Int
ern
at
i
onal
Journ
al of
P
ower E
le
ctr
on
i
cs a
n
d
Drive
S
ystem
(I
J
PE
D
S
)
Vo
l.
11
,
No.
3
,
Septem
be
r
2020
, pp.
1499
~
1507
IS
S
N:
20
88
-
8694
,
DOI: 10
.11
591/
ij
peds
.
v
1
1
.i
3
.
pp
1499
-
1507
1499
Journ
al h
om
e
page
:
http:
//
ij
pe
ds
.i
aescore.c
om
A maxi
mum po
wer poin
t tracki
ng base
d on lev
y flight
optimiz
ation
C
.
Charin
1
, D
ahaman Is
hak
2
, Muhamm
ad Ammirr
ul
A
t
iqi
M
ohd Z
ai
nuri
3
1,2
School
of El
e
ct
ri
ca
l
and
Elec
t
ronic
s E
ng
ineeri
ng,
Univer
si
ti Sa
ins Ma
la
ysi
a, Ma
la
ysi
a
1
Facul
ty
of Engi
nee
ring
T
ec
hnol
ogy,
Univer
si
ti
Mala
ysia
Perli
s
,
Mala
ysia
3
Facul
ty
of Engi
nee
ring
and
Buil
t
Env
ironm
en
t,
Univer
siti
Keba
ngsaa
n
Mal
aysia,
Mal
aysia
Art
ic
le
In
f
o
ABSTR
A
CT
Art
ic
le
history:
Re
cei
ved
Feb
4
, 2
0
20
Re
vised
M
a
r
6
,
20
20
Accepte
d
A
pr
15
, 20
20
Thi
s
p
ape
r
prop
oses
a
Le
vy
f
li
g
ht
glob
al
m
axi
m
um
power
poin
t
tra
ck
ing
for
solar
photovo
ltaic
(PV
)
sys
tem
under
p
art
i
a
l
shading
cond
it
ions.
Th
e
proposed
m
et
ho
d
comes
with
me
rit
s
such
as
simpl
icity
,
f
ast
response
and
fre
e
of
osci
ll
a
ti
o
n.
Thi
s
al
go
rit
h
m
uses
ran
do
m
sea
rch
over
the
expl
ora
ti
on
spac
e
and
co
m
par
es
the
pre
v
i
ous
and
cur
r
ent
sta
te
s
to
obtai
n
th
e
b
est
soluti
on.
For
e
val
ua
ti
on
and
com
par
at
iv
e
an
al
ysis,
p
erf
orm
anc
e
of
the
proposed
me
tho
d
is
al
so
measured
against
Pertu
rb
and
Obs
erv
e
(P&O)
and
Parti
cle
Sw
arm
Optim
izati
o
n
(P
SO
).
All
three
al
gorit
h
ms
are
simul
a
te
d
in
MA
TL
AB/S
im
uli
nk
envi
ronm
en
t.
Si
mul
a
ti
on
res
ult
s
ar
e
sati
sfa
ctory
over
th
e
conduc
t
ed
te
sts
under
un
iform
and
non
-
un
ifor
m
irra
di
ance.
T
he
proposed
al
gorit
h
m
is
ab
l
e
to
tr
ac
k
global
ma
xim
u
m
power
point
(GM
PP
)
under
par
ti
a
l
shading
cond
it
io
ns wit
h
fast
tr
acking
time
and
z
e
ro
ripple
at
st
ea
d
y
-
stat
e
.
Ke
yw
or
d
s
:
Global
ma
xim
um
powe
r p
oint
Lev
y
flig
ht
opti
miza
ti
on
M
a
ximum
pow
er
po
i
nt trac
king
Partia
l sha
ding
cond
it
io
n
This
is an
open
acc
ess arti
cl
e
un
der
the
CC
BY
-
SA
l
ic
ense
.
Corres
pond
in
g
Aut
h
or
:
Dah
a
man
I
s
ha
k,
School
of Elec
tric
al
an
d El
ect
ronic E
ng
i
neeri
ng
,
Un
i
ver
sit
i Sai
ns
M
al
a
ys
ia
, E
nginee
rin
g
Ca
m
pu
s
,
14300 Ni
bong
Tebal,
Pena
ng,
M
al
a
ys
ia
.
Emai
l:
d
aha
ma
n@usm
.m
y
1.
INTROD
U
CTION
Re
new
a
ble
e
ne
rgy
is
see
n
as
a
pro
misi
ng
al
te
rn
at
ive
in
ge
ne
rati
ng
el
ect
rical
powe
r
[1]
.
S
olar
is
one
of
the
po
te
ntial
can
did
at
es
in
powe
r
ge
ne
rati
on
du
e
to
avail
a
bili
ty
of
t
he
s
un
[
2]
and
cl
ea
nlines
s
[3]
.
M
ore
ov
e
r,
the
ope
rati
on
al
a
nd
mainte
na
nc
e
c
os
ts
of
s
ol
ar
phot
ovoltai
c
(
PV)
are
lo
w
c
ompa
red
t
o
ot
he
r
ren
e
wa
ble
ene
rgy
re
source
s
[
4]
.
This
e
mer
gi
ng
te
ch
nolo
gy
has
gaine
d
a
lot
of
interest
s
from
researc
he
rs
in
the
a
rea
of
gr
i
d
i
nteg
rati
on,
con
t
ro
l
an
d
op
ti
miza
ti
on
[5]
.
H
oweve
r,
thi
s
te
ch
nolo
gy
i
s
sti
ll
li
mit
ed
to
it
s
eff
i
ci
enc
y
[
6]
. E
ff
ic
ie
nc
y of
P
V
syst
em is sti
ll
low
, less tha
n
22.
4%
[3,7]
. Ad
diti
on
al
l
y,
non
-
li
near
pa
ra
mete
rs
su
c
h
as
te
mpe
ratur
e
a
nd
ir
ra
diance
af
fect
c
har
act
erist
ic
s
of
c
urren
t
an
d
vo
lt
age
c
urves
of
t
he
s
olar
P
V
[1]
.
Th
us
,
ma
ximum
power
po
i
nt
trackin
g
(
MPPT)
c
on
tr
ollers
a
re
nee
ded
for
harnessi
ng
the
opti
mum
powe
r
from
P
V
mod
ul
es
[8]
.
Var
i
ous
M
P
PT
met
hods
hav
e
bee
n
pro
posed
t
o
ob
t
ai
n
opti
mum
powe
r
f
r
om
s
ola
r
P
V.
Fo
r
insta
nce,
two
M
PPT
met
hods
are
po
pu
l
arly
us
e
d
s
uc
h
as
Pe
rtu
rb
at
io
n
a
nd
O
bs
e
rva
ti
on
(P
&
O)
[9]
a
nd
In
c
reme
ntal
C
onduct
ance
(
IncC
ond)
[10,1
1]
.
The
se
meth
ods
w
ork
well
unde
r
unif
orm
i
rr
a
diance.
Howev
e
r,
they
faced
s
ome
co
ns
trai
nts
unde
r
no
n
-
uniform
irra
dia
nc
es
across
P
V
modu
le
s
[12
]
.
These
c
onve
ntion
al
M
PP
T
al
gorith
ms f
ai
l
to
d
ist
inguis
h
i
n
betw
een g
l
ob
al
an
d l
ocal
ma
xima,
causim
g
t
he
syst
em
to
b
e
trap
ped
a
t
local
ma
xima
powe
r po
i
nt
[
13]
.
Pr
act
ic
al
ly,
c
onnecte
d
PV
a
r
ray
s
ca
n
be
a
f
fected
by
en
vi
ronme
ntal
c
ha
ng
e
s
s
uc
h
as
c
loud,
le
aves
and
dust
[
14,15]
.
T
his
ph
e
no
men
on
is
ine
vi
ta
b
le
an
d
unpredict
able
w
hich
ca
us
e
pa
rtia
l
sh
a
ding
c
onditi
on
(P
SC)
to
occur
[16]
.
S
of
t
c
omp
uting
meth
ods
are
c
onside
r
ed
as
the
most
po
te
ntial
cand
i
dates
to
mit
iga
te
the
Evaluation Warning : The document was created with Spire.PDF for Python.
IS
S
N
:
2088
-
8
694
In
t J
P
ow
Ele
c
&
Dr
i
S
ys
t
,
V
ol
.
1
1
, N
o.
3
,
Se
ptembe
r
2020
:
14
99
–
15
07
1500
con
ce
r
n
of
no
n
-
unifo
rm
irra
diance
[
17]
.
N
ow
a
da
ys
,
soft
com
pu
ti
ng
me
thods
s
uc
h
as
arti
fici
al
intel
l
igent
netw
ork
(
ANN
)
[15,1
6]
,
Fu
zz
y
Lo
gic
Co
ntr
ol
(F
LC
)
[16,1
8,19]
an
d
Pa
rtic
le
S
war
m
O
ptimi
zat
ion
(P
S
O)
ar
e
mo
st
prefe
ra
ble
am
ong
resear
cher
s
[22]
.
PS
O
al
gorithm
is
on
e
of
th
e
mo
s
t
le
adin
g
M
PP
T
te
ch
niques
use
d
t
o
detect
global
maxim
um
pow
er
point
(
G
M
P
P)
under
PSC
[2]
.
T
his
met
hod
c
om
e
s
with
gr
eat
a
dv
a
nta
ges
du
e
to
it
s
go
od
pe
rformance
,
simpli
ci
ty,
eas
y
to
im
plement
a
nd
le
ss
co
mp
le
x
mathe
mati
cal
struct
ur
e
[16]
.
Conve
ntion
al
PSO
al
gorith
m
w
orks
well
unde
r
P
SC.
Howe
ver,
due
to
la
r
ge
sea
rc
hing
area
,
this
meth
od
su
f
fer
s
lo
ng
e
r
com
pu
ta
ti
onal
ti
me
[16]
an
d
ste
ady
-
sta
te
osc
il
la
ti
on
.
I
n
a
ddit
ion
,
this
al
gorith
m
al
so
suffe
rs
from
inter
mit
te
nt
trac
king
sin
ce
it
s
requires
an
a
ppr
opriat
e
init
i
al
value
to
be
a
ble
to
dete
ct
certai
n
c
hanges
o
f
irrad
ia
nce
c
onditi
on
[23]
.
T
o
so
lve
t
he
af
orementi
one
d
pr
ob
le
m
s,
a
Le
vy
flig
ht
al
gorit
hm
is
propose
d.
T
he
main
featu
re
of
the
pro
pose
d
method
is
the
abse
nce
of
os
c
il
la
ti
on
at
ste
ady
-
sta
te
co
ndit
ion
,
no
sp
eci
fi
c
init
ia
l
value
is
requir
ed
a
nd
fast
c
omp
uting
ti
me.
T
his
al
go
rith
m
c
om
es
with
abili
ty
to
trac
k
extreme
e
nv
i
ronme
ntal
conditi
on
su
c
h
as
PSC.
C
ompare
d
to
oth
e
r
conve
ntion
al
t
echn
i
qu
e
s,
Le
vy
flig
ht
al
gorit
hm
has
f
ast
tra
ckin
g
sp
ee
d
a
nd
is
si
mp
le
to
be
im
plemente
d.
T
he
rest
of
this
pa
per
is
orga
niz
ed
a
s
f
ollow
s
.
Sect
ion
2
pr
e
s
ents
on
PSC.
Sect
io
n
3
intr
oduce
s
the
the
or
et
ic
al
fr
ame
w
ork
of
Lev
y
flig
ht
op
ti
miza
ti
on
.
Se
ct
ion
4
disc
usse
s
th
e
simulat
ion res
ul
ts. Finall
y, co
nclusi
on is g
i
ve
n
in
Sect
io
n 5
.
2.
PAR
TI
AL
SH
AD
I
NG C
ONDITIO
NS
Ty
pical
ly
unde
r
PSC,
di
ff
e
r
ent
photon
c
urre
nt,
I
ph
is
ge
ner
at
e
d
i
n
c
on
j
un
ct
io
n
wit
h
dif
fer
e
nt
irrad
ia
nces
pe
netrated
at
PV
ar
ray.
This
ph
enomen
on
i
nd
i
rectl
y
ge
ner
at
e
s
a
hot
s
pot
in
the
s
had
e
d
m
odule
s
and
mi
gh
t
dam
ag
e
t
he
m
odul
es.
T
he
s
olu
ti
on
to
this
af
or
e
mentio
ned
pro
blem
is
done
by
int
rod
ucin
g
bypass
diode
w
hich
is
connecte
d
in
par
al
le
l
to
eac
h
m
odule.
T
he
pr
ese
nce
of
bypa
ss
di
od
e
in
each
m
odule
create
s
mu
lt
iple
peak
s
of P
-
V
a
nd
I
-
V
curves
as s
ho
wn in Fi
gure
1
[6].
(a)
(
b)
F
igure
1. (a
)
P
-
V
a
nd (b) I
-
V
curves
unde
r u
nif
or
m
con
diti
on (UIC
)
a
nd P
SC
PV
mod
ule
B
PSX1
50
with
sp
eci
ficat
io
ns
as
sho
wn
in
T
able
1
is
us
e
d
to
ge
ner
at
e
t
he
patte
r
n
of
par
ti
al
sh
a
ding
conditi
ons
in
t
his
pa
per.
T
he
sp
eci
ficat
io
ns
are
obta
ine
d
unde
r
sta
nda
rd
t
est
conditi
on
(
STC).
Ther
e
a
re
t
hr
e
e
di
ff
e
ren
t
pat
te
rn
s
of
pa
rtia
l
sh
a
ding
co
ndi
ti
on
s
ge
ner
at
e
d
as
s
how
n
i
n
Ta
ble
2.
The
same
patte
rn
s
are
ge
ner
at
e
d
in
MATL
AB/Si
mu
l
ink
for
four
P
V
m
odules
B
PSX1
50
c
onne
ct
ed
in
se
ries
unde
r
par
ti
al
s
ha
di
ng cond
it
io
n.
Table
1.
Sp
eci
f
ic
at
ion
s
of
phot
ovoltai
c mod
ul
e
BPSX
150
Table
2.
Partia
l sha
ding
pro
file
Para
m
eters
Ratin
g
STC po
wer ratin
g
,
Pm
ax
1
5
0
W
Op
en
cir
cu
it vo
lta
g
e,
Vo
c
4
3
.5 V
Sh
o
rt
circuit cur
re
n
t,
Isc
4
.75
A
Vo
ltag
e at
m
ax
im
u
m
po
wer
,
V
m
p
p
3
4
.5 V
Cu
rr
en
t at
m
ax
im
u
m
po
wer
,
I
m
p
p
4
.35
A
Nu
m
b
er
o
f
cells
72
Cas
es
Patterns
(
W
/
m
2
)
Peak
po
wer
(W
)
Peak
1
Peak
2
Peak
3
Peak
4
PSC1
1
0
0
0
,8
0
0
,400,2
0
0
1
3
8
.94
2
4
9
.37
2
0
5
.86
1
4
2
.22
PSC2
4
0
0
,400,1
0
0
0
,20
0
1
3
8
.21
1
3
1
.26
1
3
3
.56
PSC3
1
0
0
0
,4
0
0
,200,6
0
0
1
3
8
.21
1
9
2
.86
2
0
3
.59
1
4
1
.54
3.
LE
VY
FLI
G
HT OPTI
MIZ
ATIO
N
Paul
Lev
y
is
a
Fr
e
nc
h
mathe
mati
ci
an
w
ho
introd
uced
Le
vy
fli
gh
t
distri
bu
ti
on
in
19
37
[
24]
.
Le
v
y
fligh
t
is
c
onsi
der
e
d
as
a
breakth
rou
gh
ov
er
tra
diti
on
al
Brownia
n
m
ot
ion
by
c
onsid
erin
g
distrib
ution
f
or
Evaluation Warning : The document was created with Spire.PDF for Python.
In
t J
P
ow Elec
& Dri S
ys
t
IS
S
N:
20
88
-
8
694
A maxim
um
powe
r point tr
ack
ing
base
d on le
vy fl
igh
t
op
ti
mi
za
ti
on
…
(Da
ham
an Ish
ak
)
1501
current
an
d
ne
xt
ju
mp
us
in
g
same
mathema
ti
cal
fo
r
m.
Le
vy
m
otio
n
is
di
verged
in
m
ot
ion
.
Nowa
day
s
,
th
i
s
mo
ti
on
is
bein
g
widely
ex
plored
i
n
natu
re
si
nce
Le
vy
distri
bu
ti
on
ca
n
be
a
pp
li
ed
in
t
he
m
os
t
of
a
nalyse
s
su
c
h
as
in
phys
ic
s
,
bio
lo
gy,
fina
nc
e
an
d
ec
onom
i
cs.
T
he
beh
a
vi
or
of
a
nimal
se
arch
i
ng
f
or
f
ood
is
bein
g
obse
rv
e
d.
It
is
no
ti
ced
th
at
the
a
nimal
be
hav
i
or
lo
ok
i
ng
for
foo
d
is
cl
assifi
ed
a
s
r
an
dom
or
quasi
-
r
andom
ma
nner
.
T
he
decisi
on
of
a
ni
mal
in
nex
t
m
ove
f
or
f
ood
sea
rch
is
base
d
on
curre
nt
posit
io
n
a
nd
t
ran
sit
io
n
prob
a
bili
ty.
Th
us
,
this
be
hav
i
or
c
an
be
m
od
el
e
d
mathemat
ic
al
ly.
It
is
obse
rv
e
d
that
mo
st
a
ni
mal
an
d
insect
s
ex
hib
it
Le
vy
fligh
t
beh
a
vior.
Subs
equ
e
ntly,
t
his
beh
a
vior
is
ap
plied
in
op
ti
mi
zat
ion
an
d
op
t
imal
search.
P
ow
e
r
la
w
is
us
ed
to
mathemat
ic
al
ly
m
odel
the
distribu
ti
on
of
L
evy
flig
ht
[
24]
.
Ra
ndom
wal
k
in
Lev
y
distri
buti
on
is
co
mpri
sed
of
le
ng
th
,
l
dr
a
w
n from
po
wer l
aw
f
unct
io
n
as i
n (1)
[24
,
25]
.
p
(l
)
=
l
-
(1)
wh
e
re
l
de
no
te
s by fli
gh
t l
e
ngth and
is
va
riance, 1<
<
3.
A
sim
plifie
d L
evy flig
ht is e
xpress
ed
in (2
).
x
t+1
= x
t
+
L
evy
(2)
In
he
re,
x
t+1
is
the
nex
t
sta
t
e,
x
t
is
cu
rr
e
nt
sta
te
,
is
ste
p
siz
e
an
d
is
pro
duct
of
e
ntr
ywis
e
mu
lt
ipli
cat
ion
s
.
A
sim
plifie
d st
ep
siz
e,
s
sa
mpl
e g
ene
rati
on is
expres
sed
in (
3).
s =
Le
vy(
)
= k x
(
|
|
1
)
(
−
)
(3)
wh
e
re
k
deno
te
s
as
Lev
y
mu
lt
iplyi
ng
c
oe
ff
ic
ie
nt,
is
1.5
,
wh
il
e
u
a
nd
ar
e
from
no
rmal
distrib
ution.
Given the te
rm u
is
ex
pr
esse
d
i
n (4).
u
N(0,
2
u
)
(4)
Wh
il
e
u
is e
xpress
ed
in (5
).
=
{
Γ
(
1
+
)
(
2
)
Γ
(
1
+
2
)
2
(
−
1
)
2
}
1
(5)
Give
n
e
qu
at
i
on
is d
en
oted
by e
xpressi
on (6
).
= N
(0,
2
)
(6)
Wh
il
e
is ex
pr
ess
ed
in (7
).
= 1
(7)
Figure
2(a)
shows
t
he
fl
ow
c
har
t
of
Le
vy
fligh
t
op
ti
miza
ti
on
(L
FO).
In
this
al
gorith
m,
a
set
of
par
ti
cl
es
is
ch
ose
n.
One
be
st
-
known
l
ocati
on
is
sel
ect
ed
as
sta
rting.
A
w
hole
ne
w
gen
e
ra
ti
on
is
pr
oduce
d
with
Lev
y
fligh
t
m
ot
ion
w
it
h
ra
ndom distri
buti
on.
Th
e n
e
w
g
en
e
rati
on
is
e
valua
te
d
with b
est
-
know
n
l
ocati
on.
This
process
is co
nt
inuousl
y
e
xec
ut
ed
unti
l o
ne pr
om
isi
ng
point i
s obtai
ne
d wh
i
ch
is t
he best s
olu
ti
on.
3.1.
Mech
an
ism
of
t
he
pr
oposed
MPPT se
arch
In
it
ia
ll
y,
al
l
th
e
pa
rtic
le
s
are
distrib
uted
enti
rely
at
P
-
V
c
ur
ve.
Fou
r
par
ti
c
le
s,
D
are
sel
e
ct
ed
in
this
pro
po
se
d
al
gor
it
hm
.
Fig
ur
e
2
sh
ows
the
fl
owchar
t
of
t
he
m
ov
e
ment
mec
ha
nism
of
t
he
pa
rtic
le
s.
T
he
se
le
ct
ed
par
ti
cl
es
are
in
it
ia
ll
y
distribu
t
ed
at
duty
c
ycl
e
0.1,
0.2,
0.5
and
0.7
re
sp
ec
ti
vely.
T
he
po
wer
at
the
r
es
pe
ct
ive
du
t
y
cycle
is
r
ecorde
d.
Ba
se
d
on
the
init
ia
l
distrib
ution
of
t
he
pa
rtic
le
s,
the
highest
powe
r
is
s
el
ect
ed.
T
he
sel
ect
ed
pa
rtic
le
with
t
he
high
est
powe
r
is
t
he
init
ia
l
best
pa
rtic
le
.
This
pa
rtic
le
serv
es
as
the
gu
i
dan
ce
f
or
al
l
oth
e
r
pa
rtic
le
s
for
t
heir
nex
t
moveme
nt.
Th
e
di
recti
on
a
nd
vel
ocity
of
the
m
oveme
nt
of
the
par
ti
cl
e
is
gu
i
ded
by
Lev
y fli
gh
t
as shown in
(2
).
T
he
ste
p si
ze of
t
he
pa
rtic
le
is random as t
he
p
arti
cl
e m
oves closer
t
o
the
M
P
P
the
ste
p
siz
e
is
small
er.
T
he
s
te
p
siz
e
bec
ome
s
zer
o
once
al
l
the
par
ti
cl
es
assigne
d
m
ov
e
d
t
ow
a
rds
M
P
P.
T
he
par
ti
cl
es
c
onti
nu
e
t
o
update
their
searc
h
un
ti
l
the
new
bes
t
pa
rtic
le
is
found.
On
c
e
ne
w
best
pa
rtic
le
f
ound,
Evaluation Warning : The document was created with Spire.PDF for Python.
IS
S
N
:
2088
-
8
694
In
t J
P
ow
Ele
c
&
Dr
i
S
ys
t
,
V
ol
.
1
1
, N
o.
3
,
Se
ptembe
r
2020
:
14
99
–
15
07
1502
new
best
par
ti
c
le
is
updated
.
The
ne
xt
it
erati
on
is
excec
ute
d
unti
l
the
best
s
olu
ti
on
is
f
o
und.
The
best
s
ol
ution
is t
he ou
t
pu
t
D
. A
t t
his c
onditi
on
,
the
ma
xim
um
powe
r
is
suc
cessf
ully trac
ked by t
he
al
gorithm.
(a)
(b)
Figure
2. (a
)
L
evy flig
ht
op
ti
miza
ti
on
(b)
L
evy flig
ht
GMPP
3.2.
Implem
ent
ati
on
of the pr
oposed
MP
PT
The
same
basi
c idea
of
LF
O
as d
isc
usse
d pre
viously is a
ppli
ed
to track
m
aximum p
ower
p
oi
nt o
f P
V
modu
le
.
A
flo
wch
a
rt
of
LF
O
G
M
PP
is
as
s
how
n
in
Fi
gur
e
2(b).
A
f
ull
PV
s
ys
te
m
is
s
how
n
in
Fi
gur
e
3.
In
her
e
,
this
al
gorithm
is
functi
on
e
d
t
o
e
xtract
global
ma
xim
um
powe
r
by
com
par
i
ng
wit
h
the
existi
ng
powe
r
po
i
nts
a
nd
i
de
ntify
i
ng
ph
otovo
lt
ai
c
po
wer
wh
ic
h
is
nee
de
d
to
re
gu
la
te
t
he
du
t
y
c
ycle
of
t
he
boos
t
c
onve
rter.
Lev
y
flig
ht
al
gorith
m
is
ap
plied
in
s
olar
po
wer
s
ys
te
m
t
o
trace
global
m
aximum
power
po
i
nt
un
der
pa
rtia
l
sh
a
ding
co
ndit
ion
s
.
T
he
mai
n
pro
gr
a
m
sta
rt
s
by
se
ns
in
g
t
he
vo
lt
age
an
d
curre
nt
of
P
V
mod
ule
as
s
how
n
i
n
the
flo
wchart
of
Fig
ur
e
2(b).
Each
dut
y
cycl
e
is
randoml
y
distrib
uted
within
the
sea
rc
hin
g
s
pace.
The
fitness
of
eac
h
duty
c
ycle
is
eval
uated
a
nd
t
he
best
duty
c
ycle,
pb
est
is
fou
nd.
T
hen,
f
or
eac
h
duty
cycle
po
sit
ion
is
updated
base
d
on
ra
ndom
pro
bab
il
it
y.
T
he
posit
ion
of
duty
cycle
is
up
dated
ba
sed
on
(
2).
I
f
the
c
urre
nt
du
ty
cycle
is
more
t
han
c
urren
t
pb
est
,
the
n
c
urre
nt
dut
y
c
ycle
be
comes
pbest
.
By
em
ployin
g
Lev
y
f
li
gh
t
i
n
global
powe
r
sea
rch,
i
t
en
han
ce
s
the
search
ca
pab
il
it
y
to
perform
gl
ob
al
e
xplo
rati
on
th
rou
ghout
the
sea
rch
i
ng
s
pace
.
In
Le
vy
flig
ht
,
the
mai
n
ke
y
facto
r
in
t
he
searc
h
is
determi
ned
by
par
a
mete
r.
I
n
he
re,
valu
e
of 1.5
is ch
os
e
n.
Evaluation Warning : The document was created with Spire.PDF for Python.
In
t J
P
ow Elec
& Dri S
ys
t
IS
S
N:
20
88
-
8
694
A maxim
um
powe
r point tr
ack
ing
base
d on le
vy fl
igh
t
op
ti
mi
za
ti
on
…
(Da
ham
an Ish
ak
)
1503
Figure
3.
S
olar
PV po
wer
sy
st
em
4.
SIMULATI
O
N
RESU
LT
S
In
t
his
stu
dy,
a
so
la
r
powe
r
s
ys
te
m
is
m
ode
le
d
in
MATL
AB/Si
mu
li
nk
t
o
ver
if
y
the
ef
f
ect
iveness
of
the
propose
d
a
lgorit
hm
as
de
picte
d
i
n
Fi
gur
e
3.
It
s
hows
a
dc
s
upply
s
ys
te
m
power
e
d
by
P
V
as
i
nput
so
urc
e
connecte
d
to
a
boos
t
c
onve
rter
an
d
dc
loa
d.
I
n
this
sim
ulati
on
,
t
he
P
V
model
is
re
pr
e
sented
as
a
cu
rr
e
nt
so
urce
par
al
le
l
with
a
sin
gle
diode,
a
pa
ral
le
l
sh
unt
resist
or
a
nd
a
se
rie
s
resist
or.
B
oost
conve
rter
ha
ving
par
a
mete
r
s
pec
ific
at
ion
s
of
in
du
ct
or
with
val
ue
of
1
00
uH,
ou
t
pu
t
ca
pacit
or
with
val
ue
of
100
uF
a
nd
20
kH
z
of
s
witc
hing
frequ
e
nc
y
is
desi
gn
e
d
a
nd
mode
le
d
in
M
AT
LA
B/
Simuli
nk.
T
he
bo
os
t
co
nve
rter
is
co
ntr
olled
by
a
powe
r
MOS
FET.
P
W
M
s
witc
hing
is
va
r
ie
d
base
d
on
duty
c
ycle.
A
c
on
sta
nt
loa
d
of
32
Ω
is
a
ppli
ed.
T
he
du
t
y
c
ycle
is
adjuste
d
by
t
he
pro
po
se
d
al
gor
it
hm
to
get
gl
obal
ma
ximum
powe
r
point.
F
irstl
y,
the
pro
pose
d
al
gorithm
is
te
ste
d
unde
r
unif
orm
ir
ra
diance
an
d
la
te
r
unde
r
non
-
unif
orm
irrad
ia
nces.
Fi
gure
4
s
hows
r
esults
of PV p
ower
, v
oltage a
nd c
urr
ent un
d
er
v
a
ry
i
ng irr
a
dianc
es.
(a)
(b)
(c)
Figure
4. P
ho
t
ovoltai
c un
der
varyin
g
ir
ra
diances
(a) p
ower
(b) vo
lt
age
(
c
) c
urren
t
Evaluation Warning : The document was created with Spire.PDF for Python.
IS
S
N
:
2088
-
8
694
In
t J
P
ow
Ele
c
&
Dr
i
S
ys
t
,
V
ol
.
1
1
, N
o.
3
,
Se
ptembe
r
2020
:
14
99
–
15
07
1504
Ba
sed
on
the
resu
lt
s,
al
l
al
gorithms
ar
e
able
to
track
maxim
um
power
po
i
nt
unde
r
cha
ng
i
ng
irrad
ia
nces.
H
ow
e
ve
r,
P&
O
al
gorithm
suff
ers
from
los
se
s
at
ce
rtai
n
ir
r
adiance.
T
he
os
ci
ll
at
ion
s
oc
cur
at
certai
n
i
rr
a
dia
nce
co
ntribute
d
t
o
po
wer
lo
s
ses.
As
ca
n
be
see
n,
P
&O
al
gorithm
has
os
ci
ll
at
ion
,
sl
ower
ti
me
respo
ns
e
an
d
un
sta
ble.
PS
O
al
go
rith
m
s
hows
lo
w
os
ci
ll
at
ion
,
fa
st
respo
nd
ti
me
an
d
sta
ble.
H
ow
e
ve
r,
th
e
pro
po
se
d
L
FO
sh
ows
t
he
be
st
performa
nce
as
com
par
e
d
to
both
P
&O
and
PS
O
since
LFO
has
the
lowes
t
os
ci
ll
at
ion
, fast
est
r
es
pond ti
m
e an
d
the
m
os
t
sta
ble.
Nex
t,
the
pe
rformance
of
t
he
al
gorith
m
is
te
ste
d
un
der
pa
rtia
l
sh
a
ding
conditi
ons.
Th
ree
diff
e
re
nt
patte
rn
s
of
par
t
ia
l
sh
a
ding
c
on
diti
on
s
with
di
ff
e
ren
t
irra
dian
ces
are
pen
et
ra
te
d
at
f
our
seri
es
P
V
mod
ules
.
It
is
well
kn
own
t
ha
t
P&
O
al
go
rithm
has
te
nde
nc
y
t
o
tra
p
at
lo
cal
maxima
durin
g
it
s
sea
rch.
Th
us,
this
al
gorith
m
may
fail
to
det
ect
global
ma
xi
ma
unde
r
par
ti
al
sh
a
ding
c
onditi
on
s
.
Fi
gure
5
s
hows
simul
at
ion
resu
lt
s
for
PS
O
and
L
FO
al
go
rithms.
It
ca
n
cl
early
be
ob
s
erv
e
d
that
bot
h
al
gorithm
s
a
re
a
ble
t
o
dete
ct
gl
ob
al
ma
ximu
m
powe
r
point.
I
n
te
rm
s
of
os
ci
ll
at
ion
,
both
ha
ve
lo
w
ste
a
dy
sta
te
os
ci
ll
at
ion
.
F
urt
he
r
anal
ys
is
is
do
ne
in
te
rm
s
of r
es
pond ti
m
e, it sh
ows that
LFO ha
s fast
er trac
king ti
me
com
par
e
d
t
o
P
SO
,
and
b
oth
a
lgorit
hm
s a
re st
able.
Partia
l
sh
a
ding
conditi
on
2
is
te
ste
d
w
her
e
maxim
um
pow
er
of
eac
h
pea
k
is
cl
os
e
d
to
each
oth
er
.
The
te
st
is
pe
r
forme
d
t
o
asse
ss
the
i
ntell
igence
of
both
al
gorithms
i
n
di
sti
nguish
in
g
be
tween
very
cl
os
el
y
peaks
to
eac
h
oth
e
r.
Fig
ur
e
6
sho
ws
sim
ulati
on
res
ults
un
de
r
P
SC
2
f
or
powe
r,
volt
age
and
c
urren
t.
B
ased
on
these
simulat
io
n
res
ults, b
ot
h
al
gorithms
a
re ab
le
to
disti
nguish
g
lo
bal
ma
ximum power
p
eak
e
ve
n
th
ou
gh
t
he
peaks
are
cl
ose
d
to
on
e
an
ot
her.
Both
al
go
rithms
al
s
o
pr
oduce
lo
w
ste
ady
-
sta
te
os
ci
l
la
ti
on
an
d
sta
ble.
I
n
te
rms of res
po
nd ti
me, it s
hows
th
at
LF
O h
as f
ast
er
r
es
po
nd ti
me com
pa
red to
PS
O.
Simi
la
rly,
the
te
st
is
re
peate
d
for
PSC
3.
I
n
this
c
onditi
on,
t
he
global
pea
k
is
at
t
he
rig
ht
side
of
the
P
-
V
c
urve
.
Both
al
gorit
hm
s
are
te
ste
d
un
der
this
c
ondi
ti
on
.
Ba
se
d
on
res
ults
sho
wn
i
n
Fig
ur
e
7,
PS
O
al
gorithm
ca
n
detect
the
gl
obal
maxim
um
powe
r
po
i
nt
with
s
om
e
c
hange
s
on
the
i
niti
al
value
of
duty
cycl
e
bein
g
made
.
H
ow
e
ve
r,
f
or
L
FO
,
the
same
i
niti
al
values
ar
e
us
e
d
an
d
thi
s
al
gorith
m
is
able
to
detect
global
maxim
um
po
w
er
po
i
nt.
B
oth
hav
e
lo
w
ste
ad
y
-
sta
te
os
ci
ll
at
ion.
H
ow
e
ver,
i
n
te
r
ms
of
res
pond
ti
me,
L
FO
has
faster
res
pond
ti
me
co
mp
a
r
ed
with
P
SO,
an
d
bo
t
h
al
gorith
ms
yield
sta
ble
res
ults.
Table
4
pro
vi
des
a
com
par
is
on
be
tween
c
onve
ntion
al
P
&O,
PSO
a
nd
L
FO
met
hods
in
te
rms
of
os
ci
ll
at
ion
,
re
sp
on
d
ti
me an
d st
abil
it
y.
(a)
(b)
(c)
Figure
5. Parti
al
sh
a
ding c
onditi
on
1 (a)
po
wer (
b) volt
age
(
c)
curre
nt
wa
veforms
Evaluation Warning : The document was created with Spire.PDF for Python.
In
t J
P
ow Elec
& Dri S
ys
t
IS
S
N:
20
88
-
8
694
A maxim
um
powe
r point tr
ack
ing
base
d on le
vy fl
igh
t
op
ti
mi
za
ti
on
…
(Da
ham
an Ish
ak
)
1505
(a)
(b)
(c)
Figure
6. Parti
al
sh
a
ding c
onditi
on
2 (a)
po
wer (
b) volt
age
(
c)
curre
nt
wa
veforms
(a)
(
b)
(c)
Figure
7. Parti
al
sh
a
ding c
onditi
on
3 (a)
po
wer (
b) volt
age
(
c)
curre
nt
wa
veforms
Table
4.
C
omp
ariso
n of P&
O, PS
O
a
nd LF
O
und
e
r PSCs
Alg
o
rithms
Oscillatio
n
Res
p
o
n
d
T
im
e
Stab
ility
P&
O
Fail
Fail
Fail
PSO
Low
Vary
Vary
LFO
Low
Fast
Stab
le
Evaluation Warning : The document was created with Spire.PDF for Python.
IS
S
N
:
2088
-
8
694
In
t J
P
ow
Ele
c
&
Dr
i
S
ys
t
,
V
ol
.
1
1
, N
o.
3
,
Se
ptembe
r
2020
:
14
99
–
15
07
1506
5.
CONCL
US
I
O
N
In
this
pa
per,
LF
O
al
gorith
m
has
bee
n
pro
posed
.
In
th
is
al
gorithm
,
t
he
sea
rc
h
was
ra
ndom
l
y
execu
te
d
in
be
tween
t
he
se
arch
i
ng
sp
ace
,
an
d
no
sp
eci
fic
init
ia
l
value
was
r
eq
uir
ed.
In
this
st
udy,
a
sta
nd
al
on
e
P
V
sy
ste
m
wit
h
boos
t
co
nverter
was
modele
d
in
MATL
AB
/Si
mu
li
nk
to
ve
ri
fy
the
e
ff
ect
i
ve
ness
and
the
pe
r
for
mance
of
t
he
a
lgorit
hm
.
I
n
he
re,
t
he
pro
pose
d
al
gorithm
w
as
te
ste
d
unde
r
tw
o
te
st
c
ondi
ti
on
s
wh
ic
h
we
re
unde
r
un
if
orm
and
non
-
un
if
or
m
irra
diances
.
The
al
go
rith
m
was
c
ompa
red
with
c
onve
ntion
al
P&O
an
d
PS
O.
Furthe
r
st
ud
ie
s
we
re
ca
rr
ie
d
ou
t
unde
r
non
-
un
i
form
co
ndit
ion
s
with
th
ree
dif
fer
e
nt
patte
rn
s
of
PSC.
Ba
se
d
on
the
res
ults,
it
can
be
note
d
that
LF
O
al
go
rithm
is
a
ble
to
work
well
unde
r
unif
orm
a
nd
non
-
un
i
form
i
rr
a
diances.
I
n
a
dd
i
ti
on
,
t
his
al
go
rithm
has
l
o
w
os
ci
ll
at
ion
,
f
ast
respo
nd
ti
me
an
d
sta
ble
.
It
is
exp
ect
e
d
in
fu
t
ur
e
wor
k,
this
al
gorithm
will
be
de
ploye
d
in
the
hard
war
e
and
e
xperime
nt
to
furthe
r
val
idate
it
s p
er
forma
nc
e an
d
e
ff
ect
ive
ness.
ACKN
OWLE
DGE
MENTS
The
aut
hors
w
ou
l
d
li
ke
t
o
t
ha
nk
Un
i
ver
sit
i
S
ai
ns
M
al
a
ys
ia
for
t
he
fin
anci
al
sup
port
unde
r
R
UI
gr
a
nt
scheme
w
it
h p
r
oject
num
ber
1001
/PEL
ECT/
801402
7
.
REFERE
NCE
S
[1]
M.
A.
G.
De
Bri
to,
L
.
Ga
lot
to
,
L
.
P.
Samp
ai
o
,
G.
De
Aze
v
edo
Melo,
and
C
.
A.
C
ane
sin,
“E
v
al
u
ation
of
th
e
m
ain
MP
PT
te
chni
qu
e
s for
photovoltaic
app
li
c
ations,”
I
EE
E
Tr
ans.
Ind.
El
e
ct
ron.
,
vol
.
6
0,
no
.
3
,
pp
.
115
6
–
1167,
2013
.
[2]
C.
Man
ic
ka
m,
G
.
R
.
R
aman,
G.
P.
Ra
ma
n,
S.
I
.
Gane
san,
and
C
.
Naga
ma
ni
,
“A
Hybrid
Algori
th
m
for
Tr
ac
k
ing
of
Global
MP
P
bas
ed
on
Perturb
a
nd
Obs
erv
e
and
Parti
cle
Sw
arm
Optim
izati
o
n
wi
th
Redu
ce
d
Pow
er
Os
cillation
in
String
Inve
r
te
rs,
”
IE
EE Tr
ans.
In
d.
E
lectron.
,
vol
.
0046,
no.
c, pp.
1
–
1,
2016
.
[3]
H.
Islam
et
a
l.
,
“Per
form
anc
e
Eva
lu
at
ion
o
f
Maxim
um
Pow
er
Point
Tra
ck
i
ng
Approac
hes
and
Photovolt
ai
c
Sys
te
ms,”
Ene
rg
ie
s
,
vo
l. 11, no.
2,
p
.
365
,
2018
.
[4]
S.
Haj
ighorba
ni
,
S.;
Rad
zi
,
M.A
.
M.;
Kadir
,
M.
Z
.
A.A.;
Shafi
e
,
“
Dual
Se
arc
h
Maxim
um
Pow
er
Point
(DS
MP
P)
Algorit
hm
Base
d
on
Math
ematic
al
Ana
lysis under Shade
d
Cond
it
i
ons,”
En
ergie
s
,
pp.
12116
–
1214
6,
2015
.
[5]
I.
Alh
am
rouni
,
M.
K.
R
ahmat,
F.
A.
Ism
ai
l
,
M
.
Sal
em,
A.
Jus
oh,
and
T
.
Sut
i
kno,
“D
esign
an
d
developme
n
t
of
SEP
IC
DC
-
DC
boost
conve
r
te
r
for
photovol
t
aic
appl
i
cation,”
,
”
Inte
rnationa
l
J
ournal
of
Pow
er
El
e
ct
roni
cs
and
Dr
iv
e
Syst
ems (I
JP
EDS),
vol
.
10
,
no.
1,
p.
406,
20
19.
[6]
K.
Sa
idi
,
M.
M
a
am
oun,
and
M.
Bounekhl
a
,
“A
n
ew
high
p
erf
orm
anc
e
v
ariabl
e
st
e
p
siz
e
per
turb
-
a
nd
-
observe
MP
PT
al
gorit
h
m
fo
r
ph
otovol
taic
sys
tem,
”
,
”
Inte
rnat
i
onal
Journal
of
Powe
r
E
lectroni
cs
and
Dr
ive
S
y
stems
(IJ
P
EDS),
vol.
10
,
no
.
3
,
pp
.
1662
–
1674
,
20
19.
[7]
O.
La
gd
ani
et
al.
,
“PV
arr
ay
conn
ec
t
ed
to
the
grid
with
the
im
p
lem
ent
a
ti
on
of
MP
P
T
al
go
ri
th
ms
(
I
NC
,
P
&
O
and
FL
me
thod
),”
,
” Int
ernati
ona
l Jo
urnal
of
Pow
er
El
e
ct
ronics
and
Dr
iv
e
Syst
ems (IJ
PE
DS),
vol
.
10,
no.
4
,
pp
.
2084
–
2095,
2019
.
[8]
M.
Kil
li
and
S.
Sam
ant
a
,
“Mo
difi
ed
per
turb
a
nd
observe
MP
PT
a
lgori
th
m
fo
r
drif
t
avoi
d
ance
in
pho
tovol
t
ai
c
sys
te
ms,”
I
EEE
Tr
ans.
Ind.
E
le
c
t
ron.
,
vol
.
62
,
no
.
9,
pp.
5549
–
555
9,
2015
.
[9]
M.
R.
B
engour
in
a,
M.
Rahli,
S.
Slam
i
,
and
L
.
Hass
ai
ne,
“PSO
base
d
d
ire
c
t
power
cont
rol
for
a
mu
l
ti
func
t
ional
gr
id
conne
c
te
d
photo
volt
aic
sys
te
m,”
,
”
Inte
rnat
ional
Journal
of
P
ow
er
Elec
tronic
s
a
nd
Dr
iv
e
Syst
ems
(IJ
P
EDS),
vol.
9,
no
.
2
,
pp
.
610
–
621,
2018
.
[10]
M.
Sli
mi
,
A
.
B
ouche
t
a,
a
nd
B
.
Bouch
iba,
“Ma
xim
um
power
c
ontrol
for
photo
volt
aic
sys
te
m
using
in
te
l
li
gen
t
strat
eg
ie
s
,
”
Int
ernati
onal
Journal
of Powe
r
El
e
ct
r
onic
s and
Dr
iv
e Syste
ms
(IJ
PE
D
S),
vol
.
10
,
no
.
1
,
p
.
423
,
2019
.
[11]
S.
Fara
jd
adian
and
S.
M
.
H.
Hos
seini
,
“Scienc
eDir
ect
Opti
mi
z
at
ion
of
fu
zz
y
-
base
d
MP
PT
con
trol
l
er
v
i
a
me
t
ahe
urist
ic
t
e
chni
ques
for
sta
nd
-
al
one
PV
sys
te
ms,
”
Int.
J.
Hy
drogen
En
ergy
,
vol.
44,
no
.
47,
pp.
25457
–
2547
2,
2019.
[12]
A.
Kiha
l,
F.
Kri
m,
A.
Laib,
B
.
Ta
lbi,
and
H.
A
fghoul,
“An
im
p
rove
d
MP
PT
sc
hem
e
e
mpl
oying
ad
apt
iv
e
integr
a
l
der
ivative
slid
in
g
mode
con
trol
for
photovoltaic
sys
te
ms
und
er
fast
irr
adiati
on
c
hange
s,”
ISA
Tr
ans.
,
vo
l.
87,
pp
.
297
–
306,
2
019
.
[13]
H.
Otm
ane,
M.
Yous
sef,
B.
Mokhtar,
and
A.
Inf
o,
“Com
par
at
iv
e
an
al
ysis
of
c
asc
ade
d
Fuzzy
-
PI
c
ontrol
lers
base
d
-
MP
PT
and
per
t
urb
and
observ
e
MP
PT
in
a
gr
i
d
-
conne
c
te
d
PV
sys
te
m
oper
ati
ng
under
di
ffe
r
ent
we
at
her
and
loa
ding
cond
it
io
ns,”
,
Int
ernati
on
al
Journal
of
Po
wer
E
le
c
tronic
s
and
Dr
iv
e
S
ystem
s
(IJ
P
EDS),
v
ol.
10,
no
.
4,
pp
.
1986
–
1994,
201
9.
[14]
D.
Sera
,
L
.
M
athe,
T
.
K
ere
kes
,
S.
V.
Spat
aru
,
a
nd
R.
T
eodor
esc
u,
“On
the
per
tu
rb
-
and
-
observe
and
inc
re
me
n
ta
l
conduc
t
anc
e
mp
pt
m
et
hods for
PV
sys
te
ms,”
IEE
E
J
.
Phot
ovo
lt
ai
cs
,
vol
.
3
,
no
.
3
,
pp.
1070
–
1078
,
2013.
[15]
H.
Rez
k
e
t
al
.
,
“A
novel
sta
ti
s
ti
c
al
per
for
ma
n
c
e
evalua
t
ion
of
most
mode
rn
o
pti
mization
-
base
d
globa
l
MP
PT
te
chn
ique
s
for
p
art
i
al
ly
shad
ed
PV
sys
te
m,
”
Renew.
Sustain
.
E
nergy
Rev.
,
vol.
115,
no
.
Se
pt
e
mbe
r,
p
.
10937
2,
2019.
[16]
M.
Alshare
ef
,
Z
.
Li
n,
M
.
Ma,
an
d
W.
C
ao,
“Ac
c
el
er
at
ed
Particle
Sw
arm
Optim
i
z
at
ion
for
Photov
olt
aic
Maxi
mum
Pow
er
Point
Tr
a
cki
ng
und
er
Par
t
ia
l
Shading
Con
dit
ions,
”
En
ergi
es
,
vol
.
12
,
no
.
4
,
p
.
623
,
2019
.
Evaluation Warning : The document was created with Spire.PDF for Python.
In
t J
P
ow Elec
& Dri S
ys
t
IS
S
N:
20
88
-
8
694
A maxim
um
powe
r point tr
ack
ing
base
d on le
vy fl
igh
t
op
ti
mi
za
ti
on
…
(Da
ham
an Ish
ak
)
1507
[17]
N.
Hashim
,
Z.
Sal
am,
D.
Johar
i,
and
N
.
F.
Nik
Ismail,
“DC
-
D
C
Boost
Conve
r
te
r
D
esign
for
Fast
and
Ac
cur
ate
MP
PT
Algorit
h
ms
in
Stand
-
Al
one
Photovo
lt
a
i
c
Sys
tem,
”
In
te
r
nati
onal
Journa
l
of
Powe
r
Elec
tronic
s
and
Dr
i
ve
Syste
ms
(IJ
PE
D
S),
vol
.
9
,
no
.
3
,
p.
1038
,
2018
.
[18]
H.
A.
Att
ia
and
F.
D.
Gonz
al
o
,
“Sta
nd
-
al
one
PV
sys
te
m
with
MP
PT
func
ti
on
base
d
on
fu
zz
y
logi
c
cont
ro
l
for
rem
ot
e
bui
ldi
ng
appl
i
ca
t
ions,”
In
te
rnational
Jour
nal
of
Pow
er
Elec
troni
cs
and
D
rive
S
yste
ms
(I
J
PE
DS),
vol.
10,
no.
2
,
p
.
842
,
20
19.
[19]
S.
Della
Kra
cha
i,
A.
B.
Sta
mbo
uli
,
M.
De
ll
a
K
rac
ha
i,
M.
B
ekh
ti
,
and
A.
Info
,
“E
xp
eri
m
ental
inve
stigation
of
art
if
ic
i
al
intellig
enc
e
app
li
ed
in
MP
PT
techniq
ues,
”
,
”
In
te
rna
ti
onal
Journal
of
Pow
er
E
lectronic
s
and
Dr
ive
Syste
ms
(IJ
PE
D
S),
vol
.
10
,
no
.
4
,
pp
.
2138
–
2147
,
2019.
[20]
B.
Ghit
a,
K.
Mo
ham
m
ed,
and
L. Ahme
d,
“Applic
at
ion
and
com
p
a
rison
bet
we
en
th
e
conv
ent
ion
al m
et
hods a
nd
PSO
me
thod
for
maximum
power
point
ext
r
ac
t
ion
in
photovolta
i
c
sys
tems
unde
r
par
ti
a
l
shad
in
g
conditions,
”
,
”
Inte
rnational
Jo
urnal
of Powe
r
El
e
ct
roni
cs
and
Dr
iv
e
Syst
ems (I
JP
EDS),
vol
.
9
,
no.
2
,
pp
.
631
–
6
40,
2018
.
[21]
S.
L
i,
“Opti
k
A
MP
PT
spee
d
op
t
im
izati
on
str
ateg
y
for
photovo
la
t
ic
sys
te
m
using
VWP
interva
l
b
ase
d
on
we
at
h
er
fore
ca
st
,
”
Opt.
-
Int.
J. Light E
l
ectron Opt.
,
vol
.
1
92,
no
.
June
,
p
.
162958,
2019
.
[22]
U.
Yil
maz,
O.
T
urksoy,
and
A
.
Te
ke
,
“Ele
ct
ri
cal
Pow
er
and
Energy
Sys
te
ms
Im
prove
d
MP
PT
m
et
hod
to
in
crease
ac
cur
ac
y
and
sp
ee
d
in
photovo
ltaic
sys
te
ms
und
er
v
ariabl
e
at
mo
spheric
conditio
ns,”
Elec
tr.
Pow
er
En
ergy
Syst
.
,
vol.
113
,
no
.
Jun
e,
pp
.
634
–
651
,
2019.
[23]
S.
Vee
r
ape
n
et
a
l.
,
“A
nov
el
glob
al
ma
xi
mum
po
wer
point
tr
ac
ki
ng
al
gor
it
hm
for
photovol
t
aic
sys
te
m
with
v
ariabl
e
per
turbation
f
req
uenc
y
and ze
ro
o
scil
lation
,
”
Sol.
Ene
rgy
,
vol
.
181
,
no
.
Janua
ry, pp
.
345
–
356
,
2019
.
[24]
J.
Ahmed
and
Z
.
Sal
a
m
,
“A
Ma
xim
um
Pow
er
P
oint
Tracki
ng
(
MP
PT)
for
PV
sys
te
m
using
Cu
ckoo
Se
arc
h
wit
h
par
tial
shad
ing c
apa
bi
li
ty
,
”
Appl.
Ene
rgy
,
vo
l. 11
9,
pp
.
118
–
130
,
2014.
[25]
X.
Yang
and
S.
Deb,
“Cu
ckoo
Sear
ch
via
L
ev
y
Flight
s,
”
in
2
009
World
Con
gress
on
Nature
&
Bi
o
logica
lly
Inspired
Comput
ing
(NaBIC)
,
20
09,
pp
.
210
–
214
.
BIOGR
AP
HI
ES
OF
A
UTH
ORS
Chanuri
Ch
ari
n
re
ce
iv
ed
the
B
Eng
(Hons
)
d
e
gre
e
in
elec
tri
c
al
engi
n
ee
ring
f
rom
Univ
ersit
i
Te
knologi
Ma
lays
ia
,
Johor,
Mal
aysia
and
th
e
M
Sc
in
e
lectr
i
ca
l
a
nd
e
le
c
tronics
e
ngine
er
ing
f
rom
Univer
siti
Sains
Mal
aysia
,
Pen
ang,
Mal
aysia
i
n
2008
and
201
3
respe
ct
iv
el
y
.
She
is
cur
r
ent
ly
pursuing
PhD
d
egr
ee
in
th
e
fiel
d
of
power
elec
t
ronic
s
engi
ne
ering
at
Univ
ersi
ti
Sains
Mal
aysia
.
She
is
al
so
cur
r
ent
ly
a
le
c
ture
r
at
th
e
Fa
cul
ty
o
f
Engi
n
ee
ring
T
ec
hnology
,
Univ
ersit
i
M
al
aysi
a
Perli
s,
Perli
s
,
Mala
ysia
.
Her
c
urre
nt
rese
arc
h
int
er
ests
in
cl
ude
ren
ewa
bl
e
ene
r
gy
and
power
el
e
ct
roni
c converter
s.
Daha
ma
n
Ishak
re
ceive
d
th
e
BS
c
degr
ee
in
elec
tr
ic
a
l
engi
n
eering
fro
m
Syrac
use
Univer
si
ty,
Syrac
use,
NY
,
US
A,
the
MS
c
degr
ee
in
elec
tri
c
al
power
fr
om
Univer
sity
of
Newca
stle,
Newca
stle
upon
Tyne
,
UK
,
and
t
he
PhD
d
egr
ee
i
n
elec
tri
c
al
eng
i
nee
ring
from
th
e
Univer
si
ty
of
Sh
eff
ie
ld
,
Sheffi
el
d,
UK
,
in
199
0,
2001
and
20
05
respe
ctive
ly.
He
is
cur
ren
tl
y
an
As
socia
t
e
Profess
or
with
t
he
School
of
E
l
ec
tr
ic
a
l
and
Ele
ct
roni
c
Engi
ne
er
ing,
Univer
sit
i
S
ai
ns
Mal
aysia.
His
cur
ren
t
rese
arc
h
intere
sts
in
cl
ude
pe
rma
n
ent
m
agne
t
brushl
e
ss
ma
ch
i
nes,
e
lectr
i
ca
l
dr
ive
s,
power
elec
troni
c
conve
r
te
rs
and
r
ene
wabl
e ene
rgy
.
Muhamm
ad
Am
mi
rrul
At
iqi
M
ohd
Za
inur
i
r
ecei
ved
h
is
B.
Eng
.
in
Elec
tr
ic
a
l
a
nd
El
e
ct
ron
ic
Engi
ne
eri
ng
and
Master
of
scie
n
ce
d
egr
e
e
from
t
he
Univer
si
ti
Pu
tra
Ma
la
ysia
(UP
M),
Mala
ysi
a
in
2011
and
201
3,
respe
ct
iv
el
y.
He
the
n
cont
inu
ed
and
r
ec
e
ive
d
his
Doctor
of
Philosophy
degr
e
e
in
E
le
c
trica
l
Pow
er
Engi
n
ee
r
ing
(major
in
Pow
e
r
Elec
tron
ic
s
an
d
Pow
er
Quality
)
from
UP
M
in
2017.
Curr
ent
ly
,
he
is
a
Senior
Le
c
ture
r
a
t
D
ep
art
m
ent
of
Elec
t
ric
a
l
,
E
le
c
tronic
s
and
Sys
te
ms
Engi
ne
eri
ng,
Un
ive
rsiti
Keb
angs
aa
n
Mal
aysia
(U
KM
),
Mal
aysia.
His
rese
arc
h
in
t
ere
sts
include
Pow
er
El
e
ct
ron
i
cs,
Pow
er
Quality,
R
ene
wab
le
Ene
rgy
Sys
tem,
Ene
rgy
Stor
ag
e
Sys
te
m
,
and
Artifi
c
ia
l
In
te
l
ligent.
He
is
al
so
a
rese
arc
h
m
e
mbe
r
a
t
Advanc
ed
Li
gh
tni
ng
an
d
Pow
er
Ene
rgy
Resea
rch
(ALPE
R)
at UP
M a
nd
UM
Pow
er
Ene
r
gy
Dedicat
ed
Ad
vanc
ed
Centre at
UM
.
Evaluation Warning : The document was created with Spire.PDF for Python.