Inter
national
J
our
nal
of
P
o
wer
Electr
onics
and
Dri
v
e
Systems
(IJPEDS)
V
ol.
8,
No.
4,
December
2017,
pp.
1830
–
1840
ISSN:
2088-8694
1830
Contr
ol
Strategy
of
a
Grid-connected
Photo
v
oltaic
with
Battery
Ener
gy
Storage
System
f
or
Hourly
P
o
wer
Dispatch
Mohd
Afifi
J
usoh
and
Muhamad
Zalani
Daud
School
of
Ocean
Engineering,
Uni
v
ersiti
Malaysia
T
erengg
anu,
Meng
abang
T
elipot,
21030
K
uala
Nerus,
T
erengg
anu,
Malaysia
Article
Inf
o
Article
history:
Recei
v
ed
Jun
11,
2017
Re
vised
Oct
16,
2017
Accepted
Oct
28,
2017
K
eyw
ord:
Photo
v
oltaic
Po
wer
fluctuation
Battery
ener
gy
storage
Control
scheme
Hourly
dispatch
ABSTRA
CT
The
high
penetration
of
fluctuated
photo
v
oltaic
(PV)
output
po
wer
into
utility
grid
system
will
af
fect
the
operation
of
interconnected
grids.
The
unnecessary
output
po
wer
fluctuation
of
PV
system
is
contrib
uted
by
unpredictable
nature
and
inconsistenc
y
of
solar
irr
adiance
and
temperature.
This
paper
presents
a
control
scheme
to
mitig
ate
the
output
po
wer
fluctuations
from
PV
system
and
dispatch
out
the
constant
po
wer
on
an
hourly
basis
to
the
utility
grid.
In
this
re
g
ards,
battery
ener
gy
storage
(BES)
system
is
used
to
eliminate
the
output
po
wer
fluctuation.
Control
scheme
is
proposed
to
maintain
parameters
of
BES
within
required
operating
constraints.
The
ef
fecti
v
eness
of
the
proposed
control
scheme
is
tested
using
historical
PV
system
input
data
obtained
from
a
site
in
Malaysia.
The
simulation
results
sho
w
that
the
proposed
control
scheme
of
BES
system
can
properly
manage
the
output
po
wer
fluctuations
of
the
PV
sources
by
dispatching
the
output
on
hourly
basis
to
the
utility
grid
while
meeting
all
required
operating
constraints.
Copyright
©
2017
Insitute
of
Advanced
Engineeering
and
Science
.
All
rights
r
eserved.
Corresponding
A
uthor:
Muhamad
Zalani
Daud
School
of
Ocean
Engineering,
Uni
v
ersiti
Malaysia
T
erengg
anu
Meng
abang
T
elipot,21030
K
uala
Nerus,
T
erengg
anu
Malaysia
Email:
zalani@umt.edu.my
1.
INTR
ODUCTION
In
recent
years,
grid-connected
photo
v
oltaic
(PV)
system
is
seen
to
enjo
y
the
most
rapid
gro
wth
among
the
v
arious
rene
w
able
ener
gy
sources.
Unfortunately
,
the
output
po
wer
from
the
PV
system
is
generally
unsta-
ble
and
unpredic
table
because
of
intermittent
and
uncertain
characteristics
of
solar
irradiance
and
temperature
[1].
High
penetrati
on
of
fluctuated
PV
output
po
wer
into
the
utility
grid
system
will
af
fect
operation
of
inter
-
connected
grids
[2].
Some
approaches
may
be
required
to
compensate
the
output
po
wer
fluctuations
in
order
to
ha
v
e
a
more
reliable
po
wer
system.
Recent
adv
ances
in
the
prediction
methods
enable
the
solar
radiation
profile
to
be
forecasted
with
acceptable
precis
ion
[3].
Estimating
solar
radiation
is
essential
in
order
to
generate
a
consistent
output
po
wer
of
PV
system.
There
are
man
y
prediction
models
for
prediction
of
solar
radiation
such
as
artificial
neural
netw
ork
(ANN)-based
model,
fuzzy
logic
control-based
model,
angstrom
model,
empirical
re
gression
model
and
empirical
coef
ficient
model
[3].
Se
v
eral
researchers
ha
v
e
claimed
that
the
accurac
y
of
the
forecast
model
can
be
achie
v
ed
up
to
90%
of
the
rated
resource
capacity
[4,
5].
Accurate
information
from
prediction
model
may
be
used
as
a
reference
in
mitig
ating
output
po
wer
fluctuation
of
PV
sources.
In
po
wer
system
applications,
ener
gy
storage
system
(ESS)
is
ackno
wledged
as
one
of
the
best
al-
ternati
v
e
techniques
to
mitig
ate
the
PV
output
po
wer
fluctuations
and
PV
output
po
wer
prediction
errors
[1].
V
arious
benefits
of
using
ESS
in
grid-connected
PV
system
application
are
discussed
in
[6].
Among
the
v
arious
of
ESS
technologies
capable
of
mitig
ating
the
fluctuating
and
unpredictable
output
po
wer
of
the
PV
systems
are
J
ournal
Homepage:
http://iaesjournal.com/online/inde
x.php/IJPEDS
,
DOI:
10.11591/ijpeds.v8i4.pp1830-1840
Evaluation Warning : The document was created with Spire.PDF for Python.
IJPEDS
ISSN:
2088-8694
1831
pumped
h
ydro
storage
(PHS),
compressed
air
ener
gy
storage
(CAES),
flywheel
ener
gy
storage
(FES),
super
-
conducting
magnetic
ener
gy
storage
(SMES),
supercapacitors
(SC)
and
battery
ener
gy
storage
(BES)
system.
From
the
literature,
it
has
been
found
that
BES
is
the
most
cost-ef
fecti
v
e
option
for
fluctuation
mitig
ation
purposes
compared
to
other
technologies
[7].
BES
is
also
kno
wn
as
elect
rochemical
ener
gy
storage
system
that
may
be
chosen
from
man
y
di
f
fer
-
ent
types
depending
on
the
v
arying
po
wer
application
needs
such
as
lead
acid
(LA),
v
alv
e
re
gulated
lead
acid
(VRLA),
Nick
el-cadmium
(NiCd),
Lithium-ion
(Li-ion),
sodium
sulfur
(N
AS)
and
v
anadium
redox
(VRB)
bat-
teries
[1].
The
adv
antages
and
disadv
antages
of
each
of
the
batteries
are
discussed
in
[6].
From
the
literature,
there
are
v
arious
applications
of
BES
in
grid
syst
em
in
performing
important
task
of
po
wer
fluctuation
miti-
g
ation
and
output
po
wer
dispatch
[8,
9,
10,
11,
12].
Daud
et
al.
[8]
proposed
an
optimal
controller
of
BES
to
smooth
the
output
po
wer
fluctuation
from
the
PV
sources.
The
control
system
used
the
forecasted
output
po
wer
of
PV
system
as
a
dispatching
reference.
The
controller
re
gulate
the
SOC
of
battery
according
to
the
desired
operational
constraints
and
the
output
po
wer
of
PV
system
is
dispatched
to
the
grid
system
on
hourly
basis.
T
o
optimize
the
controller
,
heuristic
optimization
is
used
to
obtain
the
optimal
parameters.
The
o
v
erall
ef
ficienc
y
of
the
controller
is
reported
as
82%
[8].
Consequently
,
a
control
scheme
based
on
the
rules
has
been
proposed
in
[9].
The
rules
in
the
control
scheme
are
b
uild
based
on
the
desired
operational
constraints
of
BES
such
as
state
of
char
ge
(SOC)
limits,
char
ge/dischar
ge
current
li
mits,
and
lifetime.
The
control
scheme
ef
fecti
v
ely
smooth
out
the
output
po
wer
of
PV
system
and
dispatch
out
the
po
wer
to
grid
system
in
hourly
basis.
Author
in
[13]
proposed
a
coordinated
control
scheme
to
reduce
the
impacts
of
wind
po
wer
forecast
errors
while
pro-
longing
the
lifetime
of
BES.
The
ener
gy
capacity
determination
method
from
the
hist
orical
data
is
proposed
in
this
w
ork.
The
control
scheme
used
forecasted
output
po
wer
data
to
impro
v
e
the
dispatchability
of
wind
po
wer
generation.
In
[10],
fuzzy-based
smoothing
control
scheme
has
been
presented.
The
fuzzy
w
a
v
elet
filtering
method
is
used
to
smoothing
the
fluctuate
output
of
wind
and
PV
systems.
Authors
in
[11]
de
v
eloped
adapti
v
e
control
of
BES
and
UC
for
smoothing
output
po
wer
of
PV
system.
The
proposed
adapti
v
e
fuzzy-based
control
scheme
manages
the
po
wer
sharing
between
BES
and
UC
based
on
the
operational
constraints
of
BES
and
UC
in
order
to
sustain
the
system
operation.
Similarly
,
an
HES
system
composed
of
Li-ion
batteries
and
UC
has
been
proposed
in
[12].
A
multimode
fuzzy
logic-based
allocator
has
been
designed
to
ensure
the
BES
and
UC
can
be
ef
ficiently
utilized
as
well
as
pre
v
enting
from
w
orking
under
e
xtreme
conditions.
Since
the
cost
of
the
lar
ge
scale
BES
is
e
xpensi
v
e,
the
ener
gy
of
the
BES
should
be
optimally
con-
trolled
particularly
in
the
BES
application
for
po
wer
fluctuation
mitig
ation
of
PV
sources.
The
optimal
con-
troller
of
BES
can
reduce
the
maintenance
cost
and
increase
the
lifetime
of
the
BES
while
pro
viding
the
contin-
uous
support
for
po
wer
fluctuation
mitig
ation.
This
paper
presents
an
optimal
control
scheme
of
grid-connected
PV
-BES
system.
The
objecti
v
e
of
the
paper
is
to
design
an
optimal
controller
of
grid-connected
PV
-BES
system
so
that
the
total
output
of
the
system
can
be
smoothed
out
and
dispatched
on
an
hourly
basis
to
the
utility
grid.
The
follo
wing
sections
of
the
paper
comprises
of
the
details
of
proposed
BES
control
scheme,
the
description
of
modelling
and
simulation
of
PV
-BES
system,
results
and
discussion,
follo
wed
by
the
conclusion.
2.
PR
OPOSED
BES
CONTR
OL
SCHEME
Inte
grating
BES
with
grid-connected
PV
system
will
reduce
the
output
po
wer
fluctuation
and
di
spatch
out
the
output
po
wer
of
PV
system
to
grid
system
in
hourly
constant.
A
typical
grid-connected
PV
-BES
system
is
illustrated
in
Figure
1.
The
BES
is
parallel
connected
to
the
system
at
the
point
of
common
coupling
(PCC)
through
po
wer
con
v
erter
.
The
purpose
of
the
con
v
erter
is
to
re
gulate
the
fluctuate
output
po
wer
of
PV
system
by
controlling
the
char
ge
and
dischar
ge
po
wer
of
BES.
In
order
to
dispatch
a
constant
po
wer
to
the
grid
system
,
a
rob
ust
control
scheme
of
BES
needs
to
be
de
v
eloped.
In
this
pape
r
,
the
control
scheme
is
included
at
the
outer
control
loop
of
the
BES-VSC
as
sho
wn
in
Figure
1.
The
control
goal
is
to
ensure
a
continuous
char
ge/dischar
ge
of
BES
po
wer
for
reducing
the
PV
po
wer
fluctuations
while
pro
viding
safety
and
economical
of
BES.
The
BES
is
subjected
to
the
operational
constraints
such
as
SOC
operating
limits
and
depth
of
dischar
ge
(DOD),
v
oltage
e
xponential
limits,
and
current
limit
at
the
des
ired
range.
The
proposed
control
scheme
is
de
v
eloped
for
hourly
PV
output
po
wer
dispatch
strate
gy
as
illustrated
in
Figure
2.
The
proposed
control
scheme
is
moti
v
ated
from
the
conceptual
design
for
output
po
wer
smoothing
used
in
[8,
9].
The
aim
of
the
control
scheme
is
to
generate
the
reference
signal
for
char
ge/dischar
ge
of
battery
po
wer
,
P
r
ef
B
E
S
while
meeting
all
required
BES
operational
constraints.
The
required
BES
operational
constraints
are
described
as
follo
ws:
S
O
C
B
E
S
min
S
O
C
B
E
S
(
t
)
S
O
C
B
E
S
max
(1)
Contr
ol
Str
ate
gy
of
a
Grid-connected
Photo
voltaic
with
Battery
Ener
gy
...
(Mohd
Afifi
J
usoh)
Evaluation Warning : The document was created with Spire.PDF for Python.
1832
ISSN:
2088-8694
Figure
1.
System
configuration
and
control
of
grid-connected
PV/BES
system
for
hourly
output
po
wer
dispatch.
I
max
ch
I
B
E
S
(
t
)
I
max
dis
(2)
V
B
E
S
min
V
B
E
S
(
t
)
V
B
E
S
max
(3)
As
sho
wn
in
Figure
2,
the
input
of
the
control
scheme
is
P
S
E
T
0
which
is
determined
from
the
hourly
a
v
erage
of
forecasted
P
P
V
.
The
details
of
P
S
E
T
0
and
P
P
V
are
described
in
the
ne
xt
chapter
.
The
SOC
feedback
signal
of
BES
(
S
O
C
B
E
S
)
e
v
ery
one
hour
is
used
to
determine
the
ener
gy
dif
ference
of
BES
to
maintain
the
S
O
C
B
E
S
at
desired
SOC
le
v
el
(
S
O
C
r
ef
B
E
S
).
In
this
case,
the
S
O
C
r
ef
B
E
S
is
set
at
0.6
p.u
which
is
the
most
ideal
SOC
starti
ng
v
alue
for
the
selected
SOC
range.
The
dif
ferent
ener
gy
of
BES
in
MWh
is
added
to
P
S
E
T
0
in
order
to
ensure
that
S
O
C
B
E
S
maintained
at
S
O
C
r
ef
B
E
S
at
the
end
of
e
v
ery
hour
.
Figure
2.
Outer
loop
control
of
BES-VSC.
T
o
reduce
the
ne
g
ati
v
e
impacts
from
the
drastic
change
of
po
wer
e
v
ery
sub-hourly
,
the
ramp
rate
limiter
is
also
applied
as
illustrated
in
Figure
2.
The
rate
limiter
(
r
r
)
of
0.03
MW/min
(up
and
do
wn
ra
mp
rates)
is
applied
to
pre
v
ent
o
v
ershooting
when
P
S
E
T
0
changes
so
as
to
a
v
oid
significant
up/do
wn
ramps
of
total
output
po
wer
to
the
grid.
Figure
3
gi
v
es
the
proposed
ramp
rate
concept.
As
illustrated
in
Figure
3,
the
line
of
B
C
F
and
area
of
O
AB
C
F
G
represent
the
po
wer
reference
(
P
S
E
T
0
)
and
total
ener
gy
reference
(
E
S
E
T
0
)
deli
v
ered
to
the
grid
system
without
ramp
rate,
while
line
of
AC
F
and
area
of
O
AC
F
G
represent
the
po
wer
reference
(
P
S
E
T
)
and
total
ener
gy
reference
(
E
S
E
T
)
deli
v
ered
to
the
grid
system
with
traditional
ramp
rate,
respecti
v
ely
.
By
comparing
of
the
total
ener
gy
reference,
the
total
ener
gy
is
reduced
if
the
traditional
ramp
rate
control
is
applied.
This
ener
gy
reduction
occurrance
is
because
of
the
cut-of
f
po
wer
during
the
ramp
rate
transition.
T
o
o
v
ercome
the
shortage
of
ener
gy
deli
v
ered,
the
fle
xible
ramp
rate
control
is
proposed
in
this
paper
.
As
sho
wn
in
Fi
g
ur
e
3,
the
solid
line
of
AD
E
and
area
of
O
AC
F
G
represent
the
P
S
E
T
and
E
S
E
T
deli
v
ered
to
the
grid
system
with
proposed
ramp
rate
strate
gy
without
ener
gy
reduction.
Based
on
the
figure,
the
P
S
E
T
ramps
up
at
0.03MW/min
rate
at
the
be
ginning
of
hour
t
and
retains
steady
until
the
end
of
the
hour
.
So,
the
ramping
duration
r
r
at
hour
t
is
calculated
by
using
Equation
4
[14].
The
hourly
ener
gy
is
e
xpressed
by
Equation
5
[14]
where
E
R
is
ener
gy
deli
v
ered
in
ramping
operation
and
E
S
is
ener
gy
deli
v
ered
during
stable
IJPEDS
V
ol.
8,
No.
4,
December
2017:
1830
–
1840
Evaluation Warning : The document was created with Spire.PDF for Python.
IJPEDS
ISSN:
2088-8694
1833
Figure
3.
Illustration
of
ramp
rate
control.
state
operation.
The
equation
represents
the
correlation
between
E
S
E
T
and
P
S
E
T
.
r
r
(
t
)
=
j
P
S
E
T
(
t
)
P
S
E
T
(
t
1)
j
r
r
(4)
E
S
E
T
(
t
)
=
E
R
(
t
)
+
E
S
(
t
)
=
r
r
(
t
)
2
(
P
S
E
T
(
t
)
P
S
E
T
(
t
1))
+
(
r
r
(
t
))
=
j
P
2
S
E
T
(
t
)
P
2
S
E
T
(
t
1)
j
2
r
r
+
j
P
S
E
T
(
t
)
P
S
E
T
(
t
1)
j
r
r
P
S
E
T
(
t
)
(5)
T
o
ensure
the
SOC
operational
of
BES
is
within
the
desired
limit,
the
rules-based
control
in
[9]
is
used.
The
de
v
eloped
rules-based
control
is
as
illustrated
in
Figure
4.
The
input
of
the
rules-based
control,
P
tar
B
E
S
is
a
de
viation
of
P
S
E
T
and
P
P
V
while
the
output
is
P
r
ef
B
E
S
.
In
the
present
study
,
the
S
O
C
B
E
S
min
and
S
O
C
B
E
S
max
are
set
to
0.3
p.u
and
0.9
p.u,
respecti
v
ely
.
T
o
ensure
that
the
output
of
the
outer
loop
control,
I
r
ef
B
E
S
stays
within
required
operational
constraint
of
I
B
E
S
,
the
current
limiter
block
is
applied.
The
maximum
char
ge/dischar
ge
current
of
BES
should
not
e
xceed
1
C
B
E
S
amperes.
Figure
4.
Flo
wchart
of
rules-based
control.
Contr
ol
Str
ate
gy
of
a
Grid-connected
Photo
voltaic
with
Battery
Ener
gy
...
(Mohd
Afifi
J
usoh)
Evaluation Warning : The document was created with Spire.PDF for Python.
1834
ISSN:
2088-8694
3.
MODELLING
AND
SIMULA
TION
OF
PV
-BES
SYSTEM
The
simulation
for
v
alidating
the
proposed
control
scheme
is
carried
out
using
Matlab/Simulink.
This
section
describes
the
method
of
obtaining
PV
output
po
wer
profile
(
P
P
V
),
hourly
set-point
po
wer
profile
(
P
S
E
T
0
),
BES
po
wer
and
ener
gy
rating.
Besides
that,
the
details
of
BES
system
model
is
also
presented.
3.1.
Output
po
wer
of
PV
system
(
PPV
)
and
determination
of
po
wer
r
efer
ence
pr
ofile
(
PSET
0
)
The
one-year
Malaysian
historical
radiation
and
temperature
data
are
used
in
pr
o
duci
ng
the
P
P
V
out-
put
po
wer
data
[8].
The
hourly
a
v
erage
of
radiation
and
temperature
are
manipulated
by
adding
random
noise
to
represent
the
actual
radiation
and
temperature
data.
The
added
random
noise
data
are
e
xtracted
according
to
the
Malaysia
weather
characteristic
where
the
interm
ittent
of
clouds
are
occurred
between
11AM
to
3PM.
Figure
5
sho
ws
the
one-day
P
P
V
of
which
e
xtracted
from
manipulated
radiation
and
temperature
data
by
using
1.2
MW
grid-connected
PV
system
model
[8].
In
this
re
g
ards,
5%
po
wer
loss
through
the
con
v
erter
is
assumed
and
maximum
po
wer
point
tracking
(MPPT)
operation
is
considered
in
the
PV
system
model.
The
P
P
V
data
obtained
are
used
to
e
v
aluate
the
proposed
control
scheme.
P
S
E
T
0
is
a
s
et-point
po
wer
profile
that
is
used
as
a
reference
to
dispatch
a
constant
po
wer
to
the
grid
system
in
a
certain
period.
F
or
this
study
,
the
one-hour
dispatch
period
is
chosen.
The
magnitude
of
P
S
E
T
0
is
determined
from
hourly
a
v
erage
of
P
P
V
.
The
ideal
P
S
E
T
0
represents
the
dispatch
reference
without
an
y
error
of
forecast
model.
T
o
represent
the
error
of
forecast
model,
10%
mean
absolute
error
(MAE)
is
added
into
P
S
E
T
0
as
illustrated
in
Figure
5.
Figure
5.
Po
wer
profile
of
P
P
V
and
P
S
E
T
0
.
3.2.
Determination
of
BES
po
wer
and
ener
gy
capacity
The
required
ener
gy
of
BES
(
E
B
E
S
)
is
determined
based
on
the
output
po
wer
profile
(
P
P
V
)
and
po
wer
set-point
profile
(
P
S
E
T
0
)
in
Figure
5
by
using
Equation
6
[9],
where
P
r
ef
B
E
S
is
po
wer
reference
of
BES
without
using
control
scheme
that
is
obtained
us
ing
Equation
7.
Figure
6
illustrates
the
BES
po
wer
rating
and
BES
ener
gy
rating,
respecti
v
ely
.
From
the
figure,
a
total
of
0.3
MWh
size
of
BES
and
0.6
MV
A
con
v
erter
rating
are
required
if
there
is
no
error
in
forecast
considered.
On
the
other
hand,
a
minimum
0.9
MWh
size
of
BES
is
required
when
10%
error
in
forecast
is
considered.
This
clearly
sho
ws
the
need
for
a
proper
control
and
management
of
BES
SOC
in
order
to
minimize
the
BES
ener
gy
rating
when
the
error
in
forecast
is
considered.
In
this
paper
,
0.3
MWh
size
of
BES
is
selected
and
used
with
the
proposed
control
scheme
in
the
simulation
study
.
E
B
E
S
(
t
)
=
E
B
E
S
(0)
+
Z
t
0
P
r
ef
B
E
S
(
t
)
dx
(6)
P
r
ef
B
E
S
(
t
)
=
P
S
E
T
0
(
t
)
P
P
V
(
t
)
(7)
IJPEDS
V
ol.
8,
No.
4,
December
2017:
1830
–
1840
Evaluation Warning : The document was created with Spire.PDF for Python.
IJPEDS
ISSN:
2088-8694
1835
Figure
6.
Po
wer
and
ener
gy
rating
for
BES.
3.3.
Modelling
of
BES
system
In
the
present
study
,
the
lithium-ion
battery
is
used
as
BES
because
of
its
e
xcellent
performance
such
as
high
ener
gy
density
and
high
capacity
.
The
lithium-ion
battery
model
is
modelled
based
on
the
dynamic
equi
v
alent
circuit
as
illustrated
in
Figure
7
[15].
The
dynamic
model
gi
v
es
the
relationship
between
v
oltage,
current
and
the
a
v
ailable
char
ge
(SOC)
of
the
battery
.
Figure
7.
Equi
v
alent
circuit
of
battery
simulation
model.
The
mathematical
equation
of
the
dynamic
model
are
described
based
on
the
follo
wing
equations:
V
B
at
=
E
B
at
R
int
I
B
at
;
(8)
S
O
C
=
100(
R
I
B
at
dt
Q
)
;
(9)
E
B
at
disc
=
E
0
[
K
(
Q
Q
it
i
)]
[
K
(
Q
Q
it
it
)]
+
Ae
(
B
)(
it
)
;
(10)
E
B
at
char
g
=
E
0
[
K
(
Q
it
0
:
1
Q
i
)]
[
K
(
Q
Q
it
it
)]
+
Ae
(
B
)(
it
)
;
(11)
it
=
Z
I
B
at
dt
(12)
where
V
B
at
is
the
battery
v
oltage
(V),
R
int
is
the
batt
ery
internal
resistance
(
),
I
B
at
is
the
battery
current
(A),
Q
is
the
cell
capacity
(Ah),
E
B
at
disc
is
battery
electromoti
v
e
force
during
dischar
ge
(V),
E
B
at
char
g
is
battery
Contr
ol
Str
ate
gy
of
a
Grid-connected
Photo
voltaic
with
Battery
Ener
gy
...
(Mohd
Afifi
J
usoh)
Evaluation Warning : The document was created with Spire.PDF for Python.
1836
ISSN:
2088-8694
electromoti
v
e
force
during
char
ge
(V),
E
0
is
battery
open-circuit
v
oltage
(V),
K
is
polarisation
resistance
(
),
it
is
actual
battery
current
(Ah),
i
is
filtered
c
u
r
rent
(A),
A
is
e
xponential
zone
v
oltage
(V)
and
B
is
e
xponential
zone
time
constant
in
v
erse
(Ah)
1
.
4.
RESUL
TS
AND
DISCUSSION
This
section
presents
the
simulation
results
and
discussions.
The
first
par
t
discusses
about
the
proposed
controller
performance
for
dispatching
the
total
output
of
PV
-BES
system.
The
second
part
pro
vides
the
case
studies
to
e
v
aluate
the
ef
fects
of
initial
v
alues
of
BES
SOC
to
the
dispatching
performance.
4.1.
Effects
of
pr
oposed
contr
oller
to
the
dispatching
perf
ormance
of
PV
-BES
Figure
8
illustrates
the
ef
fect
of
the
proposed
scheme
on
the
dispatching
performance
of
PV
-BES
system.
The
simulation
results
sho
w
graphically
which
summarises
the
output
po
wer
dispatch
curv
e,
SOC,
BES
v
oltage
and
current
profiles
of
the
PV
-BES
system.
F
or
the
case
of
uncontrolled
S
O
C
B
E
S
,
it
is
clearly
sho
wn
in
the
Figure
8(a)
that
the
po
wer
can
be
smoothly
deli
v
ered
to
the
grid
system
without
an
y
fluctuated
output.
Ho
we
v
er
,
the
parameters
of
BES
e
xceeds
the
lo
west
limit
of
desired
operational
constraints
as
e
vident
in
Figure
8(b),
(c)
and
(d),
where
the
lo
west
V
B
E
S
,
S
O
C
B
E
S
and
I
B
E
S
are
0.56
kV
,
0.1
p.u
and
0.7
kA,
respecti
v
ely
.
Therefore,
to
meet
the
acceptable
dispatching
performance
with
safe
battery
operation,
the
S
O
C
B
E
S
needs
to
be
properly
controlled.
F
or
a
controlled
S
O
C
B
E
S
,
the
po
wer
deli
v
ered
to
the
grid
system
is
consistent
with
the
fluctuati
ons
ha
v
e
been
minimized
to
a
certain
le
v
el.
Consequently
,
all
the
BES
parameters
constraints
are
satisfied
as
sho
wn
in
Figure
8(b),
(c)
and
(d),
respecti
v
ely
.
From
Figure
8(a),
there
are
some
spik
es
e
xist
mostly
between
11
AM
and
3
PM.
The
visibility
of
the
spik
es
that
occur
is
because
of
current
blocking
in
the
current
operational
limits
(
1
C
B
E
S
)
of
I
B
E
S
.
In
practice,
there
are
man
y
w
ays
to
eliminate
such
spik
es,
for
e
xample
by
installing
high
po
wer
storage
de
vice
such
as
supercapacitor
[8,
16].
Output
po
wer
in
Figure
8(a)
also
sho
ws
reduction
of
output
po
wer
dispatch
due
to
controller
setting
in
maintaining
the
S
O
C
B
E
S
at
the
end
of
sub-hourly
the
same
as
the
initial
S
O
C
B
E
S
.
As
e
vident
in
Figure
8(c),
the
S
O
C
B
E
S
le
v
el
is
maintained
to
remain
the
v
alue
close
to
initial
S
O
C
B
E
S
at
the
end
of
the
day
compared
to
the
S
O
C
B
E
S
in
pre
vious
case.
The
lo
west
S
O
C
B
E
S
is
measured
around
0.36
p.u.
Besides
that,
the
simulation
results
also
sho
w
v
oltage
and
current
profiles
of
the
PV
-BES
system
in
Figure
8(b)
and
(c),
respecti
v
ely
.
Based
on
Figure
8(b),
the
result
sho
ws
the
minimum
v
oltage
of
the
BES
which
is
0.6381
kV
does
not
e
xceed
the
lo
west
boundary
of
V
B
E
S
.
Meanwhile,
in
Figure
8(d)
sho
ws
maximum
char
ge
and
dischar
ge
current
profiles
that
has
been
limited
to
1
C
B
E
S
for
safety
purposes.
From
the
curv
e,
the
I
B
E
S
does
not
e
xceed
0.5
kA.
In
addition
to
the
results
of
the
dispatching
performances
of
PV
-BES
system,
the
ef
fecti
v
eness
of
proposed
control
scheme
is
determined
using
the
performance
inde
x
(
P
I
)
sho
wn
in
Equations
13
and
14
[17]
where,
N
x
represents
the
number
of
occurrences
of
de
viations
and
dP
is
the
dif
ference
of
the
total
output
and
the
desired
set
point.
The
total
output
po
wer
,
P
G
in
this
case
is
measured
at
the
PCC
s
ystem
b
us
where
the
PV
system
is
connected.
P
I
=
X
N
x
j
dP
x
j
(13)
dP
=
P
S
E
T
P
G
(14)
The
po
wer
de
viation,
dP
is
ill
u
s
trated
in
Figure
9
and
it
is
v
erified
that
if
the
proposed
control
scheme
is
not
used
to
smooth
out
the
P
P
V
output
and
dispatch
on
an
hourly
basis,
the
unacceptable
de
via
tion
will
occur
.
The
de
viation
without
mitig
ation
scheme
that
e
xceeds
0.12
MW
(10%)
from
the
total
PV
capacity
is
found
to
be
approximately
20%
as
illus
trated
in
Figure
9(a).
W
ith
the
proposed
control
scheme,
the
unacceptable
de
viations
greatly
impro
v
ed
as
sho
wn
in
Figure
9(b),
in
which
the
de
viations
decreased
to
less
than
1%.
4.2.
Effects
of
initial
SOCBES
to
the
dispatching
perf
ormance
of
PV
-BES
In
order
to
e
v
aluate
the
rob
ustness
and
fle
xibility
of
the
performance
of
the
proposed
control
scheme,
the
follo
wing
three
critical
cases
of
storage
initial
S
O
C
B
E
S
are
considered.
The
initial
S
O
C
B
E
S
is
set
to
nearly
full
char
ge
of
90%
(case
1),
nearl
y
full
dischar
ged
at
30%
(case
2)
and
the
a
v
erage
60%
(case
3),
respec-
ti
v
ely
.
Figure
10(a)
and
(b)
pro
vides
the
simulation
results
of
po
wer
profile
and
S
O
C
B
E
S
profile,
respecti
v
ely
.
F
or
case
1
and
2,
the
control
scheme
has
adjusted
the
P
S
E
T
0
to
increase
(for
case
1)
and
decrease
(for
case
2)
IJPEDS
V
ol.
8,
No.
4,
December
2017:
1830
–
1840
Evaluation Warning : The document was created with Spire.PDF for Python.
IJPEDS
ISSN:
2088-8694
1837
Figure
8.
Comparison
of
dispatch
performance
of
PV
-BES
using
proposed
control
scheme.
(a)
Po
wer
profile
of
P
P
V
,
P
S
E
T
and
P
G
,
(b)
V
oltage
profile
of
BES,
(c)
SOC
profile
of
BES
and
(d)
Current
profile
of
BES.
Figure
9.
Comparison
of
dP
.
(a)
Before
mitig
ation,
(b)
After
mitig
ation.
as
illustrates
in
Figure
1
0(
a).
The
adjusted
P
S
E
T
0
caused
the
dischar
ge
rate
i
s
increased
and
char
ge
rate
is
de-
creased
if
high
initial
SOC
is
set.
In
contrast,
for
the
case
2,
lo
w
starting
v
alue
of
the
SOC
causes
the
controller
to
adjust
the
P
S
E
T
0
so
that
char
ging
acti
vities
are
more
than
the
dischar
ging.
This
scenario
indicates
that,
through
the
proposed
control
scheme,
the
S
O
C
B
E
S
can
be
restored
to
its
typical
conditions
without
adding
more
ener
gy
storage
capacity
as
illustrated
in
Figure
10(b).
F
or
case
3,
60%
is
assumed
as
nominal
initial
of
S
O
C
B
E
S
.
This
case
is
e
xpected
as
re
gular
operating
condition
of
the
BES
during
clear
day
or
less
impact
of
cloud
co
v
er
to
the
PV
system
output
po
wer
fluctuation.
In
this
case,
P
S
E
T
0
is
remain
unchanged
at
early
stage
due
to
the
ability
of
the
BES
to
char
ge
and
dischar
ge.
As
illustrated
in
Figure
10(b),
S
O
C
B
E
S
is
remain
stable
within
the
controllable
range
that
reflects
the
ef
fecti
v
eness
of
the
proposed
scheme.
5.
CONCLUSION
This
paper
in
v
estig
ates
the
control
design
issues
of
lar
ge
scales
BES
to
be
inte
grated
with
grid-
connected
PV
system
so
that
the
output
po
wer
from
PV
system
can
be
smoothed
out
and
dispat
ched
on
an
Contr
ol
Str
ate
gy
of
a
Grid-connected
Photo
voltaic
with
Battery
Ener
gy
...
(Mohd
Afifi
J
usoh)
Evaluation Warning : The document was created with Spire.PDF for Python.
1838
ISSN:
2088-8694
Figure
10.
Ef
fects
of
initial
S
O
C
B
E
S
to
the
dispatching
performance
of
PV
-BES.
(a)
Po
wer
profile,
(b)
SOC
of
BES
profile.
hourly
basis
lik
e
a
con
v
entional
generator
.
F
or
such
purpose,
the
control
scheme
has
been
proposed
with
the
goal
of
using
BES
to
pro
vide
po
wer
smoothing
of
the
PV
system
while
maintaining
the
B
ES
operational
constraints
at
the
desired
le
v
el.
The
simulation
of
the
proposed
control
scheme
has
been
carried
out
using
the
historical
PV
system
input
data
of
a
site
in
Malaysia.
Besides
that,
determination
size
of
BES
and
BES
modelling
also
has
been
addressed.
From
the
simulation
results,
the
proposed
control
scheme
is
found
to
be
ef
fecti
v
e.
The
results
indicates
that
the
dispatching
performance
w
as
impro
v
ed
and
the
required
operational
constraints
for
BES
ha
v
e
been
met.
The
SOC
of
BES
is
controlled
at
le
v
el
the
same
as
initial
SOC
during
the
end
of
the
day
to
ensure
continuous
po
wer
support
for
ne
xt
day
po
wer
smoothing
operat
ion.
The
proposed
control
scheme
also
can
significant
ly
reduce
the
po
wer
fluctuation
with
the
unacceptable
de
viations
of
10%
PV
capacity
ha
v
e
been
reduced
from
20%
to
less
than
1%.
The
results
from
case
studies
of
the
ef
fects
of
initial
SOC
of
BES
indicates
the
fle
xibility
and
rob
ustness
of
the
control
scheme
that
an
y
le
v
el
of
SOC
can
be
properly
re
gulated
e
v
en
during
the
critical
conditions
of
po
wer
fluctuation
mitig
ation.
A
CKNO
WLEDGMENTS
The
authors
w
ould
lik
e
to
ackno
wledge
Uni
v
ersiti
Malaysia
T
erengg
anu,
Malaysia
and
Ministry
of
Higher
Education
Malaysia
(MOHE)
for
the
financial
support
of
this
research.
This
research
is
supported
by
MOHE
under
the
Fundamental
Researc
h
Grant
Scheme
(FRGS),
V
ot
No.
59418
(Ref:FRGS/1/2015/TK10
/UMT/02/1).
REFERENCES
[1]
S.
Shi
v
ashankar
,
S.
Mekhilef,
H.
Mokhlis,
and
M.
Karimi,
“Mitig
ating
methods
of
po
wer
fluctuation
of
photo
v
oltaic
(pv)
sources–a
re
vie
w
,
”
Rene
wable
Sustainable
Ener
gy
Re
vie
w
,
v
ol.
59,
pp.
1170–1184,
2016.
[2]
S.
K
ouro,
J.
I.
Leon,
D.
V
innik
o
v
,
and
L.
G.
Franquelo,
“Grid-connected
photo
v
oltaic
systems:
An
o
v
ervie
w
of
recent
research
and
emer
ging
pv
con
v
erter
technology
,
”
IEEE
Industrial
Electr
onics
Ma
g-
azine
,
v
ol.
9,
no.
1,
pp.
47–61,
2015.
[3]
M.
Q.
Raza,
M.
Nadarajah,
and
C.
Ekanayak
e,
“On
recent
adv
ances
in
pv
output
po
wer
forecast,
”
Solar
IJPEDS
V
ol.
8,
No.
4,
December
2017:
1830
–
1840
Evaluation Warning : The document was created with Spire.PDF for Python.
IJPEDS
ISSN:
2088-8694
1839
Ener
gy
,
v
ol.
136,
pp.
125–144,
2016.
[4]
M.
P
.
Almeida,
O.
Perpi
˜
n
´
an,
and
L.
Narv
arte,
“Pv
po
wer
forecast
using
a
nonparametric
pv
model,
”
Solar
Ener
gy
,
v
ol.
115,
pp.
354–368,
2015.
[5]
M.
C
ococcioni,
E.
D’Andrea,
and
B.
Lazzerini,
“24-hour
-ahead
forecasting
of
ener
gy
production
in
solar
pv
systems,
”
in
Intellig
ent
Systems
Design
and
Applications
(ISD
A),
2011
11th
International
Confer
ence
on
.
IEEE,
2011,
pp.
1276–1281.
[6]
X.
T
an,
Q.
Li,
and
H.
W
ang,
“
Adv
ances
and
trends
of
ener
gy
storage
technology
in
microgrid,
”
Interna-
tional
J
ournal
of
Electrical
P
ower
&
Ener
gy
Systems
,
v
ol.
44,
no.
1,
pp.
179–191,
2013.
[7]
M.
Y
.
Suberu,
M.
W
.
Mustaf
a,
and
N.
Bashir
,
“Ener
gy
storage
systems
for
rene
w
able
ener
gy
po
wer
sector
inte
gration
and
mitig
ation
of
intermittenc
y
,
”
Rene
wable
and
Sustainable
Ener
gy
Re
vie
ws
,
v
ol.
35,
pp.
499–514,
2014.
[8]
M.
Z.
Daud,
A.
Mohamed,
and
M.
Hannan,
“
An
impro
v
ed
control
method
of
battery
ener
gy
storage
system
for
hourly
dispatch
of
photo
v
oltaic
po
wer
sources,
”
Ener
gy
Con
ver
sion
Mana
g
ement
,
v
ol.
73,
pp.
256–270,
2013.
[9]
S.
T
elek
e,
M.
E.
Baran,
S.
Bhattacharya,
and
A.
Q.
Huang,
“Rule-based
control
of
batt
ery
ener
gy
storage
for
dispatching
intermittent
rene
w
able
sources,
”
IEEE
T
r
ansactions
Sustainable
Ener
gy
,
v
ol.
1,
no.
3,
pp.
117–124,
2010.
[10]
X.
Li,
Y
.
Li,
X.
Han,
and
D.
Hui,
“
Application
of
fuzzy
w
a
v
elet
transform
to
smooth
wind/pv
h
ybrid
po
wer
system
output
with
battery
ener
gy
storage
system,
”
Ener
gy
Pr
ocedia
,
v
ol.
12,
pp.
994–1001,
2011.
[11]
Y
.
Y
e,
R.
Sharma,
and
D.
Shi,
“
Adapti
v
e
control
of
h
ybrid
ultracapacitor
-battery
storage
system
for
pv
output
smoothing,
”
in
ASME
2013
P
ower
Confer
ence
.
American
Society
of
Mechanical
Engineers,
2013,
pp.
V002T09A015–V002T09A015.
[12]
X.
Feng,
H.
Gooi,
and
S.
Chen,
“Hybrid
ener
gy
storage
with
multimode
fuzzy
po
wer
allocator
for
pv
systems,
”
IEEE
T
r
ansactions
Sustainable
Ener
gy
,
v
ol.
5,
no.
2,
pp.
389–397,
2014.
[13]
F
.
Luo,
K.
Meng,
Z.
Y
.
Dong,
Y
.
Zheng,
Y
.
Chen,
and
K.
P
.
W
ong,
“Coordinated
operational
planning
for
wind
f
arm
with
battery
ener
gy
storage
system,
”
IEEE
T
r
ansactions
Sustainable
Ener
gy
,
v
ol.
6,
no.
1,
pp.
253–262,
2015.
[14]
H.
W
u,
M.
Shahidehpour
,
and
M.
E.
Khodayar
,
“Hourly
demand
response
in
day-ahead
scheduling
consid-
ering
generating
unit
ramping
cost,
”
IEEE
T
r
ansactions
on
P
ower
Systems
,
v
ol.
28,
no.
3,
pp.
2446–2454,
2013.
[15]
O.
T
remblay
,
L.-A.
Dessaint,
and
A.-I.
Dekkiche,
“
A
generic
battery
model
for
the
dynamic
simulation
of
h
ybrid
electric
v
ehicles,
”
in
V
ehicle
P
ower
and
Pr
opulsion
Confer
ence
.
IEEE,
2007,
pp.
284–289.
[16]
H.
Zhao,
Q.
W
u,
S.
Hu,
H.
Xu,
and
C.
N.
Rasmussen,
“Re
vie
w
of
ener
gy
storage
system
for
wind
po
wer
inte
gration
support,
”
Applied
Ener
gy
,
v
ol.
137,
pp.
545–553,
2015.
[17]
S.
T
elek
e,
M.
E.
Bar
an,
A.
Q.
Huang,
S.
Bhattacharya,
and
L.
Anderson,
“Control
strate
gies
for
battery
ener
gy
storage
for
wind
f
arm
dispatching,
”
IEEE
T
r
ansactions
on
Ener
gy
Con
ver
sion
,
v
ol.
24,
no.
3,
pp.
725–732,
2009.
BIOGRAPHIES
OF
A
UTHORS
Mohd
Afifi
J
usoh
recei
v
ed
his
Bachelor
de
gree
i
n
Electrical
Engineering
from
Uni
v
ersity
T
eknologi
MARA
(UiTM),
Shah
Alam,
Malaysia
in
2013.
He
is
currently
pursuing
Masters
de
gree
in
Electronic
Ph
ysics
and
Instrumentation
at
the
Uni
v
ersiti
Malaysia
T
erengg
anu,
Malaysia.
The
research
areas
of
Master
de
gree
include
battery
ener
gy
storage
applications,
distrib
uted
generation
and
rene
w
able
ener
gy
.
Contr
ol
Str
ate
gy
of
a
Grid-connected
Photo
voltaic
with
Battery
Ener
gy
...
(Mohd
Afifi
J
usoh)
Evaluation Warning : The document was created with Spire.PDF for Python.