Internati
o
nal
Journal of Ele
ctrical
and Computer
Engineering
(IJE
CE)
V
o
l.
6,
N
o
.
4
,
A
ugu
st
2016
,
pp
.
14
21
~
1
433
I
S
SN
:
208
8-8
7
0
8
,
D
O
I
:
10.115
91
/ij
ece.v6
i4.9
704
1
421
Jo
urn
a
l
h
o
me
pa
ge
: h
ttp
://iaesjo
u
r
na
l.com/
o
n
lin
e/ind
e
x.ph
p
/
IJECE
A
Feed forward Neural Networ
k MPPT Control Strategy
Applied to a Modifi
ed Cuk Converter
Mohamed T
a
har
Makhl
oufi, Yassine
A
bdessemed, Mohamed Sa
l
a
h Khireddine
Electronics
Dep
a
rtment,
Batn
a
2
University
,
050
00
Algeria
Article Info
A
BSTRAC
T
Article histo
r
y:
Received
Dec
14,
201
5
Rev
ised
Jun
8
,
2
016
Accepted
Jun
20,
2016
This
paper
pr
es
ents
an
in
tel
lige
n
t
control
s
t
rat
e
g
y
th
at
us
es
a
f
eedforwar
d
artif
icial
neur
al
network
in
order
to
improve
the
performance
of
the
MPPT
(Maximum
Po
wer
Point
Tracker
)
photovo
ltaic
(P
V)
power
sy
stem
based
on
a
modified
Cuk
converter.
Th
e
proposed
neural
network
con
t
rol
(NNC)
strateg
y
is
desig
n
ed
to
produce
regul
ated
var
i
ab
le
DC
output
v
o
ltag
e
.
Th
e
mathematical
model
of
the
Cu
k
convert
er
is
develop
e
d
and
an
artif
icial
neural
network
algorithm
is
derived.
Th
e
Cuk
con
v
erter
has
s
o
m
e
advant
ages
compared
to
other
ty
p
e
s
of
p
ower
converters
.
However
the
nonlin
ear
chara
c
t
e
ris
t
i
c
of
the
Cuk
conv
ert
e
r
due
to
the
req
u
ired
s
w
it
ching
t
echniqu
e
is
difficu
lt
to
b
e
handled
b
y
conv
en
tional
contro
ller
.
To
overcome
th
is
problem,
a
neural
network
controller
with
online
learn
i
ng
back
propag
a
tio
n
algorithm
is
elabor
ated
.
Th
e
designed
NNC
strate
g
y
tr
acks
the
conver
t
er
voltage
outpu
t
changes
and
improves
the
s
y
stem
dy
na
mic
perf
ormance
regard
less
of
the
load
disturban
ces
and
supply
v
a
ri
ations.
The
proposed
controller
effectiven
ess
dur
ing
d
y
namic
tran
sient
r
e
sponse
is
then
analy
z
ed
and
verif
i
ed
using
MATLAB-Sim
u
link.
Th
e
sim
u
lation
re
sults
confirm
the
exc
e
ll
ent
performance
of
the
proposed
NNC
tec
hnique
for
the
studied
PV
s
ystem.
Keyword:
Battery
M
a
xi
m
u
m
pow
er
poi
nt
t
rac
ker
artificial
n
eural
n
etwork
cont
rol
l
e
r
Mo
d
i
f
i
ed
C
u
k
co
nv
er
ter
p
h
o
t
ovo
ltaic
syste
m
Copyright ©
201
6 Institut
e
o
f
Ad
vanced
Engin
eer
ing and S
c
i
e
nce.
All rights re
se
rve
d
.
Co
rresp
ond
ing
Autho
r
:
Mo
h
a
m
e
d
Tahar
Makh
louf
i,
El
ect
roni
cs
De
part
m
e
nt
,
Batn
a
Un
iv
ersi
ty,
0
500
0
A
l
g
e
r
ia.
Em
a
il:
ra
m
ak
h
l
o
u
fil@yah
o
o
.
fr
1.
INTRODUCTION
In
t
h
e
era
o
f
sust
ai
na
bl
e
e
n
ergy
de
vel
o
p
m
ent
,
pho
tov
o
ltaic
(PV)
tech
no
lo
g
y
[1
]-
[3
]
h
a
s
sh
own
si
gni
fi
ca
nt
pot
ent
i
a
l
as
a
renewabl
e
ene
rgy
s
o
u
rce.
O
u
r
res
earch
w
o
rk
f
o
c
u
ses
o
n
im
pro
v
i
ng
pe
rf
orm
a
n
ce
and
efficiency
of
a
PV
syste
m
throu
gh
t
h
e
u
s
e
of
an
a
p
p
r
op
ri
at
e
al
gori
t
h
m
f
o
r
co
nt
r
o
l
l
i
ng
t
h
e
po
wer
i
n
t
e
rfac
e
.
Th
e
m
a
in
o
b
j
ectiv
e
is
to
fin
d
an
effectiv
e
an
d
o
p
tim
al
s
t
rateg
y
[4
],[5
]
for
ex
tractin
g
th
e
m
a
x
i
m
u
m
av
a
ilab
l
e
po
we
r
f
r
om
t
h
e
PV
ge
nerat
o
r.
M
o
re
ove
r,
t
h
e
st
u
d
y
,
desi
g
n
a
nd
si
m
u
l
a
tion
o
f
a
u
n
i
t
co
m
posed
o
f
a
n
M
PPT
co
n
t
ro
l
techn
i
qu
e
and
th
e
m
a
n
a
g
e
m
e
n
t
o
f
th
e
en
erg
y
tran
sm
itted
to
th
e
lo
ad
are
carried
ou
t.
Th
e
m
a
in
p
a
rts
in
out
study
are
:
the
m
odeling
of
a
P
V
system
,
the
t
opol
ogical
study
of
the
powe
r
i
n
terface,
t
h
e
study
of
m
a
xim
u
m
po
w
e
r
poi
nt
t
r
ac
ki
ng
(M
P
P
T)
al
g
o
ri
t
h
m
s
,
t
h
e
si
m
u
l
a
t
i
on,
t
h
e
desi
g
n
of
t
h
e
M
PPT
C
u
k
[
c
o
n
v
e
r
t
e
r
an
d
t
h
e
PV
outp
u
t
vo
ltag
e
reg
u
l
ation
.
In
th
i
s
in
v
e
stig
ation,
an
in
tellig
ent
co
n
t
ro
l
strateg
y
(NNC)
[6
]-[8
]
is
im
pro
v
ed
a
n
d
t
h
e
p
r
o
b
l
e
m
of
l
o
cal
m
a
xim
a
i
n
t
h
e
po
we
r
c
u
r
v
e
of
t
h
e
P
V
ge
nerat
o
r
occ
u
r
r
i
n
g
du
ri
n
g
part
i
a
l
sha
d
i
n
g
i
s
p
r
o
cessed
[4]
.
T
h
e
NNC
a
v
oi
ds
a
m
i
si
nt
erpret
at
i
on
o
f
t
h
e
l
o
cat
i
on
of
t
h
e
M
PP
u
n
d
er
ra
pi
dl
y
chan
gi
n
g
e
n
vi
r
onm
ent
a
l
co
nd
i
t
i
ons.
In
t
h
i
s
pa
per
t
h
e
C
u
k
co
n
v
e
r
t
e
r
i
s
use
d
be
t
w
een
t
h
e
PV
m
odul
e
panel
and
t
h
e
S
w
i
t
c
h.
An
d
t
h
e
b
a
ttery
is
b
e
tween
th
e
ch
arg
i
n
g
ci
rcu
it
con
t
ro
ller
b
a
ttery
an
d
t
h
e
Switch
.
Th
e
o
u
tlin
e
o
f
th
e
propo
sed
syste
m
i
s
de
pi
ct
ed i
n
F
i
gu
re
1.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
S
SN
:
2
088
-87
08
I
J
ECE
Vo
l.
6
,
N
o
.
4
,
Au
gu
st
2
016
:
14
21
–
1
433
1
422
Fig
u
re
1
.
Ou
tlin
e
o
f
th
e
p
r
op
osed
system
2.
PV
A
RRA
Y C
H
A
RAC
TER
ISTIC
S
An
id
eal
so
lar
cell
m
a
y
b
e
mo
d
e
lled
[1
]
b
y
a
curren
t
so
urce
con
n
ected
i
n
p
a
rallel
with
a
d
iod
e;
t
h
e
current
s
o
urce
repres
ents
the
gene
ra
ted
pho
to
cu
rren
t
when
th
e
sun
light
h
its
th
e
so
lar
p
a
n
e
l,
and
the
d
i
o
d
e
represen
ts
th
e
p
-
n
tran
sitio
n
a
r
ea
of
the
s
o
lar
cell.
In
practice
no
s
o
lar
cell
is
ideal
and
a
s
h
unt
resista
n
ce
Rsh
and
a
series
re
sistance
Rs
com
ponent
are
incorporated
i
n
the
m
odel
acco
rding
to
its
beha
viour.
T
h
e
basic
stru
cture
o
f
a
PV
cell
is
sh
own
i
n
Figu
re
2
,
an
d
t
h
e
equiv
alen
t
circu
it
o
f
a
so
lar
cell
co
m
p
risin
g
parasiti
resi
st
i
v
e c
o
m
pone
nt
s
[1]
,
[9]
i
s
de
pi
ct
ed
Fi
g
u
r
e
3.
Figure
2.
Basic
struct
ure
of
a
PV
cell
Fro
m
th
e
ab
ove
electrical
equiv
a
len
t
circu
it
[1
],[9
]
o
f
so
lar
cell
sho
w
n
i
n
Fig
u
re
3
,
it
is
ev
id
en
t
th
at
the
voltage
across
the
l
o
ad
re
sistance
R
and
th
e
curren
t
I
wh
ich
is
fl
o
w
i
ng
th
ro
ugh
th
is
l
o
ad
can
b
e
wri
tten
as
equat
i
o
n
(
1
) :
(
1
)
Whe
r
e
I
L
l
i
g
ht
gene
rat
e
d
cu
rr
ent
, i
s
t
h
e
di
o
d
e
cu
rre
nt
i
s
t
h
e
cu
rre
nt
w
h
i
c
h
i
s
sh
unt
e
d
t
h
r
o
ug
h R
s
h.
The c
u
r
re
n
t
di
vert
e
d
t
h
ro
u
g
h
t
h
e di
ode
i
s
gi
v
e
n
by
eq
uat
i
o
n
(2
), a
s
f
o
l
l
o
ws:
(
2
)
Here
T
is
th
e
ab
so
lu
te
tem
p
e
r
atu
r
e
i
n
Kelv
i
n
.
q
is
th
e
charg
e
o
f
a
electron
,
K
is
th
e
Bo
ltz
m
a
n
n
’
s
con
s
t
a
nt
,
n
i
s
t
h
e
di
o
d
e
i
d
eal
i
t
y
fact
or
whi
c
h
de
pe
nds
on
t
h
e
cert
a
i
n
P
V
t
echnol
ogy
a
nd
Is
i
s
t
h
e
re
verse
satu
ration
cu
rren
t
in
am
p
e
res.
Su
b
s
titu
ting
t
h
ese
in
t
o
th
e
eq
u
a
tion
(1),
pro
d
u
ces
t
h
e
ch
aracteristic
Equatio
n
(3),
of
a
typ
i
cal
so
lar
cell,
th
is
relates
so
lar
cell
p
a
ram
e
ters
to
th
e
ou
tpu
t
cu
rren
t
and
v
o
l
t
a
g
e
.
PV
Mod
u
le
Cuk P
o
w
e
r
Con
v
erter
MPPT
Neural
Net
w
ork
Battery
Charger
Battery
Sw
itc
h
Load
Evaluation Warning : The document was created with Spire.PDF for Python.
I
J
ECE
I
S
SN
:
208
8-8
70
8
A Feed
f
o
rw
ar
d N
e
ural
N
e
t
w
ork M
PPT C
o
n
t
rol
St
r
a
t
e
gy
A
ppl
i
e
d
t
o
a
..
..
(
M
oh
ame
d
Ta
h
a
r M
a
k
h
l
o
uf
i
)
1
423
(3)
So
m
eti
mes,
to
si
m
p
lify
th
e
mo
d
el,
th
e
effect
o
f
th
e
shu
n
t
resistan
ce
is
n
o
t
co
n
s
id
ered,
th
at
is
Rsh
is
in
fin
ite,
so
t
h
e
ex
pressi
o
n
of
(3
),
sim
p
lify
to
as
equ
a
tion
(4).
(4)
A
PV
pa
nel
i
s
com
posed
of
m
a
ny
sol
a
r
cel
l
s
,
whi
c
h
are
c
o
nnected
i
n
se
ries
and/or
parallel
so
the
out
put
c
u
r
r
ent
and
v
o
l
t
a
ge
o
f
t
h
e
PV
panel
a
r
e
hi
g
h
e
n
o
u
g
h
fo
r
a
cert
a
i
n
a
ppl
i
cat
i
o
n.
Ta
k
i
ng
i
n
t
o
acco
u
n
t
t
h
e
sim
p
l
i
f
i
cat
i
on
of
e
quat
i
on
(4
),
t
h
e
out
put
cur
r
ent
-
v
o
l
t
a
ge
charact
eri
s
t
i
c
of
a
P
V
p
a
n
e
l
i
s
expres
se
d
by
equat
i
o
n
(
5
),
w
h
ere
,
Np
an
d
N
s
are t
h
e
num
ber
of
so
lar
cells
in
p
a
rallel
and
series
resp
ectiv
ely.
Figure
3.
E
qui
valent
cir
c
uit
of
a
s
o
lar
cell
The
p
hot
ov
ol
t
a
i
c
sol
a
r
m
odel
l
ed
wi
t
h
M
a
l
a
b/
Sim
u
l
i
nk
i
s
de
pi
ct
ed
i
n
Fi
gu
r
e
4.
Fig
u
re
4.
Pho
t
o
v
o
ltaic
so
lar
m
o
d
e
lled
with
Malab
/
Si
m
u
lin
k
(5)
Whe
r
e
is
th
e
parallel
n
u
m
b
e
r
of
th
e
PV
cells,
ns
is
t
h
e
se
ries
num
b
er
of
t
h
e
PV
cells,
is
th
e
reve
rse-
saturat
i
on
-cu
rre
nt,
is
th
e
o
u
t
pu
t
cu
rren
t
of
th
e
PV
p
a
n
els,
q
is
the
elec
tronic
cha
r
ge,
k
is
B
o
l
t
z
m
a
nn’s
g
a
s
co
nst
a
nt
,
is
th
e
cell
te
m
p
eratu
r
e
of
th
e
PV
p
a
n
e
ls
an
d
A
is
th
e
id
eality
facto
r
o
f
t
h
e
PV
panel
s
.
Th
e
i
n
ten
s
ity
o
f
so
lar
irrad
i
an
ce
is
th
e
m
o
st
d
o
m
in
an
t
en
v
i
ron
m
en
tal
factor
wh
ich
is
stro
ng
ly
affecting
the
electrical
characteristics
of
sola
r
panel
accordi
n
g
to
the
equation
(5).
T
h
e
effect
of
the
irra
diance
on
t
h
e
v
o
l
t
a
g
e
-cu
rre
nt
(V
-I
)
and
vol
t
a
ge
-
p
o
we
r
(V
-P)
charact
e
ri
s
t
i
c
s
[10]
o
f
sol
a
r
panel
un
de
r
vari
ou
s
irrad
i
an
ce
levels
is
d
e
p
i
cted
in
Figu
re
5.
Fro
m
th
is
fi
gure
it
is
clear
that
unde
r
hi
gher
irradia
n
ce,
t
h
e
PV
cell
produces
hi
ghe
r
output
c
u
rrents
because
the
light
ge
ne
rated
curre
n
t
is
prop
ortionally
generated
by
the
fl
ux
of
ph
ot
o
n
s
.
T
h
e
m
a
xim
u
m
pow
er
p
o
i
n
t
(M
PP)
decreases
with
decreasing
irradi
ance
a
n
d
thi
s
is
indicated
on
eac
h
(V
-P)
cu
r
v
e
in
Figu
re
6.
3.
MA
X
I
MUM
POWER POI
N
T TRACKI
NG
(MPPT)
Usually
there
a
r
e
two
m
a
jor
a
p
proaches
a
dopted
f
o
r
m
a
xim
i
zi
ng
po
wer
ext
r
act
i
o
n
fr
o
m
PV
sou
r
ces
.
First
on
e
is
th
e
m
ech
an
ical
track
ing
o
f
th
e
so
lar
p
a
n
e
l.
In
th
is
case
th
e
p
a
n
e
l
is
attem
p
te
d
to
p
o
s
ition
i
n
an
y
t
e
rrai
n
at
an
an
gl
e o
f
ni
net
y
d
e
gree
wi
t
h
t
h
e
di
rect
i
o
n i
n
c
o
m
i
ng ray
of
s
u
n.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
S
SN
:
2
088
-87
08
I
J
ECE
Vo
l.
6
,
N
o
.
4
,
Au
gu
st
2
016
:
14
21
–
1
433
1
424
Fi
gure
5.
Characteristic
c
u
rves
curre
nt
of
solar
Figure
6.
Cha
r
acteristic
cur
v
e
s
powe
r
of
s
o
lar
pa
nel
, at
di
f
f
e
rent
i
r
ra
di
anc
e
l
e
vel
s
a
n
d
2
5
°C
p
a
nel
, at
di
f
f
ere
n
t
i
rra
di
an
ce l
e
vel
s
a
n
d
2
5
°C
Th
is
issu
e
is
beyo
nd
our
top
i
c
o
f
d
i
scussion.
Th
e
sec
o
nd
one
is
the
electri
cal
MPPT
whe
r
e
electrical
o
p
e
rating
p
o
i
nt
is
forced
at
t
h
e
p
eak
power
po
in
t
co
n
tinuo
u
s
ly
b
y
ad
ju
stin
g
th
e
du
ty
cycle
o
f
t
h
e
DC-DC
con
v
e
r
t
e
r
i
n
se
r
t
ed
bet
w
e
e
n
P
V
ar
ray
an
d
l
o
ad.
T
h
e
m
e
t
hods
vary
i
n
c
o
m
p
l
e
xi
ty
,
sens
ors
re
qui
re
d,
t
r
acki
n
g
efficiency,
c
onverge
nce
s
p
ee
d,
c
o
st,
a
n
d
i
n
ot
he
r
re
spect
s.
Som
e
of
t
h
e
w
e
l
l
-
kn
o
w
n
t
ech
ni
q
u
es
a
r
e
Pe
rt
ur
b&
O
b
serv
e
(
P
&O
)
[2
],
[4
],
Incr
em
en
tal
Co
n
d
u
c
tan
ce
[
10],[
11
],
Fr
action
a
l
O
p
en-
C
ircu
it,
Fraction
a
l
Sh
ort-
C
i
rcui
t
,
Fuzzy
Logi
c
[6]
an
d
Neu
r
al
Net
w
o
r
k
[7]
,
[8]
,
[1
2]
,
[
1
3
]
.
B
o
t
h
pe
r
t
ur
b
an
d
o
b
se
r
v
e,
a
nd
i
n
cre
m
ent
a
l
conductance
,
a
r
e
exam
ples
of
"hill
clim
bing"
m
e
thods
t
h
at
can
fi
nd
t
h
e
l
o
cal
m
a
xim
u
m
of
t
h
e
power
curve
f
or
th
e
op
er
atin
g
cond
itio
n
of
th
e
PV
arr
a
y
,
and
so
pr
ov
i
d
e
a
t
r
u
e
m
a
x
i
m
u
m
p
o
w
er
po
in
t.
Th
e
p
e
r
tur
b
and
o
b
s
erv
e
m
e
th
od
can
produ
ce
o
s
cillatio
n
s
of
p
o
wer
o
u
t
p
ut
aroun
d
th
e
max
i
m
u
m
p
ower
po
in
t
ev
en
un
d
e
r
steady
state
irradiance.
T
h
e
increm
en
t
a
l
cond
uct
a
nce
m
e
tho
d
has
t
h
e
a
d
vant
a
g
e
o
v
er
t
h
e
pert
ur
b
an
d
obs
er
ve
(P&O)
m
e
th
o
d
th
at
it
can
d
e
t
e
rm
in
e
th
e
m
a
x
i
m
u
m
p
o
w
er
p
o
i
n
t
withou
t
o
s
cillatin
g
arou
nd
t
h
is
v
a
lu
e
[11
]
.
It
can
perform
max
i
m
u
m
p
o
wer
po
in
t
track
i
ng
u
n
d
e
r
ra
p
id
ly
v
a
ryin
g
irrad
iation
co
nd
itio
ns
with
h
igh
er
accuracy
t
h
an
the
pert
urb
a
n
d
obs
erve
m
e
thod.
Howe
ver,
the
inc
r
em
en
tal
conductanc
e
m
e
thod
ca
n
produc
e
oscillations
and
can
pe
rform
errati
cally
under
ra
pidly
cha
ngi
ng
atm
o
spheric
conditions.
The
com
put
ational
t
i
m
e
i
s
i
n
creased
d
u
e
t
o
sl
o
w
i
ng
d
o
w
n
of
t
h
e
sam
p
l
i
ng
fre
que
ncy
res
u
l
t
i
ng
fr
om
t
h
e
highe
r
com
p
l
e
xi
t
y
of
t
h
e
alg
o
r
ith
m
co
mp
ar
ed
to
th
e
P&O
m
e
th
o
d
[2],
[
4
],
[14
].
The
ML
P
neural
m
e
thod
[8]
i
s
base
d
on
t
h
e
estim
a
ti
on
of
t
h
e
opt
i
m
i
s
ed
v
a
l
u
e
[
5
]
of
t
h
e
dut
y
cy
cl
e
fo
r
t
r
a
n
sm
i
t
t
i
ng
t
h
e
m
a
xi
m
u
m
power
pr
o
d
u
ced
by
t
h
e
sol
a
r
pa
nel
,
gr
een
char
ge.
Th
e
p
o
we
r
e
ffi
ci
enc
y
of
t
h
i
s
syste
m
in
th
is
case
is
m
a
x
i
m
u
m
.
Th
e
poin
t
of
m
a
xim
u
m
powe
r
i
n
t
h
i
s
case
is
determined
acc
urately
rega
rdl
e
ss
o
f
t
h
e
fast
va
ry
i
n
g
e
x
t
e
r
n
al
co
n
d
i
t
i
ons
suc
h
a
s
sol
a
r
ra
di
at
i
on
or
pa
nel
t
e
m
p
erat
ure
4.
THE MODIFIED
CUK
P
O
WER CO
NV
ERTER
Th
e
Cuk
conver
t
er
[1
5
]-[1
8
]
is
a
typ
e
o
f
step
-do
w
n
/
step-
u
p
conv
er
ter
[
9
]
,
[
1
9
]
b
a
sed
on
a
sw
itch
i
ng
b
o
o
s
t
bu
ck
topo
log
y
.
Essen
tially,
th
e
co
nv
erter
is
co
m
p
o
s
ed
of
two
sect
io
n
s
,
an
inpu
t
stag
e
and
an
ou
tpu
t
stage.
T
h
e
bas
i
c
schem
a
tic
of
the
Cuk
c
onverter
is
pr
esen
ted
in
Figu
re
7
,
wh
ere
Vin
is
an
inp
u
t
vo
ltag
e
source;
V
0
i
s
t
h
e
out
put
v
o
l
t
a
ge;
L
1
is
an
inp
u
t
i
n
du
ctor
and
th
e
MO
SFET
S
is
a
con
t
rollab
l
e
switch
.
C
1
is
an
energy
transfe
r
capacitor,
D
1
is
a
diode,
C
0
and
L
0
are
resp
ectiv
ely
a
filter
cap
acito
r
an
d
i
n
du
ct
or.
Th
e
resi
st
ance
R
i
s
t
h
e
l
o
ad
.
A
n
i
m
port
a
nt
adv
a
nt
age
of
t
h
i
s
t
o
pol
ogy
i
s
a
co
n
t
i
nuo
us
c
u
r
rent
at
bot
h
t
h
e
i
n
p
u
t
an
d
th
e
ou
tpu
t
of
th
e
conv
er
ter.
Th
e
d
isadv
antag
e
s
of
t
h
e
C
uk
c
o
nve
rt
er
are
t
h
e
hi
g
h
num
ber
of
r
eact
i
v
e
com
pone
nt
s a
n
d t
h
e
hi
gh
cu
rr
ent
st
resse
s
on
t
h
e s
w
i
t
c
h S
, t
h
e
di
o
d
e
D
1
,
a
n
d
the
ca
pacitor
C
0
.
Fi
gu
re
7.
C
u
k
con
v
e
rt
e
r
Evaluation Warning : The document was created with Spire.PDF for Python.
I
J
ECE
I
S
SN
:
208
8-8
70
8
A Feed
f
o
rw
ar
d N
e
ural
N
e
t
w
ork M
PPT C
o
n
t
rol
St
r
a
t
e
gy
A
ppl
i
e
d
t
o
a
..
..
(
M
oh
ame
d
Ta
h
a
r M
a
k
h
l
o
uf
i
)
1
425
During
m
o
d
e
1
(Fi
g
ure
8),
th
e
in
pu
t
vo
ltag
e
is
app
lied
wh
en
th
e
MOSFET
switch
tran
sist
o
r
S
i
s
closed
the
n
the
curre
nt
through
the
i
n
duct
o
r
L
1
rises.
At
t
h
e
sam
e
tim
e
the
volta
ge
of
capacitor
C
1
re
vers
e
bi
ases
di
od
e
D
1
and
t
u
rns
it
off.
The
ca
pacit
o
r
C
1
discha
rges
its
energy
to
the
circuit
formed
by
C
1
,
C
0
,
L
0
and
th
e
lo
ad
R.
Fi
gu
re
8.
C
u
k
con
v
e
r
t
e
r
on
st
at
e
Du
ri
n
g
m
ode
2,
t
h
e
i
n
put
v
o
l
t
a
ge
i
s
appl
i
e
d
an
d
t
h
e
swi
t
ch
S
i
s
ope
n,
t
h
en
t
h
e
di
o
d
e
D
1
is
fo
rward
biased
and
the
capacitor
C
1
is
charge
d
t
h
rough
L
1
.Th
e
en
erg
y
wh
ich
is
stored
in
th
e
ind
u
cto
r
L
0
is
tran
sferred
to
the
l
o
ad.
T
h
us,
t
h
e
diode
D
1
an
d t
h
e
swi
t
c
h S
p
r
o
v
i
d
e a s
y
nch
r
o
n
ous
co
nve
rt
er
swi
t
c
hi
ng
act
i
o
n
(
F
i
g
u
r
e
9).
Fi
gu
re
9.
C
u
k
con
v
e
r
ter
in
the
off-state
The
rel
a
t
i
ons
b
e
t
w
een
o
u
t
p
ut
and
i
n
put
c
u
rre
nt
s a
n
d
v
o
l
t
a
ge
s are
gi
ve
n i
n
t
h
e
fol
l
o
wi
n
g
:
(6)
(7)
Thu
s
th
e
ratio
n
o
f
the
inp
u
t
and
ou
tpu
t
vo
ltages
for
th
e
b
u
c
k
-
bo
o
s
t
co
nv
erter
is
th
e
sam
e
as
th
e
ratio
of
t
h
e
i
n
p
u
t
an
d
o
u
t
p
ut
cu
rre
nt
s
f
o
r
t
h
e
C
uk
converte
r.
T
h
e
adva
ntage
of
th
e
m
odi
fi
ed
C
uk
c
o
n
v
e
rt
e
r
[1
6]
i
s
th
at
th
e
i
n
pu
t
an
d
ou
tpu
t
i
n
du
ctor
s
c
r
eat
e
a
sm
oot
h
c
u
r
r
e
n
t
at
bot
h
si
de
s
o
f
t
h
e
c
o
n
v
e
r
t
e
r
whi
l
e
t
h
e
buc
k,
bo
ost
an
d
b
u
ck
-b
o
o
st
have
at
l
east
one
si
de
wi
t
h
p
u
l
s
ed
cu
rre
nt
.
The
si
m
u
l
a
t
i
on
m
odel
wi
t
h
M
a
t
l
a
b/
Si
m
u
li
n
k
of
the
overall
Cuk
converter
bases
PV
system
is
prese
n
ted
in
Figure
10.
Fi
gu
re
1
0
. M
a
t
l
ab/
S
im
ul
i
nk si
m
u
l
a
t
i
on m
ode
l
of
t
h
e C
u
k
co
nve
rt
er
Evaluation Warning : The document was created with Spire.PDF for Python.
I
S
SN
:
2
088
-87
08
I
J
ECE
Vo
l.
6
,
N
o
.
4
,
Au
gu
st
2
016
:
14
21
–
1
433
1
426
5.
BATTERY MATHEMATICAL
MODEL
B
a
t
t
e
ry
m
odel
can
us
ual
l
y
be
di
vi
ded
i
n
t
o
ex
pe
ri
m
e
n
t
al
m
odel
,
elect
roc
h
em
i
c
al
m
odel
an
d
equi
val
e
nt
ci
rc
ui
t
m
odel
.
The
equi
val
e
nt
ci
r
c
ui
t
m
odel
i
s
m
o
st
sui
t
a
bl
e
fo
r
dy
nam
i
c
sim
u
l
a
t
i
on.
B
a
se
d
o
n
she
p
h
r
ed
bat
t
e
ry
m
odel
,
re
fer
e
nce
p
r
ese
n
t
s
a
gene
ri
c
b
a
ttery
m
o
d
e
l
fo
r
dyn
amic
si
m
u
la
tio
n
,
wh
ich
assu
m
e
s
th
at
th
e
b
a
ttery
is
co
m
p
o
s
ed
of
a
con
tro
lled
-
v
o
ltag
e
so
ur
ce
and
a
series
re
sistance,
s
h
own
as
Figure
11.
Thi
s
gene
ri
c
bat
t
e
ry
m
odel
con
s
i
d
e
r
s t
h
e
st
at
e o
f
c
h
arge
(SOC)
a
s
the
only
state
va
riable.
Fi
gu
re
1
1
.
A
g
e
neri
c
bat
t
e
ry
m
odel
The e
x
pressi
on
o
f
t
h
e
co
nt
r
o
l
l
e
d
vol
t
a
ge
s
o
u
r
ce i
s
:
(8)
Whe
r
e,
Eb
is
n
o
-l
oad
v
o
ltage
(V
);
E
0
i
s
bat
t
e
ry
co
nst
a
nt
vol
t
a
ge
(
V);
K
i
s
pol
a
ri
zat
i
on
vo
l
t
a
ge
(V
);
Q
is
b
a
ttery
cap
acity
(Ah
)
;
A
is
ex
pon
en
tial
zo
n
e
am
p
litu
d
e
(V);
B
is
expo
n
e
n
tial
zon
e
ti
m
e
co
n
s
tan
t
i
n
v
e
rse
(A
h-
1)
。
.
Th
is
m
o
d
e
l
assu
m
e
s
th
e
in
tern
al
resistan
ce
o
f
the
b
a
ttery
is
k
e
p
t
co
nstan
t
du
rin
g
bo
th
ch
arge
an
d
di
scha
rge
pa
he
s.
Al
l
param
e
ters
are
de
d
u
ce
d
fr
om
t
h
e
di
schar
g
e
an
d
ass
u
m
e
d
t
o
be
sam
e
for
cha
r
ge.
Fi
gu
re
1
2
is
d
i
sch
a
rge
ch
aracteristics
of
th
e
b
a
ttery
at
rated
di
sc
ha
rge
cu
rre
nt
,
an
d
al
l
param
e
t
e
rs
can
be
cal
cul
a
t
e
d
b
y
th
ree
po
in
ts
m
a
rk
ed
i
n
th
e
figu
re,
n
a
m
e
l
y
fu
lly
ch
arg
e
d
vo
ltag
e
(E
fu
l
l
)
,
th
e
end
of
ex
pon
en
tial
zone
(
E
exp
,
Q
exp
), t
h
e e
n
d
of
n
o
m
i
nal
zon
e
(E
nom
, Q
nom
).
f
ul
l
E
ex
p
ex
p
,
EQ
,
nom
no
m
EQ
Figure
12.
Dis
charge
c
h
aracteristics
curve
of
th
e
battery
at
the
rate
d
disch
a
rge
cu
rre
nt
(
V
-Q
)
Use
MATLAB
/Si
m
u
lin
k
to
mo
d
e
l
th
e
b
a
ttery,
as
sh
own
in
Fig
u
re
13
.
The
b
a
ttery
d
i
scharg
e
cu
rv
es
at
di
ffe
re
nt
di
s
c
har
g
e c
u
r
r
e
n
t
s
ar
e
ob
tain
ed,
sh
own
in
Figur
e
14
.
Fig
u
re
13
.
Th
e
b
a
ttery
m
o
d
e
l
in
MALTAB/Si
m
u
lin
k
Figure
14.
Disc
harge
c
h
aracte
r
istics
curve
s
of
t
h
e
battery
a
t
diffe
re
nt
disc
har
g
e
c
u
r
r
e
n
t
(
V
-
Q
)
Evaluation Warning : The document was created with Spire.PDF for Python.
I
J
ECE
I
S
SN
:
208
8-8
70
8
A Feed
f
o
rw
ar
d N
e
ural
N
e
t
w
ork M
PPT C
o
n
t
rol
St
r
a
t
e
gy
A
ppl
i
e
d
t
o
a
..
..
(
M
oh
ame
d
Ta
h
a
r M
a
k
h
l
o
uf
i
)
1
427
0
1
2
3
4
5
6
7
8
10
-20
10
-15
10
-10
10
-5
8 E
poc
hs
T
r
ai
ni
ng-B
l
ue
P
e
rf
orm
anc
e
i
s
1.
74171e-
024,
G
oal
i
s
0
6.
DESCRIPTI
ON OF
THE
PROP
OSE
D
TECHNI
QUE
The
M
PPT
st
rat
e
gy
pr
o
pos
ed
here
c
o
n
s
i
s
t
s
of
a
com
b
i
n
at
i
o
n
o
f
a
neural
net
w
o
r
k
an
d
P
&
O
tech
n
iqu
es
fo
r
th
e
im
p
le
m
en
tatio
n
of
the
duty
cycle
regulator.
Whe
n
s
o
lar
rad
i
ation
ch
ang
es
slowly,
th
e
sy
st
em
cont
rol
s
t
h
e
DC
-
D
C
c
o
n
v
e
r
t
e
r
usi
n
g
t
h
e
P&
O,
an
d
t
h
e
neu
r
al
net
w
or
k
l
ear
ns
si
m
u
l
t
a
neo
u
sl
y
t
h
e
M
PP
fo
u
nd
by
t
h
e
P&O
[
2
]
.
H
o
w
e
ver
i
f
t
h
e
s
o
l
a
r
ra
di
at
i
on
va
ri
es
t
oo
ra
pi
dl
y
[4]
,
t
h
e
neu
r
al
net
w
o
r
k
co
nt
r
o
l
l
e
r
tracks
the
MPP
rapi
dly
and
adjusts
th
e
duty
cycle
of
the
DC-DC
converte
r
[
1
]
.
Ne
u
r
al
net
w
o
r
ks
us
ual
l
y
require
inde
pe
nde
nt
and
ide
n
tically
distribut
ed
sam
p
les
to
ens
u
re
s
u
cces
s
f
ul
on-line
learning.
Here
,
howeve
r
,
sim
i
l
a
r t
r
ai
ni
ng sam
p
l
e
s are
use
d
by
t
h
e
art
i
ficial
neural
netw
ork
(ANN).
The
architectur
e
of
the
ANN
use
d
is
o
f
t
h
e
m
u
lti-la
yer
p
e
rcep
t
r
on
(MLP)
typ
e
.
Th
e
m
a
in
id
ea
of
th
e
learn
i
n
g
alg
o
rith
m
is
th
at
th
e
n
e
u
r
al
n
e
t
w
ork
learns
each
sa
m
p
le
online
be
cause
it
is
diffi
cult
to
store
al
l
learning
samp
les
in
sm
all
devices.
In
Fi
gure
15,
the
ANN
lea
r
ning
technique
i
s
a
m
e
m
o
ry-ba
sed
one
a
n
d
a
llows
t
o
estim
a
t
e
at
any
instant
the
re
quire
d
optim
a
l
duty
cycle
’D’
[1].
Eve
n
wit
h
sparse
data
in
a
m
u
ltidi
m
e
n
sional
m
easur
e
m
ent
space,
t
h
e
al
gorithm
provide
s
sm
oot
h
t
ransi
t
i
ons
f
rom
one
e
s
t
i
m
a
t
e
d
val
u
e
of
D
t
o
an
ot
he
r.
T
h
e
AN
N
c
o
nsi
s
t
s
o
f
a
n
i
n
p
u
t
l
a
y
e
r
(P
p
v
)
,
t
wo
h
idd
en
layer
s
w
ith
resp
ectively
f
iv
e
n
e
ur
on
s
an
d
two
n
e
u
r
on
es
an
d
an
ou
tpu
t
layer
w
h
ich
co
n
s
ists
o
f
one
neu
r
one
w
h
i
c
h
has
t
h
e
o
u
t
p
ut
t
h
e c
o
n
v
e
rt
e
r
dut
y
rat
i
o
D s
h
ow
n i
n
t
h
e
fol
l
owi
n
g
fi
gu
re:
Fi
gu
re
1
5
. T
r
ai
ni
n
g
e
r
r
o
r
usi
n
g t
h
e
ne
u
r
al
ne
t
w
o
r
k
M
L
P
7.
R
ESU
LTS AN
D ANA
LY
SIS
Fi
gu
re
16
sh
o
w
s
t
h
e
P
V
si
m
u
l
a
t
i
on
sy
st
em
u
nde
r
M
A
TL
AB
/
S
i
m
ul
i
nk/
Sim
powe
r
Sy
s
t
em
s
i
n
or
de
r
to
v
a
lid
ate
th
e
o
n
-lin
e
learn
i
ng
ANN,
ad
equate
si
m
u
l
a
t
i
ons
t
e
st
s ha
ve
bee
n
i
m
pl
em
ent
e
d an
d ca
rri
ed
o
u
t
.
Fig
u
re
16
.
Simu
latio
n
system
in
MATLAB/Si
m
u
lin
k
B
a
sed
on
t
h
e
a
b
o
v
e
m
odel
s
a
n
d
co
nt
r
o
l
m
e
tho
d
s,
th
e
grid-co
n
n
ected
h
ybrid
PV/Battery
g
e
n
e
ration
syste
m
can
b
e
i
m
p
l
e
m
en
ted
in
MATLAB/Si
m
u
lin
k
,
as
sh
own
in
Figure
1
6
.
In
th
is
stu
d
y
,
t
h
ree
simu
latio
n
cases
are
consi
d
ere
d
,
nam
e
ly:
a.
Sim
u
latio
n
i
n
stan
d
a
rd
normal
en
v
i
ro
n
m
en
tal
con
d
ition
s
;
b.
Si
m
u
l
a
t
i
on
wi
t
h
t
h
e
occ
u
r
a
nce
of
t
w
o
di
f
f
ere
n
t
di
st
ur
ba
nces;
c.
Sim
u
latio
n
with
co
nstan
t
step
ch
ang
e
s
o
f
th
e
PV
so
lar
rad
i
atio
n
.
D
Ppv
Evaluation Warning : The document was created with Spire.PDF for Python.
I
S
SN
:
2
088
-87
08
I
J
ECE
Vo
l.
6
,
N
o
.
4
,
Au
gu
st
2
016
:
14
21
–
1
433
1
428
7.
1.
Operation in
Standard
E
n
vironmental Conditi
ons
The
Fi
g
u
re
s
1
7
,
1
8
an
d
1
9
bel
o
w
al
l
o
w
u
s
t
o
vi
sual
i
ze
t
h
e
out
put
P
V
panel
cur
r
e
n
t
,
vol
t
a
ge
an
d
p
o
wer
using
the
ANN
con
t
ro
llers
in
stan
d
a
rd
atm
o
sp
h
e
ric
co
nd
itio
ns
(100
0W/m
2
,
25
°C
).
Th
e
n
e
x
t
fo
llo
wi
n
g
Fi
gu
res
2
0
,
2
1
a
n
d
2
2
sh
o
w
t
h
e
C
u
k
c
o
nve
rt
er
[
1
]
o
u
t
put
c
u
r
rent
,
v
o
l
t
a
ge
a
n
d
po
wer
usi
n
g
t
h
e
A
N
N
cont
rol
l
e
rs
i
n
s
t
anda
rd
at
m
o
sphe
ri
c
co
n
d
i
t
i
ons :
we
ha
ve
o
b
t
a
i
n
ed
a
t
r
ai
ni
ng
e
r
r
o
r
o
f
ab
o
u
t
1.
0e
-19
.T
he
out
put
bat
t
e
ry
vol
t
a
ge
duri
ng
t
h
e
c
h
a
r
gi
ng
pe
ri
o
d
i
s
prese
n
t
e
d
i
n
Fi
gu
re
2
3
.
The
c
o
r
r
es
po
n
d
i
n
g
st
at
e
of
char
ge
of
t
h
e
bat
t
e
ry
i
s
p
r
ese
n
t
e
d i
n
Fi
gu
re
24
.
0
0.
01
0.
0
2
0.
0
3
0.
0
4
0.
0
5
0.
06
0.
07
0.
08
0.
0
9
0.
1
-2
0
2
4
6
8
10
Ti
m
e
(
s
e
c
Curr
e
n
t
(
A
)
Ip
v
0
0.
0
1
0.
0
2
0.
0
3
0.
04
0.
0
5
0.
06
0.
0
7
0.
08
0.
0
9
0.
1
30
40
50
60
70
80
90
10
0
11
0
Ti
m
e
(
s
e
c
)
V
o
l
t
ag
e(V
)
Vp
v
Figure
17.
T
h
e
output
P
V
panel
cur
r
ent
Fi
g
u
re
18
. T
h
e
o
u
t
p
ut
PV
pa
nel
v
o
l
t
a
ge
0
0.
01
0.
02
0.
03
0.
04
0.
05
0.
06
0.
07
0.
08
0.
09
0.
1
-1
0
0
0
10
0
20
0
30
0
40
0
50
0
60
0
70
0
Ti
m
e
(
s
e
c
)
Po
w
e
r(W
)
Pp
v
0
0.
0
1
0.
0
2
0.
03
0.
0
4
0.
05
0.
0
6
0.
0
7
0.
0
8
0.
0
9
0.
1
0
50
10
0
15
0
20
0
25
0
30
0
35
0
40
0
45
0
Ti
m
e
(
s
e
c
)
Po
w
e
r
(
W
)
Po
u
t
Fi
gu
re
1
9
. T
h
e
o
u
t
p
ut
P
V
pan
e
l
po
we
r
Fi
gure
2
2
.
The
C
u
k
c
o
nve
rt
er
o
u
t
p
ut
po
we
r
w
i
t
h
T
h
e
A
N
N
c
o
n
t
r
o
l
l
e
r
0
0.
01
0.
02
0.
03
0.
04
0.
05
0.
06
0.
07
0.
08
0.
09
0.
1
-1
0
1
2
3
4
5
6
7
8
Ti
m
e
(
s
e
c
)
C
ur
r
ent
(
A
)
Io
u
t
0
2
4
6
8
10
12
x 1
0
4
47.
1986
47.
1986
47.
1986
47.
1986
47.
1986
47.
1986
47.
1986
Ti
m
e
(
s
e
c
)
Vo
l
t
a
e
(
V)
V
bat
Fi
gu
re
2
0
. T
h
e
C
u
k c
o
n
v
ert
e
r
o
u
t
p
ut
Fi
gu
re
2
3
.
Th
e o
u
t
p
ut
bat
t
e
r
y
vol
t
a
ge
c
u
rre
nt wit
h
AN
N c
o
ntr
o
ller
du
rin
g
c
h
ar
gin
g
Pe
rio
d
Evaluation Warning : The document was created with Spire.PDF for Python.
I
J
ECE
I
S
SN
:
208
8-8
70
8
A Feed
f
o
rw
ar
d N
e
ural
N
e
t
w
ork M
PPT C
o
n
t
rol
St
r
a
t
e
gy
A
ppl
i
e
d
t
o
a
..
..
(
M
oh
ame
d
Ta
h
a
r M
a
k
h
l
o
uf
i
)
1
429
0
0.
01
0.
02
0.
03
0.
0
4
0.
05
0.
06
0.
07
0.
08
0.
0
9
0.
1
-1
0
0
10
20
30
40
50
60
Ti
m
e
(
s
e
c
)
V
ol
t
age
(
V
)
Vo
u
t
0
2
4
6
8
10
12
x 1
0
4
0
0.
2
0.
4
0.
6
0.
8
1
1.
2
1.
4
1.
6
1.
8
2
Ti
m
e
(
s
e
c
)
SO
C
S
t
at
e o
f
c
har
g
e
Figu
r
e
21
.
Th
e
C
u
k
conv
er
ter
o
u
t
p
u
t
vo
ltag
e
Figure
24.
T
h
e
state
of
c
h
arge
of
batt
ery
with
Th
e
ANN
con
t
ro
ller
with
ANN
contro
ller
7.
2.
Simulati
on w
i
th
Distur
banc
es
In
Fi
g
u
res
25
-
3
1
,
t
w
o
di
ffe
r
e
nt
di
st
u
r
ba
nc
es
are
ass
u
m
e
d
as
i
n
p
u
t
i
n
t
h
e
i
rra
di
an
ce
of
P
V
panel
(usi
ng
si
g
n
al
b
u
i
l
d
er
bl
oc
o
f
sim
u
l
i
nk).
T
h
e
fi
rst
one
‘
d
1
’
c
o
r
r
e
pon
ds
to
a
su
dd
en
step
in
cr
ease
o
f
1000W
/m
2
in
th
e
so
lar
radiatio
n
wh
ich
occu
rs
at
0
.
02
8s
and
th
e
second
o
n
e
‘d
2
‘
is
a
sudde
n
ste
p
de
crease
of
1000W
/m
2
whi
c
h
ha
ppe
n
s
at
0.0
68s
.
T
h
e
A
NN
c
ont
r
o
l
l
e
r
ad
just
s
t
h
e
d
u
t
y
cy
cl
e
of
t
h
e
C
u
k
con
v
e
r
t
e
r
t
o
p
r
od
uce
m
a
xim
u
m
pow
er t
o
char
ge t
h
e bat
t
e
ry
. The r
e
sp
onse t
i
m
e of t
h
e sy
st
em
at
t
h
e st
art
i
s
shor
t
about
1
0
m
s
,
whi
l
e
i
t
i
s
about
6
0
m
s
for
ot
he
r
sy
st
em
s
[3]
.
The
o
u
t
p
ut
l
o
ad
v
o
l
t
a
ge
d
u
ri
ng
t
h
e
bus/
b
at
t
e
ry
com
m
u
t
ati
on
i
s
p
r
esen
ted
in
Fig
ure
32
.
0
0.
0
1
0.
0
2
0.
0
3
0.
0
4
0.
0
5
0.
0
6
0.
0
7
0.
08
0.
0
9
0.
1
-2
0
2
4
6
8
10
Ti
m
e
(
s
e
c
)
Cu
r
r
e
n
t
(
A
)
Ip
v
0
0.
01
0.
02
0.
03
0.
04
0.
05
0.
0
6
0.
07
0.
08
0.
09
0.
1
-1
0
0
10
20
30
40
50
60
Ti
m
e
(
s
e
c
)
Vo
l
t
a
g
e
(
V)
Vo
u
t
Fi
g
u
r
e
2
5
.
The
o
u
t
p
ut
P
V
panel
c
u
rre
nt
wi
t
h
t
h
e
Fi
g
u
r
e
29
.
The
C
u
k c
o
n
v
e
rt
er
out
put
v
o
l
t
a
ge
wi
t
h
t
h
e
occu
ra
nce
of t
h
e t
w
o
di
st
ur
b
a
nces
occ
u
ra
nce
of
t
w
o s
o
l
a
r i
r
r
a
di
ance
di
st
u
r
b
a
nces
0
0.
0
1
0.
0
2
0.
0
3
0.
0
4
0.
0
5
0.
0
6
0.
0
7
0.
08
0.
0
9
0.
1
30
40
50
60
70
80
90
10
0
11
0
T
i
m
(
se
c)
V
o
l
t
a
ge(
V)
Vp
v
0
2
4
6
8
10
12
x 1
0
4
47.
1
986
47.
1
986
47.
1
986
47.
1
986
47.
1
986
47.
1
986
47.
1
986
Ti
m
e
(
s
e
c
)
Vo
l
t
a
e
(
V)
Vb
a
t
Fi
gu
re
2
6
. T
h
e
out
put
PV
p
a
nel
v
o
l
t
a
ge
wi
t
h
t
h
e
Fi
g
u
re
3
0
.
T
h
e
out
put
bat
t
e
ry
v
o
l
t
a
ge
wi
t
h
t
w
o
occ
u
ra
nce
of
t
w
o
sol
a
r i
r
rad
i
ance
di
st
ur
ba
nces
di
st
ur
ba
n
ces d
u
ri
ng
t
h
e
C
h
ar
gi
n
g
Pe
ri
o
d
d
1
d
1
d
2
d
2
d
2
d
1
d
1
d
2
Evaluation Warning : The document was created with Spire.PDF for Python.
I
S
SN
:
2
088
-87
08
I
J
ECE
Vo
l.
6
,
N
o
.
4
,
Au
gu
st
2
016
:
14
21
–
1
433
1
430
0
0.
0
1
0.
0
2
0.
0
3
0.
0
4
0.
0
5
0.
0
6
0.
0
7
0.
08
0.
0
9
0.
1
-1
0
0
0
10
0
20
0
30
0
40
0
50
0
60
0
70
0
Ti
m
e
(
s
e
c
)
Pow
e
r
(
W
)
Pp
v
0
0.
0
1
0.
0
2
0.
0
3
0.
0
4
0.
0
5
0.
0
6
0.
0
7
0.
0
8
0.
0
9
0.
1
0
50
10
0
15
0
20
0
25
0
30
0
35
0
40
0
45
0
Ti
m
e
(
s
e
c
)
P
o
w
e
r(W
)
Po
u
t
Figu
r
e
27
.
Th
e
ou
tpu
t
PV
p
a
n
e
l
po
w
e
r
with
Fi
gur
e 3
1
.
The
C
u
k
con
v
e
rt
e
r
out
p
u
t
t
h
e
occ
u
ra
nce
of
t
h
e t
w
o
di
st
ur
bance
s
p
o
w
er
wi
t
h
t
w
o
di
st
u
r
ba
nces
0
0.
01
0.
02
0.
03
0.
04
0.
05
0.
06
0.
07
0.
08
0.
09
0.
1
-1
0
1
2
3
4
5
6
7
8
Ti
m
e
(
s
e
c
)
C
u
rre
n
t
(A
)
Io
u
t
0
0.
0
5
0.
1
0.
15
0.
2
0.
2
5
0.
3
0.
3
5
0.
4
0
10
20
30
40
50
60
Ti
m
e
(
s
e
c
)
Vol
t
age(
V)
Vb
u
s
Figure
28.
The
C
u
k
conve
r
ter
out
put
c
u
rr
ent
Fi
gu
re
3
2
. T
h
e
o
u
t
p
ut
l
o
a
d
vol
t
a
ge
du
ri
n
g
t
h
e
with
the
occura
nce
of
the
two
dist
ur
bances
bus/
battery
com
m
utation
7.
3.
Si
mul
a
ti
on w
i
th
Ra
pi
d S
t
ep
Ch
an
ges
of
S
o
l
a
r R
a
di
ance
More
ove
r,
in
orde
r
to
prove
t
h
e
efficiency
of
th
e
ANN-MPPT
on
-lin
e
co
ntro
ller,
we
h
a
ve
si
m
u
lated
in
Figur
es
3
3-4
0
a
co
n
ti
n
uou
s
step
in
cr
eases
of
so
la
r
rad
iatio
n
.
Th
e
ANN
con
tro
ller
shows
th
at
it
track
s
con
v
e
n
i
e
nt
l
y
the
m
a
xim
u
m
powe
r
p
o
i
n
t
,
i
n
or
der
t
o
a
voi
d
havi
ng
t
o
m
o
v
e
rapi
dl
y
t
h
e
o
p
erat
i
o
n
p
o
i
n
t
whe
n
t
h
e
sol
a
r
ra
di
at
i
on
i
s
va
ry
i
ng
qui
c
k
l
y
or
w
h
e
n
a
di
st
u
rba
nce
or
dat
a
readi
n
g
er
r
o
r
occ
u
r
re
n
ce
[
1
]
.
T
h
e
fi
gu
re
s
ab
ov
e
m
e
n
tio
ned
sh
ow
t
h
at
th
e
MPP
con
tro
l
track
s
th
e
ch
arg
ing
o
f
th
e
b
a
ck
-up
b
a
ttery
qu
ick
ly
when
the
sha
d
i
n
g
o
f
t
h
e
PV
pa
nel
cha
nge
s
co
nsi
d
era
b
l
y
an
d
q
u
i
c
kl
y
.
It
can
be
se
en
t
h
at
a
hi
g
h
cont
rol
preci
si
on
a
n
d
stab
ility
in
ch
arg
i
n
g
fro
m
th
e
state
o
f
ch
arg
e
o
f
th
e
b
attery
as
d
e
p
i
cted
i
n
Fi
g
u
re
40
,
are
o
b
tain
ed
.
0
0.
01
0.
02
0.
03
0.
04
0.
05
0.
06
0.
07
0.
08
0.
09
0.
1
-1
0
1
2
3
4
5
6
Ti
m
e
(
s
e
c
)
C
u
rre
n
t
(A
)
Ip
v
0
0.
0
1
0.
02
0.
0
3
0.
0
4
0.
05
0.
0
6
0.
07
0.
0
8
0.
0
9
0.
1
-5
0
5
10
15
20
25
30
35
40
Ti
m
e
(
s
e
c
)
V
o
l
t
ag
e(
V)
Vo
u
t
Fi
g
u
re
3
3
.
The
o
u
t
p
ut
P
V
panel
c
u
rre
nt
wi
t
h
Fi
g
u
r
e
37
.
The C
u
k
co
n
v
e
r
t
e
r
out
put
v
o
l
t
a
ge
wi
t
h
a
s
t
ep
cha
n
ge
of
irra
diance
a
ste
p
c
h
ange
of
ir
radi
ance
d
2
d
1
d
2
d
1
d
2
d
1
Bus/Battery
Co
m
m
utation
Evaluation Warning : The document was created with Spire.PDF for Python.