Indonesi
an
Journa
l
of El
ect
ri
cal Engineer
ing
an
d
Comp
ut
er
Scie
nce
Vo
l.
23
,
No.
1
,
Ju
ly
2021
, p
p.
110
~
119
IS
S
N: 25
02
-
4752, DO
I: 10
.11
591/ijeecs
.v
23
.i
1
.
pp
110
-
119
110
Journ
al h
om
e
page
:
http:
//
ij
eecs.i
aesc
or
e.c
om
Manag
em
ent switchin
g an
gles
r
eal
-
time pr
edicti
on b
y
a
rtifi
cial
neural n
etwork
Moham
med
R
as
heed
Jub
air
A
l
-
Hiealy
, M
ohamm
ad
Sh
ahi
r Bi
n Abdul
Maje
d Shi
kh
,
Ab
d
urra
hman Bi
n Jalil
, Su
ha
il
a Ab
dul
R
ah
m
an
,
Muat
h Jarr
ah
Depa
rtmnet
o
f
C
om
pute
r
Engi
n
e
eri
ng,
Univer
s
i
t
y
Malay
si
a
of
Co
m
pute
r
Scie
n
ce
&
Eng
ine
er
ing
(
UN
IMY
),
Sela
n
gor,
Malay
s
ia
Art
ic
le
In
f
o
ABSTR
A
CT
Art
ic
le
history:
Re
cei
ved
N
ov
7
, 2
02
0
Re
vised
A
pr
27
, 2
021
Accepte
d
Ma
y
10
, 202
1
Artifi
c
ia
l
n
eur
a
l
net
works
(AN
Ns
)
is
an
eff
icie
nt
wa
y
for
diff
e
ren
t
t
y
p
e
s
of
rea
l
-
world
pre
di
ct
ion
prob
le
m
s.
In
the
p
ast
de
ca
d
e,
it
h
as
give
n
a
tre
m
endous
surge
in
a
glob
al
rese
arc
h
ac
t
i
vit
ie
s
.
AN
Ns
embody
m
uch
c
ert
a
inty
and
provide
a
gre
at
deal
of
prom
i
se
Thi
s
p
ape
r
has
pre
sent
arti
fic
i
al
n
eur
a
l
net
work
(AN
N)
te
chn
ique
anal
y
s
is
and
pre
dic
ti
on
for
m
ana
gement
sw
it
chi
ng
angl
es
re
al
-
t
ime.
The
proposes
t
o
be
used
AN
N
for
pre
diction
a
nd
sele
c
te
d
obti
ne
angl
es
for
implement
the
t
iming
dia
gra
m
f
or
m
uli
tl
ve
l
inverter
c
irc
u
it.
In
orde
r
to
cont
r
ol
the
funda
m
en
ta
l
component,
AN
Ns
are
used
to
solve
th
e
ana
l
y
sis
of
non
-
li
ne
ar
equa
t
ion
of
th
e
ou
tput
t
iming
dia
gr
am
in
orde
r
to
det
ermine
th
e
sw
it
chi
ng
ang
le
s
.
Subs
ta
nti
a
lly
,
the
num
ber
of
sw
it
chi
ng
devi
c
es
are
red
u
ci
ng
as
poss
ibl
e
basic
a
lly
for
red
uci
ng
a
sw
it
chi
n
g
loss
in
the
s
y
stem,
al
so
hav
e
bee
n
used
AN
Ns
te
chni
que
to
opti
m
iz
e
a
sw
itc
hing
angl
e
s
beha
vior
to
red
uce
tot
a
l
har
m
onic
distort
ion
(T
HD
)
at
volt
age
and
cur
ren
t
output
wave
fo
r
m
equa
l
THD
V
8.
05%
THDA
5.
1%.
For
t
he
proposed
cont
rollers,
the
per
form
anc
e
an
d
result
s
b
y
the
AN
N
s
were
o
bta
in
e
d
and
compare
d
b
y
usi
ng
MA
TL
AB software
.
Ke
yw
or
d
s
:
Ar
ti
fici
al
intel
li
gen
ce
Har
m
on
ic
s
op
t
i
m
iz
ation
Neural
netw
ork AN
N
Sw
it
chin
g
a
ngle
This
is an
open
acc
ess arti
cl
e
un
der
the
CC
B
Y
-
SA
li
ce
nse
.
Corres
pond
in
g
Aut
h
or
:
Moh
am
m
ed
Ra
sh
ee
d
J
ub
ai
r Al
-
Hieal
y
Dep
a
rtm
ent o
f C
om
pu
te
r
E
ng
i
neer
i
ng
Un
i
ver
sit
y M
al
ay
sia
o
f
C
om
pu
te
r Sci
ence
&
Enginee
rin
g (
UNIMY
)
Sela
ngor, Mal
a
ysi
a
Em
a
il
: rash
eed
.alhiea
ly
@unim
y.edu
.m
y
1.
INTROD
U
CTION
A
m
ultilevel
i
nv
e
rter
is
an
el
ect
ro
nic
ci
rc
uit
with
m
ult
i
-
le
vel
powe
r,
this
inv
e
rter/ci
rcu
it
can
be
conve
rted
i
nto
a
dir
ect
po
w
er
(
DC)
outp
ut
sup
ply
an
d
supp
li
es
a
n
al
te
rn
at
e
volt
age
(A
C)
wa
ve
form
.
”R
enew
a
ble
s
ources
of
ene
r
gy
su
c
h
as
ph
otovo
lt
ai
cs,
wind
an
d
water
tur
bin
es
a
re
now
e
ssentia
l
e
nergy
so
urces
du
e
to
their
ren
e
wa
bl
e
and
en
vir
onm
ental
fr
ie
nd
li
ness”
[
1]
-
[
3].
Most
of
these
ren
e
wa
ble
pathw
ay
s
gen
e
rate
a
DC
wav
e
f
or
m
as
a
n
outp
ut
that
is
no
t
a
pprop
riat
e
for
tran
sm
iss
ion
s
,
distri
bu
ti
on
s
a
nd
use
s.
Ba
sed
on
t
his,
the
D
C
ou
tp
ut
pow
e
r
shou
l
d
be
c
onve
rted
to
AC
powe
r
wa
veform
and
this
can
be
a
chieve
d
t
hro
ugh
the
MLI
ci
rc
uit
[2
]
-
[
5]
.
I
n
ge
ner
al
,
b
ase
d
on
their
pr
in
ci
pa
l
structur
e
,
MLI
str
uctur
e
s
ha
ve
bee
n
cat
eg
or
iz
e
d
into
three
m
ain
topolo
gies
nam
el
y
casca
ded
h”
br
i
dg
e
(CHB
-
MLI
)
w
hi
ch
is
the
m
os
t
com
m
on
and
reli
able
,
diode
-
cl
am
ped
(
DC
-
MLI
)
a
nd
fly
”capaci
tor
(F
C
-
MLI
).
A
co
nn
ect
e
d
cel
ls
in
se
ries
in
the
loa
d
c
om
po
se
a
CHB
-
MLI
t
opology,
w
hile
conve
ntion
al
st
r
uctu
re
of
CHB
-
MLI
us
es
f
our
switc
hes
pro
du
ci
ng
th
ree
le
ve
ls
for
each
cel
l
[4]
,
[
5]
.
Seve
ral
m
o
du
la
ti
on
m
et
ho
ds
f
or
el
im
inatin
g
a
ha
rm
on
ic
durin
g
the
out
pu
t
wav
e
form
,
su
c
h
as
the
pulse
wi
de
m
od
ule
(PWM
),
cal
le
d
a
high
fr
e
quen
c
y
switc
hin
g
te
chn
i
qu
e
,
inclu
ding
S
-
P
W
M
a
nd
SV
-
P
W
M
ha
ve
be
en
us
e
d.
Sec
ondly,
sel
ect
ive
har
m
on
ic
el
i
m
inati
on
(
SH
E
-
P
W
M)
m
et
ho
d,
of
te
n
cal
le
d
the
best
Evaluation Warning : The document was created with Spire.PDF for Python.
Ind
on
esi
a
n
J
E
le
c Eng &
Co
m
p
Sci
IS
S
N:
25
02
-
4752
Ma
nage
men
t s
wi
tc
hin
g a
ng
le
s rea
l
-
ti
me pre
dicti
on
by arti
fi
ci
al… (
Moha
mm
e
d Ras
hee
d Ju
ba
ir
Al
-
Hiealy)
111
m
od
ulati
on
m
e
thod,
as the
s
pa
ce vecto
r
co
nt
ro
l
(
SV
C
),
is a
fun
dam
ental
te
chn
i
qu
e
f
or
fre
qu
e
ncy
cha
nge
s
[6]
-
[11]
.
“I
n
ad
diti
on,
a
com
plex
transce
nd
e
ntal
no
t
li
near
e
qu
a
ti
on
(SHE
-
P
W
M)
can
be
f
ou
nd
wh
ic
h
has
t
o
be
reso
l
v
ed
,
w
hic
h
can
not
be
ea
sil
y
so
lved
by
si
m
ple
m
et
ho
ds
[12],
w
hile
m
eans
of
it
erat
ive
co
ntr
ol
m
e
thods
can
overc
om
e
this
com
plex
e
qu
at
io
n
(SHE
-
P
W
M)
s
uch
as
opti
m
isa
ti
on
of
arti
fici
al
ne
ur
al
netw
ork
(
ANN),
firesf
ly
al
gorithm
(F
FA
),
part
ic
le
swar
m
op
tim
iz
at
ion
(
PS
O)
,
a
nd
new
t
on
-
raphs
on
(N
R
)
[13
]
-
[
15]
to
def
i
ne
and
cal
culat
e
uniq
ue
switc
hi
ng
A
ng
le
s
f
or
lo
wer
T
HD
an
d
m
or
e
s
m
oo
th
wav
e
f
orm
of
a
m
ulti
plex”
[7]
,
[12]
-
[15]
.
TH
D
can
essenti
al
ly
b
e
def
i
ned
as the
rati
o
of
high harm
on
ic
s w
it
h
low
ha
rm
on
ic
s as sh
own,
wh
e
n
an
is
a h
igh
har
m
on
i
c and
a
1
low h
arm
on
ic
s
[16].
Ther
e
f
or
e
,
outpu
t v
oltage
ca
n be g
e
ner
at
ed b
y
a n
ine level C
HB
-
MLI
ba
s
e
d
o
n
F
i
gu
r
e
1
a
n
d
F
i
gu
r
e
2
,
THD
=
∑
(
a
n
)
∞
n
=
1
,
3
,
5
…
a
1
V
out
=
V
c1
+
V
c2
+
V
c3
+
V
c4
+
V
c5
2.
SHE
-
P
WM
TE
CHNIQ
UE
PROP
OS
E
D MET
HO
D
A
sel
ect
ive
ha
rm
on
ic
el
i
m
i
nation
S
HE
-
P
WM
te
chn
i
que
fr
om
a
fu
ndam
ental
swit
ch
fr
e
que
ncy
strat
egy
is
a
m
od
ulati
on
te
c
hn
i
qu
e
that
ha
s
b
ee
n
us
e
d
i
n
this
researc
h
pap
e
r
.
“
T
o
produce
a
sym
m
et
rical
qu
a
rter
wa
vefor
m
in
this
a
ppr
oac
h,
t
he
e
quat
ion
of
t
he
fast
f
ourier
se
ri
es
(
FFT
)
is
app
li
ed
to
the
outp
ut
vo
lt
age
sta
irca
se
wa
vefor
m
to
sat
isfy
the
be
st
switc
hing
ang
le
s
[
11
]
”.
The
(Vo
ut)
F
FT
sta
te
m
ent
can
be
com
pu
te
d
as
,
V
(
t
)
out
=
∑
a
n
sin
(
n
ω
t
)
∞
n
=
1
+
b
n
cos
(
n
ω
t
)
Fig
ure
1
.
O
ne of
cell
f
or
nin
e
level C
HB
-
MLI
Fig
ure
2
.
Pro
posal
fo
r nine
-
le
vel CHB
-
MLI
ci
rcu
it
“
W
hi
c
h
c
l
e
a
r
e
d
a
l
l
ki
nd
s
of
ha
r
m
on
i
c
s
(
b
ne
qu
a
l
z
e
r
o
)
a
nd
p
r
e
s
e
nt
e
d
on
l
y
t
he
od
d
ha
r
m
on
i
c
s
”
.
“
T
he
od
d
c
oe
f
f
i
c
i
e
nt
(
a
n)
c
a
n
t
he
r
e
f
or
e
be
de
t
e
rm
i
ne
d,
a
nd
t
he
s
w
i
t
c
hi
ng
a
ng
l
e
s
c
a
n
on
l
y
be
c
a
l
c
ul
a
t
e
d
at
t
he
i
ni
t
ia
l
qu
a
dr
a
nt
w
a
ve
f
or
m
.
”
Th
e
r
e
by
,
4
S
w
i
t
c
hi
ng
A
n
gl
e
s
a
t
l
e
ve
l
-
9
M
L
I
s
ho
ul
d
be
de
f
i
n
e
d
a
nd
c
a
l
c
ul
a
te
d
a
s
t
he
c
r
i
t
i
ca
l
a
ng
l
e
a
t
f
i
r
st
qu
a
d
r
a
nt
w
a
ve
f
or
m
(
θ
1
,
θ
2
,
θ
3
a
nd
θ
4
)
a
s
s
ho
w
n
i
n
F
i
gu
r
e
4
ba
s
e
d
on
T
a
bl
e
1
[
1
7]
.
T
he
o
dd
(
a
n
)
c
oe
f
f
i
c
i
e
nt
c
a
n,
ho
w
e
ve
r
,
be
c
a
l
c
ul
a
t
e
d
in
,
a
n
=
4
π
∫
V
sin
(
n
θ
)
d
θ
π
2
⁄
0
t
he
r
e
f
or
e
t
he
i
np
ut
s
up
pl
y
un
i
t
s
a
r
e
r
e
qu
i
r
e
d
,
a
s
s
um
i
ng
t
ha
t
t
he
M
L
I
w
hi
c
h
pr
op
os
e
s
t
h
e
s
ym
m
et
r
i
c
al
C
H
B
-
M
L
I
i
s
e
qu
i
va
l
e
nt
.
T
he
M
L
I
o
ut
pu
t
v
ol
t
a
ge
w
a
ve
f
or
m
c
a
n
t
he
n
be
de
f
i
ne
d
by
F
F
T
e
x
pa
n
s
i
on
,
Evaluation Warning : The document was created with Spire.PDF for Python.
IS
S
N
:
2502
-
4752
Ind
on
esi
a
n
J
E
le
c Eng &
Co
m
p
Sci,
Vo
l.
23
, N
o.
1
,
Ju
ly
2021
:
110
-
119
112
V
(
ω
t
)
=
4
V
dc
n
π
∑
[
cos
(
n
θ
1
)
+
cos
(
n
θ
2
)
+
cos
(
n
θ
3
)
+
cos
(
n
θ
4
)
]
∞
n
=
1
,
3
,
5
,
.
.
sin
(
n
ω
t
)
n
i
n
f
a
c
t
,
t
he
o
d
d
c
oe
f
f
i
c
i
e
nt
(
a
n
)
o
f
f
ou
r
i
e
r
s
e
r
i
e
s
(
F
F
T
)
,
“
w
hi
c
h
c
a
n
be
ob
t
a
i
ne
d
by
t
a
ki
ng
V
(
ω
t
)
a
s
t
he
c
om
m
on
f
a
c
t
or
f
o
r
e
ve
r
y
e
q
ua
t
i
on
s
i
d
e
,
ne
e
ds
t
o
be
f
ur
t
h
e
r
a
na
l
y
s
e
d
”
[18]
.
u
nt
i
l
t
he
f
i
na
l
e
qu
a
t
i
on
of
o
d
d
s
e
r
i
e
s
c
oe
f
f
i
c
i
en
t
s
f
or
ha
r
m
on
i
c
or
d
e
r
c
om
po
ne
nt
s
c
a
n
be
ob
t
a
i
ne
d
in
,
a
n
=
4
V
dc
n
π
[
cos
(
n
θ
1
)
+
cos
(
n
θ
2
)
+
cos
(
n
θ
3
)
+
cos
(
n
θ
4
)
]
n
um
e
r
i
c
al
ly
,
4
or
de
r
s
c
a
n
be
o
bt
a
i
ne
d
f
o
r
e
a
c
h
s
t
r
a
ng
e
ha
r
m
on
i
c
or
de
r
a
t
9
-
l
e
ve
l
C
H
B
-
M
LI
in
(
F
i
g
ur
e
3)
,
[
cos
(
θ
1
)
+
cos
(
θ
2
)
+
cos
(
θ
3
)
+
cos
(
θ
4
)
]
=
4π
M
i
2
[
cos
(
3
θ
1
)
+
cos
(
3
θ
2
)
+
cos
(
3
θ
3
)
+
cos
(
3
θ
4
)
]
=
0
[
cos
(
5
θ
1
)
+
cos
(
5
θ
2
)
+
cos
(
5
θ
3
)
+
cos
(
5
θ
4
)
]
=
0
[
cos
(
7
θ
1
)
+
co
s
(
7
θ
2
)
+
cos
(
7
θ
3
)
+
cos
(
7
θ
4
)
]
=
0
a
nd
m
od
ul
a
t
i
on
i
n
de
x
M
i
=
a
1
l
V
dc
w
he
n
l
“
t
he
n
um
be
r
of
i
n
de
pe
nd
e
nt
D
C
s
ou
r
c
e
s
.
,
i
n
f
a
c
t
,
a
m
a
th
e
m
a
ti
c
a
l
e
qu
a
t
i
on
c
a
n
no
t
de
t
e
rm
i
ne
t
he
c
om
pl
e
x
S
H
E
-
P
WM
e
q
ua
t
i
on
.
C
om
pl
i
c
at
e
d
a
pp
r
oa
c
he
s
A
N
N
,
F
F
A
,
P
S
O
,
a
n
d
N
-
R
m
us
t
vi
r
t
ua
l
ly
be
c
a
l
c
ul
a
te
d”
[
1
9]
,
[
2
0]
.
Table
1
.
O
utpu
t vo
lt
age
s
witc
hing
be
ha
vior
f
or n
i
ne
-
le
ve
M
-
CHB
-
ML
Is
S
1
S
2
S
3
S
4
S
5
S
6
V
o
1
0
0
1
0
1
Vd
c
1
0
0
1
0
0
V
dc
/2
0
0
0
1
1
0
V
dc
/3
0
0
0
1
1
0
V
dc
/4
0
0
1
1
0
0
0
1
1
0
0
0
1
0*
0
1
0
0
1
1
-
V
dc
/4
0
1
0
0
1
0
-
V
dc
/3
0
1
1
0
0
0
-
V
dc
/2
0
1
1
0
0
0
-
Vd
c
Fig
ure
3
.
O
utput wa
ve
form
v
ol
ta
ge of
9
-
le
ve
l M
LI
3.
ANN
PE
RF
O
RMA
NC
E E
V
ALU
ATIO
N
ANNs
a
r
e
i
ns
pi
r
e
d
by
bi
ol
o
gi
c
a
l
ne
r
vo
us
s
y
s
t
em
,
i
t
c
a
n
or
ga
ni
s
e
a
nd
pr
oc
e
s
s
i
nf
o
r
m
a
ti
on
a
nd
e
va
l
ua
t
e
t
he
de
s
i
r
e
d
da
t
a
a
s
s
ho
w
n
i
n
F
i
gu
r
e
4
”
[
15
]
,
i
t
i
s
a
po
w
e
r
f
ul
t
e
c
hn
i
qu
e
t
ha
t
h
a
s
be
e
n
s
uc
c
e
s
s
f
ul
l
y
a
pp
l
i
e
d
a
n
d
s
o
l
ve
d
va
r
i
o
us
t
y
pe
s
of
r
e
a
l
-
w
or
l
d
pr
ob
l
e
m
s
,
pa
r
t
i
c
ul
a
r
l
y
in
t
he
f
i
e
l
d
of
e
l
e
c
t
r
on
i
c
c
on
t
r
ol
.
A
f
un
da
m
e
nt
al
s
tr
uc
t
ur
e
m
od
e
l
f
or
A
N
N
s
a
r
e
c
om
pi
l
e
d
by
t
h
e
nu
m
be
r
o
f
l
a
y
e
r
s
t
ha
t
ha
ve
a
l
im
i
t
e
d
nu
m
be
r
of
Evaluation Warning : The document was created with Spire.PDF for Python.
Ind
on
esi
a
n
J
E
le
c Eng &
Co
m
p
Sci
IS
S
N:
25
02
-
4752
Ma
nage
men
t s
wi
tc
hin
g a
ng
le
s rea
l
-
ti
me pre
dicti
on
by arti
fi
ci
al… (
Moha
mm
e
d Ras
hee
d Ju
ba
ir
Al
-
Hiealy)
113
a
c
t
i
va
ti
on
f
u
nc
t
i
on
s
t
ha
t
c
a
l
l
ed
ne
ur
o
ns
;
t
he
y
a
r
e
i
nt
e
r
c
on
n
e
c
t
e
d
by
c
on
ne
c
t
i
ng
w
e
i
gh
t
s
o
t
ha
t
e
a
c
h
l
a
y
e
r
ha
s
a
bi
a
s
pa
r
am
e
te
r
,
i
n
w
hi
c
h
a
r
e
gu
l
a
t
or
y
va
l
ue
i
s
us
e
d,
t
o
c
on
t
r
ol
a
pr
ocessi
ng
m
et
ho
d
[
1
6]
A
N
N
s
ge
ne
r
a
l
l
y
ha
ve
t
hr
e
e
m
a
jo
r
l
a
ye
r
s
of
we
i
g
ht
c
on
ne
c
t
i
on
:
i
np
ut
l
a
y
e
r
t
o h
ol
d t
he
i
np
ut
pa
r
a
m
e
te
r
s
,
hi
dd
e
n
l
a
y
e
r
s
a
s
a
pr
oc
e
s
s
i
ng
l
a
y
e
r
a
nd
f
i
na
l
l
y
t
he
ou
t
pu
t
l
a
y
e
r
of
t
he
r
e
s
ul
t
f
r
om
i
np
ut
a
nd
p
r
oc
e
s
s
i
ng
l
a
y
e
r
s
[
17
]
-
[
19
]
.
Figure
4
.
A
NNs struc
ture
F
e
e
db
a
c
k
(
F
B
-
A
N
N
s
)
w
hi
c
h
c
on
t
a
i
n
f
e
e
db
a
c
k
a
nd
f
e
e
d
b
a
c
k
u
ni
t
(
F
F
-
A
N
N
s
)
a
n
d
w
hi
c
h
d
o
no
t
c
on
t
a
i
n
t
he
f
e
ed
b
a
c
k
u
ni
t
,
m
ay
al
s
o
be
di
vi
de
d
i
nt
o
t
w
o
m
ai
n
s
t
r
uc
t
ur
e
s
.
I
n
f
a
c
t
,
it
ha
s
be
e
n
e
s
t
a
bl
i
s
he
d
th
a
t
a
ll
A
N
N
s
ne
e
d
f
i
r
s
t
of
a
ll
t
o
t
r
a
in
da
t
a
t
o
un
de
r
s
t
a
nd
a
nd
t
o
f
or
m
a
li
nk
be
t
w
e
e
n
t
he
ou
t
p
ut
a
nd
i
np
ut
l
a
y
e
r
s
in
or
de
r
t
o
r
e
s
po
nd
m
or
e
qu
i
c
k
l
y
a
nd
t
em
po
r
a
r
i
l
y
[21]
.
T
he
r
e
i
s
di
f
f
e
r
e
nt
t
r
a
i
ni
ng
a
l
go
r
i
t
hm
s
us
e
d
i
n
A
N
N
s
.
Ho
w
e
ve
r
,
t
he
m
ost
a
pp
r
o
pr
i
a
t
e
a
l
go
r
i
t
hm
i
n
t
hi
s
w
or
k
i
s
L
e
ve
nb
e
r
g
-
M
a
r
q
ua
r
dt
-
a
l
go
r
i
t
hm
,
w
hi
c
h
i
s
r
e
qu
i
r
e
d
t
o
r
e
du
c
e
t
he
m
em
or
y
c
om
pa
r
ed
t
o
ot
he
r
m
e
t
ho
ds
be
c
a
us
e
o
f
f
a
s
t
e
r
t
a
nn
i
ng
[2
2]
.
”
R
e
s
ul
t
s
w
i
t
h
hi
gh
pe
r
f
or
m
a
nc
e
c
om
pa
r
e
d
t
o
ot
he
r
s
t
ud
i
e
s
t
ha
t
us
e
d
di
f
f
e
r
e
nt
c
on
t
r
ol
m
e
th
o
ds
w
e
r
e
a
c
hi
e
ve
d
by
us
i
ng
A
N
N
s
“
a
s
a
co
nt
r
o
l
s
y
s
t
em
t
o
c
al
cu
l
a
t
e
no
nl
i
ne
a
r
t
r
a
ns
c
e
nd
e
nt
a
l
e
qu
a
t
i
on
be
c
a
us
e
t
he
a
bi
li
ty
t
o
m
a
p
t
he
f
un
da
m
e
nt
a
l
r
e
l
a
ti
on
s
hi
p
be
t
w
e
e
n
i
np
ut
a
nd
o
ut
pu
t
da
t
a
du
r
i
ng
t
he
t
r
a
i
ni
ng
pr
oc
e
s
s
w
a
s
a
c
hi
e
ve
d
”
[23]
.
A
s
a
n
i
n
pu
t
ne
ur
on
,
t
he
f
i
r
s
t
ne
ur
on
of
t
he
ou
t
p
ut
l
a
y
e
r
i
s
us
e
d
t
o
ge
ne
r
a
t
e
a
s
w
i
t
c
hi
ng
s
i
gn
a
l
f
or
t
h
e
pr
o
po
s
e
d
M
-
C
H
B
M
I
[24]
.
F
or
ℎ
ne
ur
on
a
t
ℎ
a
y
e
r
t
he
ne
t
w
or
k
t
r
a
ns
f
e
r
f
u
nc
t
i
on
c
a
n
t
he
n
be
d
e
s
c
r
i
be
d
in
,
n
i
m
=
∑
W
ij
m
a
j
m
−
1
+
b
i
m
S
m
−
1
j
=
1
t
he
pa
r
a
m
e
te
r
of
bi
a
s
o
f
ℎ
l
a
y
er
a
t
ℎ
ne
ur
on
i
s
t
he
w
e
i
g
ht
r
e
l
a
t
i
on
s
pa
r
a
m
e
te
r
f
or
t
he
ℎ
a
n
d
ℎ
ne
ur
on
s
a
t
t
he
ℎ
l
a
y
e
r
a
nd
t
he
.
i
s
a
l
s
o
t
he
a
t
ℎ
l
a
y
e
r
ne
ur
on
ou
t
pu
t
f
un
c
t
i
on
t
h
a
t
c
a
n
be
de
f
i
n
e
d
a
s
:
a
i
m
=
f
m
(
n
i
m
)
I
n
t
he
hi
dd
e
n
l
a
y
e
r
,
“
a
ta
ng
e
nt
hy
pe
r
b
ol
i
c
f
un
c
t
i
on
i
s
s
e
l
e
ct
e
d
a
s
t
he
a
c
ti
va
t
i
on
f
un
c
t
i
on
(
f
)
f
or
o
ur
ne
t
w
or
k
de
s
i
gn
,
i
t
c
a
n
b
e
de
f
i
ne
d
a
s
,
f
m
(
n
i
m
)
=
2
1
+
e
−
2
n
i
m
−
1
w
he
r
e
a
s
f
i
s
un
i
t
y
a
t
t
he
ou
t
pu
t
l
ay
e
r
[25]
-
[
2
7]
.
T
he
a
bo
ve
de
s
c
r
i
be
s
a
ha
r
m
on
i
c
m
i
ni
m
i
za
t
i
on
e
qu
a
t
i
on
us
i
ng
A
N
N
s
.
F
i
gu
r
e
5
s
h
ow
s
t
he
A
N
N
ha
r
m
on
i
c
m
i
ni
m
iz
a
ti
on
pr
oc
e
du
r
e
f
l
o
w
c
ha
r
t
a
n
d
F
i
gu
r
e
6
pr
ov
i
d
e
s
t
he
r
e
s
ul
t
i
ng
op
t
i
m
a
l
s
w
i
t
c
hi
ng
a
ng
l
e
s
.
A
N
N
s
“
c
on
s
i
s
t
i
ng
of
t
w
o
hi
dd
e
n
ne
ur
o
n
s
ha
ve
be
e
n
t
r
a
i
ni
ng
t
he
o
ut
pu
t
a
r
e
s
ho
w
n
i
n
F
i
gu
r
e
7
(
a
)
a
n
d
F
i
gu
r
e
7
(
b
)
s
h
o
w
i
ng
t
he
g
r
a
d
i
e
nt
m
e
a
n
s
qu
a
r
e
d
e
r
r
o
r
a
n
d
A
N
N
”
pe
r
f
or
m
a
nc
e
va
l
i
da
t
i
on
of
9
-
l
e
ve
l
m
od
i
f
i
ed
C
H
B
-
M
L
I
t
h
a
t
ha
s
a
l
s
o
be
e
n
us
e
d
t
o
up
da
t
e
a
nd
c
or
r
e
c
t
bo
t
h
ne
t
w
o
r
k
de
vi
c
e
w
e
i
gh
t
a
n
d
bi
a
s
va
l
ue
s
,
va
l
i
da
t
i
on
i
s
us
e
d
vi
a
t
he
t
r
a
i
ni
ng
p
r
oc
e
s
s
t
o
t
r
a
c
k
t
he
e
r
r
or
.
Evaluation Warning : The document was created with Spire.PDF for Python.
IS
S
N
:
2502
-
4752
Ind
on
esi
a
n
J
E
le
c Eng &
Co
m
p
Sci,
Vo
l.
23
, N
o.
1
,
Ju
ly
2021
:
110
-
119
114
F
i
gu
r
e
5
.
A
N
N
f
l
ow
c
ha
r
t
f
o
r
ha
r
m
on
i
c
s
pr
o
bl
em
of
m
i
ni
m
i
z
a
ti
on
F
i
gu
r
e
6
.
C
a
l
c
ul
a
t
e
d
s
w
i
t
c
hi
ng
a
ng
l
e
s
θ
1
,
θ
2
,
θ
3
,
θ
4
Evaluation Warning : The document was created with Spire.PDF for Python.
Ind
on
esi
a
n
J
E
le
c Eng &
Co
m
p
Sci
IS
S
N:
25
02
-
4752
Ma
nage
men
t s
wi
tc
hin
g a
ng
le
s rea
l
-
ti
me pre
dicti
on
by arti
fi
ci
al… (
Moha
mm
e
d Ras
hee
d Ju
ba
ir
Al
-
Hiealy)
115
(
a
)
(b)
Figure
7
.
ANN Per
for
m
ance at trai
ning
with
2
n
eu
r
on
s
in hi
ding
4.
NIN
E
LE
VEL
C
HB
-
MLI
MATL
AB S
IM
ULABIO
N O
F MODI
FIED
T
he
pr
op
os
e
d
s
y
s
t
em
c
on
s
i
s
ts
of
o
ne
m
od
i
f
i
c
a
t
e
d
c
e
ll
c
on
ne
c
t
e
d
t
o
a
s
e
r
i
e
s
e
a
c
h
c
on
t
a
i
ni
ng
pu
l
s
e
ge
ne
r
a
t
i
ng
“
un
i
t
,
a
s
a
s
w
i
t
c
hi
ng
c
o
nt
r
ol
s
y
s
t
em
,
a
s
s
ho
w
n
i
n
F
i
gu
re
”
8
,
us
i
n
g
M
A
T
L
A
B
s
i
m
ul
a
ti
on
s
of
t
w
a
r
e
.
F
i
gu
r
e
8
.
M
a
t
l
ab
9
-
l
e
ve
l
C
H
B
-
M
L
I
s
im
ul
at
i
on
5.
SIMULATI
O
N
O
PTIMIZ
A
TION
RES
UL
TS FO
R
SI
N
GLE
-
PH
AS
E
MO
DIFIE
D CHB
-
MLIS
NIN
E
-
LE
VE
L
WITH
MI=
0
.
9
B
AS
E
D
O
N ANN TE
C
HNIQUES
”
I
n
or
de
r
t
o
p
r
o
du
c
e
4
i
ni
t
i
a
l
sw
i
t
c
hi
ng
a
ng
l
e
s
t
ha
t
a
r
e
us
e
d
t
o
c
on
t
r
ol
a
9
-
l
e
ve
l
m
od
i
f
i
e
d
C
H
B
-
M
L
I
,
a
n
a
r
t
i
f
i
c
i
al
ne
ur
a
l
ne
t
w
or
k
(
A
N
N
)
o
f
a
s
i
ng
l
e
-
ph
a
s
e
w
a
ve
f
o
r
m
s
y
st
em
ha
s
be
e
n
de
v
e
l
op
e
d
us
i
ng
“
M
A
T
L
A
B
/
s
of
t
w
a
r
e
c
od
e
.
F
i
ve
s
e
pa
r
a
te
D
C
i
np
ut
s
ou
r
c
e
s
,
e
a
c
h
e
qu
a
l
t
o
30
0
V
a
nd
25
0
0
H
z
s
w
i
t
c
hi
ng
Evaluation Warning : The document was created with Spire.PDF for Python.
IS
S
N
:
2502
-
4752
Ind
on
esi
a
n
J
E
le
c Eng &
Co
m
p
Sci,
Vo
l.
23
, N
o.
1
,
Ju
ly
2021
:
110
-
119
116
f
r
e
q
ue
nc
y
,
w
e
r
e
a
l
s
o
a
s
s
um
e
d.
T
hi
s
s
y
st
em
ge
ne
r
a
t
e
s
A
C
w
a
ve
f
or
m
s
on
t
he
l
oa
d
s
i
de
to
f
e
e
d
a
s
i
ng
l
e
-
ph
a
s
e
l
oa
d o
f
R
80
Ω
,
H 3
5e
-
3 L
a
n
d C
50
e
-
6m
F
.
T
he
i
ni
t
ia
l
4 sw
i
t
c
hi
ng
a
ng
l
e
s
us
i
ng
a
n A
N
N
a
pp
r
oa
c
h f
o
r
t
he
ti
m
e
f
r
a
m
e
of
M
i
=
0.
9
ha
ve
r
e
s
ul
t
e
d
θ
1
=
7.
8
71
,
θ
2
=
20
.
05
4,
θ
3=
4
4.
1
25
,
θ
4=
5
2.
8
31
θ
5=
52
.
83
1
a
nd
θ
6=
52
.
83
1
,
e
ve
n
t
he
t
im
e
f
r
am
e
f
or
e
a
c
h
l
oo
p
f
or
e
a
c
h
o
ut
pu
t
w
a
ve
f
or
m
i
s
eq
ua
l
t
o
0.
02
.
T
he
pr
e
di
c
t
i
on
d
a
t
a
a
nd
or
i
gi
na
l
da
ta
f
or
t
ot
a
l
ha
r
m
on
i
c
di
s
t
or
t
i
on
(
T
H
D
)
a
r
e
l
i
s
t
e
d
i
n
F
i
gu
r
e
9
d
ur
i
n
g
t
e
s
t
t
im
e
.
T
r
a
i
ni
ng
pe
r
f
or
m
a
nc
e
of
A
N
N
t
e
s
ti
ng
a
ng
l
e
s
s
e
t
a
r
e
i
ll
us
t
r
a
te
d
i
n
F
i
gu
r
e
1
0
t
ha
t
i
nd
i
c
a
ti
ng
e
r
r
o
r
r
e
s
u
l
t
a
nt
a
r
e
R
=
8
0.
A
9
-
l
e
ve
l
C
H
B
-
M
L
I
a
t
ha
ve
be
e
n
de
v
e
l
op
e
d
by
s
im
ul
a
t
i
on
m
od
e
in
w
hi
c
h
i
nc
l
ud
e
one
m
od
i
f
i
e
d
c
e
l
l
e
a
c
h
c
e
l
l
a
t
t
he
s
y
s
t
em
co
nt
a
i
n
s
t
w
o
u
ni
t
s
(
s
w
i
t
c
hi
ng
c
i
r
c
ui
t
s
c
he
m
e
un
i
t
a
nd
c
on
t
r
ol
un
i
t
)
e
a
c
h
c
e
l
l
c
on
s
i
s
t
in
g
4
s
w
i
t
c
hi
ng
de
vi
c
e
s
i
n
s
t
r
u
c
t
ur
e
s
c
he
m
e
w
i
t
h
on
e
pu
l
s
e
ge
ne
r
a
t
or
a
t
c
on
t
r
ol
un
i
t
f
or
e
a
c
h
o
ne
of
t
he
m
.
F
ig
ur
e
1
1
s
ho
w
s
t
he
t
im
i
ng
of
c
on
t
r
ol
pu
l
e
s
w
i
t
h
r
e
a
l
s
w
i
t
c
hi
ng
t
im
e
s
of
e
a
c
h
s
w
i
t
c
h
on
t
he
de
v
i
c
e
us
i
ng
A
N
N
op
t
i
m
i
z
at
i
on
a
ng
l
e
s
of
s
w
i
t
c
hi
ng
du
r
i
n
g
a
t
hr
e
e
-
c
e
l
l
c
e
ll
m
od
e
l
o
ne
1
,
2
,
3
4
.
T
he
r
e
s
ul
t
a
nt
t
ot
a
l
ha
r
m
on
i
c
di
s
t
or
t
i
on
(
T
H
D
)
t
ha
t
ha
v
e
be
e
n
pr
ov
i
de
u
s
i
ng
A
N
N
s
w
i
t
c
hi
ng
a
n
gl
e
s
op
t
im
i
z
a
ti
on
t
e
ch
ni
q
ue
t
o
c
on
t
r
ol
a
s
i
ng
l
e
-
ph
a
s
e
m
od
i
f
i
e
d
9
-
l
e
ve
l
C
H
B
-
M
L
I
a
r
e
r
e
du
c
e
d
a
s
c
o
m
pa
r
e
d
w
i
t
h
ot
he
r
op
t
im
i
z
at
io
n
t
e
c
hn
i
q
ue
s
l
i
ke
N
R
a
n
d
P
S
O
a
t
bo
t
h
of
c
ur
r
e
n
t
a
nd
vo
l
t
a
ge
pa
r
am
e
t
e
r
s
i
n
w
hi
c
h
a
v
ol
t
a
ge
T
H
D
ha
ve
be
e
n
r
e
a
c
he
d
i
nt
o
5%
a
nd
c
ur
r
e
nt
T
H
D
ha
ve
be
e
n
r
e
a
c
he
d
i
nt
o
1.
3%
a
s
s
h
ow
n
i
n
F
i
gu
r
e
1
2
-
1
5
.
“
F
i
gu
r
e
9
.
P
r
e
di
c
t
e
d
da
t
a
a
nd
o
r
i
gi
na
l
da
t
a
f
or
T
H
D
F
i
gu
r
e
1
0
.
T
r
a
i
ni
ng
pe
r
f
o
r
m
a
nc
e
a
ng
l
e
s
f
or
A
N
N
F
i
gu
r
e
11
.
T
i
m
e
s
w
i
t
c
hi
ng
f
o
r
1
,
2
,
3
,
4
,
5
6
0
0
.
0
0
2
0
.
0
0
4
0
.
0
0
6
0
.
0
0
8
0
.
0
1
0
.
0
1
2
0
.
0
1
4
0
.
0
1
6
0
.
0
1
8
0
.
0
2
0
0
.
5
1
T
i
m
e
S1
0
0
.
0
0
2
0
.
0
0
4
0
.
0
0
6
0
.
0
0
8
0
.
0
1
0
.
0
1
2
0
.
0
1
4
0
.
0
1
6
0
.
0
1
8
0
.
0
2
0
0
.
5
1
T
i
m
e
S2
0
0
.
0
0
2
0
.
0
0
4
0
.
0
0
6
0
.
0
0
8
0
.
0
1
0
.
0
1
2
0
.
0
1
4
0
.
0
1
6
0
.
0
1
8
0
.
0
2
0
0
.
5
1
T
i
m
e
S3
0
0
.
0
0
2
0
.
0
0
4
0
.
0
0
6
0
.
0
0
8
0
.
0
1
0
.
0
1
2
0
.
0
1
4
0
.
0
1
6
0
.
0
1
8
0
.
0
2
0
0
.
5
1
T
i
m
e
S4
0
0
.
0
0
2
0
.
0
0
4
0
.
0
0
6
0
.
0
0
8
0
.
0
1
0
.
0
1
2
0
.
0
1
4
0
.
0
1
6
0
.
0
1
8
0
.
0
2
0
0
.
5
1
T
i
m
e
S5
0
0
.
0
0
2
0
.
0
0
4
0
.
0
0
6
0
.
0
0
8
0
.
0
1
0
.
0
1
2
0
.
0
1
4
0
.
0
1
6
0
.
0
1
8
0
.
0
2
0
0
.
5
1
T
i
m
e
S6
0
0
.
0
0
2
0
.
0
0
4
0
.
0
0
6
0
.
0
0
8
0
.
0
1
0
.
0
1
2
0
.
0
1
4
0
.
0
1
6
0
.
0
1
8
0
.
0
2
0
0
.
5
1
T
i
m
e
S7
0
0
.
0
0
2
0
.
0
0
4
0
.
0
0
6
0
.
0
0
8
0
.
0
1
0
.
0
1
2
0
.
0
1
4
0
.
0
1
6
0
.
0
1
8
0
.
0
2
0
0
.
5
1
T
i
m
e
S8
0
0
.
0
0
2
0
.
0
0
4
0
.
0
0
6
0
.
0
0
8
0
.
0
1
0
.
0
1
2
0
.
0
1
4
0
.
0
1
6
0
.
0
1
8
0
.
0
2
0
0
.
5
1
T
i
m
e
S9
0
0
.
0
0
2
0
.
0
0
4
0
.
0
0
6
0
.
0
0
8
0
.
0
1
0
.
0
1
2
0
.
0
1
4
0
.
0
1
6
0
.
0
1
8
0
.
0
2
0
0
.
5
1
T
i
m
e
S
1
0
Evaluation Warning : The document was created with Spire.PDF for Python.
Ind
on
esi
a
n
J
E
le
c Eng &
Co
m
p
Sci
IS
S
N:
25
02
-
4752
Ma
nage
men
t s
wi
tc
hin
g a
ng
le
s rea
l
-
ti
me pre
dicti
on
by arti
fi
ci
al… (
Moha
mm
e
d Ras
hee
d Ju
ba
ir
Al
-
Hiealy)
117
T
he
pr
op
os
e
d
5
a
nd
9
-
l
e
ve
l
s
i
ng
l
e
-
ph
a
s
e
m
od
i
f
ie
d
C
H
B
-
M
L
I
w
a
s
pr
e
vi
ou
s
l
y
us
e
d
b
y
t
he
A
N
N
t
e
c
hn
i
qu
e
f
or
t
he
pu
r
p
os
e
of
r
e
du
c
i
n
g
t
he
T
H
D
on
t
he
v
ol
t
a
ge
a
nd
c
ur
r
e
nt
ou
t
p
ut
w
a
v
e
f
or
m
a
s
s
ho
w
n
i
n
F
i
gu
r
e
1
6
.
“
F
o
r
f
i
ve
-
l
e
ve
l
a
nd
ni
ne
-
l
e
ve
l
m
od
i
f
ie
d
C
H
B
-
M
L
I
”
,
T
a
b
l
e
2
“
di
s
pl
a
y
s
sw
i
t
c
hi
ng
a
ng
l
e
s
a
nd
di
f
f
e
r
e
nt
T
H
D
s
i
n
vo
l
t
a
ge
a
n
d
c
ur
r
e
nt
”
.
“
F
i
gu
r
e
1
2
.
O
ut
pu
t
v
ol
t
a
ge
w
a
ve
f
o
r
m
f
or
9
-
l
e
ve
l
C
H
B
-
M
L
I
s
im
ul
a
t
i
on
F
i
gu
r
e
1
3
.
O
ut
pu
t
ha
r
m
on
i
c
s
pe
c
t
r
um
f
or
vo
l
t
a
ge
w
a
v
e
f
or
m
us
i
ng
A
N
N
t
e
c
h
ni
q
ue
F
i
gu
r
e
1
4
.
O
ut
pu
t
c
ur
r
e
nt
w
a
ve
f
o
r
m
f
or
9
-
l
e
ve
l
C
H
B
-
M
L
I
s
im
ul
a
t
i
on
“
F
i
gu
r
e
1
5
.
O
ut
pu
t
ha
r
m
on
i
c
s
pe
c
t
r
um
f
or
c
u
r
r
e
nt
w
a
ve
f
or
m
us
i
ng
A
N
N
t
e
c
h
ni
q
ue
Table
2
.
C
om
par
iso
n
of
t
he 5
and 9 l
evels
of
the m
od
ifie
d
C
HB
-
MLI
b
a
s
e
d
on the
AN
N p
rop
os
al
N
-
le
v
e
l
MI
1
2
3
4
T
HD
V
T
HD
A
5
-
l
e
v
e
l
0
.
9
3
1
.
4
1
4
1
.
9
3
0
0
1
4
.
4
8
%
8
.
0
5
%
9
-
l
e
v
e
l
0
.
9
7
.
8
7
1
2
0
.
0
5
4
4
.
1
2
5
2
.
8
3
7
.
4
6
%
5
.
1
%
0
0
.
0
0
2
0
.
0
0
4
0
.
0
0
6
0
.
0
0
8
0
.
0
1
0
.
0
1
2
0
.
0
1
4
0
.
0
1
6
0
.
0
1
8
0
.
0
2
-
6
0
0
-
4
0
0
-
2
0
0
0
200
400
600
T
i
m
e
P
h
a
s
e
V
o
l
t
a
g
e
(
V
)
Evaluation Warning : The document was created with Spire.PDF for Python.
IS
S
N
:
2502
-
4752
Ind
on
esi
a
n
J
E
le
c Eng &
Co
m
p
Sci,
Vo
l.
23
, N
o.
1
,
Ju
ly
2021
:
110
-
119
118
F
i
gu
r
e
16
.
S
t
a
t
is
t
i
c
s
g
r
a
ph
f
or
s
w
i
t
c
hi
ng
a
ng
l
e
va
l
ue
s
a
nd
t
h
e
T
H
D
va
l
ue
s
of
vo
l
t
a
ge
a
nd
c
ur
r
e
nt
ba
s
e
d
o
n
A
N
N
f
o
r
9
-
l
e
v
e
l
s
M
-
C
H
B
-
ML
I
6.
CONCL
US
I
O
NS
T
he
r
e
s
e
a
r
c
h
p
a
pe
r
pr
op
os
e
s
A
N
N
s
a
s
t
he
ni
ne
-
l
e
ve
l
C
H
B
-
M
L
I
op
t
im
iz
a
t
i
on
m
e
t
ho
d
t
ha
t
r
e
du
c
e
s
T
H
D
'
s
c
ur
r
e
nt
w
a
ve
f
or
m
a
nd
ou
t
pu
t
vo
l
t
a
ge
.
T
he
a
im
of
t
hi
s
pa
pe
r
w
a
s
s
i
m
ul
a
te
d
w
i
t
h M
A
T
L
A
B
S
im
ul
i
nk
t
o
a
na
l
y
s
e
t
he
r
e
s
ul
t
i
ng
pa
r
a
m
et
e
r
s
.
T
he
r
e
s
ul
t
a
nt
T
H
D
ou
t
p
ut
w
a
ve
f
or
m
of
9
-
l
e
ve
l
ha
s
be
e
n
c
om
pa
r
e
d
w
i
t
h
a
r
e
s
ul
t
a
nt
T
H
D
ou
t
p
ut
w
a
ve
f
or
m
of
pr
e
vi
ou
s
w
or
k
of
r
e
s
e
a
r
c
h
pa
pe
r
s
f
or
5
-
l
e
ve
l
C
H
B
-
M
L
I
.
I
n
br
i
e
f
l
y
d
e
s
c
r
i
be
of
s
i
m
ul
at
i
on
r
e
s
ul
t
,
t
he
T
H
D
a
t
o
ut
pu
t
w
a
ve
f
or
m
a
r
e
r
e
du
c
e
d
w
he
n
t
he
l
e
ve
l
s
t
e
pp
e
d
o
f
C
H
B
-
M
L
I
a
r
e
i
nc
r
e
a
s
e
d
,
a
l
s
o
A
N
N
s
t
e
c
hn
i
q
ue
a
us
e
f
ul
op
t
im
i
z
a
ti
on
t
oo
l
th
a
t
pr
o
vi
de
l
ow
e
r
T
H
D
a
s
c
o
m
pa
r
e
d
w
i
t
h
NR
a
nd
P
S
O
o
pt
im
i
za
t
io
n
t
e
c
hn
i
q
ue
s
.
REFERE
NCE
S
[1]
M.
Rashee
d
,
M
.
M.
A.
Alakk
ad
,
R.
Om
ar,
M
.
Sulai
m
an,
and
W
.
A.
Ha
li
m
,
“
Enha
nc
e
th
e
accura
c
y
of
cont
r
ol
al
gorit
hm
for
m
ult
il
evel
inv
er
te
r
base
d
on
ar
ti
ficia
l
neur
a
l
n
et
work,”
I
ndone
s.
J.
El
e
ct
r.
En
g.
Comput.
Sc
i.
(
IJE
ECS)
,
vol. 2
0,
no
.
3
,
2020
,
d
oi:
10
.
11591/ije
e
cs.
v20.
i3
.
pp114
8
-
1158.
[2]
R.
Om
ar,
M.
R
ashe
ed,
M
.
Sula
iman,
and
A.
A
l
-
ja
n
ad,
“
A
Stud
y
of
a
Thr
ee
Phase
Diode
Cl
amped
Multi
l
ev
el
Inve
rte
r
Perform
anc
e
For Harmonics
Redu
ct
ion
,
”
MAGNT R
es
.
R
ep.
(
ISSN.
1
444
-
8939)
,
vol.
2
,
no
.
4
,
pp
.
62
-
71
.
[3]
P.
Chowdhur
y
,
I
.
Kole
y
,
and
S.
S
en,
“
Modelling,
sim
ula
ti
on
and
c
ontrol
of
a
gr
id
c
onnec
t
ed
non
co
nvent
ion
al
sol
ar
power
gen
era
t
io
n
s
y
st
em using
,
”
vol. 2, no. 4, pp. 1183
-
1191,
201
3.
[4]
K.
Gos
wam
i,
B.
Pande
y
,
D.
M.
A.
Hus
saia
n,
T.
K
um
ar,
and
K.
Kali
a
,
“
Input
/
output
Buffe
r
base
d
Vedic
Multi
pl
ie
r
Design
for
The
r
m
al
Aw
are
En
er
g
y
Eff
icien
t
Dig
it
al
Signal
Proce
ss
ing
on
28nm
F
PG
A,”
Indian
Jo
urnal
of
Sc
ie
nc
e
and
Technol
og
y
,
vol. 9, no. March 2016, p
p.
1
-
8,
doi:
10
.
17485/ijst/2016/
v9i10
/88
072.
[5]
G.
Vijay
krishna
and
O.
C.
Shekhar
,
“
A
Thre
e
Phase
7
-
Le
v
el
and
9
-
Le
ve
l
Reve
rsi
ng
Volta
ge
Mult
il
ev
el
Inve
rt
er,”
Indian
Journal
o
f
Sc
ie
nc
e
and
Te
chnol
ogy
,
vo
l. 8, no. Sept
ember
2
015,
pp
.
1
–
10
,
d
oi:
10
.
17485/ijst
/2
015/v8i
.
[6]
M.
M.
A.
Al
akk
ad,
Z.
R
asin,
M.
Rashee
d,
W
.
A.
Hali
m
,
and
R.
O
m
ar,
“
Rea
l
-
ti
m
e
sw
it
chi
ng
th
irt
e
e
n
-
le
ve
l
m
odifi
ed
CHB
-
m
ult
il
eve
l
inve
rte
r
using
art
ificial
neur
a
l
net
work
te
chn
ique
base
d
on
sele
c
ti
ve
har
m
o
nic
el
imin
at
ion
,
”
Indone
s. J.
Elec
t
r.
Eng
.
Co
mput.
Sci
.
(
IJE
ECS)
,
v
ol.
20
,
no
.
3
,
202
0,
doi
:
10
.
11591
/i
jeec
s.v20
.
i3
.
pp
1642
-
1652.
[7]
R.
Om
ar,
M
.
Ra
shee
d,
Z.
K.
Lo
w,
and
M.
Sul
aim
an,
“
Design
an
d
developm
ent
o
f
active
power
fi
lt
er
for
h
armonic
m
ini
m
iz
at
ion
usi
ng
s
y
n
chr
onous
ref
ere
n
ce fra
m
e
(SRF
),
”
ARPN
J
.
Eng
.
Appl.
S
ci
.
,
vol. 14, no. 2, 2
019.
[8]
J.
S.
Krism
adi
na
ta
a
,
Nasrudin
A
bd.
Rahi
m
a
He
w
W
ooi
Pinga,
“
The
3
rd
Inte
rn
at
ion
al
Confer
en
ce
on
Sus
ta
in
able
Future
for
Hum
a
n
Secur
ity
Photovolt
aic
m
odule
m
odel
ing
using
sim
uli
nk/
m
at
la
b
,
”
Proce
dia
En
vir
on.
Sci
.
,
vol.
17
,
no.
3
,
2013
,
pp
.
537
-
546,
doi
:
10
.
1016/j.proe
nv
.
2
013.
02.
069
.
[9]
N.
Nordin,
S
y
uh
ada
,
R.
Om
ar,
M.
Rashee
d
,
A.
Sabar
i,
and
K.
Krism
adi
nat
a,
“
Harm
onic
m
inim
iz
at
ion
of
a
th
ree
phase
c
asc
ad
ed hbridge
m
ultile
v
el
inve
rt
ers,
”
in
I
ET
Confe
ren
ce
Publ
ic
a
ti
ons
,
20
16
,
no
.
CP
688
,
2
016
.
[10]
H.
F.
Hashim
,
R.
Om
ar,
and
M.
Rashee
d,
“
Design
and
anal
y
s
is
of
a
thre
e
ph
a
se
serie
s
active
power
fil
t
er
(sap
f)
base
d
on
h
y
st
eres
is c
ontroller
,
” i
n
IET
Conf
ere
nc
e
Pub
li
ca
ti
ons
,
2
016,
no
.
CP
688
,
2016
.
[11]
M.
Rashee
d
,
R
.
Om
ar,
and
M.
Sulai
m
an
,
“
Com
p
ara
t
ive
p
erf
orm
a
nce
of
m
ult
ileve
l
inv
ert
er
for
h
ar
m
onic
red
uc
ti
o
n
base
d
on
N
ewto
n
rap
hson,
”
in
I
ET
Confe
ren
ce
Publ
ic
a
ti
ons
,
20
16,
no
.
CP
688
,
2
016
.
Evaluation Warning : The document was created with Spire.PDF for Python.
Ind
on
esi
a
n
J
E
le
c Eng &
Co
m
p
Sci
IS
S
N:
25
02
-
4752
Ma
nage
men
t s
wi
tc
hin
g a
ng
le
s rea
l
-
ti
me pre
dicti
on
by arti
fi
ci
al… (
Moha
mm
e
d Ras
hee
d Ju
ba
ir
Al
-
Hiealy)
119
[12]
J.
Rodrígue
z
,
St
eff
en
B
ern
et
,
Bi
n
W
u
,
Jorge
O.
Pontt
,
and
Sam
ir
Kouro
,
“
Multi
le
ve
l
Volta
g
e
-
S
ourc
e
-
Conver
t
er
Topol
ogie
s
for
Industria
l
Medi
um
-
Volta
ge
Drive
s,”
IEEE
Tr
a
ns.
Ind.
El
ectron.
,
vol.
54,
no
.
6,
pp.
2930
-
2945
,
2007
,
doi
:
10
.
11
09/T
IE
.
2007.
90
7044
.
[13]
S.
S.
Patna
ik
an
d
A.
K.
Panda,
“
Thre
e
-
le
v
el
H
-
bridge
and
thr
ee
H
-
bridge
s
-
base
d
t
hre
e
-
phase
four
-
wire
shunt
ac
t
iv
e
power
fil
ter
topo
logi
es
for
high
v
olt
ag
e
appl
i
catio
ns,”
Int.
J.
E
le
c
t
r.
Powe
r
Ene
rgy
Syst.
,
vol
.
5
1,
p
p.
298
-
306,
Oct
.
2013,
doi
:
10
.
10
16/j
.
i
je
p
es.
2013.
02.
037.
[14]
L.
M.
Tol
b
ert,
S.
Mem
ber
,
and
F.
Z.
Peng
,
“
Multi
l
eve
l
PW
M
M
et
hods
at
Low
Modulat
ion
Ind
i
ce
s,”
IEEE
Tr
ans.
POWE
R El
e
ct
ro
n.
,
vol. 15, no. 4, pp. 719
-
725,
20
00
,
doi
:
10
.
1109
/63.
849042
.
[15]
H.
H.
W
ang,
S.
Mem
ber
,
A.
M.
Kham
badkone
,
and
S.
Mem
ber
,
“
Anal
y
t
ic
a
l
Pow
er
Loss
Eva
luation
of
5
l
eve
l
H
-
Bridge
with
Co
uple
d
Induc
tor
and
Serie
s
Con
nec
t
ed
H
-
Bridg
e
for
PEBB
Applicati
ons,
”
Pow
e
r
El
ectron.
Dr
ive
Syst.
2009
.
PE
D
S
2009
,
vol. 1, p
p.
458
-
463
,
200
9
,
doi
:
10
.
1109/
PEDS
.
2009.
5385922
.
[16]
M.
S.
Rosli
Om
ar,
Moham
m
ed
Rashee
d,
“
Fundam
ent
al
Studie
s
of
a
Thr
ee
Pha
se
Casca
d
ed
H
-
Bridge
and
Dio
de
Cla
m
ped
Mult
il
e
vel
Inv
erters Using
Matlab/Sim
ul
ink,
”
Int
.
Re
v
.
A
utom.
Control
,
v
ol.
6
,
no
.
5
,
2013
.
[17]
A.
Parka
sh
and
S.
L.
S.
Ch
at
t
erj
i
,
“
Harm
onic
s
Eli
m
ina
ti
on
in
Cas
ca
de
Mul
tilevel
Inve
rte
rs
Us
ing
Newton
-
Raphson
and
Gene
tic
Alg
orit
hm
,
”
Int. J.
S
ci
.
R
es.
De
v.
,
vo
l
.
2
,
no
.
5
,
pp
.
23
6
-
239,
2014
.
[18]
M.
Sabahi,
A.
R
.
M.
Ir
ana
q
,
K.
M.
Bahra
m
i,
K.
M.
Bahra
m
i,
an
d
M.
B.
B
.
Shari
fia
n,
“
Harm
onic
s
el
imination
in
a
m
ult
i
le
ve
l
inve
rt
er
with
unequa
l
DC
source
s
u
sing
gene
ti
c
a
lgori
t
hm
,
”
2011
Int.
Conf.
El
e
ct
r.
M
ach.
Syst.
ICEM
S
2011
,
2011
,
doi
: 10.1109/
ICEMS.2011.
6073451
.
[19]
A.
Parka
sh,
S.
L.
Shim
i,
and
S.
Chat
terji
,
“
Harm
onic
s
Reduc
t
ion
in
Casca
de
H
-
Bridge
Multi
l
evel
Inve
rte
r
s
Us
ing
GA
and
PS
O,”
v
ol.
12
,
no
.
9
,
pp
.
453
–
465,
2014
,
d
oi:
10.
14445
/22
315381/IJET
T
-
V12P
287
.
[20]
C.
Y.
H’ng,
B.
Ism
ai
l,
M.
Isa,
an
d
M.
N.
K.
H.
R
ohani
,
“
Selecti
v
e
har
m
onic
el
imi
nat
ion
pulse
wid
th
m
odula
ti
on
for
five
-
l
eve
l
c
asc
ad
ed
inv
ert
e
r,
”
J
.
T
el
e
commun.
E
l
e
ct
ron.
Comput
.
Eng.
,
vol
.
10
,
no
.
1
-
14
,
pp
.
67
-
71
,
2018
.
[21]
P.
G.,
“
Optimiz
at
ion
Te
chn
iques
for
Harm
onic
s
Minim
iz
at
ion
in
Casc
ade
d
H
y
brid
Mul
ti
l
ev
el
Conver
te
rs:
a
Revi
ew,”
In
t. J. Re
s. Eng.
Techn
ol.
,
vol
.
04
,
no
.
0
3,
pp
.
188
-
193
,
2015,
doi
:
10
.
15
623/i
jret.
2
015
.
0
403033.
[22]
R.
Taleb
and
A.
Meroufe
l
,
“
Control
of
as
y
m
m
etr
ic
a
l
m
ult
ilevel
i
nver
te
r
using
artific
i
al
n
eur
a
l
netw
ork,
”
Elek
tron.
ir E
l
ek
trot
ec
hni
k
a
,
no
.
8
,
pp
.
93
-
98,
2009
,
doi
:
10
.
5755/j
01
.
e
ee
.
96
.
8.
9970
.
[23]
R.
T
al
eb
,
A.
M
ero
ufe
l
,
and
P.
W
ira
,
“
Harm
onic
e
li
m
ina
t
ion
co
ntrol
of
an
inverter
b
ase
d
on
an
artificia
l
neur
al
net
work strateg
y,
”
I
FA
C
Proc
.
,
v
ol.
2
,
no
.
p
art
1,
pp.
137
-
156
,
20
09,
doi
:
10
.
3182
/20090921
-
3
-
TR
-
3005.
00007.
[24]
M.
Mahe
sh,
K
Krant
hi,
and
P.
Singh,
“
Artifi
ci
al
Neura
l
Net
work
Based
Closed
Loop
Cont
rol
of
Multi
l
ev
el
Inve
rte
r
,
”
Int
.
J.
Mod.
Tr
ends
Sc
i
.
Techno
l.
,
no
.
2
,
2016.
[25]
W
.
A.
H.
and
M.
M.
A.
Moham
m
ad
Rashe
ed,
Rosli
Om
ar,
Mar
iza
n
Sulai
m
an,
“
Artifi
c
ial
Int
el
l
ige
n
ce
Te
chn
ique
to
Rea
l
-
Ti
m
e
Base
d
on
Sele
ct
ive
H
armonic
El
imina
ti
on
in
Modifie
d
Multi
le
v
e
l
Inve
rte
r,”
J.
Eng
.
Ap
pl.
Sci.
,
vol
.
14,
no.
24
,
p
.
9
,
201
9.
[26]
H.
A.
Moham
ed
and
H
.
M.
D.
Habbi,
“
Pow
er
qual
ity
of
du
al
t
wo
-
le
vel
inverte
r
fed
op
en
end
winding
indu
ct
i
on
m
otor,
”
Indone
s.
J.
Elec
tr
.
Eng.
Comput.
Sci
.
(
IJE
ECS
)
,
vol.
18,
no
.
2,
pp.
688
-
697,
2020,
d
oi:
10.
11591/ijeecs.
v18.
i2.
pp688
-
69
7.
[27]
M.
Rashee
d,
R.
Om
ar,
M.
Sul
ai
m
an,
W
.
A.
Hali
m
,
and
M.
M.
A.
Alakka
d,
“
Anal
y
sis
of
a
sw
it
chi
ng
ang
l
e
ca
l
cul
a
ti
on
b
y
A
NN
for
nine
le
v
e
l
inve
r
te
r
app
l
y
i
nto
expe
r
imenta
l
ca
se
stud
y
with
el
imination
of
lo
wer
and
hi
ghe
r
orde
r
har
m
onic
s
,
”
Indone
s.
J.
El
e
ct
r.
Eng
.
Co
mput.
Sci.
(
IJE
ECS)
,
vol.
20,
no.
2,
pp.
948
-
959,
2020,
doi:
10.
11591/ijeecs.
v20.
i2.
pp948
-
95
9.
[28]
Moham
ad
Ja
y
a
,
A.
S,
Jarra
h,
M,
and
Muham
ad,
M.
R
,
"
Modeli
n
g
of
Ti
N
coa
ti
ng
gra
in
size
us
ing
RS
M
appr
oac
h,
"
In
Appl
i
ed
M
e
chani
cs
and
M
ate
rials
,
vo
l.
7
54,
pp.
738
-
74
2,
Tr
ans
Tech
Public
a
ti
ons
Lt
d
,
2015
,
do
i
:
10.
4028/www
.
scie
nti
f
ic.ne
t
/AMM
.
754
-
755.
738
.
[29]
Jarra
h,
M.
I
,
Ja
ya,
A.
S.
M
,
Alqa
tt
an
,
Z.
N
,
Az
am,
M.
A,
Abdullah,
R
,
Jarra
h
,
H,
and
Abu
-
Khadrah,
A.
I
,
"
A
novel
ex
pla
n
at
or
y
h
ybrid
artifi
ci
a
l
b
ee
col
on
y
al
go
rit
hm
for
n
um
eri
c
al
fun
ct
ion
opti
m
iz
a
ti
on
.
,
"
The
Journal
of
Superc
omputing
,
pp.
1
-
25
,
2020
,
doi
:
10
.
1007/s11
227
-
019
-
03083
-
2
.
[30]
Jarra
h,
M.
I.
M
,
Ja
y
a
,
A.
S.
M
,
A
za
m
,
M.
A,
Alqat
ta
n
,
Z.
N
,
Muham
ad,
M.
R,
and
Abdulla
h,
R
,
"
Applicati
on
of
Ba
t
Algorit
hm
in
Carbon
Nanotubes
Grow
ing
P
ro
ce
ss
Para
m
et
ers
Optimiza
ti
on
,
"
In
Inte
lligent
and
Inte
ractive
Computing
,
pp
.
179
-
192,
Spring
er,
Sing
apor
e
,
2
019
,
d
o
i
:
1
0
.
1
0
0
7
/
9
7
8
-
9
8
1
-
13
-
6
0
3
1
-
2
_
1
4
.
[31]
Jarra
h,
M.
A.
I
,
Ja
y
a
,
A.
S.
M
,
Aza
m
,
M.
A
,
Alsharif
,
M.
H,
a
nd
Muham
ad,
M.
R
,
"
Inte
llige
nce
Inte
gr
ation
Of
Parti
cle
Sw
arm
Optimiza
ti
o
n
And
Phy
si
ca
l
Vap
our
Depositi
on
For
Ti
n
Grain
S
iz
e
Coa
ti
ng
Proce
ss
Para
m
et
ers,"
Journal
of
Theoretical
&
Appl
ie
d
Information
Te
c
hnology
,
vol
.
84
,
no.
3,
2016
.
[32]
Jar
rah
,
M.
I
,
Ja
ya,
A.
S.
M
,
Aza
m
,
M.
A
,
Muham
ad,
M.
R,
and
Za
in
,
A.
M,
"
Prediction
of
Grai
n
Size
in
the
T
i
N
Coat
ing
U
sing
Artifi
c
ia
l
Neur
al
Network,
"
Inte
r
nati
onal
Journal
of
Applied
Eng
i
nee
ring
R
ese
arc
h
,
vol.
11
,
no
.
1
9
,
pp.
9856
-
9869
,
2016
.
[33]
Fauzi
,
N.
F,
Abdul
S
y
ukor
Moha
m
ad
Ja
y
a
,
M.
I.
Jarra
h
,
Hus
sain
Sali
h
Akbar
,
"T
hin
fil
m
roughne
ss
opti
m
iz
at
ion
i
n
the
Ti
N
coa
t
ings u
sing g
en
et
i
c algorithms
,
"
J
The
or A
ppl
In
f
Te
ch
nol
,
vo
l.
95
,
no
.
2
4,
pp
.
6690
-
66
98
,
2017
.
Evaluation Warning : The document was created with Spire.PDF for Python.