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
)
V
o
l.
10, N
o.
1, Mar
ch 20
19,
p
p.
265~
2
7
6
IS
S
N
: 2088-
86
94,
D
O
I
:
10.11
59
1
/ij
ped
s
.
v10
.
i
1.pp
2
65-
27
6
265
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
GA-ANFIS PID compensated mode
l reference adap
tive control
for BLD
C
motor
Mu
rali
D
as
ari
1
, A
S
rin
i
vasu
l
a
Reddy
2
, M Vijaya Kumar
3
1,
3
D
e
partm
e
nt of E
l
ectrical and E
l
ectro
nics E
ng
i
n
eeri
n
g
, JNTU Col
l
ege
o
f
E
ngineeri
n
g
, India
2
CM
R
Eng
i
n
eerin
g
Co
lleg
e,
H
yderab
a
d,
T
elan
gan
a
,
I
nd
ia
Art
i
cl
e In
fo
ABSTRACT
A
r
tic
le hist
o
r
y
:
R
e
ce
i
v
e
d
Jun 17,
2018
Re
vise
d N
ov
5
, 2018
A
c
c
e
pte
d
D
ec 10,
2
0
1
8
Ad
apt
i
ve
c
o
n
trol
i
s
on
e
o
f
t
he
w
i
d
el
y
u
s
ed
c
on
tro
l
s
t
r
a
t
egi
e
s
t
o
desi
gn
adv
a
nc
ed
c
on
tro
l
s
ystems
f
o
r
b
e
t
ter
perf
orman
ce
an
d
ac
cu
r
acy.
M
odel
ref
e
rence
ad
apti
ve
c
ont
rol
(M
RAC)
i
s
a
di
rect
a
d
a
pti
v
e
strat
e
g
y
w
i
t
h
s
o
m
e
adj
u
s
t
ab
l
e
c
ont
rol
l
er
p
aram
eters
an
d
an
a
dj
us
ti
n
g
m
echan
is
m
t
o
adj
u
s
t
t
h
e
m
.
In
t
h
i
s
wo
rk
M
o
d
el
R
ef
erence
Ad
apti
ve
C
on
trol
f
or
B
L
D
C
m
o
t
o
rs
h
as
b
een
des
i
g
n
ed
w
it
h
a
P
I
D
con
t
roller
t
uned
by
G
A-ANF
IS.
GA
-Trained
A
N
FIS
f
r
am
ework
f
o
r
tu
ni
ng
t
he
P
ID
c
ont
roller
has
been
p
rop
o
s
e
d.
T
h
i
s
i
s
u
s
e
d
alo
n
g
w
i
t
h
t
he
M
RAC
t
o
d
e
l
iver
e
n
h
an
c
e
d
per
f
o
rman
ce
in
t
he
c
o
n
t
rol
of
BLD
C
m
oto
r
.
The
p
e
rfo
rm
ance
of
t
h
e
p
rop
o
s
e
d
app
r
oach
i
s
valid
at
ed
f
o
r
m
o
t
o
r
co
nt
ro
l
under
con
d
it
io
ns
o
f
chan
ge
i
n
sp
eed
,
chang
e
i
n
l
o
ad
,
ch
ang
e
i
n
in
erti
a
an
d
ch
an
ge
i
n
ph
ase
r
e
si
st
ance.
T
he
p
erf
o
rm
an
ce
is
v
ali
dated
agai
nst
con
v
en
tio
n
P
I
D
and
self
t
unin
g
P
ID
c
on
t
r
ollers
.
The
resu
lt
de
mo
n
s
t
r
ates
a
su
peri
or p
erfo
rmance o
f
th
e
pro
p
o
sed
appro
ach
K
eyw
ord
s
:
ANFI
S
BL
D
C
GA
MR
AC
PID
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:
Murali Dasari,
D
e
pa
rtme
nt
o
f
El
e
c
t
rica
l
and
El
ect
ro
ni
c
s
Eng
in
e
e
ring
,
JN
TU
Col
l
e
ge
o
f En
g
i
nee
r
in
g,
A
na
nth
pur
, A
ndhr
a,
India.
Em
ail:
Pra
d
esh,m
u
ralid
vr@g
m
a
il.c
o
m
1.
I
N
TR
OD
U
C
TI
O
N
C
o
n
t
ro
l
o
f
B
L
D
C
mot
o
r
has
alwa
ys
r
e
m
a
i
n
e
d
an
a
c
tive
a
r
ea
o
f
res
ea
rch.
E
spec
ial
l
y
t
he
P
ID
c
ont
r
o
l
of
B
LD
C
m
o
tor
has
been
s
t
u
d
i
e
d
e
x
t
e
n
s
i
v
e
l
y
f
or
t
he
o
p
t
imiz
a
tio
n
o
f
d
i
ffere
n
t
p
ar
am
eter
s
of
p
r
o
po
r
t
i
ona
l
gai
n
,
in
te
gra
l
t
ime
and
deriva
ti
ve
t
ime
.
R
ese
a
r
che
r
s
have
u
sed
d
i
ffe
r
en
t
o
p
t
i
m
i
za
tio
n
ap
proa
c
h
es
t
o
o
p
t
im
i
z
e
the
s
e
pa
ram
e
te
rs.
Rese
arc
h
ers
ha
ve
d
r
a
w
n
i
nsp
i
rat
i
on
fro
m
n
a
t
ur
al
l
y
o
cc
u
r
ri
n
g
ph
en
omen
a
i
n
s
ol
ving
t
h
e
se
op
tim
iza
t
i
o
n
p
r
ob
lem
s
.
m
i
m
i
cki
n
g
the
be
ha
vi
or
o
f
nat
u
ral
syst
e
m
s
(
o
r)
natura
lly
o
c
c
u
rrin
g
p
he
nom
en
a
h
a
ve
gi
ve
n
rise
t
o
m
u
l
t
i
p
le
o
p
tim
iza
t
i
o
n
ap
proa
c
h
es
l
i
k
e
P
a
r
t
i
c
le
S
w
arm
O
p
tim
i
z
at
i
on
(P
SO
)
[1]
A
n
t
C
o
lo
ny
Opt
i
m
i
za
t
i
o
n
(
ACO)
[
2]
G
ene
t
i
c
A
l
g
ori
t
hm
(
G
A
)
[3]
Bacter
ia
l
F
o
r
agi
n
g
O
p
tim
izat
i
on
A
l
g
o
ri
thm
(
B
F
O
A
)
D
i
ffe
re
nt
ia
l
e
v
ol
u
t
i
on
(D
E)
[
4]
I
mm
une
A
l
g
ori
t
hm
(IA)
[5]
etc.
T
h
e
s
e
al
gor
it
h
m
s
ha
ve
a
da
p
t
e
d
f
r
o
m
nat
u
ra
ll
y
occ
u
rr
in
g
pr
oc
ess.
T
he
y
ca
n
be
r
efe
rre
d
usi
n
g
di
ffere
nt
n
am
es
w
it
h
t
h
e
na
m
e
s
like
E
v
olut
i
onar
y
A
l
g
o
ri
t
h
m
s
a
n
d
m
eta
h
e
u
r
i
s
t
i
c
a
p
p
roac
hes
be
in
g
c
o
m
m
onl
y
use
d
.
T
h
e
m
e
tahe
urist
i
c
ap
proa
c
h
es
t
yp
ica
l
l
y
c
o
m
b
i
n
e
h
e
u
r
i
s
t
i
c
a
l
g
o
r
i
t
h
m
s
w
h
i
c
h
a
r
e
u
s
u
a
l
l
y
p
r
o
b
l
e
m
s
p
e
c
i
f
i
c
i
n
a
m
o
re
g
e
n
era
l
iz
ed
fram
e
w
or
k.
S
o,
me
taheur
i
s
tics
ca
n
be
c
o
n
side
red
a
s
p
roce
sse
s
w
hic
h
s
trateg
ie
s
to
f
i
n
d
a
n
o
p
t
i
m
um
(
or)
a
ne
ar
o
p
tim
um
so
l
u
t
i
o
n
.
Thes
e
me
t
a
heur
is
tic
a
ppr
oa
ches
a
r
e
a
ppr
oxim
a
te
a
n
d
n
o
n-
de
t
e
rm
ini
s
ti
c
an
d
the
y
u
sua
l
l
y
e
m
p
lo
y
me
cha
n
ism
s
to
have
a
goo
d c
o
n
v
erge
nc
e
an
d pr
ovi
de
n
ea
r op
t
i
mum
so
lu
ti
o
n
s
.
S
i
milarl
y,
A
N
F
IS
i
s
a
ver
y
e
ffec
ti
ve
m
o
d
e
l
in
g
a
p
proa
c
h
w
hic
h
c
o
mb
ines
t
h
e
a
t
t
r
i
but
e
s
o
f
both
th
e
fuz
z
y
i
nfe
r
enc
e
s
yste
m
an
d
neura
l
n
e
t
w
o
r
k
.
The
am
alga
m
a
tio
n
o
f
fuzz
y
lo
gic
w
i
t
h
a
rc
hitec
t
ur
al
d
e
s
ign
o
f
neura
l
n
e
t
w
o
rk
l
ed
t
o
cre
a
t
i
o
n
o
f
ne
uro-
f
uzz
y
s
y
s
tem
s
.
A
m
u
lti
tu
de
o
f
m
e
th
o
d
s
have
b
e
e
n
u
s
ed
t
o
o
p
t
i
m
i
z
e
the
fuzz
y
me
mbe
r
ship
f
unc
tio
ns
i
n
t
h
e
l
i
t
e
ra
t
u
r
e
.
The
s
e
me
thods
c
an
b
e
d
i
vi
d
e
d
in
to
t
wo
t
yp
e
s
i
nc
l
udi
ng
deri
va
ti
ve
b
a
s
e
d
a
nd
h
e
uris
tic
a
l
gori
t
hm
s
i
n
g
ene
r
a
l
[
6]
.
S
hoor
e
hde
l
i
e
t
a
l
[
7],
[8]
pr
o
p
o
s
e
d
h
y
b
ri
d
me
tho
d
s
c
o
mp
ose
d
p
arti
cl
e
swarm
o
p
t
i
m
i
z
a
t
ion
(PSO)
.
H
e
u
s
e
d
r
ecu
rsi
v
e
l
ea
st
s
quare
(
R
L
S
)
a
nd
ex
te
nde
d
K
a
lm
an
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
:
26
5
–
276
26
6
f
i
l
t
er
(
EK
L)
f
or
t
r
a
i
n
i
n
g.
I
n
dif
f
er
en
t
s
t
u
d
i
es,
t
h
e
y
p
r
o
p
o
se
d
fa
c
t
or
r
ecur
s
i
v
e
l
e
as
t
s
quar
e
f
or
t
r
a
in
in
g
t
h
e
c
onc
l
u
sio
n
p
a
r
am
ete
r
s
and
L
y
ap
u
n
o
v
s
ta
bi
lit
y
the
o
r
y
t
o
im
pr
ove
the
pe
r
f
or
m
a
nc
e
o
f
A
N
F
IS
.
I
n
a
dd
it
ion
t
o
th
ese,
t
h
e
y
u
s
e
d
N
SGA
-II
th
e
t
r
ainin
g
of
a
ll
p
a
rameters
o
f
ANFI
S
str
u
ctur
e
.
S
im
i
l
a
r
l
y
Z
a
n
ge
neh
e
t
a
l
[
9
]
pr
opose
d
a
n
e
w
t
y
p
e
of
t
r
a
in
in
g
A
N
F
I
S
[
10]
,
[1
1]
i
s
a
p
p
l
y
i
ng
c
o
m
p
l
e
x
t
ype
(
D
E
/c
ur
r
e
nt
-
t
o
b
e
s
t/
1
+
1
/
b
i
n
&
D
E
/r
a
n
d
/
1/
b
i
n)
on
pr
ed
ic
ti
ng
of
M
ac
ke
yg
las
s
time
ser
i
e
s
.
I
n
th
is
p
a
p
er
w
e
have
p
r
o
po
sed
G
A
-
T
r
ained
A
N
F
I
S
f
r
a
me
w
o
r
k
f
or
tun
i
ng
t
h
e
P
I
D
contr
o
l
l
e
r
.
This
c
o
n
t
r
o
ller
is us
e
d i
n
un
i
so
n with a
M
ode
l Re
fe
renc
e A
d
a
p
tive
Co
nt
r
o
l
(
M
R
A
C
)
t
o
achi
e
v
e
c
ont
rol
of
a
B
LD
C
mo
to
r.
T
h
e
p
erfo
rman
ce
o
f
the
p
r
opo
sed
con
t
r
o
ller
is
e
v
a
lu
ated
a
g
a
i
nst
a
st
a
n
dar
d
s
e
l
f
tun
i
ng
P
I
D
[
12
]
-
[13]
c
o
n
t
r
o
ller
for
t
r
a
c
ki
ng
the
sp
eed
c
ha
n
g
e
s
d
ue
t
o
sud
d
e
n
c
han
g
e
i
n
l
oad
,
sud
d
e
n
c
ha
n
g
e
i
n
r
e
fer
e
nce
spee
d.
,
sud
d
e
n
c
ha
n
g
e
in
i
ner
tia
a
n
d
pha
se
r
esis
tanc
e
.
T
h
e
d
e
t
a
ile
d
mod
e
l
o
f
BL
D
C
[
1
4
]
is
e
x
p
la
i
n
ed
i
n
se
ct
ion
2.
S
ectio
n 3
ex
p
l
ain
s
a
bo
ut G
A
-
A
N
F
IS
s
ystem
,
her
e
t
h
e
per
f
o
r
m
a
nce
of
a P
I
D
c
on
t
r
o
ller
is
c
ompa
r
e
d
f
o
r
tun
i
ng
th
r
o
ugh
ANFIS
a
n
d
GA
ANF
I
S.
Th
e
GA
ANFI
S
M
ARC
setu
p
is
e
x
p
lain
e
d
in
s
ec
ti
on
4
,
wh
ile
s
e
c
tio
n
5
b
r
ie
f
s
a
b
o
u
t
th
e
r
e
s
u
lt
s
.
2.
MO
DELI
N
G
O
F
BLDC
M
O
TOR
The
m
a
t
h
em
at
ical
m
o
d
e
l
o
f
BLD
C
h
as
l
ot
o
f
s
i
m
i
lari
ti
e
s
t
o
c
o
nve
nti
ona
l
DC
m
ot
o
r
s.
O
n
e
m
a
j
o
r
a
d
d
iti
o
n
i
s
in
r
egar
d
to
t
he
p
hases
w
h
ic
h
i
m
pac
t
t
he
o
ver
a
l
l
f
u
nc
tio
n
i
n
g
and
e
f
ficie
n
c
y
o
f
the
BLD
C
.
The
s
e
pha
se
s
ty
p
i
cal
l
y
e
xer
t
t
he
ir
i
n
f
lue
n
ce
on
r
e
s
i
stive
an
d
in
d
u
c
t
i
ve
a
rra
n
g
em
ent
o
f
B
LD
C.
T
he
s
c
h
em
at
ic
B
LD
C
is
i
ll
ustr
a
t
e
d
i
n
F
i
gur
e
1.
Fi
g
u
r
e 1
.
Sch
e
m
a
t
i
c
o
f BL
DC
T
h
e
ma
t
h
e
m
a
tic
al
m
o
d
e
l
o
f
D
C
m
o
to
r
i
s
G
.
.
.
(
1
)
T
h
e me
ch
ani
c
al
t
i
m
e
c
o
n
s
t
a
nt
i
s
(
2
)
The
ele
c
t
rica
l
tim
e
cons
tan
t
(
3
)
I
n
t
he
case
of
t
he
B
LD
C
m
o
tor
s
t
he
c
o
n
sta
n
t
s
t
a
k
e
the
s
e
f
o
r
m
s
(
4)
a
nd
(
5
)
.
.
∑
.
(
4
)
∑
∑
(
5
)
S
i
nce
t
h
er
e
i
s
a
s
ymm
e
tr
ic
al
a
r
r
angem
e
nt
a
n
d
a
t
hr
ee
p
ha
se,
the
m
echa
n
i
c
al
(
k
n
o
wn)
and
e
l
ec
trica
l
c
o
n
s
ta
nts
be
co
m
e
(
6)
a
nd
(
7
)
:
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
G
A
-ANFIS PID
com
p
e
n
s
a
t
e
d m
ode
l
re
fere
nce
ad
a
p
t
i
ve
c
o
n
t
r
o
l
f
o
r
BL
D
C
m
o
t
o
r (
M
urali D
a
sar
i
)
26
7
.
.
(
6
)
.
(
7
)
Cons
i
d
eri
ng
th
e
phase
effec
t
s
,
.
∅
.
√
.
(
8
)
.
∅
.
.
(
9
)
wher
e
i
s
t
h
e
pha
se
va
l
ue
o
f
t
h
e
E
M
F
(vol
tage)
co
nsta
n
t
;
√
(
1
0
)
A
l
so,
t
h
er
e
is
a
r
e
l
atio
ns
hi
p
betw
e
e
n
a
n
d
;
u
s
in
g
th
e
ele
c
tric
a
l
p
o
w
er
(
left
h
a
n
d
si
de)
and
m
e
c
h
an
ica
l
pow
er
(righ
t h
a
nd s
i
de)
equa
t
i
o
n
s;
√
(
1
1
)
√
(
1
2
)
√
(
1
3
)
.
(
1
4
)
wher
e
,
t
he
e
l
e
c
t
r
i
c
a
l
t
or
qu
e
and
i
s
the
m
e
c
h
a
n
i
c
a
l
t
or
que.
Co
nsi
d
er
in
g
t
h
e
effec
t
s
of
t
he
co
n
sta
n
t
s
a
nd t
h
e pha
se
,
the e
quat
i
o
n
f
or
B
LD
C
can
b
e ob
ta
i
n
e
d
as,
.
.
.
(
1
5
)
3.
GA ANFI
S
PI
D
C
O
N
TROL
LER
F
u
zz
y
lo
gic
a
s
a
n
i
d
ea
prop
o
s
ed
b
y
Za
de
h
w
a
s
fi
r
s
t
im
p
l
e
m
e
n
t
e
d
b
y
M
a
d
a
n
i
i
n
t
h
e
y
e
a
r
1
9
7
5
[
1
4
]
.
M
a
d
a
ni
d
e
m
onst
r
a
t
ed
t
he
i
d
e
a
of
i
mp
l
e
me
nt
i
n
g
t
h
e
fu
z
z
y
l
ogi
c
as
a
c
onc
ept
for
use
in
m
odel
s
t
ea
m
e
n
g
i
ne
.
S
ubseq
ue
n
t
l
y
m
a
ny
a
p
p
lic
at
ions
e
v
o
l
ve
d
u
s
ing
t
h
e
c
onc
e
p
t
o
f
f
uz
zy
l
og
i
c
.
Dif
f
e
re
nt
a
p
p
li
c
a
t
i
on
s
o
f
f
u
zzy
lo
gic
for
i
n
du
stria
l
a
n
d
h
om
e
a
p
p
l
ica
t
i
o
ns
c
a
n
b
e
f
o
u
n
d
i
n
the
l
i
ter
a
t
u
re.
Tw
o
im
p
o
r
t
a
n
t
fac
t
or
s
na
m
e
ly,
se
l
e
c
t
ion
of
k
no
wl
ed
g
e
t
echni
qu
es
a
nd
a
v
a
il
a
b
il
i
t
y
o
f
k
nowl
e
dg
e
b
a
s
e
in
fl
ue
nc
e
the
de
si
gn
of
F
uz
z
y
L
o
g
ic
Con
t
ro
llers.
T
h
ese
tw
o
fac
t
o
r
s
pr
i
m
ar
i
l
y
i
n
fl
ue
nce
the
a
p
plic
a
t
i
o
ns
o
f
F
u
z
z
y
lo
g
i
c
.
T
h
i
s
ca
n
be
o
ve
rcom
e
w
i
t
h
t
h
e
u
s
e
o
f
A
d
a
p
t
i
v
e
N
e
u
r
o
-
Fuzzy
Inference
System
(
ANF
IS).
A
n
adap
t
i
ve
N
eu
r
o
-
F
u
z
z
y
Infe
renc
e
S
y
stem
(
A
N
F
I
S
)
i
s
a
c
o
mbinat
i
on
of
a
n
A
r
t
i
f
i
c
i
a
l
N
eur
a
l
N
e
tw
ork
(
A
N
N
)
a
nd
a
fu
z
z
y
i
n
fer
e
nce
system
(F
IS).A
N
N
e
m
ula
t
e
s
t
he
f
unc
t
i
o
n
i
n
g
of
h
um
an
b
ra
in
a
n
d
i
s
form
ul
a
t
ed
a
s
col
l
ec
t
i
o
n
o
f
a
r
ti
ficia
l
n
e
u
ro
ns
.
An
a
d
a
p
t
i
v
e
n
e
t
w
o
r
k
h
a
s
m
u
l
t
i
p
l
e
l
a
y
e
r
s
o
f
f
e
e
d
f
o
r
w
a
r
d
n
e
t
w
o
r
k
.
I
n
t
h
is
t
op
o
l
o
g
y
e
ach
node
o
f
the
mul
t
il
a
y
e
r
netw
ork exec
u
t
e
s
s
pec
i
fic fu
n
c
tio
ns o
n
i
n
co
m
i
ng
s
ig
n
a
ls.
Ea
ch n
ode ha
s
i
t
s
o
w
n
spe
c
i
fi
c func
t
i
o
n
. In t
he
c
ase
of
a
da
p
tive
ne
t
w
ork
tw
o
t
y
pe
s
o
f
n
o
d
es
n
a
m
ely
ada
p
ti
ve
a
nd
fi
xe
d
n
ode
s
ar
e
pre
s
e
n
t.
I
n
t
h
e
c
a
se
o
f
S
uge
n
o
F
I
S
t
h
e
o
u
t
p
ut
m
e
m
be
rship fu
nct
i
o
n
s
ar
e
s
ingle
t
on s
p
i
k
es.
I
n
t
he
c
ase
of
G
e
n
e
tic
A
l
g
o
r
ithm
,
t
he
i
n
itia
l
c
h
romos
o
m
e
s
w
h
ic
h
a
re
r
a
n
do
mly
po
pul
a
t
ed
a
re
referr
ed
t
o
as
p
a
r
en
t
c
h
r
o
mo
som
e
s
an
d
su
b
s
eq
uen
t
g
e
n
e
r
ati
o
ns
o
f
c
hr
omosom
es
a
r
e
r
e
f
er
red
to
a
s
ch
i
l
d
(
o
r)
offspr
in
g.
T
he
p
ri
nci
p
le
b
e
h
i
n
d
ge
net
i
c
a
l
g
o
r
i
t
h
m
is
t
o
i
n
vo
l
v
e
b
e
t
t
e
r
p
a
ren
t
s
i
n
t
h
e
p
ro
ce
ss
o
f
r
e
p
roduc
t
i
on
,
so
a
s
t
o
i
mpro
ve
t
he
c
ha
nce
s
o
f
pr
od
uc
i
n
g
bet
t
er
o
ffspr
ing.
T
hr
o
u
g
h
t
h
i
s
p
r
o
ces
s
of
n
atura
l
s
e
l
ec
t
i
o
n
,
t
h
e
st
r
o
nger
chrom
o
some
s
ar
e c
a
rr
i
e
d forward
to
the
n
ex
t sta
g
e
whi
l
e
the
w
e
a
k
e
r
c
hrom
osome
s
a
re e
l
i
m
i
nat
e
d.
A
N
FIS
deli
ver
s
a
n
e
f
fi
cie
n
t
pe
rform
a
nc
e
i
n
s
y
s
tem
i
d
e
n
t
i
f
i
c
a
t
i
o
n
an
d
d
e
li
ver
s
g
ood
p
r
edic
ti
o
n
a
n
d
con
t
ro
l
per
f
orm
a
nce
t
oo.
T
h
e
p
er
form
ance
o
f
the
A
N
F
I
S
s
ys
tem
c
a
n
be
i
m
p
ro
ve
d
t
h
r
o
u
g
h
t
r
ai
n
i
n
g
a
nd
G
A
enha
nc
e
d
t
ra
i
n
in
g
i
s
o
n
e
o
f
t
h
e
m
o
st
p
referr
ed
a
nd
s
u
ita
bl
e
me
t
hod.
G
A
enha
nc
e
d
t
ra
in
in
g
i
s
m
ore
suita
b
l
e
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
:
26
5 –
27
6
26
8
b
e
ca
use
in
t
h
e
c
a
s
e
of
A
NFI
S
t
h
e
t
rai
n
in
g
has
to
i
mp
art
e
d
t
o
b
o
th
t
h
e
a
nt
ec
eden
t
part
a
n
d
c
onc
lusi
o
n
p
art
o
f
the pa
ram
e
te
rs.
The G
a
ussia
n
m
em
bership fu
nc
ti
on
is
a
s depic
t
e
d
in (1
6)
(
1
6
)
Where
{a
i
,
b
i
,c
i
}
ar
e
the
pa
ra
m
e
ter
s
o
f
MFS
whic
h
ar
e
affe
cted i
n
sha
p
e
o
f
M
Fs.
a
i
i
s
th
e
v
a
ri
an
c
e
of
the
m
e
m
b
ersh
i
p
f
u
n
ct
i
o
n
,
c
i
t
he
c
e
n
t
e
r
o
f
m
em
bershi
p
fu
n
c
ti
o
n
a
nd
b
i
i
s
u
s
u
a
l
l
y
e
q
u
a
l
t
o
1
.
I
n
t
h
e
a
n
t
e
c
e
d
e
n
t
part
t
here
a
re
3
s
et
o
f
trainabl
e
parameters
w
hic
h
h
as
N
g
enes.
N
i
s
th
e
numbe
r
o
f
M
e
m
be
rsh
i
p
F
unc
t
i
o
n
s.
The
o
p
t
imiz
at
ion
a
l
go
r
i
t
h
m
als
o
t
rai
n
s
the
c
onc
l
u
s
i
o
n
p
a
r
t
w
h
ic
h
ha
s
(I
+
1
)
×
R
g
enes,
where
R
de
n
o
t
e
s
th
e
numbe
r
o
f
r
u
l
e
s
a
n
d
I
d
e
note
s
t
he
n
um
ber
o
f
d
im
ens
i
o
n
s
o
f
d
a
t
a
i
n
p
u
t
s
.
T
h
e
f
i
t
n
e
s
s
i
s
d
e
f
i
n
e
d
a
s
R
o
o
t
M
e
a
n
Squa
re
E
rr
or
(
RMS
E
).
P
arame
t
ers
are
i
n
i
t
i
a
lize
d
r
a
n
d
o
m
l
y
in
f
ir
st
s
te
p
a
nd
the
n
a
re
b
e
i
ng
up
d
a
te
d
usi
ng
G
A
alg
o
ri
t
h
m.
I
n
ea
ch
itera
t
i
o
n
,
one
o
f
t
h
e
pa
ram
e
ter
s
s
e
t
a
re
b
ei
n
g
upda
ted.
i
.
e
.
in
f
irst
i
te
ra
t
i
on
f
or
e
xa
m
p
le
a
i
are
upda
te
d
th
en
i
n
sec
o
nd
ite
rat
i
o
n
b
i
are
up
da
ted
a
n
d
t
h
e
n
a
fter
u
pda
tin
g
a
ll
pa
ram
e
ter
s
a
gai
n
t
he
f
i
rs
t
para
me
ter
upd
ate is
c
o
n
side
r
e
d.
I
n
o
rde
r
t
o tes
t
t
h
e
p
erf
o
rma
n
ce
o
f t
h
e prop
o
s
e
d
a
ppr
oac
h
a
t
es
t
S
i
mu
li
n
k
s
ys
t
e
m
is
c
on
st
ruc
t
e
d
.
The
s
y
s
t
e
m
h
a
s
a
P
I
D
c
o
n
t
r
o
l
l
e
r
w
h
i
c
h
i
s
t
u
n
e
d
b
y
t
h
e
A
N
F
I
S
a
n
d
G
A
-
A
N
FI
S
a
pproache
s
.
The
S
i
mul
i
n
k
mode
l
of
t
he
pro
po
se
d
G
A
-
F
u
zzy
c
on
tro
l
l
e
r is
g
iv
e
n
i
n F
i
g
u
re
2.
F
i
gure
2.
S
im
ulin
k
mode
l
of t
he
G
A
-
A
N
F
I
S
c
o
n
t
r
o
ller
setu
p
The
d
i
s
t
rib
u
tio
n
of
m
ea
n
err
o
r
for
the
bo
t
h
t
he
m
et
h
ods
a
re
i
l
l
ustra
t
e
d
w
i
t
h
t
h
e
he
lp
o
f
F
i
gure
3
(
a
)
&
(b)
.
D
urin
g
trai
ni
n
g
A
N
F
IS
p
rod
u
c
e
d
a
m
ea
n
e
rror
of
-
3.82
01
w
it
h
a
st
a
n
d
a
rd
d
e
v
ia
t
i
o
n
o
f
2
5
.
34
wh
il
e
G
A
-
A
N
F
I
S
produc
e
d
a
m
ea
n error
of 4.
643
4 and
a
sta
n
d
a
r
d
de
v
i
a
ti
o
n
o
f
23.
0
026.
(a)
(b)
Figure
(3)
(a
)
Distribution of
m
ean
e
rr
or
– ANFI
S
, (b) Di
s
tribu
t
i
o
n
of mean
er
ro
r-
GA
-
ANFI
S
-1
5
0
-
100
-5
0
0
50
10
0
15
0
0
10
20
30
40
50
60
70
-
100
-5
0
0
50
100
150
0
5
10
15
20
25
30
35
,
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
G
A
-ANFIS PID
com
p
e
n
s
a
t
e
d m
ode
l
re
fere
nce
ad
a
p
t
i
ve
c
o
n
t
r
o
l
f
o
r
BL
D
C
m
o
t
o
r (
M
urali D
a
sar
i
)
26
9
I
t
ca
n
be c
l
ea
rl
y obser
ve
d fr
o
m
the F
i
g
u
r
e
4, t
ha
t
the
se
t
t
l
i
n
g
t
im
e an
d t
h
e
pea
k
o
ver
s
h
o
o
t
of
t
he G
A
-
A
N
F
I
S
t
u
n
e
d
P
I
D
i
s
m
u
c
h
l
e
s
s
e
r
w
h
e
n
c
o
m
p
a
r
e
d
t
o
t
h
e
s
e
t
t
l
i
n
g
t
i
m
e
a
nd
t
he
o
ver
s
h
o
o
t
e
xperie
n
c
e
d
b
y
t
h
e
A
N
F
IS
t
u
n
e
d
P
ID
c
ontro
l
l
e
r
for the
p
r
o
p
o
se
d BLD
C
m
otor
.
F
i
gure
4.
C
o
n
tr
ol
ler
respo
n
se
f
or
A
N
F
IS
P
ID
contro
l
l
er
a
nd
G
A
-
A
NF
IS
P
ID
c
ont
r
o
l
l
e
r
of B
L
D
C
m
ot
o
r
4.
MARC
G
A
-
A
N
FIS
PID
CO
NTROLLER
Mo
de
l
R
e
fer
e
nce
A
d
a
p
ti
ve
C
on
tro
l
(
MRA
C
)
is
a
d
ire
c
t
a
d
a
p
t
i
ve
s
t
rat
e
gy
w
i
t
h
som
e
a
djus
ta
bl
e
con
t
ro
l
l
er
p
ara
m
e
t
e
r
s
a
n
d
a
n
a
d
j
us
tin
g
me
chan
ism
t
o
a
d
j
u
s
t
t
h
em
.
A
s
c
ompa
red
to
t
he
w
e
l
l
-
know
n
a
n
d
si
m
p
le
s
truc
tur
e
d
f
i
xe
d
ga
i
n
P
ID
c
ontro
l
l
e
r
s,
a
dapt
i
v
e
co
ntr
o
ll
e
r
s
ar
e
ve
ry
e
ffe
c
t
i
v
e
t
o
h
a
n
d
l
e
t
h
e
u
n
k
n
o
w
n
para
me
ter
var
i
a
tio
ns
a
nd
e
n
v
i
ro
nm
enta
l
c
h
a
nge
s.
A
n
ada
p
tive
c
o
n
t
r
o
l
l
e
r
co
n
s
i
s
t
s
o
f
t
w
o
l
oop
s,
a
n
out
e
r
l
oop
o
r
n
o
r
mal
f
e
e
d
b
a
ck
l
oop
a
n
d
a
n
i
n
n
e
r
l
oop
o
r
p
ara
m
e
t
e
r
a
d
j
u
s
t
m
e
n
t
l
oo
p
as
i
nd
ic
ate
in
F
ig
ure
5.
F
igure
6
sh
o
w
s
bl
ock
di
ag
ra
m
o
f
t
he
p
ro
po
s
e
d
sy
st
e
m
.
M
o
d
e
l
R
e
fe
re
n
c
e
Ad
a
ptiv
e
Cont
rol
st
rat
e
gy
i
s
u
s
e
d
t
o
d
e
si
gn
t
h
e
ad
a
p
ti
ve
c
on
t
r
o
l
l
e
r
t
h
at
w
o
r
k
s
o
n
th
e
p
r
in
c
i
pl
e
of
a
dj
u
s
t
i
n
g
t
h
e
c
ontr
o
ller
pa
ram
e
ter
s
s
o
t
h
a
t
t
he
o
u
t
put
o
f
the
a
c
t
ua
l
p
l
a
n
t
t
r
ac
ks
t
he
o
u
t
p
u
t
o
f
a
refe
renc
e
mode
l
ha
vi
n
g
t
h
e
sam
e
r
efe
r
enc
e
i
np
u
t
.
The
M
I
T
ru
l
e
h
as
bee
n
em
p
l
o
ye
d
he
re
a
nd r
u
l
e
a
cost fu
nc
tio
n
i
s
d
efi
n
ed
a
s
;
(
1
7
)
To
m
ake
J
sm
a
l
l,
i
t
is
r
ea
so
na
ble
t
o
c
ha
n
g
e
the
para
me
t
e
rs
i
n
t
he
d
irec
t
i
on
o
f
t
h
e
ne
ga
tive
grad
ie
nt
o
f
J
,
t
h
a
t
i
s
,
=-
=-
(
1
8
)
wher
e;
: A
d
apta
t
i
on
ga
i
n
.
,
:
The
c
ontrol
l
er
par
a
m
eter.
e: The
e
rror
betw
ee
n t
h
e
ou
tp
ut s
pee
d
o
f
the
BLD
C
m
otor a
nd t
h
e
m
ode
l
refe
renc
e
ou
tp
u
t
.
I
t
a
ssume
d
tha
t
t
he
p
roce
ss
i
s
de
sc
ribe
d
b
y
t
he
s
i
n
g
l
e-i
n
pu
t
,
s
ingl
e-output
(
SI
S
O
)
system
a
s
shown
in Fig
ure
7.
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
:
26
5 –
27
6
27
0
F
i
gure
5.
T
he
p
ro
pose
d
MA
R
C
con
t
ro
l
l
er
s
e
t
u
p
F
igure
6.
B
l
o
c
k
d
i
a
g
ram
of the
p
ropos
e
d s
y
s
t
em
F
i
gure
7. A
ge
n
e
r
al
linea
r c
o
ntr
o
l
l
e
r wi
th
two
d
eg
r
e
e
s
o
f
fr
e
e
d
om
.
(
1
9
)
wher
e:
A
&
B
a
r
e
poly
nom
i
a
ls
d
epe
n
d on t
h
e
BLD
C
m
otor.
,(t): The
ou
tpu
t
o
f c
o
n
trol
ler
.
(
):
Th
e
o
ut
pu
t
sp
e
e
d
of
B
LDC mo
t
o
r.
(
): The
proc
e
ss di
stur
ba
nce
.
The
co
ntro
l
l
e
r
is de
scri
be
d
i
n
(
20)
-
(
2
0
)
wher
e:
R
,
T
and
S
a
r
e
contro
l
l
e
r
pol
yn
om
i
a
l
s
.
,
: The
de
sire
d
spee
d of
B
LD
C m
o
t
o
r.
S
ubst
i
t
ut
i
ng (1
1)
i
n
t
o
(10)
w
i
l
l r
e
sul
t
(
12)
+
(
2
1
)
A
ssume
the
m
ode
l
refere
nce
is
d
escr
ibe
d
by
the
si
ngle
-
i
n
put,
si
ng
l
e
-o
ut
put
(
S
I
S
O
)
syst
e
m
(
22)
wher
e:
a
r
e
po
ly
n
o
mia
l
s
dep
e
nd
o
n
t
he
r
e
f
er
ence
m
od
e
l
.
,
:
The
out
pu
t
of
m
ode
l
reference. Ass
um
ing
,
(
V
(
)=0) the
fo
l
low
i
ng
con
d
i
t
i
on
mu
st h
ol
d
:
(
2
3
)
(
2
4
)
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
G
A
-ANFIS PID
com
p
e
n
s
a
t
e
d m
ode
l
re
fere
nce
ad
a
p
t
i
ve
c
o
n
t
r
o
l
f
o
r
BL
D
C
m
o
t
o
r (
M
urali D
a
sar
i
)
27
1
wher
e:
=
,
,
,
,
: The mod
el r
efe
r
e
n
ce t
r
ansfe
r
fun
ct
i
o
n
coeffic
i
e
n
t.
A
ssume
the
t
r
a
nsfe
r func
t
i
o
n
o
f t
h
e
BLD
C
m
otor
i
s
(
25)
W
h
ere;
,
,
BLD
C
m
otor tra
ns
fer
fu
nc
ti
o
n
c
oeffic
i
e
n
t. The
D
io
pha
n
t
i
n
e e
q
ua
t
io
n
is
+
=
(
2
6
)
W
h
ere:
,
A
nd
i
s
a
ga
i
n
.
,R a
n
d
S
:
contr
o
l
l
e
r
pol
yn
om
ia
ls.
Whe
r
e
de
g is t
he
p
o
l
yn
om
ia
l de
gree
.
(
2
7
)
+
P
(
2
8
)
+
-
=2 + 1
–
2=1
(
2
9
)
S
i
m
i
l
a
r
l
y
T
=
P
(
3
0
)
Subst
i
t
ut
i
ng (2
7)
t
o
(21)
w
ill
resul
t
(31)
P
u=
P
-
(
3
1
)
-
(
3
2
)
F
r
om
(
10)
a
nd a
s
sum
e
=0
(
)=
(
33)
S
ubsti
t
u
ti
ng
(3
2) in
t
o (33)
w
il
l
re
sul
t
(
3
4
)
y
b
y
(
34)
=>
b
Mo
dif
y
in
g
(34
)
to
b
e
com
e
(
35)
(
3
5
)
(
3
6
)
S
ubst
i
t
ut
i
ng (2
4),
(
3
5
)
in
t
o (3
6)
w
il
l
resu
lt (37)
(
37)
(
3
8
)
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
:
26
5 –
27
6
27
2
(39)
F
r
om
(
15)
(
4
0
)
(
4
1
)
wher
e
Sim
i
larly
y
(
4
2
)
(
4
2
)
wher
e:
1
.
5
0
0
.
5.
RESULT
S
A
N
D
DISCU
SSIO
N
I
n
o
rde
r
t
o
va
l
i
d
a
te
t
he
p
er
fo
rm
ance
o
f
t
h
e
pro
pose
d
c
on
trolle
r
s
et
up
i
s
su
bj
ect
e
d
t
o
diff
ere
n
t
t
e
st
ca
se
s,
l
ike
su
d
d
e
n
c
ha
n
g
e
i
n
l
oa
d a
nd
su
d
d
e
n
c
han
g
e
in
s
p
eed.
T
he
r
e
s
ult
s
of
w
h
ic
h a
r
e pre
s
e
n
t
e
d he
re.
The
F
i
g
u
re
8
i
ll
ustra
t
es
t
he
ope
n
l
o
op
r
e
s
p
o
n
se
o
f
the
des
i
gne
d
B
LD
C
m
ode
l
.
I
n
order
to
v
a
l
i
d
at
e
the
m
ode
l
a
sudde
n
loa
d
c
ha
nge
i
s
ap
pl
ie
d
at
0
.1
s
ec
o
n
d
s.
I
niti
a
l
l
y
t
he
m
o
t
or
i
s
run
a
t
n
o
loa
d
a
nd
the
spee
d
curve
poi
n
t
s
t
o
t
he
n
o
loa
d
o
p
e
rati
on.
T
he
n
a
t
t
=0.1
s
e
c
o
nd
a
l
oa
d
e
q
u
i
va
l
e
nt
t
o
t
h
e
5
0
%
o
f
t
h
e
r
a
ted
lo
a
d
i
s
app
l
ied.
I
t
c
a
n
b
e
o
b
se
rve
d
t
h
a
t
t
h
ere
is
a
s
u
dde
n
dro
p
i
n
s
p
e
e
d
a
t
t
hat
in
sta
n
t
a
n
d
t
h
e
m
o
tor
se
tt
l
e
s
a
t
l
e
sser
spee
d
from
t=
0
.14 se
co
nds.
F
i
gure
8.
P
lot of
o
pe
n lo
o
p
spe
e
d
re
s
p
onse
at
s
udde
n cha
n
ge in
loa
d
The
F
i
gur
e
9
i
l
l
u
s
t
rate
s
t
h
e
s
p
e
e
d
r
e
sp
onse
of
t
he
c
on
tro
l
le
r
f
or
d
i
f
fer
e
n
t
a
d
a
pta
t
i
o
n
rat
e
g
a
i
n,
t
he
spee
d
re
gula
t
i
o
n
char
ac
ter
i
s
tic
s
are
stu
d
i
e
d
w
i
t
h
t
he
s
u
dde
n ch
an
ge
i
n
l
o
ad.
The
l
o
a
d
t
orq
u
e
is var
i
e
d
to
50 %
of
t
he
r
a
t
e
d
v
a
l
ue
a
t
t=
0.2
5
s
ec
on
d.
T
h
i
s
is
a
fter
i
n
i
t
i
al
s
e
t
t
li
ng
t
i
m
e
f
o
r
th
e
c
o
nt
rol
l
e
r.
T
h
i
s
g
r
a
ph
dem
o
n
s
t
r
ate
s
t
he
s
u
ita
b
ili
t
y
o
f
hi
g
h
er
a
da
p
t
a
tio
n
ra
t
e
o
n
t
h
e
p
erf
o
rma
n
ce
o
f
the
c
o
n
t
ro
ller
.
I
t
ca
n
be
obse
r
ved
from
the
fig
ure
tha
t
hi
g
her
ad
a
p
t
a
t
i
o
n
g
a
i
n r
e
duce
s
the
rise
ti
me
.
It
c
an
a
lso
be
i
n
f
erred
th
a
t
it
al
so
red
uc
es
t
h
e
overs
ho
o
t
a
nd
stea
dy
sta
t
e
er
ror.
0
0.
02
0.
04
0.
06
0.
08
0.
1
0.
12
0.
14
0.
1
6
0.
1
8
0.
2
0
500
1
000
1
500
2
000
2
500
3
000
3
500
4
000
4
500
Ti
m
e
(
S
)
R
o
t
o
r
S
peed
(
R
P
M
)
O
p
en
l
oop
r
es
pon
s
e
f
or
s
udd
en
c
h
ang
e
i
n
l
o
a
d
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
G
A
-ANFIS PID
com
p
e
n
s
a
t
e
d m
ode
l
re
fere
nce
ad
a
p
t
i
ve
c
o
n
t
r
o
l
f
o
r
BL
D
C
m
o
t
o
r (
M
urali D
a
sar
i
)
27
3
F
i
gure
9. S
pee
d
r
e
s
po
nse
of
t
he
pro
pose
d
c
o
n
tr
ol
l
e
r for di
ff
ere
nt
a
da
pt
a
tio
n r
a
te
g
ai
n
It
c
an
b
e
i
n
fe
rred
f
ro
m
t
h
e
fig
u
r
e
t
h
at
p
ro
po
sed
c
ont
ro
ll
er
h
as
b
e
t
ter
per
f
orma
nce
at
s
u
dde
n
s
p
e
e
d
cha
nge
w
he
n
c
o
mpa
r
ed
t
o
t
h
e
con
v
en
t
i
ona
l
con
t
ro
l
l
e
r
.
T
h
e
rise
t
im
e
is
l
e
ss
in
t
he
c
a
s
e
of
s
ud
de
n
c
h
a
nge
i
n
loa
d
a
n
d
h
enc
e
the
c
ontr
o
ller
can
a
ct
s
uita
b
l
y
fa
s
t
t
o
re
gu
la
t
e
t
he
s
pee
d
.
Th
is
i
s
of
p
a
r
am
oun
t
i
m
porta
nc
e
as
the pr
o
p
o
s
e
d
c
on
tro
l
ler has th
e
abi
l
ity to re
cover
t
h
e los
t
s
p
e
ed
a
t
a
mu
ch
fast
e
r
ra
t
e
.
Th
e
c
o
nt
roll
e
r
o
ut
pu
t
f
o
r
the pr
opose
d
con
tro
l
ler
and t
h
e
con
v
e
n
t
i
o
n
a
l
c
on
t
r
o
ller
is il
l
u
stra
t
e
d i
n
t
he
F
igure
10.
F
i
gure
1
0
. Co
m
pariso
n of
c
o
n
tr
ol
ler
out
pu
t be
t
w
ee
n
t
h
e
pr
op
ose
d
a
n
d
c
o
nve
n
tio
na
l
se
lf
t
un
i
ng P
I
D
co
ntr
o
l
I
n
o
rder
t
o
fu
rthe
r
eva
l
ua
t
e
t
he
r
esp
o
n
se
o
f
t
h
e
pr
opo
se
d
co
n
t
r
ol
l
e
r
c
o
nf
igu
r
at
i
on
i
t
h
as
b
e
e
n
eva
l
ua
t
e
d
cha
n
gi
n
g
m
otor
p
ar
a
m
e
t
e
r
s
li
ke
i
n
e
rtia
a
nd
p
h
ase
r
e
si
s
t
a
n
ce
i
ndiv
i
du
al
ly
a
n
d
simu
l
t
a
n
e
o
usl
y
.
Th
e
iner
tia
i
s
inc
r
ea
sed
by
1
5
%
w
h
ile
t
he
p
h
a
s
e
r
esist
a
nce
is
d
ec
re
a
s
ed
b
y
15
%.
W
he
n
s
i
multa
ne
ou
s
t
e
s
t
i
n
g
is
ma
de,
both a
r
e
i
n
c
r
ea
sed b
y
5
0 %.
F
i
gure
11
a
n
d
F
i
g
u
r
e
1
2
il
lu
strate
s
an
d
c
o
m
p
are
s
t
he
p
er
form
anc
e
of
t
h
e
p
ro
po
sed
co
ntr
o
l
l
er
a
n
d
con
v
e
n
t
i
ona
l
s
e
lf
t
un
i
n
g
P
I
D
c
o
n
t
r
o
l
w
h
e
n
t
her
e
i
s
a
s
u
dde
n
c
h
a
n
g
e
in
i
ner
tia
.
Wh
ile
F
igure
11
de
p
i
cts
th
e
sp
e
e
d
res
p
on
se,
Fi
g
u
r
e
12
p
rese
nt
s
t
h
e
cont
ro
l
l
e
r
ou
tpu
t
.
It
c
a
n
be
obser
ve
d
fr
om
t
he
f
i
g
ur
e
th
a
t
t
he
p
ro
pos
ed
con
t
ro
l
l
er
d
el
i
v
e
r
s a better
re
sponse
i
n
c
omp
a
r
i
so
n.
0
0.
0
5
0.
1
0.
1
5
0.
2
0.
25
0.
3
0.
3
5
0.
4
0.
45
0.
5
0
500
1000
1500
2000
2500
3000
3500
4000
4500
Ti
m
e
(
S
)
R
o
t
o
r
S
p
ee
d
(
R
P
M
)
S
peed
R
e
s
pons
e
at
a
d
apt
at
i
o
n
r
a
t
e
gai
n
=
0
.
1
5
S
peed
R
e
s
pons
e
at
a
d
apt
at
i
o
n
r
a
t
e
gai
n
=
0
.
3
0
0
0.
05
0.
1
0.
15
0.
2
0.
25
0.
3
0.
35
0.
4
0.
45
0.
5
0
5
10
15
20
25
30
35
40
Ti
m
e
(
S
)
C
o
n
t
ro
l
l
e
r
O
u
t
p
u
t
(
V
)
P
r
opos
ed
C
ont
rol
l
er
C
o
nv
ent
i
onal
S
e
l
f
t
u
ni
ng
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
:
26
5 –
27
6
27
4
F
i
gure
1
1
.
S
p
e
e
d c
ontro
l
resp
ons
e w
i
t
h
sud
d
e
n
ch
a
n
g
e
i
n
i
n
erti
a
F
i
gure
1
2
.
Con
t
ro
ller
ou
t
put
w
i
t
h
su
dde
n ch
a
nge i
n
rotor
iner
tia
F
i
gure
1
3
d
e
p
ic
t
s
t
he
c
o
n
tro
l
l
e
r
o
u
t
p
u
t
f
or
s
ud
de
n
c
h
a
nge
i
n
ph
a
s
e
res
i
s
t
ance
;
i
t
c
an
b
e
o
b
se
rv
e
d
from
t
he
f
ig
ur
e
t
h
a
t
c
o
n
tr
ol
le
r
out
p
u
t
v
a
r
i
e
s
r
apid
l
y
w
it
h
the
c
h
a
n
g
e
i
n
i
n
erti
a.
T
h
e
r
i
s
e
t
i
me
i
s
s
udden
a
n
d
ca
n hel
p
i
n
en
h
a
nc
in
g
t
h
e
s
p
e
e
d
of
resp
onse
of t
he
c
o
n
tro
l
l
e
r i
n
m
a
in
t
a
in
t
h
e
c
on
s
t
a
n
t
sp
e
e
d
.
F
i
gure
1
3
.
Contr
o
ller
ou
t
p
ut
o
f the
pro
p
o
se
d
con
t
ro
lle
r for sud
de
n
chan
ge
i
n pha
se
re
s
i
s
ta
nc
e
The
spe
e
d
r
e
s
p
ons
e
of
t
he
p
r
opos
e
d
c
o
n
t
r
o
ll
er
a
s
com
p
are
d
w
i
t
h
c
o
nve
n
t
i
ona
l
sel
f
t
un
in
g
c
o
n
t
ro
l
l
e
r
is
d
e
p
ic
te
d
us
i
ng
F
i
g
u
re
14.
T
he
c
ompa
ris
on
is
m
ade
f
o
r
5
0
%
c
h
a
nge
i
n
pha
se
r
e
s
i
s
ta
nc
e
an
d
i
n
er
tia
from
t
h
e
in
itia
l
c
o
nd
it
i
o
ns.
It
can
b
e
obs
e
r
ved
fr
om
t
he
f
i
g
ur
e
t
h
er
e
is
no
o
v
ersho
o
t
i
n
t
h
e
case
of
t
he
p
ro
pos
e
d
con
t
ro
l
l
er,
whe
r
e
as in the
ca
s
e
of conve
n
tional
P
ID
t
here i
s a
s
i
g
n
i
fic
a
n
t
over
s
ho
o
t
from
the r
e
fe
renc
e
val
u
e.
F
i
gure
1
4
. S
pe
e
d
con
tro
l
r
esp
onse
for 50
% c
h
an
ge
i
n
pha
s
e
r
esi
stance
an
d
50 %
cha
nge
in iner
t
i
a
0
0.
0
5
0.
1
0.
15
0.
2
0.
2
5
0.
3
0.
3
5
0.
4
0.
4
5
0.
5
0
50
0
10
00
15
00
20
00
25
00
30
00
35
00
Ti
m
e
(
S
)
R
o
t
o
r
S
peed
(
R
P
M
)
P
r
o
pos
ed
C
o
nt
r
o
l
l
e
r
C
o
nv
e
n
t
i
o
na
l
S
e
l
f
t
u
n
i
ng
0
0.
05
0.
1
0.
15
0.
2
0.
2
5
0.
3
0.
3
5
0.
4
0.
4
5
0.
5
0
5
10
15
20
25
30
35
40
Ti
m
e
(
S
)
C
o
nt
r
o
l
l
e
r
O
u
t
put
(
V
)
P
r
op
o
s
ed
C
on
t
r
o
l
l
e
r
C
o
n
v
ent
i
onal
S
el
f
t
uni
ng
0
0.
05
0.
1
0.
1
5
0.
2
0.
25
0.
3
0.
3
5
0.
4
0.
45
0.
5
0
5
10
15
20
25
30
35
40
Ti
m
e
(
S
)
C
o
n
t
r
o
l
l
e
r
O
u
tp
u
t
(
V
)
P
r
o
pos
ed
C
o
n
t
rol
l
e
r
out
pu
t
f
o
r
s
udd
en
c
h
n
age
i
n
P
h
a
s
e
r
e
s
i
s
t
a
n
c
e
0
0.
0
5
0.
1
0.
1
5
0.
2
0.
25
0.
3
0.
35
0.
4
0.
45
0.
5
0
50
0
10
00
15
00
20
00
25
00
30
00
35
00
Ti
m
e
(
S
)
S
p
eed(
R
P
M
)
S
p
e
ed
r
e
s
p
ons
e
of
c
on
v
ent
i
onal
c
o
n
t
r
ol
l
e
r
S
p
e
ed
R
e
s
pons
e
of
t
he
pr
opos
ed
c
o
n
t
r
ol
l
e
r
Evaluation Warning : The document was created with Spire.PDF for Python.