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.
4, D
e
c
e
m
ber
201
9,
pp.
1742~
17
50
ISSN: 2088-
8694,
DOI
:
10.11591
/ijpeds.
v10.
i
4.pp1742-1750
1742
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
Speed p
erformance evaluation
of BLDC
motor based on
dynamic wavelet neural n
etw
o
rk and PSO algorithm
Adel
A
h
m
e
d
O
bed
1
,
Am
eer Latee
f
S
aleh
2
,
Abb
a
s Kareem Kad
h
im
3
1
D
e
p
a
rtm
e
n
t
o
f El
ectri
c
a
l
E
ngi
n
eeri
n
g
, Mid
dl
e
Tech
nical
Un
i
v
e
rsi
ty
,
Iraq
2
Dep
a
rt
m
e
nt of E
l
ect
rical En
g
ineer
ing, U
n
i
vers
i
t
y of Mis
an,
Iraq
3
D
e
p
a
rtm
e
n
t
o
f El
ectrical
P
o
w
er En
g
i
n
eerin
g Techn
i
q
u
es,
M
i
d
d
le T
ech
nical
Univ
e
rsit
y
,
Iraq
+
Art
i
cl
e In
fo
ABSTRACT
A
r
tic
le hist
o
r
y
:
R
e
ce
i
v
e
d
Dec
31,
2018
Re
vise
d Mar
1
1
,
201
9
Ac
ce
p
t
ed
J
un
1
7
,
2
019
In
t
his
pap
e
r,
s
ev
eral
m
eth
ods
a
re
d
ev
el
op
ed
t
o
con
t
rol
th
e
bru
sh
le
s
s
DC
(BLD
C)
m
otor
s
peed.
S
i
nce
it
i
s
d
i
ff
icu
l
t
t
o
g
e
t
a
g
oo
d
sh
owi
n
g
b
y
ut
il
i
z
in
g
clas
si
ca
l
P
I
D
cont
roll
er,
t
h
e
Dy
n
a
m
i
c
W
a
vel
e
t
Ne
u
r
al
N
etwo
rk
(
D
W
NN)
i
s
th
e
pro
pos
ed
w
ork
in
t
his
p
a
per,
w
it
h
p
a
ra
l
l
el
P
ID
c
on
troll
e
r
t
o
o
b
t
a
i
n
a
no
vel
con
t
ro
ller
n
am
ed D
W
NN-P
I
D
con
t
roller.
I
t collects
t
h
e
art
i
f
ici
a
l
neural
abi
l
i
t
y
of
i
ts
n
etw
o
rks
f
o
r
im
partin
g
f
r
om
m
o
t
o
r
o
f
BL
DC
w
i
t
h
d
rive
s
ys
t
e
m
and
t
h
e
abi
lity
o
f
iden
t
i
fi
ca
t
i
on
f
or
t
he
w
avelet
d
ecompositi
on
a
n
d
c
o
ntro
l
of
th
e
dy
nam
i
c
system
f
u
r
th
ermo
re
t
o
hav
e
a
bi
lity
f
or
a
d
a
pting
an
d
s
e
l
f
-
learn
i
n
g
.
Th
e
s
u
g
g
est
e
d
con
t
ro
lle
r
method
i
s
u
t
i
lizing
to
c
ont
r
o
l
t
h
e
s
p
e
e
d
o
f
B
L
D
C
m
o
t
o
r
o
f
w
h
i
c
h
s
u
p
p
l
y
a
b
e
t
t
e
r
s
h
o
w
i
n
g
t
h
a
n
u
t
i
l
i
z
i
n
g
c
l
a
s
sical
con
t
ro
llers
wit
h
a
w
id
e
ran
g
e
of
c
on
tro
l
.
Th
e
p
r
o
p
os
ed
c
on
troll
er
p
aram
eters
are
m
a
tched
con
tin
uo
u
s
ly
u
s
i
ng
P
articl
e
Swarm
Opti
mizati
on
(PSO
)
algorithm.
T
he
s
imula
tion
results
b
as
ed
o
n
pro
p
o
s
ed
D
W
NN-P
I
D
co
nt
rol
l
er
dem
o
nst
r
ate
a
s
u
peri
or
i
n
th
e
stab
ility
a
n
d
p
erf
o
rm
an
ce
com
p
are
d
at
u
ti
li
zi
ng
clas
si
ca
l
W
N
N
-
P
I
D
an
d
con
v
ent
i
o
n
al
P
ID
c
on
tro
l
lers.
T
h
e
s
i
mula
t
ion
resu
l
t
s
are
acco
m
p
l
i
sh
ed
u
sin
g
M
atl
a
b/S
i
m
u
lin
k
.
I
t
s
hows
that
t
he
p
ropo
sed
cont
rol
scheme h
as a su
p
e
ri
o
r
perfor
m
an
ce.
K
eyw
ord
s
:
Al
go
rit
h
m o
f
P
SO
BL
D
C
m
otor
Dy
n
a
mi
c RWNN
S
i
m
u
l
a
t
i
o
n
a
nd
m
odeli
n
g
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:
A
bbas
K
a
re
em K
a
dhi
m
,
D
e
pa
rtme
nt
o
f
El
e
c
t
rica
l
P
o
w
e
r Engi
nee
r
i
n
g
Tec
h
n
i
que
s,
Mi
ddl
e
Te
ch
n
i
ca
l
Un
iv
e
r
s
ity
,
Ba
gh
da
d,
I
ra
q.
Em
ail:
alzu
ba
i
d
y
_
a
bba
s@
ya
ho
o.com
1.
I
N
TR
OD
U
C
TI
O
N
T
h
e
t
y
p
e
o
f
B
L
D
C
m
o
t
o
r
h
a
s
a
d
i
f
f
e
r
e
n
t
o
f
a
p
p
l
i
c
a
t
i
o
n
s
i
n
a
e
r
o
s
pa
ce,
a
utom
ot
iv
e,
i
nd
u
s
tria
l
aut
o
m
a
t
i
on,
m
il
i
t
a
r
y,
h
ouse
h
old
pr
oduc
t
s
,
com
p
u
t
e
r
s,
e
tc
.
beca
u
se
o
f
i
t
s
h
i
g
h
e
r
e
f
f
ic
i
e
nc
y,
h
i
g
her
t
o
rq
ue
,
low
e
r
w
e
ig
h
t
,
go
o
d
e
ffic
i
enc
y
,
incre
a
si
n
g
p
ow
er
d
ens
ity,
simp
le
i
n
c
o
nstruc
t
i
o
n
,
ease
of
c
o
n
t
r
o
l
an
d
less
o
f
ma
int
e
na
nce
[1,
2]
.
A
BLD
C
m
otor
i
s
a
typ
e
o
f
perm
ane
n
t
sync
h
r
o
n
o
u
s
m
otor
w
i
t
h
r
o
tor
po
si
t
i
o
n
f
ee
dba
c
k
.
It
i
s
ge
nera
l
l
y
dri
v
en
b
y
ut
i
l
iz
in
g
a
thr
ee-
ph
a
s
e
p
o
w
e
r
sem
i
c
o
n
du
ct
o
r
b
ridg
e
.
T
h
e
B
LD
C
mo
t
o
r
i
s
com
m
uta
t
i
o
n
e
l
e
c
tr
on
ica
l
l
y
b
y
i
t
s
t
h
re
e-pha
se
i
nv
e
r
ter
bas
e
d
o
n
r
ot
or
f
ee
d
b
a
c
k
p
o
s
i
tio
n
ins
t
e
a
d
o
f
usi
ng
t
h
e
brus
hes.
T
he
pos
i
tio
n
of
t
h
e
r
otor
i
s
req
u
ire
d
f
or
m
ot
or
s
tart
i
ng
a
n
d
t
o
s
u
ppl
y
a
pro
p
er
s
eq
ue
nce
o
f
com
m
uta
t
i
o
n
t
o
c
o
n
t
r
ol
t
he
p
ow
e
r
d
ev
ice
s
i
n
t
h
e
i
nver
t
er
b
r
i
dge
.
The
m
o
t
o
r
co
nsi
s
ts
o
f
a
r
o
tor
o
f
p
e
r
m
a
nent
ma
gnet
a
n
d
sta
t
or
w
ind
i
n
g
d
istrib
u
t
ed
t
o
ge
nera
te
t
he
t
rapez
o
i
d
a
l
b
a
c
k-
EMF
w
a
ve
for
m
s
a
nd
a
r
e
fe
d
w
i
th
rec
t
an
gu
lar
sta
t
or
c
ur
ren
t
s,
w
hic
h
g
i
v
e
a
t
h
eor
e
t
i
c
a
l
l
y
c
o
n
s
t
a
n
t
tor
que.
T
h
e
o
p
er
at
i
o
n
o
f
t
he
B
LD
C
m
o
tor
esta
bl
ishe
d
up
on
the
sw
i
t
c
h
ing
the
phas
e
s
by
t
h
e
in
ver
t
er
w
hen
t
u
rn
in
g
O
N
t
w
o
phase
s
at
a
ny
t
i
me
w
hi
le
t
he
rest
phase
i
s
f
l
oa
t
w
h
e
r
e
eve
r
y
6
0
d
e
g
r
ees
t
he
t
w
o
pha
se
s
a
r
e
e
nergize
d
a
nd
c
h
an
ge
d
seque
n
tia
l
l
y
e
st
a
b
lis
he
d
up
o
n
t
he
r
o
t
or
pos
it
io
n
[1,
3].
D
u
e
t
o
e
lec
t
ro
ni
c
c
o
mm
uta
t
i
on,
th
is
t
y
p
e
o
f
m
otor
h
a
s
a
m
ore
c
o
mp
le
x
c
ontro
l
a
l
go
rith
m
c
o
mp
a
r
e
d
w
i
t
h
othe
r
mo
t
o
rs
.
To
g
et
a
c
o
m
pl
et
e
an
d
exq
uis
i
t
e
c
on
tro
l
s
c
h
e
m
e
for
BLD
C
m
o
t
or
,
an
ac
cura
t
e
m
o
d
el
i
s
requ
ired.
T
h
e
mot
o
r
m
ode
l
is
t
hen
the
he
a
r
t
o
f
it
s
co
ntr
o
l
dr
ive
si
nce
the
de
ve
lo
pe
d
t
o
rq
ue
Evaluation Warning : The document was created with Spire.PDF for Python.
Int J
P
o
w
E
l
e
c
&
D
ri S
yst
IS
S
N
:
2088-
86
94
Speed per
f
or
m
a
nce evaluation of
BL
DC m
o
tor
b
a
sed on dy
nam
i
c
wavel
e
t
neural ... (Adel
Ahmed O
b
ed)
1
743
rela
t
e
d
t
o
t
r
a
p
e
z
o
ida
l
b
ac
k
e
m
f
and
curr
en
t
is
p
r
odu
c
e
d
[3].
A
rotor
position
is
n
eces
sary
u
s
i
ng
s
ensors
t
o
d
e
t
e
rmin
e
th
e
in
t
e
r
v
al
o
f
th
e
n
e
xt
c
o
mmu
t
a
ti
on
.
Th
e
B
L
D
C
m
o
t
o
r
c
a
n
b
e
c
on
t
r
ol
led
b
y
va
ryi
ng
t
h
e
D
C
b
us
vo
lta
ge or
b
y
c
on
t
r
o
lli
n
g
the
dut
y
cycle
o
f
t
he
P
WM m
eth
od [5,
6].
The
sta
t
or
h
a
s
t
hre
e
w
i
ndi
ng
s
an
d
t
h
e
ro
to
r
has
perm
ane
n
t
m
a
gne
ts
t
o
bu
i
l
d
u
p
B
LD
C
m
o
t
o
r.
A
trape
z
o
i
d
a
l
s
ha
pe
o
f
th
e
mu
tu
al
i
nduc
t
a
nc
e
a
m
ong
the
tw
o
m
a
in
p
ar
t
s
o
f
t
h
e
mot
o
r,
s
ta
t
o
r
a
nd
r
o
tor
,
m
akes
a
t
r
a
p
ezo
id
al
b
ack
-e
mf
i
n
th
e
w
i
nd
ing
of
t
h
e
s
t
a
to
r.
T
h
e
a
nal
ogo
u
s
t
h
re
e-pha
se
c
i
r
cu
i
t
,
BLD
C
m
otor
o
f
s
t
a
r
-
con
n
ec
ted u
s
ed
i
n th
i
s
w
ork dr
i
v
e
n
b
y it
s thre
e
-
phase
i
nver
t
e
r
is show
n i
n
F
ig
ure
1 [1
-
6
].
F
i
gure
1.
T
hree
-phase
i
nv
e
r
te
r w
ith
a
n
anal
o
g
o
u
s c
i
r
c
u
i
t
of
B
LD
C m
o
tor d
r
ive
The
de
ve
lo
pm
ent
i
n
t
he
i
n
d
u
str
i
a
l
f
i
e
l
d
d
e
m
and
hi
g
h
r
el
i
a
ble,
h
igh
a
ccu
ra
cy
,
rap
i
d
resp
on
s
e
,
and
go
o
d
e
ffic
ie
nt
d
ri
ves;
t
her
e
fo
re
,
the
spee
d
con
t
ro
l
of
t
he
m
oto
r
sys
t
em
d
r
i
ve
i
s
es
senti
a
l.
C
lass
ical
c
ontr
olling
by
P
I
D
is
s
tab
l
e,
s
im
pl
e
an
d
ea
sy
p
ara
m
e
t
e
r
a
dj
ust
i
n
g
.
U
n
l
e
ss
i
n
g
e
n
e
ra
lity
i
n
d
u
s
t
r
i
a
l
d
r
iv
e
s
w
i
t
h
a
d
i
f
fe
ri
ng
degre
e
o
f
n
o
n
l
inear
,
unce
r
tai
n
ty
i
n
m
a
th
em
atica
l
m
o
d
el
in
g
o
f
t
h
e
sy
st
em
p
a
r
amet
e
r
v
ari
a
bi
l
i
t
y
a
n
d
ad
ju
sti
ng
p
a
r
a
m
e
t
e
r
s
o
f
P
I
D
c
o
n
t
r
o
l
l
e
r
a
r
e
d
i
f
f
i
c
u
l
t
,
i
n
f
e
r
i
o
r
r
o
b
u
s
t
n
e
s
s
,
su
bse
que
ntl
y
,
i
t
i
s
n
o
t
easy
to
r
ea
l
i
ze
t
he
op
tim
al
c
ase
u
nde
r
leve
l
c
o
n
d
it
io
ns
i
n
the
tr
ue
p
ro
duc
ti
on
.
To
ge
t
m
o
re
a
cc
urate
a
nd
b
e
tter
perfor
ma
nce,
A
Fr
actio
nal
Or
d
e
r
PID
contro
l
l
er
i
s
u
t
i
l
i
zed
h
ere
to
c
o
n
t
ro
l
t
h
e
s
pee
d
o
f
th
is
m
otor
.
Thi
s
c
on
t
r
o
ller
add
s
fle
x
ib
i
l
i
t
y
a
nd
m
a
kes
t
h
e
sys
t
e
m
m
o
r
e robus
t
thus,
e
n
ha
nc
i
ng i
t
s
dyn
a
m
i
c
p
erf
o
rma
n
ce
[
6-9
]
.
In
t
he
p
ast
dec
a
d
es,
in
te
lli
gen
t
c
o
n
tr
ol
t
ech
n
i
que
s
suc
h
a
s
fuz
z
y
l
o
g
i
c,
n
e
u
ral
ne
tw
ork(
N
N
)
,
w
a
vele
t
netw
ork
a
n
d
t
h
e
w
a
ve
l
e
t
ne
ur
al
n
e
t
w
o
rk(
W
N
N
)
have
b
ee
n
use
d
t
o
c
o
n
tr
ol
t
he
B
LD
C
m
o
t
o
r
dri
v
e
system
[1
0,1
1
]
.
T
he
B
LD
C
mot
o
r
i
s
c
on
side
red
as
m
ulti
var
i
a
b
l
e
s
ys
t
e
m
a
nd
n
o
n
-
li
nea
r
.
There
f
ore
it
i
s
no
t
ea
sy
t
o
rea
l
i
z
e
a
fu
lfil
l
r
e
sult
for
BL
DC
m
ot
or
u
ti
l
i
zin
g
l
i
n
ea
r
clas
si
ca
l
co
ntr
o
l
m
e
t
h
o
d
s.
A
W
N
N
i
s
impro
v
e
d
i
n
t
h
i
s
pape
r
to
a
c
h
i
e
ve
h
i
gh
c
ontr
o
l
o
f
a
c
c
u
ra
cy
,
hi
gh
p
e
rfor
m
anc
e
,
a
n
d
dy
n
a
mi
c
cha
r
ac
t
e
ri
st
i
c
a
n
d
p
owerf
u
l
ro
bu
stn
e
ss.
T
he
W
NN
-P
ID
c
o
n
t
r
ol
l
e
r
b
a
sed
o
n
PS
O
t
e
c
h
ni
qu
e
i
s
d
e
ve
lo
pe
d
here
.
It
m
e
r
ge
s
t
h
e
a
b
i
l
i
t
y
of
t
h
e
N
N
for
l
e
arni
n
g
f
r
o
m
t
h
e
dri
v
e
sys
t
em
a
n
d
t
he
a
b
i
l
i
t
y
of
w
ave
l
e
t
dec
o
m
pos
it
io
n
to
i
de
nt
i
f
y
a
nd
gu
i
d
anc
e
o
f
the
d
y
n
am
ic
s
ys
tem
a
nd
a
l
s
o
h
as
the
g
u
i
dance
o
f
s
e
l
f-ada
p
tin
g
a
n
d
self-
l
e
a
rni
n
g.
T
w
o
t
y
p
es
o
f
w
avele
t
netw
ork
ar
e
modi
fied
i
n
th
is
p
a
p
er,
fee
d
f
or
w
a
rd
w
a
v
e
l
et
n
e
u
ral
n
etwork
a
nd
r
ec
urrent
w
a
v
elet
n
e
u
ral
netw
ork
w
ith
o
n
l
i
n
e
t
u
n
i
ng
op
t
i
miz
a
tio
n
u
s
i
ng
P
S
O
a
l
go
r
i
t
h
m
.
M
o
re
ov
er,
an
i
m
p
rove
d
WN
N
ne
tw
o
r
k
w
i
t
h
IIR
f
ilter
is
i
m
p
lem
e
n
t
e
d
t
o
p
r
oduc
e
a
n
o
v
el
ty
c
ontr
o
l
l
er
c
alle
d
dyna
mic
w
a
ve
l
e
t
ne
ura
l
n
etw
o
rk
(
D
W
N
N
)
.
The
no
ve
lt
y
ne
tw
ork
ha
s
a
s
u
peri
or
c
apa
b
i
l
i
t
y
t
o
t
he
c
onve
nti
o
n
al
W
NN
in
a
n
effi
ci
en
t
le
a
r
n
i
n
g
me
ch
an
is
m
and
dy
na
mic
re
sp
onse
.
I
n
t
his
pa
per,
t
he
frac
t
iona
l
order
P
I
D
(F
O
P
ID
)
c
o
n
t
r
o
l
l
er
,
w
a
v
elet
n
eural
ne
t
w
or
k
(W
N
N
-P
I
D
)
con
t
ro
l
l
er
a
nd
the
n
ove
l
t
y
d
y
n
am
ic
w
ave
l
et
n
e
u
ra
l
ne
twork
(DWNN
-P
ID)
c
ontro
l
l
er
a
re
u
se
d
t
o
c
ompa
re
t
h
e
ac
cura
cy
a
n
d
a
dap
t
a
b
le
s
pee
d
c
on
t
r
ol
a
nd
he
nc
e
to
m
inim
iz
e
effe
c
t
ive
l
y
c
o
m
m
u
t
a
ti
o
n
i
n
torq
ue
r
ip
p
l
e
w
ithi
n
a
n
e
n
o
r
m
o
u
s
s
p
e
e
d
r
a
n
g
e
.
I
n
a
d
d
i
t
i
o
n
,
t
h
e
P
S
O
a
l
g
o
r
i
t
h
m
i
s
u
t
i
l
izi
n
g
for
t
u
nin
g
a
nd
o
p
tim
i
z
ing
al
l
t
h
e
para
me
ters
f
or all
c
o
n
t
ro
ller
type
s ut
ilize
d
i
n
thi
s
w
ork.
2.
SPEED CONTROL OF
BL
DC M
O
TOR
BY INTEL
LIGEN
T
TECH
N
IQU
E
S
2.1.
Frac
t
iona
l or
der
PID
con
t
r
o
l
l
er
The
FOP
I
D
c
o
ntroller
P
I
D
i
s
a
stre
t
c
h
of
t
he
c
lass
i
c
al
c
on
tro
l
l
e
r
of
P
I
D
b
a
s
ed
o
n
frac
t
i
on
ca
l
c
u
l
us.
T
h
e
in
te
gra
l
a
n
d
t
he
d
e
r
i
v
at
ive
or
ders
d
eve
l
op
t
he
f
l
e
x
ib
il
i
t
y
a
n
d
gu
id
an
ce
t
h
e
sy
st
e
m
t
o
l
e
ss
se
n
s
i
t
i
v
e
t
o
c
h
ang
e
s
of
p
a
r
amet
e
r
s.
T
h
e
p
a
r
amet
ers
P
I
D
o
f
t
h
e
con
t
ro
lle
r
c
a
n
be
quali
fied
b
y
th
is
di
ffe
re
nt
ia
l equ
a
tio
n
bel
o
w
[6-
9
]:
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 V
ol.
10,
N
o.
4
, Dec
201
9 :
1
7
4
2
– 1
750
1
744
u
t
K
e
t
K
D
e
t
K
D
e
t
(
1
)
H
e
r
e
e
(
t
)
r
e
p
r
e
s
e
n
t
s
t
h
e
e
r
r
o
r
o
f
t
i
m
e
w
h
i
c
h
i
s
a
d
i
f
f
e
r
e
n
c
e
b
e
tw
ee
n
the
re
qu
ire
d
s
e
t
poi
n
t
a
n
d
t
he
m
ea
s
u
red
dri
v
e
o
u
t
put
v
aria
b
l
e
;
u
t
i
s
the
ou
tp
u
t
o
f
t
h
at
c
on
tro
lle
r.
T
herefore,
t
h
e
Laplace
t
r
a
ns
fer
fu
nc
tio
n
of
t
h
e
fra
c
ti
o
n
a
l
P
ID
c
on
tro
lle
r w
ill
be
a
s:
G
s
K
K
S
K
S
(
2
)
2.2.
Wave
let
n
e
u
r
al n
e
t
wor
k
A
w
a
v
e
l
e
t
n
e
u
r
a
l
n
e
t
w
o
r
k
(
W
N
N
s
)
i
s
a
g
r
a
d
e
o
f
t
h
e
n
e
u
r
a
l
n
e
t
w
o
r
k
w
h
i
c
h
is
c
o
n
s
t
ruct
ed
a
cc
or
d
i
n
g
to
w
a
v
e
l
et
t
r
a
n
s
form
.
The
c
o
m
b
ina
t
io
n
o
f
t
hi
s
ne
tw
or
k
is
b
e
t
w
e
e
n
t
h
e
n
e
u
r
al
n
et
wo
rk
a
n
d
w
av
el
e
t
t
ra
n
s
fo
rm,
the
WN
N
c
o
m
p
le
te
ly
e
n
t
ai
ls
t
he
f
eat
ure
s
o
f
the
bo
t
h
.
It
p
er
form
s
a
neura
l
n
e
t
w
o
r
k
b
y
fe
ed-forw
ar
d
w
i
t
h
o
ne
o
u
t
p
ut
l
a
y
er
a
n
d
hi
dd
en
l
a
y
er
b
as
ed
o
n
m
o
re
o
r
on
e
inpu
t
.
T
h
e
n
e
u
ro
ns
o
f
the
h
i
d
d
e
n
la
ye
r
t
h
a
t
h
ave
a
wa
v
e
l
e
t
b
a
si
s
to
c
on
st
ru
ct
a
c
tiv
atio
n
f
u
nctions
a
re
p
erform
e
d
t
he
n
e
u
ral ne
tw
ork.
T
hese
n
eur
o
n
s
o
f
the
n
e
ural
n
e
t
w
o
r
k
a
r
e
almo
s
t
c
all
e
d
as
“
wa
velo
ns”,
w
hose
i
t
's
p
ara
m
e
t
ers
c
o
nt
a
i
n
coe
ffic
i
en
ts
o
f
t
h
e
di
l
a
tio
n
w
a
vele
t
(a)
and
trans
l
a
t
i
o
n
(b)
[1
0-1
3
]
.
I
n
WN
N
,
b
o
t
h
t
h
e
d
i
la
ti
on
a
n
d
th
e
t
r
a
n
sl
a
t
io
n
(p
os
itio
n
)
a
re
o
pt
i
m
i
z
e
d
a
s
w
e
ll
a
s
t
h
e
w
e
ig
h
t
s.
E
ac
h
c
e
rta
i
n
s
i
gna
l
f(t)
of
t
h
i
s
ne
tw
or
k
i
s
app
r
ox
i
m
a
t
ed
b
y
po
p
u
l
a
ri
zi
ng
a
l
in
ea
r
col
l
ec
t
i
o
n
of a su
i
t
da
u
gh
ter wave
l
e
t
s
ψ
,
,
w
he
re
ψ
,
a
re
cr
eated by di
la
ti
o
n
and t
r
ansla
t
i
o
n
fr
om
t
he
m
other
wav
e
l
e
t
ψ
a
s
b
e
l
o
w
:
ψ
,
ψ
,
(
3
)
The
w
a
vele
t
ne
ura
l
n
e
t
w
o
rk
o
utp
u
t i
s
gi
v
e
n
b
y
:
y
∑
w
ψ
,
(
4
)
wher
e
w
r
epresents
the
h
eav
in
ess
o
f
t
h
e
n
node
i
n
t
h
e
h
i
dde
n
layer
to
t
he
o
u
t
p
u
t
l
a
t
er.
a
an
d
b
ar
e
the
d
i
l
a
t
i
on
a
nd
t
ra
n
s
l
a
ti
o
n
fac
t
ors
re
sp
e
c
t
i
v
e
l
y;
x
r
e
p
rese
nt
s
t
h
e
i
np
ut
o
f
t
h
e
n
e
t
w
o
r
k;
ψ
,
i
s
t
h
e
w
a
v
e
l
e
t
fu
nc
ti
o
n
.
Thi
s
p
aper
uses t
h
e
Me
xica
n ha
t w
a
vele
t
as a
w
a
v
e
l
et
fu
nc
ti
on,
w
hic
h
i
s
g
i
ve
n as
f
ol
low
s
:
ψ
x
1x
e
(
5
)
RW
NN
i
s
s
i
m
il
ar
t
o
t
h
at
f
eed
f
o
r
wa
rd
W
NN
wi
th
f
e
e
db
a
c
k
co
nn
ect
i
ons
.
The
fee
dba
ck
i
s
o
b
ta
i
n
ed
via
lin
ki
n
g
s
i
g
na
l
o
f
t
he
o
ut
put
s
trat
um
t
o
t
h
e
i
n
p
u
t
s
t
rat
u
m
o
r
in
one
s
t
r
at
um
t
ha
t
is
c
a
lle
d
sl
ig
h
t
l
y
fe
ed
bac
k
Th
e
inp
u
t
o
f
t
h
e
w
a
vele
t
ne
tw
ork
is
c
onsis
t
of
p
revi
ous
i
np
u
t
o
f
the
sys
t
em
s
am
ples
x
a
nd
the
ou
tpu
t
o
f
s
y
st
e
m
y
(t)
as
a
pp
e
a
r
in
F
ig
ure
2.
A
s
ize
of
i
np
u
t
s
t
o
A
WN
N
is
i
nc
rea
s
e
d
w
ith
i
n
c
rea
s
ing
the
syste
m
i
n
orde
r
to be
re
dupl
ic
a
t
e
d
.
There
f
ore
, e
ac
h
layer
out
pu
t
ca
n be
d
ee
m
e
d
as
[
11-1
4
]:
ψ
ψ
(
6
)
Thi
s
l
ay
er i
n
put
s f
o
r ti
me t
i
s
d
eno
t
ed
a
s:
u
t
x
t
ψ
t
1
∗∅
(
7
)
wh
ere
∅
i
n
d
i
c
a
t
e
d
the
se
l
f-
fee
dba
ck
l
o
op
w
e
i
g
ht
.
The
o
u
t
p
ut
o
f t
h
e
ne
tw
ork
is
t
he
n
:
y
∑
w
ψ
(
8
)
u
t
x
tD
y
t
D
∗r
(
9
)
w
h
e
r
e
x
i
s
t
he
w
anted
sig
n
a
l
,
N
is
t
he
h
i
d
den
stra
tum
neuro
n
s
n
um
be
r,
w
:
the
o
u
t
p
ut
w
eight,
D
,D
: the
d
e
la
y n
u
m
ber
for
the out
p
u
t an
d
inp
u
t
ne
t
w
o
rk a
n
d
r
:
the
ou
t
p
u
t
f
e
e
d
bac
k
l
o
op w
e
i
g
ht.
Evaluation Warning : The document was created with Spire.PDF for Python.
Int J
P
o
w
E
l
e
c
&
D
ri S
yst
IS
S
N
:
2088-
86
94
Speed per
f
or
m
a
nce evaluation of
BL
DC m
o
tor
b
a
sed on dy
nam
i
c
wavel
e
t
neural ... (Adel
Ahmed O
b
ed)
1
745
1
x
1
Ψ
2
Ψ
y(t
)
1
w
2
w
N
w
2
x
M
x
N
Ψ
1
φ
2
φ
N
φ
1
r
2
r
N
r
F
i
gure
2.
S
truc
t
u
r
e
of
RWN
N
The
IIR
f
i
l
t
ers
c
a
n
be
acc
om
pl
is
he
d
as
a
n
a
n
al
o
g
f
or
m
or
d
i
g
i
t
a
l
f
o
rm
f
i
lter
s
.
D
e
sign
of
d
igi
t
a
l
I
I
R
f
ilt
ers
whi
c
h
are
t
a
rd
il
y
d
e
p
e
nd
e
n
t
on
t
h
a
t
of
t
h
e
i
r
a
n
a
l
og
c
o
u
n
t
e
r
par
t
s
be
c
a
use
t
h
e
r
e
is
a
w
eal
th
o
f
r
e
s
ource
s,
w
o
rks
an
d
s
t
ra
i
g
h
t
forw
ard
de
si
gn
m
e
t
h
ods
r
elate
d
a
na
lo
g
fe
edbac
k
fi
lter
des
i
g
n
.
The
ge
nera
l
e
qua
t
i
o
n
o
f
a
n
IIR fi
l
ter
ca
n be
expr
essed a
s
[1
5
]
.
H
⋯
⋯
(
10)
H
∑
∑
(
11)
wher
e
c
a
n
d
d
a
re
t
he
f
il
t
e
r
coeffic
i
en
ts.
F
o
r
t
h
e
im
pl
e
m
e
n
t
a
tio
n
of
t
he
a
bo
v
e
e
qua
ti
on
t
h
e
fol
l
ow
i
n
g
di
ffe
re
nce
e
q
u
a
tio
n
i
s
n
ee
de
d
:
y
n
∑
c
x
nk
∑
d
y
nk
(
12)
w
h
ic
h is of or
d
e
r m
a
x[N
,
M
]. A
b
loc
k
d
i
a
gra
m
for the
I
IR filter
w
it
h N
=
2, M=2
i
s show
n
i
n
F
ig
ure
3.
The
WN
N
i
s
a
r
egi
o
nal
net
w
ork
w
h
e
r
e
t
h
e
func
t
i
o
n
o
f
ou
tpu
t
i
s
w
e
l
l
fo
c
u
se
d
in
b
o
t
h
fre
que
nc
y
doma
i
n
a
n
d
t
i
m
e
dom
ai
n.
M
oreo
ver
,
a
d
ual
l
o
c
a
l
ne
t
w
or
k
m
a
y
be
c
om
p
l
e
t
ed
b
y
i
n
t
e
gra
t
i
n
g
th
is
W
N
N
d
e
s
ign
ca
sc
ade
d
w
ith
a
r
egio
na
l
I
n
fi
nite
I
mpu
l
se
R
es
po
nse
(
I
IR)
[16,
1
7].
F
i
gure
3
s
h
ow
s
t
h
e
s
t
r
u
c
t
u
r
e
th
a
t
appr
ox
im
ate
a
ny
w
a
n
t
ed
s
ig
nal
y(
t)
w
ith
pop
ular
i
z
i
ng
a
li
ne
ar
s
yn
thes
is
o
f
a
se
ries
o
f
da
u
g
h
t
er
w
ave
l
e
t
s
ψ
,
casc
a
ded
b
y
th
e
t
op
ica
l
IIR r
e
c
urrent
n
e
t
w
o
r
k
s.
T
he sa
c
rific
i
a
l
ne
tw
or
k
sig
n
al
y
t
can
b
e giv
e
n
by
:
y
t
∑
c
y
ti
∑
d
y
tj
v
t
(
13)
t
∑
w
ψ
,
(
14)
wh
ere
w
represe
nts
the kt
h coeffi
cients
o
f
weight, M
a
nd c
i
a
re the fee
d forw
ard dela
ys n
u
m
ber and
the
II
R
fi
lter
coe
f
fic
i
e
n
ts,
respec
t
i
ve
l
y
,
N
and
d
j
a
re
t
h
e
f
ee
d
b
ack
d
el
ays
n
u
m
b
e
r
a
n
d
t
h
e
c
o
e
ffi
ci
en
ts
o
f
th
e
rec
u
rsive
fi
l
t
e
r
,
respec
tive
l
y.
T
he
s
igna
ls
v
(
t
)
and
x(t)
a
r
e
t
h
e
c
o
-
i
n
p
u
t
a
n
d
i
n
p
u
t
t
o
t
h
e
s
y
s
t
e
m
a
t
t
i
m
e
t
,
respe
c
t
i
ve
l
y
.
Co-I
nput
v
(t)
is
c
om
monl
y
pr
eserve
d
sm
all
be
ca
use
o
f
sta
b
il
ity
g
oal
in
t
he
f
eed
ba
c
k
.
Th
e
di
l
a
t
i
o
n
a
nd
trans
l
at
i
o
n
fac
t
or
s,
w
e
i
gh
ts
l
i
n
k
a
ge
f
or
W
N
N
,
c
oef
f
i
cients
o
f
the
II
R
f
ilter
a
nd
the
par
a
m
e
ters
o
f
P
I
D
are
l
e
arnin
g
a
nd t
u
ni
ng
o
n
-l
ine i
n
a
n a
l
g
o
ri
thm
of
P
S
O
.
Figure 3.
W
N
N
wi
t
h
I
IR
f
i
lter str
u
ct
ure
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 V
ol.
10,
N
o.
4
, Dec
201
9 :
1
7
4
2
– 1
750
1
746
2.3.
Tech
nique
o
f alg
o
r
i
t
h
m
for par
t
icle
s
w
a
r
m
op
t
imizati
o
n
P
S
O
i
s
a
n
e
w
i
n
h
a
bi
t
a
n
c
e
ba
sed
e
v
o
l
ut
iona
ry
c
o
m
put
a
t
i
o
n
whi
c
h
t
urn
s
o
n
th
e
s
o
cia
l
b
eha
v
io
r
pri
n
ci
p
l
e
l
i
k
e fis
h
s
ch
o
o
l
i
ng o
r
bi
r
d
f
l
oc
ki
ng.
P
S
O
t
e
c
hn
iq
u
e
i
s ce
rtif
i
e
d fr
om rese
a
rc
hi
n
g
o
n
sw
a
r
m
s
like
f
i
s
h
in
struc
tio
n
o
r
b
ir
d
crow
d
i
n
g
.
D
u
e
to
t
he
r
e
s
u
l
t
s
o
f
t
h
e
re
se
arc
h
for
a
bir
d
s
gr
oup,
b
ird
s
can
d
etec
t
fo
od
b
y
con
g
re
ga
t
i
n
g
(
no
t
w
i
thi
n
e
ac
h
part
ic
ular
).
T
h
u
s,
a
ny
gu
ide
i
n
t
he
i
n
h
a
b
i
t
a
n
c
e
c
al
l
e
d
a
partic
le,
w
h
ic
h
is
nam
e
d
sw
a
r
m
bri
llia
nc
e
[7,
1
1
].
T
he
a
l
gor
ithm
o
f
P
S
O
i
s
a
t
y
pe
of
t
he
e
vo
lut
i
onar
y
r
e
c
ko
n
i
n
g
m
e
th
o
d
s
for
resol
v
i
n
g
pr
o
b
l
e
m
s
o
f
o
pt
imi
z
a
tio
n.
T
his
w
a
y
c
a
n
b
e
u
t
ili
ze
d
to
t
he
n
o
n
-
li
nea
r
o
p
t
im
iz
ati
o
n
pro
b
lem
th
a
t
guara
nty
co
ns
tr
ain
t
s
w
a
n
tin
g
the
c
o
s
t
f
unc
t
i
on
grad
uate
.
T
h
e
co
n
s
i
d
ered
c
o
s
t
fu
n
c
ti
o
n
s
a
re
e
st
abl
i
sh
ed
u
pon
the
de
s
i
r
e
d
c
r
it
erio
n. The c
h
o
s
en of t
h
e cr
ite
r
i
a
de
pen
d
s o
n
t
h
e
c
o
ntro
ller
a
n
d
t
h
e dr
ive
.
The basic
p
r
i
nc
iple
o
f
the
P
S
O
a
l
g
o
r
i
t
hm
c
an
b
e
descr
i
be
d
t
h
rou
g
h
t
he
f
lo
w
c
hart
b
e
l
ow
.
W
h
e
r
e
e
a
c
h
p
a
r
t
i
c
l
e
i
n
t
h
e
P
S
O
alg
o
ri
t
h
m
has
tw
o
pr
ima
r
y
o
p
er
at
ors'
v
e
l
oc
ity
a
nd
p
os
iti
on
t
ha
t
upda
te
d
d
u
ring
i
ts
f
lig
ht.
F
o
r
a
mult
i
d
i
m
e
n
si
o
n
al
t
r
o
u
b
l
e
,
t
h
e
po
si
t
i
o
n
a
nd
ve
loc
ity
f
or
a
ny
par
tic
l
e
i
n
the
sw
arm
are
up
d
a
ted
b
y
u
t
iliz
in
g
the f
o
l
l
ow
i
n
g e
q
u
a
t
i
o
n
s
:
v
w
∗
v
c
∗R
∗
l
b
e
s
t
x
c
∗R
∗
g
b
e
s
t
x
(
15)
x
x
v
(
16)
wh
er
e
v
,
x
ar
e
the
spe
e
d
i
n
sta
n
t
and
p
o
s
iti
o
n
o
f
pa
rtic
le
i
a
t
i
t
era
tio
n
k
r
e
s
pe
ct
iv
ely
,
w
i
s
t
h
e
weig
h
t
o
f
iner
tia,
c
a
nd
c
a
re
the
c
on
sta
n
ts of
acc
eler
a
tio
n
and
R
,
R
r
epre
se
nt ra
ndom
v
a
r
iable
s
b
etw
e
e
n
0
a
nd 1.
w
w
(
17)
wher
e
w
a
nd
w
r
eprese
nt
t
he
w
e
i
g
h
t
s
of
e
lem
e
n
t
a
r
y
and
eve
n
t
u
a
l
,
it
e
r
r
epr
e
se
nt
s
the
prese
n
t
itera
t
i
on
ti
m
e
an
d
it
er
is its itera
tio
ns ma
x
im
um num
b
e
r.
B
y t
h
i
s
pa
p
er
, a mult
i
-
fit
n
e
ss fu
nc
t
i
o
n
is ut
i
l
i
z
e
d to de
t
ec
t
the
o
p
tim
al
s
et
t
l
e
m
e
n
t
w
i
t
h
a
l
ow
er
e
r
r
or
o
f
spee
d
ba
se
d
on
t
he
I
n
t
eg
ra
l
of
S
q
u
a
re
d
E
r
r
o
r
(I
S
E
)
c
r
i
t
e
r
i
on
a
n
d
ma
ximum
ove
rsho
ot
(
M
p
)
criterion
a
s bel
o
w
:
C
o
st fu
n
ct
i
on
= m
i
n(IS
E)+
m
in(
M
(
18)
IS
E
e
t
dt
(
19)
M
n
n
(
20)
e(i
)
=
ni
n
i
(
21)
w
h
er
e
n
is
t
he
p
rese
nt
s
pee
d
,
and
n
i
s
t
h
e
re
quire
d
va
lue
of
t
he
s
pee
d
.
In
t
hi
s
pa
per
,
t
hi
s
m
e
tho
d
i
s
u
s
e
d
t
o
f
i
nd
t
h
e
o
pti
m
a
l
v
a
l
u
e
f
o
r
t
he
c
o
n
t
r
oll
e
rs'
pa
ramet
e
rs
t
o
co
nt
r
o
l
B
LD
C
m
o
tor
s
p
ee
d.
R
efe
rri
ng
t
o
t
h
e
on
t
h
e
to
p, the
a
l
g
orit
hm
o
f P
S
O
ca
n be
s
pecifie
d
as
a flow
char
t
sh
ow
n
i
n F
i
gur
e 4.
F
i
gure
4.
F
low
c
h
art
of P
SO
Evaluation Warning : The document was created with Spire.PDF for Python.
Int J
P
o
w
E
l
e
c
&
D
ri S
yst
IS
S
N
:
2088-
86
94
Speed per
f
or
m
a
nce evaluation of
BL
DC m
o
tor
b
a
sed on dy
nam
i
c
wavel
e
t
neural ... (Adel
Ahmed O
b
ed)
1
747
3.
CONT
ROL OF S
PE
ED
F
O
R
B
LDC
MO
T
O
R
B
A
S
E
D
ON PROPOS
E
D DY
N
A
MI
C
W
NN
CONT
ROLLER
The
schem
a
t
i
c
di
a
g
ram
of
t
h
e
B
LD
C
mot
o
r
drive
ba
se
d
on
dy
nam
i
c
W
N
N
c
ont
r
o
l
l
er
i
s
sh
ow
n
in
F
i
gure
5.
I
t
co
ns
ists
o
f
B
L
D
C
m
o
t
or
d
r
i
ve
n
u
t
i
lizi
n
g
t
h
r
e
e-
phas
e
i
nve
rte
r
w
it
h
a
co
n
t
r
o
l
c
i
r
c
u
it
to
p
r
ovi
de
t
he
pro
p
er
s
eque
n
c
e
to
t
he
m
ot
or.
A
step-do
w
n
ch
op
per
u
til
ize
d
t
o
c
on
t
r
ol
t
he
i
n
put
vol
ta
ge
o
f
t
h
e
inve
r
t
er
ac
cord
in
g
to
t
h
e
n
o
v
e
l
d
ynam
i
c
WN
N
co
ntr
o
l
l
e
r
w
it
h
IIR
f
i
lter
.
The
P
S
O
alg
o
rithm
is
u
se
d
to
t
un
ing
a
l
l
the
coe
f
fic
i
e
n
ts of
the
p
r
op
ose
d
con
tro
l
ler.
The
ro
tor
s
p
ee
d
i
s
d
irec
tly
c
on
tro
l
led
b
y
s
ta
t
o
r
vo
l
t
age
.
T
her
e
fore
t
he
d
iffer
e
nce
be
t
w
een
t
he
s
e
t
spee
d
an
d
ro
to
r
spe
e
d
w
h
i
c
h
de
t
e
c
t
e
d
b
y
the
rotor posi
t
i
o
n
tec
h
n
i
q
u
e is
u
s
e
d
t
o
d
riv
e
t
he con
tro
l
l
e
r.
T
he
r
otor
pos
it
io
n
sch
e
me
d
ete
c
t
e
d
b
y
the
rot
o
r
p
o
sit
i
on
s
t
rate
g
y
,
it
's
t
o
gener
a
te
t
he
p
ro
pe
r
c
o
mm
utat
i
o
n
s
i
gna
l
s
f
or
sw
it
c
h
i
n
g
t
h
e
i
nver
t
e
r
.
H
e
r
e
t
he
o
u
t
put
o
f
t
h
e
dy
nam
i
c
W
N
N
-
P
I
D
contr
o
lle
r
c
o
n
t
r
o
l
s
t
he
d
u
t
y
c
y
c
l
e
of
t
he
st
e
p
-dow
n
c
h
oppe
r.
T
his
co
n
t
r
o
l
l
e
r-base
d
o
n
P
S
O
a
l
gor
i
t
h
m
i
s
ut
il
ize
d
t
o
co
ntr
o
l
BLD
C
m
ot
or
s
pee
d
i
n
a
n
exte
ns
i
v
e
r
a
ng
e.
I
t
provi
des
b
e
t
t
e
r
s
h
o
w
i
ng
t
ha
n
c
l
a
ssic
a
l
P
ID
contro
l
l
er
a
nd
c
l
a
s
si
ca
l
WN
N
-
PID
contro
ller
as w
i
ll be shown in the res
ults.
F
i
gure
5.
S
c
h
em
atic d
ia
g
r
am
of B
L
D
C
m
ot
o
r
w
ith
t
he
pro
p
o
se
d
con
tro
l
l
e
r
4.
SIMU
L
A
TION
R
ES
ULT
S
FOR
A
P
ROPOSED DWNN-PID
CONTROLLE
R
W
ITH IIR
FILT
E
R
The
o
v
e
r
al
l
BLD
C
m
ot
o
r
S
im
ul
i
n
k
m
ode
l
w
ith
i
ts
d
ri
ve
s
ys
tem
i
s
gi
ve
n
in
F
i
g
ur
e
6.
T
h
e
S
im
ulin
k
mode
l
o
f
t
hi
s
m
o
tor
is
c
o
n
st
ruc
t
e
d
i
n
M
a
tl
ab
fro
m
it
s
m
a
t
h
e
m
at
i
c
a
l
equ
a
tio
n
.
T
he
o
v
e
ra
ll
d
ri
ve
s
yst
e
m
of
B
L
D
C
m
o
t
o
r
b
a
s
e
d
o
n
D
W
N
N
-
P
I
D
c
o
n
t
r
o
l
l
e
r
w
i
t
h
I
I
R
f
i
l
t
e
r
i
s
c
o
n
s
tr
ucte
d
i
n
S
imul
in
k
/
Ma
tla
b
pr
o
g
r
a
m
,
versi
on
2
0
1
5
b
w
it
h
sam
p
lin
g
peri
od
t
h
at
a
ssum
e
d
i
n
t
his
m
odel
i
s
1s.
T
h
e
sim
u
la
t
i
o
n
r
esul
ts
s
ho
w
t
h
at
t
he
D
W
N
N
is
c
a
p
ab
le
t
o
dri
v
e
t
h
e
s
y
ste
m
w
ith
a
l
a
r
ger
num
ber
of
ite
ra
t
i
o
n
s
com
p
ar
ed
w
it
h
t
h
e
con
v
e
n
t
i
ona
l
WN
N
a
nd
by
e
m
p
lo
y
i
n
g
t
he
s
am
e
w
a
ve
l
e
ts
f
unc
t
i
o
n
s
t
yp
e
a
nd
give
s
be
tter
per
f
orm
a
nce
t
h
an
the o
t
her
m
e
thods.
F
i
g
u
r
e
6
.
S
i
muli
n
k
m
odel
of
B
LD
C
motor
dr
iv
e
F
i
gure
7
(a)
show
s
the
ste
p
c
han
g
e
for
a
BLD
C
m
ot
or
s
pee
d
r
espo
nse
.
T
he
c
han
g
e
of
t
he
s
te
p
in
spee
d
for
a
ref
e
re
nce
is
5
0
0
r
pm
a
nd
i
n
cr
eases
5
0
0
rpm
any
0.2se
c
.
F
i
g
ure
7
(
b
)
sh
ow
s
t
h
e
s
p
ee
d
res
p
o
n
se
w
ith
n
o
l
o
ad
a
nd
fu
l
l
l
oad
(
2
N
.
m
)
i
s
adde
d
a
t
t
=
0
.4sec
at
r
e
f
er
e
n
c
e
s
p
e
e
d
i
s
2
0
0
0
r
p
m
.
T
h
e
n
o
l
o
a
d
a
n
d
f
u
l
l
loa
d
c
on
d
iti
on
for
t
h
e
de
ve
l
o
p
e
d
tor
que
i
s
s
how
n
i
n
F
ig
ur
e
7
(c
).
F
ig
ur
es
8
(
a,
b
)
sho
w
s
the
c
u
rrent
i
n
phase
A
a
nd
the
bac
k
e
m
f
v
ol
tage
i
n
p
h
ase
A
respe
c
t
i
ve
l
y
;
F
i
gu
res
9
(
a
,
b
)
show
s
the
l
i
ne
v
ol
ta
ge
(
v
)
a
nd
the
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 V
ol.
10,
N
o.
4
, Dec
201
9 :
1
7
4
2
– 1
750
1
748
pos
it
io
n
s
i
g
n
a
l
o
f
th
e
rot
o
r
respec
t
i
ve
l
y
for
t
h
a
t
sa
me
c
o
ndi
ti
ons
a
t no
l
o
a
d
a
nd ful
l
loa
d.
T
hese
r
e
s
u
lts
e
nsur
e
t
h
e
v
a
li
dity
o
f
t
h
e
mo
d
e
l
,
t
h
e
s
p
eed
,
c
u
rre
n
t
an
d
to
rqu
e
r
each
to
i
ts
n
e
w
v
a
l
ue
s
qu
ic
kl
y.
T
he
b
ac
k
e
m
f
r
e
vea
l
s
the tra
p
ezo
i
d
al
s
ha
pe
a
n
d
the
r
ot
or pos
it
ion
rea
l
iz
es six sec
to
r
s
duri
n
g the
2
per
i
od.
(a)
(b
)
(c)
F
i
gure
7.
(
a) A
s
t
e
p cha
nge i
n
se
t spee
d. (
b) S
pee
d
r
e
s
po
ns
e
fo
r
direc
t
s
t
a
r
tin
g.
(c) D
evel
ope
d
tor
que
f
or
no l
o
ad a
n
d
l
oad
c
o
n
d
i
t
i
on
s.
(a)
(b
)
Fi
g
u
re
8
.
(a
)
P
h
as
e
A
c
u
r
r
e
n
t
o
f
B
L
D
C
mot
o
r. (b) P
ha
se A
bac
k-e
m
f
v
o
lta
ge
of B
L
D
C
m
ot
or.
(a)
(b
)
F
i
gure
9.
(
a) Line
vo
lta
ge (
v
)
of BLD
C m
o
t
o
r. (b)
R
ot
or
p
os
i
t
i
o
n
s
ig
na
l
o
f
B
LD
C m
o
tor.
Evaluation Warning : The document was created with Spire.PDF for Python.
Int J
P
o
w
E
l
e
c
&
D
ri S
yst
IS
S
N
:
2088-
86
94
Speed per
f
or
m
a
nce evaluation of
BL
DC m
o
tor
b
a
sed on dy
nam
i
c
wavel
e
t
neural ... (Adel
Ahmed O
b
ed)
1
749
5.
COMPARISO
N
O
F
D
IFF
E
RE
N
T
CONTROL
M
ETHODS FO
R
SP
EED CONTRO
L OF
A
BLDC
MOT
O
R D
R
I
V
E
A
c
o
mpa
r
iso
n
am
ong
var
i
o
u
s
w
a
ve
le
t
ne
ur
al
n
e
t
w
o
rk
s
chem
es,
frac
t
i
o
n
a
l
a
n
d
c
l
a
s
s
i
c
a
l
P
I
D
con
t
ro
l
l
ers
i
s
i
l
l
u
s
t
r
a
te
d
i
n
T
ab
le
1
.
Fr
om
T
h
i
s
c
o
m
p
ari
s
on,
a
s
a
r
e
s
ul
t
,
s
ho
ws
t
h
a
t
a
p
r
opo
sed
DW
NN
con
t
ro
l
l
er
w
i
t
h
IIR
f
i
l
t
er
i
s
the
pe
rfe
c
t
w
a
y
t
o
bea
t
on
t
h
e
no
n
l
i
n
ea
ri
ty
i
n
this
s
ys
tem
w
i
t
h
g
o
o
d
r
el
ia
b
ili
ty,
add
i
tio
na
l
p
o
w
e
rful
a
n
d
be
t
t
e
r
achie
v
e
me
nt
c
om
par
e
t
ha
n
a
not
he
r
w
a
y
a
s
ap
pe
are
d
i
n
F
i
gure
10
w
i
t
h
l
oa
d
cha
nge
c
o
ndi
t
i
ons.
Tab
l
e
1.
A
c
ompa
risi
on
in
s
pe
ed
p
erfor
m
a
n
ce
w
ith
d
i
ffe
r
e
nt c
o
n
t
ro
l
m
e
t
h
od
s
Per
f
o
r
man
c
e
Ri
s
e
ti
me
Se
ttli
ng
ti
me
St
e
a
d
y
sta
t
e
er
r
o
r
Ove
r
shoot
%
W
N
N
-
PID
0.
0035
0
.
0
3
3*
10
8
*
10
RW
NN
-PID
0
.
0038
0
.
0
4
2*
10
0
D
W
N
N
-P
ID
0
.
0
0
3
0
.
0
3
2.
5*
10
0
R
D
WNN
-
PID
0.
005
0
.
0
2
2*
10
4
*
10
F
O
P
I
D
0.
015
0
.
2
5
2*
10
5
*
10
P
I
D
0.
03
0
.
3
5
5*
10
5
.
6
*
10
F
i
gur
e 1
0
.
S
p
e
e
d
performa
n
c
e
w
ith d
iffe
ren
t
c
on
t
r
o
l
m
e
t
h
ods
a
n
d
w
i
t
h loa
d
c
ha
n
g
e
c
o
ndi
ti
o
n
s.
6.
CONCL
U
S
IONS
I
n
t
hi
s
p
a
per,
C
l
a
ssi
ca
l
P
I
D,
F
O
P
ID,
WNN-
P
I
D,
R
WNN-PI
D
,
D
W
N
N
-
PI
D
an
d
R
D
W
NN-P
I
D
con
t
ro
l
l
er
h
a
s
b
ee
n
u
til
ize
d
t
o
c
o
n
t
r
o
l
t
h
e
s
p
eed
o
f
BLD
C
m
ot
or
.
PS
O
tec
hni
que
h
a
s
b
ee
n
used
f
o
r
t
un
n
i
n
g
the
pa
ram
e
ter
s
o
f
t
h
e
s
ugge
s
t
e
d
c
o
n
tro
l
ler.
T
he
c
om
pari
s
on
be
tw
e
e
n
t
h
e
differ
en
t
m
e
th
o
d
s
show
s
tha
t
a
pro
pose
d
D
WN
N
-
P
I
D
control
l
er
i
s
t
h
e
be
t
t
er
m
etho
d
t
o
c
on
tro
l
t
h
e
spe
e
d
of
B
LD
C
motor
an
d
t
h
e
r
e
sults
show
s
t
h
a
t
the
B
LD
C mot
o
r
pe
rfor
ma
nce u
n
d
er d
i
f
fer
e
n
t
o
per
a
t
i
ng
c
o
n
d
i
tio
ns is
im
pro
v
e
d.
M
ore
ove
r, a dri
ve
syste
m
p
rov
i
de
s a robus
t a
nd
flex
i
b
le d
ue
t
o t
h
e
effec
t
o
f c
h
an
g
i
ng i
n
set
s
pe
ed a
n
d
l
oad
torq
ue
.
REFE
RENCES
[1]
R.
N
arm
a
da,
M.
A
ro
massal
a
ne,
"
D
esig
n
an
d
p
e
rf
orm
a
nce
ev
aluat
i
o
n
o
f
f
r
acti
onal
ORDER
co
nt
ro
ll
er
f
o
r
b
r
us
hl
ess
DC
m
ot
or,
"
In
tern
a
tiona
l
Jo
ur
nal o
f
El
ectrical
E
n
g
i
n
e
e
r
in
g an
d
Info
rm
atics
,
vo
l.
6
,
no.
3
,
pp
.
6
0
6
-
617,
20
14
.
[2]
S
.
K
h
u
b
a
lk
a
r
,
A.
C
h
opad
e
,
A
.
S
.
Ju
ng
hare,
M.
V
.
Aw
re,
"Des
ig
n
a
nd
tu
ni
ng
of
f
rac
t
io
na
l
o
r
d
e
r
P
I
D
con
t
ro
ller
f
or
s
p
eed
c
on
trol
o
f
p
e
rman
ent
m
a
gnet
bru
s
h
l
e
s
s
DC
m
oto
r
,
"
IE
EE
Inter
nati
o
n
a
l Co
n
f
eren
ce
on Co
nt
ro
l,
M
e
as
ur
e
m
ent and In
stru
ment
a
tion
(
C
M
I
)
, 2
01
6
.
[3]
C.
L
ucas
,
F.
T
o
o
t
o
o
n
chi
a
n,
Z
.
N
a
s
i
ri
-Gh
e
idari
,
"
Mu
lti-o
b
j
e
ctiv
e
d
es
ign
op
tim
izati
o
n
o
f
a
l
i
n
ear
b
ru
sh
les
s
p
e
rm
anent
m
a
gn
et
m
o
t
or
u
s
i
n
g
p
art
i
cl
e
sw
arm
optim
i
zati
on,"
Ir
an
ia
n Jo
urn
a
l
of
El
ectr
i
c
a
l
&
Electr
o
n
i
c
En
gine
e
r
ing
,
vol.
6,
n
o.
4
,
201
0.
[4]
T
a
sh
ako
r
i
A.,
Ek
tes
a
bi
M
.,
H
o
s
s
e
in
zadeh
N
.
,
"
M
o
d
e
li
ng
o
f
BLD
C
m
o
t
or
w
i
t
h
i
d
eal
b
ack-em
f
f
o
r
a
uto
m
oti
v
e
ap
pl
icat
ions,
"
P
r
o
ceedin
gs
of
t
h
e W
o
r
l
d
Co
ng
res
s
o
n
E
ngin
eerin
g,
vo
l
.
2,
2
01
1.
[5]
A
d
e
l
A
.
O
b
e
d
,
A
b
b
a
s
K
.
K
a
d
h
i
m
,
"
S
p
e
e
d
a
n
d
c
u
r
r
e
n
t
l
i
m
i
t
i
n
g
c
o
n
t
r
o
l
s
t
r
a
te
g
i
e
s
f
or
B
L
D
C
mo
to
r
d
r
iv
e
sy
s
t
e
m
:
A
co
m
p
arati
v
e
stud
y,"
In
ter
n
a
t
io
na
l Jou
r
na
l of
A
d
van
ced En
gin
eeri
ng Resear
ch and
S
c
ien
c
e
,
v
o
l
.
5
(
2
)
,
p
p
.
119
-13
0
,
2018.
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 V
ol.
10,
N
o.
4
, Dec
201
9 :
1
7
4
2
– 1
750
1
750
[6]
L
.
S
a
l
eh
,
A.
A
.
Ob
e
d
,
"S
peed
c
o
n
tro
l
o
f
b
r
us
hles
s
D
C
m
oto
r
b
ase
d
o
n
fra
c
t
io
na
l
o
r
de
r
PID
c
o
ntro
lle
r,"
In
ter
n
a
t
io
nal Jour
na
l o
f
Co
mp
u
t
er A
pplicati
ons
,
vol
.
9
5
,
no.
4
,
pp.
1
-
6
,
2
0
1
4
.
[7]
L
.
S
aleh,
M.
A
.
H
u
s
s
ai
n
,
S
.
M.
K
li
m,
"
O
p
tim
a
l
t
r
a
j
ecto
r
y
tracki
ng
c
o
n
t
r
o
l
f
o
r
a
w
h
eeled
m
o
b
ile
r
obo
t
usin
g
f
r
act
io
na
l
o
r
der
P
I
D
c
o
nt
rol
l
er,
"
Journal
of Uni
versity of Ba
bylon, Eng
i
neering Sci
e
nces
,
vo
l.
2
6,
n
o.
4
,
p
p
.
292
-30
6
,
2018.
[8]
A
k
r
a
m
H
.
A
hm
e
d
,
A
b
d
E
l
S
a
m
i
e
B
.
K
ot
b,
A
ym
a
n
M
.
A
li
,
"
C
o
m
p
a
r
i
s
o
n
b
e
tween
f
uzzy
lo
g
i
c
and
PI co
nt
ro
l f
o
r
t
h
e
s
p
e
e
d
o
f
B
L
D
C
m
o
t
o
r
,
"
Int
e
rn
a
t
io
nal
Jo
urn
a
l
o
f
P
o
wer
El
ectro
n
i
cs
and
D
r
i
v
e
Sys
t
em (
I
JPED
S
)
, v
o
l
.
9
,
no.
3
,
pp
.
1
1
1
6
-1
12
3,
S
ep
.
2
018
.
[9]
Ri
n
k
u
S.,
Su
bh
r
a
nsu
P.,
Gag
a
n
d
eep
K., "
Design
o
f
f
r
acti
o
nal
ord
er P
ID
con
tro
ller
f
o
r s
p
eed
c
on
tro
l
of
DC
m
ot
or,
"
Internat
ional Journal
of
Scient
ific and R
e
search
P
u
bl
i
c
ations
,
vol.
2,
2
01
2.
[10]
M
a
o
h
u
a
Z
hang
,
Chan
gli
a
ng
X
ia,
Y
a
n
g
T
i
a
n,
D
an
L
i
u
,
Zhiq
iang
L
i,
"S
peed
c
o
n
t
r
ol
o
f
b
r
us
h
l
es
s
D
C
m
o
t
or
b
as
ed
o
n
s
i
n
g
l
e
n
e
u
r
o
n
P
I
D
a
n
d
w
a
v
e
l
e
t
n
e
u
r
a
l
n
e
t
w
o
r
k
,
"
IEE
E
In
ternat
io
nal Con
f
er
e
n
ce on
Control
and Automa
tion
Gu
an
gzh
o
u
,
C
hina,
2
00
7.
[11]
A
d
e
l
A
.
O
b
e
d
a
n
d
A
m
e
e
r
L
.
S
a
l
e
h
,
"
S
p
e
e
d
c
o
n
t
r
o
l
o
f
B
L
D
C
m
o
t
o
r
b
as
ed
o
n
recu
rren
t
wav
e
let
n
e
ural
n
etw
o
rk,
"
Ir
aq
J.
Elect
r
i
ca
l a
nd El
ectr
onic
E
n
g
i
neer
in
g
,
vo
l.
1
0
,
n
o
.
2,
p
p.
1
18-12
9,
2
01
4
[12]
S
u
ng
J
in
Y
oo
,
J
i
n
Bae
Park
,
Yoon
Ho
C
h
o
i,
"
Di
rect
a
dap
tiv
e
co
nt
rol
usin
g
se
lf
-re
cu
rrent
w
avelet
n
eural
n
e
t
w
o
r
k
v
i
a
adapt
i
v
e
l
earn
i
n
g
r
at
es
f
o
r
s
t
a
bl
e
pa
t
h
t
racki
n
g
o
f
m
o
b
i
l
e
robots,"
IEE
E
A
m
erican
Co
ntrol
Confer
ence.
P
r
o
ceedin
gs
o
f
the
, vo
l
.1
. p
p.
28
4
8
-
2
8
5
2
, 2
00
5.
[13]
S
u
ng
J
in
Y
oo
,
Yo
on
H
o
Cho
i
,
J
i
n
Bae
P
a
rk,
"Gen
e
r
al
ized
p
redi
cti
v
e
c
o
n
t
r
o
l
b
ased
o
n
self
-recurrent
w
avel
et
n
e
ural
n
etw
o
rk
f
o
r
s
tabl
e
path
t
racki
n
g
o
f
m
o
b
ile
rob
o
t
s
:
A
d
apt
iv
e
learni
ng
r
at
es
a
pp
roach,
"
IEEE Tran
sac
t
io
n’
s
on
C
i
rc
ui
ts an
d
S
y
ste
m
s I
,
v
o
l
.53
, pp
. 2
88
-2
93
, 2
00
6.
[14]
L
.
S
a
l
e
h
,
B
.
A
.
O
b
a
i
d
,
A
.
A
.
O
b
e
d
,
"
M
o
t
i
o
n
c
o
n
t
r
o
l
o
f
l
i
n
e
a
r
i
n
duct
i
o
n
m
oto
r
b
as
ed
o
n
o
p
t
i
m
a
l
recurren
t
w
avel
et
neural
N
et
wor
k
-PID
c
ont
r
oll
e
r,"
I
n
te
r
n
a
t
io
na
l J
o
ur
na
l o
f
E
n
g
i
n
e
e
r
in
g
&
T
e
c
h
no
lo
gy
,
vo
l.
7,
n
o.
4,
p
p
.
20
28-2
0
3
4
,
20
18.
[15]
Ranjit
Singh,
S
andeep
K
.
Arya,
"O
pti
m
i
zat
i
on
of
IIR
d
igi
t
al
f
i
l
ters
u
s
i
n
g
p
articl
e
s
w
arm
optim
i
zatio
n,
"
I
EEE
In
ter
n
a
t
io
nal
Conf
erence on Co
mm
un
ica
t
io
n
,
In
fo
rm
a
tio
n
&
Compu
t
i
n
g
T
echno
lo
gy
(
I
CCICT)
,
M
u
mb
ai,
India,
pp
.
1
-
7,
2
01
2.
[16]
Che
n
g-Jia
n
L
in
,
Ch
i-Yu
ng
Le
e
,
C
he
n
g
-
C
hu
ng
C
hin,
"
d
y
n
a
mic
re
c
u
rr
ent
wavelet
network
cont
rollers
f
o
r non
li
near
sy
ste
m
c
on
tr
o
l
,"
J
o
ur
na
l of
t
h
e Ch
in
ese In
s
t
it
u
t
e of
En
gin
eers
, vo
l
.
2
9
,
no
.
4
,
pp
. 7
47
-75
1
,
2
00
6.
[17]
A
d
el
A
.
O
b
ed
,
Ab
bas
K
.
K
ad
him
,
"
Mu
lti-res
o
luti
on
w
avel
et
P
ID
s
p
e
e
d
and
cur
r
ent
con
t
ro
llers
of
B
LD
C
m
o
t
o
r
b
a
sed
o
n
i
nv
asi
v
e
w
eed
opt
im
i
zatio
n
tech
niq
u
e,"
In
ter
n
a
t
i
onal Jou
r
na
l of
Ap
p
l
ied
E
ngi
neer
in
g
R
e
sea
r
ch
,
vo
l.
13,
n
o
.
8
,
p
p
.
6
2
34-624
3,
201
8.
Evaluation Warning : The document was created with Spire.PDF for Python.