Internati
o
nal
Journal of Ele
c
trical
and Computer
Engineering
(IJE
CE)
Vol.
5, No. 6, Decem
ber
2015, pp. 1275~
1
283
I
S
SN
: 208
8-8
7
0
8
1
275
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 Smart Management Approach
I
n
vesti
g
at
ion for Hyb
r
id
Autonomous Power System
Nasri
Sihem
*
,
Ben Sl
ama S
a
mi**,
Cherif
Adn
a
ne
*
* Department of
Ph
y
s
ics, Faculty
of Scien
ces of
Tunis, Farhat
Hache
d
Uni
v
e
r
sity
,
T
uni
si
a
** Departmen
t
o
f
Information
S
y
stem, King Ab
d
u
l-Aziz
Universi
t
y
, Jedd
ah,
Saud
i Arabi
a
Article Info
A
B
STRAC
T
Article histo
r
y:
Received
May 26, 2015
Rev
i
sed
Ju
l 24
,
20
15
Accepted Aug 10, 2015
A novel design of sy
st
em
m
a
nagem
e
nt
based
on m
u
lti-agent
s
approach
applied to
an au
tonomous hy
br
id
power
s
y
stem is proposed
and in
vestigated.
The s
y
stem under stud
y
integ
r
ates fe
w elements, some serve to provide
power requir
e
ments, while th
e
others
used to
store en
erg
y
. A
m
ong these
ite
ms,
we
c
a
n
me
ntion a
Sola
r Powe
r Sourc
e
na
me
ly
(SPS) whi
c
h works a
s
prim
ar
y
s
ource
to feed a DC elec
tric lo
ad. Th
e s
y
s
t
em
integr
ates
als
o
a
secondar
y
power source namel
y
Power
Recover
y
Source (PRS) based on
a
fuel cell technolog
y
used to co
mpensate
the po
wer deficit if required. Mor
e
than two kinds of energ
y
stor
age,
th
e first c
a
ll
ed H
y
drogen
Generati
o
n
Elem
ent (HGE)
including a water el
ectro
l
y
z
e
r to s
t
ore the
energ
y
in
h
y
drogen fo
rm
, while the s
e
co
nd us
es
an Ultracap
aci
tor El
em
ent (UE) t
o
s
t
ore th
e en
erg
y
in i
t
s
el
ec
tric
al
form
. To r
e
a
c
h
the w
e
ll
funct
i
oning of th
e
s
y
s
t
em
in order to s
a
tis
f
y
the load requir
e
m
e
nts
whatever the
facts
,
a
n
intel
ligen
t en
erg
y
m
a
nagem
e
n
t
approach b
a
sed on m
u
lti-agent
m
odeling is
im
plem
ented an
d verified
. Henc
e, the r
e
li
abi
lit
y
and the eff
e
c
tive
n
es
s
of the
applied management stra
teg
y
,
which allows th
e coordin
a
tion
between
the
differen
t
energ
y
s
ources
and pro
t
ects
the s
y
s
t
em
agains
t
an
y f
l
uc
t
u
ation
,
ar
e
proved b
y
the ob
tain
ed re
sults from Ma
tla
b/Simulink.
Keyword:
Electrolyzer
Fu
el Cell
Hy
dr
o
g
e
n
M
a
nagem
e
nt
Mu
lti-ag
en
t
Solar
Ultracapacitor
Copyright ©
201
5 Institut
e
o
f
Ad
vanced
Engin
eer
ing and S
c
i
e
nce.
All rights re
se
rve
d
.
Co
rresp
ond
i
ng
Autho
r
:
Nasri Sihem
,
Depa
rt
m
e
nt
of
Phy
s
i
c
s,
Faculty of Scie
nces
of T
unis,
F
a
rh
a
t
H
a
ch
ed
U
n
iv
e
r
s
ity,
E
l
Ma
n
a
r, PB
2092, Bel
v
ede
r
e,
Tunisia.
Em
a
il: n
a
sri_
si
h
e
m
@
liv
e.fr
1.
INTRODUCTION
En
vi
ro
nm
ent
a
l
l
y
fri
endl
y
,
t
h
e
rene
wabl
e e
n
e
r
gy
so
u
r
ces p
r
esent
rel
i
a
bl
e a
nd
pr
om
i
s
i
ng sol
u
t
i
o
ns t
o
replace the fossil fuels caus
e
of the
steady increase of
its price along w
ith conce
r
ns about e
m
issions
of
gree
n
h
o
u
se
ga
ses. T
hus
, t
h
e
use
of s
u
c
h
re
newa
bl
e e
n
er
g
y
reso
urces
i
s
becom
i
ng
very
l
a
rge
by
i
n
t
e
g
r
at
i
n
g
t
h
em
i
n
t
o
t
h
e equi
pm
ent
s
us
ed f
o
r st
an
dal
one
o
r
gri
d
-t
i
e
d sy
st
em
s. Ho
weve
r, e
v
e
n
i
t
s
im
port
a
nce,
t
h
ese
so
urces d
ecline o
f
its in
ter
m
i
tten
t
ch
aracteri
s
tic th
at re
qui
r
e
s t
h
e use of e
n
er
gy
st
ora
g
e sy
st
em
s. For t
h
at
, t
o
per
f
o
r
m
t
h
e fu
nct
i
oni
ng
o
f
s
u
ch
sy
st
em
based
on
re
new
a
bl
e ene
r
gy
res
o
u
r
ce, i
t
i
s
ne
cessary
t
o
c
o
m
b
i
n
e
sev
e
r
a
l
k
i
nd
s of
so
ur
ces and
mean
s of
st
o
r
ag
e to ob
tain a
h
ybr
id
system
[
1
].
In
t
h
is
way, in
t
e
g
r
ating
so
lar
p
o
wer so
urce
with
th
e u
s
e of h
ydr
og
en
a
s
en
e
r
g
y
s
t
or
ag
e
sys
t
e
m
,
le
a
d
s
t
o
a no
n-
p
o
l
l
u
t
i
ng rel
i
a
bl
e ener
gy
so
urce
and
red
u
ces
th
e to
tal
m
a
in
t
e
n
a
n
ce co
sts.
In
su
ch
system
,
th
e
hy
d
r
o
g
e
n
can be ge
nerat
e
d vi
a an
electrol
y
zer supplied
by the excess
el
ectrical ene
r
gy issue
d
from the
rene
wa
bl
e ener
gy
so
urce
. The
H
2
gas ca
n t
h
en be st
ore
d
f
o
r f
u
t
u
re
use
by
a fuel
cel
l
(F
C
)
, w
h
i
c
h
w
o
r
k
s as
a
seco
nda
ry
po
w
e
r so
urce se
rvi
ng
fo
r t
h
e ene
r
gy
reco
very
if necessa
ry
[2]. The
com
b
in
atio
n
o
f
Fu
el cell with
th
e u
ltra cap
a
cito
r is an
attractiv
e cho
i
ce d
u
e
t
o
th
ei
r
hig
h
efficien
cy, fast lo
ad-respo
n
s
e, flex
ib
ility and
m
odul
ar st
r
u
ct
ure
f
o
r
t
h
e
use
wi
t
h
ot
her
al
t
e
rnat
i
v
e
s
o
u
r
ce
s suc
h
as
PV
s
y
st
em
s or
wi
n
d
t
u
r
b
i
n
es
[
3
]
.
Hence
,
the prese
n
ce of ultraca
pacitor would
be
ne
eded t
o
e
n
sure
the ene
r
gy st
ora
g
e a
nd t
h
e
energy rec
overy if
Evaluation Warning : The document was created with Spire.PDF for Python.
I
S
SN
:
2
088
-87
08
IJECE
Vol. 5, No. 6, D
ecem
ber
2015 :
1275 –
1283
1
276
necessa
ry
. F
o
r
t
h
at
, t
o
en
su
r
e
t
h
e co
nt
r
o
l
and m
a
nagem
e
nt
o
f
suc
h
a
u
t
o
n
o
m
ous hy
b
r
i
d
p
o
w
er sy
st
em
, an
efficien
t and
reliab
l
e o
p
eratio
n techn
o
l
o
g
y
is requ
ired
. In
th
is con
t
ex
t, the app
licatio
n
of m
u
lti-ag
en
t syste
m
app
r
oach
ap
pe
ars r
e
l
i
a
bl
e an
d
pr
op
er t
o
c
ont
rol
hy
b
r
i
d
sy
st
em
s t
h
i
s
k
i
nd.
T
h
i
s
st
rat
e
gy
o
f
t
h
e sy
st
em
m
a
nagem
e
nt
h
a
s bee
n
ex
pa
n
d
ed
t
o
i
n
cl
u
d
e
seve
ral
a
ppl
i
cat
i
on a
r
eas l
i
ke t
r
a
n
s
p
ort
a
t
i
o
n
,
po
we
r sy
s
t
em
s.
Hen
ce t
h
is tech
no
log
y
is classified
as th
e
m
o
st in
tel
lig
en
t and
effectiv
e to
m
o
d
e
l an
d
t
o
so
lv
e co
m
p
lex
p
r
ob
lem
s
relat
e
d
to th
e con
t
ro
l of system
s [4
].
Seve
ral
researc
h
es an
d
devel
o
pm
ent
wor
k
s h
a
ve bee
n
achi
e
ved
o
n
t
h
e hy
b
r
i
d
p
o
w
er sy
st
em
as wel
l
as th
e strateg
i
es o
f
m
a
n
a
g
e
men
t
an
d
a larg
e nu
m
b
er o
f
p
u
b
licatio
n
s
are n
o
w av
ailab
l
e in
th
e literat
u
re. In
part
i
c
ul
a
r
, co
n
s
i
d
era
b
l
e
resea
r
ch
wo
r
k
s are
i
n
t
e
rest
i
ng
wi
t
h
t
h
e use
of
hy
d
r
o
g
e
n
as a
st
ora
g
e m
e
dium
for
ren
e
wab
l
e en
erg
y
system
s were laun
ch
ed and
ad
d
e
d
t
o
the
literatu
re.
The authors in [5] pres
ent
an approach t
o
op
t
i
m
i
ze a
real
t
i
m
e
dem
a
nd re
sp
o
n
se m
odel
for a
micro
g
r
i
d
w
o
rk
ing
in
t
h
e islan
d
m
o
d
e
. H
e
nce, th
ey propos
ed
m
u
lti-ag
ent strateg
y
at th
e resid
e
n
tial an
d
grid
l
e
vel
f
o
r
i
m
pl
em
ent
i
ng i
t
s
dy
nam
i
c dem
a
nd
res
p
o
n
se
. T
h
e
w
o
r
k
p
r
o
p
o
se
d
by
[
6
]
,
pr
ese
n
t
s
a
n
a
u
t
o
no
m
ous
PV/
F
C
sy
st
em
i
n
orde
r t
o
d
e
t
e
rm
i
n
e opt
i
m
u
m
desi
gn,
cont
rol
st
rat
e
g
y
,
econ
o
m
i
c and
per
f
o
r
m
a
nce of a
PV/FC
h
ybrid
p
o
wer g
e
n
e
ratio
n system
wit
h
ou
t
b
a
ttery sto
r
ag
e tak
i
ng
into
accoun
t all lo
sses i
n
th
e
syste
m
.
Th
is wo
rk
treats a stan
d
a
rd
alg
o
rith
m
to
m
a
n
a
g
e
t
h
e
system
.
In fact, the
ove
rall
efficiency achieve
d
by the
propose
d
work reaches at m
a
xim
u
m
30%. The aut
h
ors in
[7] de
velop a multi-age
n
t syst
em to control of a PV
base
d m
i
crogr
i
d
i
n
cl
u
d
i
n
g s
o
l
a
r
ph
ot
o
v
o
l
t
a
i
c
(PV
)
sy
st
e
m
coupl
ed
wi
t
h
bat
t
e
ry
st
or
age. T
h
ey
t
r
ea
t
an
d
d
e
m
o
n
s
trate the ag
en
ts’ ab
ilities to
islan
d
t
h
e PV-b
ased
m
i
crog
ri
d
in
the
ev
en
t
o
f
an
extern
al fau
lt, secu
re
critical loads,
and re
sync
hronize the
m
i
crogri
d
t
o
t
h
e
m
a
in
gri
d
a
f
t
e
r t
h
e
fa
ul
t
i
s
cl
eare
d
.
The
wo
rk
gi
ven
by
[8]
foc
u
s
o
n
a
dece
nt
ral
i
zed
m
a
nagem
e
nt
sy
st
em
of
a hy
bri
d
a
g
ent
pa
r
a
di
gm
-bas
ed e
l
ectrical syste
m
. The
pr
o
pose
d
sy
st
e
m
i
s
based o
n
t
w
o
rene
wa
bl
e sou
r
ces
whi
c
h
are p
hot
ov
ol
t
a
i
c
panel
a
nd
wi
nd t
u
r
b
i
n
e
gen
e
rat
o
r
wi
t
h
a
bat
t
e
ry
fo
r e
n
er
gy
st
ora
g
e.
He
nce,
i
t
ado
p
t
e
d a
m
a
nagem
e
nt
st
rat
e
gy
t
h
at
ai
m
s
at
qua
nt
i
f
y
i
ng
a
n
d
co
n
t
ro
lling
prod
u
c
tion
sou
r
ces to
ad
ap
t th
e en
erg
y
con
s
u
m
e
d
b
y
co
nsu
m
p
tio
n
sou
r
ces t
o
th
at sup
p
lied
by th
e
sy
st
em
renewa
bl
e p
r
o
d
u
ct
i
on
sou
r
ces
whi
l
e
r
e
duci
ng t
h
e o
p
e
rat
i
ng c
o
st
o
f
t
h
e sy
st
em
. A m
u
lt
i
-
agent
sy
st
em
,
whe
r
e eac
h p
r
od
uct
i
o
n a
nd
con
s
um
pt
i
on s
o
u
r
ce i
s
m
ode
l
e
d by
a
n
age
n
t
,
i
s
p
r
o
p
o
se
d t
o
rep
r
ese
n
t
t
h
e
el
ect
ri
cal
sy
stem
. The aut
h
ors i
n
[
9
]
p
r
op
ose a
n
d de
vel
o
p a m
u
l
t
i
-age
nt
sy
st
em
fo
r di
st
ri
but
i
on
g
r
i
d
congestion m
a
nagem
e
nt with electric
vehicles. They pres
ent a two leve
l hierarc
h
ical control m
e
th
od
for
i
n
t
e
grat
i
n
g E
V
s i
n
t
o
t
h
e di
st
r
i
but
i
o
n net
w
o
r
k t
o
co
or
di
nat
e
t
h
e sel
f-i
nt
er
est
s
and o
p
e
r
a
t
i
onal
con
s
t
r
ai
nt
s of
two
actors, th
e EV
o
w
n
e
r and
Distri
b
u
t
i
o
n
syste
m
o
p
e
rat
o
r
(DSO). Th
en
, th
ey
bu
ilt a
m
u
lti-ag
en
t syste
m
(M
AS
) t
h
at
i
s
base
d
on
t
h
e c
o
-si
m
ul
at
i
on e
nvi
ro
nm
ent
of
JAC
K
, M
a
t
l
a
b
and
Si
m
u
l
i
nk.
Thi
s
pape
r f
o
c
u
s
on
anal
y
s
i
s
and
de
vel
o
pm
ent
o
f
a
hy
b
r
i
d
po
we
r sy
st
em
i
n
t
e
grat
i
n
g a
n
e
w st
rat
e
gy
o
f
system
m
a
n
a
g
e
m
e
n
t
b
a
sed o
n
m
u
lti-ag
ents
m
o
d
e
lin
g
t
h
at respo
n
d
s
t
o
sev
e
ral con
s
train
t
s en
co
un
tered
b
y
th
e work
s cited prev
i
o
u
s
ly as:
1)
R
e
pl
aci
ng t
h
e use o
f
bat
t
e
ri
e
s
by
st
ora
g
e m
eans m
o
re effi
ci
ent
and e
n
vi
ro
nm
ent
a
l
l
y
harm
l
e
ss
l
i
k
e hy
d
r
oge
n.
2)
Using ultracapacitor
de
vice whic
h,
tha
n
ks to
its
fast dyna
m
i
c response, allows m
a
naging the
p
r
ob
lem
o
f
storag
e on
on
e
h
a
n
d
and
on
th
e
o
t
h
e
r settlin
g th
at related to
t
h
e lo
ad
t
r
an
siti
o
n
s
.
3)
Ap
pl
y
i
ng t
h
e
st
rat
e
gy
o
f
sy
s
t
em
m
a
nagem
e
nt
t
o
an
aut
o
nom
ous
hy
b
r
i
d
p
o
w
er sy
st
e
m
t
h
at
i
s
dedicate
d
to re
m
o
te area appl
ication.
4)
The a
d
o
p
t
e
d
cont
rol
st
rat
e
g
y
whi
c
h i
s
b
a
sed
on m
u
l
t
i
-age
nt
can ac
hi
eve t
h
e
f
o
l
l
o
wi
n
g
per
f
o
r
m
a
nce:
a)
Fast
sy
st
em
respo
n
se
t
h
a
nks
t
o
c
o
o
r
di
nat
i
o
n
bet
w
ee
n a
g
ent
s
.
b)
A
better efficiency up t
o
40%
whic
h is
not re
ached by t
h
e c
o
nve
n
tional
control strategies
.
The pa
pe
r has
been
or
ga
ni
zed
i
n
t
h
e fol
l
o
wi
ng
way
.
T
h
e overall hybrid Powe
r system
(
H
PS) and its
b
e
h
a
v
i
or are
describ
e
d
and
detailed
in
Sectio
n
2. S
ection 3
p
r
esen
ts the
m
u
lti-ag
en
t en
erg
y
m
a
n
a
g
e
m
e
n
t
approach accom
p
anied by t
h
e
descri
ption of the
cont
rol algorithm
.
The sim
u
lation
results
ha
ve
bee
n
di
scuss
e
d
i
n
Se
ct
i
on 4,
a
n
d fi
n
a
l
l
y
,
Sect
i
on 5 est
a
bl
i
s
hes
t
h
e concl
u
si
o
n
s.
2.
DESC
RIPTI
O
N OF
THE
WHOLE
H
P
S
The
proposed
hybrid
powe
r s
y
ste
m
(HPS) c
onsists
of
a
fe
w elem
ents tha
t
each
one is
re
prese
n
ted
by
an
a
g
ent (see Figure 1).
T
h
es
e
ele
m
ents are i
d
entifie
d
by:
1)
Sol
a
r P
o
wer
S
o
u
r
ce(
SPS
):
ba
sed o
n
a s
o
l
a
r
panel
wi
t
h
m
a
xi
m
u
m
power
poi
nt
t
r
acke
r
s
(M
PPT
)
and used as a
prim
ary energy
source.
2)
H
y
d
r
og
en
G
e
ner
a
tio
n Elem
en
t (
H
G
E
)
:
ch
ar
a
c
terized
by the
use electrolyzer for H
2
pr
odu
ctio
n
.
3)
Press
u
rize
d
tan
k
fo
r H
2
gas st
ora
g
e.
4)
Power Recove
ry Source (PRS) th
at includes a fuel cel
l with a DC-DC conve
rter for
H
2
u
tilizatio
n
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
J
ECE
I
S
SN
:
208
8-8
7
0
8
A S
m
art
M
a
n
a
g
eme
n
t
A
p
pro
a
c
h
Invest
i
g
at
i
o
n f
o
r H
y
bri
d
A
u
t
o
no
m
o
u
s
P
o
w
e
r Syst
e
m
(
N
asri
Si
he
m)
1
277
5)
Ultracapacitor ele
m
ent
(UE
)
: for
a short
-
time electricity
energy buffer.
Ag
e
n
t
Su
pe
r
v
i
s
o
r
Ag
e
n
t
P
R
S
Ag
e
n
t
U
E
Ag
e
n
t
S
P
S
Ag
e
n
t
H
G
S
Ag
e
n
t
Ta
n
k
Co
n
t
r
o
l
Ex
c
e
s
s
Co
n
t
r
o
l
Ex
c
e
s
s
Co
n
t
r
o
l
De
f
i
c
i
t
Co
n
t
r
o
l
Co
n
t
r
o
l
Co
n
t
r
o
l
Co
n
t
r
o
l
Co
n
t
r
o
l
Co
n
t
r
o
l
Co
n
t
r
o
l
Co
n
t
r
o
l
De
c
i
s
i
o
n
De
ci
s
i
o
n
De
c
i
s
i
o
n
De
c
i
s
i
o
n
Fi
gu
re
1.
Sc
he
m
e
of t
h
e
wh
ol
e HP
S
All th
e rep
r
esen
tativ
e ag
en
ts
are
m
o
n
ito
red
b
y
th
e ag
en
t su
p
e
rv
isor th
at co
n
t
ro
ls and
man
a
g
e
s th
e
system
by the coordi
nation
and
t
h
e com
m
unication be
tween the
different a
g
ents.
Thus, the a
g
e
n
t SPS
co
n
t
ro
ls th
e po
wer co
m
i
n
g
fro
m
th
e so
lar in
o
r
d
e
r to
supp
ly a DC lo
ad
an
d
send
s th
e ex
cess po
wer to
th
e
stora
g
e elem
ents. The st
ora
g
e
agents
eithe
r
t
h
e HGE
or UE are rea
d
y to
store t
h
e exce
ss powe
r
whe
n
they
receive t
h
e
dec
i
sion
order from
the age
n
t supervisor a
f
ter t
h
e re
quirem
ents analysis.
At the sam
e
tim
e
,
bot
h
HGE a
n
d UE
agents c
ont
rol
the inlet power com
i
ng from
the SPS a
nd se
nd the c
h
ec
king res
u
lts to the agent
su
perv
isor.
The b
a
cku
p
system
is in
terv
en
ing
i
n
th
e
d
e
ficit po
wer case.
At th
is m
o
m
e
n
t
, eith
er th
e ag
ent PR
S
or the a
g
e
n
t UE receives the
activation
orde
r from
th
e agent supervisor to
satisfy the loa
d
dem
a
nd.
2.
1. T
h
e
Age
n
t SP
S
The age
n
t SPS
is dedicated to co
ntrol the energy ge
nerate
d by th
e sola
r source.
Indee
d
,
The SPS i
s
com
posed m
a
inl
y
by
a sol
a
r
cel
l
whi
c
h i
s
t
h
e ba
si
c u
n
i
t
o
f
p
h
o
t
o
vol
t
a
i
c
m
odul
es t
h
at
con
v
e
r
t
t
h
e su
n
’
s ray
s
directly into e
l
ectrical energy. He
nce, t
h
e
SPS is m
a
de
up of a
series
and
parallels
ra
nge
of s
o
la
r cells
connected toge
ther to provi
de
the de
sired
outp
u
t
term
in
al vo
ltag
e
an
d curren
t. Ind
e
ed
, the cu
rren
t of a
so
lar
cel
l
i
s
defi
ned
as a f
u
nct
i
o
n
o
f
vol
t
a
ge
[
10]
:
VR
I
SP
S
S
SP
S
(V
R
I
)
V
SPS
S
S
P
S
t
II
-
I
[
e
-
1
]
-
SP
S
P
H
S
R
P
(1)
The a
g
ent
SPS contains a
DC
/DC boost c
onverter w
ith
M
P
PT wh
ich en
ables th
e SPS to
work at th
e
m
a
xim
u
m
pow
er poi
nt
i
n
a hi
ghl
y
fl
uct
u
at
ed
envi
r
o
nm
ent
.
In
dee
d
, t
h
e SE
agent
n
o
rm
al
ly
uses a
m
a
xim
u
m
p
o
wer po
in
t track
ing
(MPPT) tech
n
i
q
u
e
to
co
n
tinuo
usly
d
e
liv
er th
e h
i
ghest p
o
wer throu
g
h
th
e co
nv
erter to
the loa
d
whe
n
t
h
ere
are
va
riations
in irra
diation and tem
p
erature
[11].
Tabl
e
1.
C
h
a
r
act
eri
s
t
i
c
s of
S
O
LAR
E
X M
S
X-
60
Para
m
e
ters
Valu
es
Max
i
m
a
l
Po
wer
:
P
ma
x
60 W
Maxi
m
a
l
Voltage:
V
ma
x
17.
1 V
Max
i
m
a
l
Cu
rren
t:
I
ma
x
3.
5 A
Short-circuit Curre
n
t: I
SC
3.
8 A
Open-circuit Volta
ge: V
OC
21.
1 V
Cell Nu
m
b
ers
N
S
=3; N
p
=6
2.
2. T
h
e
Age
n
t HP
S
The a
g
ent PR
S is characte
r
ized by the
use
of a
pr
ot
o
n
e
x
cha
n
ge m
e
mbra
ne f
u
el
cel
l
(PEM
FC
)
.
I
n
fact, the PRS works as a ba
ckup syst
em that is dedicated to control
t
h
e ene
r
gy
rec
ove
ry
by
co
nv
ert
i
n
g
hy
d
r
o
g
e
n
i
n
t
o
el
ect
ri
ci
t
y
. Thus, t
h
e cal
cul
a
t
i
on
of
t
h
e
hy
d
r
oge
n c
o
nsum
pt
i
on
vari
at
i
o
ns
due
t
o
l
o
a
d
c
h
ange
s
is realized
b
y
Equ
a
tio
n
(3) an
d
th
e to
tal h
y
d
r
og
en
con
s
um
p
t
io
n
is calcu
lated
b
y
th
e su
mmatio
n
o
f
th
e in
stan
t
hy
d
r
o
g
e
n
c
ons
um
pt
i
on val
u
es
[
12]
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
S
SN
:
2
088
-87
08
IJECE
Vol. 5, No. 6, D
ecem
ber
2015 :
1275 –
1283
1
278
N
C
CE
L
L
Q.
I
HP
R
S
PR
S
2.
F
.
2
F
(3)
Tabl
e
2. C
h
ara
c
t
e
ri
st
i
c
s of
P
E
M
F
C
(PR
S
)
Para
m
e
ters
Valu
es
Max
i
m
a
l
Po
wer
L
o
ad
: P
ma
x
40 W
Boost eff
i
ciency
90%
Nu
m
b
e
r
of cells
30
Active Ar
ea per
ce
ll
10 cm
2
Range of oper
a
ting
cur
r
e
nt
0-
20 A
Oper
ating tem
p
er
atur
e
20°
-
80°
2.
3. T
h
e
Age
n
t H
G
E
The age
n
t HGE is a kind of
energy
sto
r
ag
e u
n
it th
at is d
e
d
i
cated
to
cont
rol and ensure
the energy
stora
g
e in the
chem
ical form (Chec
k
i
n
g
hy
dr
o
g
en
p
r
o
d
u
c
t
i
on)
. It
i
s
cha
r
act
eri
zed
by
t
h
e
use
of
a
pr
ot
o
n
m
e
m
b
rane exc
h
an
ge
wat
e
r el
ect
rol
y
si
s w
h
o
s
e
m
a
i
n
fu
nct
i
on i
s
t
o
ge
nera
t
e
hy
dr
og
en
ga
s by
dec
o
m
pos
i
ng t
h
e
water m
o
lecules into hydrogen an
d oxyge
n
. The
HGE use the electri
c curre
nt com
i
ng from
the
SPS to
p
r
od
u
ce th
e
h
y
d
r
og
en
th
at wi
ll b
e
th
en
co
n
t
ro
lled
b
y
th
e ag
en
t tank
in
ord
e
r to
be sto
r
ed
. Usi
n
g
th
e law o
f
Fara
day
,
t
h
e
ra
t
e
of
hy
dr
oge
n
ge
nerat
e
d
by
an el
ect
r
o
l
y
zer
i
s
de
ri
ve
d
fr
o
m
t
h
e el
ect
ri
cal
cur
r
ent
a
n
d i
t
can
be
defi
ned
as
fol
l
ows
[
1
3]
:
.
2.
2
HG
E
N
C
P
F
QI
H
HG
E
F
(4)
Table
3. C
h
ara
c
teristics of
El
ectrolyzer
(HGE)
Para
m
e
ters
Valu
es
Max
i
m
a
l
Po
wer
L
o
ad
: P
ma
x
30 W
Nu
m
b
e
r
of cells
15
Active Ar
ea per
ce
ll
5 cm
2
Range of oper
a
ting
cur
r
e
nt
0-
20 A
Oper
ating tem
p
er
atur
e
60°
-
120°
2.
4. T
h
e
Age
n
t T
a
nk
The m
a
i
n
obje
c
t
i
v
e of t
h
i
s
ag
ent
i
s
t
o
cont
ro
l
t
h
e i
n
l
e
t and t
h
e o
u
t
l
e
t
am
ount
of hy
dr
oge
n
st
ored i
n
a
hi
g
h
pre
ssu
re t
a
nk st
o
r
a
g
e. I
n
fact
, t
h
e am
ount
of H
2
gas
ne
eded t
o
be e
x
p
l
oi
t
e
d by
t
h
e PR
S i
s
di
rect
l
y
issue
d
from
the HGE com
pone
nt bas
e
d on th
e proportionality between t
h
e output
powe
r and the re
qui
red am
ount
of
hy
d
r
o
g
e
n
use
d
by
t
h
e PR
S
com
pone
nt
.
The r
e
m
a
i
n
i
ng hy
dr
o
g
en a
m
ount
(t
he
di
ffe
rence
bet
w
een t
h
e
gene
rat
e
d a
nd
t
h
e used am
ou
nt
of H
2
g
a
s) is tran
sm
i
tted
to
th
e tan
k
storage. In
th
is stud
y, th
e d
y
n
a
m
i
c
o
f
th
e
t
a
nk
st
ora
g
e i
s
obt
ai
ne
d a
s
f
o
l
l
ows
[
14]
:
IN
QR
T
H
T
2
P-
P
Z
TT
i
M
V
H
T
2
(5)
2.
5. T
h
e
Age
n
t UE
The UE a
g
e
n
t
i
s
a ki
nd o
f
ener
gy
st
ora
g
e
uni
t
t
h
at
i
s
dedi
cat
ed t
o
co
nt
r
o
l
t
h
e be
ha
vi
o
r
of t
h
e
ultracapcitor.
Obviously, t
h
e
UE
ha
s
two st
atuses
(cha
rgi
n
g a
n
d disc
ha
rging c
o
rres
p
onding t
o
the
re
ne
wable
energy source
and the loa
d
,
respec
tively).
In fact, whe
n
the power,
s
e
nt from
the SPS and the
PRS, is
in
sufficien
t t
o
su
pp
ly th
e lo
ad
, t
h
e
UE is d
i
sch
a
rg
ed
to meet th
e lo
ad
d
e
man
d
as an energ
y
sup
p
lier.
In the
opposite case,
whe
n
the s
u
pply
from
the SPS excee
ds t
h
e loa
d
dem
a
nd
and the HGE becom
e
s unable to
pr
o
duce m
o
re
H
2
gas
,
the UE
is charged a
nd viewe
d
as the
lo
ad.
Us
ually, there are t
w
o
factors related to the
perform
a
nce of the syste
m
, state of
charge
(SOC
) and the
float charge c
u
rren
t.
Hen
c
e,
th
e SOC is th
e in
d
e
x
whi
c
h
wo
ul
d
p
r
eve
n
t
t
h
e
UE
f
r
om
ove
rcha
r
g
i
n
g
an
d
u
nde
rc
har
g
i
n
g.
It
ca
n
be e
x
p
r
esse
d a
s
f
o
l
l
o
ws
[
1
5]
:
Evaluation Warning : The document was created with Spire.PDF for Python.
I
J
ECE
I
S
SN
:
208
8-8
7
0
8
A S
m
art
M
a
n
a
g
eme
n
t
A
p
pro
a
c
h
Invest
i
g
at
i
o
n f
o
r H
y
bri
d
A
u
t
o
no
m
o
u
s
P
o
w
e
r Syst
e
m
(
N
asri
Si
he
m)
1
279
2
2
ma
x
V
UE
SO
C
UE
V
UE
(6)
The V
UE
and V
UEm
a
x
are defi
ned as t
h
e U
E
vol
t
a
ge an
d
t
h
e USC
m
a
xim
u
m
vol
t
a
ge respect
i
v
el
y
.
Hence
,
t
h
e
U
E
vol
t
a
ge
ca
n
be
cal
cul
a
t
e
d f
r
o
m
Equat
i
on
(
7
)
[
16]
.
1
..
-
.
(
0
)
0
t
DH
VR
I
I
I
d
t
V
U
E
UE
UE
UE
U
E
U
E
C
(7)
Table
4. C
h
ara
c
teristics of
Ul
tracapacitor (UE)
Para
m
e
ters
Valu
es
Maxi
m
u
m
State
of
charge: SOC
ma
x
0.
87
Mini
m
u
m
State
of
charge : SOC
mi
n
0.
05
Resistance: R
UE
25 m
Ω
Capacitance: C
50F
2.
6. T
h
e
Age
n
t L
o
ad
Th
e l
o
ad ag
ent is to
m
a
n
a
g
e
th
es lo
ad
t
o
ma
ke i
t
a co
nt
rol
l
a
bl
e e
n
er
g
y
reso
urc
e
.
In
aut
o
nom
ous
syste
m
, th
e lo
ad
also
p
a
rticip
ates in
th
e co
m
p
etitio
n
to
i
n
fo
rm
o
n
th
e cu
rren
t lo
ad
p
o
wer flu
c
tu
ation
.
3.
AGENT
BASED SYSTEM
FO
R EN
ER
GY
M
ANA
GEM
E
N
T
Our
work
is sp
ecialized
b
y
a n
e
w
en
erg
y
man
a
g
e
m
e
n
t
ap
pro
ach
wh
ich is b
a
sed on
a
m
u
lti-ag
en
t
technique. T
h
is approac
h
ca
n
be cl
assified as an i
n
telligent m
e
thod
us
ed to m
a
nage
the rec
o
very a
n
d the
sto
r
ag
e
o
f
th
e
en
erg
y
.
Ind
e
ed, m
u
lti-ag
en
t syste
m
is a syst
e
m
co
m
p
risin
g
in
tellig
en
t ag
en
ts th
at
wo
rk
t
o
g
e
t
h
er
to
ach
iev
e
a g
l
o
b
a
l go
al. Th
is is a
m
o
st
ex
cit
i
n
g
and
fastest g
r
o
w
i
n
g
research
field
.
Mu
lti-ag
en
t syste
m
h
a
s
a
g
r
eat
p
o
t
en
tial for m
o
d
e
lin
g
of au
t
o
no
m
o
u
s
d
ecision
m
a
k
i
n
g
en
tities, wh
ich
can
b
e
u
s
ed
to
m
o
d
e
l an
d
op
erate
an
off-gri
d
syste
m
. He
nce, t
h
e m
u
lti-agent
m
odeling ca
n be done
by re
presenting
each i
m
portant
elem
ent
in
th
e syste
m
as
a
n
in
tellig
en
t agen
t. In
tellig
en
t ag
en
t p
r
ov
id
es p
l
atfo
rm
to
i
m
p
l
e
m
en
t co
mp
u
t
ation
a
l in
tellig
en
t
to
o
l
s and
m
a
th
e
m
atical
to
o
l
s for d
ecision
mak
i
ng
o
f
th
e
d
i
fferen
t
en
tities [17
]
. In
ad
d
ition
,
In
o
r
d
e
r to
m
o
d
e
l
the m
u
lti-agent
approach
st
ru
ct
ure, we used t
h
e
l
a
ng
ua
ge
A
g
ent
-
UM
L (A
UM
L) whi
c
h
i
s
i
n
spi
r
ed of U
n
i
f
i
e
d
M
odel
i
n
g La
n
gua
ge (
U
M
L
) t
h
at
i
s
speci
fi
ed
by
t
h
e use
of
agent
s
. I
n
fact
,
t
h
i
s
l
a
ng
uage
,
gra
phi
cal
m
odel
i
n
g
,
i
s
dedi
cat
ed t
o
m
odel
an
d de
si
gn sy
st
em
s based o
n
a
g
ent
s
i
n
t
e
ract
i
ons
. I
n
t
h
i
s
case, t
h
e Age
n
t
-
UM
L
bi
d
s
di
ffe
re
nt
ki
n
d
s
of m
odel
i
ng t
o
re
prese
n
t
t
h
e
sy
st
em
and t
h
e i
n
t
e
ract
i
ons
bet
w
ee
n i
t
s
el
em
ent
s
. Am
ong
t
h
em
,
we can cite the state-chart UML diagram
th
at is c
l
assifi
ed as t
h
e
m
o
st
pro
p
er f
o
r
descri
b
i
ng age
n
t
’
s be
havi
or
.
There
b
y
,
t
h
i
s
ki
n
d
o
f
di
ag
ra
m
hel
p
s t
o
un
derst
a
nd t
h
e
p
r
i
n
ci
pl
e
of i
n
t
e
ract
i
ons
bet
w
een ag
ent
s
a
n
d t
h
e
activity of each of them
(Figure 2). He
nce
,
the agents cha
nge thei
r state
fr
om
one to anothe
r relying e
ither on
the ha
ppene
d
events
or on t
h
e r
ecei
ved messages
.
He
nc
e, each a
g
e
n
t
changes its sta
t
e on t
h
e
basis
of t
h
e
interactions tha
t
are
occurre
d
betwee
n the
di
ffe
re
nt a
g
e
n
ts
according t
o
the syste
m
require
m
ents.
Fi
gu
re
3,
prese
n
t
s
t
h
e
UM
L s
e
que
nce
di
a
g
ra
m
t
h
at
i
s
used
t
o
descri
be a
n
d
t
o
re
prese
n
t
t
h
e di
f
f
ere
n
t
interactions be
tween t
h
e syste
m
agents
during t
h
e e
n
ergy
s
t
ora
g
e
an
d rec
ove
ry
eve
n
t
s
. Thi
s
ope
rat
i
o
n
i
s
do
n
e
b
y
co
m
p
aring th
e lo
ad
power
v
a
lu
e
with
th
e SPS
pow
er
on
e. In
ad
d
ition
,
t
h
e su
p
e
rv
isor ag
en
t send
s
fun
c
tion
i
ng
n
a
t
u
re t
o
th
e
HGE ag
en
t and
at the sam
e
t
i
m
e
two
o
t
h
e
r m
e
ssages to
th
e
UE and
PRS ag
en
ts.
Th
e
messag
e
sen
t
to
th
e
HGE agen
t is a requ
est to
sto
r
e
the
energy. T
h
e message se
nt to the PRS age
n
t is a
requ
est to
sup
p
ly th
e en
erg
y
by sp
eci
fy
i
ng t
h
e dem
a
nded
po
wer i
n
t
h
e
defi
ci
t
powe
r
case.
W
h
i
l
e
t
h
e m
e
ssage
sent to t
h
e
UE
agent m
a
y be a
request
t
o
s
u
pply energy in the case
of a
po
wer
rec
ove
ry
o
r
to st
ore
the e
n
er
g
y
in the e
x
cess
powe
r case
[18].
Evaluation Warning : The document was created with Spire.PDF for Python.
I
S
SN
:
2
088
-87
08
IJECE
Vol. 5, No. 6, D
ecem
ber
2015 :
1275 –
1283
1
280
Fi
gu
re
2.
St
at
e
C
h
art
Di
ag
ram
o
f
t
h
e
HE
S
be
havi
or
Fi
gu
re
3.
St
at
e
C
h
art
Di
ag
ram
o
f
t
h
e
HE
S
be
havi
or
To
sum
m
ari
z
e t
h
e a
g
e
n
t
base
d sy
st
em
beha
vi
o
r
,
w
h
ich i
n
terests in the
syste
m
response
way to c
ope
wi
t
h
cha
n
ges;
part
i
c
ul
arl
y
t
h
e deci
si
o
n
s t
a
ken t
o
ens
u
r
e
t
h
e rapi
di
t
y
and t
h
e sm
oot
h sy
st
em
func
t
i
oni
n
g
t
h
r
o
u
g
h
t
h
e co
or
di
nat
i
o
n bet
w
een t
h
e ot
he
r
agent
s
, t
h
e
di
f
f
ere
n
t syste
m
s
t
ates and transitions are
descri
bed a
s
fo
llows:
St
at
e 1:
E
x
ces
s of
power
T
r
ansi
ti
on 1:
SOC
H2
<SOC
M
Ev
ent 1:
H
ydro
g
e
n
pr
odu
ction
HG
E is on (
D
HGE
=1
)
T
r
ansi
ti
on 2:
SOC
H2
>SOC
M
and SOC
UE
<SOC
ma
x
Ev
ent 2:
UE i
s
cha
r
gi
ng
H
G
E
is o
ff (
D
HGE
=0) a
n
d
UE is
o
n
(D
UE
=1)
T
r
ansi
ti
on 3:
SOC
H2
>SOC
M
and SOC
UE
>SOC
ma
x
Ev
ent 3:
Stopp
ing
Stor
ag
e pr
o
cess
HG
E is o
f
f
(D
HGE
=0) and
UE is off
(D
UE
=0).
St
at
e 2:
Deficit o
f
power
T
r
ansi
ti
on 4:
SOC
H2
>SOC
Mn
Ev
ent 4:
Hy
d
r
oge
n
C
o
nsum
pt
i
o
n
PRS
is
on (D
PRS
=1)
T
r
ansi
ti
on 5:
SOC
H2
<SOC
Mn
and SOC
UE
>SOC
mi
n
Ev
ent 5:
UE i
s
di
sc
har
g
i
n
g
PRS is
off
(D
PR
S
=0
)
an
d UE i
s
on
(
D
UE
=1
)
T
r
ansi
ti
on 6:
SOC
H2
<SOC
Mn
and SOC
UE
<SOC
mi
n
Ev
ent 6
:
St
op
pi
n
g
R
eco
very
pr
ocess
P
R
S is off (
D
PR
S
=0) and UE is
of
f (D
UE
=0).
We sh
o
u
l
d
be
not
ed t
h
at
w
e
have o
p
t
i
m
ize t
h
e
syste
m
sizing with the aim
of not
encounte
r
a
pr
o
b
l
e
m
eit
h
er
i
n
st
ora
g
e
or
r
ecove
ry
l
e
vel
s
.
Thi
s
c
a
n
be
al
so
o
n
e
of t
h
e i
m
provem
e
nt
s t
h
at
ha
ve
occ
u
r
r
ed
o
n
the system
.
4.
R
E
SU
LTS AN
D ANA
LY
SIS
B
y
appl
y
i
ng t
h
e st
rat
e
gy
o
f
m
u
lt
i
-
agent
,
t
h
e sy
st
em
beco
m
e
s abl
e
t
o
m
a
nage e
v
e
r
y
p
r
obl
em
due t
o
th
e lack
o
f
electricity o
r
th
at related
to
en
ergy sto
r
ag
e.
In
fact, th
e m
u
lti-ag
en
t algo
rith
m
allo
ws th
e system
to
i
m
p
r
ov
e p
e
rfo
r
man
ce an
d
achiev
e a b
e
tter yield
.
To
d
e
m
o
nstrate th
e reliab
ility an
d
th
e effectiv
en
ess an
d
t
o
sho
w
s t
h
e i
n
n
ovat
i
v
e m
odel
i
ng as
pect
o
f
t
h
e st
udi
ed sy
st
e
m
, we have
us
ed M
a
t
l
a
b–Si
m
uli
nk en
vi
r
o
nm
ent
wh
ich
offers t
h
e p
o
ssib
ility to
d
e
sign
th
e ag
en
t’s
b
e
h
a
v
i
o
r
t
h
rou
g
h
Statefl
o
w m
o
d
e
ling
.
Th
is latter is used
to
m
o
d
e
l co
m
p
lex
syste
m
an
d
it see
m
s su
itab
l
e to
i
m
p
l
e
m
en
t
o
u
r syste
m
. Hen
ce, th
e system is si
m
u
lated
u
s
ing
expe
ri
m
e
nt
al
dat
a
pr
ofi
l
e
s
of
sol
a
r
radi
at
i
o
n,
sol
a
r t
e
m
p
erat
ure a
n
d
user
de
m
a
nds c
o
n
s
um
pt
i
o
n
t
h
at
have
be
e
n
measured and
collected in t
h
ree co
nsec
ut
i
v
e day
s
of t
h
e
m
ont
h of
Jan
u
a
ry. All th
is
data
are ext
r
acted from
Tun
i
sian
m
e
te
o
r
o
l
og
ical d
a
tab
a
se
wh
ich
lead
s
u
s
t
o
ach
i
eve th
e
fo
llowing resu
lts.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
J
ECE
I
S
SN
:
208
8-8
7
0
8
A S
m
art
M
a
n
a
g
eme
n
t
A
p
pro
a
c
h
Invest
i
g
at
i
o
n f
o
r H
y
bri
d
A
u
t
o
no
m
o
u
s
P
o
w
e
r Syst
e
m
(
N
asri
Si
he
m)
1
281
Acco
r
d
i
n
g t
o
t
h
e
obt
ai
ne
d
re
sul
t
s
,
we can
clearly obse
r
ve that the a
g
e
n
t SPS plays t
h
e role
of a
p
r
im
ary so
urce for
feed
ing
a
co
n
tinuo
us lo
ad
.
Ad
d
ition
a
lly
, th
e l
o
ad pro
f
i
l
e u
s
ed
is sim
i
l
a
r to an aud
i
ovisu
al
load i
n
stalled
in a
house
.
T
hus
, t
h
e a
g
ent
SPS contro
ls h
i
s en
tire
power to
b
e
com
p
ared
with th
e lo
ad
dem
a
nds (see
Figure 4).
W
e
can rem
a
rk that, cause of
the
weathe
r condit
i
ons va
riations, som
e
tim
es
the agent
SPS ca
nnot only satisfy the load
dem
a
nd.
While, i
n
the
other case,
we c
a
n see the
SPS
deliveri
ng a
n
exces
s
of
p
o
w
e
r
w
h
i
c
h m
u
st
be m
a
naged
t
h
e
n
st
ore
d
.
From
Fi
gu
re
5,
we ca
n
de
duc
e t
h
e
deci
si
o
n
m
a
de by
t
h
e
a
g
ent
base
d sy
st
em
rel
y
i
ng t
o
s
t
at
us sh
o
w
n
i
n
Fi
g
u
re
4
.
H
e
nce, t
h
e
beha
vi
o
r
o
f
t
h
e
ag
ent
res
p
ons
i
b
le for system
managem
e
nt is analyzed i
n
the ne
xt
su
bsectio
ns.
Figure
4.
Age
n
t SPS c
u
rre
nt c
ont
rol
Figure
5.
Age
n
t HGE a
n
d PR
S curre
nt
profil
e
4.
1. M
a
n
age
m
e
nt of
th
e Defi
ci
t
Pow
e
r
Th
is su
b
s
ection
d
eals
with
t
h
e treatm
e
n
t
o
f
th
e
in
sufficien
cy of th
e en
erg
y
su
pp
lied
by th
e
m
a
in
so
urce
for th
e
syste
m
. Hen
c
e, th
e ap
p
lied
man
a
g
e
m
e
n
t
ap
pro
ach
all
o
ws, in
tellig
en
tly and
rap
i
d
l
y, en
suring
t
h
e rem
a
i
n
ed
m
i
ssi
ng ene
r
gy
wi
t
h
out
a
ffecti
n
g the
operation
of the l
o
ad.
For m
o
re e
xpl
anation,
we t
r
eat
th
e tim
e i
n
terv
al b
e
t
w
een
0
h
an
d 10h
. In
th
is tim
e,
th
e system
declines
of a
deficit of power that
re
q
u
i
r
es
a
ra
pi
d i
n
t
e
r
v
en
t
i
on
of
t
h
e
rec
ove
ry
e
n
er
gy
s
o
u
r
ces.
I
n
t
h
i
s
case,
th
e ag
en
t sup
e
rv
iso
r
h
a
s to
ch
eck th
e en
erg
y
flow
rates
p
r
esen
ting
in th
e tank
and
t
h
e
UE
resp
ectiv
ely.
Accord
ing
to
Fig
u
re 6,
wh
ich
p
r
esen
ts th
e statu
s
of th
e
tank
storag
e,
we can
rem
a
rk
that in
th
e sam
e
in
terv
al
ti
m
e
, th
e tan
k
is e
m
p
t
y wh
ich
m
ean
s th
e i
m
p
o
s
sib
ility o
f
H
2
g
a
s
d
e
liv
eri
n
g
and
cau
ses
th
e d
eactiv
ation
of the
PRS. To
m
a
in
t
a
in
th
e lo
ad
d
e
man
d
s
, th
e ag
en
t su
p
e
rv
isor sen
d
s requ
est to th
e ag
en
t UE in
o
r
d
e
r to
activ
ate it
(see Fi
gu
re
7)
.
Hence
,
t
h
i
s
l
a
t
t
er i
s
used
i
n
i
t
s
di
scha
r
g
i
n
g m
ode
w
h
i
c
h
cau
ses t
h
e
decrea
s
e
of
ene
r
gy
gat
h
ere
d
in
th
e
UE as
well as th
e state
o
f
ch
arg
e
is red
u
c
i
n
g fro
m
its in
itial v
a
lu
e
wh
ich
is 80
%.
In
o
t
h
e
r tim
es
, we can
see t
h
e in
terv
en
ing o
f
t
h
e ag
en
t
PRS to
rectify th
e d
e
ficit th
an
ks to
the
suf
f
i
c
i
e
nt
hy
dr
oge
n st
ore
d
am
ou
nt
.
Fi
gu
re
6.
A
g
e
n
t
Tan
k
st
at
us
c
h
ecke
r
Fi
gu
re
7.
A
g
e
n
t
UE
St
at
us c
h
ecker
4.
2.
M
a
n
age
m
e
nt
of
th
e E
x
c
e
ss P
o
w
er
In
t
h
i
s
s
u
bsect
i
o
n
,
we t
r
e
a
t
t
h
e be
ha
vi
or
of t
h
e sy
st
em
al
on
g
wi
t
h
t
h
e e
x
c
e
ss o
f
e
n
e
r
gy
pr
o
v
i
d
e
d
by
the SP
S a
g
ent.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
S
SN
:
2
088
-87
08
IJECE
Vol. 5, No. 6, D
ecem
ber
2015 :
1275 –
1283
1
282
In th
e i
n
terv
al ti
m
e
b
e
tween 10
h an
d 15h
, th
e syst
em
pr
ovi
des a
n
e
x
ce
ss o
f
p
o
we
r t
h
at
has t
o
be
store
d
. For t
h
at, the tank status checki
ng
step m
u
st be
achieved. He
nce, a
ccording to
t
h
e
Fi
gu
re 6
,
t
h
e t
a
nk i
s
ab
le to
sto
r
e hyd
ro
g
e
n
in
th
is ti
me. Co
n
s
equen
tly, th
e agent based syste
m
turn
on the agent HGE
whic
h can
ex
p
l
ain th
e i
n
crease
o
f
t
h
e
h
y
d
r
og
en
sto
r
ed
q
u
a
n
tity as wel
l
as th
e tank
st
ate o
f
ch
arg
e
(SOC
H2
). I
n
a
d
di
t
i
on,
we can see t
h
at
the syste
m
does not
require t
h
e use
of
t
h
e a
g
ent
UE t
o
m
a
nage t
h
e p
r
obl
em
of ene
r
gy
s
t
ora
g
e
in
th
is stag
e
(see Figu
re 7).
Fo
r th
at, t
h
e ag
en
t UE still in
operab
l
e i
n
th
e case of ap
pro
p
riate tan
k
statu
s
.
Fin
a
lly, th
e
Fig
u
re
8 is
d
e
d
i
cated
to sh
ow th
e
d
ecision
made by t
h
e
agent
supe
rvis
or which i
s
assigne
d t
o
the
activation a
n
d
deactiva
tion
of each system
com
pone
nt.
It ha
s bee
n
ass
u
m
e
d that t
h
e a
g
e
n
t SP
S
rem
a
i
n
ed act
i
v
e du
ri
n
g
t
h
e en
t
i
r
e sim
u
l
a
t
i
on pha
se w
h
i
l
e
t
h
e beha
vi
o
r
of
o
t
her age
n
t
s
va
r
i
es depe
ndi
ng
on t
h
e
sy
st
em
requi
re
m
e
nt
s vari
at
i
o
n. A
d
di
t
i
onal
l
y
, i
n
t
h
e sam
e
fi
gu
re,
we can s
ee t
h
e perf
o
r
m
a
nce p
r
o
v
i
d
e
d
by
t
h
e
sy
st
em
whi
c
h seem
s t
o
be best
co
m
p
ared t
o
t
h
at
usi
ng ene
r
gy
m
a
nagem
e
nt
al
gori
t
hm
ot
her t
h
a
n
m
u
l
t
i
-
agent
strategy. Henc
e, accordi
ng t
o
the gi
ven
res
u
lt, the syst
e
m
adopted by
our work reac
hes
at
m
a
xim
u
m
47% as
effi
ci
ency
w
h
i
l
e i
n
t
h
e sam
e
case of
st
u
d
y
wi
t
h
t
h
e cl
assi
cal
m
e
t
hod
of
m
a
nagem
e
nt
usi
ng si
m
p
l
e
algo
ri
t
h
m
the system
efficiency does
ne
ver excee
d
20%.
Figure 8.
HPS
System
Status
We ca
n
de
duc
e that the
sys
t
e
m
ensures t
h
e c
o
ordination
betwee
n eac
h a
g
ent
according to the
decision value
s
. He
nce, the syste
m
r
eacts ra
pidly against a
n
y fluctuation
or
tra
n
sition
by virtue of the
agents’
co
mm
u
n
i
catio
n
s
t
h
at is clearly d
e
m
o
n
s
trat
ed
and
j
u
stified
b
y
th
e ob
tai
n
ed sim
u
latio
n
resu
lts. For t
h
is,
we
h
a
v
e
p
r
o
v
e
d
that o
u
r
app
lied
man
a
g
e
m
e
n
t
a
p
pro
ach
su
cceed
ed
to
r
e
spond
sm
ar
tl
y to
th
e en
er
g
y
r
e
qu
ire
m
en
ts
i
n
s
u
ch
a
u
t
o
no
m
ous hy
b
r
i
d
p
o
we
r sy
st
em
.
5.
CO
NCL
USI
O
N
In
th
is
p
a
p
e
r,
we propo
sed
an
in
tellig
en
t ap
pro
ach
o
f
energ
y
m
a
n
a
g
e
men
t
b
a
sed
on
m
u
l
ti-ag
en
t
sy
st
em
dedi
cat
ed t
o
c
ont
rol
t
h
e
be
havi
or
o
f
an
of
f-
gri
d
hy
b
r
i
d
p
o
we
r sy
st
em
. The
ai
m
of t
h
i
s
co
nt
r
o
l
st
rat
e
gy
is to
an
alyze th
e
p
r
ob
lem
s
o
f
electricity p
r
o
d
u
c
tion
i
n
tegratio
n fro
m
fluctu
atin
g ren
e
wab
l
e en
erg
y
sou
r
ces,
into the electricity supp
ly. The m
u
lti-agent approach
c
a
n be accom
m
odated with the interaction am
ong
di
ffe
re
nt
ki
nds
of
a
g
ent
s
.
T
h
e
ap
pl
i
e
d st
rat
e
gy
res
p
ect
s
t
h
e
specificity of t
h
e a
g
ent m
o
dule. He
nce, t
h
e
use
of
t
h
ese a
g
ent
s
,
k
eeps t
h
e
opt
i
m
al
beh
a
vi
or
o
f
t
h
e l
o
a
d
.
The
m
odel
of t
h
e st
udi
e
d
sy
st
em
was est
a
bl
i
s
he
d
usi
n
g
MATLAB/Simu
lin
k.
Th
e ob
tain
ed
sim
u
lat
i
o
n
resu
lts p
r
ov
ed th
e reliab
i
lity an
d
feasibilit
y o
f
su
ch
hyb
ri
d
electric syste
m
based
on sola
r-hydrog
e
n
en
ergy
p
r
od
uct
i
o
n dest
i
n
ed t
o
s
t
and-alone application that can be
applied for a
re
m
o
te area.
REFERE
NC
ES
[1]
J. Lagorse
, D. Paire
,
and A. Mir
a
oui, “
A
m
u
lti-a
g
ent s
y
st
em
for energ
y
m
a
nage
m
e
nt of distribut
ed power source
s,
”
Renewab
l
e Ener
gy
, Vol. 35
, pp
.
174–182, 2010
.
[2]
M.
Ca
sta
n
eda
,
A.
Ca
no,
F
.
Jurado,
H.
Sanch
e
z,
and L. M. Fernandez, “Siz
ing op
timization, d
y
namic modeling and
e
n
e
r
gy
ma
na
geme
nt
st
ra
te
gi
e
s
of a stand-
alon
e
PV/h
y
d
rogen/b
a
tter
y
b
a
sed h
y
br
id s
y
stem,”
International Journal
of
Hydrogen Energ
y
,
Vol. 38
, pp
. 3
830-3845, 2013
.
[3]
M. Uzunoglu, O.C. Onar
,
and M
.
S. Alam
, “Mod
eling
,
con
t
rol
an
d simulation of
a PV/FC/UC based h
y
br
id power
generation s
y
stem for sta
nd-alon
e app
lications,”
Renewab
l
e Ener
gy,
Vol. 34
, pp
.
509–520, 2009
.
[4]
M. W
ooldridge
,
“
M
ultiAgent S
y
s
t
em
s”, W
i
l
e
y,
20
02.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
J
ECE
I
S
SN
:
208
8-8
7
0
8
A S
m
art
M
a
n
a
g
eme
n
t
A
p
pro
a
c
h
Invest
i
g
at
i
o
n f
o
r H
y
bri
d
A
u
t
o
no
m
o
u
s
P
o
w
e
r Syst
e
m
(
N
asri
Si
he
m)
1
283
[5]
R. Fazal
,
J. Solanki,
and S. K. Solanki
,
“Dem
and Response
using Multi-ag
ent
S
y
st
em
,”
I
EEE
,
201
2.
[6]
A.
A.
Elbaset,
‘‘Desi
gn, Modeling and Control Strateg
y
of PV/FC Hy
brid Power S
y
stem,”
J. Ele
c
trica
l
Syste
m
s,
Vol.
7,
No. 2
,
pp
. 270-286
, 2011
.
[7]
M. Pipatt
anaso
m
porn, H. Fero
ze,
and S
.
Rah
m
an, “
S
ecuri
ng
criti
ca
l lo
ads in
a PV-based m
i
c
r
ogrid with
a m
u
lti-
agent
s
y
s
t
em
,”
Renewable En
erg
y
,
pp
. 1-9
,
2
011.
[8]
A.
K.
Bodian,
M.
L.
Day
a
,
M. Ndiay
e
b,
and P.
A.
Ndia
y
e
, “Operation Optimal D
y
namics
of a Hy
br
id Ele
c
tr
ic
a
l
S
y
stem
: Mult
i A
g
ent Approa
ch,
”
Proced
ia Comp
uter Science,
Vo
l. 36
, pp
. 454–46
, 2014
.
[9]
J.
Hua,
A.
Saleem,
S.
You,
L.
Nordst
röm, M.
Lind,
and J. Ster
gaard, “A multi-agent s
y
s
t
em for distribution g
r
id
congestion management with
electric vehicles,”
Engine
ering Ap
plicat
ions of Artific
ial Int
e
lli
gen
ce,
Vol. 38, pp. 45–
58, 2015
.
[10]
C. Li, X. Zhu
,
G. Cao, S. Sui, an
d M. Hu, “Dy
n
amic
modeling and sizing optimizati
on of stand-alone photovoltaic
power s
y
s
t
em
s
us
ing h
y
b
r
id
ener
g
y
s
t
orage
t
echn
o
log
y
,”
Renewa
ble Energy,
Vol. 34, pp. 815–826
, 2009
.
[11]
A. A. Elbaset, “Design, Mode
ling and Control Strateg
y
of PV/FC Hy
brid Power S
y
stem”,
J.
E
l
ec
t
r
i
c
al
Sy
st
ems.
Vol. 7
,
No. 2, pp
. 270-286
, 2011
.
[12]
P. M. Sudhakar, and B. M.
Mohan, “Predictive and Optimizing
Energ
y
Manag
e
ment Of P
hotovoltaic Fuel Cell
H
y
brid S
y
s
t
em
s
W
ith S
hort Tim
e
Energ
y
S
t
or
age
,
”
International
Journal of Eng
i
ne
ering Res
e
arch
and Applications
(
I
JERA)
,
Vol. 2,
No. 6, pp. 550-5
56, 2012
.
[13]
O. C. Onar, M. Uz
unoglu, and M.S. Alam
, “
Modeling, control and simula
tion of an autonomous
wind
turbine/pho
tovoltaic/fuel cell/ultr
a-
cap
acitor h
y
br
id power s
y
s
t
em,”
Journal of
Po
wer Sources,
Vo
l. 185
, pp. 1273
–
1283, 2008
.
[14]
T.
Lajn
ef, S. A
b
id,
and A. Ammous,
“Modeling,
Control, and Si
mulation of
a
Solar H
y
drog
en/Fuel Cell H
y
brid
Energ
y
S
y
s
t
em
f
o
r Grid-Connect
ed Applic
ations
,
”
Hindawi Pub
l
ishing Corporation Ad
vances in
Powe
r Electronics,
2013.
[15]
M
.
Hadartz
,
an
d M
.
J
u
lander, “
B
atter
y
-S
uperc
apac
itor Energ
y
S
t
orage,” M
a
s
t
er of S
c
ience T
h
es
is
in Electr
i
c
a
l
Engineering, Department of En
erg
y
a
nd
Environ
m
ent, Division o
f
Electric Po
wer Engineering Ch
almers University
Of Techno
log
y
,
Göteborg, Swed
en, 2008
.
[16]
L. Wei, Z. Xin-jian, C. Guang-
y
i
,
“Modeling and control of a small solar
fuel c
e
ll h
y
brid
energ
y
s
y
s
t
em
,”
Journal of
Zhejiang Univer
sity S
c
ien
c
e A,
V
o
l. 8
,
No
. 5
,
pp
.
734-74, 2007
.
[17]
A. Yah
y
aou
y
,
J. Sabor, H. Gualous, and M.
Lam
r
ini, “
M
odeling And Im
plem
enta
tion O
f
A Multi-Agen
t
Archite
cture Fo
r Intell
igen
t En
erg
y
M
a
nagem
e
nt In An Ele
c
tr
ic Vehi
cle
,
”
Jo
urnal Of Theoretical And
Applied
Information Technology,
Vol 37
,
No. 2
,
2012
.
[18]
A. Yah
y
aou
y
, J
.
Sabor, H. Gualous, and M.
Lam
r
ini,
“
I
nte
l
l
i
gent en
erg
y
m
a
nagem
e
nt b
a
se
d on m
u
lti-age
n
t
approach
in
a h
ybrid vehicle,”
IJCSNS Internatio
nal Journal o
f
C
o
mputer Science and
Network S
e
curity,
Vol. 9,
N
o
.
11, 2009
.
BIOGRAP
HI
ES
OF AUTH
ORS
Sihe
m Nasr
i
was born in
Tunis,
Tunisia, in
1986.
S
h
e rece
ived
the
M
a
s
t
er’s
degre
e
in E
l
ec
troni
cs
from
the F
acult
y of S
c
i
e
nc
es
of
Tunis
(F
S
T
) in
2011. Curren
t
ly
, she is pursuing
the Ph.D. degr
ee in
Electronics
with the Faculty of Scien
ces of
Tunis, in
the
lab
o
rator
y
of Innov
ation
of
communican
t and
coop
erative mobiles (
Innov’Com
), th
e
Higher School o
f
Communication of
Tunis (SUPCOM). Her research
interests in
clude
electrical
power s
y
s
t
em
s
integra
ting
energ
y
s
t
o
r
ag
e dev
i
ces and power
s
y
stem management.
Sami Be
n Slama
rec
e
iv
ed th
e
e
ngineer
, m
a
s
t
er
and doc
torat
e
d
e
grees
,
in
ele
c
tro
n
ics
from
F
acul
t
y
of scien
ces of
Tunis (FST), resp
ectively
in 2005
, 2009 and
2014.
He is assistant p
r
ofessor in King
Abdul-Aziz University
, Jeddah
Saudi Arabia. H
e
fiel
d of interes
t
concerns the p
hotovoltaic power
, en
erg
y
s
y
s
t
em
,
and M
a
t
l
ab
m
odeling
.
Adna
ne
Che
r
i
f
rece
ived the eng
i
neer
, m
a
s
t
er and doctorat
e
degr
ees
from
National Engine
ering
School of Tunis (ENIT), in Tunisi
a. He is a university
teacher
in Electron
i
cs at Faculty
o
f
Sciences of Tun
i
s (FST). He is responsible of
the master specialty
communication s
y
stems and
networks. His f
i
eld of
interest concerns photov
o
ltaic power
s
y
stem, digi
tal sig
n
al pro
cessing,
energ
y
s
y
stem d
e
signing
and modeling
.
Evaluation Warning : The document was created with Spire.PDF for Python.