Int
ern
at
i
onal
Journ
al of Ele
ctrical
an
d
Co
mput
er
En
gin
eeri
ng
(IJ
E
C
E)
Vo
l.
10
,
No.
4
,
A
ugus
t
2020
,
pp. 337
5~33
83
IS
S
N: 20
88
-
8708
,
DOI: 10
.11
591/
ijece
.
v10
i
4
.
pp3375
-
33
83
3375
Journ
al h
om
e
page
:
http:
//
ij
ece.i
aesc
or
e.c
om/i
nd
ex
.ph
p/IJ
ECE
Loss all
ocation in
distribu
tion netwo
rks with
distri
bu
ted
ge
n
erato
rs
un
de
rgoin
g network
reco
nfi
guration
Ambik
a
Pr
asa
d H
ota
1
, Si
vkumar
M
ishr
a
2
1
Inte
rna
ti
ona
l
In
stit
ute of
In
form
at
ion
Technol
og
y
Bhub
ane
sw
ar
,
India
2
Cent
re
for
Adv
anc
ed
Pos
t
Grad
uat
e
Studie
s
,
Bi
j
u
Patnaik
Univ
er
sit
y
of Te
chnol
o
g
y
,
Indi
a
Art
ic
le
In
f
o
ABSTR
A
CT
Art
ic
le
history:
Re
cei
ved
A
pr
3,
2019
Re
vised Jan
13
, 2020
Accepte
d
Ja
n 3
0,
2020
In
thi
s
pap
er,
a
bra
nch
exc
h
an
ge
base
d
heur
is
ti
c
n
et
work
r
econfigura
t
ion
m
et
hod
is
proposed
for
obta
i
n
in
g
an
opti
m
al
n
etw
ork
in
a
der
eg
ula
t
ed
power
s
y
stem.
A
uniq
ue
bus
ide
nti
f
i
ca
t
ion
sche
m
e
is
emplo
y
ed
w
hic
h
m
ake
s
the
loa
d
f
low
a
nd
loss
calc
ul
ation
fast
er
du
e
t
o
it
s
r
educed
s
ea
rch
ti
m
e
under
var
y
ing
net
work
topol
o
gic
a
l
envi
ronm
ent
.
Th
e
propo
sed
power
loss
al
loc
a
ti
on
te
chn
ique
e
li
m
i
nat
es
the
eff
e
ct
of
cro
ss
-
te
rm
ana
l
y
t
ical
l
y
from
the
loss
f
orm
ula
ti
on
with
out
an
y
assum
pti
ons
and
appr
oximati
ons.
The
eff
ective
n
e
ss
of
the
proposed
rec
onfigu
r
at
ion
and
loss
al
lo
ca
t
ion
m
et
hods
are
inv
esti
gated
b
y
co
m
par
ing
t
he
r
esult
s
obtained
b
y
the
pr
ese
n
t
appr
oac
h
with
t
hat
of
the
ex
isti
ng
“
Quadra
tic
m
et
hod”
using
a
3
3
-
bus
rad
ial
distri
buti
on
s
y
st
em wit
h/wit
hou
t
DG
s.
Ke
yw
or
d
s
:
Distrib
uted ge
ner
at
or
Loss
al
locat
ion
Netw
ork reco
nfi
gurati
on
Ra
dial dist
rib
ut
ion
netw
ork
Copyright
©
202
0
Instit
ut
e
o
f Ad
vanc
ed
Engi
n
ee
r
ing
and
S
cienc
e
.
Al
l
rights re
serv
ed
.
Corres
pond
in
g
Aut
h
or
:
Am
bik
a Pr
asa
d H
ota,
In
te
r
natio
nal Inst
it
ute of
Inform
ation
Tec
hn
ology
Bh
ubane
swar,
Bhuba
nesw
a
r,
Od
is
ha, I
nd
ia
.
Em
a
il
:
c11
7001@ii
it
-
bh.ac
.in
1.
INTROD
U
CTION
In
a
de
regulat
ed
en
vir
onm
e
nt
an
d
with
the
em
erg
ence
of
sm
art
gr
id
s,
it
is
the
co
ns
um
ers
a
nd
distrib
uted
ge
ne
rators
(DGs)
of
the
distri
bu
t
ion
syst
em
(D
S)
,
w
ho
will
ul
tim
at
el
y
pay
or
get
pai
d
for
thei
r
ind
ivi
du
al
sh
a
r
e
of
net
wor
k
util
iz
at
ion
.
Acti
ve
powe
r
lo
ss
in
el
ect
rici
ty
distrib
ution
is
a p
r
om
inent
com
po
ne
nt
of
this
netw
ork
util
iz
at
ion
cost.
He
nce,
it
is
to
be
rec
ov
e
red
f
r
om
the
ne
twork
us
ers
by
su
it
ably
and
fairly
al
locat
ing
it
a
m
on
g
them
.
T
he
non
li
nea
r
r
el
at
ion
sh
i
p
between
po
wer
loss
a
nd
i
nj
ect
ed
powe
rs
m
akes
l
os
s
al
locat
ion
(
L
A
)
process
dif
fic
ult
an
d
c
om
pli
cat
ed
[
1]. S
im
ultaneo
us
ly
,
i
n
el
ect
ric
po
wer
distri
bu
ti
on n
e
tworks
(EPDNs
),
the
r
e
is
a
gr
owin
g
tren
d
in
th
e
pen
et
rati
on
of
D
Gs
beca
use
op
ti
m
al
pl
ace
m
ent
of
D
Gs
in
the
EP
DNs
of
t
en
le
ad
to
s
ub
sta
ntial
reducti
on
in
act
ive
powe
r
lo
ss
[2
]
.
Ther
e
f
or
e,
it
is
essenti
al
to
al
locat
e
this
decr
ease
in
powe
r
loss
a
m
on
g
the
network
par
ti
ci
pa
nts
j
udic
io
us
ly
as
per
their
act
ual
con
tri
buti
on
towa
rd
s
syst
em
loss
re
duct
ion
(
SLR).
F
ur
t
her,
L
A
to
the
end
use
rs
get
a
ff
ect
ed
du
e
to
the
ch
an
ging
ne
twor
k
topolo
gy
i.e.
ne
twork
rec
onfi
gurati
on
(
NR)
,
w
hich
is
al
s
o
an
est
a
blished
m
et
ho
d
of
loss
reducti
on
i
n
E
PDNs.
Th
us
,
base
d
on
one
’s
c
hangi
ng
relat
ive
po
s
it
ion
in
the
net
wo
rk
an
d
the
ty
pe
of
loa
d,
a
consum
er/DGO
m
ay
ei
ther
ha
ve
to
get r
ewa
r
ded
or p
e
naliz
ed.
He
nce,
the
re is a scope to
an
al
yz
e the i
m
pact o
f
NR on
act
ive p
owe
r
loss
al
locat
ion
in
m
od
er
n
dist
rib
ution
netw
orks
,
pa
rtic
ularl
y,
wh
e
n
ve
ry
li
tt
le
wo
r
k
has
been
repor
te
d
wh
e
re
bo
t
h NR a
nd L
A
a
re c
on
si
dered to
gethe
r.
Lit
eratur
e
surv
ey
rev
eal
s
tha
t
m
os
t
of
the
LA
m
et
ho
ds
wer
e
init
ia
ll
y
pro
po
se
d
f
or
transm
issi
on
syst
e
m
s
[3
]
,
and
la
te
r
,
LA
m
et
ho
ds
e
xclu
sively
m
eant
fo
r
distrib
utio
n
syst
e
m
s
are
rep
ort
ed
i
n
Re
f
s
[4
-
11]
.
W
it
h
th
e
pen
et
rati
on
of
DGs,
the
LA
m
et
ho
ds
are
m
od
ifie
d
an
d
pr
opos
e
d
acco
rd
i
ng
ly
.
A
powe
r
s
umm
at
ion
base
d
LA
al
go
rithm
is
dev
el
op
e
d
in
[5
]
w
her
e
the
c
r
os
s
-
te
rm
of
powe
r
loss
eq
uation
is
bifurcate
d
a
m
on
g
the
respo
ns
ible
nodes
us
i
ng
a
qu
a
dr
at
ic
sche
m
e.
This
te
chn
iqu
e
is
furthe
r
exten
ded
for
energy
cal
culat
ion
in
Re
f.
[
6]
thr
ough
a
sta
ti
sti
cal
a
naly
sis
of
daily
load
an
d
generati
on
c
urves
of
DGs.
I
n
[7
]
,
a
three
ste
ppe
d
L
A
Evaluation Warning : The document was created with Spire.PDF for Python.
IS
S
N
:
2088
-
8708
In
t J
Elec
&
C
om
p
En
g,
V
ol.
10
, No
.
4
,
A
ugus
t
2020
:
3375
–
3383
3376
proce
dure
is
di
scusse
d
w
here
in
the
fir
st
ste
p;
losses
are
assigne
d
to
l
oads
the
n
in
t
he
ne
xt
ste
p
t
o
D
Gs
.
In
t
he
thi
rd
ste
p
norm
al
iz
a
tio
n
te
ch
nique
i
s
us
e
d
for
fin
al
set
tl
e
m
ent
of
t
he
lo
sses.
Howe
ver,
the
m
et
hod
discusse
d
i
n
[
8]
do
es
no
t
require
a
ny
rec
onci
li
at
ion
proce
dure
f
or
al
loca
ti
on
of
lo
sses
wh
e
re
the
c
r
oss
-
te
rm
s
are
s
har
e
d
us
i
ng
a
ge
om
et
r
ic
schem
e
of
L
A.
Kas
hyap
an
d
De
[9
]
ha
ve
propose
d
a
pr
opor
ti
onal
s
har
i
ng
base
d
LA
sc
hem
e
for
distri
bu
ti
on of
losses
am
on
g t
he
c
on
s
um
ers
and
D
G
unit
s.
In
this
a
ppr
oac
h,
l
oa
ds
a
re
as
s
ign
e
d
losses
acco
rd
i
ng
to
their
c
ontrib
utio
n
an
d
then
DGs
ar
e
al
locat
ed
by
con
side
rin
g
a
con
trib
utio
n
m
at
rix
fo
ll
owe
d
by
powe
r
s
ha
rin
g
m
at
rix.
T
he
m
et
hod
de
velo
pe
d
i
n
[10]
us
e
s
a
par
ti
ci
pati
on
m
at
rix
to
i
den
ti
fy
the
par
ti
ci
patio
n
of
nodes
in
the
loss
of
each
br
anc
h.
An
a
na
ly
ti
cal
coo
pe
r
at
ive
gam
e
theor
y
base
d
te
ch
nique
is
fo
ll
ow
e
d
in
the
de
velo
pe
d
m
et
ho
d
[
12
]
w
it
h
Sh
a
pley
val
ue
to
ove
rco
m
e
com
pu
ta
ti
onal
bur
den
al
on
g
with
fair
al
locat
io
n.
K
um
ar
et
al
.
[13]
de
velo
ped
a
br
a
nc
h
-
or
ie
nted
ci
rc
uit
theo
r
y
base
d
LA
te
ch
nique
w
hich
use
s
a
con
t
rib
ution ba
sed
l
os
s all
ocat
ion
facto
r for
bi
fu
r
cat
io
n o
f
th
e cr
os
s
-
te
rm
s p
resen
t i
n
the
power l
os
s
eq
uation.
More
ov
e
r,
Mi
s
hr
a
et
a
l.
[
14
]
pr
ese
nted
a
c
om
pr
ehen
sive
r
eview
of
var
i
ous
NR
m
et
ho
ds
up
t
o
20
16.
Heurist
ic
-
base
d
NR
m
e
tho
ds
converge
ve
ry
qu
ic
kly
as
com
par
ed
to
the
m
e
ta
-
heuri
s
ti
c
search
bas
ed
NR
m
et
ho
ds
[
15
]
.
Howe
ver,
he
uri
sti
c
m
et
ho
ds
do
not
al
ways
gu
ara
ntee
a
glob
al
m
ini
m
u
m
co
nf
i
gurati
on.
I
n
[16
]
,
a
br
anc
h
exc
ha
ng
e
based
he
ur
ist
ic
NR
al
go
rithm
is
pr
opos
e
d
for
the
f
ulfillm
ent
of
reli
abili
ty
and
powe
r
qu
al
i
ty
base
d
obj
ect
ives
.
G
ha
sem
i
[1
7
]
s
ugge
s
te
d
a
he
uri
sti
c
NR
al
go
rithm
fo
r
reli
a
bili
ty
based
in
dices.
Gam
apa
and
Das
[
18
]
us
e
d
a
he
ur
ist
ic
m
et
hod
f
or
a
f
uz
zy
m
ulti
-
ob
je
ct
ive
NR
with
DGs.
Das
et
al
.
[19]
pro
po
se
d
a
he
ur
ist
ic
,
ci
rcu
la
r
m
echan
ism
fo
r
N
R
in
the
presence
of
D
Gs
.
In
[
20
]
,
a
two
-
sta
ge
he
ur
ist
ic
NR
m
et
ho
d
us
in
g
op
ti
m
al
load
f
low
is
intr
oduc
ed.
Ty
agi
et
al
.
[21
]
al
so
pro
posed
a
tw
o
-
sta
ge
heurist
ic
NR
al
gorithm
fo
r
l
oad
a
bili
ty
enhancem
ent.
Jasthi
an
d
Das
[
22
]
propose
d
a
loss
form
ula
ba
sed
he
ur
ist
ic
m
et
hod
for
NR.
A
s
NR
is
pr
oven
to
be
the
m
o
st
eff
ect
ive
m
eans
of
m
ini
m
iz
ing
act
ive
powe
r
loss
for
EPDNs
,
so
,
it
is
essenti
al
to
st
ud
y
t
he
im
pact
of
NR
on
th
e
loss
al
locat
i
on
m
et
ho
ds
in
a
distri
bu
ti
on
syst
e
m
,
wh
e
r
e
it
is
assum
ed
that
the
co
nsum
ers
hav
e
t
o
pay
f
or
the
losse
s
or
so
m
et
i
m
es
g
et
rew
a
rd
e
d
.
Oliv
ei
ra
et
al
.
[23
]
wer
e
the
fir
st
to
c
on
sider
NR
a
nd
LA
t
og
et
her
f
or
distrib
ution
s
yst
e
m
s
with
D
Gs,
w
her
e
the
NR
is
e
xecu
te
d
us
in
g
a
he
ur
ist
ic
m
et
hod
a
nd
the
loss
al
l
ocati
on
us
in
g
Z
-
bu
s
m
e
tho
d
[
24]
,
but
the
fin
di
ng
s
wer
e
not
ve
ry
convinci
ng.
S
a
vier
a
n
d
Das
[
25
]
st
ud
ie
d
t
he
i
m
pact
of
N
R
on
L
A
i
n
a
m
uch
bette
r
w
ay
,
w
her
e
a
he
ur
ist
ic
br
a
nc
h
ex
cha
nge
based
al
gori
thm
is
e
m
plo
yed
f
or
opti
m
al
reconfi
gurati
on,
a
nd
the
n
quadr
at
ic
los
s
al
locat
io
n
schem
e
is
us
e
d
to
al
locat
e
l
os
ses
to
va
rio
us
c
onsu
m
ers
in
a
rad
i
al
dist
rib
ution
net
work
(R
DN)
befo
re
a
nd
after
the
rec
onfi
gurati
on.
R
ecentl
y,
a
m
etah
eu
ri
sti
c
base
d
fi
ref
ly
al
gor
it
h
m
[2
6
]
,
a
m
od
ifie
d
FBS
base
d
te
chn
iq
ue
[27
]
and
a
G
rey
Wo
l
f
opti
m
iz
a
ti
on
m
et
ho
d
[
28
]
are
al
s
o
intr
oduce
d
f
or
ob
ta
ini
ng
a
n
e
ff
ic
ie
nt
reconfi
g
ure
d p
ow
e
r dist
rib
ution net
wor
k wit
h
im
pr
oved
vol
ta
ge
pr
of
il
e.
In
t
he
li
gh
t
of
the
ab
ov
e
de
velo
pm
ents,
this
pap
e
r
car
ries
the
analy
sis
furthe
r.
F
or
fa
ir
al
locat
ion,
a
new
loss
al
locat
ion
te
ch
ni
qu
e
is
em
plo
yed
in
sect
ion
-
2.
Sim
ultaneou
sly
,
syst
e
m
l
os
s
is
r
ed
uce
d
us
in
g
a
he
ur
ist
ic
bra
nch
exc
ha
ng
e
(BE)
base
d
N
R
ap
proach,
w
hich
is
desc
rib
ed
in
sect
io
n
-
3.
As
the
LA
s
chem
e
us
es
the
c
onve
rg
e
d
loa
d
fl
ow
resu
lt
s,
a
nd
in
NR,
s
uccessiv
e
load
flo
ws
ar
e
perform
ed
for
cha
ngin
g
net
w
or
k
topolo
gies
wit
h
suc
cessi
ve
br
a
nc
h
exc
ha
nges,
a
for
ward
-
bac
kw
a
rd
s
weep
base
d
lo
ad
fl
ow
te
c
hniqu
e
as
descr
i
bed
in
[
2]
is
i
m
ple
m
e
nted
with
a
uniq
ue
bus
i
de
ntific
at
ion
sc
hem
e.
In
sect
io
n
-
4,
t
he
ef
fici
ency
of
the
pro
po
se
d
m
et
ho
d
is
eval
uated
by
com
par
in
g
the
LA
r
esults
with
the
existi
ng
Q
ua
dra
ti
c
m
et
ho
d
at
bo
th
befor
e
and a
fte
r
NR
consi
der
i
ng a
33
-
bu
s
R
DN w
it
h/
without D
Gs. Co
ncl
us
ive
r
em
ark
s
are m
ade in
se
ct
ion
-
5.
2.
NEW
METH
OD OF LO
SS
A
LL
O
CA
TI
ON
The
pro
pose
d
LA
schem
e
is
exp
la
ine
d
by
c
on
si
der
i
ng
a
sa
m
ple
10
-
bus
RDN
as
s
how
n
in
Fig
ur
e
1.
Fo
r
the
sa
ke
of
sim
plicity
in
the
a
naly
sis,
th
e
substat
io
n
bus
is
in
de
xed
as
‘
1’
an
d
t
he
s
ubse
qu
e
nt
bu
s
es
al
ong
the
m
ai
n
feed
e
r
an
d
la
te
ral
fe
eder
s
a
re
in
de
xed
i
n
the
i
ncrea
sing
orde
r.
The
br
a
nc
h
nu
m
ber
is
on
e
unit
le
ss
than
that
of
it
s
receivi
ng
e
nd
node
nu
m
ber
.
I
n
this
ty
pe
of
RD
N,
the
tot
al
num
ber
of
branc
hes
(
nbr)
i
s
one
un
it
less t
han t
hat of t
he
total
nu
m
ber
of no
de
s (nb) i
n
the
net
work.
2.1.
Prop
os
ed
bu
s
identifica
tio
n
scheme
To
identify
th
e
adj
ace
nt
bu
s
es
and
the
sub
seq
uen
t
buses
in
a
RDN,
sev
eral
arr
ay
s
are
pr
op
os
e
d.
These
a
rr
ay
s
m
ake
the
loa
d
flo
w
faste
r
due
to
it
s
re
du
ce
d
sea
rch
ti
m
e.
The
a
dj
ac
ent
buses
in
the
R
DN
a
re
detect
ed
an
d
s
aved
in
a
n
arra
y
adb
[]
of
dim
ension
twic
e
the
nb
r
.
Tw
o
po
i
nter
ar
rays
of
dim
ension
e
qu
al
to
the
num
ber
of
bu
s
es
‘
nb’
ar
e
pr
ese
nted
,
w
he
re
the
ar
rays
m
f
[i]
and
m
t[i
]
are
us
e
d
to
point
the
sta
rti
ng
an
d
end
i
ng
m
e
m
or
y
locat
ion
s
i
n
t
he
a
db[]
arr
ay
,
r
especti
vely
. Sim
i
la
rly
,
an
ar
r
ay
pb[]
is
us
e
d
to
ide
ntify
an
d
store
al
l
the
previ
ous
buses
of
the
RDN.
T
he
c
onstructio
ns
of
these
a
rr
ay
s
a
re
perf
or
m
ed
us
i
ng
net
work
da
ta
a
nd
si
m
ple p
r
ogra
m
m
ing
techn
i
ques.
The
adjac
ent and
previ
ous
bus
data of t
he 10
-
bus
RD
N
as
sho
wn
i
n
Figure
1
are
sto
red
i
n
th
e
arr
ay
s
m
f,
m
t
,
ad
b
an
d
pb
as
sh
ow
n
in
Fi
gure
2
to
desc
rib
e
the
s
tora
ge
a
nd
pointer
ope
rati
on
of
these
a
rr
ay
s
.
Evaluation Warning : The document was created with Spire.PDF for Python.
In
t J
Elec
&
C
om
p
En
g
IS
S
N: 20
88
-
8708
Lo
s
s
alloc
atio
n i
n dist
ribu
ti
on
n
et
work
s wi
th
distri
bu
te
d gene
ra
tors
un
der
goin
g…
(
A
mb
ik
a
Pr
asad H
ot
a
)
3377
Si
m
il
arly
,
tw
o
ot
her
ar
rays
nsb[]
a
nd
sb
[]
ar
e
pro
pose
d
f
or
pro
vid
in
g
i
nfor
m
at
ion
a
bout
the
s
ub
se
quent
buses.
All
th
e
subse
qu
e
nt
bu
s
es
c
orres
pond
i
ng
to
eac
h
branc
h
of
the
RDN
are
sto
r
ed
i
n
the
arr
ay
s
b[
]
, an
d
t
heir
re
sp
e
ct
ive
nu
m
ber
s (
i.e.
the n
um
ber
of
subse
quent
bu
se
s)
a
re
sto
r
ed
in
the
a
rr
ay
n
sb[]
(d
im
ension
e
qual
to
nbr
).
Th
e
sta
rting
an
d
end
i
ng
m
e
m
or
y
locat
ion
s
of
t
he
s
b[
]
a
rr
ay
a
re
po
i
nted
out
by
tw
o
po
i
nter
a
rr
ay
s
m
fs
and
m
ts,
r
especti
vely
.
T
he
s
ubse
qu
e
nt
bu
s
data
of
the
10
-
bus
R
DN
as
sho
wn
in
Fi
gure
1
al
ong
with
the
represe
ntati
on
of
br
a
nc
h
-
(2)
are
sp
eci
fied
in
the
ar
rays
s
b,
nsb,
m
ts
and
m
fs
as
sh
own
i
n
Figure
3.
Thes
e
arr
ay
s
a
re
form
ed
us
in
g
the
netw
ork
data
and
sim
ple
pr
ogram
m
ing
te
ch
niques
in
MA
TLA
B
(R2018
b)
e
nv
i
ronm
ent.
The
util
iz
at
ion
of
t
he
p
rop
os
ed
a
rr
ay
s
in
both
load
fl
ow
(LF)
analy
sis
and
loss
al
locat
ion
redu
ce the c
om
pu
ta
ti
on
al
tim
e and h
e
nce, wor
ks
m
or
e eff
ect
ivel
y for lar
ger
R
D
Ns.
In each
sta
ge of
NR,
the
ar
rays
are
rec
on
st
ru
ct
ed
base
d
on
th
e
reconfi
gured
RDN
a
nd
syst
e
m
loss
cal
culation
is
pe
rfor
m
ed
f
or
ob
ta
ini
ng opti
m
al
r
econfig
urat
ion
.
Figure
1. A
Sa
m
ple 1
0
-
Bu
s R
DN
Figure
2. Sto
ra
ge
a
nd pointe
r op
e
rati
ons
of
m
f,
m
t, p
b
a
nd
adb ar
rays
Figure
3
.
Sto
ra
ge
a
nd pointe
r op
e
rati
o
ns
of
m
fs,
m
ts, sb
and
ns
b ar
rays
S
/
S
2
1
3
4
5
7
8
10
(
2
)
(
5
)
(
7
)
9
6
(
1
)
(
3
)
(
4
)
(
6
)
(
8
)
(
9
)
t
l
=[
4
,
9
]
ss
=[
3
,
4
]
mf
(
i
)
mf
(
3
)=
5
mt
(
3
)=
7
a
d
b
(
5
)
a
d
b
(
7
)
(
i
)
1
2
3
4
5
6
7
1
2
5
8
10
11
13
8
9
10
(
i
)
1
2
3
4
5
6
7
1
4
7
9
10
12
13
8
9
10
mt
(
i
)
14
17
18
16
17
18
Bu
s
(
i
)
1
2
3
4
5
6
7
8
9
10
Bu
s
(
i
)
1
2
3
4
5
6
7
Pb
(
i
)
0
1
2
3
4
8
9
10
3
6
2
8
8
s
1
2
3
4
5
6
7
a
d
b
(
s
)
2
1
3
8
2
4
6
8
9
10
11
3
5
4
3
7
12
13
14
15
16
17
18
6
2
9
10
8
8
m
f
s
(
2
)=
10
m
t
s
(
2
)=
14
sb
(
10
)=
3
sb
(
14
)=
7
n
s
b
(
2
)=
5
i
1
2
3
4
5
6
sb
(
i
)
2
3
4
5
6
7
7
9
10
11
12
13
14
10
3
4
5
6
7
8
8
9
15
16
17
4
5
5
18
20
21
22
23
24
25
7
8
9
10
9
10
19
6
7
n
s
b
(
b
)
j
1
2
3
4
5
6
9
5
2
1
2
1
7
8
9
3
1
1
m
f
s
(
b
)
b
1
2
3
4
5
6
1
10
15
17
18
20
7
8
9
21
24
25
m
t
s
(
b
)
b
1
2
3
4
5
6
9
14
16
17
19
20
7
8
9
23
24
25
Evaluation Warning : The document was created with Spire.PDF for Python.
IS
S
N
:
2088
-
8708
In
t J
Elec
&
C
om
p
En
g,
V
ol.
10
, No
.
4
,
A
ugus
t
2020
:
3375
–
3383
3378
2.2.
Form
ula
tion
of t
he
LA m
et
ho
d
In
t
his
pa
per,
a
forw
a
r
d
bac
kwar
d
s
weep
(FB
S)
ba
sed
l
oa
d
fl
ow
m
et
ho
d
as
discu
ssed
i
n
Re
f.
[
2]
is
carried
out
for
obta
inin
g
the
conve
rg
e
d
val
ues
of
bus
volt
ages
as
these
a
re
the
prere
quisi
te
s
for
power
loss
al
locat
ion
.
F
ur
t
her,
the
inj
ect
i
on
of
DG
pow
er
aff
ect
s
powe
r
loss
of
a
syst
e
m
.
Hen
ce,
in
order
to
in
cl
ud
e
DG
s
into
the
com
puta
ti
on
al
proces
s
the
neg
at
ive
l
oad
m
od
el
ing
of
D
Gs
[
8]
is
fo
ll
ow
e
d.
T
hus,
con
si
der
in
g
D
Gs
as
neg
at
ive
loa
ds
,
the
eq
uiv
al
e
nt
curre
nt
injec
ti
on
(EC
I)
at
a
ny
arb
it
rar
y
bus
-
i
of
the
R
DN
with
net
node
powe
r
inj
ect
io
n
“
SLi
= PLi +
jQLi”
and bus
volt
ag
e “Vi” ca
n be c
om
pu
te
d
as:
(1)
The
c
urren
t
of
br
a
nc
h
-
b
can
be cal
culat
ed
by
ad
di
ng
t
he
EC
Is
of the s
ubse
qu
e
nt buses wi
th the hel
p of
sb
,
mts
and
mfs
ar
rays
as:
(2)
Her
e
,
I
Lsb(
i)
rep
r
esents the
ECI
of the s
ubse
quent cons
um
er of br
a
nch
-
b
wh
i
ch
is c
onnecte
d
at
node
-
i
The bra
nc
h
c
urren
t ca
n be e
xp
resse
d
in
term
s
of c
om
plex
po
wer o
f
the
s
ub
s
equ
e
nt loa
d p
oi
nt
s as:
(3)
The
re
al
or
act
ive
po
wer
l
os
s
of
t
he
bra
nch
-
b
can
be
eval
uated
in
te
rm
s
of
sen
ding
en
d
volt
age
(
V
s
)
,
rece
ivi
ng
end volt
age
(
V
r
)
a
nd the
br
a
nc
h
c
urren
t
I(b)
as:
(4)
Wh
e
re,
‘
*’ r
e
presents
the c
on
j
ugat
e
value
of
a co
m
plex n
um
ber
.
Substi
tu
ti
ng (3
)
in
(4)
(5)
Re
arr
a
ng
i
ng
,
(6)
(7)
The
e
xpressi
on of Real
po
wer l
os
s
for
t
he br
anch
-
b
ca
n be
wr
it
te
n
as:
(8)
It
is
ob
ser
ve
d
fr
om
(8
)
tha
t,
the
custom
e
rs
beyo
nd
bra
nch
-
b
of
th
e
RDN
are
respon
si
ble
f
or
the
act
ive
or
tr
ue
powe
r
lo
ss
of
the
branc
h
-
b
.
He
nce,
this
loss
ca
n
be
al
l
ocated
to
t
he
c
on
s
um
ers
th
ose
ar
e
connecte
d
to
t
he
s
ub
se
quent
buses
of
th
e
br
a
nc
h
-
b
(i.e
.
al
l
sb
(
i
)
,
wh
er
e
i
=
mfs(
b)
to
mts(
b)
).
T
he
refor
e
,
the contri
bu
ti
on
of no
des
to
branc
h p
ow
e
r
lo
ss can be
co
m
pu
te
d
a
s:
,
2
,
3
,
.
.
.
.
.
.
.
,
L
i
L
i
L
i
L
i
Li
i
i
P
j
Q
P
j
Q
I
f
o
r
i
n
b
V
V
m
t
s
b
L
s
b
i
i
m
f
s
b
I
b
I
m
t
s
b
L
s
b
i
L
s
b
i
s
b
i
i
m
f
s
b
P
j
Q
Ib
V
Re
sr
P
L
o
s
s
b
a
l
V
V
I
b
Re
m
t
s
b
L
s
b
i
L
s
b
i
sr
s
b
i
i
m
f
s
b
P
j
Q
P
L
o
s
s
b
a
l
V
V
V
Re
m
t
s
b
sr
L
s
b
i
L
s
b
i
s
b
i
i
m
f
s
b
VV
P
L
o
s
s
b
a
l
P
j
Q
V
,
sr
s
b
i
s
b
i
s
b
i
VV
L
e
t
A
j
B
V
m
t
s
b
s
b
i
L
s
b
i
s
b
i
L
s
b
i
i
m
f
s
b
P
L
o
s
s
b
A
P
B
Q
Evaluation Warning : The document was created with Spire.PDF for Python.
In
t J
Elec
&
C
om
p
En
g
IS
S
N: 20
88
-
8708
Lo
s
s
alloc
atio
n i
n dist
ribu
ti
on
n
et
work
s wi
th
distri
bu
te
d gene
ra
tors
un
der
goin
g…
(
A
mb
ik
a
Pr
asad H
ot
a
)
3379
(9)
Th
us
,
t
he
t
otal
loss
al
locat
io
n
for
a
co
nsum
er
at
i
th
bus
ca
n
be
cal
c
ulate
d
a
s
the
a
ddit
ion
of
the
c
orres
po
nd
i
ng
loss all
ocati
ons
f
r
om
each
of t
he bra
nc
h
-
b
of
the RD
N, an
d
i
s r
e
pr
ese
nted
in (1
0).
(10)
Hen
ce
, th
e
tota
l real
po
wer
l
oss o
f
th
e sys
te
m
can
be
cal
cu
la
te
d
as:
(11)
3.
PROP
OSE
D HE
URISTI
C NET
WO
RK RECO
NFI
GURATI
ON (
NR)
METHO
D
In
t
he
pro
po
se
d
heurist
ic
NR m
et
ho
d,
an
opt
i
m
al
RDN
is
o
btained
after
a series
of b
ra
nc
h
e
xch
a
nge
s
(BEs)
.
Tw
o
he
ur
ist
ic
ru
le
s
a
re
f
ollow
e
d
in
the
pro
po
se
d
BE
te
chn
iq
ue
to
achieve
a
m
ini
m
u
m
loss
RDN
.
The
first
r
ule
decides
the
ti
e
li
ne
(i.e.
the
first
m
e
m
ber
of
the
BE
pair)
to
be
cl
os
e
d
for
form
ing
a
weak
ly
m
esh
ed
netw
ork
.
T
he
rad
ia
l
natu
re
of
the
ne
twork
is
re
gai
ned
by
ope
ning
a
sel
ect
ed
branch
(secti
on
al
iz
ing
switc
h)
from
t
he
lo
op
s
o
for
m
ed
by
us
in
g
t
he
seco
nd
he
ur
ist
ic
ru
le
.
The
pro
po
se
d
BE
a
lgorit
hm
is
dev
el
op
e
d
to
m
a
intai
n
the
rad
ia
l
str
uctur
e
of
t
he
RD
N
aft
er
eac
h
BE
so
that
loa
d
flo
w
ca
n
be
perform
ed
eff
e
ct
ively
.
Th
us
,
t
he
ar
ray
s
Fig
ur
e
s
2
an
d
3
a
re
to b
e
m
od
ifie
d
acco
rd
i
ng
to n
e
w
R
DN
after
eac
h
BE
s
te
p
f
or
LF
a
nd loss
cal
culat
ion
.
Th
e
NR
m
et
ho
d
i
s
ex
plained
in
detai
l
in
al
gori
thm
-
1.
I
n
this
pap
e
r,
t
he
nor
m
al
l
y
op
e
n
sw
it
ch
or
ti
e li
ne
is de
no
t
ed
as
‘
tl
’, an
d
t
he norm
al
l
y clo
se
d
s
witc
h or
sect
ion
al
iz
ing
switc
h
is
de
no
t
ed
as
‘
ss
’.
Algo
rit
hm
-
1:
STEP
-
1
:
Re
ad
the
in
pu
t
data of t
he
RD
N
a
nd assig
n,
Ntl = N
um
ber
of ti
e li
ne
in t
he
RDN.
STEP
-
2
:
Assign
j
=
1
STEP
-
3
:
Ex
ec
ute
load
fl
ow
a
nd
cal
cula
te
the
real
pow
er
loss
(TPL
oss)
of
the
base
netw
ork
by
the
propose
d
m
et
ho
d.
STEP
-
4:
Find the
pote
nt
ia
l diff
e
ren
ce
a
cro
ss
all
the tie
li
nes
(i.e
. VDi
, for i =
j
,....
.N
t
l ) a
nd ide
ntify
the
ti
e
li
ne
with
m
axi
m
u
m
po
te
ntial
di
ff
e
re
n
ce
(VDm
ax)
.
Als
o,
fin
d
t
he
no
de
vo
lt
age
s
(
Vp
a
nd
Vq) of
the tie
l
ine (
tl
=
[p,
q])
h
a
ving m
axi
m
um
p
otentia
l di
ff
e
ren
ce
VDm
ax,
Let
say V
p < V
q.
STEP
-
5
:
Cl
os
e
the
ti
e
l
ine
(tl)
ha
ving
m
axi
m
u
m
vo
lt
age
diff
e
re
nc
e
(VDm
ax)
an
d
the
n
open
t
he
branc
h
or
sect
io
nal
iz
ing
s
witc
h
(ss
)
adj
ace
nt
to
t
he
node
-
p
of
the
(tl)
to
m
ain
ta
in
rad
ia
l
struct
ur
e
of
the
net
wor
k.
STEP
-
6
:
Com
pu
te
the
powe
r
loss
(TPL
os
s
-
ne
w)
of
t
he
ne
wly
form
ed
RD
N
by
th
e
propose
d
m
e
thod
us
i
ng
m
od
ifie
d
ar
rays.
STEP
-
7
:
Check
TPL
os
s
> TPL
os
s
-
ne
w, if ye
s s
te
p
-
8
is
ex
ec
uted el
se
ste
p
-
11 is foll
owed
.
STEP
-
8:
Assign T
PL
os
s
= TPL
os
s
-
new
STEP
-
9
:
Check
for
al
l
branc
hes
of
t
he
loop,
if
T
PL
oss
>
TPLoss
-
ne
w,
ste
p
-
11
is
e
xecu
te
d
ot
herwise
ste
p
10
is fo
ll
owe
d.
STEP
-
10
:
Sele
ct
the
br
an
ch
ad
j
ace
nt
to
the
pr
e
vious
br
anc
h
as
sect
io
na
li
sing
switc
h
for
the
sam
e
tie
li
ne
(tl),
and f
ollo
w
ste
p
-
6.
STEP
-
11
:
Che
ck,
i
f j=Ntl
pro
ceed to
step
-
12
else i
ncr
e
ase
j
by one
unit
and e
xecu
te
ste
p
-
4.
STEP
-
12
:
The
networ
k
ob
ta
ined
is
the
op
tim
u
m
con
fi
gure
d
net
work
and
the
c
orres
pondin
g
TPL
o
ss
is
the m
ini
m
u
m
l
os
s
of the
n
et
w
ork.
4.
RESU
LT
S
AND DI
SCUS
S
ION
The
pro
po
se
d
BE
base
d
heur
ist
ic
NR
te
chn
i
qu
e
s
al
ong
wit
h
pr
ese
nt
L
A
s
chem
e
are
i
m
ple
m
ented
on
a
12.66
kV,
10
0
kVA
,
33
-
bu
s
RDN
as
s
how
n
in
Fi
gure
4
[
4]
in
MA
TLAB
(R
2018
b)
e
nv
i
ronm
ent
fo
r
evaluat
in
g
ef
f
ect
iveness
of
t
he
de
velo
pe
d
LA
procedu
re
s
against
the
est
ablished
Q
ua
dr
at
ic
Sc
hem
e
[25]
with/wit
hout
DG
s
.
Th
e
rela
te
d
li
ne
an
d
load
data
are
c
ollec
te
d
from
Re
f.
[
4]
and
posit
ion
s
of
D
Gs
are
identifie
d
at
th
e
bu
ses
14,
18
and
32
based
on
the
l
os
s
sen
sit
ivit
y
analy
si
s
as
discuss
e
d
in
[2
9
]
.
The
D
Gs
at
these
buses
in
je
ct
act
ive
pow
ers
into
the
sys
tem
i.e.
58
9.7
kW
at
bus
-
14,
189.5
kW
at
bus
-
18
an
d
1014.
6k
W
at
bus
-
32.
A
ne
w
RDN
is
obta
ined
in
each
BE
ste
p
hen
ce
,
the
cal
culat
i
on
of
po
wer
loss
is
perform
ed
wi
t
h
the
ne
wly
obta
ined
R
DN
at
each
ste
p
of
NR
an
d
c
om
par
ed
with
t
he
pr
e
vious
value
by
f
ollo
wing
the
L
A
,
s
b
i
L
s
b
i
s
b
i
L
s
b
i
p
l
o
s
s
b
s
b
i
A
P
B
Q
1
,
2
,
3
,
.
.
.
.
n
b
r
b
T
p
l
o
s
s
i
p
l
o
s
s
b
s
b
i
w
h
e
r
e
i
n
b
1
nb
i
T
P
L
o
s
s
T
p
l
o
s
s
i
Evaluation Warning : The document was created with Spire.PDF for Python.
IS
S
N
:
2088
-
8708
In
t J
Elec
&
C
om
p
En
g,
V
ol.
10
, No
.
4
,
A
ugus
t
2020
:
3375
–
3383
3380
proce
dure
as
di
scusse
d
in
al
go
rithm
-
1
in
orde
r
to
achie
ve
an
op
ti
m
al
RDN.
The
detai
l
ti
e
l
ine
data
of
the b
ase
and
opti
m
ise
d
33
-
bu
s
RD
N
with/wit
hout
DG
a
re
presen
te
d
in
Ta
ble
1
.
The
t
otal
real
powe
r
loss
of
t
he
base
netw
ork
is
f
ound
to
be
20
2.67k
W
in
th
e
absen
ce
of
DG
un
it
s,
a
nd
it
gets
red
uc
ed
to
14
5.96k
W
f
or
the
rec
onfig
ur
ed
RD
N
a
s
obt
ai
ned
us
i
ng
t
he
pro
posed
BE
base
d
NR
te
c
hn
i
qu
e
.
Mo
re
over
,
it
can
be
n
otice
d
from
the
LA
com
par
ison
Tab
le
2
,
the
R
DN
loss
gets
reduc
ed
to
88.67
kW
at
base
case
due
to
i
nj
ect
io
n
of
D
G
powe
r
int
o
the
syst
e
m
.
The
i
m
ple
m
entat
ion
of
the
pro
pose
d
NR
te
ch
niqu
e
again
de
creas
es
this
lo
ss
to
70.
2k
W
du
e
to
im
pr
ove
m
ent
in
vo
lt
age
pro
file
of
m
axi
m
u
m
load
points.
The
vo
lt
age
m
agn
it
ud
e
s
of
the
ne
twor
k
befor
e
an
d
a
f
te
r
NR
in
th
e
pr
e
sence/a
bse
nce
of
D
Gs
are
c
om
pu
te
d
a
nd
t
he
c
orres
pondin
g
grap
h
is
plo
tt
ed
i
n
Fi
gu
re
5.
It
can
be
ob
se
r
ved
from
this
figure
that
bot
h
NR
an
d
DG
powe
r
in
j
ect
io
n
hav
e
sig
nifican
t
i
m
pact
on
the
im
pr
ov
em
ent
of
volt
age
pro
file
an
d
c
on
seq
uen
tl
y,
on
syst
e
m
loss
re
du
ct
io
n.
T
hu
s
,
the
e
ff
ect
of
PDNR
m
ay
be
reali
zed
f
ro
m
the
i
m
pr
ovem
ent
in
vo
lt
age
pro
file
s
at
bu
s
-
18
a
nd
bus
-
33.
F
urt
herm
or
e,
the
va
riat
ion
of
vo
lt
age
s
f
r
om
su
bst
a
ti
on
bus
to
en
d
buses
a
r
e
m
or
e
in
the
base
R
DN
wit
hout
DG
s
,
wh
i
ch
ca
n
be
c
ons
idere
d
as
an
unhealt
hy
sit
uation
f
or
a
RDN
.
Al
so
,
it
can
be
no
ti
ced
,
the
fluctuati
ons
in
vo
lt
age
s
decr
e
ase
in
the r
ec
onfig
ure
d netw
ork.
Figure
4
.
33
-
B
us
test
Syste
m
b
ef
or
e
N
R
w
it
hout
DG
s
Table
1
.
Detai
l rec
onfig
ur
at
io
n data o
f
t
he
c
on
si
der
e
d 3
3
-
bus R
DN w
it
h/
without
DGs
1
Tie
-
lin
e data
of
the b
ase 3
3
-
b
u
s RDN
b
ef
o
re
NR with
o
u
t DGs
3
3
,
3
4
,
3
5
,
3
6
,
3
7
2
Total activ
e po
we
r
los
s o
f
the b
ase 3
3
-
b
u
s RDN b
ef
o
re
NR with
o
u
t DGs
2
0
2
.67
kW
3
Tie
-
lin
e data
of
the b
ase 3
3
-
b
u
s RDN
af
ter
NR with
o
u
t
DGs
8
,
1
4
,
2
8
,
3
2
,
3
3
4
Total activ
e po
we
r
los
s o
f
the b
ase 3
3
-
b
u
s RDN af
ter
N
R with
o
u
t DGs
1
4
5
.96
kW
5
Tie
-
lin
e dat
a
of
the b
ase 3
3
-
b
u
s RDN
b
ef
o
re
NR with
D
Gs
3
3
,
3
4
,
3
5
,
3
6
,
3
7
6
Total activ
e po
we
r
los
s o
f
the b
ase 3
3
-
b
u
s RDN b
ef
o
re
NR with
DGs
8
8
.67
kW
7
Tie
-
lin
e data
of
the b
ase 3
3
-
b
u
s RDN
af
ter
NR with
DGs
7
,
2
8
,
3
2
,
3
4
,
3
5
8
Total activ
e po
we
r
los
s o
f
the b
ase
33
-
b
u
s RDN af
ter
N
R with
DGs
7
0
.2 k
W
Figure
5. V
oltage P
r
of
il
e of
33
-
B
us
test
syst
e
m
w
it
h/with
out D
Gs be
fore
and after
N
R
2
1
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
(
2
)
(
5
)
(
7
)
(
12
)
(
14
)
(
17
)
(
18
)
(
21
)
(
22
)
(
24
)
(
25
)
(
27
)
(
32
)
33
(
33
)
(
35
)
(
34
)
(
36
)
(
37
)
Evaluation Warning : The document was created with Spire.PDF for Python.
In
t J
Elec
&
C
om
p
En
g
IS
S
N: 20
88
-
8708
Lo
s
s
alloc
atio
n i
n dist
ribu
ti
on
n
et
work
s wi
th
distri
bu
te
d gene
ra
tors
un
der
goin
g…
(
A
mb
ik
a
Pr
asad H
ot
a
)
3381
Table
2
. L
os
s
a
ll
ocati
on
res
ults o
f
the
33
-
bus
test
syst
e
m
w
it
h/without
DGs
Bu
s
No
.
Los
s Allo
catio
n
Resu
lts with
o
u
t DGs
(in k
W
)
Los
s Allo
catio
n
Resu
lts with
DGs
(i
n
kW)
Prop
o
sed
L
A sch
e
m
e
Qu
ad
ratic
sch
e
m
e
Prop
o
sed
L
A sch
e
m
e
Qu
ad
ratic
sch
e
m
e
Bef
o
re
NR
Af
ter
NR
Bef
o
re
NR
Af
ter
NR
Bef
o
re
NR
Af
ter
NR
Bef
o
re
NR
Af
ter
NR
1
0
.31
2
9
0
.30
8
9
0
.21
6
4
0
.21
5
8
0
.23
1
4
0
.23
0
1
0
.18
8
3
0
.18
8
8
2
1
.63
0
3
1
.33
0
8
0
.87
4
3
0
.55
5
9
1
.10
4
2
0
.93
5
6
0
.72
7
7
0
.48
3
7
3
3
.25
0
1
2
.11
3
6
2
.71
1
.49
5
6
2
.23
9
5
1
.56
6
3
2
.20
4
7
1
.28
1
1
4
2
.10
2
6
1
.16
7
9
0
.83
2
3
0
.37
9
1
.30
7
9
0
.83
3
5
0
.69
4
1
0
.33
9
6
5
3
.22
9
5
1
.41
5
3
1
.17
8
5
0
.44
7
2
1
.84
2
4
0
.96
9
7
0
.93
4
9
0
.39
6
7
6
1
1
.19
9
3
4
.84
9
7
1
2
.52
6
2
5
.13
8
9
7
.21
3
4
3
.38
5
1
8
.91
5
8
4
.02
8
1
7
1
2
.35
3
1
5
.08
9
1
4
.34
3
3
5
.38
1
7
7
.62
0
4
5
.27
8
5
9
.91
5
3
8
.42
2
2
8
4
.11
6
9
2
.72
9
1
1
.76
2
7
1
.91
7
5
2
.19
1
.65
3
7
1
.31
3
5
1
.32
2
9
9
4
.49
3
1
2
.52
4
1
2
.04
4
4
1
.73
6
2
2
.26
3
8
1
.76
5
3
1
.47
7
7
1
.51
4
7
10
3
.36
3
3
1
.63
2
1
.43
5
4
1
.17
7
7
2
.08
6
1
1
.52
1
6
1
.10
7
7
1
.15
9
4
11
4
.61
4
1
2
.15
8
3
2
.59
9
1
1
.93
1
7
2
.65
0
7
1
.97
7
1
.91
4
3
1
.99
2
5
12
4
.98
2
1
2
.31
7
4
2
.92
3
2
2
.02
8
9
2
.70
4
4
2
.08
6
4
2
.05
9
9
2
.18
8
8
13
1
0
.11
3
3
4
.55
8
1
1
1
.47
8
5
7
.37
6
4
-
9
.54
1
2
-
9
.55
6
5
-
1
5
.52
7
9
-
1
1
.98
7
14
5
.34
7
9
3
.26
5
2
.42
1
2
2
.05
1
1
1
.94
2
9
1
.69
9
6
1
.55
9
7
1
.66
8
1
15
5
.36
3
3
.18
2
2
.68
2
6
2
.30
5
4
2
.31
2
6
1
.96
3
3
1
.73
8
9
1
.88
5
1
16
5
.47
7
8
3
.34
2
8
2
.76
8
8
2
.42
5
4
2
.35
1
9
2
.05
6
2
1
.74
7
8
1
.93
1
6
17
8
.18
1
6
4
.89
4
8
6
.69
1
6
5
.74
3
6
-
1
.25
9
-
1
.81
2
8
-
2
.65
5
-
2
.90
8
7
18
0
.31
8
3
0
.40
6
0
.19
7
6
0
.29
9
6
0
.23
7
6
0
.30
5
4
0
.17
6
5
0
.25
7
19
0
.59
6
6
1
.53
2
3
0
.47
6
1
.54
8
3
0
.51
5
1
1
.20
6
3
0
.45
4
1
1
.28
7
6
20
0
.64
7
2
1
.81
8
6
0
.52
7
1
.87
3
4
0
.56
5
6
1
.45
3
4
0
.50
5
1
.54
8
3
21
0
.69
1
1
2
.27
2
6
0
.57
1
4
2
.39
9
6
0
.60
9
4
1
.49
8
5
0
.54
9
2
1
.59
4
4
22
1
.95
5
4
2
.06
5
1
1
.01
1
2
0
.76
9
9
1
.45
6
5
1
.50
1
5
0
.85
8
3
0
.66
4
5
23
1
1
.71
1
1
6
.17
6
1
5
.03
0
5
1
8
.01
3
9.
2212
1
1
.55
3
1
1
.94
5
4
1
3
.08
7
24
1
3
.03
4
2
2
1
.58
3
1
6
.39
5
2
2
3
.32
7
1
0
.51
8
6
1
5
.03
3
1
3
.27
4
2
1
6
.24
8
1
25
3
.38
8
6
1
.43
3
1
1
.28
9
2
0
.49
0
4
2
.00
2
6
0
.99
9
8
1
.02
9
8
0
.43
7
3
26
3
.58
6
3
1
.44
8
7
1
.34
2
1
0
.50
6
4
2
.08
8
1
1
.01
5
1
1
.06
8
9
0
.45
3
27
4
.37
4
4
1
.47
6
5
1
.43
8
5
0
.50
6
8
2
.38
4
8
1
.02
9
5
1
.11
8
0
.45
5
3
28
1
0
.17
0
4
6
.92
4
9
7
.39
4
9
2
.82
9
1
6
.41
4
9
5
.00
3
5
5
.47
6
2
.21
9
7
29
2
2
.55
1
7
1
5
.47
0
5
5
4
.40
2
5
3
3
.57
7
8
3
0
.26
9
9
2
0
.72
5
7
4
6
.84
4
3
3
0
.27
4
5
30
1
3
.97
8
9
9
.72
6
7
1
1
.16
6
2
4
.82
5
9
7
.57
7
8
5
.95
5
6
7
.65
1
3
3
.39
7
7
31
1
9
.77
2
1
1
3
.75
3
8
1
9
.69
7
7
9
.52
6
7
-
1
8
.03
2
3
-
1
6
.16
5
2
-
2
2
.29
7
1
-
1
8
.31
7
6
32
5
.74
8
1
2
.99
2
6
2
.22
0
9
3
.15
5
8
3
.58
6
4
2
.53
5
9
1
.69
5
3
2
.68
6
5
Total LA
→
2
0
2
.67
1
4
5
.96
2
0
2
.67
1
4
5
.96
8
8
.67
7
0
.2
8
8
.67
7
0
.2
More
ov
e
r,
it
c
an
be
obser
ve
d
f
ro
m
Table
2
,
the
co
nsum
ers
co
nnect
ed
from
bu
s
-
2
to
bus
-
11
a
r
e
hav
i
ng
a
fair
al
locat
ion
a
fte
r
re
co
nf
i
gurati
on
as
re
gard
t
o
th
ei
r
res
pecti
ve
loa
d
value
s
an
d
ge
ogra
ph
ic
al
locat
ion
s
.
All
the
custom
ers
in
this
gr
ou
p
get
ben
e
fite
d
du
e
to
NR
by
the
pr
opose
d
LA
ap
proac
h.
But,
the
custom
er
at
bu
s
9
is
al
locat
ed
with
m
or
e
loss
by
Qu
a
drat
ic
m
et
ho
d,
w
hich
is
unfair
.
Custom
ers
at
bu
s
-
19
to
22
get
hi
gher
al
locat
ion
s
by
both
the
L
A
m
et
ho
ds
be
cause,
th
e
rela
ti
ve
locat
ion
s
of
the
se
co
nsu
m
ers
changes
due
t
o
rec
onfig
ur
at
io
n
wh
ic
h
le
a
ds
t
o
decr
ease
in
the v
oltag
e
m
agn
it
ud
e
s
f
or
the
se
cust
om
ers.
Du
e
t
o
the
sam
e
abov
e
reas
on,
th
e
c
us
tom
ers
at
bu
s
-
23
t
o
25
get
higher
al
locat
ion
s
.
B
ut,
th
e
pro
posed
L
A
te
chn
i
qu
e
al
ways al
locat
es lesse
r
l
os
ses
to heavil
y l
oa
de
d
c
us
tom
ers
as
com
par
ed
t
o Q
uadrati
c LA
m
et
hod.
The
e
ff
ect
i
veness
of
t
he
pro
po
s
ed
m
et
ho
d
durin
g
t
he
sim
ultaneo
us
im
p
act
of
DG
an
d
NR
ca
n
be
evaluate
d
f
ro
m
the
LA
data
of
Table
2
.
The
loss
al
locat
io
n
of
t
he
D
G
c
onnected
c
onsu
m
er
at
bus
-
14
is
first
analy
sed.
The
relat
ive
locat
io
n
of
t
his
c
on
s
um
er
is
no
t
a
f
fected
by
the
NR.
T
he
pro
pose
d
m
et
ho
d
a
ssign
s
alm
os
t
an
eq
ua
l
a
m
ou
nt
of
lo
ss
(
-
9.5
5kW)
a
t
bo
th
t
he
co
nd
it
ion
s
of
the
ne
twork
whereas
,
Q
uadrati
c
sch
e
m
e
al
locat
es
le
ss
ben
efit
(
-
11.
98kW)
to
the
loa
d
po
int
afte
r
N
R
as
com
par
ed
to
befor
e
NR
(
-
15.52
kW),
w
hich
is
undesire
d.
T
he
DG
co
nn
ect
e
d
con
s
um
ers
at
bu
s
es
-
18
a
nd
32
ha
ve
be
nef
it
te
d
beca
us
e
of
NR
by
bo
t
h
th
e
LA
m
et
ho
ds.
H
ow
ever,
the
pr
ese
n
t
ap
proac
h
pr
ov
i
des
a
n
ade
qu
at
e
am
ount
of
i
ncen
ti
ve
s
to
the
DG
co
nnect
e
d
nodes
18
an
d
32
afte
r
NR
a
ga
inst
Q
uadrati
c
m
et
ho
d.
It
is
id
entifi
ed
th
at
th
e
co
ns
um
ers
at
bus
-
19
to 2
5
a
nd
32
hav
e
higher
L
A
after
NR
by
bo
t
h
the
m
e
thods.
But,
al
l
re
m
ai
nin
g
co
nsu
m
ers
are
al
locat
ed
with
highe
r
losses
after
reconfi
gu
rati
on b
y
Q
uadrati
c LA w
hile
the pr
opos
e
d m
et
ho
d
al
l
ocates l
esser l
os
ses
.
Fu
rt
her,
in
ord
er
to
a
naly
se
r
esp
on
se
of
the
RDN
as
re
gard
to
ge
ographi
cal
locat
ion
s,
two
set
s
of
consum
ers
wit
h
e
qual
loa
d
dem
and
s
a
re
i
de
ntifie
d.
T
he
fir
st
set
co
ns
ist
s
of
tw
o
c
onsu
m
ers
sit
uate
d
far
away
from
each
oth
e
r
i.e.
at
nodes
6
an
d
28.
Bot
h
m
et
ho
ds
a
re
al
locat
ing
m
or
e
l
os
s
to
co
nsum
e
r
at
bus
28
t
ha
n
that
of
at
node
6
but,
the
disc
rim
i
nation
bet
ween
their
L
A
is
m
or
e
prom
inent
i
n
the
present
proce
dure
wh
i
ch
c
a
n
be
ve
rified
fro
m
Figu
re
6.
Sim
il
arly
,
the
discrim
inati
on
be
tween
tw
o
cl
os
e
c
us
tom
ers
connecte
d
at
node
16
and
17
as
show
n
in
Fig
ure
7
is
bette
r
in
the
propos
ed
schem
e
as
com
par
ed
to
Qu
a
dr
at
ic
m
et
hod
a
t
befor
e/
a
fter
re
confi
gurati
on
of
th
e
netw
ork
with/wit
hout
DG
s
.
At
bo
t
h
scenari
os
,
t
he
dev
el
op
e
d
m
eth
od
is
pro
vid
in
g
pr
om
isi
ng
resu
lt
s
against
Q
ua
drat
ic
m
et
ho
d.
Hen
ce
,
the
present
ap
proac
h
of
loss
al
lo
cat
ion
is
fou
nd
su
it
able
to
be
im
ple
m
e
nted
i
n
the
pra
ct
ic
al
fiel
d
of
a
pp
li
cat
io
n
f
or
f
ai
r
al
locat
ion
with
op
ti
m
al
network
reconfi
gurati
on.
Evaluation Warning : The document was created with Spire.PDF for Python.
IS
S
N
:
2088
-
8708
In
t J
Elec
&
C
om
p
En
g,
V
ol.
10
, No
.
4
,
A
ugus
t
2020
:
3375
–
3383
3382
Figure
6. Dif
fe
ren
ce
in
L
A be
tween
node
s
6 and 2
8
Figure
7. Dif
fe
ren
ce
in
L
A be
tween
node
s
16 a
nd 17
5.
CONCL
US
I
O
N
The
im
pact
of
NR
on
syst
em
loss
al
locat
ion
has
bee
n
analy
sed
i
n
th
is
pap
e
r
with
a
ne
w
L
A
te
chn
iq
ue.
The
propose
d
BE
base
d
NR
te
c
hn
i
qu
e
is
sim
ple
to
unde
rstand
a
nd
easy
to
im
ple
m
ent.
Ag
ai
n,
the
loss
al
loca
ti
on
m
et
ho
d
de
velo
ped
el
im
i
nates
the
ef
fe
ct
s
of
cr
os
s
-
te
rm
m
at
he
m
ati
cal
ly
fr
om
the
loss
form
ulati
on
wi
thout
a
ny
ass
um
pt
ion
s
a
nd
a
ppr
ox
im
at
ion
s.
The
ef
fecti
ve
ne
ss
of
t
he
pro
pose
d
L
A
te
c
hniqu
e
is
ver
ifie
d
by
co
m
par
ing
the
L
A
res
ults
with
the
est
ablishe
d
“Q
ua
dr
at
ic
Me
thod”
with/
without
D
Gs
a
t
bo
th
sit
uations
of
th
e
netw
ork,
i.e.
befor
e
a
nd
afte
r
NR.
T
he
res
ul
ts
are
found
to
be
fa
ir
a
nd
prom
isi
ng
.
It
al
locat
es
losses
to
the
ne
twork
par
ti
ci
pan
ts
with
due
con
si
der
at
io
n
to
their
load
dem
and
s
an
d
ge
ogra
ph
ic
al
loc
at
ion
s.
Si
m
ultaneo
us
ly
,
it
al
so
dis
tribu
te
s
t
he
be
nef
it
of
NR
an
d
D
G
power
i
nj
ect
i
on
am
on
g
the
netw
or
k
par
ti
ci
pa
nts, j
udic
io
usl
y.
REFERE
NCE
S
[1]
A.
G.
Exposi
to,
J.
R.
San
tos,
T
.
G.
Garc
i
a,
and
E.
R
.
Vel
asc
o,
“
Fair
Allocat
ion
of
Tra
nsm
ission
Pow
er
Losses
,
”
IEE
E
Tr
ansa
ct
io
ns on
Powe
r
Sys
te
ms
,
vol
.
15
,
no
.
1
,
pp
.
184
-
188
,
2000.
[2]
S.
Mishra,
D.
Das,
and
S.
Paul,
“
A
Sim
ple
Algorit
hm
fo
r
Distribut
ion
S
y
s
te
m
Loa
d
Flow
with
Distribut
e
d
Gene
ration
,
”
In
I
nt
.
Con
f
.
on
Re
c
ent
Adv
an
.
and
I
nnovat
ions
in
En
gine
ering
,
pp
.
1
-
5,
2014
.
[3]
A.
J.
Cone
jo,
J.
M.
Arro
y
o
,
N.
Alguac
i
l,
and
A
.
L
.
Guij
arr
o,
“
Tra
nsm
ission
Loss
Alloc
at
ion
:
A
Com
par
ison
of
Diffe
ren
t
Prac
ti
c
al
Algor
it
hm
s
,”
I
EE
E
Tr
ansacti
o
ns on
Powe
r
Sys
te
ms
,
vol
.
17
,
no
.
3
,
pp
.
571
-
576
,
2002.
[4]
E.
M.
B
ara
n
,
a
nd
F.
F.
W
u,
“
Network
Rec
o
n
figura
t
ion
in
Di
stribut
ion
S
y
s
tem
s
for
Loss
Reduc
ti
on
and
Lo
a
d
Bal
an
ci
ng
,
”
I
EEE
Tr
ans P
ower
Del.
,
vol
.
4
,
no
.
2,
pp
.
1
401
-
7
,
1
989.
[5]
M.
Atana
sovs
ki,
and
R
Ta
le
ski,
“
Pow
er
Summ
at
ion
Method
for
Loss
Alloc
at
ion
in
Radi
al
Distri
buti
on
Networks
with
DG
,
”
IEEE
Tr
ansacti
ons on
Powe
r S
yste
ms
,
vol.
26
,
no
.
4
,
pp
.
2491
-
2499
,
20
11
.
[6]
M.
Atana
sovs
ki,
and
R
Taleski
,
“
Ene
rg
y
Sum
m
at
ion
Met
hod
for
Loss
Alloc
ation
in
Radial
Distrib
uti
on
Networks
with
DG
,
”
IEEE
Tr
ansacti
ons on
Powe
r S
yste
ms
,
vol.
27
,
no
.
3
,
pp
.
1433
-
1440
,
20
12.
[7]
Z.
Ghofran
i
-
Jah
rom
i,
Z.
Mahm
oodza
deh
,
and
M.
Ehsan,
“
Distribut
ion
Loss
Alloc
a
ti
on
for
Radi
al
S
y
st
ems
inc
ludi
ng
DG
s,”
IEE
E
Tr
ansacti
o
ns on
Powe
r
Del
iv
ery
,
vo
l. 29, no
.
1,
pp.
72
-
80,
20
14.
[8]
K.
M.
Jagta
p,
an
d
D.
K.
Khatod,
“
Loss
Alloc
a
t
ion
in
Radi
al
Distri
buti
on
Networks
with
Diffe
ren
t
L
oad
Models
and
Distribut
ed
Gen
era
t
ions
,
”
IET
G
ene
ration
,
Tr
ansm
ission
&
Distri
buti
on,
vo
l
.
9
,
n
o.
12
,
pp
.
1275
-
1291,
2015
.
[9]
S.
S.
Kash
y
ap
,
and
M.
De
“
A
Novel
Loss
All
oca
t
ion
Method
for
Rad
ia
l
Dist
ribut
ion
S
y
stem
with
Distribu
ted
Gene
rations,”
1st I
E
EE Int
.
Conf
.
on
Pow
er
E
le
c
tronic
s.
In
telli
g
ent Cont
rol an
d
En
ergy
Syst
ems
,
pp
.
1
-
6
,
2016
.
[10]
S.
Ghae
m
i,
and
K.
Za
re,
“
Loss
Alloc
at
io
n
in
Restruc
tu
red
Radi
a
l
Distribut
ion
Netw
orks
Consideri
ng
the
Con
tra
c
tual
Pow
er
,”
I
ET
Ge
nerati
on,
Tr
ansm
ission
&
Distri
buti
on,
vol
.
11
,
no.
6
,
pp
.
1389
-
1397,
2016
.
[11]
K.
M.
Jag
ta
p
,
a
nd
D.
K.
Khato
d,
“
Novel
Appr
oac
h
for
Lo
ss
Alloc
a
ti
on
of
D
istri
buti
on
Netw
orks
with
DG
s,”
El
e
ct
ric
Pow
er
S
yste
m Research,
vol.
143
,
pp
.
303
-
311,
2017
.
[12]
S.
Sharm
a,
and
A.
R
.
Abh
y
an
kar
,
“
Loss
Allo
ca
t
ion
for
W
e
a
kl
y
Meshed
Dis
tri
buti
on
S
y
st
e
m
using
Anal
y
t
ic
a
l
Form
ula
ti
on
of S
hapl
e
y
Va
lue
,
”
IEE
E
Tr
ans
act
io
ns on
Powe
r
Sys
te
ms
,
vol
.
32
,
no
.
2
,
pp
.
1369
-
13
77,
2017
.
[13]
P.
Kum
ar,
N.
G
upta
,
K.
R
.
Niazi,
and
A.
Sw
arn
kar
,
“
A
Circ
uit
The
or
y
-
base
d
L
oss
Alloc
at
ion
Method
for
Active
Distribut
ion
S
y
st
ems
,
”
IE
EE
Tr
a
nsacti
ons on
Sm
art Grid.
,
vol. 1
0,
no
.
1
,
pp
.
100
5
-
12
,
2019
.
[14]
S.
Mishra,
D.
Das,
and
S.
Pa
ul,
“
A
Com
pre
hensive
Surve
y
o
n
Pow
er
Distribut
ion
Network
Rec
onfigur
at
ion
,
”
Ene
rgy
S
yste
ms
,
vol.
8
,
no
.
2
,
pp
.
227
-
284,
2017
.
0
0
.2
0
.4
0
.6
0
.8
1
1
.2
1
.4
W
ith
o
u
t D
Gs
b
efore
NR
W
ith
o
u
t D
Gs
af
ter
NR
W
ith
DG
s
b
efore
NR
W
ith
DG
s
af
ter
NR
(
kW
)
Diff
erence
in LA
bet
w
een nodes
6 a
nd
2
8
Pr
o
p
o
sed
Met
h
o
d
Qu
ad
rati
c M
eth
o
d
0
0
.02
0
.04
0
.06
0
.08
0
.1
0
.12
0
.14
0
.16
0
.18
W
ith
o
u
t D
Gs
b
efore
NR
W
ith
o
u
t D
Gs
af
ter
NR
W
ith
DG
s
b
efore
NR
W
ith
DG
s
af
ter
NR
(
kW
)
Diff
erence
in LA
bet
w
een nodes
16
and 1
7
Pr
o
p
o
sed
Met
h
o
d
Qu
ad
rati
c M
eth
o
d
Evaluation Warning : The document was created with Spire.PDF for Python.
In
t J
Elec
&
C
om
p
En
g
IS
S
N: 20
88
-
8708
Lo
s
s
alloc
atio
n i
n dist
ribu
ti
on
n
et
work
s wi
th
distri
bu
te
d gene
ra
tors
un
der
goin
g…
(
A
mb
ik
a
Pr
asad H
ot
a
)
3383
[15]
M.
A.
Muham
m
ad,
et
al
.
,
“
I
nte
gra
te
d
Dat
ab
ase
Approac
h
i
n
Multi
Objecti
ve
Network
Re
conf
igurat
ion
fo
r
Distribut
ion
S
y
stem
using
Dis
cre
t
e
Optimis
ation
Te
chn
ique
s,
”
IET
Gene
rat
ion
Tr
ansm
is
sion
Distributi
on,
vol.
12
,
no
.
4
,
pp
.
976
-
986
,
2018
.
[16]
Y.
Ch,
S.
K.
Go
sw
ami,
and
D.
Chat
terje
e,
“
Eff
ec
t
of
Network
Rec
onfigur
at
ion
on
Po
wer
Quali
t
y
of
Distribut
io
n
S
y
stem,”
Int
ernati
onal Journal of
E
le
c
tric
al
Po
wer
&
Ene
rgy
Syste
ms
,
vol
.
83
,
pp.
87
-
95
,
2016
.
[17]
S.
Ghasem
i,
“
B
al
an
ce
d
and
Un
bal
an
ce
d
Distrib
uti
on
Networks
Rec
onfigur
at
ion
Consideri
ng
Re
li
ability
Ind
ic
es
,
”
in
Shams
Engi
n
e
ering
Journal
,
v
ol.
9
,
no
.
4
,
pp
.
1
5
67
-
15
79,
2016
.
[18]
S.
R.
Gam
pa
,
an
d
D.
Das,
“
Multi
-
Objec
t
ive
Appr
oac
h
for
Re
conf
i
gura
ti
on
of
Distr
ibut
ion
S
y
stems
with
Distributed
Gene
rations,”
Elec
tri
c
Pow
er
Co
mponents
and
S
y
stems,
vol
.
45
,
n
o.
15
,
pp
.
1678
-
1690,
2017
.
[19]
S.
Das,
D.
Das,
and
A.
Patra,
“
Rec
onfi
gur
at
ion
of
Distribut
ion
Networks
with
Optimal
Placem
ent
of
Distribu
ted
Gene
rations
in
t
he
Presenc
e
of
Remote
Volta
g
e
Control
le
d
Bus
,
”
R
ene
wabl
e
a
nd
Sustainabl
e
Ene
rgy
R
ev
i
ews,
vol.
73
,
pp
.
772
-
781,
2017
.
[20]
G.
Gutié
rr
ez
-
Al
c
ara
z
,
and
J.
H.
T
ovar
-
Herná
nd
ez
,
“
Two
-
Stage
He
uristi
c
M
et
hodol
og
y
for
Optimal
Rec
onfigur
at
io
n
and
Volt/
VA
r
Control
in
the
Opera
ti
on
of
El
e
ct
ri
ca
l
Distri
buti
on
S
y
stems
,
”
IET
Gene
ration
Tr
an
sm
issio
n
Distributi
on,
vol
.
11
,
no
.
16
,
pp
.
3946
-
3954,
201
7.
[21]
A.
T
y
agi,
A.
Verm
a,
and
R.
P
.
Bij
we
,
“
Re
co
nfigura
t
ion
for
Loa
dab
il
i
t
y
Li
m
it
Enha
n
ce
m
ent
of
Distribut
io
n
S
y
stems
,
”
I
ET
G
ene
ration
Tr
ans
miss
ion
Distributi
on,
vol.
12
,
no
.
1,
pp.
88
-
93,
20
18.
[22]
K.
Jasthi,
and
D
.
Das,
“
Sim
ult
an
eous
Distribut
io
n
S
y
stem
Re
con
figura
t
ion
and
DG
sizi
ng
Algori
t
hm
W
it
hout
Lo
a
d
Flow Soluti
on,
”
IET
Gene
ration
Tr
ansm
i
ss
ion
Distributi
on,
vol. 1
2,
no
.
6
,
pp
.
130
3
-
1313,
2018
.
[23]
M.
E.
Olive
ir
a,
e
t
al.
,
“
Network
Rec
onfigur
at
ion
and
Loss
Alloc
ation
in
a
Dere
gul
at
ed
Env
ironment
of
Distribut
io
n
S
y
stems
,
”
In
Pr
oc.
o
f the
18th
In
t.
Con
f. on
El
e
ctr
ic
it
y
Distribut
io
n,
pp
.
6
-
9
,
2005
.
[24]
A.
J.
Conej
o,
F.
D.
Gali
an
a,
a
nd
I.
Kocka
r,
“
Z
-
Bus
Loss
All
oca
t
ion,
”
I
EE
E
Tr
ansacti
ons
on
Powe
r
Syste
ms
,
vol.
16
,
no
.
1
,
pp
.
105
-
110
,
2001
.
[25]
J.
S.
Savie
r,
and
D.
Das,
“
I
m
pac
t
of
Network
Rec
onfiguration
on
Loss
All
oca
ti
on
of
Radi
al
Distri
buti
on
S
y
stems
,
”
IEE
E
Tr
ansacti
o
ns on
Powe
r
Del
iv
ery
,
vo
l. 22, no
.
4
,
pp
.
2473
-
24
80,
2007
.
[26]
C.
Djab
al
i
,
A.
B
oukar
oura
,
N.
K
et
fi
,
and
T
.
Bouk
ti
r, “
Optimum
D
istri
buti
on
Netw
ork
Rec
onf
igura
t
ion
using Fire
f
l
y
Algorit
hm
,
”
Proc.
o
f the
In
t. Con
f.
on
R
ec
en
t Adv
ance
s in
E
le
c
trical
Syst
ems
,
pp.
2
87
-
283,
2016
.
[27]
M.
Quinte
ro
-
Duran,
E
.
J.
Candelo,
and
J.
Soto
-
Ortiz
,
“
A
Modifie
d
Bac
kward
/Forward
Sw
ee
p
-
base
d
Method
for
Rec
onfigur
at
ion
of
Unbala
n
ced
Distribut
ion
Networks,”
Int
e
rnational
Journ
al
of
E
le
c
tric
a
l
and
Computer
Engi
ne
ering
(
IJ
ECE
)
,
vol
.
9
,
no
.
1,
pp.
85
-
101,
2
019.
[28]
S.
A.
Redd
y
,
D.
M.
Redd
y
,
a
nd
K.
S.
M.
R
edd
y
,
“
Network
Rec
onfiguratio
n
of
Pri
m
ar
y
Distribut
ion
S
y
st
e
m
Us
ing
GW
O
Algorit
hm
,
”
Int
ernati
onal
Journal
of
Elec
tri
cal
an
d
Computer
En
gin
ee
ring
(
IJE
C
E)
,
vol.
7,
no.
6
,
pp.
3226
-
32
34
,
2017.
[29]
S.
R.
Rao,
et
a
l.,
“
Pow
er
Loss
M
ini
m
iz
ation
in
D
istri
buti
on
S
y
ste
m
using
Networ
k
Rec
onfigur
at
io
n
in
the
Presen
c
e
of
Distributed
G
ene
ra
ti
on,
”
I
EEE
Tr
ansacti
ons
on
Powe
r S
ystem
s,
vol. 28, no.
1,
pp
.
317
-
3
25
,
2013.
BIOGR
AP
H
I
ES
OF
A
UTH
ORS
Amb
ika P
rasad H
ota
r
ecei
ved
h
is B.
E
degr
ee
in El
e
ct
ri
ca
l
Eng
in
ee
ring
in
2005
fr
om
Bij
u
Patn
ai
k
Univer
sit
y
of
T
ec
hnolog
y
,
Indi
a,
and
his
M
.
Tech
degr
ee
in
El
e
ct
ri
ca
l
Engi
n
ee
r
i
ng
in
2013
from
India
n
Insti
tute
of
Technol
og
y
,
Khara
gpur,
Ind
i
a.
He
is
cur
r
entl
y
pursuing
his
PhD
progra
m
in
Inte
rna
ti
ona
l
Instit
ut
e
of
Inform
at
ion
T
ec
hno
log
y
,
Bhub
ane
sw
ar
,
India
in
Elec
tr
ic
a
l
Engi
n
ee
ring
with
emphasis
in
Sm
art
Grid
ma
nage
m
ent
s
y
s
tem
s.
His
rese
arch
int
ere
sts
inclu
de
oper
ation
and
co
ntrol of
power
s
y
st
ems
,
ren
ewa
ble
ene
rgi
es,
and
net
work m
ode
lling
Sivk
umar
Mi
sh
ra
rec
e
ive
d
hi
s
B.
E
degr
e
e
in
El
e
ct
ri
ca
l
Enginee
ring
from
Mala
v
i
y
a
Reg
ional
Engi
ne
eri
ng
Col
le
ge
,
Jaipur
aff
i
li
ated
to
Univer
sit
y
of
Raj
ast
an
in
1995,
M.T
ec
h
degr
ee
from
In
dia
n
Insti
tut
e
of
Te
chno
log
y
,
Khara
gpur,
Ind
i
a,
and
his
PhD
degr
ee
from
Jada
vpur
Univer
sit
y
,
India
.
He
is
cur
r
ent
l
y
working
as
As
socia
te
Prof
e
ss
or
in
the
d
epa
r
tment
of
E
lectr
i
c
al
Eng
ineeri
ng
,
CAP
G
S,
BP
UT,
Rourkel
e
,
Indi
a.
His
rese
arc
h
int
er
ests
inclu
de
Ele
ctric
al
P
ower
distri
but
io
n
s
y
stem
anal
y
s
is,
Distribut
ed
Ge
ner
ation
and
Mi
cro
Grid,
Sm
art
Grid,
Applicat
i
on
of
IOT
&
Bi
g
dat
a
an
aly
t
ic
s in
El
e
ct
ri
ca
l
power
s
y
st
ems
.
Evaluation Warning : The document was created with Spire.PDF for Python.