TELKOM
NIKA Indonesia
n
Journal of
Electrical En
gineering
Vol. 14, No. 3, June 20
15, pp. 376 ~ 3
8
0
DOI: 10.115
9
1
/telkomni
ka.
v
14i3.790
0
376
Re
cei
v
ed Ma
rch 2, 2
015;
Re
vised Ap
ril
26, 2015; Accepte
d
May 1
5
, 2015
An Advanced High Performance Maximum Power Point
Tracking Method with Ant Colony and Particle Swarm
Optimization Method Using Interleaved Boost
Converter
Niv
e
tha V*,
G. Vija
y
a
Go
w
r
i, Elamathy
A
K.S.Rangas
am
y Co
lle
ge of T
e
chno
log
y
, T
i
ruchen
go
de, Indi
a
*Corres
p
o
ndi
n
g
author, e-ma
i
l
: nivetham
ep
e
d
@gma
il.com
A
b
st
r
a
ct
T
o
obtain effic
i
ent maxi
mum pow
er point tracki
n
g
oper
ati
on un
der vary
ing a
nd stea
d
y
state
envir
on
me
ntal
cond
itions w
h
ic
h is base
d
on
Ant Colo
ny
Optimi
z
a
tion (ACO
) com
b
ined with Particle Swar
m
Optimi
z
a
t
i
o
n
(
PSO) that cont
rols a
n
i
n
terle
a
v
ed D
C
-DC c
o
nverter co
nn
ec
ted at the
out
p
u
t of PV arr
a
y
and
ma
inta
ins a co
nstant inp
u
t-po
w
e
r load. T
he
oper
ation
of maxi
mu
m p
o
w
e
r point trackin
g
is to adjust th
e
pow
er int
e
rfac
es so th
at the
oper
atin
g
char
acteristics of t
he l
o
a
d
a
nd th
e ph
otovo
l
taic
array
match
at
th
e
max
i
mu
m p
o
w
e
r poi
nts. T
he simu
lati
on is
base
d
on
the
compute
d
al
go
rithms w
h
ich s
how
very goo
d
persp
ectives.
T
he searc
h
in
g
routin
e is d
o
n
e
in l
e
ss th
a
n
1
5
steps a
nd ta
kes less th
an
5
0
ms. T
h
e track
i
ng
accuracy is b
e
tter than those f
oun
d in co
nven
tiona
l metho
d
s.
Ke
y
w
ords
:
PV
system
, m
a
x
i
m
u
m
power point tra
cking, int
e
rleaved boost
converter
Copy
right
©
2015 In
stitu
t
e o
f
Ad
van
ced
En
g
i
n
eerin
g and
Scien
ce. All
rig
h
t
s reser
ve
d
.
1. Introduc
tion
The in
stallati
on of p
hotovo
l
taic (PV) ge
n
e
rati
on
sy
ste
m
s i
s
rapidly
gro
w
ing
a
s
a
n
ene
rgy
se
curity, a
n
d
ca
uses e
n
vironm
ental i
ssues such
as
global
warmi
ng. PV g
ene
ration
system
is
clea
n an
d e
c
o-frie
ndly sou
r
ce
of en
ergy
and it is
ope
rated eithe
r
in
stand
alon
e or grid
co
nne
ct
ed
mode
s. But the po
we
r–vo
ltage charact
e
risti
c
s of
ph
otovoltaic p
a
nel (PV)
hav
e a no
n-li
ne
ar
output due to
some tem
p
e
r
ature and i
r
radiation o
perating und
er p
a
rtial-sh
ading
conditio
n
s. S
o
,
there i
s
a
n
eed of MPP
T
system
to
sam
p
le
the
output of th
e cell
s a
nd
apply the p
r
oper
resi
stan
ce
(l
oad) to o
b
ta
in maximum
po
wer for
any given
e
n
vironm
ental
co
ndition
s.
The
maximum po
wer p
o
int tra
cki
ng (MPPT
) is the auto
m
atic co
ntrol
algorithm to
adju
s
t the po
wer
interfaces a
nd harve
st the greate
s
t
possible
p
o
we
r, durin
g
variations
of temperatu
r
e,
photovoltai
c
module
cha
r
a
c
teri
stics, ligh
t
level and sh
ading fo
r eve
r
y moment. It has b
e
come
an
essential
co
mpone
nt to evaluate the
desig
n
perf
o
rma
n
ce of photovoltai
c
panel. Maxi
mum
Powe
r Point Tra
cki
ng (MP
P
T) usually is implem
ente
d
with po
we
r electroni
c converte
rs
whi
c
h
act as a
n
interface betwe
en PV array
s
. Usually
a PV system ma
y be a stand
alone
system
or
grid
con
n
e
c
te
d. In the first ca
se it is ne
cessar
y to
kee
p
the output
voltage of the system in t
he
stand
ard
ra
n
ge for
avoidi
ng po
we
r
cut
.
A module o
f
MPPT controller i
s
sho
w
n in Fig
u
re
1
in
whi
c
h it con
s
i
s
ts of a DC-DC interle
a
ve
d
boost converter and a mi
croco
n
trolle
r.
The maximu
m powe
r
poin
t
(MPP) of a
PV m
odule can be dete
c
te
d by a microcontrolle
r
whi
c
h is
driv
en by an MP
PT algorith
m
. Once the M
PP is obtain
ed, a trigg
e
ri
ng sig
nal
with a
spe
c
ific d
u
ty cycle i
s
gen
e
r
ated an
d u
s
ed to tri
gge
r the boo
st co
nverter
swit
ches in o
r
de
r
to
ensure that th
e conve
r
ter o
perate
s
a
s
cl
ose a
s
po
ssib
le to the PV
MPP.
To obtain
an
efficient maxi
mum po
we
r
point tr
a
cki
ng
, two co
nsi
d
e
r
ation
s
can b
e
made,
firstly by developin
g
an im
proved
algo
rithm for
maxim
u
m po
we
r poi
nt tracking
an
d by desi
gnin
g
an efficient in
terleaved b
o
o
s
t conve
r
ter
circuit whic
h is an importa
nt comp
one
nt in the controller.
Many MPPT t
e
ch
niqu
es ha
ve bee
n p
r
op
ose
d
in
the lit
eratu
r
e; exa
m
ples a
r
e th
e
Pertu
r
b
and Ob
se
rve
method
s, Incremental
Con
ducta
nce (IC)
method
s, Fu
zzy Lo
gic
M
e
thod, etc. In th
is
pape
r two m
o
st p
opula
r
o
f
MPPT opti
m
ization
te
ch
nique (Ant colony
an
d
p
a
r
ticle swa
r
m)
are
introdu
ce
d. Ant colo
ny alg
o
rithm o
p
timi
zation
(ACO)
is a
nonlin
ea
r pro
b
lem b
a
sed techniq
ue
not
Evaluation Warning : The document was created with Spire.PDF for Python.
TELKOM
NIKA
ISSN:
2302-4
046
An Advan
c
e
d
High Perfo
r
m
ance Ma
xi
m
u
m Power
Point Tra
ckin
g
Method wit
h
… (Ni
v
eth
a
V)
377
only ensures
the ability to
find t
he gl
obal maximum
power
point
(MPP), but al
so gives
a lower
system
co
st and simpl
e
r
co
ntrol sche
me.
Figure 1. MPPT Controller
PSO is a population ba
se
d stocha
stic optimizat
io
n tech
niqu
e whi
c
h is initiali
ze
d with a
popul
ation of rand
om soluti
ons a
nd search
e
s
for opti
m
a by updati
ng gen
eratio
n
s
.
This
pap
er
p
r
opo
se
s th
e
combi
ned
alg
o
rithm
s
to fin
d
ing MPP
s
d
u
ring
o
scill
ation an
d
power
de
cre
a
se
s
co
nditio
n
. Initially, the ACO
is
em
ployed to t
r
a
c
k the m
a
ximum target v
a
lue
rapidly. And
PSO is empl
oyed to obtai
n the exac
t target value.
The global
MPPs value is
obtaine
d in shorte
st path in DC-DC con
v
erter.
2. Conv
entional Method
For the impl
ementation
of MPPT system
, the choice of the appro
p
ri
ate
DC-DC
conve
r
ter i
s
based on th
ree ba
sic to
p
o
logie
s
of
three differe
nt DC-DC
conv
erters (B
uck
and
Boost
conve
r
ter)
and
MPPT tra
c
ker. T
h
e characte
ri
stics an
d p
r
op
e
r
ties
of
DC-
DC
c
onver
ters
is
esp
e
ci
ally as
rega
rd
s the i
nput imped
an
ce that
they p
r
esent un
der
certai
n ope
ra
ting con
d
ition
s
.
Based o
n
three topolo
g
ie
s the best co
nfiguratio
n is to
be use
d
.
While u
s
in
g buck convert
e
r in MPPT
controll
er it
has a maxi
mum po
wer
point at
minimum irra
diation ho
urs. Also bu
ck
DC–
DC
conv
e
r
ter ha
s a discontinuo
us in
p
u
t current a
n
d
a
contin
uou
s
o
u
tput current
. So, bu
ck topolo
g
y
re
qu
ires a
large
and
expen
si
ve ca
pa
citor to
smooth
the
d
i
scontinu
o
u
s
input cu
rre
nt from
the
p
hot
ovoltaic
mod
u
le a
n
d
to h
a
ndle
sig
n
ifica
n
t
curre
n
t rip
p
l
e
. On
co
ntrary, the b
o
o
s
t conv
erte
r ha
s a
co
ntinuou
s in
put
cu
rrent an
d a
discontin
uou
s output
cu
rre
n
t. But the p
h
o
tovoltaic
cu
rr
ent
in
the bo
ost conve
r
ter is
a
s
smo
o
th as
its ind
u
cto
r
current, with
ou
t any input
ca
pacito
r
.
But t
he ri
pple
cont
ent is high
du
e to the
usag
e
of powe
r
ele
c
tronic
swit
che
s
and in
du
cto
r
.
In conve
n
tion
al metho
d
m
a
ximum po
wer p
o
int tra
cking was a
c
hi
eved by u
s
in
g so
me
popul
ar alg
o
r
ithms li
ke, increme
n
tal condu
ctan
ce
algorith
m
, Perturb-a
n
d
-
ob
serve Alg
o
rit
h
m
(P&O) combi
ned with Parti
c
le Swa
r
m O
p
timization
Al
gorithm (PSO
).But it has the disadva
n
ta
ge
of longer
co
nverge
nce time and failu
re to tr
ack g
l
obal maxim
u
m point, when the pa
n
e
l is
subj
ecte
d to sha
de or cl
ou
dy condition
s. And
both P&O and INC
algorith
m
s a
r
e pron
e to failure
in ca
se
of large chan
ge
s i
n
irradia
n
ce. Some
othe
r
disa
dvantag
e
of c
onve
n
tio
nal pe
rturb a
n
d
observe m
e
thod is, it can produce
oscillations of power output
around the maxim
u
m power poi
nt
even u
nde
r
steady
state
illuminatio
n. The
in
cre
m
ental
con
d
u
c
tance
metho
d
al
so
pro
d
u
c
e
s
oscillation
s a
nd ca
n perfo
rm errati
cally unde
r ra
pidly cha
ngin
g
atmosp
heri
c
con
d
itions.
3. Proposed
Metho
d
In
propo
se
d
system DC-DC conve
r
ter
is
a
interlea
ved boo
st co
nverter top
o
l
ogy for
redu
cin
g
the
rippl
e
cu
rre
nt, it also
i
m
prove
the
reliability of
the sy
stem
and i
n
crea
se
it
s
efficien
cy. The schemati
c
diagram of interleav
e
d
boo
st conve
r
ter i
s
sh
own in Figure 2.
In prop
osed
system, to o
v
erco
me the
difficu
lties in
conve
n
tional
method, pe
rt
urb a
nd
observe
algo
rithm is
repl
aced by Ant
Col
ony Op
timi
za
tion (ACO) al
gorithm,
whi
c
h ca
n track th
e
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN: 23
02-4
046
TELKOM
NI
KA
Vol. 14, No. 3, June 20
15 : 376 – 38
0
378
global
maxim
u
m po
we
r
po
int effectively in
sho
r
te
st
path. Th
e An
t Colo
ny Alg
o
rithm
(ACO) is
combi
ned
with Particle Swarm Optimi
za
tion algorit
hm
(PSO). The ant colony op
timization (A
CO
)
algorith
m
is i
n
spi
r
ed
by re
al ant beh
avior, whi
c
h
i
s
u
s
ed to fin
d
the global
opti
m
al sol
u
tion f
o
r a
nonlin
ear
pro
b
lem. ACO
m
i
mics the fora
ging b
ehavio
r of the ants to achieve o
p
timization
of the
path in
a g
r
a
ph. The
colle
ctive be
havio
rs
of a l
a
rg
e
numbe
r of
a
n
ts form a
p
o
sitive fee
d
b
a
ck
phen
omen
on
and
ant
s initi
a
lly sea
r
ch th
e path ran
d
o
m
ly and lay d
o
wn
phe
rom
o
ne for oth
e
r
a
n
ts
to follow. If a
n
ts find the
hi
gher den
sity
of phe
rom
o
n
e
on the
path
,
then more a
n
ts that travel
on
the same p
a
th, and as a
re
sult, the sub
s
eque
nt ants
will ch
oo
se th
e path.
Figure 2. Sch
e
matic Di
ag
ram of Propo
sed System
Finally, the trail path is foll
owe
d
by most of
the ants u
n
til individual
ants find the
sho
r
test
path throug
h
the exchan
g
e
of such inf
o
rmatio
n.
Initially, the ACO algorithm is
us
ed to tack
le
combi
nation
a
l
proble
m
s.
Figure 3. ACO & PSO Controlle
r for In
terfaci
ng PV Panel DC-DC Conve
r
ter
The ACO i
s
a
combin
ation
of positive feedba
ck mech
anism, di
strib
u
ted co
mputi
ng, and
a gre
edy se
a
r
ch
algo
rithm
.
To sea
r
ch the optimal
solution it has a stron
g
abil
i
ty such a
s
, the
positive fee
d
back me
ch
an
ism en
su
re
s t
hat the ant
colony algo
rith
m is capa
ble
of detectin
g
the
optimal
soluti
on in e
a
rli
e
r
stage. By u
s
i
ng the g
r
e
e
d
y
sea
r
ch the
accepta
b
le
solution i
s
qui
ckly
found an
d efficien
cy of the system is i
m
prov
e
d
. The MPPT pro
b
lem in PV system
s is n
o
w
solved throug
h modified A
C
O
-
ba
sed o
p
t
imization.
Th
e sy
st
em
st
ru
ct
ure i
s
shown in Figure 3.
The flow
cha
r
t of the propo
sed A
C
O-ba
sed M
PPT alg
o
rithm for PV
system
s is
shown in
Figure 4.
The PSO
is
an o
p
timization the
o
ry i
n
spired
by th
e foragin
g
b
ehaviou
r
of
bird
s
and
probl
em
s rel
a
ted to search and o
p
timization,
this phenom
eno
n is suitabl
e
to resolve that
probl
em. PSO is de
rived f
r
om the b
eha
viors of bi
rd
fl
ocking. Li
ke,
a gro
up of bi
rds a
r
e
sea
r
ch
ing
for a foo
d
ra
ndomly in
an
are
a
; inste
a
d
ther
e is o
n
ly one pi
ece of food i
n
the area b
e
i
n
g
sea
r
ched. All
the bi
rd
s
do
not kno
w
wh
ere
the fo
od
i
s
, but t
hey
kn
ow
dista
n
ce
of food fo
r
ea
ch
iteration. The
capabl
e one
is to follow the bird whi
c
h i
s
nea
re
st to the food. In PSO, each si
n
g
le
solutio
n
i
s
a
"
b
ird" i
n
the
search
sp
ace. We
call
it "p
article". All
of
pa
rticle
s
hav
e fitness val
u
es
whi
c
h are evaluated by th
e optimiz
e
d
fitness functio
n
and have v
e
lo
citie
s
which dire
ct the flying
Evaluation Warning : The document was created with Spire.PDF for Python.
TELKOM
NIKA
ISSN:
2302-4
046
An Advan
c
e
d
High Perfo
r
m
ance Ma
xi
m
u
m Power
Point Tra
ckin
g
Method wit
h
… (Ni
v
eth
a
V)
379
of the particle
s
. By following the curre
n
t optimum
pa
rticle
s, the parti
cle
s
fly through the pro
b
le
m
spa
c
e.T
w
o
memory valu
es influen
ce
the moveme
nt of the particles: Pbe
s
t and Gb
est. The
particl
e up
da
tes its velo
ci
ty and po
sitions, afte
r fin
d
ing the t
w
o
best valu
es. Pbest is th
e
individual opti
m
um of particle i; and Gbest is
the swa
r
m or glob
al optimum. Stop tracking if the
stop conditio
n
s are met. Otherwise, the pro
c
e
ss i
s
repeate
d
ag
ain. The sto
p
con
d
itions
are
either lo
catin
g
the global o
p
timum.
The
sea
r
ch e
fficiency a
nd
su
ccess
rate
of PSO are
d
e
termin
ed p
r
i
m
arily by the
values
assign
ed fo
r
the weight
s
and th
e le
arning fa
cto
r
s.
Whe
n
the
weight i
s
too
high, the
pa
rticle
sea
r
ch might
lack accu
ra
cy due to the large step
sizes
Figure 4. Ant Colo
ny Optimization Algo
rit
h
m
Flow Chart
Figure 5. Particle Swa
r
m O
p
timization
Algorithm Flo
w
Ch
art
While fa
cing the multipea
k
values the we
ight
beco
m
e
s
low, particl
e movement be
come
s
slo
w
, and the local optim
um trap mig
h
t be unavoi
dable. Based
on the obje
c
tive function
the
weig
hting is a
l
ways d
e
ci
de
d.
4. Results a
nd Analy
s
is
Figure 6. Output Powe
r of Ordin
a
ry Boo
s
t
Conve
r
ter
Figure 7. Output Powe
r of Interleaved B
oost
Conve
r
ter
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN: 23
02-4
046
TELKOM
NI
KA
Vol. 14, No. 3, June 20
15 : 376 – 38
0
380
The result a
nalysi
s
sho
w
s the trackin
g
time and
a
v
erage
outpu
t powe
r
for
both the
conve
n
tional
boo
st conve
r
ter and inte
rle
a
ved boo
st converte
r. The
tracki
ng tim
e
and averag
e
output po
we
r of the interl
eaved bo
ost
conve
r
te
r i
s
246.4
W
and
tracking time
for 1KW/m2
to
0.4KW/m2 i
s
0.5m
s/10
cycle. The
tra
cki
ng time
and
averag
e o
u
tp
ut po
wer of t
he
conve
n
tio
nal
boo
st conve
r
ter is 24
4W a
nd trackin
g
time
for 1KW/
m2 to 0.4KW/m2 is 1m
s/20
cycle.
By compari
n
g these two t
opolo
g
ies, int
e
rleav
e
d
boo
st conve
r
ter
make the
system more
efficient and f
a
ster.
5. Conclusio
n
Und
e
r rapidly
chan
ging e
n
v
ironme
n
tal conditi
on
s fast
MPPT is re
quire
d, while
under
steady e
n
vironmental
co
ndition
s very
accu
rate
a
n
d
efficient M
PPT without
Oscillation
s are
requi
re
d. The develop
ed
MPPT algorithm is pro
g
rammed in a
micro
c
o
n
trol
ler ba
sed M
PPT
controlle
r and
tested on PV system. Simulation re
su
lt
s sh
ows that the prop
osed
MPPT gives an
averag
e tracking effici
en
cy of 89.
2%.An
advantag
e of
the p
r
o
p
o
s
e
d
MPPT
cont
rolle
r i
s
that it
is
co
st effective and sim
p
le
beca
u
se it doe
s
not re
q
u
ire
current
measurement
s du
ring MP
P
tracking.
Referen
ces
[1]
Ali Bidr
am, Ali
Davou
d
i, Ro
b
e
rt S Balog. C
ontro
l a
nd C
i
rc
uit T
e
chniqu
es
to Mitigate Pa
rtial Sha
d
i
n
g
Effects in Photovoltaic Array
s
.
IEEE Journal of Photovolt
a
ic
s.
2012; 2(4): 5
32-5
46.
[2]
Azade
h Safari,
Saad Mekh
ilef
.
Simulatio
n
an
d Ha
rd
w
a
re Implem
entatio
n o
f
Incremental Con
ducta
nc
e
MPPT
w
i
th Dir
ect Contr
o
l Me
thod Us
in
g C
u
k Conv
erter.
IEEE Transactions on Industrial Electronics
.
201
1; 58(4): 11
54-1
161.
[3]
Bader
N Al
ajm
i
, Khal
ed
H A
h
med, Ste
p
h
e
n
J Fin
n
e
y
, B
a
rr
y W
W
illi
am
s. A Ma
ximum
Po
w
e
r
Poi
n
t
T
r
acking T
e
chniq
ue f
o
r Parti
a
ll
y
Sha
d
e
d
P
hotovo
l
taic
S
ystems in
Micro
g
rids.
IEEE
T
r
ansacti
ons
o
n
Industria
l Elect
r
onics.
20
13; 6
0
(4): 159
6-1
6
0
6
.
[4]
Dezso S
e
ra,
Laszl
o Mathe,
T
a
mas Kerekes, Se
rgi
u
Vio
r
el Sp
ataru, R
e
mus T
eodore
scu. On th
e
Perturb-a
nd-O
b
serve
an
d Inc
r
ementa
l
Co
nd
uctance MPPT
Methods f
o
r P
V
S
y
stems.
IE
EE Journ
a
l
of
Photovo
l
taics.
201
3; 3(3): 107
0-10
77.
[5] Dzun
g N
g
u
y
e
n
, Brad
L
e
h
m
an. An
Ad
aptive
Sol
a
r Photovo
l
taic
Arra
y
Usi
n
g
Mode
l-Base
d
Reco
nfigur
atio
n Algor
ithm.
IEEE Transactions on Industrial Electronics.
2
008; 55(
7): 264
4-26
54.
[6]
Eftichios Koutroulis, Frede
Bl
aab
jerg. A N
e
w
T
e
c
h
n
i
qu
e f
o
r T
r
acking th
e Glob
al Ma
xi
mum Po
w
e
r
Point of PV Arra
y
s
Op
eratin
g Und
e
r Partial-
Shad
ing C
o
n
d
i
t
ions.
IEEE Journal of Photovolt
a
ics.
201
2; 2(2): 18
4-18
9.
[7]
Kashif Isha
qu
e, Z
a
ina
l
Sal
a
m, Muhamma
d Amja
d, Saa
d
Mekh
ilef. A
n
Improve
d
P
a
rticle S
w
a
r
m
Optimizatio
n
(
PSO)–Based
MPPT
for PV
With
Re
duce
d
Stead
y-State
Oscillatio
n
.
IE
EE T
r
ansacti
o
n
s
on Pow
e
r Elect
r
onics.
20
12; 2
7
(8): 362
7-3
6
3
8
.
[8]
Lia
ng-R
u
i C
h
e
n
, Chi
h
-Hu
i
T
s
ai, Yua
n
-Li
Li
n, Yen-Sh
in
Lai.
A Biol
ogic
a
l S
w
a
rm Ch
asi
ng
Algorit
hm f
o
r
T
r
acking the PV Maximum P
o
w
e
r Point.
IEEE Transactions on Energy
Conv
ersion.
20
10; 25(
2): 484-
493.
[9]
Soren
Ba
ekho
j
Kja
e
sr. Eval
u
a
tion
of th
e “H
i
ll
C
limb
i
n
g
” a
n
d
the“Incr
emen
tal C
ond
uctanc
e” Ma
xim
u
m
Po
w
e
r P
o
int T
r
ackers for Ph
otovolta
ic Po
wer S
y
st
ems.
IEEE Transactions on Energy
Conv
ersion.
201
2; 27(4): 92
2-92
9.
[10]
W
e
ido
ng
Xi
ao
, Nathan Oz
o
g
, W
illiam G
Dunfor
d.
T
opolog
y Stu
d
y
of
Photovo
l
taic
Interface fo
r
Maximum Po
w
e
r Point T
r
acking.
IEEE Transactions on Industrial Electronic
s.
2007; 54(
3): 169
6-17
03.
[11]
Ar
y
u
anto S
oet
edj
o. Mod
e
li
ng
of Ma
ximum
Po
w
e
r
P
o
int T
r
ackin
g
Co
ntroll
er
for Solar Pow
e
r S
y
stem
.
T
E
LKOMNIKA Indon
esi
an Jou
r
nal of Electric
al Eng
i
ne
eri
n
g
.
2012; 1
0
(3): 4
19-4
30.
Evaluation Warning : The document was created with Spire.PDF for Python.