Indonesi
an
Journa
l
of El
ect
ri
cal Engine
er
ing
an
d
Comp
ut
er
Scie
nce
Vo
l.
23
,
No.
1
,
Ju
ly
2021
, p
p.
98
~
109
IS
S
N: 25
02
-
4752, DO
I: 10
.11
591/ijeecs
.v
23
.i
1
.
pp
98
-
109
98
Journ
al h
om
e
page
:
http:
//
ij
eecs.i
aesc
or
e.c
om
Intellig
en
t voltage
regulator f
or distribu
ted
ge
n
eration
-
bas
ed
network
Zw
awi Ham
adouc
he
1
, M
ouni
r Khia
t
2
,
M
uha
m
ma
d
As
ad
I
qb
al
3
1,
2
Scamre
L
abor
at
or
y
,
Depa
r
tment
of El
ec
tr
ic
a
l E
ngine
er
ing, Nat
i
onal
Pol
y
t
ec
hn
ic
School
of
Oran
-
Mauric
e
Audin
(ENPO
-
MA
),
Alger
ia
3
Depa
rtment
MS
Elec
tr
ical Engi
n
ee
ring
,
Com
sats
Univer
sit
y
,
Paki
stan
Art
ic
le
In
f
o
ABSTR
A
CT
Art
ic
le
history:
Re
cei
ved
Feb
8
, 2
021
Re
vised
Jun
8
,
2021
Accepte
d
J
un
1
8
, 202
1
Pow
er
grids
are
bei
ng
tra
nsfor
m
ed
int
o
a
sm
art
distri
bu
ti
on
n
et
work
that
inc
orp
ora
te
s
m
ult
iple
distri
but
ed
ene
rg
y
r
esourc
es
(DERs),
ensuring
stabl
e
oper
ation
and
i
m
prove
d
power
qual
ity
at
th
e
sam
e
ti
m
e.
Ma
n
y
rese
arc
h
pape
rs
have
b
een
publi
shed
in
r
ec
en
t
y
ea
rs
th
at
discuss
the
volt
a
ge
viol
a
ti
o
n
issues
tha
t
emer
ge
from
the
hig
h
pene
tr
a
ti
on
of
distri
bute
d
gene
r
at
ion
(DG
)
.
In
thi
s
pap
er,
w
e
propose
a
new
opti
m
al
vol
ta
g
e
cont
rol
techniq
ue
base
d
on
fee
dforward
n
eu
ral
net
works
(FF
NN
)
to
m
ai
nta
in
a
stable
voltage
profi
le.
MA
TL
AB®/S
imulink®
has
bee
n
used
to
c
a
rr
y
out
th
e
sim
ula
ti
on
.
Th
e
sim
ula
ti
on
resul
t
s
show
the
eff
ici
ency
of
thi
s
m
ethod
in
voltage
c
ontrol
.
The
proposed
appr
oac
h
ensure
d
a
stabl
e
voltage
profil
e
for
the
conside
red
sche
m
es
.
Ke
yw
or
ds:
Ar
ti
fici
al
n
e
ur
a
l netw
ork
Distrib
ution sy
stem
So
la
r
p
a
nel
Vo
lt
age
contr
ol
W
i
nd tu
r
bin
e
Th
is i
s an
open
acc
ess arti
cl
e
un
der
the
CC
B
Y
-
SA
l
ic
ense
.
Corres
pond
in
g
Aut
h
or
:
Zwa
wi H
am
ado
uc
he
Scam
re Lab
ora
tory,
De
par
tm
e
nt of Elect
rical
Enginee
rin
g
Nati
on
al
Po
ly
t
echn
ic
Scho
ol
of Oran
-
M
au
ri
ce A
ud
i
n (ENP
O
-
M
A)
,
A
l
ger
i
a
Em
a
il
: ha
m
ado
uch
e
92zo
ua
ou
i
@g
m
ai
l.co
m
1.
INTROD
U
CTION
The
ea
rlie
r
el
e
ct
rical
syst
e
m
s
le
ver
a
ge
a
ce
nt
rali
zed
m
od
el
.
I
n
this
m
od
el
, th
e
ass
ociat
ed
l
arg
e
po
wer
plants
em
plo
y
hig
h
power
gen
e
rato
rs.
T
he
gen
er
at
ors
are
interco
nnec
t
ed
to
the
unde
rly
ing
tran
sm
issi
on
netw
ork.
Also
,
in
this
m
od
el
,
the
transm
issi
on
syst
em
is
e
m
plo
ye
d
in
the
trans
portat
ion
of
the
ge
ne
rate
d
energy
that
is
us
ua
ll
y
fr
om
t
he
central
po
w
er
plants
to
th
e
resp
ect
ive
c
on
s
um
ers.
It
is
no
te
w
ort
hy
th
at
the
energy
is
usua
ll
y
trans
m
itted
ov
e
r
lo
ng
distances.
Ba
se
d
on
this,
high
-
vo
lt
age
values
a
r
e
dem
and
ed
f
or
the
eff
ect
ive
ness
of
the
syst
em
.
At
the
substat
ion
s
that
a
re
i
n
cl
os
e
pro
xim
i
ty
to
the
c
ons
um
ption
ce
nters,
t
he
energy
ca
n
be
distrib
uted
via
seve
ral
li
nes
with
r
el
at
ively
lowe
r
-
rated
powe
r
a
nd
s
hort
er
le
ngth
.
I
n
ge
ner
al
,
the
co
nv
e
ntio
na
l
distribu
ti
on
syst
e
m
m
od
el
exp
loit
s
a
powe
r
fl
ow
th
a
t
is
based
on
a
on
e
way
fro
m
the
transm
issi
on
to
the
distrib
utio
n netw
ork (D
N
)
[
1
]
, [
2].
More
ov
e
r,
w
it
h
c
urren
t de
velop
m
ent,
the
e
xi
sti
ng
ce
ntrali
zed
m
od
el
co
nc
ept
ha
s
bee
n
re
viewe
d
with
grow
i
ng
at
te
nti
on
in
the
distri
bu
te
d
m
od
el
.
This
is
i
n
a
n
e
ffor
t
t
o
harnes
s
va
rio
us
ene
r
gy
res
ource
s
i
n
wh
ic
h
sm
a
ll
po
wer
plants
can
be
de
plo
ye
d
i
n
a
dis
per
se
d
m
ann
er
.
This
idea
is
ty
pical
ly
ref
err
ed
to
as
a
distr
ibu
te
d
gen
e
rati
on
(DG)
.
It
sho
uld
be
no
te
d
t
hat
the
a
forem
entio
ne
d
e
ne
rg
y
r
eso
ur
ces
are
usual
ly
associat
ed
wit
h
ren
e
wa
ble
ene
rg
ie
s
with
a
ty
pical
po
we
r
rati
ng
lo
wer
than
50
MW
[
1
]
-
[
3].
Furtherm
or
e,
the
re
are
po
s
sibil
it
ie
s
of
D
G
un
it
c
onne
ct
ion
s
t
o
the
syst
e
m
.
This
c
an
be
re
al
iz
ed
at
diff
e
re
nt
vo
lt
age
le
vels
th
at
are
us
ua
ll
y from
the low to hig
h v
oltage.
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
In
te
ll
igent v
oltag
e
re
gu
l
ato
r f
or
distri
bute
d gen
e
ra
ti
on
-
bas
ed netw
or
k (
Z
wawi
Hama
douch
e)
99
Additi
on
al
ly
,
t
he
DG
i
nteg
rati
on
t
o
the
netw
ork
c
hanges
t
he
syst
e
m
con
tr
ol.
Be
sides
,
it
changes
t
he
qu
al
it
y
of
el
ect
rical
power,
r
esulti
ng
i
n
ce
rtai
n
co
ns
trai
nts
on
the
operati
on
an
d
functi
onin
g
of
t
he
networ
k
[1
]
-
[
4].
This
c
an
be
at
trib
ute
d
to
the
fact
t
hat
the
ne
w
producti
ons
by
t
he
DG
we
re
not
co
ns
i
der
e
d
in
the
desig
n
of
c
urre
nt
syst
em
s
[4
]
.
In
ce
rtai
n
cas
es,
the
im
ple
m
entat
ion
ca
n
bri
ng
a
b
ou
t
a
n
el
ect
ric
powe
r
qual
it
y
ind
ic
es
dete
rio
rati
on.
O
n
the
oth
e
r
ha
nd,
it
can
res
ult
in
in
dices
i
m
pr
ove
m
ent.
Fo
r
i
ns
ta
nce,
the
al
te
r
at
ion
in
the
syst
e
m
ca
n
res
ult
in
the
occurre
nce
of
reacti
ve
ene
r
gy
flo
w,
as
w
el
l
as
the
network
a
nd
tra
nsfo
rm
er
ov
e
rloa
ds
.
Be
sides,
the
syst
e
m
is
pr
one
to
ove
r
-
vo
lt
age
s
and
subse
quently
,
it
resu
lt
s
in
powe
r
qual
it
y
deterio
rati
on
[
5].
I
n
co
ntrast
,
de
plo
ym
ent
of
the
bac
k
-
up
un
it
s
nea
r
en
ough
to
the
c
enter
w
he
re
th
ey
are
dem
and
ed hel
ps i
n
sa
ving t
he t
ran
sm
issi
on
pow
e
r
e
xpense
as
well
as t
he r
el
at
ed
tran
sm
is
sion l
os
ses
[1].
As
a
forem
entio
ne
d,
the
re
is
a
cl
os
e
c
orrelat
ion
bet
wee
n
D
G
a
nd
re
ne
wabl
e
energies.
I
n
this
co
ntex
t
,
there
has
been
a
nota
ble
de
vel
op
m
ent
in
the
wind
a
nd
sun
r
enew
a
ble
e
nergies.
Also,
thei
r
em
plo
ym
ent
in
the
sys
tem
pr
esents
a
chall
eng
e
no
t
only
reg
a
r
ding
the
eff
ic
i
ent
and
sa
fe
operati
on
but
al
so
on
the
unde
rly
ing
netw
ork
c
ontr
ol.
T
o
s
om
e
extent,
t
he
iss
ue
ca
n
be
a
ddress
ed
by
m
eans
of
m
ic
ro
gr
i
ds
.
Mi
cr
ogr
ids
are
el
e
m
ents
that
m
anag
e
res
our
ces
in
a
distrib
uted
ene
rg
y
in
a
m
or
e
dep
e
ndable
an
d
dece
ntrali
zed
way.
Ba
sed
on
this
,
not
on
ly
the
con
tr
ol
bur
den
on
the
gr
i
d
can
be
re
du
ce
d
bu
t
al
so
their
fu
ll
be
ne
fits
can
be
ef
fe
ct
ively
exp
l
oited
[1
]
-
[
6].
Mo
re
ov
e
r,
dep
e
ndin
g
on
t
he
gen
e
rati
on
a
nd
loa
d
ty
pe
be
ing
de
plo
ye
d,
m
ic
ro
gri
ds
ca
n
be
ac
or
dc
[
7].
Als
o,
powe
r
flo
w
can
be
c
ontr
ol
le
d
in
op
ti
m
a
l
m
ann
ers
bet
ween
the
m
ic
ro
gri
d
an
d
t
he
public
netw
ork
us
i
ng
an
appr
opria
te
interface.
Be
sides,
a
sm
art
gr
id
has
been
prese
nt
ed
to
facil
it
at
e
sel
f
-
m
anag
em
ent in th
e
netw
ork [8]
.
Con
ce
pt
ually
,
a
s
m
art
gr
id
offe
rs
dy
nam
ic
op
ti
m
iz
ation
usi
ng
real
-
ti
m
e
m
easur
em
ents
to
en
han
ce
the
syst
em
per
form
ance
and
ens
ur
e
e
ff
ect
iv
e
m
anag
em
ent
in
va
rio
us
as
pects
su
c
h
as
losses,
sec
uri
ty
,
an
d
vo
lt
age
le
vels.
Ba
sed
on
t
he
data
gat
her
e
d
t
hro
ugh
the
sm
art
gri
d
a
nd
it
s
subsyst
em
s,
t
he
be
st
ap
proa
ch
to
ens
ur
e
s
uitable
syst
e
m
op
erati
on
by
the
syst
em
op
erat
or
s
ca
n
be
prom
ptly
identifie
d
[8
]
,
[
9].
T
her
e
a
re
so
m
e
effor
ts
in
w
hic
h
the
insertio
n
of
D
G
energ
y
pr
oduce
rs
in
to
the
DN
s
ha
ve
bee
n
co
ns
idere
d
to
analy
ze
the
s
yst
e
m
per
for
m
ance.
O
ne
of
su
c
h
s
hows
the
be
nef
it
s
of
D
G
instal
la
ti
on
s
f
or
po
wer
ge
ner
at
io
n
in
t
he
distrib
ution
sy
stem
.
The
w
ork
al
s
o
pr
ese
nt
s
nota
ble
en
ha
ncem
ent
that
c
an
be
ac
hieve
d
in
the
dist
ribu
ti
on
syst
e
m
’s
vo
lt
a
ge
prof
il
e
as
well
as
a
redu
ct
ion
in
th
e
el
ect
ric
syst
e
m
l
os
ses
[
10
]
.
Als
o,
t
he
e
ff
ect
s
of
wind
energy
tu
rb
i
ne
s
co
nn
ect
io
n
on
the
volt
age
pro
file
are
c
on
si
der
e
d
in
[
11
]
.
Be
sides,
the
lo
ad
fl
ow
pro
ba
bi
li
stic
te
chn
iq
ue
is
e
m
plo
ye
d
to
stu
dy
the
relat
ed
eff
ect
s
on
volt
age
qual
it
y
when
wi
nd
t
urbin
es
are
interc
on
nected
with
the
D
Ns
[
11
]
,
[
12]
.
Li
ke
wise,
the
analy
ti
cal
m
e
tho
d
im
ple
m
entat
ion
s
for
in
flue
nce
s
of
the
wind
powe
r
plant
(
WPP
)
on th
e
syst
em
’s
reli
abili
ty
h
ave
b
ee
n pr
ese
nte
d [12].
More
ov
e
r,
vol
ta
ge
con
tr
ol
can
be
ind
e
pe
nd
e
nt
-
base
d
and
c
ooper
at
i
ve
-
base
d
[
13
]
.
Likewise,
a
gen
et
ic
c
ode
-
ba
sed
al
gorithm
can
al
so
be
e
m
plo
ye
d
for
volt
age
regulat
ion
[
14]
.
Also,
vo
lt
age
c
on
t
ro
l
can
be
achieve
d
by
e
m
plo
yi
ng
an
a
ppr
oach
that
is
base
d
on
reacti
ve
po
wer
(RP
)
c
om
pen
sat
io
n.
Sim
il
arly
,
a
m
et
ho
d
that
em
plo
ye
d
st
at
ist
ic
al
analy
sis
and
ca
n
be
us
ed
f
or
re
gu
la
ti
ng
the
volt
age
a
fter
th
e
ene
rg
y
pro
ducer
’
s
insertio
n
ha
s
be
en
prese
nted
[15]
Be
sides,
an
al
gorithm
t
hat
can
be
em
plo
ye
d
f
or
c
ontrolli
ng
t
he
vo
lt
age
us
in
g
RP
of
en
erg
y
pro
ducers
has
been
pr
es
ented
[
16
]
.
Als
o,
i
t
sh
ould
be
no
te
d
that
in
a
giv
en
netw
ork,
t
he
op
ti
m
al
sitting
of the
ge
ner
at
ors is im
per
at
iv
e to im
pr
ove th
e volt
age
prof
il
es [17
]
.
In
this
pa
per
,
we
prese
nt
a
ge
ner
al
iz
ed
m
od
el
f
or
a
DN
t
ha
t
is
interconne
ct
ed
with
a
s
olar
pa
nel
-
ty
pe
and
wi
nd
tur
bi
ne
po
we
r
ge
ne
rators
to
anal
yz
e
the
va
ried
natu
re
of
the
volt
age.
I
n
a
dd
it
io
n,
we
c
onside
r
eff
ect
ive
m
eans
for
optim
al
r
egu
la
ti
on
by
le
ver
a
ging
an
a
rtific
ia
l
neu
ral
ne
twork
(
A
NN) techn
iq
ue.
E
xt
ensiv
e
si
m
ulati
on
s
ar
e
carried
out
with
MATL
A
B®/
Si
m
ulink
®
to
substanti
a
te
and
s
upport
the
pr
e
sente
d
con
t
ro
l
m
od
el
s.
The
rest
of
this
pap
e
r
is
or
ga
nized
a
s
fo
ll
ows:
in
Sect
ion
2,
the
e
m
plo
ye
d
m
a
the
m
at
ic
al
m
od
el
s
fo
r
el
ect
rical
ener
gy
producti
on
with
the
m
a
in
fo
c
us
on
the
s
olar
pa
nel
-
ty
pe
ener
gy
ge
ne
ra
tor
as
well
as
a
wind
powe
r
pro
duce
r
are
c
onside
re
d.
Also,
dif
fere
nt
de
velo
pm
e
nt
sta
ges
of
i
nt
el
li
gen
t
vo
lt
ag
e
regulat
ors
f
or
the
PD
Es
a
re
co
nsi
der
e
d
in
Sect
i
on
3.
In
Sect
io
n
4,
we
pr
e
sen
t
resu
lt
s
an
d
di
scussion.
I
n
Se
ct
ion
5,
we
present
the concl
udi
ng r
em
ark
s.
2.
RESEA
R
CH MET
HO
D
Th
is
sect
io
n
presents
t
he
em
plo
ye
d
m
at
hem
at
ic
al
m
od
els
for
el
ect
rical
energy
pr
oduc
ti
on
.
In
t
hi
s
reg
a
rd,
we
f
oc
us
on
the
so
la
r
pan
el
-
ty
pe
e
ne
rg
y
ge
ner
at
or
that
is
cou
ple
d
to
the
gri
d
vi
a
the
inv
erter
syst
e
m
and a
wind
power
pro
ducer t
hat is co
uple
d
t
o
the
gr
id
us
i
ng a sy
nchr
onou
s g
e
ner
at
or. A
l
so
, we
de
velo
pe
d
a
nd
m
od
el
ed
a
ty
pi
cal
con
fi
gurati
on
of
these
generator
s
us
in
g
MATLAB
®/Si
m
ul
ink
®
platf
or
m
.
Fu
rthe
rm
or
e
,
w
e
pr
ese
nt a
nd explai
n
a
sim
plifi
ed diag
ram
o
f a
ty
pical
contr
ol for eac
h ge
ne
rator.
2.1
.
Win
d
tur
bine m
od
el
The
interact
io
n
betwee
n
the
wind
ro
t
or
an
d
wind
is
usu
al
ly
e
m
plo
ye
d
fo
r
represe
nting
the
wind
tur
bin
e
ae
rody
nam
ic
s.
The
fe
at
ur
es
of
t
his
a
erodynam
ic
ca
n
be
desc
ribe
d
us
in
g
the
disc
t
heory.
Ba
se
d
on
the
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
:
98
-
109
100
theo
ry,
the
a
va
il
able
wind
powe
r,
,
ai
m
e
d
at
a
pa
rtic
ul
ar
disc
s
wep
t
by
m
eans
of
the
r
otor
ca
n
be
determ
ined.
B
esi
des,
the
c
onnecti
on
betwee
n
t
he
an
d
the
extracte
d
pow
er
from
the
r
oto
r
,
,
is
def
i
ne
d
thr
ough the
d
is
c theory
. Also
, t
he
insta
ntane
ous
powe
r,
, ca
n be e
xpres
sed
a
s [
4
]
=
1
2
ρ
3
(1)
wh
e
re
ρ
de
no
t
es
the
ai
r
den
s
it
y
in
kg
/m
3
,
a
re
pr
ese
nts
th
e
swe
pt
area
by
the
ro
t
or
in
m
2
,
an
d
v
de
no
t
es
the
wind
velocit
y i
n
m
/s. A
lso,
th
e w
in
d
t
urbine
’
s pow
e
r
c
oe
ff
i
ci
ent is d
e
fine
d as
(2)
=
(2)
m
or
eov
er
, usin
g (
1
)
a
nd (
2
)
, t
he
e
xtracted
po
wer f
ro
m
the roto
r
ca
n be
defi
ned
a
s [1]
=
1
2
2
3
(3)
wh
e
re
r den
ote
s the
ro
t
or r
a
diu
s
of the
wind
tur
bin
e i
n
m
.
Also
,
the
ex
e
rted rotat
io
nal to
rque
on the
turbine
by the
w
i
nd can
b
e
ex
pr
essed de
fine
d
a
s
(
4)
=
Ω
(4)
wh
e
re
Ω
repr
esents
the
tur
bin
e
r
otor’s
a
ngular
m
echani
cal
velocit
y
i
n
rad
/s
.
It
sho
uld
be
no
te
d
that
the
aforem
entione
d param
et
ers
are em
plo
ye
d
in
the
wind tu
r
bin
e m
od
el
in
g.
In
gen
e
ral,
the
conve
ntion
al
s
yst
e
m
dep
en
ds
on
t
he
el
ect
rici
ty
delivery
from
the
po
we
r
pro
duce
r
plant
to
res
pe
ct
ive
custom
ers
thr
ough
the
DN.
T
his
res
ul
ts
in
un
i
direc
ti
on
al
flo
ws
.
I
n
co
ntra
st,
the
D
G
integrati
on
res
ults
in
bid
irect
ion
al
flo
ws
of
the
associat
ed
act
ive
power
(
AP
)
i
n
the
D
N
s.
Furthe
rm
or
e,
in
a
scenari
o
w
he
re
the
pro
duct
io
n
s
urpasses
t
he
consum
ption,
the
fl
ows
ca
n
be
t
owar
d
th
e
trans
port
net
works.
This
sit
uation
can
prese
nt
a
con
si
der
a
ble
inf
luence
on
the
m
at
erial
s
that
are
norm
al
ly
un
idirect
io
nal
s
uch
a
s
equ
i
pm
ent
prot
ect
i
on
,
an
d
m
easur
em
ent
de
vi
ces,
in
the
D
N
s.
Mo
re
ov
e
r,
a
powe
r
flo
w
reversal
in
the
gri
d
ca
n
be
induce
d
by
the
DG
c
onnec
ti
on
,
res
ulti
ng
in
bid
irect
io
nal
flow
s
.
Co
ns
eq
uen
tl
y,
D
G
pre
sents
com
patibil
ity
issue
bet
wee
n
the
cu
rr
e
nt
ne
twork
a
nd
th
e
introd
uced
e
nergy.
T
his
de
m
and
s
m
od
ific
at
ion
of
t
he
e
xisti
n
g
el
ect
rical
n
et
w
ork prote
ct
io
n plan
[
4
].
More
ov
e
r,
t
he
DG
i
nteg
rati
on
can
le
a
d
to
a
r
ise
in
the
te
ns
i
on
w
hich
can
r
esult
in
a
n
ov
e
rvolta
ge
i
n
the
net
wor
k.
F
or
instance
,
ba
sed
on
(
2),
for
co
nn
ect
i
on
s
be
tween
a
sin
gl
e
powe
r
ge
nerat
or
a
nd
the
node
,
N,
the volt
age
drop
betwee
n
the
N
c
onnecti
on
point a
nd the
sour
ce
stat
ion
ca
n be e
xpresse
d as [
4
]
∆
=
(
−
)
+
(
±
−
±
)
(5)
=
−
(6)
=
±
−
±
(7)
wh
e
re
,
,
a
nd
are
the
AP
at
no
de
N,
A
P
s
uppl
ie
d
by
t
he
ge
ne
rator,
a
nd
AP
consum
ption
,
r
especti
vely
,
,
,
and
rep
r
ese
nt
the
RP
at
no
de
N,
RP
sup
plied
by
th
e
ge
ne
rator,
a
nd
R
P
c
on
s
um
ption
,
re
sp
ect
ivel
y,
and
denotes t
he
RP c
om
pen
sat
ion
dev
ic
e
.
More
ov
e
r,
to
dem
on
strat
e
th
e
i
m
pact
of
th
e
D
G
inte
gr
at
i
on,
we
c
onsid
er
a
sc
ena
rio
with
a
ty
pical
netw
ork
that
ha
s
N
nodes
a
nd
wit
h
co
nnect
ed
N
loa
ds
as
dep
ic
te
d
in
Fi
gure
1.
D
ur
i
ng
distrib
ution,
th
ere
will
be
vo
lt
a
ge
dro
ps
am
ong
the
s
ource
a
nd
c
onne
ct
ion
point, Nj.
Th
e
ass
ociat
ed
vo
lt
a
ge
dro
p
ca
n
be
e
xpre
ssed
a
s
[
18
]
.
(
)
=
∑
(
∑
)
.
+
∑
(
∑
)
.
<
=
1
=
1
<
=
1
=
1
(8)
wh
e
re
P
j
an
d
Q
j
den
ote
the
AP
an
d
RP
at
the
node
N
j
,
U
so
urce
is
the
upstream
vo
lt
age
of
the
short
-
ci
rcu
it
i
m
ped
ance
(
R
1
; X
1
), j
=
1
,
2
,
….
n,
a
nd
n de
no
te
s
the
node
s.
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
In
te
ll
igent v
oltag
e
re
gu
l
ato
r f
or
distri
bute
d gen
e
ra
ti
on
-
bas
ed netw
or
k (
Z
wawi
Hama
douch
e)
101
Figure
1
.
LV
net
work with
out D
G
inte
gr
at
io
n
2.2
.
P
V
cel
l
m
od
el
The
PV
cel
l
can
be
m
od
el
ed
us
ing
a
diod
e,
a
ph
oto
c
urre
nt
so
urce,
a
pa
rall
el
resist
or
,
and
a
series
resist
or.
In this
context, t
he o
utput cu
rr
e
nt
of the
P
V
cel
l c
an be
def
i
ned a
s [
19
]
, [
20
]
=
ℎ
−
[
exp
(
+
)
−
1
]
−
+
ℎ
(9)
wh
e
re
Iph
is
t
he
s
hort
-
ci
rc
ui
t
current
due
to
s
un
li
ght
(
ph
otons),
I
d
de
note
s
the
s
hunt
ed
c
urren
t
t
hro
ugh
t
he
diode,
V
denot
es
the
vo
lt
age
on
the
lo
ad
,
I
i
s
the
c
urren
t
t
hr
ou
gh
t
he
loa
d,
Rs
an
d
Rs
h
re
pr
ese
nts
the
pa
rasit
ic
series a
nd shu
nt
r
esi
sta
nce
of t
he
a
rr
ay
, res
pe
ct
ively
, and V
T is g
i
ven as
[
20
]
=
(10)
wh
e
re
k
de
not
es
the
Bolt
zm
a
n’
s
c
onsta
nt
(1.38×1
0−
23
J/K
),
q
is
the
el
ec
tro
n
cha
r
ge
(
1.6×10−
19
C
)
a
nd
T
c
denotes t
he
te
m
per
at
ur
e of t
he op
e
rati
ng m
odule
(
°
C)
[
21
]
.
Fu
rt
her
m
or
e,
the
photo
c
urre
nt
Iph
is
a
f
un
ct
ion
of
the
s
ola
r
irra
dia
nce
re
cei
ved
by
the
so
la
r
cel
l
as
well
as its t
em
per
at
ur
e a
nd ca
n be e
xpresse
d as [
19
]
-
[
22
]
ℎ
=
[
+
(
−
)
]
.
(11)
wh
e
re
α
=
0:0012
×
Isc
denotes
the
te
m
per
at
ur
e
coe
ff
ic
i
ent
in
(
A/K
),
I
sc
is
the
s
hort
-
ci
rcu
it
c
urren
t
,
Tre
f
represe
nts at re
fer
e
nce c
onditi
on
s
(2
98 K) a
nd
G denote
s t
he
so
la
r
irr
a
dia
nc
e level
(k
W
/
m
2
).
More
ov
e
r,
to
de
velo
p
a
P
V
m
o
dule
,
there
is
a
nee
d
f
or
so
la
r
cel
ls.
These
a
re
us
ua
ll
y
inter
connecte
d
in
series
a
nd
pa
rall
el
.
W
it
h
re
gards
t
o
a
sin
gl
e
-
cel
l
ci
rcu
it
m
od
ule,
the
rel
at
ion
s
hip
betw
een
t
he
P
V
m
od
ule
’
s
vo
lt
age
and
cu
rr
e
nt can be
e
xpress
ed
as
[
19
]
=
ℎ
−
[
{
(
+
)
}
−
1
]
−
+
ℎ
(12)
wh
e
re
an
d
re
pr
ese
nt the
s
olar
cel
ls wit
hi
n a m
od
ule that a
re c
onnected
in
ser
ie
s a
nd
paral
le
l.
2.3
.
Im
pa
c
t o
f
t
he inser
tio
n
of DG
on the
vo
l
tage
profi
le
In
t
his
s
ubsect
ion,
we
dem
on
s
trat
e
the
DG
im
ple
m
entat
ion
i
m
pact
us
in
g
a
3−
ph
a
se
L
V
syst
e
m
.
The
syst
e
m
is
m
od
el
ed
as
well
a
s
sim
ulate
d
on
the
M
ATL
A
B®/
Si
m
ulink
®
platf
or
m
as
il
lustrate
d
in
Figure
2.
Also
,
we
em
plo
y
(8)
to
e
val
ua
te
the
volt
age
dro
p
at
the
resp
ect
ive
no
de.
F
ur
t
her
m
ore,
the
net
wor
k
unde
r
consi
der
at
io
n
com
pr
ise
s
14
nodes
a
nd
10
loads
(P
a
nd
Q)
.
Mo
reover
,
a
160−
kVA
,
20
=
0:4
−
kV
ste
p
-
dow
n
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
:
98
-
109
102
trans
form
er
is
us
e
d
to
s
upply
the
loa
ds
.
We
us
e
this
netw
ork
as
a
case
st
udy
due
to
t
he
e
asy
identific
at
ion
of
the
surge
ph
e
no
m
enon.
Be
s
ides,
at
the
co
ns
ide
red
volt
age
le
vel,
the
r
especti
ve
li
nea
r
resist
ance
is
m
or
e
i
m
per
at
ive
co
m
par
ed
with
t
he
relat
ed
li
ne
ar
reacta
nc
e
(i
.e.
R
X
).
It
s
houl
d
be
note
d
that
we
co
ns
i
de
r
tw
o
zon
e
s
(i.e.
1
and
2)
a
nd
t
he
r
especti
ve
zo
ne
denotes
a
distrib
ution
li
ne.
Also
,
t
o
dem
on
strat
e
the
D
G
i
m
pact,
we
c
onsider
sc
enar
i
os
withou
t and with
DG
insertio
n.
Figure
2
.
MAT
LAB®/Si
m
uli
nk® sim
ulati
on
of LV net
work with
out D
G i
ntegr
at
io
n
a)
W
i
t
h
o
u
t
D
G
:
T
h
e
v
o
l
t
a
g
e
d
r
o
p
a
t
t
h
e
r
e
s
p
e
c
t
i
v
e
n
o
d
e
i
s
e
v
a
l
u
a
t
e
d
u
s
i
n
g
t
h
e
v
o
l
t
a
g
e
p
r
o
f
i
l
e
w
i
t
h
o
u
t
D
G
i
n
F
i
g
u
r
e
2
.
I
n
t
hi
s
s
c
e
n
a
r
i
o
,
5
0
%
a
n
d
1
0
0
%
l
o
a
d
s
t
a
t
e
s
a
r
e
c
o
n
s
i
d
e
r
e
d
.
S
i
n
c
e
n
o
D
G
i
s
e
m
pl
o
y
e
d
,
t
h
e
n
e
t
w
o
r
k
i
s
b
a
s
e
d
o
n
a
c
o
n
v
e
n
t
i
o
n
a
l
o
p
e
r
a
t
i
o
n
i
n
w
h
i
c
h
v
o
l
t
a
g
e
d
r
o
p
s
o
c
c
u
r
b
e
t
w
e
e
n
t
h
e
s
o
ur
c
e
s
t
a
t
i
o
n
a
n
d
t
h
e
c
o
n
s
um
p
t
i
o
n
p
o
i
n
t
s
.
S
i
m
u
l
a
t
i
o
n
r
e
s
u
l
t
s
p
r
e
s
e
n
t
e
d
i
n
F
i
g
u
r
e
3
s
h
o
w
s
t
h
a
t
f
o
r
a
f
u
l
l
l
o
a
d
(
1
0
0
%
)
s
i
t
u
a
t
i
o
n
i
n
b
o
t
h
Z
o
n
e
1
(
3
(
a
)
)
a
n
d
Z
o
n
e
2
(
3
(
b
)
)
,
t
h
e
v
o
l
t
a
g
e
g
e
t
s
t
o
l
o
w
v
a
l
u
e
s
t
o
w
a
r
d
s
t
h
e
e
n
d
o
f
t
h
e
l
i
n
e
.
b)
W
it
h
D
G:
In
this
par
t
,
we
c
onside
r
the
inse
rtion
of
a
so
la
r
pan
el
-
ty
pe
(
no
de
11)
an
d
wi
nd
tur
bin
e
(
node
14)
po
wer
generator
s
into
th
e
netw
ork
at
zon
e
2.
T
he
vol
ta
ge
dro
p
at
the
ind
ivi
du
al
node
is
e
valuate
d
us
in
g
the
volt
age
pr
of
il
e
with
DG
in
Fi
gure
4.
M
or
e
over,
half
(50%
)
an
d
f
ull
(
100%
)
load
sta
te
s
a
re
consi
der
e
d
in
the
si
m
ulati
on
.
The
resu
lt
s
presente
d
in
Figure
5
sho
w
that
apar
t
from
the
su
bs
ta
ntial
su
r
ge
i
n
the
volt
age
at
the
po
i
nts
w
he
re
t
he
DG
is
co
uple
d,
the
re
is
a
nota
ble
inc
r
ease
acr
os
s
t
he
neig
hbori
ng
node
s
as
well
.
Be
sides,
r
ega
r
ding
the
half
(50%
)
loa
d
st
at
es,
the
vo
lt
a
ge
e
xceed
s
th
e
acce
ptable
volt
age
lim
it
.
The
crit
ic
al
le
vel
is
exp
e
rience
d
from
no
de
10
to
node
14
as
de
pi
ct
ed
in
Figure
5(b)
in
Z
one
2.
The
e
xcess
vo
lt
age
is
owin
g
to
the
D
G
im
pl
e
m
entat
ion
a
nd
propo
r
ti
on
al
to
the
delive
re
d
powe
r.
C
onseq
uen
tl
y, the
DG
i
m
ple
m
entat
ion
ca
n resu
lt
i
n ov
e
r
vo
lt
age
on
the
vo
lt
ag
e
prof
il
e.
Figure
3. V
oltage
prof
il
e
for 5
0%
a
nd
100%
l
oad stat
es w
it
hout
D
G
;
(a
)
z
one
1
a
nd (b) zo
ne 2
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
In
te
ll
igent v
oltag
e
re
gu
l
ato
r f
or
distri
bute
d gen
e
ra
ti
on
-
bas
ed netw
or
k (
Z
wawi
Hama
douch
e)
103
Figure
4
.
MAT
LAB®/Si
m
uli
nk® sim
ulati
on
of LV net
work with
integ
ra
ti
on
of so
la
r pa
nel
-
ty
pe (n
ode
11)
and w
i
nd tu
rb
i
ne (n
ode
14) D
G
Figure
5
.
V
oltage
prof
il
e
for 5
0%
a
nd
100%
l
oad stat
es w
it
h D
G
;
(a
)
z
on
e
1
a
nd (b) zo
ne 2
2.4. In
tell
ige
nt
vo
l
tage re
gul
ators f
or
t
he
pd
es
The
sect
io
n p
r
es
ents
diff
e
re
nt d
e
velo
pm
ent stages
of
i
ntell
igent
volt
age
re
gu
la
to
rs f
or
t
he
PD
E
s.
2.4.1
.
Ar
tifici
al
neural
netw
or
k
(
ANN
)
The
3
-
P
hase
L
V
syst
e
m
si
mu
la
ti
on
prese
nt
ed
in
subsect
i
on
2.3
sho
ws
that
the
conne
ct
ion
of
D
G
un
it
s
bet
ween
loads
ca
n
caus
e
crit
ic
al
vo
lt
a
ge
re
gu
la
ti
on
pro
blem
s.
Con
seq
uen
tl
y,
to
offe
r
a
prop
e
r
volt
ag
e
regulat
ion
in
t
he
D
N,
we
e
m
plo
y
an
ANN
te
chn
i
que.
An
ANN
is
a
bio
lo
gical
ly
m
otivate
d
c
om
pu
ta
ti
on
al
relat
ed
m
od
el
[
23
]
.
It
com
pr
ise
s
neur
on
s
th
at
are
pr
oc
essing
el
em
ents.
The
ne
uro
ns
are
con
nected
to
ge
the
r
and
there
are
c
oeffici
ents
(
we
igh
ts)
that
are
placed
on
the
l
ink
s
.
I
n
this
w
ork,
fee
dforwa
rd
ne
ur
al
netw
orks
(F
F
NN)
is
em
plo
ye
d
beca
us
e it
is
the
m
os
t
widely
us
e
d
m
od
el
in
a num
ber
of
pract
ic
al
app
li
cat
io
ns
[
24
]
.
Th
e
schem
at
ic
o
f
F
FNN
with
the
backp
ropa
gation al
gorithm
is d
e
picte
d
in
Fi
gure
6
. A
n
F
F
NN co
ntains
a
chain o
f
la
ye
rs.
For
instance,
it
com
pr
ise
s
diff
e
re
nt
tiers
su
c
h
as
the
inp
ut,
hidden
,
and
outp
ut
la
ye
rs.
The
act
iv
at
ion
functi
ons
for
com
pu
ti
ng
the
op
ti
m
u
m
weigh
ti
ng
of
inpu
t
bo
un
ds
re
ga
rd
i
ng
the
ta
r
ge
t
ou
tp
ut
valu
es
are
con
ta
ine
d
in
t
he
hidde
n
la
ye
rs.
Furthe
rm
or
e,
t
o
a
da
pt
w
ei
gh
ts
a
nd
bia
ses
withi
n
t
he
in
div
id
ual
la
ye
r,
a
backp
ropa
gation
al
go
rithm
is adopted fo
r
th
e trai
nin
g
sc
he
m
e. Th
is i
s in
an
effo
rt to
le
ssen
the er
rors bet
wee
n
the
net
work
outp
ut
an
d
ta
r
ge
t
value
for
th
e
input
-
base
d
factors
[
24
]
,
[
25
]
.
M
or
e
over
,
the
em
plo
ye
d
FFNN
m
od
el
co
m
pr
ise
s
one
in
pu
t
la
ye
r,
tw
o
hi
dd
e
n
la
ye
rs,
a
nd
one
outp
ut
la
ye
r
.
Also,
the
i
nput
as
well
as
outp
ut
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
:
98
-
109
104
la
ye
r
di
m
ension
s
are
al
ike
to
the
input
and
ta
rg
et
pa
ram
et
e
r
nu
m
ber
s.
On
the
oth
er
ha
nd,
dim
ension
s
of
th
e
hidden
la
ye
r
a
r
e
ada
pted
m
an
ually
with
respec
t
to
the
m
od
el
perform
ance
[
23
]
.
The
net
work
m
od
el
w
ei
gh
ts
and
a
sso
ci
at
e
d
biases
are
i
niti
al
iz
ed
by
m
eans
of
t
he
MAT
L
AB
®
Neural
Netw
ork
T
oo
l
box.
Al
s
o,
the
m
od
e
l
is
update
d
t
hro
ugh
t
he
c
onfi
gure
d
functi
ons
su
c
h
as
ada
ption
le
ar
ning,
tr
ai
nin
g,
act
ivati
on,
a
nd
perfor
m
ance.
In
a
ddit
ion,
si
nce
the
de
vel
op
e
d
m
od
el
is
base
d
on
FF
NN
with
the
backp
ropa
gation
al
gorithm
,
and
the
al
gorithm
us
ua
ll
y
exp
erie
nces
no
t
only
a
slo
w
to
c
onve
r
ge
bu
t
al
s
o
a
n
ove
rf
it
ti
ng
pro
ble
m
[
23
]
,
a
le
ve
nb
e
r
g
-
m
arq
ua
rd
t
(L
M)
re
gula
rizat
ion
te
c
hn
i
que
[
25
]
,
[
26
]
is
e
m
plo
ye
d
to
a
ddress
the
issue
s.
T
his
is
du
e
to
it
s
relat
ively
swifter
co
nver
gence
in
the
ba
ck
pro
pag
at
io
n.
Like
wise,
the
“t
r
ai
n
lm
”
trai
nin
g
f
un
ct
io
n
is
em
plo
ye
d
wh
il
e
the
“m
s
e”
and
“l
ear
ngdm
”
in
the
t
oo
l
box
are
ch
os
e
n
for
the
pe
rfor
m
ance
functi
ons
an
d
adap
ti
on
le
arn
in
g,
re
sp
e
ct
ively
.
Be
sides,
to
achieve
it
s
ou
tp
ut,
any
di
ff
ere
ntiable
act
ivati
on
f
un
ct
i
on
ca
n
be
em
plo
ye
d
by
each
ne
u
r
on.
T
he
act
ivati
on
f
un
ct
io
ns
a
r
e
locat
ed
in
t
he
hidden
la
ye
r
and
outp
ut
la
ye
r
ne
uro
ns
,
but
no
t
in
the in
pu
t l
ay
er
neur
on
s
.
Figure
6.
Sc
he
m
at
ic
o
f
fee
dfo
rw
a
rd n
e
ural
net
work
s
w
it
h b
ackpr
op
a
gatio
n
al
go
rithm
2.4.2
. M
od
el
p
erfo
rm
an
ce
as
sessment
The
re
s
pecti
ve
m
od
el
pe
rform
ance
is
evaluated
by
ex
plo
it
ing
the
m
ean
-
s
quare
e
rror
(MSE)
a
nd
coeffic
ie
nt
of
de
te
rm
inati
on
(
R2)
t
hat can
be
d
e
fine
d,
res
pe
ct
ively
as [
23
]
-
[
27
]
=
1
∑
(
̂
−
̅
)
2
=
1
(13)
2
=
∑
(
̂
−
̂
)
2
=
1
∑
(
−
̂
)
2
=
1
(14)
wh
e
re
̂
re
pr
ese
nts
t
he
predict
ed
outp
ut
us
in
g
t
he
ne
ur
al
ne
twork
,
̅
represe
nts
t
he
m
ean
of
ta
r
get
val
ues,
denotes t
he ob
t
ai
ned
ta
rg
et
va
lue from
d
at
a s
et
s,
an
d n rep
re
sents the
num
ber
of
sim
ulate
d
scena
rios.
3.
RESU
LT
S
AND DI
SCUS
S
ION
The
res
ults
ob
t
ai
ned
s
how
th
at
the
vo
lt
age
con
t
ro
l
is
eff
e
ct
ive
since
the
set
po
int
val
ue
s
are
ind
ee
d
i
m
po
sed
on
th
e
co
nn
ect
i
o
n
node
of
the
ene
r
gy
pro
ducers
a
s
sho
wn
i
n
Fi
gure
7
.
I
n
a
ddit
ion,
it
is
evi
dent
that
there
is
a
go
od
c
orres
ponde
nce
betwee
n
t
he
e
xcita
ti
on
a
nd
ge
ne
rator
volt
ages.
Als
o,
as
sho
wn
in
Figure
8
,
there is a
noti
c
eable co
rrel
at
ion bet
ween t
he
elec
tric
pow
e
r
s an
d
t
he
c
on
tr
ol volt
age.
T
h
e
M
A
T
L
A
B
n
e
u
r
a
l
n
e
t
w
o
r
k
t
o
o
l
k
i
t
i
s
e
m
p
l
oy
e
d
f
o
r
n
e
u
r
a
l
n
e
t
w
o
r
k
t
r
a
i
ni
n
g
a
n
d
l
e
a
r
n
i
n
g
.
T
h
e
t
r
a
i
ni
n
g
s
t
a
t
e
d
i
a
g
r
a
m
o
f
t
h
e
b
a
c
k
p
r
o
p
a
g
a
t
i
o
n
n
e
u
r
a
l
n
e
t
w
o
r
k
i
s
s
h
o
w
n
i
n
F
i
g
u
r
e
9
.
T
h
e
f
i
gu
r
e
i
l
l
u
s
t
r
a
t
e
s
t
h
a
t
t
he
m
o
d
e
l
r
e
a
c
h
e
s
t
h
e
c
o
n
v
e
r
g
e
n
c
e
p
r
e
c
i
s
i
o
n
a
t
s
t
e
p
1
0
1
.
M
o
r
e
o
v
e
r
,
t
h
e
t
r
a
i
n
e
d
n
e
u
r
a
l
n
e
t
w
o
r
k
i
s
t
h
e
n
e
m
p
l
oy
e
d
f
o
r
p
r
e
d
i
c
t
i
o
n
a
n
d
v
e
r
i
f
i
c
a
t
i
o
n
.
T
h
e
v
a
l
i
d
a
t
i
o
n
p
e
r
f
o
r
m
a
n
c
e
s
f
o
r
t
h
e
s
o
l
a
r
p
o
w
e
r
i
n
v
e
r
t
e
r
m
o
d
e
l
i
s
p
r
e
s
e
n
t
e
d
i
n
F
i
g
u
r
e
1
0
.
T
h
e
m
o
s
t
e
x
c
e
l
l
e
n
t
v
a
l
i
d
a
t
i
o
n
p
e
r
f
o
r
m
a
n
c
e
i
s
r
e
a
l
i
z
e
d
a
t
t
h
e
e
p
o
c
h
1
0
1
w
i
t
h
a
c
o
r
r
e
s
p
o
n
d
i
n
g
M
S
E
v
a
l
u
e
o
f
0
:
0
0
0
3
8
6
6
7
.
I
t
i
s
n
o
t
e
w
o
r
t
h
y
t
h
a
t
t
he
v
a
l
i
d
a
t
i
o
n
,
t
r
a
i
n
,
a
n
d
t
e
s
t
c
ur
v
e
s
a
r
e
r
e
l
a
t
i
ve
l
y
c
om
pa
r
a
b
l
e
.
A
l
s
o
,
i
t
s
h
o
u
l
d
b
e
n
o
t
e
d
t
h
a
t
t
h
e
p
l
o
t
d
o
e
s
n
o
t
i
n
c
r
e
a
s
e
a
f
t
e
r
c
o
n
v
e
r
g
e
n
c
e
,
v
e
r
i
f
y
i
n
g
t
h
e
s
t
a
b
i
l
i
t
y
o
f
t
h
e
m
o
d
e
l
.
T
h
i
s
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
In
te
ll
igent v
oltag
e
re
gu
l
ato
r f
or
distri
bute
d gen
e
ra
ti
on
-
bas
ed netw
or
k (
Z
wawi
Hama
douch
e)
105
i
m
pl
i
e
s
t
h
a
t
t
h
e
r
e
i
s
n
o
o
v
e
r
f
i
t
t
i
n
g
i
s
s
u
e
i
n
t
h
e
s
y
s
t
e
m
b
y
t
h
e
A
N
N
a
n
d
t
h
a
t
t
h
e
a
d
o
p
t
e
d
n
e
t
w
o
r
k
t
r
a
i
n
i
n
g
i
s
e
f
f
i
c
i
e
n
t
.
A
l
s
o
,
t
h
e
e
r
r
o
r
h
i
s
t
o
g
r
a
m
f
o
r
f
o
r
e
c
a
s
t
i
n
g
t
h
e
v
o
l
t
a
g
e
r
e
g
u
l
a
t
i
o
n
w
i
t
h
t
h
e
d
e
v
e
l
o
p
e
d
A
N
N
i
s
r
e
p
r
e
s
e
n
t
e
d
i
n
F
i
g
u
r
e
1
1
.
T
h
i
s
p
l
o
t
e
m
p
h
a
s
i
z
e
s
t
h
a
t
m
o
s
t
o
f
t
h
e
e
r
r
o
r
s
f
a
l
l
b
e
t
w
e
e
n
−
1
:
0
4
4
5
4
a
n
d
0
:
1
1
9
3
,
w
h
i
c
h
i
s
a
n
a
r
r
o
w
i
n
t
e
r
v
a
l
.
A
l
s
o
,
t
h
e
m
a
j
o
r
i
t
y
o
f
t
h
e
e
r
r
o
r
s
h
a
v
e
a
v
a
l
u
e
o
f
a
b
o
u
t
0
:
0
3
7
3
6
.
C
o
n
s
e
q
u
e
n
t
l
y
,
t
h
e
e
r
r
o
r
h
i
s
t
og
r
a
m
i
l
l
u
s
t
r
a
t
e
s
t
h
a
t
t
h
e
f
o
r
e
c
a
s
t
i
n
g
a
c
c
u
r
a
c
y
o
f
o
u
r
p
r
o
p
o
s
e
d
m
o
d
e
l
i
s
v
e
r
y
g
o
o
d
.
Figure
7.
V
oltage
prof
il
e
with
DG im
ple
m
entat
ion
a
nd contr
ol m
od
el
(
F
FNN)
Figure
8.
P
ow
e
r profile
with
DG im
ple
m
entat
ion
a
nd contr
ol m
od
el
(
F
FNN)
The
trai
ni
ng
m
od
el
of
the
tw
o
-
ne
uro
n
bac
kpr
op
a
gatio
n
ne
ural
network
f
or
synch
ron
ous
ge
ner
at
or
is
sh
ow
n
in Figur
e 1
2
. T
he
fi
gur
e d
epict
s that the m
od
el
r
each
es the co
nve
rgence preci
sio
n
at
step 9
65. Be
sides,
the
validat
io
n
perform
ances
f
or
t
he
sy
nchronous
ge
ner
at
or
m
od
el
is
sh
own
in
Fi
gure
1
3
.
T
he
best
val
idati
on
perform
ance
is
at
ta
ined
at
an
MSE
value
o
f
1:0756
×
10
−
5
and
e
poch
96
5.
Also
,
base
d
on
t
he
sim
i
la
rity
of
the
set
s,
there
is
no
over
fitt
ing
issue
in
the
s
yst
e
m
.
Si
m
il
ar
ly
,
the
err
or
histogram
fo
r
forecast
ing
the
volt
age
regulat
ion
base
d
on
the
de
velop
e
d
A
N
N
is
rep
rese
nted
in
Figure
1
4
.
T
his
plo
t
i
ll
us
trat
es
that
the
m
ajo
rity
of
the
er
rors
fall
i
n
a
relat
ively
c
onfine
d
ra
ng
e
of
−1:
00173
a
nd
0:
004775,
a
nd
nea
rly
al
l
th
e
erro
rs
ha
ve
a
valu
e
of
ab
out
0:00
1521.
F
ur
the
r
m
or
e,
the
reg
r
essions
e
xp
e
ri
enced
within
the
outp
uts
an
d
netw
ork
ta
r
gets
are
com
pu
te
d
an
d
represente
d.
The
ove
rall
predict
ion
perfor
m
ances
for
th
e
synch
ron
ou
s
gen
e
rato
r
m
od
el
is
dep
ic
te
d
i
n
Figure
1
5
by
m
eans
of
plo
ts
of
ou
t
pu
t
an
d
ta
rg
et
values
.
The
fitt
ing
pe
rfor
m
ances
f
or
the
consi
der
e
d
m
od
el
are
good,
a
s
the
data
poin
ts
are
ti
gh
tl
y
dist
ribu
te
d
dow
n
the
fitt
ing
li
ne
.
Also,
the
ob
ta
ined
correla
ti
on
coe
ff
ic
ie
nts
a
re
ve
ry
cl
os
e
to
1
(
0:
9998).
T
his
in
dicat
es
that
t
he
co
ns
ide
re
d
sc
hem
e
of
fe
rs
a
cl
os
e
-
fitt
ing
betwee
n
the
netw
ork
ta
rg
et
s
a
nd
rela
te
d
ou
t
puts.
Be
sides,
the
virt
ually
li
near
distrib
utio
n
of
th
e
data
po
i
nts alo
ng
w
it
h
the
ou
t
put v
ersu
s
the
targ
et
li
ne
s
hows
t
ha
t t
he
m
od
el
’s p
red
ic
ti
on acc
uracy
is good.
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
:
98
-
109
106
Figure
9. Trai
ni
ng
sta
te
d
ia
gra
m
o
f
the
backp
ropa
gation ne
ural
n
et
work f
or so
la
r
po
wer i
nverter
Figure
10. FF
NN m
od
el
v
al
idati
on p
e
rfo
rm
ance
for
so
la
r p
ow
e
r
i
nverter
Figure
11. E
rro
r hist
ogram
f
or so
la
r
po
wer i
nverter
Figure
1
2.
T
rai
ning stat
e
diag
ram
o
f
the t
wo
-
ne
uro
n back
pr
op
a
gatio
n ne
ural
n
et
w
ork for
synch
ron
ous
ge
ner
at
or
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
In
te
ll
igent v
oltag
e
re
gu
l
ato
r f
or
distri
bute
d gen
e
ra
ti
on
-
bas
ed netw
or
k (
Z
wawi
Hama
douch
e)
107
Figure
1
3.
FF
NN m
od
el
v
al
idati
on p
e
rfo
rm
ance
for
synch
ron
ous
ge
ner
at
or
Figure
1
4.
E
rro
r hist
ogram
f
or sync
hrono
us
gen
e
rato
r
Figure
1
5.
Re
gressi
on m
od
el
perform
ance f
or sync
hro
nous
gen
e
rato
r.
The
top
le
ft b
l
ue plot de
picts t
he
trai
ning f
it
ti
ng,
the top
rig
ht
gre
en
p
l
ot il
lustr
at
es the
validat
ion
fitt
ing
,
the
bo
tt
om
left re
d plot
sho
ws
the
te
sti
ng
f
it
ti
ng, a
nd the
bott
om rig
ht
gr
ay
plo
t
d
e
picts t
he
m
od
el
f
it
ti
ng p
e
rfor
m
ance for al
l data set
s
4.
CONCL
US
I
O
N
In
this
pa
per,
we
have
propo
sed
a
ge
ne
ric
m
od
el
for
a
D
N
wit
h
i
nterc
onnecte
d
s
olar
pa
nel
-
ty
pe
a
nd
wind
tur
bin
e
powe
r
ge
ner
at
ors.
Also,
we
ha
ve
co
ns
ide
red
and
dem
on
stra
te
d
the
var
ie
d
natu
re
of
the
volt
age
unde
r
distrib
uted
gen
e
rati
on.
In
this
c
on
te
xt,
we
ha
ve
dev
e
lop
e
d
c
on
t
ro
l
m
od
el
s
for
vo
l
ta
ge
re
gula
ti
on.
T
he
si
m
u
la
ti
on
res
ults
sho
w
that
de
velo
ped
co
ntr
ol
m
od
el
s
can
help
i
n
disturbance
s
el
im
inati
on
in
t
he
D
N
.
More
ov
e
r,
we
hav
e
de
velo
ped
a
nd
e
valuated
FF
N
N
m
od
el
s
with
t
he
bac
kpr
op
a
gation
al
go
rithm
an
d
Evaluation Warning : The document was created with Spire.PDF for Python.