Internati
o
nal
Journal of P
o
wer Elect
roni
cs an
d
Drive
S
y
ste
m
(I
JPE
D
S)
Vol
.
6
,
No
. 2,
J
une
2
0
1
5
,
pp
. 18
5~
19
5
I
S
SN
: 208
8-8
6
9
4
1
85
Jo
urn
a
l
h
o
me
pa
ge
: h
ttp
://iaesjo
u
r
na
l.com/
o
n
lin
e/ind
e
x.ph
p
/
IJPEDS
Advanced Techniques in Harmon
ic Suppression via Active
Power F
ilter: A Review
E
khl
as Mh
aw
i
,
Ha
md
an D
a
ni
yal
,
M
o
h
d
Herw
an Sul
ai
man
S
u
s
t
ainable
En
er
g
y
& P
o
wer
El
e
c
troni
cs
Res
e
arc
h
Cl
uster
(SuPER), FKEE, Univ
ersiti
Mala
ysi
a
P
a
hang
Article Info
A
B
STRAC
T
Article histo
r
y:
Received
Ja
n 31, 2015
Rev
i
sed
Ap
r
27
, 20
15
Accepted
May 10, 2015
This
paper int
e
n
d
s
to pres
ent the
recent
dev
e
lop
m
ent of artific
ial
intell
igen
ce
(AI) appli
cat
ions in a
c
tiv
e power
filte
r
(APF). As a result
of th
e de
velopm
en
t
in power electro
n
ic technolog
y
,
(APF)
continues to attract ample attention
.
Com
p
ared with the tradi
tion
a
l
react
iv
e LC fi
lter
,
ac
tive po
wer filter
is
cons
idered to be
m
o
re effectiv
e i
n
com
p
ens
a
ting harm
onic curren
t
genera
te
d
b
y
nonlinear lo
ads. APF, can
correct
the pow
er quality
and
improve the
reli
abil
it
y
and
stabili
t
y
on po
wer util
ity
.
A
brief
explan
atio
n of some
important ar
eas in AI and a comprehens
ive survey
of th
e liter
a
tur
e
along
th
e
m
a
in ca
tegori
e
s of AI is pr
esen
ted to
introdu
ce
the r
ead
ers int
o
the wid
e
-
ranging topics
t
h
at
AI encom
p
a
s
s
e
s
.
P
l
en
ty
of
r
e
lev
a
nt
literatures have b
een
selected in th
e
review, mostly
empha
sized on
better accuracy
, robustness,
effic
i
enc
y
,
s
t
abi
l
i
t
y
and tr
ack
ing
abili
t
y
of
th
e s
y
s
t
em
.
Keyword:
Activ
e p
o
wer
filter
Artificial In
tellig
en
ce
Harm
oni
c di
st
ort
i
o
n
No
n-l
i
near l
o
a
d
PW
M
Copyright ©
201
5 Institut
e
o
f
Ad
vanced
Engin
eer
ing and S
c
i
e
nce.
All rights re
se
rve
d
.
Co
rresp
ond
i
ng
Autho
r
:
Ekhlas Mha
w
i
Sustaina
ble E
n
ergy
& Power
Electr
onics
Re
search Cluster
(SuPER),
Faku
lti Kejuru
t
e
raan Elek
tri
k
& Elek
tron
ik,
Un
i
v
ersity Malaysia Pah
a
n
g
,
Peka
n, Paha
n
g
,
M
a
l
a
y
s
i
a
.
Em
a
il: ek
h
l
asm
h
awi@g
m
ai
l
.
co
m
1.
INTRODUCTION
Th
ere are m
a
n
y
stu
d
i
es about h
a
rm
o
n
i
c d
i
sto
r
tion
w
ith
tech
n
i
q
u
e
s to
i
m
p
r
o
v
e
p
o
w
e
r q
u
a
lity an
d
com
p
ensat
e
di
st
ort
e
d si
g
n
al
. Usual
l
y
, w
h
e
n
a passi
ve LC
(I
nd
uct
o
r a
nd C
a
paci
t
o
r
)
p
o
w
er
fi
lter isconnect
ed in
p
a
rallel with
t
h
e lo
ad
,
o
f
p
a
rallel o
r
series reson
a
n
ces b
ecau
s
e of wh
ich
th
e p
a
ssi
v
e
filter cann
o
t
p
r
ovid
e
a
co
m
p
lete so
lu
tio
n
it is
u
s
ed
t
o
elim
in
ate cu
rren
t
h
a
rm
o
n
i
cs. Th
is co
m
p
ensatio
n
equ
i
p
m
en
t h
a
s so
m
e
d
e
fect
s
main
ly related
to
th
e ap
p
e
aran
ce [1
]. Fo
r elimin
ati
ng ha
rm
oni
c
p
o
l
l
u
t
i
on
i
n
p
o
w
e
r sy
st
e
m
s, t
h
e act
i
v
e po
we
r
filter (APF) is
a v
e
ry su
itab
l
e to
o
l
. APF
h
a
s
to
resp
ond
in
st
an
tan
e
o
u
sly and
wo
rk
with
h
i
g
h
con
t
ro
l
p
r
ecisio
n
i
n
cur
r
e
n
t
t
r
acki
n
g, si
nce t
h
e
l
o
ad ha
rm
oni
cs
m
a
y
be very
com
p
l
i
c
at
ed and c
h
a
nge
ra
nd
om
l
y
and qui
ckl
y
.
M
a
ny
adva
nce
d
co
nt
rol
an
d s
i
gnal
-
pr
ocessi
n
g
t
echni
que
s h
a
ve bee
n
ap
pl
i
e
d, suc
h
as p
u
l
s
e wi
dt
h m
odu
l
a
t
i
on
(PW
M
),
h
y
steresis b
a
n
d
curren
t con
t
ro
l
(HBCC), slid
in
g
-
m
o
d
e
co
n
t
ro
l, fu
zzy-l
o
g
i
c con
t
ro
l, n
e
ural-network
theory, a
n
d adaptive si
gnal
processi
ng
and etc. [2]. Fi
gure
1 s
h
ows t
h
e
basic
pri
n
c
i
ple of a
n
APF. T
h
e
wav
e
fo
rm
s o
f
th
e APF i
n
th
e situ
atio
n
wh
ere th
e lo
ad
is
as
sum
e
d to be
a
three-phas
e
d
i
o
d
e
rectifier with
an
in
du
ctiv
e lo
ad
is sho
w
n
in Fi
g
u
re
2
.
Th
e
APF co
n
t
ro
lled
t
o
m
a
k
e
th
e so
urce curren
t
is i
n
th
e
sinu
so
i
d
, wh
ich
in
j
ects th
e co
mp
ensatin
g
cu
rren
t
i
c
into the s
o
urce t
o
cancel t
h
e ha
rm
oni
cs containe
d in t
h
e
load c
u
rrent
il
[3]
.
Th
e targ
et
o
f
th
e activ
e filter is to
co
m
p
en
sate d
i
st
o
r
ted cu
rren
t drawn
b
y
th
e
n
o
n
-
l
i
n
ear load
s from
th
e
source, s
o
that only the fu
nda
m
ental frequency co
m
pone
nts
rem
a
in in
the source curre
nt. The active filter and
its cu
rren
t con
t
ro
l m
u
st precisely track
th
e sud
d
e
n
slope c
h
a
nge
in t
h
e
refe
rence c
u
rrent
[4].
The l
o
ad
c
u
r
r
e
nt
c
onsi
s
t
s
o
f
f
u
n
d
am
ent
a
l
com
pone
nt
i
1
and
h
a
rm
oni
c com
pone
nt
i
h
.
To
g
e
t
th
e
refe
rence
harm
onic signal, fi
rst the load current is m
easured. T
h
e funda
m
ental curre
nt is extracted from
the
m
easured sy
st
em
l
o
ad curre
nt
by
usi
n
g t
h
e ban
d
pass fi
l
t
er, with appropriate cut-
of
f f
r
eq
ue
nci
e
s. Th
e l
o
ad
cur
r
ent
i
s
co
m
p
ared t
o
t
h
e
fu
ndam
e
nt
al
com
pone
nt
by
usi
n
g a com
p
arat
or;
t
h
e e
r
r
o
r i
s
t
h
e re
f
e
rence
Evaluation Warning : The document was created with Spire.PDF for Python.
I
S
SN
:
2
088
-86
94
I
J
PED
S
Vo
l. 6,
No
.
2,
Ju
ne 20
15
:
185
–
1
95
18
6
harm
onics si
gnals
[5]. The
AI tech
niques
usi
n
g in electric powe
r
has
receive
d large conside
r
ation
from
rev
i
ewer i
n
t
h
e electric po
wer area and
the literatu
re on
t
h
ese app
licatio
n
s
h
a
s b
e
co
m
e
en
orm
o
u
s
in
v
o
lu
m
e
.
M
o
t
i
on c
o
nt
r
o
l
an
d
po
wer
el
ect
ro
ni
cs are
ot
her a
r
eas
w
h
er
e ne
ural
net
w
o
r
ks
an
d
fuzzy
l
ogi
c
(FL
)
have
be
e
n
appl
i
e
d
[6]
.
T
h
i
s
pap
e
r ai
m
s
t
o
prese
n
t
a
sur
v
ey
of
har
m
oni
c co
m
p
ensat
i
o
n usi
n
g
shu
n
t
AP
F wi
t
h
an
ap
p
lication
of AI.
In
t
h
e p
a
st
few years, m
a
n
y
AI ap
p
licati
o
n
s
i
n
th
e activ
e power filter APF
h
a
v
e
expan
d
e
d.
Th
e fo
llo
wi
n
g
is
rev
i
ew o
f
wo
rk
do
cu
m
e
n
t
ed
in
th
e
literature
in
cl
u
d
e
th
e ap
p
lication
o
f
AI
techn
i
qu
es
t
o
APF
related
issu
es.
Fi
gu
re
1.
B
a
si
c p
r
i
n
ci
pl
e
of
A
P
F
Fi
gu
re
2.
The
o
ret
i
cal
wave
f
o
r
m
s
2.
H
A
R
M
O
N
I
C
M
I
TI
GA
TI
ON
A
P
PR
OA
CH
ES
I
n
pow
er d
i
st
r
i
bu
tio
n system
s h
a
r
m
o
n
i
c mit
i
g
a
tio
n
can b
e
carr
i
ed
ou
t thr
oug
h the fo
llo
w
i
ng
tech
n
i
qu
es:
(1) Passi
v
e
filter;
(2)
APF; (3)
Hyb
r
i
d
AP
F.
Du
e t
o
rem
a
rkab
le growth
i
n
p
o
wer electronics, th
e
u
s
e of
APF
h
a
s b
e
co
m
e
th
e d
y
n
a
mic so
lu
tion
fo
r m
itig
atio
n
of h
a
rm
o
n
i
cs. Th
e
fund
am
e
n
tal ru
le
o
f
APF is to
u
tilize ad
v
a
n
c
es in
p
o
wer electron
ics switches to
p
r
o
d
u
ce
eq
u
a
l and
o
ppo
site cu
rren
t si
g
n
a
ls th
at eli
m
in
ate
the ha
rm
onic c
u
rrents
from
the nonlinea
r l
o
a
d
s
[7].
2.
1.
P
a
ssi
ve Fi
l
t
eri
ng of
H
a
rm
oni
c
s
To
m
i
t
i
g
a
te th
e h
a
rm
o
n
i
c
d
i
sto
r
tion
p
a
ssive filterin
g
is th
e sim
p
lest trad
itio
n
a
l
reso
l
u
tio
n. The
p
a
ssiv
e
h
a
rm
o
n
i
c filters wo
rk
s is as fo
llo
ws: A filter conn
ected
in
p
a
ral
l
el with
th
e lo
ad
and
in
series wit
h
inductance
and capacitance
is
a c
u
rr
ent acce
ptor. T
h
e
ha
rm
onic fre
que
ncy
m
u
st be
re
duced to be
equal the
reson
a
n
t
frequen
c
y of th
e circu
it. Th
e im
p
e
d
a
n
ce
of
th
e
network
and
the lo
w im
p
e
d
a
nce o
f
t
h
e filter th
us
eli
m
in
ate th
e harm
o
n
i
c cu
rrent. Th
e
g
e
n
e
ral
k
i
nd
s
o
f
p
a
ssi
ve filters and
their ord
e
r are sho
w
n
i
n
Figure
3
[8
].
Th
e u
s
e
o
f
a passiv
e
filter h
a
s
m
a
n
y
d
i
sad
v
an
tag
e
s
wh
ich
d
ecrease th
e
reliab
ility
an
d
flex
ib
ility o
f
th
e filter
devi
ces
suc
h
a
s
, t
u
ni
n
g
,
l
a
rge
si
ze an
d
ri
sk
o
f
re
so
na
nce
pr
obl
em
s [9]
.
Fig
u
re
3
.
Co
mm
o
n
typ
e
s of
passiv
e
filters an
d th
eir con
f
i
g
u
r
ation
s
2.
2.
Acti
ve
Fi
l
t
eri
ng of
H
a
r
m
oni
c
s
APF i
s
a ki
n
d
of st
abl
e
,
hi
g
h
l
y
efficient, flexible op
tim
i
zation power qual
ity
machine, which plays
an
i
m
p
o
r
tan
t
ro
le in
i
m
p
r
ov
in
g
p
o
wer qu
ality. APF
is o
n
e
of m
o
st
i
m
p
o
r
tan
t
h
a
rm
o
n
i
c m
i
tig
atio
n
and
reactiv
e power co
m
p
en
satio
n electro
n
i
c
d
e
vices [10
]
. Acti
v
e
filters can
b
e
d
i
v
i
d
e
d
in
t
o
sing
le-p
hase activ
e
fi
lters and thre
e-phase (3
φ
) activ
e
fi
l
t
e
rs [1
1
]
, [12]
, [
1
3]
. R
e
search
on si
n
g
l
e
-
phase
(1
φ
) active
fi
lters h
a
s b
een
carried out, a
n
d the
resultant pape
rs
have
appeare
d
in tec
h
nical literature. Because
1
φ
version
s
are li
mited
to
lo
w-po
wer ap
plicatio
n
s
ex
cept fo
r electric tractio
n
or ro
lling
sto
c
k
,
1
φ
AFs attract
m
u
ch
less atten
tion
th
an
3
φ
AFs [13
]
.
Fo
r prov
id
i
ng h
a
rm
o
n
i
c com
p
en
satio
n
to a syste
m
, act
iv
e filters are co
mm
o
n
l
y u
s
ed
b
y
co
n
t
ro
lling
cu
rren
t
h
a
rm
o
n
i
cs in
sup
p
l
y
n
e
twork
s
at th
e
lo
w t
o
m
e
d
i
um
v
o
ltag
e
d
i
st
ribu
tio
n lev
e
l
o
r
fo
r
Evaluation Warning : The document was created with Spire.PDF for Python.
I
J
PED
S
I
S
SN
:
208
8-8
6
9
4
Adv
ance
d
Techniques
i
n
Harmonic
Suppression via Active
Power Filte
r (
APF)
: A Review (Ekhlas
Mhawi)
18
7
react
i
v
e
po
we
r o
r
v
o
l
t
a
ge c
ont
rol
at
hi
g
h
v
o
l
t
a
ge
di
st
ri
but
i
o
n l
e
vel
.
To ac
hi
eve
t
h
e va
ri
o
u
s f
u
n
c
t
i
ons
mentioned a
bove or in se
para
te active filters whic
h can
att
ack each a
s
pec
t
indivi
dually, these functions
m
a
y
b
e
pu
t tog
e
th
er in
a sing
le circu
it [14
]
. Accord
i
n
g
t
o
its procedure c
o
nnect
ion in
t
h
e
power system
, there are
two
typ
e
s
o
f
APFs: Series
APF and
p
a
rallel APF. Resear
ch a
n
d applications s
h
ow t
h
at the series
APF is
selected to com
p
ensate the harm
onic
v
o
ltage-source, wh
ile sh
un
t APF is
m
o
re ap
p
r
op
ri
ate to
co
m
p
en
sate th
e
harm
oni
c cu
rre
nt
so
urce
[1
5]
. APF
whi
c
h i
s
at
m
o
st
used as
a curre
nt
o
r
v
o
l
t
a
ge ha
rm
oni
c com
p
ensat
i
on, c
a
n
b
e
d
i
v
i
d
e
d
in
to
AC and
DC filters. Th
e
APF can
b
e
cl
assified
as top
o
l
o
g
y
-b
ased
classificatio
n
,
co
nv
erter-
base
d cl
assi
fi
cat
i
on, a
nd
su
p
p
l
y
-sy
s
t
e
m
-
based cl
assi
fi
cat
i
o
n, a c
o
m
b
i
n
at
ion
of
va
ri
o
u
s t
y
pes o
f
PP
F an
d A
P
F
is HPF
[
1
6]
.
2.
3.
Shu
nt APF
As APF eliminate the harm
onics and c
o
m
p
ensate th
e reacti
v
e power, it is also called as active power
line conditione
rs (APLC
)
. T
h
e shunt
or
parallel connectio
n is the m
o
re usual APF
configuration. T
h
e
basic
schem
e
of t
h
e co
nnect
i
o
n
i
s
sho
w
n i
n
Fi
gu
re
4,
where the Metal
–
Oxide
–
Sem
i
c
o
ndu
ctor
Field
-
Ef
f
e
ct
Tran
si
st
or M
O
SFET s
w
i
t
c
hi
n
g
de
vi
ce re
pre
s
ent
s
t
h
e A
PF
p
o
we
r bl
oc
k. T
h
e l
o
ads
wi
t
h
cu
rre
nt
ha
rm
oni
cs can
b
e
co
m
p
en
sated
b
y
th
is APF co
nfigu
r
ation
[1
7
]
. In
m
o
st
ind
u
s
t
r
ial p
r
o
cesses, p
a
rallel activ
e filters are
b
y
far
the
m
o
st vastly accept and
pre
dom
inant filter of choice.
In parallel at the point of
comm
o
n
coupling PC
C, the
activ
e filter is co
nn
ected
and
is sup
p
lied
fro
m
th
e
ma
in
p
o
wer circu
it. Th
e SAF obj
ectiv
e is to
prov
id
e
op
p
o
si
n
g
ha
rm
oni
c cu
rre
nt
t
o
t
h
e no
nl
i
n
ear
l
o
ad, e
x
cel
l
e
nt
l
y
resul
t
i
ng i
n
a pu
re harm
on
i
c
curre
nt
. Thi
s
ki
n
d
o
f
filter can
com
p
en
sate h
a
rm
o
n
i
c cu
rren
t
s g
e
n
e
rating
from th
e so
urce
I
hs
a
n
d the l
o
ad
I
h1
with
ou
t
affectin
g
the m
a
in com
p
one
nt of load c
u
rrent
I
f
. In
case o
f
id
eality, th
e p
a
rallel activ
e filter g
e
n
e
rates a cu
rren
t
equ
a
l t
o
(
I
hs
−
I
h1
).
As t
h
e internal rea
c
tance of the
s
o
urce a
g
ains
t
harm
onic com
p
one
n
ts of loa
d
c
u
rrent c
r
eates high
im
pedance
,
these harm
onic com
pone
nts
are not able to cross t
h
e im
peda
nce
of the
source.
In addition,
harm
oni
c com
p
o
n
e
n
t
s
of t
h
e
source curre
n
t are not able
to affect nonli
n
ear loa
d
.
In s
u
ch a case, the load
v
o
ltag
e
will be in
sin
e
form [10
]
, [1
8
]
, and
[1
9
]
.
Fi
gu
re 5
shows th
e
Prin
ci
p
l
e con
f
i
g
uration
of Vo
ltag
e
So
urce
I
nve
rt
e
r
V
S
I
base
d
Sh
unt
AP
F.
Fi
gu
re
4.
B
a
si
c di
ag
ram
of s
h
unt
AP
F
Fi
gu
re
5.
Pri
n
c
i
pl
e co
nfi
g
u
r
at
i
o
n
o
f
VS
I
base
d S
h
unt
AP
F
2.4.
Series
APF
As sh
own
i
n
Fig
u
re 6, th
e
series po
wer filter co
m
p
en
sates fo
r t
h
e h
a
rm
o
n
i
cs an
d
im
p
r
o
v
e
s the
p
o
wer qu
ality in
th
e system
. Th
e series activ
e filte
r id
eally b
e
h
a
v
e
s as a con
t
rolled
v
o
ltag
e
so
urce;
su
bsequ
e
n
tly t
h
e lo
ad
vo
ltage will h
a
v
e
only a p
o
s
itiv
e-seq
u
e
n
ce at th
e fun
d
a
m
e
n
t
al frequ
en
cy co
mp
on
en
t
[
2
0
]
. Figu
r
e
6 sh
ow
s th
e
basic co
nn
ection
d
i
ag
r
a
m
for
seri
es AP
F and
Fi
g
u
re
7 s
h
o
w
s t
h
e
pri
n
ci
pl
e
con
f
i
g
urat
i
o
n
f
o
r a
VSI
-
b
ase
d
AP
F. B
y
pr
o
duci
ng a P
W
M
vol
t
a
ge
wa
vef
o
rm
whi
c
h
i
s
subt
ract
e
d
o
r
ad
ded
ag
ain
s
t th
e supp
ly v
o
ltag
e
wav
e
fo
rm
, th
e m
a
in
ob
j
ect
o
f
th
e series activ
e fi
lter is to
m
a
in
t
a
in
a n
e
t sinu
so
id
al
v
o
ltag
e
w
a
v
e
fo
r
m
acr
o
ss th
e lo
ad
is carr
i
ed o
u
t
. Th
e
v
o
ltag
e
-f
ed
PW
M i
n
v
e
r
t
er
w
ith
out a cu
r
r
e
n
t
m
i
n
o
r
loop
i
s
t
h
e choi
ce
of
po
wer ci
rc
ui
t
use
d
i
n
m
o
st
cases. Th
e
s
e Se
ries APFs are
not c
o
mm
only
use
d
in power
syste
m
lik
e th
e sh
un
t
APF [2
1
]
.Series activ
e filters
act as vo
ltag
e
so
urces con
n
ect
ed
in series
with
th
e electrical g
r
i
d
,
and
t
h
ey
ca
n c
o
m
p
ensat
e
f
o
r
vol
t
a
ge
ha
rm
oni
cs. T
h
re
e-
ph
ase series
APFs can als
o
c
o
mpensate unbalances
i
n
th
e ph
ase vo
l
t
ag
es. Th
e co
m
p
en
satio
n
cap
ab
ilities of
th
e series
APF in
creases, allowing
also
th
e
com
p
ensat
i
on
of v
o
l
t
a
ge sa
gs
, vol
t
a
ge s
w
el
l
s
and fl
i
c
ke
r, I
f
t
h
e dc l
i
nk o
f
t
h
e powe
r
i
n
v
e
rt
er i
s
con
n
ec
t
e
d t
o
o
t
h
e
r po
wer su
pp
lies [2
2
]
, [2
3
]
. Series
activ
e po
wer filters were in
tro
d
u
ced
t
o
op
erat
e
m
a
in
ly as a
v
o
ltage
regu
lato
r and
as a h
a
rm
o
n
i
cs iso
l
ato
r
b
e
tween
th
e
n
o
n
lin
ear lo
ad
an
d
the utili
ty syste
m
. Th
e series co
nnected
filter pro
t
ects t
h
e system
from
an
in
ad
equ
a
te sup
p
l
y
v
o
ltag
e
q
u
a
lity [2
4
]
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
S
SN
:
2
088
-86
94
I
J
PED
S
Vo
l. 6,
No
.
2,
Ju
ne 20
15
:
185
–
1
95
18
8
Fig
u
re
6
.
Series Activ
e Po
wer filter
Fig
u
re
7
.
Pri
n
cip
l
e con
f
i
g
uratio
n of
Vo
ltag
e
Sou
r
ce
In
vert
e
r
(
V
S
I
)
base
d APF
2.
5.
Hy
bri
d
Acti
ve
P
o
w
er Fi
l
t
er
A co
m
b
in
atio
n o
f
static an
d
p
a
ssiv
e
filters
u
s
ing
to red
u
c
e th
e cost of t
h
e static co
m
p
en
sation
is
called
as h
ybri
d
activ
e
p
o
wer filter. To can
cel th
e m
o
st
rel
e
v
a
n
t
h
a
rm
o
n
i
cs o
f
t
h
e lo
ad
,
th
e p
a
ssi
v
e
filt
ers are
u
s
ed
an
d th
e activ
e filter is ded
i
cated
t
o
cancel o
t
h
e
r
h
a
rm
o
n
i
cs co
m
p
on
en
ts
o
r
to
im
p
r
ov
ing
t
h
e
p
e
rforman
ce
o
f
passiv
e
filters and
th
u
s
, the to
tal co
st d
ecreases
w
ithout d
ecrease of efficien
cy [1
7
]
. Fig
u
r
e 8
show
s th
e
m
o
re usual
hy
bri
d
t
o
pol
ogi
es
[2
5]
. F
o
rm
erl
y
,
m
a
ny
of t
h
e
cont
rol
l
e
rs
f
o
r
APF
o
p
erat
i
o
n we
re i
m
pl
em
ent
e
d
base
d
on
a
n
al
og
ue
ci
rcui
t
s
.
Due
t
o
t
h
i
s
,
t
h
e
per
f
o
rm
an
ce
of
t
h
e APF is
affecte
d
by the
signal de
viation.
Dig
ital con
t
ro
llers using
m
i
cro
c
on
tro
llers
o
r
Dig
ital Si
gnal
Process
o
r DSPs are
selected, prim
arily due to its
i
m
m
u
n
ity
to
n
o
i
se an
d
flexib
ility. B
y
u
s
i
n
g
d
i
g
ital
m
e
t
h
od
s, the h
i
gh-ord
e
r h
a
rm
o
n
i
cs are n
o
t
filtered
effectiv
ely,
b
e
cau
se
of th
e hardware limita
tio
n
o
f
sam
p
li
n
g
rate i
n
real
-ti
m
e ap
p
licatio
n.
In add
ition
,
t
h
e
u
tilizatio
n
o
f
fast switch
i
ng p
o
wer electro
n
i
c switch
e
s
(i.e. In
su
lated
-
Gate Bip
o
l
ar Tran
sisto
r
(IGBTs),
M
e
t
a
l
–
Oxi
d
e–
Sem
i
cond
uct
o
r
Fi
el
d-Ef
fect Transistor (M
OSFE
Ts)) in
APF a
pplication ca
uses swi
t
chi
ng
freq
u
e
n
c
y no
ise to
ap
p
e
ar in
th
e co
m
p
en
sated
sou
r
ce cu
rr
en
t. Ad
d
ition
a
l filterin
g
circu
it is req
u
i
red
to
redu
ce
th
is switch
i
n
g
freq
u
e
n
c
y
n
o
i
se and
to prev
ent in
te
rferen
c
e
with
o
t
h
e
r sen
s
itiv
e eq
u
i
p
m
en
t [26
]
[27
]
.
Fig
u
re
8
.
Hy
b
r
id
activ
e
po
wer filter
(HAPF) Topo
log
i
es
(a) series activ
e po
wer
filter +shu
n
t
p
a
ssi
v
e
filter,
(b) sh
un
t active po
wer
filter
+ shu
n
t
p
a
ssive filter,
(c
) activ
e
p
o
wer filter
co
nn
ected in series with sh
unt
p
a
ssiv
e
filter
3.
ARTIFICIAL
INTELLIGE
NCE
(AI)
To
d
e
tect and
co
m
p
en
sate h
a
rm
o
n
i
c cu
rren
t th
ere are m
a
n
y
classical tec
h
n
i
q
u
e
s.
A
g
e
n
e
ral
o
u
tline
for t
h
e
research
literature
o
n
h
o
w t
o
so
lv
e
harm
o
n
i
c d
i
stor
t
i
o
n
prob
lem
s
u
s
in
g AI is
presen
ted in
t
h
is
pap
e
r.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
J
PED
S
I
S
SN
:
208
8-8
6
9
4
Adv
ance
d
Techniques
i
n
Harmonic
Suppression via Active
Power Filte
r (
APF)
: A Review (Ekhlas
Mhawi)
18
9
Th
e au
to
m
a
tio
n
o
f
activ
ities th
at are associated
with
human
th
ink
i
ng
is ab
ro
ad
defi
n
itio
n
o
f
AI,
su
ch as
l
earni
n
g
,
deci
si
on m
a
ki
ng
, pe
r
cept
i
o
n
,
p
r
obl
e
m
sol
v
i
n
g, a
nd r
eason
ing
.
The A
I
too
l
s
of
benefit to the el
ectric
po
we
r com
m
u
n
i
t
y
i
n
cl
ude expe
rt
sy
st
em
s, Fuzzy
Logi
c
(FL
)
, ada
p
t
i
v
e
FL, genet
i
c
a
l
go
ri
t
h
m
s
(GA
s
) an
d
artificial n
e
ural n
e
tw
or
k
s
(A
NN
s) [6
].
3.
1.
An Ap
pl
i
c
ati
o
n of FL
i
n
AP
F
FL, t
h
e
l
o
gi
c
of a
p
pr
o
x
i
m
ate reas
oni
ng
, c
ont
i
n
ues t
o
gr
ow
i
n
i
m
port
a
nce, as
i
t
s
ap
p
l
i
cat
i
on t
o
a
num
ber of
p
r
a
c
t
i
cal
pro
b
l
e
m
s
f
u
rt
h
e
r dem
onst
r
at
es of
i
t
s
use
f
ul
ness.
FL
h
a
s been
use
d
i
n
areas
suc
h
as
cont
rol
,
p
r
oce
ss, di
a
g
nost
i
c
s, est
i
m
at
i
on,
m
e
di
ci
ne,
id
en
tificatio
n
,
ag
ricu
ltu
re, th
e sto
c
k m
a
rk
et, etc.
Ho
we
ver
,
p
r
oc
ess co
nt
rol
i
s
b
y
far i
t
s
m
o
st
i
m
port
a
nt
an
d
vi
si
bl
e ap
pl
i
cat
i
on [
2
8]
. F
u
zz
y
m
odel
i
ng
pr
ovi
des
th
e ab
ility to
l
i
n
gu
istically sp
ecify ap
p
r
ox
imate relatio
n
s
b
e
tween
th
e inp
u
t an
d
ou
tpu
t
, wh
en
a system i
s
to
o
com
p
lex or too poorly unde
rstood to
be descri
bed in a
ccurate m
a
the
m
atical ter
m
s.
Recently, in
m
o
re
appl
i
cat
i
o
ns, F
L
co
nt
r
o
l
l
e
rs (
F
LC
s) ha
ve be
en of
i
n
t
e
rest
as
a good
alte
rnative
.
T
h
e
fuzzy system
s
do not
n
eed
a precise math
e
m
atica
l
m
o
d
e
l;
th
e ad
van
t
ag
e of it is t
h
at: Th
ey can
h
a
nd
le no
n-linearity; can
wo
rk
with
im
preci
se i
n
pu
t
s
an
d t
h
ey
are
m
o
re ro
bu
st
t
h
an t
r
a
d
i
t
i
onal
n
onl
i
n
ea
r c
o
nt
r
o
l
l
e
rs [
29]
.
A
F
L
C
di
vi
de
d i
n
t
o
fo
u
r
areas: knowledge
base, Fuz
z
ification, infe
rence m
ech
ani
s
m
,
and def
u
z
z
i
f
i
cat
i
on.T
h
e kn
o
w
l
e
d
g
e ba
se i
s
desi
g
n
e
d
t
o
o
b
t
a
i
n
a
g
o
o
d
dy
nam
i
c respo
n
se
un
de
r u
n
c
e
rt
ai
nt
y
i
n
p
r
ocess
param
e
t
e
rs a
nd
f
o
r e
x
t
e
r
n
al
di
st
ur
ba
nces a
n
d
i
s
c
o
m
posed
of
a
dat
a
ba
s
e
an
d
rul
e
bas
e
an
d t
h
e
dat
a
base c
o
nsi
s
t
i
n
g
of
i
n
put
a
n
d
o
u
t
p
ut
m
e
m
b
ershi
p
f
unct
i
o
ns
p
r
o
v
i
des i
n
fo
rm
ati
on
f
o
r t
h
e
appropriate
fuzzificatio
n op
eration
s
,
i.e. th
e
defuzzification and infere
nc
e m
echanis
m
.
To c
o
nve
rt
t
h
e inpu
t conditio
n
s
in
t
o
a
fu
zzif
i
ed ou
tput, th
e
in
feren
ce m
e
c
h
an
ism
u
s
es a
co
llectio
n
of lin
gu
istic ru
les.
At last, d
e
fu
zzi
ficatio
n
is u
s
ed to
co
nv
ert th
e fu
zzy
out
put
s i
n
t
o
co
nt
r
o
l
si
gnal
s
. T
h
e f
o
rm
ul
at
i
o
n
of f
u
zzy
r
u
l
e
s
e
t
pl
ay
s a key
rol
e
i
n
i
m
pro
v
e
m
e
nt
of t
h
e s
y
st
e
m
per
f
o
r
m
a
nce i
n
desi
g
n
i
n
g
of i
t
s co
nt
rol
sy
st
e
m
[30]
,
[3
1]
. S
t
ruct
u
r
e o
f
a F
L
C
can
be see
n
i
n
Fi
gu
re
9.
I
n
fact
,
wi
t
h
FL
f
o
r
d
i
ffere
nt
w
o
rki
ng
co
n
d
i
t
i
ons,
i
t
’
s p
o
ssi
bl
e
t
o
desi
gn
a c
ont
rol
sy
st
em
adj
u
st
i
n
g t
h
e
cont
rol
surface, s
o
the control can
follow th
e re
ference c
u
rrent
even whe
n
very
high pea
k
s occur. Besides, dc
cap
acito
r
vo
ltag
e
s can
b
e
m
a
in
tain
ed at con
s
tan
t
lev
e
ls
with fu
zzy co
n
t
ro
l
[32
]
, [33
]
. Fuzzy arith
m
e
tic i
s
u
s
ed
for adj
u
stm
e
n
t
o
f
th
e pro
p
o
r
ti
o
n
a
l
-
in
tegral co
efficien
ts in
a ti
mely
fash
ion, wh
ile th
e g
e
neralised
in
tegrato
r
is
use
d
t
o
di
vi
de
fre
q
u
ency
i
n
t
e
gral
c
o
nt
rol
.
R
e
sul
t
s
ha
ve
sho
w
n t
h
at
t
h
e ne
w
di
vi
di
n
g
f
r
e
que
ncy
c
ont
rol
m
e
t
hod i
s
eas
y
t
o
cal
cul
a
t
e
, im
pl
em
ent
,
and i
s
ve
ry
eff
ect
i
v
e i
n
re
d
u
c
i
ng
harm
oni
c
s
. Fi
g
u
re
1
0
.
S
h
o
w
s a
n
adapt
i
v
e
fuzzy
di
vi
di
n
g
f
r
eq
u
e
ncy
cont
rol
m
e
t
h
o
d
[
34]
. T
h
e
m
o
st
co
m
m
o
n
l
y
used c
u
r
r
e
n
t
cont
rol
st
rat
e
gy
i
s
t
h
e fi
xed
ba
n
d
hy
st
eresi
s
m
e
tho
d
.
B
u
t
i
t
ha
s
t
h
e
di
sad
v
ant
a
ge
of
u
n
co
nt
r
o
l
l
a
bl
e hi
g
h
s
w
i
t
chi
n
g
f
r
eq
ue
n
c
y
.
To
im
prove this cont
rol, a
n
ada
p
tive hysteresis band curre
nt
c
ont
rol
t
ech
ni
q
u
e
can be
progra
mm
ed as a function
o
f
t
h
e activ
e filter an
d supp
ly p
a
ram
e
ters to
m
i
n
i
mize th
e in
fl
u
e
n
c
e of curren
t
d
i
st
o
r
tio
n
s
on
m
o
d
u
lated
wave
f
o
rm
. [3
5
]
. Fi
g
u
re
1
1
s
h
ows
t
h
e
schem
a
t
i
c
di
agram
of
f
u
zzy
co
nt
r
o
l
schem
e
.
Fi
gu
re
9.
St
r
u
c
t
ure
of
a
fuzzy
l
ogi
c c
ont
rol
l
e
r
Fi
gu
re
1
0
. T
h
e
A
d
apt
i
v
e
F
u
zz
y
Di
vi
di
ng
Co
n
t
ro
ller C
o
nfigu
r
ation
Evaluation Warning : The document was created with Spire.PDF for Python.
I
S
SN
:
2
088
-86
94
I
J
PED
S
Vo
l. 6,
No
.
2,
Ju
ne 20
15
:
185
–
1
95
19
0
Fi
gu
re
1
1
. T
h
e
schem
a
t
i
c
di
agram
of
f
u
zzy
cont
rol
sc
hem
e
3.
2.
Neur
al
Netw
o
r
ks Ap
pl
i
c
ati
o
n
The
no
n-
si
n
u
s
o
i
d
al
c
u
r
r
ent
d
i
st
ort
e
d
by
ha
r
m
oni
cs and t
h
e l
o
w
p
o
we
r-
f
act
or a
r
e p
r
o
d
u
ced
by
t
h
e
non-linear loa
d
s in electricity suppl
y net
w
orks, [36], [37]. AI appli
cations res
o
lvi
n
g the powe
r
quality
pr
o
b
l
e
m
m
e
nt
ione
d ab
o
v
e by
usi
n
g t
h
e par
a
l
l
e
l
APF st
rat
e
gy
i
n
t
w
o
-
wi
re
di
st
ri
but
i
o
n sy
st
em
s. The pr
o
pos
ed
AI ado
p
t
ed
is an
ANN respo
n
s
i
b
le fo
r t
h
e d
e
tectio
n
o
f
cu
rren
t
h
a
rm
o
n
i
cs
for th
e
activ
e p
o
wer filterin
g
process
.
Fi
gure 12
presents
the c
o
n
f
iguratio
n of t
h
e m
o
d
i
fied
propo
sed
ANN al
g
o
rithm
wh
ich
is li
mited
in
com
p
lexity com
p
ared to
t
h
e
cl
assi
cal
al
gori
t
h
m
[3
6]
. A m
odi
fi
ed
AN
N
was ef
f
ect
i
v
el
y
devel
ope
d i
n
det
ect
i
ng
harm
oni
c c
o
m
pone
nt
s an
d a P
W
M
was al
so i
m
pl
em
ent
e
d i
n
g
e
nerat
i
n
g s
w
i
t
c
hi
n
g
st
rat
e
gi
es
fo
r t
h
e
filter [3
8
]
. To eli
m
in
ate th
e
h
a
rm
o
n
i
cs, an
ad
ap
tiv
e radial b
a
sis fu
n
c
t
i
o
n
(RBF) n
e
u
r
al n
e
twork
co
n
t
ro
l
sy
st
em
for a t
h
ree
-
p
h
ase
AP
F i
s
pr
o
p
o
s
ed
.
Thi
s
m
e
t
hod
feat
ure
s
hi
gh
cont
rol
preci
si
on
, a wi
de ra
nge
o
f
ap
p
lication
s
, an
d real-tim
e o
p
eratio
n, th
en
t
h
e to
tal
ha
rm
onic distortion c
u
rrent ca
n
be
dec
r
ease e
ffectivel
y.
Fi
gu
re 1
2
. A
m
odi
fi
ed A
N
N
e
x
t
r
act
i
o
n
t
o
p
o
l
ogy
3.
3.
Gene
ti
c Al
g
o
ri
thm
s
GAs
we
re
fi
rs
t
l
y
i
n
t
r
od
uce
d
and
so
p
h
i
s
t
i
cat
ed as
a sy
st
e
m
aim
e
d at
m
odel
i
n
g a
n
d e
xpl
ai
ni
ng
t
h
e
adapt
a
t
i
o
n
of
n
a
t
u
ral
sy
st
em
s. A
basi
c
GA
i
n
cl
udes
sel
ect
i
o
n, m
i
xi
ng,
an
d
m
u
t
a
t
i
on o
f
c
o
m
ponent
s.
Sel
e
ct
i
o
n
is d
r
iv
en
b
y
an
o
r
g
a
n
i
sm
’s a
b
ility
lik
e b
i
o
l
o
g
i
cal syste
m
s, to
su
rv
iv
e in its en
v
i
ron
m
e
n
t. Mix
i
n
g
is u
s
ually
i
m
p
l
e
m
en
ted
by co
m
b
in
in
g
gen
e
tic in
fo
rm
atio
n
fro
m
two
o
r
m
o
re
p
a
ren
t
s. Mu
tation
is a m
ech
an
ism
fo
r
rei
n
t
r
od
uci
n
g
i
n
f
o
rm
at
i
on t
h
at
m
a
y
not
h
a
ve bee
n
c
o
nt
ai
ned i
n
t
h
e
po
p
u
l
a
t
i
on.
A
s
an
or
gani
s
m
,
t
h
e
po
p
u
l
a
t
i
on se
r
v
es wi
t
h
di
st
ri
but
e
d
k
n
o
wl
e
d
ge t
h
ro
u
g
h
o
u
t
t
h
e ge
nes o
f
t
h
e ent
i
r
e
po
p
u
l
a
t
i
on.
Fo
r fi
ndi
ng
optim
al
minimum
,
GAs are e
s
pecially
good, whe
r
e the fit
n
ess s
u
rface is
nonlinear, highly
convol
uted with
m
a
ny
l
o
cal
m
i
ni
m
a
and dep
e
nde
nt
o
n
se
ver
a
l
param
e
t
e
rs sim
u
l
t
a
neousl
y
.
To reac
h t
h
es
e sol
u
t
i
o
ns, t
h
e t
i
m
e
tak
e
n
b
y
GAs is j
u
st a
p
o
rtio
n
o
f
th
at
o
f
o
t
h
e
r trad
ition
a
l tech
n
i
q
u
e
s [39
]
, [4
0
]
.
A GA, im
p
r
o
v
e
d
later,
decrease
d
c
o
n
v
er
ge
nce t
i
m
e
but
was
no l
e
ss com
p
l
e
x or
easi
e
r t
o
im
plem
ent
,
har
d
wa
re-
w
i
s
e or as
onl
i
n
e
co
n
t
ro
l.
Hysteresis cu
rren
t con
t
ro
l, prin
cip
l
ed
on
ad
ap
tiv
e
n
o
i
se can
celling
,
n
eeds bo
th
lo
g
i
c and
an
al
o
g
u
e
ci
rcui
t
s
, c
o
m
p
l
i
cat
i
ng
har
d
w
a
re
devel
opm
en
t
.
A
n
i
m
pro
v
e
d
a
n
d
fast
AN
N c
o
nt
rol
sche
m
e
, fo
r i
t
s
l
e
a
r
ni
n
g
,
si
m
p
licit
y, g
e
n
e
ralizatio
n
ab
ility, an
d
also
fo
r po
ssi
b
l
e co
n
t
ro
l sch
e
me
m
o
d
i
ficati
o
n
thro
ugh
a weigh
t
s-
up
dat
i
n
g al
go
r
i
t
h
m
t
h
at
ex
pe
di
t
e
ha
rm
oni
cs ext
r
act
i
o
n,
t
h
e AP
F’s
re
sp
o
n
se a
n
d a
d
a
p
t
a
t
i
on
[4
1]
,
[
4
2
]
. T
o
reduce the %
THD
of t
h
e s
o
urce curre
nt, the GA is
us
ed
to
d
e
sign
th
e con
t
ro
llers. Th
e p
a
ram
e
te
rs fo
r
Evaluation Warning : The document was created with Spire.PDF for Python.
I
J
PED
S
I
S
SN
:
208
8-8
6
9
4
Adv
ance
d
Techniques
i
n
Harmonic
Suppression via Active
Power Filte
r (
APF)
: A Review (Ekhlas
Mhawi)
19
1
searchi
ng a
r
e
the inductor fi
lter (
L
f
)
,
d
c
bu
s vo
ltag
e
(
V
dc
), an
d t
h
e
hy
st
eresi
s
ba
nd
(
HB
) [4
3]
.
Fig
u
re 13
illu
strate ho
w to
search th
e
p
a
ram
e
ters o
f
APF u
s
i
n
g th
e GA m
e
th
o
d
.
3.
4.
G
A
an
d
a F
u
z
z
y
Neur
al
Ne
tw
or
k
Fig
u
r
e
14
sh
ow
s th
e stru
cture o
f
APF
w
ith
f
u
zzy n
e
ur
al netw
or
k
p
r
ed
ictiv
e con
t
ro
l. In
th
e strateg
y
,
t
o
pre
d
i
c
t
fut
u
r
e
harm
oni
c co
m
p
ensat
i
ng cu
rre
nt
, t
h
e fu
zzy
neural
net
w
o
r
k i
s
em
pl
oy
ed,
i
n
orde
r t
o
m
a
ke t
h
e
pre
d
ictive m
o
del com
p
act and
precise
i.
e. to
op
ti
m
i
ze th
e m
o
d
e
l p
a
ram
e
t
e
r,
a
GA
with an
e
fficient se
arch is
devel
ope
d [
4
4
]
. C
ont
rol
vec
t
or
u
in
th
e fo
rm
o
f
g
a
tin
g p
a
ttern
s of th
e in
v
e
rter switch
e
s b
a
sed o
n
th
e
p
r
ed
ictiv
e ou
tpu
t
, can
b
e
fi
g
u
red
o
u
t
b
y
th
e
m
o
d
e
l p
r
ed
ictiv
e con
t
ro
l algorith
m
,
wh
ich
main
tain
s track
ing
of
d
y
n
a
m
i
c referen
ce cu
rren
t
wit
h
ou
t tim
e d
e
lay. All resu
lts
prove t
h
at the
GA and t
h
e
fuzzy neural
network -
b
a
sed
pred
ictiv
e con
t
ro
ller can
cel vo
ltag
e
d
i
sto
r
tion
an
d
supp
ly cu
rren
t g
r
eatly
and the pe
rformance
of
h
a
rm
o
n
i
c co
mp
ensatio
n b
e
tter th
an
a PI con
t
ro
ller.
Figu
re
1
3
. T
h
e
G
A
a
p
p
r
o
ach
fo
r
APF
desi
gn
Figu
re
1
4
. T
h
e
struct
ure
o
f
A
P
F
4.
SUM
M
AR
Y
All AI techn
i
qu
es
p
r
esen
ted
in
th
is
pa
per are summ
arized in Figure 15
, to
give t
h
e rea
d
er an idea
of
a larg
e p
e
rcen
tag
e
o
f
recen
t
dev
e
lop
m
en
ts o
f
artificial in
tellig
en
ce i
n
APF.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
S
SN
:
2
088
-86
94
I
J
PED
S
Vo
l. 6,
No
.
2,
Ju
ne 20
15
:
185
–
1
95
19
2
Figure
15.
Sum
m
a
ry of A
I
i
n
APF
application
5.
CO
NCL
USI
O
N
The m
a
i
n
foc
u
s o
f
t
h
e
re
vi
ew
was t
h
e
use
of
FL,
ne
u
r
al
net
w
or
ks
, an
d
G
A
s i
n
AP
F
appl
i
cat
i
o
ns
.
Thi
s
pa
per s
u
g
g
est
s
di
f
f
ere
n
t
AI a
p
p
r
oac
h
es
t
o
adva
nce si
gnal
p
r
ocessi
n
g
t
echni
qu
es t
o
im
pro
v
e t
h
e
po
w
e
r
q
u
a
lity an
d
t
h
e syste
m
p
e
rforman
ce. An
ex
t
e
n
s
iv
e literature
su
rv
ey of AI tech
n
i
qu
es i
n
APF is
p
r
esen
t
e
d
to
provide a clear pers
pective
on va
rious
as
pe
ct
s of
AI t
o
t
h
e researc
h
er
s a
nd e
n
gi
nee
r
s
wo
rki
n
g i
n
t
h
i
s
fi
el
d.
These m
e
t
hod
s ha
ve t
h
e a
d
vant
a
g
es
of
q
u
i
c
k
resp
o
n
se, good current
tracking
acc
uracy of the
APF, a
n
d
m
i
nim
a
l curre
nt
ri
p
p
l
e
. Si
m
u
l
a
t
i
on an
d ap
pl
i
cat
i
on res
u
l
t
s
have s
h
ow
n t
h
at
usi
ng
AI i
n
cont
rol
m
e
t
hod
s i
s
not
onl
y
easy
t
o
c
a
l
c
ul
at
e and
i
m
pl
em
ent
,
but
al
so
very
e
f
f
ect
i
v
e i
n
re
d
u
c
i
ng
harm
oni
c
s
. F
r
om
t
h
e r
e
vi
ews
d
e
scri
b
e
d in
t
h
e p
a
p
e
r, it is rev
ealed
t
h
at mo
st of FLC h
a
s b
e
en
ap
p
lied to
regu
latin
g
th
e d
c
bu
s capacito
r
voltage
of a
n
APF filter,
while
m
o
st of ANN
has bee
n
a
pplied t
o
detec
ting curre
nt ha
rm
onics for t
h
e active
p
o
wer filtering p
r
o
cess. Th
e
GA
h
a
s
b
e
en
ap
p
lied to
o
p
t
i
m
ize th
e p
a
rameters o
f
th
e si
m
u
latio
n
m
o
d
e
l o
f
th
e
APF
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
J
PED
S
I
S
SN
:
208
8-8
6
9
4
Adv
ance
d
Techniques
i
n
Harmonic
Suppression via Active
Power Filte
r (
APF)
: A Review (Ekhlas
Mhawi)
19
3
REFERE
NC
ES
[1]
P.
Sa
lme
r
on,
S
.
P.
A L
i
tra
n
, “c
ontrol stra
teg
y
for h
y
br
id pow
er fil
t
ers to
co
m
p
ensate four-
w
ire thre
e-phas
e
s
y
ste
m
s”
,
Power
El
ectr
oni
cs
,
I
E
EE Transactions
on
, Vol. 25, pp.
1923-1931, 201
0.
[2]
L. Xia, D. Taih
ang and
T.Shen
gxue, “A
fuzzy
logic v
a
riable h
y
ster
esis ba
nd current control technique
fo
r
thr
e
e
phase shunt active power f
ilter”,
In
Con
t
rol,
Au
tomation and S
y
stems Engi
neering (
C
ASE)
, IEEE, In
ternationa
l
Conference on
, pp.
1-4
,
July
201
1.
[3]
D. Chen and
S
.
Xie
,
“
R
eview
of the
control
s
t
rateg
i
es
app
l
ied to
ac
tive
p
o
wer filt
ers
”
,
I
n
El
ectr
i
c
Uti
lit
y
Deregulation
,
R
e
structuring and Power Technologies,(
DRPT 2004)
. Proceedi
ngs of
the IEEE International
Conference on
, Vol.2,
pp.666-6
70,
April
2004.
[4]
J.
F.
Pe
tit,
H
.
A
m
a
r
ı
s and G. R
obles, “Control
schemes for shunt activ
e
fi
lt
ers
to m
itiga
t
e
har
m
onics inje
cted
b
y
inverted-fed motors”,
15th Proton Synchrotron
&
S
ynch
r
o
Cyclotr
on Committee PSCC
, Lieg
e, 22-26 August,
Session 22, 2005
.
[5]
C.
J.
Msigwa
,
B.
J.
Kundy
a
nd
B.
M.
Mwiny
i
wiwa
,
“C
ontrol A
l
gorithm
for Sh
unt Activ
e Power Filt
er using
S
y
nchronous R
e
ference Frame Theor
y
”,
World
Academy o
f
S
c
ien
ce, Eng
i
neering
and Technolog
y
, Vol. 58, 2009.
[6]
W
.
A. Ibrahim
a
nd M.M. Morcos, “
A
rtificia
l i
n
tell
igen
ce and
advanced mathematic
al tools fo
r power quality
applications: a s
u
rvey
”,
Pow
e
r
Deli
ver
y,
IE
EE Transactions on
, V
o
l. 17
, pp
. 668-6
73, 2002
.
[7]
S. Kim, G. Yoo
and J. Song
, “Abifunc
tion
a
l utility
conn
ected ph
otovoltaic s
y
st
e
m
with power f
a
ctor
corre
ction
a
n
d
UPS facilit
y”
,
In
Photovoltaic Sp
ecia
lists Conference, Ma
y 1996.
Conference Record of the Twenty Fifth IEEE
, pp.
1363-1368.
[8]
L.S. Czarneck
i,
“An overview
of methods of har
m
onic suppre
ssion in distributio
n sy
stems”,
In Power
Engineerin
g
Society Summer
Meeting, 2000. I
EEE
, Vol. 2
,
pp
. 800-805, 2000.
[9]
H.L. Jou,
J.C.
W
u
and H.Y.
Chu,
“
N
ew single-pha
se act
ive pow
er f
ilter
”
,
In proc. I
EE Electric Pow
e
r Applications
,
Vol. 141
, pp
. 12
9-13, 1994
.
[10]
H. Akagi, “New
trends in activ
e filters fo
r power condition
i
ng.
In
dustr
y
Applicati
ons”,
IEEE T
r
ansactions on
, V
o
l.
32, pp
. 1312-13
22, 1996
.
[11]
P.C. Tan,
Z. S
a
la
and A.
Jusoh, “A single-p
h
ase h
y
brid
active pow
er filter
using extensio
n pq theor
e
m for
photovoltaic ap
plication
”
,
In Power Electronics and Drives System
s, PEDS
2005. Internatio
nal Conference
on IEEE
, Vol. 2, pp. 1250-1255,
Decem
ber 2005
.
[12]
R. Dehini, B. B
e
rbaoui
, C. Ben
achai
ba and O.
Harici
, “Shunt
H
y
brid Activ
e Power Filter Im
provem
e
nt Based On
Passive Power Filters S
y
n
t
he
si
s b
y
Genet
i
c
Algorithm
”
,
International Jour
nal of
Engin
e
ering Scien
ce a
nd
Technology
, Vol. 2
,
pp
. 1185-11
93, 2010
.
[13]
Z. Xiao
, Y. Ch
e and X. D
e
ng,
“A Shunt Active Powe
r Filter
with Enhan
ced
D
y
nam
i
c p
e
rfor
m
ance using Du
al-
Repeti
tiv
e Cont
roller
and Pred
ictiv
e Com
p
ensation
”
,
International Journal of
Power Electro
n
ics and Drive
Sy
ste
m
s (IJ
PED
S)
,Vol. 3, pp. 20
9-217, 2013
.
[14]
G. Adam , A.
Zbant
and G.
Livint,
“New Simulink control
block for sing
le
phase shunt
active power
filter
s”,
In Advan
ced
Topics in
Electrica
l
Eng
i
neerin
g
(
A
TEE)
,8th International S
y
mposium on IEEE
, pp. 1-4, May
2013
.
[15]
I.I. Abdalla, K.S
.
R. R
a
o
and
N.
Pe
rumal, “Three-phase four-
l
eg s
hunt activ
e pow
er filter
to
comp
ensate harmonics
and reactive po
wer”,
Computer
s
&
Info
rmatics (ISCI)
,
IEEE Sym
posium,
20-23 March, pp. 495-50
0, 2011
.
[16]
H. Akagi
,
“
A
cti
v
e harm
onic
fi
lt
ers
”
,
Proceeding
s of th
e I
EEE
, V
o
l. 93
, pp
. 2128-
2141, 2005
.
[17]
A. Boukadoum, T. Bah
i
, A. Bou
guerne , Y
.
Soufi and S. O
udina, “Hy
s
teresis ban
d
curre
nt
and fu
zzy
log
i
c con
t
ro
l
for activ
e power
filter
”
,
In E
c
olo
g
ical V
e
hic
l
es a
nd Renewabl
e E
n
ergi
es (
EVER)
, 8th Internationa
l Conferenc
e
an
d
Exhibition
on I
E
EE
, pp. 1-6, Mar
c
h 2013.
[18]
S. Janpong, K.L
.
Areerak
and K.N. Areer
ak, “A liter
a
tur
e
su
rvey
of neur
al network applicat
ions for shunt activ
e
power fil
t
ers
”
,
J. World Acad. Sci., Eng. Tech
., Vo
l. 60
, pp
. 392-39
8, 2011
.
[19]
K. Chaturved
i
,
A. Mahor and A
.
D. Dwi
vedi, “Activ
e Power Filt
er Techniqu
es for Harmonics Suppression in Non
Line
ar Lo
ads”,
I
n
ternational Jou
r
nal of S
c
ien
tific Engin
eering
an
d Technolog
y
, Vol. 1
,
pp
. 123-12
7, 2012
.
[20]
F
.
Iranne
zhad
,
S
.
H. Hos
s
e
inian
,
M
.
Abedi
and
B. Vahid
i
, “Using the
Instantaneous Power Theor
y
in ord
e
r to
control
the
curre
nt in
the p
a
ral
l
e
l
act
ive
fil
t
er
to c
o
m
p
ens
a
te re
act
i
v
e power
and re
duction of
harm
onics
”,
In Pow
e
r
Ele
c
tronics, Dri
ves and Energy
Systems (
PED
ES)
,
IEEE Int
e
rnational Conf
erence on
IE
EE
, p
p
. 1-5, December
2012.
[21]
K. Sarasvathi and R.
Rajalakshmi, “Performance Anal
y
s
is Of Shunt Activ
e Filter
Using Differ
e
nt Contro
llers”,
International Jo
urnal of
Engineer
ing Trends and
Technology (
I
JETT)
,Vol. 4
,
pp
.
1-6, 2013
.
[22]
N. Kum
a
r, K.B. Thapa and S. Khalid
, “
A
novel series ac
tive
power filte
r sch
e
m
e
using adaptive tabu sear
ch
algorithm for
harmonic compensation
”
,
Int. J.
Ap
pl. Inno
v. Eng
.
Manage, (
I
JAI
E
M)
,
Vol. 2
,
pp
. 1
55-159, 2013
.
[23]
J.G. Pinto, H.
Carneiro
, B
.
Ex
posto, C. Couto
and J.L.
Afons
o, “
T
rans
form
er
les
s
s
e
ries
act
iv
e power f
ilt
er
t
o
com
p
ens
a
te vo
lt
age dis
t
urb
a
nc
es
”,
In
Power Electronics
and Ap
plications
, Pro
c
eedings of the August 2011-14th
European Conference on
IEEE
, 1-6,
2011
.
[24]
M.A. Mulla, R.
Chudam
a
n and A. Chow
dhur
y
,
“Series Active Power Filter Usi
ng Generalised In
stantan
e
ous Power
Theor
y
”
,
Proceedings of
the Worl
d Congress on
Engineering
, Vol. 2
,
pp
. 1-6
,
201
2.
[25]
A.
A.
Dutta, M.
Sabley
, B.
S. Sudame
and
A.N.
Kadu, “
H
arm
onic Com
p
ens
a
tion
in Power S
y
ste
m
using Acti
v
e
Power Filters”
,
I
n
t.
J. of Multidis
ciplinary and Cu
rrent research
, p
p
. 188-192
, 201
3.
[26]
T. Demirdelen,
M. Inci, K.C.
Bay
i
ndir
and M. Tumay
,
“Review of h
y
br
id activ
e power filter topolog
ies an
d
controllers”,
In
Power Eng
i
neeri
ng, En
ergy and
Ele
c
trica
l
Dr
ive
s
(
P
OWERENG), Fourth
Internat
ional Conf
eren
c
e
on IEEE
, pp. 58
7-592, May
201
3.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
S
SN
:
2
088
-86
94
I
J
PED
S
Vo
l. 6,
No
.
2,
Ju
ne 20
15
:
185
–
1
95
19
4
[27]
I.
A.
Altawil,
K.A.
Mahafzah and A.
A.
Smadi, “H
y
b
rid active p
o
wer filter
based on diode clamped inver
t
er an
d
h
y
ster
esis
band current controller”,
In
Advan
ces
in Computation
a
l Tools
for Engineering
Applications (
A
CTEA)
,
2
nd
International C
onferen
ce on I
EEE
, pp
. 198-20
3, December 20
12.
[28]
G.C.D. Sousa and B.K. Bose
, “Fuzzy
Logic Applications to Power
Electro
nics
and Drives
An Overview”
,
Industrial Electr
onics, Control a
nd Inst
rumentation, Proceedings
of th
e IECO
N 2
1
st Internationa
l Conference on
,
IEEE
, Vol. 1
,
pp
. 57-62
, 1995
.
[29]
C. Salim
and B.M. Toufik
, “Intellig
ent Contro
ll
ers fo
r Shunt Activ
e Filter to
Com
p
ensate Current Harm
onics
Based on SRF and SCR C
ontrol Strategies”,
International Journ
a
l on Electric
al
Engineering and
Informatics
, V
o
l.
3, pp
. 372-393
,
2011.
[30]
U. Mahesh and M.A. Shai
k, “
F
uzz
y
Logic C
ontrolle
r of a
Series
Active P
o
wer Filter for
Power Qualit
y
Improvement”,
I
n
ternational Jou
r
nal of Mod
e
rn
Engineering Research (
I
JMER)
,
Vol. 2
,
pp
. 412-
415, 2012
.
[31]
M. Chau, A. Lu
o and V. Chau,
“PID-
F
uzzy
Co
ntrol Me
thod with Time Delay
Compensation for H
y
brid Activ
e
Power Filter
wit
h
Inje
ction
Cir
c
u
it”
,
Internationa
l Journal o
f
Com
puter App
lica
tio
ns
. Vol. 36, 201
1.
[32]
H.
Benalla and H.
Djeghl
oud
, “Shunt Active Filter Contro
lled b
y
Fuzzy
Logic”,
J. King
Saud Univ.
Vol. 18, pp.
231-247, 2005
.
[33]
J.W. Dixon, J.M. Contardo an
d L.A. Mor´
an, “A Fuzzy
-
Con
t
rolled
Activ
e
Front-End Recti
fi
er with Curre
nt
Harm
onic Filteri
ng Charact
eristi
cs and Minim
u
m
Sensing
Variables”
,
IEEE Transactions On
Power Elec
tronic
s
,
Vol. 14
, pp
. 724
-729, 2013
.
[34]
N.S. Rao
,
J. A
y
y
a
pp
a
and A.K
.
Babu,
“Develop
ment of H
y
brid
Active Powe
r Filter
Based on
th
e Adaptive Fu
zzy
Dividing Frequency
Contro
l Method”,
Internatio
nal Journal of E
ducation and Ap
plied Research
(
IJEAR
), Vol. 4,
pp. 1-7
,
2014
.
[35]
D.A. Gadanay
a
k
and P.
C. Panda, “A Novel Fuzzy
Variable-Band
Hy
st
eresis Curr
ent Controller For Shunt Active
Power Filters”
,
ACEEE Int.
J. o
n
Control S
y
stem and Instrumentation,
Vol. 2
,
p
p
. 24-28
, 2011
.
[36]
A. Sabo, N.I
.
Abdul Wahab,
M. Radzi, M. Amran
and N.F.
Mailah
,
“A modified
artif
icial n
e
ural network (ANN)
algorithm to
co
ntrol shunt
active power filter
(SAPF) for curr
ent harmonics r
e
duction”,
In
C
l
ean Energ
y
an
d
Technology (
C
EAT)
,
IEEE
Conference on
IEEE
,
pp. 348-352
November, 2013
.
[37]
R. Dehini, A. B
a
ssou and B.
Ferdi, “Artificial n
e
ural networks applic
ation to im
prove shunt acti
v
e power filter
”
,
International Jo
urnal of Comput
er and Information Engin
eering
,
Vol. 3
,
pp
.
247-
254, 2009
.
[38]
J. Fei and
Z. W
a
ng, “Adaptiv
e
control of
ac
tiv
e power filter u
s
ing RBF neural network”,
In
Mechatronics
and
Automation
(
I
CMA)
, IEEE In
ter
national Con
f
erence on
IEEE,
pp
. 767-772, August 2013.
[39]
M. El-Hab
rouk
and M.K. Darwi
s
h, “A
new con
t
rol t
echniqu
e fo
r active power
f
ilters
using a co
m
b
ined genetic
algorithm/conventional an
aly
s
is”,
Industrial Electronics,
IEEE Transactions on
, V
o
l. 49
, pp
. 58-66
, 2002
.
[40]
F. Cupertino
,
M. Marinelli, P. Za
nchetta
and M.
Sumner, “Modelling and d
e
si
gn
of shunt active p
o
wer filters using
genetic algor
ith
ms”,
In Pow
e
r Electronics and
Applicat
ions, European
C
onf
er
en
ce on
IE
EE
, pp
.
1-5, 2005
.
[41]
K. Saifullah
,
D. Bharti and S. Bhim
, “New
Optim
um
Three Phase Shunt Active Power
Filter based on Adaptive
Tabu Sear
ch an
d Genetic Algor
ithm using ANN control
in un
balan
ced
and dist
ributed supply condition”,
UTM
Jornal of Electr
ical Engineering
, Vol. 14, 2012.
[42]
G.J. Krishna , N.R. Raju and C.
Sw
apna, “Genetic Algorithm tuned PID cont
roller based Shunt Active Filter for
harmonic r
e
duction”,
I
PAS
J In
ter
national
Journal of
Electrical En
gineering(
IIJEE
), Vol. 2
,
pp
. 1-9
,
2014
.
[43]
T. Narongri, K.L. Areerak and
K.N. Areerak, “Design of
an active power filter using genetic al
gorithm
techniq
u
e”,
Proceed
ings of the 9th WSEAS i
n
terna
tional
con
f
erenc
e
on Artifi
cial int
e
l
ligen
ce
, knowledge eng
i
neering and dat
a
bases, World S
c
ientific and
Engin
eering
Acad
emy and So
ciety (
W
SEAS
), PP. 46-5
0
, 2010
.
[44]
Y. Weike, L. Bin and X.
Yong, “Fuzzy
Neur
al Networks and GA ba
sed Predictiv
e Control fo
r Active Power
Filter”
,
M
e
asuring Technolog
y
and Mecha
t
roni
cs Automat
i
on
(
I
CMTMA)
, Fifth
Internati
onal C
onference on
, p
p
.
598-601, 2013
.
[45]
A. Bhattach
ar
y
a
and C. Ch
akraborty
, “A Shunt Ac
tive Power
Filter With
Enh
a
nced Perfo
rmance Using ANN-
Bas
e
d P
r
edict
i
v
e
and Adaptiv
e
Controllers
”,
IEEE Transactions on I
ndustrial Electronics
, Vol.
58, pp. 421-428
,
2011.
[46]
B.
Vasumathi and S.
Moorthi,
“Harmonic estimation
using M
odified ADALINE algorithm with Time-Variant
Widrow—Hoff
(TVWH) learning rule”,
In
Computers
&
Informatics (
I
SCI)
, IE
EE Symposium on
IEEE
, pp. 11
3-
118, Mar
c
h 201
1.
[47]
E. Almaita and
J.A. Asu
m
a
du, “Harmonic content extr
action in convert
er waveforms u
s
ing radial basis function
neural n
e
tworks (RBFNN) and p
q
power theor
y
”,
In
Power and Energy Conferen
ce at Illinois (
PECI)
, IEEE
, pp
. 1
-
7, Februar
y
201
1.
[48]
A. Bhattachar
y
a
and C. Chakraborty
, “A shunt a
c
tiv
e
power filter with enhanced
performance using ANN-based
predic
tive
and
a
d
aptiv
e con
t
roll
e
r
s
”
,
IndustrialElectronics
,
I
EEE Transactions
on
, Vol. 58, pp. 421
-428, 2011
.
[49]
J.
Talla,
Z. Pe
routka
, S.
J.
Ova
s
ka
a
n
d J.
Ste
h
lik, “G
enetic
Algor
ithm
Based Opt
i
m
i
zation
of M
G
P-FIR Current
Referen
ce Gene
rator for Active
Power
Filters”,
In
Internationa
l Joint Conferen
ce CISIS’12-IC
EUTE´
12-SOCO´
12 Special Sessions, Spri
nger Be
rlin
Heid
elberg
,
pp. 429-438
, Jan
u
ar
y
2013.
[50]
N.L. Surasm
i and M.R. Si
ndhu, “Optim
um Allocation of Act
i
ve Filte
rs In A 4-Bus
Sy
st
em
Using Genetic
Algorithm”,
International Journal
of
Emerging
Technology and
Ad
vanced
Engin
eer
ing
, Vol. 2
,
pp
. 2
51-256, 2012
.
[51]
H. Has
h
em
i De
zak, M
.
M
oham
m
a
d alizad
eh-S
habes
t
ar
y, H
.
As
karian-Ab
y
ane
h
and M
.
Rezae
i-J
e
garlu
e
i, “
N
e
w
Approach to Opt
i
m
i
ze the Apfs
P
l
ac
em
ent Bas
e
d
on Ins
t
antan
e
ous
React
ive P
o
wer
Theor
y
b
y
Gene
tic Algori
t
hm
”,
Journal of Electr
ical Engineerin
g,
Vol. 65
,
pp. 1
2
-20, 2014
.
Evaluation Warning : The document was created with Spire.PDF for Python.