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.
9
, No
.
6
,
Decem
ber
201
9
, p
p.
4507
~
4515
IS
S
N:
20
88
-
8708
,
DOI: 10
.11
591/
ijece
.
v
9
i
6
.
pp4507
-
45
15
4507
Journ
al h
om
e
page
:
http:
//
ia
es
core
.c
om/
journa
ls
/i
ndex.
ph
p/IJECE
A new d
esig
n
alg
or
ithm
f
or hybri
d activ
e power
filter
Chau
Minh
Thuy
e
n
Facul
t
y
of Electr
ic
a
l
Eng
ineeri
ng
Technol
og
y
,
In
dustria
l
Univer
si
t
y
of
Ho
Chi
Mi
nh
Cit
y
,
Vi
et
na
m
Art
ic
le
In
f
o
ABSTR
A
CT
Art
ic
le
history:
Re
cei
ved
Ma
r 3
, 2
01
9
Re
vised
A
pr 17
, 2
01
9
Accepte
d
J
un
2
5
, 201
9
The
cor
re
ct
d
et
e
rm
ina
ti
on
of
the
par
amet
ers
of
H
y
brid
A
ct
iv
e
P
ower
Filt
e
r
(HA
PF
)
pla
y
s
a
decisive
rol
e
in
it
s
per
form
anc
e
.
Th
ere
for
e
,
thi
s
pap
er
proposes
a
new
design
al
gorit
h
m
for
H
APF
b
ase
d
on
the
Social
Spider
Algorit
hm
(SS
A
).
Thi
s
al
gori
th
m
has
the
adv
a
nta
ge
tha
t
it
is
poss
ibl
e
to
det
ermine
the
p
ara
m
et
ers
o
f
bo
th
the
power
circui
t
par
t
and
t
he
con
trol
ci
rcu
it
par
t
of
HA
PF
.
The
achie
ved
r
esult
s
are
m
ult
i
-
purpo
se,
such
as:
m
ini
m
um
tot
al
h
armonic
distort
i
on
of
the
suppl
y
cur
ren
t
and
sou
rce
voltage
,
the
m
axi
m
um
re
ac
t
ive
power
co
m
pensa
ti
on
int
o
the
s
y
stem
and
sati
sf
y
m
a
n
y
constra
in
ts
such
as:
s
y
stem
stab
il
ity
,
resona
n
ce
condi
ti
ons
of
th
e
bra
nch
es
and
the
li
m
it
s
of
the
p
ara
m
ete
rs.
Com
par
ed
t
o
tra
di
ti
ona
l
de
sign
m
et
hod
using
the
Part
icle
Sw
arm
Optimiza
ti
on
al
gori
t
hm
,
the
propose
d
al
gori
thm
show
s
the
adva
nta
ges
of
sm
al
le
r
tot
al
har
m
onic
distort
ion
of
suppl
y
cur
r
ent
and
source
voltage,
and
highe
r
react
iv
e
power
compensa
ti
on
i
nto
th
e
gr
i
d
while
st
il
l
m
ee
t
i
ng
the c
onstr
ai
nt
s.
Ke
yw
or
d
s
:
Acti
ve powe
r
f
il
te
r
Hybr
i
d
act
ive
powe
r
filt
er
Mult
i
-
obj
ect
iv
e optim
iz
at
ion
Passive
powe
r fil
te
r
So
ci
al
s
pid
er
a
lgorit
hm
Copyright
©
201
9
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
:
Chau
Mi
nh T
huye
n,
Faculty
of Elec
tric
al
Engineer
ing
Tech
nolo
gy
,
Ind
us
tria
l U
nive
rsity
o
f
Ho C
hi Mi
nh Ci
ty
,
12 Ng
uyen Va
n
Ba
o
Roa
d, 4
Ward,
Go
Va
p Dist
rict
, Ho
C
hi
Mi
nh
Ci
ty
70
0000, Vie
tnam
.
Em
a
il
:
chau
m
i
nh
t
hu
ye
n@
i
uh.edu.
vn
1.
INTROD
U
CTION
To
day,
m
or
e
and
m
or
e
pow
e
r
el
ect
ronic
de
vices
are
co
nn
ect
ed
to
the
el
ect
rical
syst
e
m
,
resu
lt
in
g
in
poor
pow
er
qual
it
y.
Acco
r
din
gly,
to
im
pr
ove
the
powe
r
qu
al
it
y,
the
Pa
ssive
P
ow
e
r
F
il
te
rs
(P
PFs
)
[
1]
ar
e
connecte
d
in
pa
rall
el
with
th
e
load
to
s
uppress
the
h
a
rm
on
ic
s
ge
ne
rated
by
the
nonli
ne
ar
loa
d.
H
owever,
the
PP
F
has
t
he
disad
va
ntag
e
that
it
is
eas
y
to
res
onat
e
with
the
syst
em
i
m
ped
ance
and
the
non
-
fl
exible
com
pen
sat
ion
capab
il
it
y.
Sin
ce
the
n,
t
he
A
ct
ive
P
ow
e
r
Fil
te
r
(
AP
F
)
was
bor
n
[
2],
it
is
consi
der
e
d
as
a
go
od
so
luti
on
f
or
har
m
on
ic
filt
erin
g
an
d
re
act
ive
powe
r
com
pen
sa
ti
on.
Howe
ver
,
the
AP
F
a
lso
ha
s
the d
isa
dva
ntage
of large
in
ve
st
m
ent co
sts, d
iffic
ult t
o
appl
y fo
r
high
vo
lt
age grids.
F
r
om
the sh
or
tc
om
ing
s o
f
PPF
an
d
A
PF
that
Hybr
i
d
A
ct
ive
Po
we
r
Fi
lt
er
(H
A
PF
)
w
as
born
as
a
ne
cessi
ty
.
HA
P
F
is
a
co
m
bin
at
ion
of
bo
t
h
PP
F
an
d
AP
F
,
so
it
ha
s
bo
t
h
ad
va
ntag
es
of
PPF
a
nd
AP
F
.
T
he
Hyb
rid
A
ct
ive
P
ower
Fil
te
r
(HA
PF)
is
consi
der
e
d
as
t
he
m
os
t
eff
ect
ive
s
olu
ti
on
in
har
m
on
ic
filt
er
ing
a
nd
reacti
ve
powe
r
com
pen
sat
io
n.
How
ever,
the
w
orkin
g
eff
ic
ie
ncy
of
H
APF
de
pe
nds
on
m
any
facto
rs
s
u
ch
as:
co
rrec
t
determ
inati
on
of
the p
ar
a
m
et
ers
[3
]
, c
on
tr
ol m
e
thod [4
-
8], h
a
r
m
o
nic cu
rr
e
nt d
et
ect
ion
m
et
ho
d [9
, 10], con
trol strategy [
11]
and
sta
bili
zat
ion
of
the D
C
-
bus
vo
lt
age for
t
he
in
ver
t
er
[12
]
.
In
wh
ic
h,
t
he
c
orr
ect
d
et
erm
inatio
n o
f
the
p
a
ra
m
et
ers
of
the
HAPF
is
the
m
os
t
i
m
p
or
ta
nt
factor,
i
t
decides
the
ef
fici
ency
of
harm
on
ic
s
filt
ering
an
d
reacti
ve
powe
r
com
pen
sat
ion
of HAPF
.
The
st
ru
ct
ur
e
of
H
AP
F
c
on
s
ist
s
of
tw
o
pa
rts:
the
powe
r
ci
rcu
it
par
t
a
nd
the
co
ntr
ol
ci
rcu
it
par
t
.
Un
ti
l
now
,
m
ul
ti
-
obj
ect
ive
op
t
i
m
iz
ation
stu
dies
for
the
par
a
m
et
ers
determ
i
nation
of
powe
r
ci
rcu
it
pa
rt
c
an
be
su
m
m
arized
as
fo
ll
ows:
the
a
pp
li
cat
io
n
of Ge
netic
Algo
rith
m
(G
A)
to
m
ul
ti
-
obj
ect
ive optim
iz
ation
de
sign
f
or
passive
powe
r
filt
ers
(P
P
Fs)
i
s
pro
po
se
d
by
[13].
H
oweve
r,
the
G
A
has
t
he
disad
va
ntage
of
c
om
pu
ta
ti
on
al
ly
exp
e
ns
i
ve
i.e.
tim
e
-
con
su
m
ing
.
A
no
t
her
al
gorithm
i
s
al
so
com
m
on
ly
us
ed
for
H
APF
desig
n,
cal
le
d
the
Partic
le
Sw
a
rm
Op
tim
iz
ation
(
PS
O)
al
gorithm
.
T
he
PS
O
a
l
gorithm
is
us
ed
t
o
desig
n
a
n
A
PF
in
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.
9
, N
o.
6
,
Dece
m
ber
201
9
:
4507
-
4515
4508
four
-
wi
re
three
-
phase
syst
em
in
case
of
balanced
a
nd
unba
la
nced
loa
ds
but
on
ly
co
ns
id
er
to
desi
gn
f
or
PP
F
s
and
optim
iz
ati
on
for
the
APF
bu
t
no
t
f
or
the
H
AP
F
[
14
-
17]
.
A
pp
li
cat
ion
of
t
he
bat
al
gorithm
to
m
ul
ti
-
obj
ect
ive
opti
m
iz
at
ion
desig
n
of
passi
ve
powe
r
filt
ers
set
with
ti
m
e
-
vary
ing
ine
rtia
we
igh
ts
[18]
a
nd
a
no
t
her
stud
y
t
o
m
ulti
-
obj
ect
ive
desi
gn
on
H
APF
is
us
in
g
a
n
a
nt
c
olony
al
gorith
m
[1
9]
but
on
l
y
asl
o
m
ulti
-
obj
ect
ive
op
ti
m
iz
ation
de
sign
f
or
PPF
s.
A
hm
ed
Fa
he
e
m
Zob
aa
ha
s
us
e
d
the
F
ort
ran
Fea
sible
Sequentia
l
Q
ua
dr
at
i
c
Pr
og
ram
m
ing
al
gorithm
to
so
lve
the
m
ulti
-
ob
j
ect
ive
opti
m
iz
at
ion
pr
obl
e
m
fo
r
H
AP
F
with
the
aim
of
fin
ding P
PF pa
ram
et
ers
[2
0].
In
su
m
m
ary,
the
pre
vious
researc
h
on
m
ul
ti
-
obj
ect
ive
opti
m
iz
at
ion
f
or
H
AP
F
was
t
he
on
ly
m
ul
ti
-
obj
ect
ive
op
ti
m
iz
at
ion
desig
n
of
t
he
powe
r
ci
r
cuit
par
t
(par
a
m
et
ers
of
th
e
PPF)
.
Me
a
nwhile
,
the
opti
m
iz
ed
desig
n
f
or
the
par
am
et
ers
of
t
he
co
ntr
ol
ci
rc
uit
par
t
has
no
t
been
stu
died
a
nd
t
he
cal
culat
ion
of
the abo
ve para
m
et
ers
wit
hout
conside
rin
g
th
e co
nd
it
io
n of t
he
syst
em
stabil
it
y.
To
ove
rco
m
e
this
disad
va
ntage,
in
this
pa
per,
the
Ro
ut
h’
s
sta
bili
ty
sta
nd
a
rd
is
use
d
to
c
hec
k
the
sta
ble
co
nd
it
ion
of
the H
A
PF
syst
em
.
Then
a n
ew
m
ulti
-
ob
j
ect
ive
opti
m
iz
at
ion
desig
n
al
gorithm
based o
n
the
So
ci
al
S
pi
der
Algorithm
(S
S
A)
[21]
is
pro
po
se
d
to
de
te
rm
ine
the
be
st
set
of
par
am
et
ers
f
or
t
he
H
AP
F
.
The
pro
pose
d
al
gorithm
has
the
adv
a
nta
ge
of
fi
nd
i
ng
a
quic
k
s
olu
ti
on
with
li
tt
le
loops
an
d
it
can
f
ind
al
l
the
pa
ram
et
ers
of
both
pow
er
ci
rc
uit
par
t
a
nd
co
ntr
ol
ci
rcu
it
pa
rt.
T
he
re
su
lt
s
ac
hi
eved
will
be
global
op
ti
m
iz
ation
s
uch
as
t
he
m
ini
m
u
m
of
the
to
ta
l
har
m
on
ic
di
stortion
of
t
he
source
volt
ag
e
an
d
s
upply
c
urren
t,
m
axi
m
u
m
r
eact
ive pow
e
r
c
om
pen
sat
ion
int
o gr
i
d
in
stead
y
-
sta
te
an
d sat
isfy t
he st
able c
o
ndit
ions.
The
st
ru
ct
ur
e
of
the
pa
pe
r
is
di
vid
ed
int
o
fi
ve
pa
rts.
Pa
rt
1
is
over
view
intr
oductio
n
of
the
i
ssu
es
that
need
t
o
be
in
ve
sti
gated.
T
he
sel
ect
ed
Hyb
rid
Acti
ve
P
owe
r
Fil
te
r
topolo
gy
is
pr
ese
nte
d
in
Pa
rt
2.
Pa
rt
3
is
m
ul
ti
-
obj
ect
ive
op
ti
m
iz
a
ti
on
al
g
ori
thm
fo
r
H
ybrid
Acti
ve
P
ow
e
r
Fil
te
r.
T
he
si
m
ulati
on
re
su
lt
s
an
d
disc
ussi
on
are
pr
ese
nted
in
Pa
rt
4,
a
nd th
e co
nclusi
on
s
a
re sum
m
arized in
Pa
rt
5.
2.
THE
SELE
C
TE
D
H
YBRID
AC
TI
VE P
OWER
FILT
ER
TOP
OLO
GY
The
sel
ect
ed
Hybr
i
d
Acti
ve
Power
Fil
te
r
topolo
gy
is
s
hown
i
n
Fi
gure
1.
It
co
ns
ist
s
of
five
par
ts:
U
s
an
d
Z
s
are
t
he
s
ource
vo
lt
a
ge
a
nd
t
he
resist
ance
of
t
he
s
ource.
C
1
a
nd
L
1
are
ca
pacit
ance
a
nd
i
nduct
ance
at
the
f
undam
ental
fr
e
qu
e
ncy.
C
F
is
the
a
dd
e
d
ca
p
aci
ta
nce
t
o
c
om
pen
sat
e
reacti
ve
powe
r
.
T
he
tra
ns
f
or
m
er
has
rati
o
1
:1
t
o
protect
,
ins
ulate
betwee
n
the
s
ource
a
nd
t
he
vo
lt
age
sour
ce
inv
e
rter.
L
0
is
the
outp
ut
filt
er
of
the volt
age
sou
rce in
ver
te
r.
A
ccordin
g
t
o
[
3]
, th
e c
ontr
ol b
l
ock d
ia
gr
am
o
f
HAPF
is s
ho
w
n
in
Fig
ure
2.
Figure
1.
To
po
log
y
of
the
hybri
d
act
ive
po
w
er f
il
te
r
X
X
L
h
i
)
(
s
G
c
)
(
s
G
i
n
v
)
(
s
G
o
u
t
s
h
i
-
-
F
h
I
i
n
v
U
Fig
ur
e
2. Co
nt
ro
l
blo
c
k diag
r
a
m
o
f H
AP
F
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
A n
ew
desi
gn al
go
rit
hm f
or
hy
br
id
a
ct
iv
e
powe
r fi
lt
er
(
Cha
u
Mi
nh T
hu
ye
n
)
4509
Fr
om
Fig
ur
e
2, we ha
ve
the
tr
ansf
e
r funct
io
n o
f
loa
d harm
on
ic
curre
nt
I
Lh
to s
upply
har
m
on
ic
c
urre
nt
I
sh
as:
)
(
).
(
).
(
1
1
s
G
s
G
s
G
I
I
s
G
out
i
n
v
c
Lh
sh
(1)
Wh
e
re:
G
c
(
s)
is
the
trans
fe
r
functi
on
of
t
he
tra
nd
it
io
nal
PI
c
on
t
ro
ll
er
,
G
inv
(
s)
is
the
trans
fer
funct
ion
o
f
the in
ver
te
r
a
nd
G
out
(
s
)
is t
he t
ran
s
fer
f
un
ct
io
n of t
he o
utput
ci
rcu
it
.
All
of
the
a
bo
ve
tra
ns
fe
r
f
unct
ion
s
are
dete
rm
inate
d
accor
ding
to
[
3].
T
he
ope
rati
ng
pri
nciple
of
HAPF
co
ntr
ol
ci
rcu
it
is
al
so
based
on
the
load
cu
rr
e
nt:
the
load
c
urrent
i
L
(
i
La
,
i
Lb
,
i
Lc
)
will
be
sepa
rated
the
har
m
on
ic
com
po
nen
t
of
the
load
c
urre
nt
as
i
Lha
,
i
Lhb
,
i
L
hc
.
The
se
com
po
ne
nts
will
be
cho
s
en
as
the
ref
e
re
nce
c
om
po
ne
nts
an
d
they
are
c
om
par
ed
to
the
com
pen
sat
e
cu
rr
e
nt
com
po
ne
nts
into
t
he
gri
d
i
c
of
HAPF
are
i
ca
,
i
cb
and
i
cc
.
Th
e
error
of
the
above
c
om
par
ison
will
be
ta
ken
t
hroug
h
th
e
PI
(Pr
opor
ti
on
al
-
In
te
gr
al
)
c
on
t
r
oller,
thr
ough
pu
lse
width
m
odulati
on
t
o
co
ntr
ol
ON
-
O
FF
switc
hes
of
the
inv
erter
t
o
ge
ner
at
e
com
pen
sat
e sig
nals
on the
gr
i
d.
The
pa
ram
et
ers
need
t
o
be
de
te
rm
ined
in
t
he
m
od
el
in
Figure
1
i
nclu
din
g:
passive
ci
r
cuit
C
F
-
C
1
-
L
1
,
outp
ut
filt
er
L
0
,
volt
age
bus
-
DC
of
inv
e
rter
U
d
c
,
par
a
m
et
ers
of
PI
c
on
t
ro
ll
er
are
K
p
and
K
i
.
Curr
ently
,
these p
a
ram
et
e
rs
are
us
ually
det
erm
ined
base
d
on expe
rienc
e and
loc
al
calc
ulati
on
s
, r
e
ga
rd
le
ss
of
the st
abili
ty
of
the
syst
em
.
T
her
e
fore,
this
pap
e
r
pro
po
s
es
a
ne
w
m
ul
ti
-
obj
ect
ive
opti
m
iz
at
ion
al
gori
thm
that
ha
s
consi
der
e
d
t
o
t
he
sta
bili
ty
o
f
t
he
syst
em
.
3.
NEW
MU
LT
I
-
OBJE
CTIVE
OPTI
MIZ
ATION
DESI
GN ALG
ORI
T
H
M
F
OR
H
YBRID ACTI
VE
POWER
FIL
TE
R
3.1.
C
on
s
tra
in
ts a
nd obj
ec
tive
functi
on
Wh
e
n desi
gn a
HAPF
syst
em
,
all
o
f
the
fo
ll
owin
g
c
onstrai
nt
s n
ee
d
to
b
e
c
on
si
der
e
d:
+ Syst
em
stabi
li
ty
co
ns
trai
nt:
Fr
om
(1), we
have the
cha
ract
erist
ic
eq
ua
ti
on of the
contr
ol
trans
fer
f
un
ct
io
n
as
foll
ows:
1
1
2
2
3
3
4
4
5
5
6
6
)
(
s
a
s
a
s
a
s
a
s
a
s
a
s
D
(2)
Accor
ding to
Rou
t
h’
s
sta
bili
ty
stand
a
rd
, foll
ow
i
ng con
diti
on
s m
us
t be
sat
isfie
d.
-
-
-
-
0
0
0
0
0
3
1
4
1
3
2
3
5
3
1
1
4
2
3
3
6
4
5
b
a
b
b
b
b
b
a
a
b
a
a
a
a
a
a
a
a
(3)
W
it
h:
5
4
6
3
1
5
a
a
a
a
b
a
-
;
3
2
4
1
3
3
a
a
a
a
b
a
-
;
1
3
5
3
2
1
b
a
a
b
b
b
-
;
31
4
3
ba
b
b
;
3
2
1
4
1
2
b
b
b
b
c
b
-
;
24
cb
;
6
1
1
1
1
0
1
0
()
i
F
s
s
a
T
T
L
C
C
L
L
L
L
L
L
L
;
5
1
1
0
1
1
1
1
0
1
1
0
(
)
(
)
i
s
F
s
F
i
F
s
F
F
s
a
T
T
R
L
C
C
R
L
C
C
T
C
C
L
L
L
L
C
C
C
C
L
L
;
4
0
1
1
0
1
1
1
0
1
1
1
(
)
(
)
)
i
F
s
F
i
s
F
s
F
F
i
n
v
i
a
T
T
L
C
L
C
L
C
L
C
T
R
L
C
C
R
L
C
C
C
C
L
K
K
T
;
3
0
1
1
0
1
1
1
()
i
F
s
i
F
s
F
i
n
v
F
a
T
T
C
R
T
L
C
L
C
L
C
L
C
K
K
C
L
C
;
2
i
i
F
s
i
i
n
v
F
a
T
T
T
C
R
T
K
K
C
;
1
i
i
n
v
F
a
T
K
K
C
;
0
.
0
1
;
0
.
0
0
0
0
1
;
1
0
;
1
i
i
n
v
T
m
s
T
K
K
Wh
e
re
K
inv
is i
nv
e
rter
gain
,
T
is t
i
m
e con
sta
nt
,
K
is
pr
opor
ti
on
al
c
oe
ff
ic
ie
nt
,
T
i
is i
nte
gr
al
const
ant.
-
Con
st
raints on
resona
nce con
diti
on
s i
n
PPF
s
: The
L
an
d
C
par
am
et
ers
in a b
ra
nc
h
m
us
t be r
eso
nated
at
a
certai
n
f
re
qu
e
nc
y.
1
n
n
L
C
(4)
-
Con
st
raint
of
pa
ssive
powe
r
f
il
te
r
values:
pa
ssive
powe
r
fil
te
r
values
m
us
t
be
po
sit
ive
a
nd
m
eet
syst
e
m
sta
bili
ty
an
d re
so
na
nce
con
diti
on
s
.
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.
9
, N
o.
6
,
Dece
m
ber
201
9
:
4507
-
4515
4510
m
a
x
m
a
x
m
a
x
0
(
,
,
)
(
,
,
)
i
i
i
R
L
C
R
L
C
(5)
The val
ues
of
R
max
,
L
max
an
d
C
max
are d
et
e
r
m
ined
un
der sta
ble con
diti
on
s
(2)
.
-
PPFs
m
us
t be
c
om
pen
sat
ed wi
th a m
axi
m
u
m
capacit
y b
ut no
t exceed
the
r
e
qu
i
red m
axi
m
u
m
lim
it
m
i
n
m
a
x
c
c
c
Q
Q
Q
(6)
-
Con
st
raint
of
DC
-
bu
s
volt
ag
e v
al
ue:
0
<
U
dc
<
U
dc
-
m
ax
(7)
-
Con
tr
oller
para
m
et
ers
con
str
ai
nts:
Con
tr
oller
pa
ram
et
ers
m
us
t
be
po
sit
ive
an
d
sat
isfy
syst
e
m
sta
bilit
y
conditi
ons (3
).
0 < K
p
< K
pmax
; 0 < K
i
< K
imax
(8)
Accor
ding to
[8], the
v
al
ue of
K
p
an
d
K
i
is t
oo sm
al
l or
to
o
l
arg
e
to
ca
us
e
s
yst
e
m
instabil
ity.
-
Objecti
ve
f
unct
ion
:
He
re
w
e
co
ns
ide
r
th
e
three
m
ai
n
obj
ect
ives
are:
m
i
n
s
T
H
D
i
,
m
i
n
s
T
H
D
u
an
d
m
a
x
c
Q
.
We tra
nsfer t
hree o
bj
ect
i
ves p
roblem
to
sing
l
e obj
ect
i
ve pr
oble
m
b
y t
he
f
ollow
in
g f
om
ula
r:
m
in(
1
/
)
s
s
c
F
T
H
D
i
T
H
D
u
Q
(9)
Accor
ding
to
I
EEE
Std.
519
-
1992
Re
com
m
end
e
d
P
racti
ces
and
Re
qu
i
re
m
ents
fo
r
harm
on
ic
con
trol
in
el
ect
rical
po
we
r
syst
em
s
[22],
the
n
t
otal
har
m
on
ic
distor
ti
on
(T
H
D)
of
c
urren
t
is
le
ss
than
or
e
qual
5%.
THD
of
sup
ply
vo
lt
age
at
po
i
nt
of
com
m
on
couplin
g
is
le
ss
than
or
eq
ual
5%
and
c
Q
is
determ
ined
base
d
on the l
oad.
3.2.
So
ci
al sp
ider
algorithm
and
its appli
ca
tio
n
t
o
h
yb
ri
d
ac
tive powe
r
filt
er
desig
n
In
o
r
de
r
to d
esi
gn
a
Hy
br
id Act
ive
Power
Fil
te
r
syst
e
m
,
a
ne
w
m
ulti
-
ob
j
ec
ti
ve
optim
iz
ati
on
al
gorithm
base
d
on
S
ocia
l
Sp
ide
r
Al
gori
thm
(S
SA
)
is p
rop
os
ed
.
I
n
S
S
A
[
21]
,
the
s
ol
ution
of
the SS
A
is
sim
ulate
d
by
an
arti
fici
al
sp
ider
'
s
po
sit
ion
on
a
sp
ide
r
we
b.
S
pid
e
rs
in
a
popula
ti
on
(po
p)
m
ov
e
on
t
he
w
eb
an
d
s
hare
lo
cat
io
n
inf
or
m
at
ion
throu
gh
it
s
vi
brat
ion
.
Dep
e
nd
ing
on
t
he
vi
br
at
io
n
recei
ve
d
f
ro
m
the
sp
ide
rs,
it
will
direct
the
sp
i
der
s
t
o
the
opti
m
a
l
posit
ion
.
Ar
ti
fici
al
sp
i
ders
are
a
m
ajo
r
facto
r
in
the
SS
A
al
gorithm
.
Each
sp
ide
r
po
s
sesses
a
l
oc
at
ion
on
t
he
sp
ide
r
we
b
m
ap
an
d
t
he
m
os
t
appropr
ia
te
va
lu
e
of
the
po
sit
ion
is
locat
e
d
on
the
sp
ide
r
m
a
p.
Eac
h
sp
i
der
occupies
the
best
sta
te
,
includi
ng
:
c
urren
t
po
sit
io
n,
c
urr
ent
valu
e
,
vi
brat
ion
intensit
y,
pitch
,
pr
e
vious
act
i
on
a
nd
the
siz
e
of
the
s
pid
e
r.
All
of
the
se
ch
aracte
risti
cs
will
direct
the
sp
i
der
to
fin
d
the
optim
al
so
luti
on.
I
n
each
it
erati
on,
al
l
sp
ider
s
on
t
he
web
m
ov
e
to
a
ne
w
posit
ion
an
d
e
valuat
e
thei
r
fitness
value
s.
The
al
gorit
hm
first
cal
culat
es
the
fitnes
s
val
ues
of
al
l
the
a
rtific
ia
l
sp
i
der
s
on
di
ff
e
ren
t
po
sit
io
ns
on th
e w
e
b.
The
n
t
he
se sp
i
der
s
g
e
ne
rate vi
br
at
io
ns at
their
posit
ion
s
usi
ng the
foll
ow
i
ng equat
ion
.
m
a
x
m
i
n
1
/
(
(
)
)
f
o
r
m
a
x
i
m
i
z
a
t
i
o
n
(
,
,
)
1
/
(
(
)
)
f
o
r
m
i
n
i
m
i
z
a
t
i
o
n
-
-
s
ss
s
C
f
P
I
P
P
t
f
P
C
(10)
Wh
e
re
C
m
ax
is
a
c
o
nfi
de
ntly
la
rg
e
c
onsta
nt
sel
ect
ed
s
uch
that
al
l
po
s
sible
fitnes
s
val
ues
of
the
m
axi
m
iz
ati
on
pro
blem
is
sm
a
ll
er
than
C
m
a
x
,
C
m
in
is
a
c
onfide
ntly
s
m
a
ll
con
sta
nt
su
c
h
that
al
l
po
ss
ibl
e
fitness
value
s
of
the
m
inim
iz
at
ion
pr
ob
lem
is
la
rg
er
than
C
m
in
,
(
,
,
)
ss
I
P
P
t
is
i
ntensity
of
th
e
vib
rati
on
gen
e
rated
b
y s
pid
e
r
s
at
t
he
s
ource
posit
ion
,
f(
P
s
)
is fit
ness of t
he p
os
it
ion
of sp
i
der.
Af
te
r
al
l
the
vibrat
ion
s
a
re
gen
e
rated
,
the
al
go
rit
hm
si
m
ula
te
s
the
prop
a
gatio
n
pr
oc
ess
of
t
hese
vibrat
ion
s
usi
ng the
f
ollo
wing
equati
on.
m
a
x
(
,
)
(
,
,
)
(
,
,
)
.
e
x
p
.
-
ab
a
b
a
a
a
D
P
P
I
P
P
t
I
P
P
t
Dr
(11)
Wh
e
re
(
,
)
ab
D
P
P
is
distance
bet
ween
sp
ide
r
a
a
nd
b
,
m
a
x
D
is
m
axi
m
um
distance
be
tween
tw
o
po
i
nts
in
the searc
h spac
e an
d
(0
,
1
)
a
r
.
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
A n
ew
desi
gn al
go
rit
hm f
or
hy
br
id
a
ct
iv
e
powe
r fi
lt
er
(
Cha
u
Mi
nh T
hu
ye
n
)
4
511
In
t
his
pro
cess
,
e
ach
sp
i
der
s
will
receive
(
pop
–
1)
dif
fere
nt
vi
br
at
io
ns
gen
e
rated
by
ot
her
s
pid
e
rs.
Each
sp
i
der
s
will
sel
ect
th
e
strongest
vi
br
at
io
n
v
best
from
(s
-
1)
sp
iders
an
d
com
par
e
it
s
intensit
y
with
the
inten
sit
y
of
the
ta
r
get
vi
brat
ion
v
tar
sto
re
d
in
it
s
m
e
m
or
y.
s
will
sto
re
v
best
as
v
tar
if
th
e
intensit
y
of
v
best
is
la
rg
er
,
ot
herwise
the
or
i
gin
al
v
tar
is
retai
ne
d.
T
he
al
gorithm
then
m
ani
pu
la
te
s
s
to
pe
rfor
m
a
ran
do
m
walk
towa
rd
s
v
tar
.
T
he
al
go
rithm
rep
eat
s
this
pro
cess
f
or
al
l
the
sp
id
ers
i
n
pop
.
T
o
a
void
S
SA
get
ti
ng
stu
ck
in
a
local
op
ti
m
u
m
,
each
sp
ider
in
pop
,
rig
ht
after
the
rando
m
walk
ste
p,
has
a
s
m
all
pr
obab
il
ity
to
decide
no
t
t
o
fo
ll
ow it
s
pr
es
ent tar
get and
jum
p
away f
rom
it
s cu
rr
e
nt
posit
ion.
Flow
c
ha
rt
of
m
ul
ti
-
obj
ect
ive
op
ti
m
iz
ation
al
gorithm
based
on
S
SA
f
or
HAPF
s
how
n
in
Fig
ur
e
3
(appe
nd
i
x)
.
Be
gin
,
we
ente
r
t
he
uppe
r
a
nd
lowe
r
li
m
i
ts
of
the
pa
ram
et
ers
C
F
,
C
1
,
L
1
,
L
0
,
U
dc
,
K
p
an
d
K
i
,
the
n
the
SSA
al
gori
thm
beg
ins
by
creati
ng
the
or
iginal
dataset
of
C
F
,
C
1
,
L
1
,
L
0
,
U
dc
,
K
p
and
K
i
.
This
dataset
wil
l
be
chec
ke
d
f
or
sta
bili
ty
of
th
e
syst
e
m
,
if
it
do
e
s
not
m
eet
the
sta
bili
ty
con
diti
on,
th
e
syst
e
m
will
retur
n
to
create
a
ne
w
da
ta
set
,
if
it
sat
isfie
s
the
sta
bl
e
conditi
on
of
t
he
syst
em
,
then
these
pa
ram
e
te
rs
will
be
est
i
m
at
e
d
by
the
fi
tness
functi
on
c
orre
sp
on
ding
to
th
e
curre
nt
posit
ion
of
the
s
pi
d
ers
on
t
he
s
pi
der
web,
a
nd
then
the
s
pid
e
rs
will
create
vib
rati
on.
T
he
sp
i
ders
will
pr
op
a
gat
e
an
d
sel
ect
vi
br
at
io
n
.
I
f
this
vibrat
ion
is
str
onge
r
than
the
pr
e
vious
ta
r
get
the
s
pid
e
r
will
us
e
the
sel
ect
ed
be
st
vib
rati
on
as
the
ne
w
ta
rg
et
,
oth
er
wise
at
te
nu
at
e
the
pr
e
vi
ou
s
t
arg
et
vi
br
at
io
n.
Ba
sed
on
the
vib
rati
on
fro
m
the
sp
iders
it
will
direct
t
he
sp
i
der
s
t
o
a
bette
r
po
sit
io
n.
T
his
new
posit
ion
will
be
com
par
ed
t
o
the
obj
ect
ive
f
unct
io
n,
if
sat
isfie
d,
then
st
op
a
nd
pri
n
t
the b
e
st res
ult
set
, o
t
herwise r
et
urn
to
the c
re
at
e an or
i
gin
al
dataset
.
Figure
3
.
Flo
w
char
t
of n
e
w
m
ulti
-
obj
ect
ive
optim
iz
at
ion
d
es
ign
al
gorithm
f
or HAPF
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.
9
, N
o.
6
,
Dece
m
ber
201
9
:
4507
-
4515
4512
4.
SIMULATI
O
N RESULTS
AND DIS
C
USSION
To
com
par
e
th
e
eff
ect
ive
ness
betwee
n
the
pro
posed
desig
n
al
g
or
it
hm
and
tradit
ion
al
de
s
ign
.
Let
us
consi
der
a
H
A
PF
m
od
el
as
s
how
n
in
Fi
gur
e
1.
T
hr
ee
-
ph
a
se
sou
rce
vo
lt
a
ge
is
380V
-
50
Hz.
N
on
li
near
load
is
bu
il
t
by
three
-
ph
a
se
unco
ntr
ol
le
d
br
id
ge
rect
ifie
r
with
loa
d
R
L
=
3Ω,
L
L
=
0.3m
H.
THD
of
the
loa
d
curr
ent
is
sh
ow
n
in
Fig
ure
4.
T
he
cal
culat
ion
of
H
A
PF
pa
ram
et
ers
is
do
ne
in
t
w
o
ways:
the
fir
st
way
is
to
ca
lc
ulate
accor
ding
t
o
t
he
tra
diti
on
al
desig
n
us
in
g
P
SO
al
gorithm
and
the
seco
nd
way
is
to
ca
lc
ulate
the
pro
po
s
ed
al
gorithm
.
Figure
4. TH
D
of
the
loa
d
c
urren
t
W
it
h
the
tra
diti
on
al
de
sig
n
m
et
hod
usi
ng
th
e
PSO
al
gorith
m
,
the
par
am
et
ers
of
the
HAPF
syst
em
after
100
it
er
at
ion
s
a
re
s
um
m
arized
in
Table
1.
Acc
ordin
gly,
the
de
fine
d
par
a
m
et
ers
on
ly
i
nclu
de
the
par
am
et
ers
of
the
powe
r
ci
rcu
it
par
t
a
re
C
F
,
C
1
,
L
1
and
L
0
.
I
n
ad
di
ti
on
,
the
c
ontr
oller
pa
ram
et
e
rs
are
rand
om
l
y
sel
e
ct
ed
with
the
best
pa
ram
et
er
s
of
K
p
=
30
and
K
i
=
0.1
,
switc
hing
f
re
quency
f
=
10kHz
an
d
DC
-
bu
s
volt
ag
e is 6
00V
.
Table
1.
H
AP
F
p
a
ram
et
e
rs
by trad
it
ion
al
desi
gn m
et
ho
d
C
F
(µF)
C
1
(µF)
L
1
(
m
H
)
L
0
(
m
H
)
U
dc
(V)
K
p
K
i
1
1
7
.27
4
7
9
.83
2
1
.83
1
.83
600
30
0
.1
Fr
om
the
pa
ra
m
et
ers
in
Tabl
e
1,
we
ob
ta
i
n
the
wa
vefo
rm
s
of
t
he
HAPF
syst
e
m
sh
own
in
Fig
ur
e
5.
Fr
om
Fig
ur
e
5,
we
see
t
ha
t:
THD
of
t
he
sup
ply
cu
rr
e
nt
i
s
decr
eas
es
from
28.33%
to
2.3
2%,
T
HD
of
the
sou
rce
volt
age
u
s
is
2.04%,
wh
il
e
th
e
reacti
ve
po
wer
dec
rease
d
t
o
740var
from
30
05
var,
that
is
,
the
capaci
ty
com
pen
sat
ed
by
PPFs
is
2265va
r.
T
H
D
of
the
supp
ly
cu
r
ren
t
i
s
show
n
in
Fig
ur
e
6.
T
HD
of
t
he
sou
rce
vol
ta
ge
u
s
is
sho
wn
in
Fi
gure
7.
With
t
he
pro
posed
al
gori
thm
,
the
best
par
am
et
ers
of
both
the po
wer ci
rc
uit pa
rt and c
ontr
ol circuit
pa
rt
after
50 itera
ti
on
s a
re sh
own
in
Ta
ble 2.
Figure
5.
W
a
ve
form
s o
f HA
P
F w
it
h t
ra
diti
on
al
d
e
sig
n usin
g
P
SO
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
A n
ew
desi
gn al
go
rit
hm f
or
hy
br
id
a
ct
iv
e
powe
r fi
lt
er
(
Cha
u
Mi
nh T
hu
ye
n
)
4513
Figure
6. TH
D
of the
sup
ply current
i
s
Figure
7. TH
D
of the
sou
rce
vo
lt
age
u
s
Table
2.
H
AP
F
p
a
ram
et
ers
with the p
rop
os
e
d
m
ulti
-
obj
ect
ive
op
ti
m
iz
ation
desig
n
al
gorithm
C
F
(µF)
C
1
(µF)
L
1
(
m
H
)
L
0
(
m
H
)
U
dc
(V)
K
p
K
i
1
1
9
.53
3
3
9
.978
3
0
.84
9
0
.19
3
6
6
2
.061
3
2
.11
8
0
.69
9
Fr
om
the
pa
ra
m
et
ers
in
Tabl
e
2,
we
ob
ta
i
n
the
wa
vefo
rm
s
of
t
he
HAPF
syst
e
m
sh
own
in
Fig
ur
e
8.
Fr
om
Figure
8,
we
ca
n
see
that:
THD
of
the
sup
ply
curr
ent
i
s
decr
eas
e
s
from
28
.33
%
to
1.7
2%,
THD
of
the
so
urce
volt
age
u
s
is
1.
36%
wh
il
e
the
reacti
ve
power
d
ec
r
eased
to
700Var
f
ro
m
3005Va
r
,
so
the
c
om
pen
sat
ion
po
wer
by
PPFs
is
2305Va
r,
T
HD
of
the
su
pply
cu
rr
e
nt
i
s
in
ste
ady
-
sta
te
is
sh
own
i
n
Figure
9, TH
D
of the
sou
rce
vo
lt
age
u
s
in st
eady
-
sta
te
is s
how
n
i
n
Fi
gure
10.
Figure
8.
The
wav
e
f
or
m
s o
f HA
PF wit
h
t
he
prop
os
ed
alg
ori
thm
Figure
9. TH
D
of the
sup
ply current
i
s
Figure
10.T
H
D of
t
he
s
ource
volt
age
u
s
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.
9
, N
o.
6
,
Dece
m
ber
201
9
:
4507
-
4515
4514
Fr
om
the
above
analy
sis,
we
can
see
that:
the
propose
d
a
lgorit
hm
is
m
or
e
eff
ect
ive
tha
n
the
tra
diti
on
al
desig
n
i
n
re
du
ci
ng
t
he
total
ha
rm
on
ic
disto
rtion
a
nd
the
m
a
xim
u
m
reacti
ve
powe
r
c
om
pen
sat
e
into the
grid
in
steady
-
sta
te
a
nd m
eet
IEEE St
d.
519
-
1992.
5.
CONCL
US
I
O
N
The
pa
pe
r
ha
s
pro
vid
e
d
a
new
m
ulti
-
obj
ect
ive
op
ti
m
iz
a
ti
on
design
al
gorithm
fo
r
H
AP
F
.
This
al
go
rithm
can
determ
ine
the
par
am
et
ers
of
bot
h
the
powe
r
ci
rc
uit
par
t
a
nd
the
c
on
t
ro
l
ci
rc
uit
par
t
of
HAPF.
T
he
ac
hieve
d
res
ults
are
m
ulti
-
pu
r
pose,
su
c
h
as:
m
ini
m
iz
e
the
total
ha
rm
on
ic
distor
ti
on
of
th
e
sup
ply
current
a
nd
source
volt
age,
m
axi
m
iz
e
the
reacti
ve
po
wer
com
pen
sat
ion
into
the
gri
d
and
sat
isfyi
ng
m
an
y
const
raint
c
onditi
on
s
,
s
uch
a
s:
syst
e
m
sta
bi
li
ty
,
reso
na
nce
co
nd
it
io
ns
,
pa
ssive
po
wer
fi
lt
er
values
,
D
C
-
bu
s
vo
lt
age
value
and
c
on
t
ro
ll
er
par
am
et
ers
con
strai
nts
.
This
researc
h
can
be
app
l
ie
d
to
de
sign
f
or
al
l
diff
ere
nt
ty
pes
of
HAPF
.
ACKN
OWLE
DGE
MENTS
This
w
ork
is
par
ti
al
su
pport
ed
by
the
sci
entifi
c
research
s
ubj
ect
N
o.
50
48/Q
Đ
-
BC
T
of
the
Mi
nistry
of
Ind
us
try
a
nd
T
ra
de
of
the
so
ci
al
ist
repub
li
c
of
Viet
Nn
am
with
the
ti
tl
e:
"R
esearch,
desig
n
and
m
anu
fact
ur
e
thr
ee
-
ph
a
s
e
act
ive
c
ompen
sat
ion
m
od
el
wit
h
5kW
ca
pacit
y
f
or
ha
rm
on
ic
filt
ering
of no
nlinear l
oa
ds
"
.
REFERE
NCE
S
[1]
Am
ol
Anandra
o
Pati
l,
et
al
.
,
"
Harm
onic
m
et
ig
at
ion
using
passive
filter
,
"
Inte
r
nati
onal
Journa
l
of
Engi
n
ee
rin
g
Tr
ends
and
Tech
nology
,
vol
.
45
,
pp.
83
-
86
,
2009
.
[2]
C.
Madth
ara
d
an
d
S.
Prem
rude
ep
ree
ch
ac
h
arn
,
"
A
ct
iv
e
power
f
il
t
e
r
for
three
-
phase
four
-
wire
el
e
ct
r
ic
s
y
s
te
m
s
using
neur
al ne
tworks,
"
El
e
ct.
Powe
r S
yst.
Re
s.
,
vo
l. 60
,
pp
.
1
79
-
192
,
2
002.
[3]
Chau
Minh
Thuy
en
,
"
A
new
appr
oac
h
in
desig
n
for
hy
br
id
act
ive
power
fil
t
er
,"
ICIC
Ex
press
le
tters
,
vol.
12,
pp.
897
-
904
,
20
18.
[4]
C.
Madth
ara
d
an
d
S.
Prem
rude
ep
ree
ch
ac
h
arn
,
"
A
ct
iv
e
power
f
il
t
e
r
for
three
-
phase
four
-
wire
el
e
ct
r
ic
s
y
s
te
m
s
using
neur
al ne
tworks,
"
El
e
ct.
Powe
r S
yst.
Re
s.
,
vo
l. 60
,
pp
.
179
-
192
,
2
002.
[5]
M.
Chau,
et
al
.
,
"
Online
cont
rol
m
et
hod
with
ti
m
e
-
dela
y
compensat
ion
for
h
y
brid
ac
t
ive
power
filt
er
with
inj
e
ction
ci
rcu
it
,
"
IET
Po
wer
Elec
troni
cs,
vol.
5
,
pp
.
1472
-
1482,
2012
.
[6]
Minh
Thu
y
en
Chau,
"
Adapti
v
e
cur
ren
t
con
tr
ol
m
et
hod
for
h
y
brid
a
ct
iv
e
power
fil
te
r
,"
Jo
urnal
of
El
ec
tri
cal
Engi
ne
ering
,
vol
.
67
,
pp
.
343
-
35
0,
2016
.
[7]
An
Luo,
et
al
.
,
"
Deve
lopment
of
h
y
brid
ac
t
ive
po
wer
fil
t
er
base
d
on
the
ada
p
ti
v
e
f
uzzy
divi
d
ing
fr
eque
nc
y
-
cont
rol
m
et
hod,
"
IE
EE
Tr
ansacti
ons on Power
Del
iv
er
y,
vol.
24
,
pp
.
424
-
432,
2009
.
[8]
Z.
Shuai
,
A.
Lu
o,
C.
Tu
,
D.
Liu
,
"
New
cont
rol
m
et
hod
of
inj
e
ct
ion
-
t
y
p
e
h
y
bri
d
ac
ti
v
e
power
f
il
te
r
,
"
IET
Power
El
e
ct
ronics
,
vol
.
4,
pp.
1051
-
105
7,
2011
.
[9]
Chau
Minh
Thu
y
en
,
"
Im
prove
d
p
-
q
Harm
onic
Detect
ion
Me
thod
for
H
y
bri
d
Act
ive
Pow
er
Filt
er
,"
Inte
rnat
ional
Journal
of
Elec
t
rical
and
Computer
Eng
ine
ering
(
IJE
CE)
,
vol. 8
,
no.
5
,
pp
.
2910
-
2919,
2018
.
[10]
H.
Akagi,
e
t
al
.
, "
Gene
ral
i
ze
d
theor
y
of
inst
antane
ous
rea
ctive
pow
er
and
it
s a
pplic
at
ion
,"
El
e
ct
ri
ca
l
Engi
ne
ering
of
Japan
,
vo
l.
103
,
pp.
483
-
490
,
19
83.
[11]
Y.
K.
La
tha
,
et
al.
,
"
Control
strat
eg
y
for
t
hre
e
phase
shunt
ac
t
ive
powe
r
fil
te
r
with
m
ini
m
um
cur
ren
t
m
ea
surem
ent
s,"
Inte
rnational
Jo
urnal
of
Elec
trical
and
Compute
r
Engi
nee
ring
(
IJE
CE)
,
vol
.
1
,
n
o.
1
,
pp
.
31
-
42
,
Oct
2011.
[12]
An
Luo,
et
a
l.
,
"
Design
conside
r
at
ions
for
m
ai
nt
ai
nin
g
DC
-
side
volt
ag
e
of
h
y
bri
d
ac
t
ive
power
f
il
te
r
wi
th
injec
tion
ci
rcu
it
,
"
IEEE
Tr
ansacti
ons on P
ower
Elec
troni
cs
,
vol
.
24
,
pp
.
75
-
84,
2009
.
[13]
C.
C.
M.
Mour
a,
et
al.
,
"
De
te
r
m
i
nat
ion
of
the
R
-
L
-
C
p
ara
m
eters
of
a
passiv
e
har
m
onic
fi
lt
e
r
using
gen
eti
c
al
gorit
hm
,
"
in
10
th
Int
ernati
onal
conf
ere
n
ce on
h
arm
onic
s and
qu
ali
ty of
power,
2
002,
pp
.
495
-
50
0.
[14]
Suresh
ku
m
ar,
R
amesh
red
d
y
an
d
Archa
na,
"
Th
e
appl
icati
on
of
p
so
to
hy
b
rid
acti
ve
po
wer
fil
t
er
design
for
3
phase
4
-
wire
s
y
stem
wi
th
balanc
ed
&
u
nbal
an
ce
d
lo
ads,
"
Inte
rnational
j
ournal
of
advanc
es
in
engi
ne
eri
ng
&
te
chnol
ogy
,
vol.
2
,
pp
.
32
-
42
,
2012
.
[15]
N.
S.
Hasan,
et a
l
.
,
"
Harm
onic
Suppress
ion
of
Shunt H
y
b
r
id
Filt
er using L
QR
-
PS
O
base
d,
"
I
n
te
rnat
ional
Journal
o
f
El
e
ct
rica
l
and
C
omputer
Engi
n
e
ering
(
IJE
CE)
,
v
ol.
7
,
no
.
2
,
pp
.
8
69
-
876
,
2017
.
[16]
Shengqing
Li
,
et
al.
,
"
Mult
i
-
objecti
v
e
opti
m
al
de
sign
for
passive
power
fil
t
ers
in
h
y
brid
power
f
ilter
s
y
st
em
base
d
on
m
ult
i
-
isla
nd
par
ticle
sw
arm
opti
m
iz
ation
,
"
i
n
7th
Int
ernati
o
nal
Confe
ren
ce
on
Powe
r
E
le
c
tronic
s
and
Mot
io
n
Control
,
pp
.
285
9
-
2863,
2012
.
[17]
Suresh
kum
ar
and
K.
Ramesh,
"
The
application
PS
O
hy
brid
a
ctive
power
filter
design
for
3phase
4
-
wire
s
y
st
em
with
var
ia
bl
e load,
"
Int
ernati
onal
jour
nal
o
f engi
n
ee
ring
inv
en
ti
on
s
,
vol. 1, pp. 39
-
46,
2012
.
[18]
Nien
-
Che
Yang
and
Minh
-
Du
y
Le
,
"
Mult
i
-
objec
ti
ve
ba
t
al
gor
it
h
m
with
ti
m
e
-
var
y
ing
ine
rt
ia
wei
ghts
for
opti
m
al
desig
n
of
p
assive
power
filters
s
et
,
"
IET
Gen
erat
ion,
Tr
ansm
ission
&
Distributi
on
,
vol
.
9
,
pp
.
644
-
654,
2015
.
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
A n
ew
desi
gn al
go
rit
hm f
or
hy
br
id
a
ct
iv
e
powe
r fi
lt
er
(
Cha
u
Mi
nh T
hu
ye
n
)
4515
[19]
Rrac
h
id
Dehini a
nd
Slim
ane
Sef
ia
ne
,
"
Pow
er
qu
al
ity
and
cost
improvem
ent
b
y
p
a
ss
ive
power
fi
lt
e
rs
s
y
nthe
sis
usin
g
ant
col
on
y
al
gor
i
thm,"
Journal
o
f
the
ore
ti
ca
l
and
appli
ed
inf
orm
at
ion
t
ec
hnolog
y
,
vol.
1
,
pp
.
70
-
79
,
2012
.
[20]
Ahm
ed
Fahee
m
Zoba
a,
"
Optim
al
m
ult
i
-
obj
ec
t
i
ve
design
of
h
y
brid
ac
t
ive
power
fil
t
ers
conside
ring
a
distorte
d
envi
ronm
ent
,
"
I
EE
E
transacti
on
s on
industrial
elec
troni
cs
,
vo
l. 6
1,
pp
.
107
-
114
,
2014.
[21]
Jam
es
J.Q.
Yu,
et
al
.
,
"
A
soci
al
spider
al
gor
it
h
m
for
globa
l
op
ti
m
iz
ation
,"
Te
c
hnic
al
Re
port
N
o.
TR
-
2003
-
004,
Dept.
o
f El
e
ct
ri
c
al
&
El
ec
tronic
Engi
ne
ering, t
he
Unive
rs
ity
o
f
H
ong
Kong
,
pp
.
1
-
16
,
2013
.
[22]
IEE
E
Std.
519
-
1
992.
IEE
E
R
ec
o
mm
ende
d
Pract
i
ce
s
and
Re
quire
ments
for
Harm
onic
Control
in
El
e
ct
rica
l
Power
Syste
ms
,
1993
.
BIOGR
AP
H
I
ES
OF
A
UTH
ORS
Ch
au
Mi
nh
Th
uy
en
was
born
i
n
Binh
Dinh,
Vi
et
nam,
on
June
6,
1977.
He
r
ec
e
ive
d
his
B
.
S.
an
d
M.S.
from
the
Da
Nang
Univer
sit
y
of
Technol
og
y
,
Da
Nang,
Viet
Nam
,
and
the
Univer
sit
y
o
f
Te
chn
ic
a
l
Edu
c
at
ion
Ho
Chi
Minh
Cit
y
,
Ho
Chi
Minh
Cit
y
,
Vi
et
Nam
,
i
n
2001
and
20
05,
respe
ctively
,
an
d
Ph.D.
from
H
unan
Univer
sit
y
,
Changsha,
Chi
na,
in
2012.
,
Si
nce
2004,
he
h
as
bee
n
a
Lectu
r
er
at
Industri
al
U
nive
rsit
y
of
Ho
Chi
Minh
Cit
y
,
Viet
Nam
.
His
cur
ren
t
rese
ar
ch
int
er
ests
include
el
e
ct
ri
c
power
sa
vings,
re
active p
ower
compensat
i
on,
and
a
ct
iv
e
p
ower
filters.
Evaluation Warning : The document was created with Spire.PDF for Python.