Indonesi
an
Journa
l
of
El
ect
ri
cal Engineer
ing
an
d
Comp
ut
er
Scie
nce
Vo
l.
23
,
No.
1
,
Ju
ly
2021
, p
p.
63
~
74
IS
S
N: 25
02
-
4752, DO
I: 10
.11
591/ijeecs
.v
23
.i
1
.
pp
63
-
74
63
Journ
al h
om
e
page
:
http:
//
ij
eecs.i
aesc
or
e.c
om
Experim
ental
d
esign for
an
enhanced p
aram
etr
i
c modelin
g of
super
capacitor e
qu
ivale
n
t circuit
m
odel
Ali M
oh
se
n
A
lsabari
, M.K
Ha
ss
an,
Az
ura CS
,
Rib
ha
n
Z
af
ir
a
Depa
rtment
o
f
E
le
c
tri
c
al a
nd
Ele
ct
roni
c
Eng
ineer
ing,
Univ
ersit
i
P
utra
Ma
lay
sia
,
Serda
ng,
Sel
angor
,
Mal
a
y
s
ia
Art
ic
le
In
f
o
ABSTR
A
CT
Art
ic
le
history:
Re
cei
ved
Ma
r 22
, 202
1
Re
vised Ju
n
9
,
2021
Accepte
d
J
un
1
5
, 202
1
The
m
odel
l
ing
of
the
super
ca
p
ac
i
tor
(SC)
pl
a
ys
an
important
role
for
the
industri
al
appl
i
c
at
ion
with
m
an
y
m
odel
rep
r
ese
nta
ti
ons
such
as
el
e
ct
ri
ca
l
,
che
m
ic
a
l,
and
elec
tro
che
m
ical
m
odel
s.
Am
ong
one
of
those
m
odel
s
are
th
e
equi
va
le
nt
c
irc
u
it
m
odel
which
has
bee
n
used
to
desc
ribe
the
rea
l
-
ti
m
e
(ch
arg
ing
/di
scha
rging)
oper
a
ti
o
n
cha
ra
cteri
sti
c
s
of
the
SC
.
Apart
of
it
s
m
at
hemati
c
al
co
m
ple
xity
,
the
t
i
m
e
-
consum
ing
expe
rimentall
y
is
al
so
a
r
eal
cha
l
le
nge
for
ob
ta
ini
ng
th
e
int
e
r
nal
par
amet
ers
val
ues
for
the
SC
.
Choic
es
of
te
st
equ
ipment
with
a
struc
ture
design
of
experim
ent
al
so
pl
a
y
import
ant
cri
t
eri
a
aff
ec
t
t
he
a
cc
ur
acy
of
the
m
odel
.
Th
is
rese
ar
ch
emphasis
on
a
struct
ure
d
of
ex
per
imental
desig
n
for
SC
m
odel
li
ng
b
y
using
n
e
ware
bat
t
e
r
y
te
ster
.
The
exp
e
rimenta
l
ex
ercise
to
at
t
ai
n
intern
al
par
amet
ers
of
t
he
SC
are
desc
ribe
d
and
di
scuss
ed
in
th
e
p
ape
r.
Th
e
find
in
gs
were
b
enc
hm
ark
ed
wi
th
an
empiric
a
l
m
odel
of
pre
vious
rese
arc
h
ers.
The
te
rm
ina
l
voltage
of
SC
was
val
id
at
ed
via
ex
per
iment
wi
th
m
axi
m
um
rel
ative
err
or
of
0
.
045%.
The
m
ode
l
succ
essfull
y
r
epr
oduce
t
he
SC
d
y
namic
beha
vior
during
the
cha
rg
e/
disch
arg
e
phase
which
ind
i
ca
t
es
the propos
ed
m
et
hod
and m
odel
accurac
y
.
Ke
yw
or
d
s
:
Char
ge/disc
harge
beh
a
viou
r
In
te
r
nal
resist
ance
New
a
re BT
S4000
Param
et
ric
m
od
el
ing
Superca
pacit
or
Term
inal vo
lt
age
This
is an
open
acc
ess arti
cl
e
un
der
the
CC
B
Y
-
SA
l
ic
ense
.
Corres
pond
in
g
Aut
h
or
:
M.K.
Hassa
n
Dep
a
rtm
ent o
f El
ect
rical
an
d
Ele
ct
ro
nic
Eng
ineerin
g
Un
i
ver
sit
i P
utr
a Ma
la
ysi
a
Ser
dang,
Sela
ngor,
43
300
Ma
la
ysi
a
Em
a
il
: kh
ai
r@up
m
.ed
u.m
y
1.
INTROD
U
CTION
Ma
ny
resea
rch
cases
a
re
ca
r
ryi
ng
out
to
di
scov
e
r
a
ne
w
su
sta
ina
ble
powe
r
s
ource
a
ns
we
rs
f
or
conve
ntion
al
powe
r
s
ources
in
va
rio
us
i
ndus
tria
l
a
ppli
cat
ion
s.
T
he
energy
st
or
a
ge
syst
em
s
(ESS
)
is
consi
der
e
d
on
e
of
th
os
e
ar
rangem
ents
as
su
sta
inable
powe
r
so
urces
[
1]
-
[3].
Su
pe
rcapa
ci
tor
is
one
of
the
m
ai
n
com
po
ne
nts
in
ESS
w
hich
ha
ve
t
he
at
trib
ute
s
of
high
power
de
ns
it
y,
lo
ng
l
ife
spa
n
w
hich
can
be
c
ha
rg
e
d
an
d
release
d
in
a
c
ouple
of
sec
onds
[4
]
-
[
6]
.
O
ne
of
t
he
a
dv
a
nc
ed
powe
r
c
om
po
nen
ts
i
n
s
uch
a
ca
pacit
or
is
t
he
el
ect
ric
do
uble
-
la
ye
r
capaci
to
rs
(E
DLC)
wh
i
ch
store
s
the
ener
e
gy
by
thou
san
ds
tim
es
com
par
ed
with
ty
pical
regular
cap
aci
tor
[
7].
It
com
pr
ise
s
of
tw
o
equ
al
plate
s
th
at
hav
e
posit
iv
e
and
ne
gative
charges
isolat
ed
by
protect
or as a
ppeare
d
i
n
Fi
gur
e
1
[
8].
Fund
am
ental
ly,
SCs
sto
re
el
e
ct
rical
energy
thr
ough
the
de
velo
pm
ent
of
the
double
la
ye
r
ca
pacit
or
structu
re
at
the
interface
betw
een
the
a
nodes
and
t
he
el
ect
ro
ly
te
[9
]
.
The
us
es
of
s
up
e
rc
apacit
or
c
a
n
be
var
ie
d
betwee
n
el
ect
r
ic
veh
ic
le
s
(
E
Vs),
s
olar/win
d
po
wer
a
ppli
cat
ion
s
rem
ote
sens
or
node
s
and
hybri
d
el
e
ct
rical
veh
ic
le
s
[
1
0
]
.
Superca
pacit
or
s
hav
e
a
n
ess
entia
ll
y
lower
energy
densi
ty
and
hi
gh
e
r
powe
r
de
ns
it
y
wh
e
n
com
par
ed
wit
h
tradit
io
nal
batte
ries
[1
1
].
Co
m
bin
at
ion
of
thebatt
ery
and
superc
a
pacit
or
has
integ
ral
char
act
e
risti
cs
and
gi
ves
a
n
ast
oundin
g
ar
r
ang
em
ent
that
can
c
over
a
wide
sc
ope
of
intensit
y
an
d
vital
it
y
necessit
ie
s
[1
2
]
.
Evaluation Warning : The document was created with Spire.PDF for Python.
IS
S
N
:
2502
-
4752
Ind
on
esi
a
n
J
E
le
c Eng &
Co
m
p
Sci,
Vo
l.
23
, N
o.
1
,
Ju
ly
2021
:
63
-
74
64
Figure
1.
S
up
e
rcap
aci
to
r
cel
l
s
tructu
re
Accor
dingly
,
upgra
de
of
i
nten
sit
y
qu
al
it
y
util
iz
ing
b
at
te
ries and
supe
rcap
a
ci
tors
is
e
ff
ect
ively
sought
after
i
n
the
fiel
d
of
s
us
ta
inabl
e
powe
r
source
an
d
it
was
s
ho
wn
that
t
his
hybr
i
dizat
ion
has
lowe
r
batte
ry
costs
,
an
overall
inc
r
e
m
ent
in
batte
ry
li
fe
and
higher
e
ff
ic
ie
ncy
w
hic
h
m
ay
no
t
be
fu
l
fill
ed
by
the
sing
le
st
or
a
ge
dev
ic
e
[
1
3
],
[1
4
]
hen
ce
t
he
hy
br
idiza
ti
on
w
as
gro
wn
e
ff
ec
ti
vely
in
nu
m
ero
us
a
ppli
cat
ions
li
ke
batte
ry
el
ect
ric
veh
ic
le
,
hybr
i
d
el
ect
ric
vehi
cl
e
and
unint
errup
ti
ble
po
wer
s
upply
as
pr
ese
nted
i
n
[
1
5
]
-
[1
7
]
.
I
n
ESS
s
app
li
cat
io
ns
,
t
hey
ordi
nar
il
y
us
e
SC
te
rm
i
nal
volt
age
as
the
crit
ic
ism
of
re
gu
la
to
r
to
accom
plish
con
sta
nt
con
t
ro
l
for
th
e
SC
[1
8
]
.
N
on
et
heless,
t
he
SC
te
rm
inal
vo
lt
ag
e
w
ou
l
d
be
in
fluen
ce
d
by
nu
m
erous
el
em
ents,
T
o
viably
assess
t
he
SCs
te
rm
inal
vo
lt
age
a
nd
accom
plish
the
exact
co
ntr
ol
f
or
th
e
SCs
in
t
he
ESS
s,
it
is
pivotal
to
bu
il
d
up
a
pr
eci
se
m
od
el
for
the
SC
[3
]
.
Disp
la
yi
ng
of
SCs
has
happ
ened
to
gr
eat
e
st
sign
ific
ance
wh
e
n
plan
ning
a
nd
dim
ension
in
g
S
C
est
ablishm
e
nts
since
it
is
the
best
ap
proa
ch
to
kn
ow
a
he
ad
of
ti
m
e
abo
ut
t
he
cond
uct and e
xe
cution o
f
t
he vit
al
it
y energ
y
stora
ge devices
when ap
plied t
o
s
pecific
hybr
idiza
ti
on
[
19
].
Con
tr
ol
syst
em
s
of
superca
pa
ci
tor
or
op
e
rati
on
al
bounda
rie
s
an
d
cut
off
po
ints
can
be
acq
uire
d
f
ro
m
a
m
od
el
,
am
p
li
fy
ing
the
li
f
et
i
m
e
of
the
capaci
ty
and
hen
ce
accom
plishing
a
m
or
e
sign
i
ficant
l
evel
of
unwa
ver
i
ng
qu
al
it
y and
co
m
petit
iveness
[2
0
].
N
um
ero
us
S
C
m
od
el
s h
a
ve
b
een
distrib
ut
ed
in the lit
erat
ur
e
for
var
i
ou
s
purpos
es,
inclu
ding
de
scribin
g
el
ect
rical
un
iq
ue
be
hav
i
or,
wh
ic
h
is
of
m
os
t
extre
m
e
sign
ific
an
ce
fo
r
the
industrial
a
pp
li
cat
io
ns
[2
1
]
.
The
m
od
el
s
that
descr
i
be
e
le
ct
rical
beh
av
ior
of
SCs
ca
n
be
cha
racteri
z
ed
in
three
pr
im
ary
cl
asses:
el
ect
ro
c
hem
ic
al
mo
dels
,
intel
li
ge
nt
m
od
el
s
and
eq
uiv
al
e
nt
ci
rcu
it
m
od
el
s.
Th
e
pro
po
se
d
m
odel
s
var
y base
d
on
the
s
pecific app
li
cat
io
n
tha
t
has
bee
n
ad
dr
essed
a
nd
inte
nded
f
or
co
nven
ie
nce
in
that
a
ppli
cat
ion
[
2
2
]
.
The
m
od
el
sh
ould
li
kew
ise
a
void
com
plexity
so
that
it
te
nd
s
to
be
e
ffor
tl
essly
j
oi
ned
into
real
tim
e
con
t
ro
ll
ers
[2
3
]
.
Eq
ui
valent
ci
rcu
it
m
od
el
s
i
m
per
sonat
e
the
el
ect
rical
beh
avi
or
of
SC
s
with
par
am
et
rized
capaci
tors,
in
duct
ances,
a
nd
re
sist
or
s
(R
LC).
They
f
ocus
on
effor
tl
ess
ness,
su
bbi
ng
P
DEs
wit
h
custom
ary
diff
ere
ntial
equ
at
ion
s
(
O
DEs),
wh
ic
h
trem
endo
us
ly
encou
rag
e
their
u
sa
ge
an
d
m
ake
the
m
especial
ly
ap
pr
opriat
e f
or in
dustry a
pp
li
cat
ion i
nvest
igati
on a
nd stu
dies
[2
4
].
Thr
ee
s
ubcat
e
gories
can
be
consi
der
e
d
for
equ
i
valent
ci
rc
uit
m
od
el
s:
RC
m
od
el
s,
tran
sm
issi
on
li
ne
mo
dels,
and
dy
nam
ic
m
od
el
s
[1
6
]
.
M
odel
powe
r
un
der
tra
ns
ie
n
t
a
nd
dif
f
erin
g
co
ndit
ions
is
of
m
os
t
ex
trem
e
sign
ific
a
nce
[
2
5
]
.
Ne
ver
t
heless,
the
diff
e
rent
m
od
el
s
of
s
uperca
pacit
or
stud
ie
d
in
the
li
te
ratur
e
,
ch
oosi
ng
t
he
m
od
el
dep
en
din
g
on
ap
plica
ti
on
of
the
m
od
el
.
In
this
pap
e
r
the
two
bran
ches
m
od
el
wer
e
us
e
d
for
el
ect
rical
beh
a
vior
a
ppli
cat
ion
s
si
nce
i
t
is
con
si
der
e
d
as
one
of
the
m
os
t
widespr
ead
m
od
el
[
1
5
]
.
Eve
n
th
ough
the
chosen
m
od
el
was
st
ud
ie
d
previo
us
ly
,
the
re
are
m
any
ha
r
dw
a
re
t
oo
ls
use
d
for
par
am
ter
s
ide
ntific
at
io
n
a
nd
the
m
et
ho
d
of
identify
in
g
th
e
internal
par
a
t
m
te
rs
of
s
up
e
rcap
aci
to
r
is
m
at
he
m
at
ic
a
lly
com
plex
us
ing
le
as
t
sq
ua
re
m
et
ho
d.
H
ow
e
ve
r,
i
n
this
resea
rch
c
ase
an
e
ff
ic
ie
nt
too
l
of
te
sti
ng
de
vice
(n
e
w
are)
only
will
be
us
e
d
with
pro
po
se
d sim
ple an
d fast
em
pirical
eq
ua
ti
on
s
for para
m
et
er id
entifi
c
at
ion
s
for
m
odel
ing
pur
po
se
.
This
work
em
ph
a
sis
on
the
te
rm
inal
vo
lt
ag
e
dynam
ic
s,
the
cal
culat
ion
of
the
i
nter
nal
pa
ram
et
ers
of
SC.
N
ot
on
ly
that,
bu
t
al
s
o
the
ex
pe
rim
ents
(
pulse
re
la
xation)
f
or
validat
io
n
pur
po
s
es
wer
e
de
sign
e
d
structu
rall
y.
The
rob
us
tne
ss
te
sts
of
the
m
o
del
has
bee
n
be
nch
m
ark
e
d
w
it
h
Ma
tl
ab
m
o
del
and
E
rror
a
naly
sis
validat
io
n.
T
he
fo
ll
ow
i
ng
sect
ion
s
of
this
pa
per
are
orga
ni
zed
as;
m
at
eri
al
s
and
m
et
ho
ds
,
m
od
el
struc
ture,
identific
at
ion o
f
m
od
el
p
a
ram
et
ers,
sim
ulati
on
m
od
el
, result
s w
it
h disc
us
si
on
an
d
c
oncl
usi
on
.
2.
MA
TE
RIA
L
S
AND MET
H
ODO
L
OG
Y
2.1.
Model
structure
The
m
od
el
was
bu
il
t
in
this
pap
er
based
on
two
br
anch
es
ci
rcu
it
wh
ic
h
has
the
fo
ll
ow
ing
structur
e
pr
eci
sel
y
disp
la
ye
d
in
Figu
re
2.
As
in
Figu
re
2
the
m
ai
n
br
anch
cor
resp
on
ds
to
the
im
m
ediat
e
resp
on
se
of
the
Evaluation Warning : The document was created with Spire.PDF for Python.
Ind
on
esi
a
n
J
E
le
c Eng &
Co
m
p
Sci
IS
S
N:
25
02
-
4752
Experime
nta
l
desi
gn for a
n
e
nhance
d para
m
et
ric
m
odel
in
g of s
up
erc
a
pa
ci
tor
… (
Ali
Moh
sen
Als
abar
i)
65
su
per
capaci
tor
du
ring
the
char
ge
or
dischar
ge
even
t
in
the
ti
m
e
ran
ge
of
m
inu
te
s.
In
the
m
ai
n
br
anch
,
R1
is
the series r
esi
sta
nce an
d
rep
resen
ts t
he
waste po
wer
f
or
intern
al
h
eat
in
g
on
ch
arg
ing
an
d
dischar
gin
g
(Ω)
.
Figure
1
.
S
up
e
rcap
aci
to
r
t
wo
br
a
nc
hes
m
od
e
l
[1
3
]
The
vo
lt
ag
e
of
the
s
up
e
rcap
a
ci
tor
m
od
ule
a
nd
ot
her
basic
relat
ion
s
hip
s
de
scribe
d
in
[
6]
an
d
[
7]
ca
n
be
represe
nted
by (1
)
-
(
8)
as
f
ol
lows
:
=
(
1
+
1
)
(1)
Wh
e
re
is
the
num
ber
of
cel
ls
in
series,
is
t
he
num
ber
of
cel
ls
in
par
al
le
l,
Isc
is
the
ch
arg
e/
discha
rge
current.
Since
on
ly
one
sin
gl
e
cel
l
wer
e
m
od
el
e
d
the
num
ber
of
cel
ls
in
pa
rall
el
and
series
was
set
to
one.
The v
oltage
(
1
)
acr
os
s t
he
ca
pa
ci
tor
1
on the
m
ai
n
cel
l can
be
d
esc
ribe
d
as:
1
=
−
0
+
√
0
2
+
2
1
(2)
The
ca
pacit
or
C1
dep
e
nds
on the
vo
lt
ag
e
1
an
d
ca
n be e
xpre
ssed
as:
1
=
0
+
1
(3)
Wh
e
re
C
0
is
th
e
co
ns
ta
nt
ca
pa
ci
ta
nce
in
Fa
r
ads
(F
)
an
d
is
the
c
on
sta
nt
pa
ram
et
er
(F
/V)
.
Wh
e
re
1
is
the
instanta
ne
ous c
harge
of
1
and c
an be cal
c
ulate
d by:
1
=
0
1
+
1
2
1
2
(4)
The
slo
w
br
a
nc
h
determ
ines
the
internal
en
erg
y
distri
bu
ti
on
at
the
en
d
of
the
charge
or
discha
rg
e
cy
cl
e
in
the
tim
e ran
ge of m
inu
te
s.
T
he p
arall
el
r
esi
sta
nc
e R desc
ribes t
he
le
aka
ge
c
urren
t ca
n be
negl
ect
ed
for fas
t
charge
and
discha
rg
e
cy
cl
es.
Co
nc
ern
i
ng
the
slo
w
cel
l,
t
he
volt
age
2
in
the
seco
ndary
ca
pacit
y
C
2
can
be
expresse
d by:
2
=
1
2
⁄
∫
2
ⅆ
=
1
2
⁄
∫
1
2
(
1
−
2
)
ⅆ
(5)
Let
2
the insta
nt
aneous
ch
a
rg
e
of
C
2
, we
hav
e
:
2
=
∫
2
ⅆ
(
6
)
The
c
urre
nt
2
going
i
n
the
sec
onda
ry ca
pacit
or
C
2
re
pr
ese
nte
d by:
2
=
−
1
(
7
)
The
c
urre
nt
1
going
i
n
the
m
ai
n
capaci
to
r
C
2
is
expres
sed
as
:
1
=
1
ⅆ
1
ⅆ
=
ⅆ
1
ⅆ
=
(
0
+
1
)
ⅆ
1
ⅆ
(
8
)
2.
2
.
P
ar
amt
e
ri
c exp
eri
men
ta
l s
trcu
tu
re
The
inter
nal
pa
ram
et
ers
of
the
supe
rcap
aci
tor
are
t
he
res
ist
ance
an
d
ca
pacit
ance
during
t
ran
sie
nt
respo
ns
e
a
nd
t
hey
are
un
def
i
ned.
In
orde
r
t
oid
e
ntify
th
os
e
pa
ram
et
ers,
the
f
ollow
i
ng
ex
per
im
ent
has
be
en
set
as in
Fig
ur
e
3.
Evaluation Warning : The document was created with Spire.PDF for Python.
IS
S
N
:
2502
-
4752
Ind
on
esi
a
n
J
E
le
c Eng &
Co
m
p
Sci,
Vo
l.
23
, N
o.
1
,
Ju
ly
2021
:
63
-
74
66
(a)
(b)
Figure
2
.
The
s
e
fig
ur
e
s ar
e;
(
a) s
c
hem
at
ic
d
i
agr
am
o
f
t
he
e
xp
e
rim
ental
set
up
,
(
b)
e
xperi
m
ental
set up
The
ex
per
im
ental
set
up
was
base
d
on
new
a
re
te
sti
ng
de
vice
with
ei
gh
t
ind
e
pe
nd
e
nt
ch
ann
el
s
a
nd
interm
ediat
e
machine,
host
co
m
pu
te
r
with
th
e
new
a
re
soft
war
e
v7.5.6
a
nd
Ma
tl
ab
2019
a.
The
te
ste
r
ha
s
the
functi
on
of
ch
arg
e
an
d
disc
ha
rg
e
by
a
pply
ing
certai
n
s
of
t
war
e
co
ding
a
lgorit
hm
to
th
e
Ne
war
e
s
of
t
war
e
instal
le
d
in
the
host
c
om
pu
te
r
as
in
Fig
ure
3.
The
ne
wa
re
in
te
rn
al
te
ste
r
ha
s
tw
o
s
witc
hes
to
operate
i
n
e
it
he
r
charge
or
disc
harge
base
d
on
the
data
collected
from
the
current
an
d
vo
lt
age
sensors
ci
rcu
it
s
throu
gh
DAQ
,
so
tra
ns
m
it
s the contr
ol co
m
m
and
s t
hroug
h t
he
s
of
t
war
e
to
the
SC c
onne
ct
ed
to it
.
The
te
ste
r
wor
ks
on
th
e
ra
nge
of
(0
-
10
V
)
a
nd
c
urre
nt
of
(
0
-
6A),
with
sa
m
pl
ing
f
reque
ncy
of
1
H
z
and
m
easur
i
ng
error
s
of
le
ss
than
0.5%.
T
he
m
idd
le
m
a
chine
has
the
functi
ons
of
t
he
f
ollo
wing;
ne
twork
connecti
ons,
r
ecei
vin
g
the
con
t
ro
l
c
omm
ands
f
ro
m
the
host
com
pu
t
er,
c
ontrol
t
he
batte
ry
cy
cl
er
a
nd
up
l
oad
i
ng
the
acqu
i
red
data
f
ro
m
the
real
ti
m
e
exp
e
rim
ent
.
T
he
c
om
pu
te
r
has
t
he
f
unct
ion
of
c
ontr
olli
ng
the
cy
cl
er
thr
ough the
et
he
rn
et
ca
bles
a
nd
sto
re
t
he
data
in
the
s
of
t
war
e
f
ro
m
volt
age
a
nd
c
urr
ent
se
nsors.
T
he
SC
us
e
d
in
this e
xperim
ent is Ma
xw
el
lB
CP1
00
F and its im
po
r
ta
nt p
a
ram
et
ers
are
s
umm
arize
d
in
Ta
ble 1.
Table
1
.
s
uperc
apacit
or p
a
ram
et
ers
Rated
Voltag
e
V
Rated
Cap
acitance
F
Ty
p
ical
E
SRDC
,
Mω
Maxi
m
u
m
L
e
ak
ag
e Cu
rr
en
t
mA
Maxi
m
u
m
Peak
C
u
rr
en
t
A
2
.7
100
8
0
.26
61
2.
3
.
Pr
opose
d metho
d
of
p
arameters
ide
nt
ific
ati
on
The
inter
nal
pa
ram
te
rs
of
the
su
pe
rca
pacit
or
re
pr
ese
nt
the
resist
ance
a
nd
capaci
ta
nce
in
charge
a
nd
discha
rg
e
phas
es
as
tw
o
bran
ches
el
em
ents.
The
basic
no
nlinear
relat
io
ns
hi
p
betwee
n
cha
rg
e
,
c
urre
nt
an
d
vo
lt
age
ca
n be
expresse
d
[6]
a
s foll
ow
:
=
ⅆ
ⅆ
∗
ⅆ
ⅆ
=
(
+
)
ⅆ
ⅆ
(9)
In
te
ger
ti
ng
e
qu
at
ion
with
respec
t t
o
ti
m
e
:
∫
0
ⅆ
=
∫
(
+
)
ⅆ
ⅆ
0
ⅆ
(10)
Since
the
curr
e
nt is c
on
sta
nt yi
el
ds
:
=
(
)
=
+
1
2
1
2
(11)
Since
the
tw
o
-
br
a
nc
h
m
od
el
hav
e
t
wo
ca
pa
ci
ta
nces
in
(
11
)
we
can
as
sum
e
that
the
values
of
th
os
e
co
ns
ta
nts
represe
nted fr
om
(
11
)
by
:
C
1
=
(12)
C
2
=
2
∗
(13)
So
rear
ra
ng
i
ng
in
(
11
)
-
(
13
)
yi
el
ds
the
foll
owing:
Evaluation Warning : The document was created with Spire.PDF for Python.
Ind
on
esi
a
n
J
E
le
c Eng &
Co
m
p
Sci
IS
S
N:
25
02
-
4752
Experime
nta
l
desi
gn for a
n
e
nhance
d para
m
et
ric
m
odel
in
g of s
up
erc
a
pa
ci
tor
… (
Ali
Moh
sen
Als
abar
i)
67
=
C
1
+
C
2
2
(14)
Ba
sed
on
(
14
)
the
volt
age
du
r
ing
c
ha
rg
i
ng
phase
has
tw
o
unkn
own
c
oe
ff
i
ci
ent
C
1
a
nd
C
2
.
Ba
se
d
on
vo
lt
age
data
obta
ined
e
xp
e
ri
m
ental
ly
the
va
lues
of
these
two
coe
ff
ic
ie
nt
s
C
1
an
d
C
2
c
an
be
c
om
par
e
d
with
(
14
)
as
dis
play
ed
in
F
i
gure
4
.
Fr
om
figure
t
her
e
is
m
inu
es
sign
on
the
disp
la
ye
d
eq
uatio
n
in
Fi
gure
4
a
nd
t
he
capaci
ta
n
ce
ca
nnot
negat
ive
r
at
her
t
han
c
onf
irm
the
relat
ion
sh
i
p
physi
cal
ly
,
so
that
t
he
t
wo
points
ha
ve
bee
n
sel
ect
ed
f
or ob
ta
ining
t
he val
ues of C
1a
nd
C2 as i
n
Fi
gure
5
.
Figure
3
.
Ex
pe
rim
ental
ch
arg
i
ng ph
a
se
d
at
a
Figure
4
.
Ex
pe
rim
ental
ter
m
i
nal volt
age
d
at
a of
su
pe
rca
pacit
or
Ba
sed
on
Fi
gure
5,
t
wo
point
s
of
c
hargin
g
phase
we
re
sel
e
ct
ed
to
cal
cula
te
the
values
of
i
m
m
ediate
br
a
nc
h
du
rin
g
chargin
g
phase
.
The
sel
ect
ed
po
i
nts
we
re
ba
sed
on
the
c
ha
ra
ct
erist
ic
s
of
SC
sp
eci
fie
d
on
data
sh
eet
that
the
pr
ese
ntati
on
of
capaci
ta
nces
c
an
be
picke
d
a
t
0.
5,
0.9
5
of
S
C
rated
volt
ag
e
or
0.4,
0.8
of
rated
vo
lt
age
.
Ba
se
d
on
(
14
)
,
t
he
t
wo
points
in
Figure
5
m
us
t
be
eq
ual
since
th
ey
are
in
th
e
sa
m
e
chargin
g
phase
a
s
fo
ll
ow:
t
1
=
C
1
1
+
C
2
1
2
(15)
t
2
=
C
1
2
+
C
2
2
2
(16)
By
s
olv
in
g
i
n (
15
)
a
nd
(
16
)
yi
el
ds
:
1
=
[
1
∗
2
−
2
∗
1
1
∗
2
2
−
2
∗
1
2
]
∗
ℎ
(17)
2
=
[
1
∗
2
−
1
∗
1
1
∗
2
2
−
2
∗
1
2
]
∗
ℎ
(18)
So
ba
sed
on
t
he
tw
o
points
obta
ined
,
the
va
lues
of
1
is
32.7604
F
a
nd
2
44
.
9576
F
.
T
he
resis
ta
nce
R1
is
cal
culat
ed base
d on
:
R
1
=
D
v
I
c
h
arg
e
(19)
D
v
is t
he discha
r
ge
drop
volt
age
in the fi
rst
10
m
s o
f
the
vo
lt
a
ge discha
rg
e
r
e
pr
ese
nted
as
fol
low
s:
D
v
=
V
1
−
V
2
t
1
−
t
2
(20
)
R
1
ob
ta
ine
d
a
s
0.0
097 O
hm
, s
o
t
hat the
only
p
a
ram
et
er left is R
2
by apply
ing fit
ti
ng
c
urve
a
pp
li
cat
yi
on in
m
at
la
b
fo
r
ter
m
inal vo
lt
age
durin
g rest t
he
expo
nen
ti
al
fu
nction ca
n b
e e
xpreese
d
in
(2
1).
T
he op
e
n
ci
cuit
vo
lt
age
(OCV
)
dur
i
ng d
isc
harge in
the
seco
nd
ph
a
se
disp
la
ye
d
in
F
ig
ur
e
6.
=
∗
−
(21)
Fr
om
F
ig
ur
e
6 t
he
Ma
tl
ab
fitt
ing cu
r
ve
e
qu
at
ion
with RM
S
E er
ror of
0.00
8204V
r
e
pr
e
nt
ed
as:
=
0
.
5017
∗
−
0
.
04
13
9
t
(22)
Evaluation Warning : The document was created with Spire.PDF for Python.
IS
S
N
:
2502
-
4752
Ind
on
esi
a
n
J
E
le
c Eng &
Co
m
p
Sci,
Vo
l.
23
, N
o.
1
,
Ju
ly
2021
:
63
-
74
68
So
t
hat
the m
od
el
w
a
s obtai
ne
d
a
nd and tim
e co
ns
ta
nt
(
2
)
obta
ined by c
om
par
in
g
(
22
)
an
d
(
21
)
, so
R
2
can
be
cal
culat
ed
ba
s
ed
on
t
he
f
ollow
i
ng
(
23).
T
he
al
l
obta
ine
d
inter
nal
para
m
et
ers
cal
culat
ed
an
d
ta
bul
at
ed
in
Table
2.
2
=
2
∗
2
(23)
2
=
2
2
(24)
Table
2
.
Id
e
nti
ficat
ion
res
ults o
f
the
tw
o
-
br
a
nch
equ
i
valent circ
uit m
od
el
1
O
h
m
1
F
2
2
Oh
m
Kv
0
.00
9
7
3
2
.76
0
4
4
4
.95
7
6
0
.00
0
8
1
0
.92
0
1
Figure
5
.
T
erm
inal v
olatge
du
rin
g discha
rg
e
curve
2.
4
.
Sim
ulat
i
on
m
od
el
Af
te
r
ob
ta
ini
ng
the
identifie
d
par
am
et
ers
usi
ng
the
te
st
procedu
res,
the
s
i
m
ulati
on
m
od
el
w
as
bu
il
t
via
Ma
tl
ab
Sim
ul
ink
.
Li
ne
transm
issi
on
m
od
el
was
us
ed
to
be
c
om
par
ed
with
t
he
m
od
el
pro
po
se
d
in
th
i
s
pap
e
r
a
nd
fina
ll
y
the
resu
lt
was
ver
ifie
d
with
ex
pe
rim
e
ntal
te
rm
inal
vo
lt
age
data
of
SC
.
T
he
tw
o
-
branc
h
m
od
el
s,
li
ne
transm
issi
on
m
od
el
an
d
m
easur
ed
vo
lt
age
a
r
e
sh
ow
n
in
Fi
gure
7.
D
if
fere
nt
ex
per
im
ents
with
d
iffe
re
nt
cu
rr
e
nt
pr
of
il
e
we
re
us
e
d
in
sim
ula
ti
on
to
valid
at
e
the
the
tw
o
-
branch
m
od
el
.
T
he
detai
le
d
m
od
el
of
the tw
o
-
br
a
nc
h m
od
el
u
se
d
in
si
m
ulati
on
is
di
sp
la
ye
d pr
eci
s
el
y i
n
Fig
ur
e
8.
Figure
6
.
Mat
la
b
m
od
el
and
two bra
nch m
od
el
sim
ulati
on
The
tw
o
-
branc
h
m
od
el
was
base
d
on
the
basic
pri
nciple
of
Kirc
hhoff
la
ws
us
i
ng
(1)
-
(
8).
The
num
ber
of
par
al
le
l
and
s
eries
cel
ls
was
set
to
on
e
f
or
the
pu
rpos
e
of
cel
l
inv
e
sti
gating
m
od
el
.
The
m
a
them
at
ic
a
l
represe
ntati
on
in
(
1)
-
(
8)
were
us
e
d
t
o
cal
c
ulate
the
te
rm
i
nal
volt
age
as
m
a
in
pa
ram
eter
s
of
supe
rca
pacit
or
m
od
el
.cu
rr
e
nt
prof
il
e
data
w
ere
im
ple
m
ent
ed
in
sim
ulatio
n
ba
sed
on
extracte
d
ex
pe
rim
ental
data
from
new
a
re.
Also
t
erm
inal
vo
lt
age
data
from
exp
erim
ents
expor
te
d
to
w
orks
pace
in
Ma
tl
ab
so
that
it
will
b
e
us
e
d
Evaluation Warning : The document was created with Spire.PDF for Python.
Ind
on
esi
a
n
J
E
le
c Eng &
Co
m
p
Sci
IS
S
N:
25
02
-
4752
Experime
nta
l
desi
gn for a
n
e
nhance
d para
m
et
ric
m
odel
in
g of s
up
erc
a
pa
ci
tor
… (
Ali
Moh
sen
Als
abar
i)
69
in Mat
la
b
Si
m
ulink
for
m
od
el
v
al
idati
on
and co
m
par
iso
n
of
s
up
e
rca
pacit
or. Th
e sim
ulatio
n
m
od
el
in
Figure
7
and
8
us
e
d
di
ff
e
ren
t
pro
file
s
fo
r
diff
e
re
nt
validat
ion
pu
rpose
w
hich
will
be
hig
hli
gh
te
d
an
d
disc
us
se
d
in
resu
lt
s a
nd
discuss
io
n.
Figure
7
.
S
ub
s
yst
e
m
Si
m
ulin
k for s
uperca
pa
ci
tor
tw
o br
a
nc
h
m
od
el
3.
RESU
LT
S
The
m
od
el
v
al
idati
on
m
ai
nly f
ocused
on
the t
erm
inal vo
lt
age o
f
supe
rcp
aci
tor
as r
e
gu
la
to
r
p
aram
e
te
r
for
co
ntr
oller
desig
n
purpo
ses.
Di
ff
e
ren
t
current
prof
i
le
us
ed
t
o
te
sti
fy
the
m
od
el
validai
ty
with
the
com
apr
ison
of
Ma
tl
ab
bu
il
t
i
n
m
od
el
and
m
easur
ed
volt
age
data.
T
he
first
te
st
us
ed
t
o
cha
rg
e/
disch
arg
e
th
e
su
pe
rca
pacit
ors for
t
wo
re
pea
te
d
cy
cl
es of
re
st
-
cha
rg
e
-
disc
ha
rg
e
as s
how
n i
n
Fig
ure
9.
(a)
(b)
(c)
Figure
8
.
The
s
e figure
s ar
e;
(
a)
c
urren
t
pro
file
(
b) s
uperca
pa
ci
tor
e
xp
e
rim
ental
te
rm
inal
vo
lt
age
data
,
si
m
ulati
on
tw
o b
ran
c
h
m
od
e
l
and
sim
ulati
on
of li
ne
tra
ns
m
i
sion m
od
el
(c)
error bet
ween
m
easur
ed
and
si
m
ulate
d
m
od
el
s
Evaluation Warning : The document was created with Spire.PDF for Python.
IS
S
N
:
2502
-
4752
Ind
on
esi
a
n
J
E
le
c Eng &
Co
m
p
Sci,
Vo
l.
23
, N
o.
1
,
Ju
ly
2021
:
63
-
74
70
The
c
urre
nt
pro
file
in
Fi
gure
9
(
a
)
was
s
et
in
ne
wa
re
dev
ic
e
with
c
om
plete
cy
cl
e
s
of
(
rest
-
charge
/
disc
harge
-
rest)
f
or
t
he
pur
pose
of
s
up
e
rca
pacit
or
te
rm
inal
vo
lt
ag
e
validat
io
n.
T
he
sim
ulati
on
sta
rted
with
pe
rio
d
re
st
of
50
seco
nds
to
chec
k
se
lf
-
disc
harge
of
su
pe
rcap
a
ci
to
r
with
init
ia
l
vo
lt
age
m
easur
e
m
ent.
The
n,
f
ollo
we
d
by
30
se
co
nd
s
of
co
ns
ta
nt
cha
rg
i
ng
phase
un
ti
l
it
reached
t
he
rat
ed
volt
age
of
2.7V.
Af
te
r
that,
the
SC
was
in
rest
f
or
30
sec
onds
f
ollow
e
d
by
on
e
m
inu
te
co
ns
ta
nt
disch
ar
ge
phase
.F
ur
t
he
rm
or
e
the
sam
e
tim
e
fr
am
e
and
phases
for
the
fi
rst
cy
cl
e
was
rep
eat
e
d.
I
n
F
igure
9
(
b
)
the
te
r
m
inal
vo
lt
age
of
su
pe
rca
pacit
or
(
durin
g
rest
-
c
hargin
g
/
disc
ha
rg
i
ng
)
phase
s
for
two
cy
cl
es
was
com
par
ed
with
ex
per
i
m
ental
curve
data.
T
he
init
ia
l
vo
lt
a
ge
was
ab
out
0.442
7
V
si
nce
s
up
e
rca
pacit
or
cannot
reach
ed
to
ze
r
o
vo
lt
ag
e
w
hic
h
confirm
s
the
s
el
fd
isc
ha
r
ge
prop
e
rty
of
supe
rcap
aci
to
r
.
In
F
igure
9
(
b
)
by
loo
ki
ng
to
the
do
tt
ed
re
d,
bl
ue
an
d
black
li
nes
w
hi
ch
re
pr
ese
nt
e
xp
e
rim
ental
data,
two
br
a
nc
h
si
m
ulati
on
m
od
el
m
od
el
res
pecti
vely
,
it
is
cl
ear
that
al
l
m
od
el
s
ha
ve
sim
i
la
rit
y
in
tren
d
but
diff
e
re
nt
prece
ci
sion
.
The
rel
at
ive
erro
r
wa
s
show
n
i
n
F
i
gure
9(
c
)
.
for
m
easur
ed
and
sim
ulate
d
te
rm
inal
vo
lt
a
ge
of
t
wo
bra
nch
an
d
a
nd
li
ne
tra
ns
im
isi
o
n
(Mat
la
b)
m
od
el
.
Of
su
pe
rca
pacit
or.
The
tw
o
br
a
nc
h
m
od
el
stud
i
ed
in
t
his
pa
pe
r
ha
ve
a
m
axim
u
m
relat
ive
error
of
0.0
8
V
wh
il
e
the
li
ne
transi
m
isi
on
(Mat
la
b)
m
od
el
reac
hed
up
to
0.1
V.
Ba
se
d
on
r
el
at
ive
err
or
va
lues,
the
tw
o
br
a
nc
h
m
od
el
hav
e
bette
r
te
rm
ina
l
vo
lt
age
res
pons
e
a
nd
f
ol
lowed
m
easur
ed
data.
A
nothe
r
par
m
at
er
of
charge/disc
harge
validat
io
n
of
superca
pacit
or
is
load
fluctu
at
ion
f
or
fast
charge
an
d
disc
harge
as
disp
la
ye
d
in
Figure
10.
(a)
(b)
(c)
Figure
9
.
The
s
e figure
s ar
e
;
(
a) c
urren
t
pro
file
,
(
b)
superca
pacit
or
ex
pe
rim
ental
te
r
m
inal vo
lt
age
da
ta
,
si
m
ulati
on
tw
o b
ran
c
h
m
od
e
l
and
sim
ulati
on
of li
ne
tra
ns
m
i
sion m
od
el
,
(c
)
erro
r betwee
n m
easur
ed
and
si
m
ulate
d
m
od
el
s
As
sh
ow
n
in
Figu
re
10
(
a
)
,
the
cur
ren
t
pr
of
il
e
us
ed
was
by
init
ia
ti
ng
the
su
per
capaci
tor
to
rest
fo
r
20
secon
ds
and
then
fo
ll
ow
ed
by
char
ge
and
dischar
ge
fo
r
30
secon
ds
in
each
ph
ase
of
char
gin
g
and
dischar
gin
g.
I
n
Figu
re
10
(
b
)
it
is
cl
early
sh
ow
n
that
du
ring
char
gin
g
and
dischar
g
ing
ph
ase
the
te
rm
inal
vo
lt
age
com
pr
om
ise
d
the
m
easur
ed
vo
lt
age
data
of
te
rm
inal
vo
lt
age
of
su
per
capaci
tor
.
As i
n
Figu
re 1
0
(
c
)
, th
e
m
axim
um
r
el
at
ive
err
or
f
or
Ma
tl
ab
m
od
el
was
up
to alm
os
t
0.
07
5
V
wh
il
e the
two
-
br
anch
m
od
el
inv
est
igate
d
in
this
stud
y
was
u
p
to
al
m
os
t
0.
04
V
.
Fr
om
the
resu
lt
s
ob
ta
ined,
the
sim
ulate
d
two
br
anch
m
od
el
sh
ow
s
bette
r
accuracy
fo
r
te
rm
inal
vo
lt
age
resp
on
se
of
su
per
capaci
tor.
An
oth
er
m
et
ho
d
fo
r
sel
fd
isc
har
ge
check
ing
of
su
per
capaci
tor
within sh
or
t per
iod
o
f
ti
m
e is
disp
la
ye
d
in Fig
ur
e 1
0.
The r
est
/discharg
e
as
sh
ow
n
in Fig
ur
e
11
(a)
the cu
rr
ent p
ro
file
o
f
6A
o
f
con
ti
n
uo
us
cu
rr
ent w
as u
sed
to
dischar
ge
the
su
per
capaci
tor.
The
su
per
capaci
tor
was
un
der
rest
wh
il
e
the
su
per
capaci
tor
was
in
fu
ll
y
char
ged
ph
ase
then
dischar
ged
fo
r
30
secon
ds
and
being
in
rest
fo
r
the
rem
ai
n
sim
ula
ti
on
ti
m
e
of
30
secon
ds
.
In
Figu
re
11
(b)
the
two
br
anch
sim
ulati
on
m
od
el
and
exp
erim
ental
data
of
te
rm
inal
vo
lt
age
of
su
per
capaci
tor
Evaluation Warning : The document was created with Spire.PDF for Python.
Ind
on
esi
a
n
J
E
le
c Eng &
Co
m
p
Sci
IS
S
N:
25
02
-
4752
Experime
nta
l
desi
gn for a
n
e
nhance
d para
m
et
ric
m
odel
in
g of s
up
erc
a
pa
ci
tor
… (
Ali
Moh
sen
Als
abar
i)
71
wer
e
com
par
ed
and
validat
ed.
T
he
err
or
between
the
sim
ulate
d
two
br
anch
m
od
el
and
m
easur
ed
data
hav
e
m
axim
um
relat
ive err
or
value
of
al
m
os
t
0.
02
V
as
sh
ow
n
in Fig
ur
e 1
1
(c)
.
(a)
(b)
(c)
Figure
10
.
T
he
se f
ig
ures a
re
;
(a)
c
urre
nt pr
ofi
le
(
b) r
est
-
disc
hargin
g
te
rm
inal vo
lt
ag
e
validat
ion (c
)
e
rror
for
rest an
d dis
charge
validat
ion o
f
s
uperca
pa
ci
tor vo
lt
age
4.
DISCU
SSI
ON
In
over
al
l,
th
e
inv
est
igate
d
par
am
at
ers
of
the
m
od
el
of
s
up
e
rca
pacit
or
diff
e
rs
with
re
gard
to
t
he
app
li
cat
io
n
of
the
m
od
el
it
sel
f
but
sti
ll
accuracy
of
the
m
od
el
co
nsi
der
e
d
as
m
a
in
fact
or
f
or
m
od
el
evaluati
on.i
n
e
le
ct
rical
m
od
el
,
the
m
od
el
accuracy
depen
ds
on
t
he
ide
ntific
at
ion
pro
file
,
as
well
a
s
th
e
rob
us
tness
of
s
up
e
rca
pacit
or
vo
lt
age
respo
nse
.
In
this
w
or
k,
the
ide
ntific
at
ion
proces
s
was
base
d
on
pro
posed
e
m
pirical
eq
ua
ti
on
s
t
hat
s
howed
the
ef
fecti
ven
ess
of
det
erm
ining
the
inter
nal
pa
ram
t
ers
of
s
up
e
rca
pacit
or
su
ccess
fu
ll
y.
The
pr
ocedure
p
r
opos
e
d
to
e
valuate
the
pa
ram
te
rs
hav
e
two
m
easur
em
ents
points
in
charge
curve
instea
d
of
four
m
easur
e
m
ents
us
ed
in
charge
an
d
dis
charge
phases
i
n
li
te
ratur
e.
T
he
se
two
points
cov
e
r
as
m
uch
of
th
e
chargin
g
pha
se
to
represe
nt
m
at
he
m
atical
equ
at
io
ns
of
te
rm
inal
vo
lt
age
with
captu
rin
g
th
e
ph
ysi
cs
relat
io
ns
hi
p
for
resis
ta
nce
an
d
ca
pa
ci
ta
nce.
T
hes
e
res
ults
of
in
te
rn
al
par
am
ter
s
ob
ta
ine
d
a
r
e
m
or
e
pr
eci
se
by lo
ok
ing
i
nto
t
he
e
rror eval
uated
i
n res
ults sect
ion.
To
a
dd
m
or
e,
i
n
this
pa
per
on
ly
on
e
de
vice
ne
wa
re
BT
S40
00
us
e
d
f
or
m
od
el
li
ng
proce
dures
w
hich
has
the
abili
ty
to
captur
e
m
any
i
m
po
rtant
da
ta
wh
il
st
is
us
ed
to
cha
r
ge
and
discha
r
ge
the
SC
com
par
ed
to
ty
pical
m
et
ho
ds
.Ty
pical
m
e
thod
nee
d
to
us
e
osc
il
lc
op
e,
vo
lt
m
et
er,
sh
un
t
bo
a
r
d,
da
ta
colle
ct
ion
de
vic
e
,
interface
d
e
vic
es for
host c
ompu
te
r
wh
ic
h
m
ay
lead to
t
he
i
ncr
ease
of t
he e
rror i
n
te
rm
inal vo
lt
ag
e
of
S
C.
By
Loo
ki
ng
t
o
the
li
te
rature
m
any
stud
ie
d
an
d
inv
e
sti
ga
te
d
the
m
od
e
l
of
superca
pa
ci
tor
widely
,
howe
ver
the
ac
cur
acy
of
the
m
od
el
hav
e
dif
fer
e
nt
fact
or
s
f
or
eval
uation.
This
st
ud
y
c
om
pa
red
an
d
e
va
luate
d
the
m
od
el
studi
ed
her
e
base
d
on
to
ols
use
d
f
or
m
od
el
li
ng
a
nd
acc
ur
acy
er
ror
analy
sis.
T
akin
g
so
m
e
previo
us
m
od
el
li
ng
work s
uc
h
as t
he
st
ud
y
done
b
y
[
2
6
]
in
w
hich
t
he
ir m
axi
m
u
m
r
el
at
ive error wa
s up
t
o
al
m
os
t
2%.
Also
an
oth
e
r
s
tud
y
done
by
[6
]
too
m
any
h
ardware
to
ols
us
e
d
and
thei
r
m
axi
m
u
m
rela
ti
ve
err
or
of
te
rm
inal
vo
lt
ag
e
was
0.25
V.
Be
sd
ie
s,
the
stud
y
done
by
[
1
0
]
introd
uce
d
the
m
od
el
ing
of
superca
pacit
or
with
i
m
pr
oved
t
wo
-
br
a
nc
h
ci
rcu
it
m
od
el
by
add
i
ng
c
urre
nt
co
nt
ro
ll
ed
s
ource
to
i
m
pr
o
ve
sel
f
-
discha
rg
e
rat
e
of
su
pe
rca
pacit
or
vo
lt
age
.
I
n
t
his
stud
y
they
us
e
d
ne
war
e
te
sti
ng
de
vice
but
th
e
m
axi
m
u
m
rel
at
ive
error
s
rang
e
d
from
0.
19
an
d
up
to
2.35
%
f
or
si
m
il
ar
exp
er
i
m
ents
te
st
pr
ofi
le
s
app
li
ed
in
this
wo
r
k.
H
oweve
r,
in
this
pap
e
r
the m
axi
m
u
m
relat
ive error w
as only
up t
o 0.045
V for te
rm
inal v
oltage
va
li
dation of su
pe
rcap
aci
to
r.
More
ov
e
r
[
1
0
]
in
their
stu
dy
the
sel
f
-
discha
rg
e
rate
was
i
m
pr
ov
e
d
in
th
ei
r
m
od
el
bu
t
sti
ll
there
are
long
an
d
co
m
plex
m
e
tho
d
of
the
ide
nt
ific
at
ion
proc
ess
of
inte
rn
a
l
par
am
et
ers
by
us
in
g
filt
ers
a
nd
Evaluation Warning : The document was created with Spire.PDF for Python.
IS
S
N
:
2502
-
4752
Ind
on
esi
a
n
J
E
le
c Eng &
Co
m
p
Sci,
Vo
l.
23
, N
o.
1
,
Ju
ly
2021
:
63
-
74
72
op
ti
m
iz
ation
al
gorthim
s.
Howev
e
r
in
this
stud
y,
the
sel
f
-
disc
harge
rate
durin
g
rest
w
as
relat
ively
l
ow
a
nd
bette
r
than
li
ne
transm
issi
on
(Mat
la
b)
m
od
el
with
s
m
al
l
var
ia
ti
on
erro
r,
so
that
it
is
n
ot
m
od
el
le
d
a
s
m
ai
n
par
am
et
er
s
in
t
his
pa
per
.
In
a
dd
it
io
n
[
28
]
sta
te
d
in
their
stud
y
that
it
is
diff
ic
ult
to
ha
ve
a
un
ifie
d
sta
ndar
d
to
com
par
e
the
s
el
f
-
disc
harge
r
at
es
du
e
t
o
th
e
diff
e
re
nt
syst
e
m
s
(includ
i
ng
el
ect
rode
m
at
erial
s,
el
ect
ro
ly
te
s,
separ
at
or
s
,
an
d
init
ia
l
ce
ll
vo
lt
ages)
em
plo
ye
d
by
resea
rchers
f
or
sel
f
-
dis
charge
te
sts.
Howe
ver,
due
to
the
ph
e
nom
eno
n
of
s
up
e
rcaict
or
sel
f
-
discha
r
ge
in
pr
act
ic
al
app
li
cat
ion
s
on
ce
th
e
supe
rcacit
or
a
pp
li
c
at
ion
requires
bein
g
in
rest
f
or
day
s
an
d
lo
ng
ti
m
e
without
ch
ar
ge
or
discha
rge,
it
is
reco
m
m
end
ed
f
or
m
od
el
i
ng
sel
f
disc
harge
accuratel
y.
T
o
co
nclu
de,
bas
ed
on
the
er
ror
analay
sis,
t
he
m
od
el
inv
est
i
gated
he
re
has
bette
r
error
value
w
hi
ch
co
nf
i
rm
s
t
hat
the
to
ols
a
nd
m
et
hod
use
d
f
or
inter
nal
par
am
te
rs
est
im
at
ion
an
d
val
idati
on
are r
ec
omm
end
ed
m
et
h
od
for
super
ca
pacit
or
m
od
el
li
ng
.
5.
CONCL
US
I
O
N
The
tw
o
-
bran
ch
ci
rc
uit
m
o
del
was
i
nv
e
s
ti
gated
in
dif
fer
e
nt
cu
reen
t
prof
il
es
a
nd
the
m
od
el
represe
nted
t
he
dynam
ic
beh
a
vior
of
s
uperca
pacit
or
.
I
nter
na
l
par
am
et
ers
of
the
m
od
el
we
r
e
identifie
d
an
d
the
si
m
ulati
on
m
od
el
was
buil
t
via
Ma
tl
ab
/
Sim
ul
ink
s
ucces
sfu
ll
y.
T
he
res
ults
of
the
sim
ulati
on
a
nd
m
easur
e
d
vo
lt
age
was
com
par
ed
an
d
the
resu
lt
s
wa
s
com
pr
m
ise
d.
The
validat
ion
er
r
or
bet
w
een
m
easur
ed
an
d
si
m
ulate
d
m
od
el
s w
as inv
est
i
gate in each
sc
enar
i
o
discu
sse
d.
T
he
error
w
as u
p
to
0.04V
as
m
axi
m
u
m
r
el
at
ive
error
wh
ic
h
shows
t
he
ef
fecti
ven
e
ss
an
d
va
li
dity
of
the
m
od
el
c
om
par
ed
to
the
li
te
ratur
e.
Finall
y,
the
sel
f
-
discha
rg
e
rate
of
s
uperca
paci
tor
was
im
pr
oved
a
nd
f
ollo
wed
t
he
m
eas
ur
e
d
vo
lt
age
.
I
t
is
reco
m
m
en
ded
f
or
fu
t
ur
e
w
ork
t
o
stu
dy
the
e
ff
e
ct
of
sel
f
-
disc
ha
rg
e
rate
durin
g
lo
ng
per
i
od
of
rest
i
n
pr
act
ic
al
app
li
cat
io
ns
for
us
in
g
the
tw
o
-
br
a
nc
h
ci
rcu
it
m
od
el
.
The
n
e
war
e
te
sti
ng
de
vice
us
e
d
m
ain
ly
for
ba
tt
ery
te
sti
ng
an
d
it
can
be
us
e
d
f
or
s
uper
capaci
tor
as
w
el
l
for
c
harge
and
disc
ha
rg
e
,
howe
ve
r
it
is
rec
omm
end
ed
to
buy
an
d
use
it
if
m
any cel
ls no
t
on
ly
on
e
cell
nee
ded to
be
m
odel
le
d for c
os
t
savin
g.
ACKN
OWLE
DGE
MENTS
The
aut
hors
w
ou
l
d
li
ke
to
thank
the
Fac
ulty
of
En
gin
ee
ring,
U
niv
er
sit
y
Pu
tra
Ma
la
ysi
a
(U
PM)
f
or
pro
vid
in
g
t
he
Gr
a
nt
Vot
95
91
00
an
d
fa
ci
li
ti
es
and
co
nducive
le
ar
nin
g
en
vir
onm
e
nt
in
c
onduct
i
ng
the
researc
h
.
REFERE
NCE
S
[1]
Y.
Parvini
,
J.
Siege
l
,
A.
Stef
anopoul
ou,
and
A.
Vahidi,
“
Superc
ap
acitor
E
le
c
tri
c
al
and
T
her
m
al
Modeli
n
g,
Ide
nti
f
ic
a
ti
on
,
a
nd
Vali
dation
for
a
W
ide
Range
of
Te
m
per
ature
and
Pow
er
Applic
a
ti
ons,”
I
EE
E
Tr
ans.
Ind.
El
e
ct
ron.
,
vol
.
6
3,
no
.
3
,
pp
.
157
4
-
1585
,
2015
,
d
oi:
10
.
1109/T
I
E.2015.2494868
.
[2]
A.
Berr
ueta,
I.
S
an
Martí
n,
A.
He
rná
ndez,
A.
Urs
úa,
and
P.
Sanch
is,
“
El
ec
t
ro
-
the
r
m
al
m
odel
li
ng
of
a
superc
apaci
to
r
and
exp
eri
m
enta
l
va
li
da
ti
on,
”
J.
Powe
r Source
s
,
vol.
259
,
pp
.
154
–
165,
2014
,
doi
:
10.
1016/j.jpowsour.201
4.
02
.
089
.
[3]
K.
Popoola,
“
Modell
ing
and
Si
m
ula
ti
on
of
Superc
ap
acitor
for
E
ner
g
y
Stor
age
A
ppli
c
at
ions,
”
vol
.
7,
no.
1
,
pp.
2
–
7,
2018.
[4]
K.
Li
a
,
Y.
W
ahi
ba,
and
E.
E
vgueniy
,
“
H
y
bri
d
bat
te
r
y
/sup
ercapa
c
it
or
en
erg
y
storage
s
y
stem
for
the
el
e
ct
ri
c
vehi
c
le
s,”
J
.
Po
wer
Source
s
,
vol
.
374
,
pp
.
237
-
2
48,
2018
,
doi
:
10
.
1016/j.jpowsour.2017.
11
.
040
.
[5]
I.
Ji
y
a
,
N
.
Gur
usinghe,
and
R
.
Gouw
s,
“
El
ect
ric
a
l
Cir
cui
t
M
odel
li
ng
of
Dou
ble
Lay
er
Cap
a
ci
tors
for
Pow
er
El
e
ct
roni
cs
and
Ene
rg
y
Stor
ag
e
Applic
a
ti
ons:
A
Revi
ew,
”
El
e
ct
ron
ics
,
vol
.
7,
no.
11,
p
.
268,
2018
,
doi:
10.
3390/electron
ic
s7110268
.
[6]
P.
Saha
and
M.
Khanra
,
“
Equi
v
al
en
t
ci
rcu
it
m
odel
of
superc
apaci
tor
for
self
-
dis
cha
rge
an
aly
sis
-
A
compara
ti
ve
stud
y
,
”
In
.
2016
Inte
rnationa
l
C
onfe
renc
e
on
Si
gnal
Proce
ss
ing
,
Comm
unic
a
ti
o
n,
Powe
r
and
E
mbedde
d
Syste
m
(
SCOPES
)
,
2016
,
pp
.
1381
–
1386
,
doi:
1
0.
1109/S
COP
ES.
2016.
7955667
.
[7]
A.
La
h
y
a
ni,
P.
Vene
t,
A.
Guer
m
az
i,
and
A.
T
roudi,
“
Batter
y
/
Superc
apaci
tors
Com
bina
ti
on
i
n
Uninte
rrupt
ible
Pow
er
Supply
(UPS),
”
IEEE
Tr
ans.
Pow
er
El
ectron.
,
vol.
28,
no.
4,
pp.
1509
–
1522,
2013
,
doi:
10.
1109/T
PE
L.
2
012.
2210736
.
[8]
L.
Zha
ng
,
X.
Hu,
Z
.
W
ang
,
F.
Sun,
and
D
.
G.
Dorrel
l,
“
A
rev
i
ew
of
superc
apaci
tor
m
odel
ing,
esti
m
at
ion
,
a
n
d
appl
i
ca
t
ions:
A
cont
rol/
m
an
agem
ent
per
spec
t
iv
e,
”
Re
new
.
Susta
in.
Ene
rgy
Rev
.
,
vol.
81,
pp.
186
8
-
1878,
2018
,
d
oi:
10.
1016/j.rser.
20
17.
05.
283
.
[9]
L.
E
.
Helse
th,
“
Modell
ing
super
ca
pa
ci
tors
using
a
d
y
n
amic
equ
iva
l
ent
c
irc
ui
t
with
a
distri
bu
tion
of
rel
ax
at
io
n
ti
m
es,
”
J. E
n
ergy
Storage
,
vo
l. 2
5
,
p
.
100912
,
20
19
,
doi
:
10
.
1016
/j
.
est.
201
9
.
1009
12
.
[1
0
]
D.
Xu,
L.
Zha
n
g,
B.
W
ang,
an
d
G.
Ma,
“
Modeli
ng
of
Super
capac
i
tor
Beha
v
ior
W
it
h
an
Im
prove
d
Two
-
Branch
Equi
valent
Circ
u
it
,
”
I
EE
E
A
cc
ess
,
vol
.
7
,
pp
.
2637
9
-
26390,
2019
,
doi:
10
.
1109/AC
CESS
.
2019.
290
1377
.
[1
1
]
K.
Li
u,
C.
Zhu,
R.
Lu,
and
C.
C.
Chan,
“
Im
prov
ed
stud
y
of
te
m
per
at
ur
e
depe
nd
enc
e
equ
iva
l
ent
ci
rcu
it
m
odel
for
superc
apaci
to
rs,
”
IE
EE Tr
ans.
P
lasma Sc
i.
,
vo
l.
41,
no
.
5
,
pp
.
12
67
-
1271,
2013
,
doi:
10
.
1109/T
P
S.2013.
2251363
.
[1
2
]
Y.
W
ang,
L
.
W
a
ng,
M.
L
i,
and
Z
.
Chen
,
“
A
rev
iew
of
ke
y
issues
for
cont
rol
and
m
ana
gement
in
bat
t
er
y
and
u
lt
r
a
-
ca
pa
ci
tor
h
y
brid
ene
rg
y
stor
age sy
stems
,
”
eTrans
portati
on
,
vol. 4
,
2020
,
doi:
10.
10
16/j
.
et
ran
.
2020.
1
00064
.
Evaluation Warning : The document was created with Spire.PDF for Python.