,
a
n
in
d
u
ct
i
on
mo
tor
c
a
n
t
o
l
e
r
a
t
e
rea
s
ona
bl
e
l
e
vel
s
o
f
vol
t
a
g
e
unb
a
l
an
ce
b
ut
w
h
e
n
the
un
ba
lanc
e
be
c
o
me
s
exc
e
s
s
iv
e
t
h
e
m
o
tor
m
u
s
t
b
e
de
ra
ted
to
p
r
e
ven
t
ea
rly
d
a
mag
e
d
u
e
t
o
vol
t
a
g
e
-unba
l
a
n
c
e
in
duc
ed
h
ar
mo
nic
c
u
rren
t
s
[7
].
I
n
t
h
e
s
t
u
dy
by
[15]
,
a
N
e
u
r
al
N
e
t
w
o
r
k
c
o
n
tr
olle
r
w
a
s
pr
op
ose
d
f
or
r
ed
uc
in
g
tor
que
r
i
p
p
l
e
a
nd
c
u
rre
nt
h
ar
monic
s
.
The
d
e
ra
t
i
ng
fac
t
or
o
f
an
i
ndu
c
tion
mo
t
o
r
i
s
d
etermi
n
e
d
by
a
na
l
y
s
i
ng
com
p
ara
t
i
v
e
l
y
the
perfor
m
a
n
ce
o
f
t
he
m
oto
r
und
er
u
n
b
a
l
a
n
ced
a
n
d
b
al
anc
e
d
vol
t
a
g
e
o
pe
ra
t
i
o
n
a
l
co
ndi
t
i
on
s,
and
i
t
is
c
al
cu
late
d
as
t
he
r
ati
o
o
f
t
h
e
m
echa
n
i
c
a
l
o
ut
p
u
t
p
o
we
r
d
u
rin
g
u
nba
la
nc
e
v
o
lta
ge
t
o
tha
t
u
nde
r
bala
nc
e
d
s
up
p
l
y
[
1
6],
[17].
Pow
e
r
sup
p
l
y
q
u
a
l
i
t
y
is
a
m
aj
or
i
nd
uc
ti
o
n
m
ot
or
p
e
rform
anc
e
d
e
t
e
r
m
i
na
n
t
[
1
8
]-
[2
0],
and
as
s
uch,
a
de
qua
te
e
ffor
t
m
u
st
b
e
pu
t
in
p
lac
e
t
o
m
a
na
ge
p
ow
er
qua
li
ty
i
ss
u
e
s
by
u
s
in
g
m
oder
n
t
e
c
hni
qu
e
s
[
21
]
t
o
g
u
a
ra
nt
e
e
qu
al
ity
p
o
w
e
r
s
upp
ly
i
n
o
r
d
e
r
to
e
ns
ure
mo
tor
re
liab
i
l
ity
a
nd
op
tim
al pe
rfor
m
a
n
ce.
Whe
n
a
n
ind
u
c
t
i
o
n
m
ot
or
i
s
oper
a
tin
g
e
i
t
h
e
r
under
ba
lanc
ed
o
r
u
nb
al
a
n
ced
v
o
l
t
a
g
e
c
ond
iti
on
s
,
t
h
e
perform
ance
m
ea
surem
e
nt
p
ara
m
e
t
e
r
s
of
t
he
m
o
t
or
s
uc
h
as
t
he
r
o
t
or
a
nd
s
t
a
t
or
c
ur
rent
s,
t
he
n
e
g
a
t
ive
a
n
d
pos
it
ive
se
que
nce
t
o
rque,
t
h
e
elec
trom
ag
n
e
t
i
c
pow
er
,
t
h
e
air
ga
p
pow
e
r
,
the
r
o
tor
a
n
d
the
s
t
at
or
c
op
p
e
r
wi
nd
i
n
g
lo
ss
es,
t
h
e
re
al
a
nd
r
e
a
c
t
i
v
e
i
npu
t
po
wer,
t
h
e
p
o
w
e
r
f
a
ctor
e
t
c
.
cha
n
ges
acc
ord
i
ngl
y
w
i
t
h
t
he
v
ol
t
a
ge
su
pp
ly
c
on
dit
i
o
n
s
.
In
t
hi
s
stu
d
y
,
t
h
e
s
i
m
ul
at
e
d
o
p
e
ra
t
i
ona
l
d
a
t
a
of
a
t
hre
e
-p
h
a
se
i
n
duc
t
i
o
n
mo
to
r
ope
rat
i
n
g
w
ithi
n
the
m
ot
ori
n
g
s
lip ran
g
e
(
0 <
sl
i
p
<
1
)
un
der bala
nc
e
d
(B
V
)
a
n
d
unb
a
l
an
ce
vol
t
a
ge
s
u
ppl
y
(U
B
=
1%
t
o
5%)
is
c
ol
lec
t
ed
a
nd
p
r
o
ces
sed
for
pred
ic
ti
ve
m
o
d
e
l
li
ng
u
s
i
n
g
d
a
t
a
m
i
ni
ng.
A
p
r
e
dic
t
iv
e
mo
de
l
w
a
s
deve
l
ope
d
usi
ng
K
N
I
ME
(
K
ons
ta
nz
I
nfor
m
a
tio
n
M
i
ner
)
A
naly
t
i
cs
P
l
a
t
for
m
t
o
ana
l
y
s
e
the
da
t
a
set
toward
deve
l
o
p
i
ng
a
f
unc
t
i
o
n
a
l
m
od
e
l
t
ha
t
ca
n
de
t
e
r
m
ine
the
na
t
u
re
o
f
the
v
o
l
t
a
ge
s
u
p
p
l
y
w
h
ether
ba
lance
d
o
r
no
t
us
i
n
g
the
m
o
tor
’
s pe
rform
a
nc
e
hi
s
t
orica
l
d
a
t
a.
2.
NATIO
NAL ELE
CTRICAL MA
NUFA
C
T
URES A
S
SO
CIATIO
N
(NEMA)
The
da
t
a
se
t
de
pl
o
y
ed
i
n
t
h
i
s
s
t
u
d
y
w
as
g
e
n
era
t
e
d
u
si
n
g
N
EMA
M
G
1
(
1
9
9
3
)
de
fi
n
iti
on
o
f
vo
l
t
a
g
e
un
ba
lanc
e.
A
cc
ord
i
ng
t
o
N
EMA
,
vol
t
a
ge u
nba
l
a
nce
is
d
e
f
ine
d
a
s:
m
a
x.
d
ev
i
a
t
i
o
n
from
a
v
erage
l
i
ne
v
olt
a
ge
=
×
100%
av
er
age
l
i
ne
v
ol
t
a
ge
m
a
gni
t
ude
V
o
lt
ag
e
u
n
b
a
lan
ce
(
1
)
(
2
)
a
b
L
a
vg
b
c
La
vg
c
a
La
v
g
La
v
g
M
a
x
[
|
V
-
V
|,|V
-
V
|
,|V
-
V
|
]
×
100%
V
Evaluation Warning : The document was created with Spire.PDF for Python.
Int J
P
o
w
El
e
c
&
D
ri S
yst
I
S
S
N
:
2088-
86
94
D
e
term
i
n
in
g t
h
e
ope
rat
i
o
n
a
l
st
a
t
us
o
f
a th
re
e
phas
e
in
d
u
ct
io
n m
o
tor usin
g …
(Aderi
b
i
g
b
e
I
s
rae
l
Adek
it
a
n
)
95
(
3
)
Th
e
op
e
r
at
i
o
n
a
l
mot
o
r
p
e
rfo
r
ma
n
c
e
d
a
t
a
s
e
t
d
e
pi
ct
s
a
415V
t
h
r
ee
ph
as
e
i
ndu
cti
o
n
moto
r
h
a
vi
ng
t
h
e
fo
l
l
ow
i
n
g
per
un
i
t
p
a
r
am
eter
s
X
m
=
2
.41
Ω
,
X
s
=
0
.
1
2
Ω
,
X
r
=
0
.
1
2
Ω
,
R
r
=
0
.
0
84Ω
,
R
s
=
0
.073Ω
w
i
t
h
a
bas
e
impe
da
nce
of
3
.3
04.
T
he
v
o
l
ta
ge
v
aria
ti
o
n
s
c
ons
ide
r
erd
a
r
e
the
b
a
l
a
nc
e
d
v
olt
a
g
e
(
B
V
),
1
%
un
ba
l
a
n
ce
(1%U
B),
2%
u
nba
la
nc
e
(2%
U
B),
3%
u
n
b
a
l
anc
e
(
3%U
B
),
4
%
un
ba
lanc
e
(%U
B
),
a
nd
5
%
u
nba
la
nc
e
(5%U
B
)
sup
p
l
y
c
o
n
d
i
tions,
i
n
l
i
n
e
w
i
t
h
N
EMA
rec
o
mm
enda
t
i
on
of
5
%
ma
xi
m
um
u
n
b
ala
n
c
e
.
The
pe
r
phase
equ
i
vale
n
t
c
ircui
t
d
iagra
m
o
f
a typ
i
ca
l
TPIM
i
s
prese
n
t
e
d i
n
F
i
gure
2.
I
s
I
r
R
s
jX
M
jX
S
jX
r
I
M
V
s
R
r
1
r
s
R
s
S
t
a
t
o
r
C
op
pe
r
Lo
s
s
Ro
t
o
r
Co
p
p
e
r
Lo
s
s
(P
Co
n
v
)
F
i
gure
2.
P
e
r
p
hase
eq
u
i
v
a
l
e
n
t
di
a
g
ram
of a TP
I
M
3.
RESEARCH
M
ETH
O
D
D
a
ta
m
ini
n
g
is
a
f
i
e
l
d
o
f
st
udy
tha
t
e
nc
om
passes
bo
th
s
t
a
t
i
st
i
c
s
and
ma
ch
i
n
e
l
e
arn
i
ng
,
a
n
d
it
i
s
a
sub
f
iel
d
o
f
c
o
m
put
e
r
s
c
i
e
n
ce
t
ha
t
en
a
b
l
e
s
i
n
t
e
lli
ge
n
t
e
xt
r
a
cti
o
n
of
u
se
ful
i
n
form
ati
o
n
[
2
2
,
23],
pa
tter
n
s
a
n
d
kn
ow
le
d
g
e
[2
4
]
f
rom
dataset
tow
a
rds
cre
a
tin
g
m
o
de
ls
t
hat
r
e
pre
s
e
n
t
t
h
e
k
n
o
w
l
e
dge
a
c
q
u
i
re
d
from
the
d
a
tase
t
there
b
y
ma
k
i
n
g
s
uc
h
kn
ow
l
e
dge
r
e
u
sab
l
e
for
ma
kin
g
d
e
c
is
ions
on
s
i
m
ilar
case
s
.
The
K
N
I
ME
A
na
l
y
tic
s
P
l
atfor
m
w
as
d
ep
l
oye
d
t
o
a
c
h
ie
ve
t
he
m
o
t
o
r
s
upp
l
y
-sta
tu
s
pr
edic
t
i
v
e
m
o
d
e
l
l
i
n
g
.
K
N
I
M
E
i
s
t
h
e
o
p
e
n
s
o
u
r
c
e
softw
a
re
w
i
t
h
ca
paci
t
y
t
o
ha
n
d
le
l
ar
ge
v
o
l
u
m
e
of
d
ata
;
e
q
u
i
p
ped
w
i
t
h
e
x
t
e
n
sive
t
o
o
ls
a
nd
r
e
so
urce
s.
K
N
I
ME
has
fo
un
d
a
p
pl
ica
t
i
on
in
v
ar
ious
a
s
p
ec
t
s
o
f
data
m
i
n
in
g
p
r
oje
c
t
s
ha
n
d
le
d
b
y
m
ore
tha
n
6
0
00
profe
s
s
i
o
n
a
l
s
gl
oba
l
l
y
[2
5],
[26].
K
N
IME
is
t
he
m
odu
lar
da
ta
i
nte
g
rat
i
on
a
n
d
pr
ocess
i
ng
pla
t
for
m
t
ha
t
e
n
a
b
les
use
r
s
to
v
isua
ll
y
cre
a
t
e
data
f
l
o
w
s
f
or
d
a
t
a
a
n
a
l
ys
i
s
a
nd
e
x
p
l
or
at
ion
[2
7].
I
n
t
he
s
tud
y
b
y
[
26]
,
a
mode
l
for
pre
d
ic
tin
g
the
in
te
rnal
fa
ul
ts
o
f
a
n
o
i
l
-imm
er
se
d
po
w
e
r
t
r
ansform
e
r
usin
g
his
t
or
i
c
al
f
a
ult
d
a
t
a
w
as
d
ev
el
op
ed
u
sing
K
NIM
E
.
Th
e
mode
l
deve
lop
e
d
us
i
n
g
pro
b
a
bi
l
i
s
tic
n
e
u
ral
ne
tw
ork
a
c
h
i
eved
a
n
a
c
c
u
r
acy
o
f
8
0
%
.
D
a
t
a
pr
ocess
i
n
g
a
n
d
ana
l
ys
is
i
s
sig
n
i
fica
n
t
i
n
de
ve
l
opi
ng
a
da
ta
m
i
n
i
ng
w
o
rk
flo
w
,
t
he
m
ot
or
p
ar
am
eters
for
the
six
v
o
l
t
a
g
e
sup
p
l
y
scena
r
i
o
s wer
e
a
ppro
p
riate
l
y
sort
e
d
a
nd pre
p
are
d
for
s
u
p
ervi
sed
l
e
a
rn
i
n
g us
in
g K
N
I
M
E
w
orkflow
.
4.
DATA BASED
PRE
D
ICTIVE MODELLING
OF TH
R
EE PHASE INDU
CT
ION
M
O
TOR
VOLTAG
E
S
TAT
U
S U
S
ING KNIME
I
n
t
he
s
t
udy
[28],
an
A
rti
f
i
c
ial
N
e
utra
l
N
e
t
w
or
k
(A
NN
)
mode
l
w
a
s
tra
i
ne
d
t
o
d
e
t
ect
v
ol
ta
g
e
un
ba
lanc
e
in
t
he
m
o
t
or
’s
o
p
e
r
a
tiona
l
da
tas
e
t
us
i
n
g
the
h
i
stor
i
c
a
l
v
o
l
t
a
g
e
d
a
t
a
s
e
t
a
s
a
t
a
r
g
e
t
f
o
r
t
r
a
i
n
i
n
g
t
h
e
f
e
ed
-fo
r
w
ard
net
w
o
r
k
ANN
mo
d
e
l
.
T
he
a
ccu
r
acy
o
f
t
h
e
ANN
mo
d
e
l
w
as
a
ssessed
usi
ng
the
m
e
a
n
s
qua
re
err
o
r.
T
h
e
u
se
o
f
ANN
and
ad
ap
tiv
e
n
e
u
r
o
-
f
u
zz
y
in
f
e
r
e
n
c
e
sy
st
em
f
o
r
p
redic
t
in
g
t
h
e
par
a
m
e
ters
o
f
a
n
in
duc
t
i
on
m
ot
or
w
as
p
ro
pos
ed
[
2
9
].
A
lso,
a
n
o
n
l
ine
fa
ul
t
det
e
ct
ion
a
n
d
perfo
r
m
a
n
ce
eva
l
ua
tio
n
sim
u
la
t
i
o
n
w
a
s
de
ve
l
ope
d
[3
0]
u
s
i
ng
t
he
pha
se
c
urre
n
t
s
,
t
he
v
o
lta
ge
a
nd
t
h
e
m
o
tor
s
p
eed
for
asses
s
m
e
nt.
Likewi
s
e
,
the
fe
asib
il
i
t
y
of
u
si
ng
na
i
v
e
ba
y
e
s
da
ta
m
in
in
g
a
l
gor
it
hm
f
or
i
de
n
ti
fica
t
i
on
a
nd
c
la
ssi
fica
tion
o
f
motor
bearing
fa
ul
ts
w
as
d
em
ons
trate
d
[
3
1
]
,
w
hile
i
n
the
st
ud
y
[32]
f
uzzy
l
o
g
i
c
w
as
a
ppl
i
e
d
fo
r
i
d
ent
i
f
y
i
ng
s
ho
rt
a
nd
o
p
e
n
circui
t
TPIM
f
aults.
I
n
t
hi
s
s
t
u
d
y
,
a
K
N
I
M
E
w
or
kfl
o
w
s
how
n
in
F
ig
ur
e
3
w
a
s
d
e
vel
o
ped
f
or
d
ata
m
i
nin
g
t
he
opera
tio
na
l
motor
perfor
m
a
n
ce
da
tase
t
t
o
w
a
r
d
e
nab
l
in
g
a
pr
edic
t
i
o
n
of
t
he
n
a
t
ure
of
t
he
v
o
l
ta
g
e
s
up
pl
y
i.e
.
w
he
t
h
e
r
b
a
l
a
n
ced
o
r
unb
a
l
an
ced
.
Fo
r
co
mp
ara
t
i
v
e
an
aly
s
i
s
,
t
h
ree
p
r
edi
c
t
ive
al
go
rithm
s
w
er
e
appl
ie
d,
a
nd
t
h
es
e
ar
e
–
Pr
ob
a
b
ilis
tic
N
e
u
ral
N
e
tw
o
r
k
(PN
N
)
,
N
a
ïve
Ba
ye
s
Pr
edic
t
o
r
a
nd
D
e
c
i
sio
n
T
ree
P
r
edic
tor.
T
he
m
otor
opera
tio
nal
da
ta
set
co
n
t
ai
ns
t
he
m
ot
or
s
li
p
,
t
he
n
e
g
a
t
iv
e
a
nd
p
os
i
tive
se
que
nc
e
c
u
rre
nt
a
nd
t
o
rq
ue,
t
h
e
rotor
and
s
t
a
t
or
c
urrent
p
e
r
p
hase,
the
t
o
t
a
l
ro
tor
and
s
t
a
t
or
r
e
s
i
s
t
i
v
e
c
o
p
p
e
r
l
os
se
s,
t
he
r
eal,
rea
c
tive
a
nd
a
ppa
ren
t
i
n
p
u
t
p
o
w
er,
th
e
a
i
r
g
a
p
po
wer
a
n
d
th
e
e
l
ect
ro
me
ch
an
i
c
al
p
o
w
er.
T
he
v
olta
ge
s
up
pl
y
st
atu
s
f
or
e
ac
h
s
a
m
p
l
e
ab
b
c
c
a
La
vg
(V
+V
+
V
)
W
h
ere V
=
3
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SSN: 2088-
8694
I
nt
J
P
ow
Elec
& Dr
i
S
y
st, Vol. 10,
N
o.
1, Mar
c
h 2
0
1
9
:
93
–
103
96
c
a
s
e
w
a
s
c
l
a
s
s
i
f
i
e
d
a
s
B
V
,
1
%
U
B
,
2
%
U
B
,
3
%
U
B
,
4
%
U
B
,
a
n
d
5
%
U
B
i
n
the
da
tase
t.
246
da
ta
s
a
m
p
l
e
s
f
or
ea
ch
o
f
t
h
e
pa
ram
e
ter
s
w
as
s
hare
d
i
n
t
he
r
a
tio
7
:3
u
sin
g
s
tra
t
i
fie
d
s
am
plin
g
;
7
0%
f
or
t
ra
in
i
n
g
an
d
30
%
for
pred
ic
ti
ve
e
val
u
a
t
i
on.
F
i
gure
3.
S
upp
ly v
o
l
t
a
ge
s
ta
t
u
s pr
edic
ti
ve K
N
I
ME w
orkflo
w
5.
RESULT
S
A
N
D
DISCU
SSIO
N
Th
e
d
e
scrip
t
ive
st
a
t
i
s
ti
cs
o
f
t
h
e
v
a
lu
e
s
o
f
the
mo
to
r
p
a
ra
me
t
e
r
s
ar
e
pr
ese
n
te
d
i
n
T
a
b
le
1
.
The
da
ta
mi
n
i
ng
w
o
r
k
f
lo
w
i
m
pl
emen
ted
,
d
ev
el
op
ed
v
ario
u
s
s
t
a
ti
st
ic
al
p
rop
ert
i
es
f
or
e
ach
o
f
t
h
e
pa
ram
e
ter
s
a
n
d
usi
ng
the
u
n
i
que
ne
ss
o
f
eac
h,
a
r
ep
re
senta
t
i
v
e
m
o
del
w
a
s
a
u
tom
a
tica
l
l
y
co
mp
ut
ed
w
hi
ch
d
e
p
ic
t
s
t
h
e
r
e
l
ati
onsh
i
p
betw
ee
n the
v
o
l
ta
ge
sta
t
u
s an
d
the m
o
t
o
r pa
ram
e
ter
s
.
Tab
l
e
1.
D
escr
i
p
t
i
ve
S
tat
i
stic
s
of t
he
M
otor
P
ara
m
e
t
e
r
M
in
M
ax
M
e
a
n
S
t
d
. d
ev
i
a
t
i
o
n
V
ar
i
a
n
c
e
S
k
e
w
n
e
s
s
K
u
r
t
o
s
i
s
S
lip
0
1
0
.
5
0
.
296
4
0.
0879
0
-
1.
20
14
Ise
qpos
(
A
)
35.
089
4
259.
10
1
95.
95
63.
17
3990.
3
8
-1.
09
0.
08
Ise
qne
g
(A)
0
8.
06
3
.
9
2
2.
68
7.
20
0
.
0
0
-
1
.
2
6
Ira
(
A
)
20.
105
9
246.
78
1
86.
64
61.
28
3755.
2
1
-1.
15
0.
29
Ir
b
(
A
)
20.
105
9
252.
87
1
87.
57
63.
07
3977.
9
0
-1.
14
0.
25
Irc
(
A
)
14.
133
1
246.
68
1
81.
77
62.
35
3887.
9
0
-1.
13
0.
24
Isa
(
A
)
35.
089
4
259.
21
1
97.
04
62.
58
3916.
7
4
-1.
10
0.
13
Isb
(A
)
35.
089
4
265.
61
1
98.
69
63.
53
4036.
3
8
-1.
07
0.
05
Isc
(
A
)
27.
708
259.
10
1
92.
23
63.
41
4020.
5
9
-1.
08
0.
07
P
r
-
T
ota
l
(
W
)
336.
57
83
5070
6.
82
3
181
3.
92
15866.
79
2517
549
86.
50
-
0.
62
-
0
.
9
5
P
s
-
Tot
a
l
(
W
)
890.
91
39
4861
7.
81
3
067
5.
13
15047.
41
2264
246
40.
37
-0.
61
-
0
.
9
5
P
i
n
(W)
1435
4.
047
4
1064
61.
27
9
212
5.
81
22812.
30
5204
009
20.
13
-2.
07
3.
35
P
i
n
(VA
R
)
2073
9.
461
6
1577
32.
27
1
049
19.
53
43619.
64
1902
672
799.
81
-
0.
58
-
1
.
0
1
P
i
n
(VA
)
2522
2.
290
6
1863
99.
71
1
409
07.
65
45407.
87
2061
874
313.
23
-1.
09
0.
08
A
i
rga
p
P
ow
e
r
1346
3.
133
5
7587
2.
50
6
145
0.
68
13577.
87
1843
584
66.
84
-
1.
56
2.
89
E
l
ec
t
Mec
h
P
owe
r
0
5492
0.
08
2
963
6.
76
17886.
50
3199
267
09.
18
-0.
0
8
-
1
.
3
7
P
o
s
Se
q T
(Nm
)
85.
709
482.
84
3
92.
89
86.
83
7539.
9
5
-1.
63
3.
05
N
e
g S
e
q
T
(Nm
)
-
0.
26
43
0
.
0
0
-
0
.
0
7
0.
07
0.
01
-
0.
82
-
0
.
4
8
pf
0
.
5329
0
.
8
6
0
.
68
0
.
11
0.
01
0
.
2
8
-
1
.
3
2
The
sta
tis
tic
al
v
aria
t
i
o
n
s
o
f
t
he
m
otor
’s
r
ot
or,
stat
or
a
nd
se
q
u
e
n
c
e
c
u
rren
t
s
in
a
mp
e
r
e
f
o
r
al
l
th
e
vo
lta
ge
s
u
p
p
l
y
m
odes,
b
o
t
h
ba
la
nce
d
a
nd
un
ba
lanc
e
d
a
r
e
s
how
n
in
F
i
gure
4.
T
he
b
ox
plo
t
s
r
e
vea
l
t
he
minim
u
m,
t
he
l
ow
e
r
q
uart
i
l
e
,
t
he
m
ed
ian,
t
h
e
u
pper
qua
r
til
e
a
n
d
t
he
m
axi
m
um
v
al
ue
s
fo
r
eac
h
o
f
t
he
c
urre
nt
para
me
ters.
In F
igure
5,
t
he re
a
l
(W), rea
c
ti
v
e
(
V
A
R
), appar
e
nt
(
V
A
)
,
a
i
r
gap
(W)
a
n
d
e
l
e
c
trom
ag
net
i
c po
w
e
r
(W)
of
t
he m
ot
or is dis
p
laye
d
as a
box p
l
o
t
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t
J
P
o
w
Elec
&
D
r
i
S
y
st
I
S
S
N
:
2088-
86
94
D
e
te
rm
i
n
i
n
g th
e
oper
a
t
i
on
a
l
sta
t
us
o
f
a three
phas
e
i
n
d
u
c
t
i
o
n m
o
t
o
r us
in
g …
(
A
deri
b
i
g
b
e
I
srael
Adeki
t
a
n)
97
F
i
gur
e
4.
A
box
plo
t
s
h
o
w
i
ng
t
he
m
agni
tu
de
s
pr
e
a
d
of
t
he
m
otor
curr
ents
F
i
gur
e
5.
A
box
plo
t
s
h
o
w
i
ng
t
he
m
agni
tu
de
s
pr
e
a
d
of
t
he
m
otor
’
s
pow
er
c
om
po
ne
nt
s
I
n
F
ig
ur
e
6,
t
h
e
s
ta
t
i
st
ica
l
s
pr
ea
d
of
t
he
v
a
l
u
e
s
o
f
t
he
r
otor
c
o
pp
er
l
o
s
se
s
an
d
th
e
st
ato
r
c
o
p
p
e
r
l
o
sse
s
in
w
a
tt
is
p
re
se
nte
d
a
s
a
bo
x
p
l
o
t
.
The
ro
to
r
losse
s
i
n
cr
ea
sed
fr
om
3
36.
58
W
t
o
5
0
7
06.
82
W
w
i
t
h
i
n
c
r
ea
sin
g
sl
ip
a
n
d
v
olta
g
e
u
n
b
a
l
a
n
c
e
,
whi
l
e
the
t
o
tal
s
t
at
or
w
i
ndi
n
g
c
op
p
er
l
osse
s
i
n
cr
e
a
se
d
fr
om
8
90.
9
1
W
t
o
48
6
17.
8
1
W
.
Fig
u
r
e
7
p
r
e
se
nt
s
t
h
e
var
i
a
tion
in
t
he
m
a
g
n
itu
de
o
f
th
e
po
si
tive
a
n
d
ne
ga
ti
ve
s
e
que
nc
e
tor
que
i
n
N
m
.
The var
i
a
tio
n
o
f
t
he
n
e
g
a
tive
s
e
que
nc
e
tor
q
ue
i
n
N
m
f
or
t
he
B
V
,
1
%
U
B,
2
%
U
B,
3
%
U
B,
4
%
U
B
a
n
d
5%U
B
v
o
l
tag
e
c
on
dit
i
ons
i
s
di
sp
la
yed
in
F
i
gur
e
8.
T
he
b
o
x
p
lo
t
re
v
e
a
l
s
t
h
at
a
t
5%U
B
t
her
e
i
s
a
si
gn
i
f
ican
t
incr
ease
in
t
he
m
agn
i
t
u
de
o
f
t
h
e
n
e
ga
t
i
ve
s
e
que
nc
e
t
o
r
q
ue
a
s
c
o
mpa
r
e
d
w
it
h
t
h
e
va
l
u
e
w
h
en
t
he
v
o
l
ta
ge
w
as
ba
l
a
nce
d
.
S
i
milar
l
y,
F
igur
e
9
pr
esents
a
b
ox
p
l
o
t
o
f
t
h
e
seque
n
c
e
c
u
r
r
e
n
t
(
A
)
f
o
r
t
h
e
B
V
,
1
%
U
B
,
2
%
U
B
,
3%U
B
,
4%U
B
a
nd
5%
U
B
vol
ta
ge
c
o
n
d
i
tio
ns.
The
seq
u
e
n
ce
cur
r
e
nt
h
a
s
t
h
e
max
i
mu
m
v
a
lu
e
at
5
%
vo
lt
ag
e
un
ba
la
nce
c
o
n
d
i
tio
n.
The
c
h
an
ges
i
n
t
h
e
r
o
t
or
w
ind
i
ng
c
o
pper
l
o
sses
for
the
B
V
,
1%U
B
,
2%UB,
3
%
UB
,
4
%
UB
a
nd
5%U
B
v
olta
ge
c
on
dit
i
o
ns
i
s
di
sp
l
a
yed
in
t
h
e
b
o
x
p
l
o
t
of
F
igur
e
10.
F
ig
ur
e
1
1
d
e
t
a
ils
t
he
v
a
r
ia
t
i
ons
i
n
t
h
e
sta
t
or
w
i
n
di
ng
c
o
pper
l
o
sse
s
f
or
t
he
b
a
l
ance
d
vo
l
t
age
(
B
V
)
a
n
d
t
h
e
u
n
b
a
la
nce
d
(
1%U
B
t
o
5
%
U
B
)
vo
l
t
age
co
n
d
iti
ons.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
S
S
N: 2
0
8
8
-
86
94
I
nt
J
P
ow
E
l
e
c
&
Dr
i
S
y
st,
Vol.
10,
N
o.
1
,
Mar
c
h
2
0
1
9
:
93
–
10
3
98
F
i
gur
e
6.
A
b
o
x
p
l
o
t
of
t
he
r
otor
a
n
d
s
tat
o
r
w
i
n
d
in
g
cop
p
e
r
l
os
ses
F
i
gur
e
7.
A
box
plo
t
s
h
o
w
i
ng
t
he
m
agni
tu
de
s
pr
ea
d
of
t
he
pos
i
t
i
v
e
and
n
eg
ativ
e
sequ
en
c
e
to
r
qu
e
F
i
gur
e
8.
A
box
pl
o
t
s
h
o
w
i
ng
t
he
n
e
g
a
t
ive
seque
nce
t
o
r
q
ue
i
n
N
m
f
r
o
m
balance
d
t
o
5%
u
n
b
a
l
anc
e
d
v
o
l
tage
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t
J
P
o
w
Elec
&
D
r
i
S
y
st
I
S
S
N
:
2088-
86
94
D
e
te
rm
i
n
i
n
g th
e
oper
a
t
i
on
a
l
sta
t
us
o
f
a three
phas
e
i
n
d
u
c
t
i
o
n m
o
t
o
r us
in
g …
(
A
deri
b
i
g
b
e
I
srael
Adeki
t
a
n)
99
F
i
gur
e
9.
A
b
ox
pl
o
t
s
h
o
w
i
ng
t
he
n
ega
t
i
v
e
se
que
nc
e
c
u
r
r
e
nt
i
n
N
m
f
or
b
ala
n
ce
d
to
5
%
u
nba
l
a
nce
d
v
o
l
t
a
ge
F
i
gur
e
1
0
.
A
box
p
l
ot
o
f
t
h
e
rot
o
r
c
o
p
p
er
l
osse
s
w
ith
i
ncr
e
a
s
i
n
g
u
n
b
ala
n
c
e
vol
ta
ge
F
i
gur
e
1
1
.
A
box
p
l
ot
o
f
t
h
e
stat
or
c
op
per
lo
sses
w
ith
i
ncr
e
asi
ng
u
n
b
a
l
a
n
c
e
volta
g
e
Evaluation Warning : The document was created with Spire.PDF for Python.
I
S
S
N: 2
0
8
8
-
86
94
I
nt
J
P
ow
E
l
e
c
&
Dr
i
S
y
st,
Vol.
10,
N
o.
1
,
Mar
c
h
2
0
1
9
:
93
–
10
3
10
0
5.
1.
T
h
e
n
aïve
b
ayes
p
red
ict
o
r
r
e
su
l
t
s
Usin
g
the
Naï
v
e
Ba
ye
s
pre
d
ict
o
r
a
n
acc
ur
ac
y
o
f
9
8.
6
4
9
%
w
as
a
c
h
i
ev
ed.
The
sc
at
t
e
r
plo
t
o
f
t
h
e
cl
assi
fi
e
d
s
amp
l
es
a
s
sho
w
n
i
n
F
i
g
u
r
e
12
.
Th
e
co
nf
usion
ma
t
r
i
x
of
t
he
N
a
ï
ve
B
a
y
es
p
r
e
dic
t
or
i
s
pr
esente
d
i
n
Ta
b
l
e
2
.
O
ut
o
f
the
t
o
t
a
l
73
s
am
ples
r
an
d
o
m
l
y
se
lec
t
e
d
f
or
p
er
f
or
m
a
nce
e
v
al
ua
ti
o
n
,
on
ly
o
ne
s
am
p
l
e
w
a
s
m
i
sclass
ifie
d.
F
igure
13
s
h
o
w
s
t
h
e
ROC
c
u
rve
for
t
h
e
1
0
0
%
corr
e
ct
l
y
p
redic
t
ed
B
V
sa
mple
s
wh
i
l
e
Fi
gure
1
4
pr
esen
ts
t
he
R
C
c
u
r
v
e
for
t
h
e
2%
u
n
b
a
l
a
n
c
e
v
o
lta
ge
p
r
e
dic
t
io
n
w
hi
c
h
h
a
s
94.
2%
acc
ur
ac
y
du
e
to
t
he
m
i
sclass
ifica
t
i
on o
f
a
s
a
m
ple
.
F
i
g
u
r
e
12
. S
cat
t
er p
lo
t of
t
h
e
cl
a
ss
i
f
i
e
d
samp
les
Ta
b
l
e
2.
C
on
f
u
sion
Ma
tr
i
x
o
f
th
e
N
a
ï
v
e
Ba
y
e
s
Pr
edict
o
r
B
V
1%
UB
2
%
U
B
3%
UB
4
%
U
B
5%
U
B
B
V
1
2
0
0
0
0
0
1%UB
0
1
2
0
0
0
0
2%UB
0
0
1
1
1
0
0
3%UB
0
0
0
1
2
0
0
4
%
U
B
0
0
0
0
1
2
0
5
%
U
B
0
0
0
0
0
1
4
F
i
gur
e
1
3
.
R
O
C
cur
v
e
sh
ow
i
ng
t
h
e
acc
ur
ac
y
of
t
he
b
ala
n
ce
d vo
l
t
a
g
e
sam
p
l
e
cla
ssific
a
tio
ns
F
i
gur
e
1
4
.
RO
C
cur
v
e
sh
ow
i
ng
t
h
e
acc
ur
ac
y
of
t
he
2%UB sam
p
l
e c
l
ass
i
f
i
c
a
t
i
o
ns
Evaluation Warning : The document was created with Spire.PDF for Python.
Int J
P
o
w
El
e
c
&
D
ri S
yst
I
S
S
N
:
2088-
86
94
D
e
term
i
n
in
g t
h
e
ope
rat
i
o
n
a
l
st
a
t
us
o
f
a th
re
e
phas
e
in
d
u
ct
io
n m
o
tor usin
g …
(Aderi
b
i
g
b
e
I
s
rae
l
Adek
it
a
n
)
10
1
5.2.
The
de
c
isi
o
n t
ree p
redictor
r
e
sults
The
co
n
f
us
i
o
n
m
a
trix
o
f
the
D
ecisi
on
Tr
ee
pre
d
ic
t
o
r
i
s
p
re
se
n
t
e
d
in
T
able
3
.
A
l
l
the
74
sa
mp
le
s
rand
om
ly
s
e
l
ec
te
d
for
per
f
orm
a
nce
eva
l
ua
ti
on
w
e
re
a
ccur
a
tel
y
c
l
a
ssi
fied
a
s
show
n
b
y
t
he
d
iag
o
n
al
e
l
e
me
nt
s
o
f
T
a
b
l
e
3
.
Tab
l
e 3.
Con
fu
si
o
n
m
atrix o
f
t
he
de
c
i
s
io
n
t
r
ee
predic
tor
B
V
1
%
UB
2
%UB
3
%
UB
4
%
U
B
5
%
UB
B
V
1
2
0
0
0
0
0
1%U
B
0
1
2
0
0
0
0
2%U
B
0
0
1
2
0
0
0
3%U
B
0
0
0
1
2
0
0
4
%
U
B
0
0
0
0
1
2
0
5
%
U
B
0
0
0
0
0
1
4
5.3.
Th
e
P
N
N p
r
e
d
ictor
re
su
lts
V
o
l
t
age
s
u
p
p
l
y
s
t
at
us
p
red
i
cti
o
n
us
i
n
g
tra
i
ne
d
P
r
o
b
ab
i
l
i
s
t
i
c
N
e
ur
al
N
etwork
n
o
d
e
was
a
l
s
o
perform
ed,
the
con
f
usi
on
ma
tri
x
o
f
t
h
e
P
N
N
predict
o
r
is
p
r
e
se
nt
ed
i
n
Ta
bl
e
4
.
O
f
al
l
t
h
e
73
s
amp
l
e
s
rand
om
ly sel
ec
ted o
n
l
y o
n
e
w
a
s
miscla
ssi
f
ie
d.
Tab
l
e
4. Con
fu
si
on m
a
trix o
f t
h
e P
N
N
predic
tor
B
V
1
%
UB
2
%UB
3
%
UB
4
%
U
B
5
%
UB
B
V
1
2
0
0
0
0
0
1%U
B
0
1
2
0
0
0
0
2%U
B
0
0
1
2
0
0
0
3%U
B
0
0
0
1
2
0
0
4
%
U
B
0
0
0
0
1
1
1
5
%
U
B
0
0
0
0
0
1
4
5.4.
S
u
mmar
y
of
mod
e
l p
r
e
d
iction
s
Th
e
co
mp
a
r
a
t
i
v
e
p
e
rfo
r
ma
n
ce
o
f
t
he
t
h
r
e
e
p
re
di
c
t
o
r
s
i
s
p
re
se
nt
e
d
i
n
T
a
b
l
e
5
.
The
d
e
cisio
n
t
re
e
pred
ic
tor ha
d the
h
i
g
h
est pe
rform
a
n
ce
w
i
t
h
a
cc
u
r
ac
y
o
f
100
%
fo
r
the
BV
,
1%U
B,
2
%U
B,
3%U
B,
4%
U
B an
d
5%U
B
v
ol
t
a
ge
s
a
m
ples
c
ons
ide
r
ed.
The
ac
c
u
ra
cy
o
f
the
m
ode
l
is
s
i
g
n
i
f
i
c
a
n
t
l
y
h
i
g
h
b
e
c
a
u
s
e
a
l
o
t
o
f
m
o
t
o
r
opera
tio
nal
pa
ra
me
ters
w
ere
c
o
nside
r
ed
i
n
t
h
e
mode
l
.
A
l
l
t
he
s
i
m
u
la
t
e
d
pa
ram
e
ter
s
m
ay
n
ot
b
e
re
a
d
il
y
ava
ila
b
l
e
or
e
asy
t
o
m
easure
in
p
ra
c
tica
l
s
t
u
d
i
e
s
,
a
nd
a
s
s
uc
h,
t
h
e
e
xp
e
c
t
e
d
a
ccu
rac
y
f
o
r
e
xp
eri
m
en
t
a
l
l
y
gene
ra
ted
da
ta
set w
i
l
l
be
q
u
i
t
e
low
e
r
.
Tab
l
e
5.
C
om
p
a
riso
n of t
he
p
erfor
m
a
n
ce
of
t
he
t
hree
da
t
a
m
i
nin
g
p
re
di
c
t
o
r
s
N
aï
v
e
B
ay
es
D
e
c
i
s
i
o
n
Tr
e
e
P
N
N
P
r
e
d
i
ct
o
r
C
o
r
r
e
c
t
Cl
a
s
s
i
f
i
e
d
7
3
7
4
7
3
Ac
c
u
r
a
cy
98.
649%
100%
9
8.
649%
C
ohe
n’s
Ka
pp
a
(
k
)
0.
984
1
0.
984
W
r
ong
C
l
a
ssifie
d
1
0
1
Erro
r
1.
351%
0
%
1.
351%
6.
CONCL
U
S
ION
I
n
t
h
i
s
s
t
ud
y,
d
ata
mi
ni
n
g
w
a
s
a
pp
lie
d
t
o
a
cqu
i
re
k
n
o
w
l
e
dge
f
ro
m
the
da
t
a
set
ge
nera
te
d
from
t
h
e
si
m
u
late
d
ope
r
a
ti
o
n
o
f
a
t
h
re
e
pha
se
i
nd
u
c
ti
on
mot
o
r
u
n
d
e
r
ba
l
a
nce
d
a
nd
un
ba
la
nc
ed
v
o
lta
ge
s
u
p
p
l
y.
A
pred
ic
ti
ve
K
N
I
ME
m
ode
l
w
a
s
deve
l
ope
d
a
n
d
thr
ee
da
t
a
m
i
n
i
ng
a
l
g
o
ri
t
h
m
s
;
the
Naï
v
e
Ba
ye
s,
D
ec
isi
o
n
Tre
e
and
P
N
N
Pre
d
ic
t
o
r
w
e
re
t
r
a
ine
d
u
s
i
n
g
7
0%
o
f
t
h
e
to
t
a
l
sam
p
les
w
h
i
ch
w
ere
ra
nd
o
m
l
y
s
e
l
ect
ed
.
Th
e
k
n
o
w
l
e
dg
e
acq
ui
re
d
f
r
o
m
t
h
e
t
rai
n
ing
w
a
s
appl
i
e
d
in
p
redi
ct
in
g
t
h
e
t
ype
o
f
s
u
pp
ly
t
ha
t
pr
o
d
u
c
ed
t
h
e
rem
a
ini
ng 30
% of the
m
ot
or ope
rat
i
ona
l da
ta
sam
ples. Th
e
t
hr
ee
pre
d
ic
tor
s
had acc
urac
ies o
f
9
8.64
9
%
,
10
0
%
and
9
8
.64
9
%
r
e
spect
ive
l
y
w
h
ic
h
i
n
di
c
a
t
es
t
ha
t
t
h
e
mode
l
w
a
s
a
d
e
q
u
a
t
e
l
y
a
bl
e
to
a
cqui
re
s
u
f
f
i
cient
kn
ow
le
d
g
e
fr
o
m
t
he
o
pe
rat
i
o
n
al
m
ot
or
d
a
t
a
s
et,
a
nd
th
is
e
n
a
ble
d
th
e
c
o
rrec
t
p
re
di
c
tio
n
of
t
h
e
t
y
p
e
o
f
vo
lt
ag
e
sup
p
l
y
c
l
a
ssi
fi
e
d
a
s
ba
la
nc
e
d
(
BV
),
a
nd
u
nba
la
nc
ed
(
1%
U
B
,
2%U
B
,
3
%U
B,
4
%U
B
and
5%U
B
)
v
o
l
t
a
g
e
sup
p
l
y
.
The
mode
l
deve
lope
d
w
a
s
e
xpor
ted
usin
g
t
h
e
P
M
ML
w
rit
e
r
a
nd
t
h
i
s
c
re
a
t
es
a
n
o
p
p
o
r
t
uni
t
y
f
or
r
eu
se
eve
n
o
n
o
t
her
pla
t
form
s.
T
he
p
re
di
c
t
i
v
e
a
c
c
u
r
acy
a
ch
ie
ve
d
in
t
h
is
w
ork
i
s
i
n
d
i
c
a
t
i
ve
o
f
the
su
i
t
ab
i
lit
y
o
f
d
a
t
a
mini
n
g
a
ppr
oa
ch
f
or
m
ot
or
p
e
r
form
anc
e
m
on
i
t
or
i
n
g.
T
h
i
s
stud
y
op
e
n
s
u
p
f
urt
h
e
r
r
e
s
ear
ch
o
pp
ort
u
n
i
ti
es
f
or
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SSN: 2088-
8694
I
nt
J
P
ow
Elec
& Dr
i
S
y
st, Vol. 10,
N
o.
1, Mar
c
h 2
0
1
9
:
93
–
103
10
2
dep
l
oy
ing
sim
ilar
da
ta
m
i
n
in
g
m
o
dels
o
n
p
r
ac
t
i
c
a
l
m
o
t
ors
for
vo
lta
ge
q
ual
i
t
y
m
on
i
t
or
ing
us
i
n
g
rea
l
t
im
e
motor
o
p
era
tio
na
l
da
ta.
ACKNOW
LEDG
E
MEN
T
S
The
p
u
b
lica
tion
of
t
h
i
s
pa
per
w
a
s
spon
sored
b
y
C
o
v
e
nan
t
U
n
i
vers
it
y
Ce
n
t
r
e
f
or
R
ese
a
rc
h,
In
no
v
a
ti
on
a
nd Di
s
c
o
v
e
ry
.
REFE
RENCES
[1]
K.
V
i
n
o
t
h
Kumar,
S
.
Su
resh
K
um
ar,
an
d
A.
I
m
m
a
n
u
el
S
e
l
v
a
kum
ar,
"
A
rev
i
ew
o
f
v
o
l
t
ag
e
and
cu
rren
t
si
gnature
d
i
a
g
no
sis
in
i
nd
ustr
ia
l
d
r
iv
e
s
,"
Int
e
rna
t
i
o
n
a
l
Jou
r
n
a
l
o
f
P
o
wer
El
ectro
n
i
cs
and
D
r
i
ve Sys
t
ems
(IJP
E
DS)
,
v
o
l
.
1
,
pp
.
7
5-82
,
2
0
1
1
.
[2]
Z.
M
.
S.
E
l-Bar
b
ary,
"
Si
ngle-to-three
phase
ind
u
ct
ion
mo
t
o
r
se
n
sor
l
ess
dr
ive
system,
"
Al
exan
d
r
i
a E
ngineer
in
g
J
o
urna
l
, v
ol
.
5
1
,
p
p. 77
-
83
, 20
1
2
.
[3]
O.
O
lat
unji,
S
.
Aki
n
l
a
bi
,
A.
O
lus
e
yi,
A.
A
b
i
oye,
F
.
I
s
hola,
M
.
P
e
ter,
e
t
al.,
"Electric
P
o
wer
Crisi
s
i
n
Nig
e
ria:
A
S
t
rat
e
gic
Call
f
o
r
Ch
ange
o
f
F
o
cu
s
to
R
enewab
le
S
o
u
rces
,"
i
n
2
n
d
In
ter
natio
na
l Co
nf
erence
on
Eng
i
n
eeri
ng fo
r
Su
st
a
i
nab
le W
o
rld
(
I
CESW
2
018)
,
20
18
.
[4]
A.
A
d
e
kitan,
A
.
Og
un
juyig
b
e,
R
.
A
y
o
d
ele,
a
nd
I
.
Sam
u
el,
"
T
he
p
e
rf
orm
a
n
c
e
of
a
3
-P
h
a
se
I
n
d
u
c
tio
n
Mach
in
e
un
de
r
Un
balan
ce Vo
lt
a
g
e Reg
i
m
e
,"
Jou
r
n
a
l of En
g
i
neering
Sci
e
nce a
n
d
T
echn
o
lo
gy
Review
,
vol.
1
0
,
pp.
13
6
-1
43
,
2
0
1
7
.
[5]
P
.
P
il
lay
a
n
d
M
.
M
an
yage,
"
D
efi
n
iti
ons
o
f
v
o
lt
age
u
nbal
a
nce,
"
IEE
E
Po
wer En
gi
neer
ing Revi
ew
,
vol
.
2
1
,
pp
.
5
0-51
,
2
0
0
1
.
[6]
J.
F
a
i
z
and
H
.
E
brah
im
p
our,
"
P
reci
se
d
era
t
in
g
o
f
t
h
r
ee-phas
e
i
n
duct
ion
motors
w
it
h
u
nbal
a
n
ced
v
o
ltag
e
s,
"
in
Fo
urtie
t
h
IAS An
nu
al M
eeting.
Co
nf
erence
Recor
d
of
th
e 20
05
Ind
u
s
t
r
y
Ap
pl
icatio
n
s
Confer
ence
,
v
o
l
.
1,
pp
.
4
85-4
9
1
,
2
0
0
5.
[7]
P
.
G
nacin
s
ki,
"Eff
ect
o
f
u
n
b
a
lanced
v
olt
a
ge
on
wi
nd
ing
s
t
em
pera
ture,
op
e
r
at
io
nal
lif
e
and
l
o
ad
car
ryi
ng
capacit
y
o
f
in
du
ctio
n mach
in
e,"
En
ergy Conver
si
on an
d
M
a
n
a
gem
e
nt
,
vol.
49,
p
p
.
7
61
-770
,
200
8.
[8]
B.
N
.
Gafford
,
W
.
C
.
Duesterh
o
e
f
t
,
an
d
C.
C
.
Moshe
r
,
"He
a
t
i
n
g
o
f
In
du
ctio
n
M
o
tors
on
U
nbalan
c
ed
,
"
P
o
w
e
r
Apparatus
and
Syst
e
m
s
,
P
ar
t
III.
T
r
ansact
ions of t
h
e
A
m
eric
an Institute of
Elect
r
i
cal Eng
i
neers
,
vo
l.
7
8,
pp
.
2
82-2
8
6
,
1
9
5
9.
[9]
A.
I
.
Adek
it
an,
B.
A
det
o
k
u
n
,
T
.
S
h
om
ef
un,
a
nd
A
.
Ali
g
b
e
,
"Co
s
t
im
p
l
i
c
atio
n
o
f
L
in
e
Volt
age
vari
ati
o
n
on
T
h
ree
P
h
as
e
In
du
ctio
n
M
o
tor
o
p
erati
on,"
T
E
L
K
OM
NIKA
(T
el
ecomm
unica
ti
on Co
mputi
n
g
Elect
ro
n
i
c
s
and Con
t
r
o
l)
,
v
o
l
.
16
,
p
p.
1
40
4-1
4
1
2
,
2
01
8.
[10]
J.
F
ai
z,
E
.
H.,
and
P.
P
illay,
"
Influence
of
U
nbal
a
n
ced
V
olt
a
ge
o
n
the
S
t
ead
y-S
t
ate
P
e
rf
orm
a
nce
of
a
T
h
r
ee
-P
h
a
se
S
q
u
i
rrel-Cag
e
I
n
d
uct
i
on
M
ot
or,
"
IEE
E
Tran
sac
t
io
ns
on
En
e
r
g
y
Co
n
v
e
r
sion
,
v
o
l
. 19
,
p
p.
p
p.65
7
-
62
, 2
00
4
.
[11]
C.-Y.
Lee,
B
.
-
K.
C
h
e
n,
W
.-J.
L
ee,
a
nd
Y
.-F
.
H
su,
"Eff
ects
of
v
a
riou
s
u
n
b
a
la
n
c
e
d
v
olta
ge
s
on
t
h
e
o
pe
r
a
tion
perf
o
r
m
a
n
ce
of
a
n
i
n
d
u
cti
o
n
m
o
t
o
r
und
er
t
he
s
am
e
vo
lt
age
u
n
b
a
la
nce
fact
or
c
ondition
,
"
E
l
ectric Power
S
y
s
t
em
s
Research
, vo
l
.
4
7
, pp
. 1
53
-1
63
,
1
9
9
8
.
[12]
P
.
Pi
l
lay
,
P
.
H
ofman
n
,
a
nd
M
.
M
a
ny
age,
"
D
e
rat
i
ng o
f
in
d
u
c
ti
on
m
ot
ors operat
i
n
g
w
it
h a com
b
i
n
at
ion
o
f
un
b
alan
ce
d
vo
lt
ages
a
nd
ove
r
or
u
nd
e
r
v
o
ltag
e
s,
"
IEEE T
r
ans.
Energy Con
s
ervation
,
vol.
17,
p
p
.
4
85
-
49
1,
2002.
[13]
Eb
adi,
S
.
A
l
i,
M
i
r
zaie,
M
.
Zareinej
ad,
an
d
S
.
Asgh
a
r
Gh
olam
i
a
n,
T
orqu
e
anal
ysi
s
o
f
th
ree-p
h
as
e
ind
u
ct
io
n
m
o
t
o
r
un
der
v
o
l
t
age
u
n
balan
ce
us
in
g
2D
F
EM,
Inter
n
a
tio
nal Jo
ur
nal
o
f
Eng
i
n
eeri
ng Sci
e
nce a
nd
T
ech
no
lo
gy
(
I
J
E
ST)
,
vo
l.
3
,
p
p
.
8
7
1
-876,
2
0
1
1
.
[14]
C.
L
.
F
o
rt
escue,
"
M
e
th
od
o
f
s
y
mm
etri
cal
c
oo
rdinates
a
p
p
l
i
ed
t
o
t
h
e
so
lu
tio
n
o
f
p
oly
p
h
a
se
n
e
t
w
o
r
k
s,"
A
I
EE
Tr
ans
,
vo
l.
37,
p
p
.
1
02
7
-
14
0,
1
9
18.
[15]
A.
S
i
v
akum
ar
a
n
d
N
.
B.
M
uthu
S
e
lv
an,
"Red
u
c
tio
n
of
s
ou
rce
cu
rre
n
t
h
a
r
m
o
n
i
c
s
i
n
A
N
N
c
o
ntro
lle
d
in
du
c
tion
mo
to
r
,
"
Alexan
d
r
i
a En
g
i
n
eeri
ng
Jou
r
n
a
l
,
v
o
l
.
57,
p
p
.
1
48
9-14
99,
S
ep
tem
b
er
2
01
8.
[16]
W.
H
.
Kers
t
i
ng
a
nd
W
.
H.
P
h
illips
,
"
Phase
Frame
Anal
ysi
s
o
f
t
h
e
E
ff
ect
s
o
f
V
o
l
t
a
ge
U
n
b
alan
ce
on
In
du
ct
io
n
M
achi
n
es,
"
IEE
E
Tran
sa
c
t
io
n
s
on
I
n
d
u
stry
Ap
pl
ic
at
io
ns
,
vol.
3
2
,
p
p
.
pp.
4
15-2
0
,
19
97.
[17]
J.
F
a
i
z
an
d
H.
E
brah
im
po
ur,
"P
recise
d
eratin
g
of
t
hree
ph
ase
in
du
c
t
io
n
m
o
t
o
rs
w
it
h
un
ba
l
a
nced
v
ol
tages
,
"
En
e
r
g
y
Co
nv
e
rsio
n
a
n
d
Ma
na
ge
me
n
t
,
v
o
l
.
4
8
,
p
p
.
2
5
79-258
6,
200
7.
[18]
A.
I
.
Adekit
an,
"S
up
pl
y
inst
ab
i
l
ity
i
nd
uced
t
o
r
q
u
e
v
a
riat
ion
s
o
f
a
three
p
h
as
e
a
s
yn
chro
nou
s
m
o
tor,"
International
Jou
r
n
a
l
o
f
M
echa
n
i
c
al
En
g
i
neering
an
d
T
ech
no
logy
, vo
l
. 9
, p
p
. 57
2
-5
8
3
, 20
1
8
.
[19]
P
.
A
m
a
i
z
e
,
K
.
O
.
I
g
n
a
t
i
u
s
,
E
.
S
.
O
l
u
w
a
s
o
g
o
,
A
.
S
.
A
l
a
y
a
n
d
e
,
a
n
d
A
.
E.
A
i
r
o
b
om
an,
"
I
n
f
luen
ce
of
P
ow
er
Q
u
a
l
i
t
y
P
r
ob
lem
on
t
he
P
erf
o
rm
an
ce
o
f
a
n
Ind
u
ct
ion
M
o
t
o
r,"
Am
erica
n
J
o
urn
a
l
o
f
Elect
r
i
cal P
o
wer
a
n
d
E
n
er
gy Sys
t
em
s
,
v
o
l.
4
,
pp
. 3
9-4
4
, 2
01
7.
[20]
S
.
S
hah
b
u
d
i
n
,
Z.
F
.
M
ohmad
,
S
.
I
.
S
u
l
i
m
a
n,
M
.
K
a
ssi
m
,
an
d
R.
M
o
ham
a
d
,
"
Class
i
f
i
catio
n
of
p
ower
qual
i
t
y
di
st
urban
ces
a
t
tran
sm
is
si
on
s
ystem
us
ing
sup
p
o
r
t
vect
or
m
achin
es,"
Ind
ones
i
a
n
Jou
r
na
l of
Elect
rical
E
ngi
neerin
g
an
d Co
mp
u
t
er
Sc
i
e
nce (
I
J
E
ECS
)
,
vol.
6,
p
p.
3
1
0
-3
17,
2
0
1
7
.
[21]
S
.
S
uraya,
P
.
S
.
S
uj
atha,
an
d
B.
K
um
ar,
"A
N
o
v
el
C
on
tro
l
S
tra
t
e
gy
f
o
r
Com
p
ensatio
n
o
f
V
olt
a
ge
Q
u
a
lit
y
P
roblem
in A
C Dr
ives,"
In
tern
atio
nal Jo
u
r
nal o
f
P
o
wer
Elec
t
r
o
n
i
c
s a
nd Drive S
y
stem
s (IJPED
S
)
,
vo
l.
9,
pp
.
8-1
6
, 2
01
8
.
[22]
R
.
M
y
u
n
g
,
H
.
Y
u
,
a
n
d
D
.
L
e
e
,
"
O
p
t
i
m
i
z
i
n
g
p
a
r
a
l
l
e
l
i
s
m
o
f
b
i
g
d
a
t
a
analy
t
i
c
s
at
d
i
s
tri
buted
c
o
m
p
u
tin
g
sy
stem
,"
Int
e
rna
t
i
o
n
a
l
Jo
u
r
n
a
l on
Ad
vanced
Sci
e
nce,
Eng
i
n
eeri
ng
an
d Informa
ti
on T
echn
o
lo
gy
,
v
o
l
. 7
, pp
.
1
71
6
-
17
21
, 20
1
7
.
Evaluation Warning : The document was created with Spire.PDF for Python.