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. 4, Decem
ber
2015, pp. 897~
905
I
S
SN
: 208
8-8
6
9
4
8
97
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
Modified Synchronous Referen
ce Frame based Harmonic
Extraction for Shunt Active Filter
CS.
Su
bas
h
K
u
mar
1
,
V. G
o
palakrishn
an
2
1
Depart
em
ent o
f
El
ectr
i
cal
and
Electroni
cs Eng
i
neering
,
PSG
Institut
e
of
T
echnol
og
y
and Appl
ied
Resear
ch, Coim
batore
2
Depart
em
ent o
f
El
ectr
i
c
a
l
and
Ele
c
troni
cs
Eng
i
neer
ing
,
Govern
ment College of
Techno
log
y
, Coimbatore
Article Info
A
B
STRAC
T
Article histo
r
y:
Received Apr 24, 2015
Rev
i
sed
O
c
t 20
, 20
15
Accepted Nov 14, 2015
This paper presents the modified sy
n
c
hronous referen
ce frame
based Shunt
Active Fi
lter
(SAF) for the inst
antan
e
ous com
p
ensation of
har
m
onic curren
t
present
at
the P
o
int of Common Coupli
ng. Th
e
harmonics gener
a
ted
b
y
th
e
non linear lo
ad
are
extracted using the
positive f
r
ame and n
e
gative frame of
the inpu
t signal
using the modified s
y
n
c
hronous referen
ce frame theor
y
with
extend
ed Multip
le Ref
e
ren
ce Fr
am
e based PLL
(EMRFPLL). B
a
sed on th
e
harmonics extracted
,
pulse width m
odulation signals are gen
e
r
a
ted using
Artific
ial Neur
a
l
Network based Space Vector
Pulse width Modulation
technique (ANNSVP
WM).
U
s
ing this sw
itching technique the losses can be
reduced
and co
m
p
ensation can
be done m
o
re
accur
a
t
e
l
y
Th
e
concep
t was
verified using
MATLAB / Simulink Simu
lati
on and the
resul
t
s confirm
the
THD at po
int of
common couplin
g is below th
e r
e
quired standards
.
Keyword:
AN
NS
VP
WM
EMRFPLL
M
odi
fi
e
d
Sy
nc
hr
o
n
o
u
s
Refere
nce Fra
m
e
THD
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
:
C.S.Su
bash Kumar,
Depa
rtem
ent of Electrical a
n
d
El
ect
ro
ni
cs E
n
gi
nee
r
i
n
g,
PSG In
stitu
te of Tech
no
log
y
an
d App
lied
research,
Co
im
b
a
to
r
e
–
6
410
62
.
Em
a
il: su
b
a
shku
m
a
rcs@g
m
ail
.
co
m
1.
INTRODUCTION
Th
e
nu
m
b
er of electric u
tilitie
s u
s
ed fo
r con
t
ro
lling
an
d com
p
u
tin
g
d
e
v
i
ces are in
creasin
g every
d
a
y,
wi
t
h
t
h
e
a
dve
n
t
of
t
ech
nol
og
y
fo
r c
ont
rol
o
f
m
o
t
o
rs
us
ing dr
iv
es
fo
r pow
er conv
er
si
on
th
er
e is an essen
tial
necessi
t
y
t
o
us
e po
we
r sem
i
con
d
u
ct
o
r
de
vi
c
e
s i
n
t
h
e el
ect
ri
cal
sy
st
em
. Swi
t
c
hi
ng
ON a
n
d O
FF t
h
e si
n
u
s
oi
dal
su
pp
ly d
i
storts th
e fu
nd
am
en
tal sig
n
a
l g
i
v
e
n
to
t
h
e
electrical device. T
h
ese
devices
have non linea
r load
characte
r
istics and these
non
l
i
n
ear l
o
a
d
cu
r
r
ent
s
c
a
use
di
s
t
ur
bance
i
n
ot
her system
s connected to t
h
e
mains
[1]
.
The
harm
oni
cs
whi
c
h a
r
e
gene
rat
e
d
d
u
e
t
o
t
h
e s
w
i
t
c
hi
ng of
nonlinea
r l
o
ad leads
to
increased los
s
es
in the
electrical syste
m
,
poor powe
r fact
o
r
, i
n
cre
a
se i
n
react
i
v
e
po
wer co
ns
u
m
pti
on an
d i
n
duce
s
di
st
u
r
ba
nce i
n
electrically sen
s
itiv
e lo
ad
s.
The
harm
oni
cs
can
be
com
p
ensat
e
d
by
p
r
o
v
i
di
ng
pa
ssi
ve
fi
l
t
e
rs, w
h
i
c
h
c
o
m
p
ensat
e
s t
h
e
harm
oni
c
of
a part
i
c
ul
ar
or
der a
n
d t
h
e
ot
her
ha
rm
oni
cs rem
a
i
n
i
n
t
h
e
sy
st
em
. B
y
t
h
e pr
o
v
i
s
i
o
n o
f
passi
ve el
em
ent
s
f
o
r
filterin
g
resonan
ce prob
lem
o
ccurs in
an
electrical syste
m
an
d
th
e syste
m
b
eco
m
e
s
b
u
l
k
y
[2
]. In
o
r
d
e
r t
o
o
v
e
rco
m
e th
es
e d
r
awb
a
ck
s
activ
e filters were in
t
r
odu
ced
fo
r
d
y
n
a
m
i
c co
m
p
en
satio
n
wh
ich
elimin
ates
h
a
rm
o
n
i
cs of an
y o
r
d
e
r
p
r
esen
t in
th
e syste
m
an
d
reson
a
n
ce prob
lem
is
av
o
i
d
e
d
to
k
e
ep
th
e po
wer
q
u
a
lity
i
ndi
ces
bel
o
w t
h
e
req
u
i
r
e
d
l
e
v
e
l
s
as s
p
eci
fi
ed
by
st
a
nda
rd
o
r
gani
zat
i
o
ns s
u
c
h
as
IEE
E
51
9.
Activ
e filters
are classified
as series typ
e
, sh
un
t typ
e
and
Hy
b
r
id
active filter. Shun
t typ
e
activ
e
filters are u
s
ed
for th
e eli
m
in
atio
n
of curren
t h
a
rm
o
n
i
cs p
r
esen
t in
th
e syste
m
. Hyb
r
id
activ
e filter is a
co
m
b
in
atio
n
of sh
un
t typ
e
an
d series ty
pe filter us
ed fo
r t
h
e elim
in
a
tio
n
o
f
bo
th
v
o
ltag
e
and
cu
rrent
harm
onics
pres
ent in t
h
e syste
m
.
SAF m
a
k
e
s th
e so
urce cu
rren
t sin
u
s
o
i
d
a
l irresp
ectiv
e
o
f
harm
o
n
i
c p
r
esen
t in
th
e n
o
n
lin
ear lo
ad
and
d
u
e
t
o
d
i
storted
sou
r
ce vo
ltag
e
cond
itio
n
s
.
Th
e d
y
n
a
m
i
c
p
e
rform
a
n
ce o
f
shu
n
t
activ
e
filter is b
a
sed
o
n
how
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
.
4
,
D
ecem
b
er
2
015
:
89
7 – 905
89
8
No
n Li
near
Loa
d
3 P
h
ase
AC
Source
I
B_
I
C_
I
B_
I
A
I
C_
I
A_
Harm
oni
c
Ex
traction
PW
M
Si
gnal
I
C
I
A
I
B
V
DC
qui
c
k
l
y
t
h
e ha
rm
oni
c com
ponent
s a
r
e ext
r
act
ed fr
om
t
h
e l
o
ad cu
rre
nt
. Fo
r t
h
e ext
r
act
i
on o
f
ha
r
m
oni
cs,
m
e
thods
suc
h
as Instanta
neous Reactive
powe
r th
e
o
ry, PQ t
h
eory
, sym
m
e
tr
ical com
pone
nt t
h
eory,
sy
nch
r
o
n
o
u
s
r
e
fere
nce
fram
e
theo
ry
(
S
RF)
,
In
stantane
ous
active and
rea
c
tive com
pone
n
t the
o
ry
[3],
PI a
nd
Fu
zzy lo
g
i
c b
a
sed
con
t
ro
ller [4
] are u
s
ed
for id
eal so
u
r
ce co
nd
itio
ns and
th
e SRF can
be u
s
ed
fo
r
u
n
b
a
l
a
n
ced
sou
r
ce v
o
l
t
a
ge
con
d
i
t
i
on i
n
whi
c
h spee
d o
f
refe
rence
fra
m
e
i
s
not
con
s
t
a
nt
due t
o
v
o
l
t
a
ge ha
rm
oni
cs and
un
bal
a
nce
.
In t
h
i
s
pa
pe
r a
m
odi
fi
ed SR
F
t
h
eo
ry
wi
t
h
E
M
R
FPLL i
s
us
ed f
o
r t
h
e e
x
t
r
act
i
on
of
ha
rm
oni
cs
. T
h
i
s
m
e
t
hod m
i
nim
i
zes t
h
e det
ect
i
o
n
er
ro
r
of t
h
e
p
h
ase a
ngl
e t
o
ge
nerat
e
qui
c
k
a
nd
p
r
eci
se r
e
fere
nce si
gnal
s
[
5
]
.
The P
W
M
si
g
n
al
s f
o
r t
h
e
V
o
l
t
a
ge S
o
urce
In
vert
e
r
(
V
S
I)
are ge
ne
rat
e
d
by
A
N
N
S
V
P
WM
de
pe
ndi
n
g
o
n
t
h
e
ext
r
act
ed
ha
r
m
oni
cs as sh
ow
n i
n
Fi
gu
r
e
1.
The
S
A
F pr
odu
ces cu
rr
en
t i
n
o
ppo
site d
i
r
ection to
th
e
disturba
nces t
o
cancel the
harm
onics and
makes the s
ource curre
nt sinusoi
d
al an
d f
r
ee fr
om
harm
oni
cs at
PCC.
Fi
gu
re
1.
B
l
oc
k
Di
ag
ram
of S
h
u
n
t
Act
i
v
e Fi
l
t
er
2.
MO
DIFIE
D
S
Y
N
C
H
R
O
N
O
U
S
REFERE
NCE
FR
A
M
E BASE
D PLL
For t
h
e ext
r
act
i
on o
f
ha
rm
oni
cs from
t
h
e l
o
ad cu
rre
nt
vari
o
u
s al
g
o
ri
t
h
m
s
have
been
pr
o
pos
ed i
n
t
h
e
literatu
re, th
e
m
o
st p
o
p
u
l
ar i
s
syn
c
h
r
o
nou
s reference
frame th
eo
ry
(SR
F
) an
d it h
a
s
ad
v
a
n
t
ag
e of ease of
ex
ecu
tion
,
robu
st perform
a
n
ce fo
r i
d
eal cond
itio
n
s
wh
en
t
h
e lo
ad
curren
t
is h
a
rm
o
n
i
cally less d
i
storted [6
]. In
unbalance
d
a
n
d distorted loa
d
cond
itions S
R
F
m
e
thod of harm
onic current
ext
r
action is less effective. T
o
t
r
o
unce t
h
ese
dra
w
backs
EM
R
FPLL
has
be
en p
r
op
ose
d
i
n
t
h
i
s
pa
per a
s
s
h
o
w
n i
n
Fi
g
u
r
e
2. T
h
e l
o
ad c
u
r
r
en
t
and s
o
urce
vol
tage are m
easured
by
m
eans of c
u
rrent
and
pote
n
tial transform
e
r. The m
e
asure
d
three phase
l
o
ad cu
rre
nt
ar
e con
v
ert
e
d i
n
t
o
t
w
o axi
s
st
at
i
onary
co
or
di
nat
e
sy
st
em
usi
ng C
l
arke t
r
a
n
sfo
r
m
a
t
i
on as
sho
w
n
in
equ
a
tion
s
(1) to (3
).
I
α
=
)
2
1
2
1
(
3
2
_
_
_
L
C
L
B
L
A
I
I
I
(1
)
I
β
=
L
C
L
B
I
I
_
_
2
3
2
3
3
2
(2)
Evaluation Warning : The document was created with Spire.PDF for Python.
I
J
PED
S
I
S
SN
:
208
8-8
6
9
4
Mod
ified
S
y
n
c
h
r
ono
us Referen
ce Frame
b
a
sed
Ha
rmo
n
i
c Extra
c
tio
n fo
r
Sh
un
t .... (CS
.
Su
ba
sh
K
u
ma
r)
89
9
(3
)
Fi
gu
re
2.
B
l
oc
k
Di
ag
ram
of
M
odi
fi
e
d
Sy
nc
hr
o
n
o
u
s R
e
fere
nce F
r
am
e based P
L
L
The
refe
ren
c
e
fram
e
s i
n
t
h
e
st
at
i
onary
c
o
or
di
nat
e
sy
st
e
m
are t
r
ansf
or
m
e
d t
o
t
h
e
sy
nch
r
on
o
u
sl
y
rotatin
g
refe
re
nce
fram
e
usin
g e
q
uations
(
4
)
an
d
(5
).
T
h
e c
u
r
r
ents
I
d
and
I
q
re
prese
n
t
t
h
e
di
rect
a
n
d
qua
drat
ure
axi
s
of t
h
e
di
st
ort
e
d l
o
a
d
c
u
r
r
e
nt
.
I
d
=
I
α
co
s
θ
+
I
β
sin
θ
(4
)
I
q
=
-I
α
sin
θ
+
I
β
cos
θ
(5)
Th
e two
ax
es syn
c
hrono
u
s
ly ro
tatin
g
reference fram
e
is p
a
ssed
th
rou
g
h
a lo
w pass filter to
filter o
u
t
t
h
e avera
g
e co
m
ponent
o
f
t
h
e
curre
nt
an
d pa
ss t
h
ro
u
gh t
h
e pul
sat
i
n
g com
p
o
n
e
n
t
of t
h
e c
u
r
r
ent
p
r
esent
i
n
t
h
e
lo
ad
cu
rren
t [7
]. Th
e filtered
h
a
rm
o
n
i
c curren
t
is tran
sfo
r
m
e
d
to
stati
o
n
a
ry referen
c
e fram
e
as g
i
v
e
n i
n
eq
u
a
tion
s
(6
)
an
d (7
)
I
adc
= I
d
1
co
s
θ
+
I
q
1
sin
θ
(6)
I
bdc
= -I
d
1
sin
θ
+
I
q
1
cos
θ
(7)
Th
e referen
ce
cu
rren
t in
two
ax
is is tran
sformed
in
to three
axes by re
vers
e Clarke trans
f
orm
a
tion as
gi
ve
n i
n
eq
uat
i
ons
(
8
)
t
o
(
1
0
)
.
I
sa
=
adc
o
I
I
2
1
3
2
(
8
)
I
sb
=
)
2
3
2
1
2
1
(
3
2
bdc
adc
o
I
I
I
(9
)
I
sc
=
)
2
3
2
1
2
1
(
3
2
bdc
adc
o
I
I
I
(1
0)
The c
u
rrent i
n
the three
phas
e
coordinates
represe
n
ts the
e
x
tracted ha
rm
onic content
of
the supply
currents a
n
d the current to
be
i
n
ject
ed t
o
t
h
e
VSI
de
pen
d
s
up
o
n
t
h
e
di
ffe
r
e
nce bet
w
een t
h
e ha
rm
oni
c cont
e
n
t
o
f
th
e si
g
n
a
l
and
th
e inj
ected cu
rren
t
of th
e activ
e filter.
3.
E
X
TE
NDE
D MR
F
PLL
The angle
Ɵ
f
o
r t
h
e Pa
rk a
n
d In
ve
rse pa
rk
t
r
ansf
orm
a
t
i
on i
s
gene
rat
e
d
by
t
h
e Phase
Loc
k
ed L
o
o
p
(PLL) wh
ich
syn
c
hron
izes th
e ou
tpu
t
signal in
frequ
e
n
c
y
and p
h
ase
wi
t
h
t
h
e i
n
put
. In
u
nbal
a
nce
d
an
d
I
q
1
I
d
1
I
bdc
I
adc
I
s
c
I
s
b
I
sa
Co
s
Sin
Ɵ
I
d
I
q
I
β
I
α
V
C
S
V
B
S
V
A
S
I
C
I
A
I
B
Clark
e
Transf
orm
a
t
i
on
Park
Transf
orm
a
t
i
on
LPF
LPF
Rev
e
rse
Park
Transfo
rm
ation
Rev
e
rse
Clark
e
Transfo
rm
ation
Extende
d M
R
F
PLL
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
.
4
,
D
ecem
b
er
2
015
:
89
7 – 905
90
0
di
st
ort
e
d c
o
n
d
i
t
i
ons t
o
e
n
a
b
l
e
fast
an
d ac
curat
e
phase
and freque
ncy detec
tio
n EMRFPLL is
used
for
obt
ai
ni
ng
t
h
e
i
n
st
ant
a
neo
u
s
i
n
f
o
rm
at
i
on
of
t
h
e p
h
ase
an
gl
e an
d t
h
e m
a
gni
t
ude
of t
h
e s
i
gnal
[8]
.
The
PL
L
d
e
tects th
e p
h
a
se, filters o
u
t
t
h
e un
wan
t
ed
si
g
n
a
l and
th
e
oscillatio
n
s
are p
r
od
u
c
ed
b
y
the v
o
ltag
e
co
n
t
ro
lled
o
s
cillato
r. Th
e in
pu
t
vo
ltag
e
s wh
ich
are u
n
b
a
lan
ced
ar
e t
r
ansform
e
d
in
to
dq
sync
hrono
us referen
c
e
frame.
V
abc
+
=
V
abc
–V
abc
-
.1
(1
1)
V
abc
-
=
V
abc
–
V
abc
+
.1
(1
2)
)
(
)
(
)
(
)
(
)
(
t
V
t
V
t
V
T
t
V
t
V
c
b
a
(13)
T
αβ
=
2
3
2
1
2
3
2
1
0
1
3
2
(14)
)
(
)
(
)
(
)
(
t
V
t
V
T
t
V
t
V
dq
q
d
(15)
T
dq
=
(1
6)
Th
e feed
back
l
o
op
regu
lates th
e dq
referen
c
e fram
e
. Th
e p
o
s
itiv
e sequ
en
ce an
d
n
e
g
a
tiv
e
sequ
en
ce
o
f
t
h
e harm
oni
cs
are prese
n
t
at
t
h
e i
n
p
u
t
.
In
EM
R
FPLL h
a
s t
w
o sy
nc
hr
on
o
u
s re
fere
nc
e fram
e
s rot
a
t
i
ng
i
n
o
ppo
site d
i
rectio
n
s
at th
e eq
u
i
v
a
len
t
an
gu
lar sp
eed
.
Th
e fun
d
a
m
e
n
t
al p
o
s
itiv
e and
n
e
g
a
tiv
e
seq
u
e
n
ce
com
pone
nt
s ar
e separat
e
d f
r
o
m
t
h
e fu
ndam
e
nt
al
fre
que
nc
y
[9]
.
Th
e er
r
o
r ca
use
d
by
t
h
e ne
gat
i
v
e se
que
nce
com
pone
nt
i
s
el
im
i
n
at
ed.
V
d
gives an e
s
tim
a
tion
of t
h
e am
plitude
of the
positive sequence
com
pone
nt and V
q
gi
ves i
n
f
o
rm
at
ion
of t
h
e
negat
i
ve seq
u
e
n
ce c
o
m
pone
nt
, b
o
t
h
r
o
t
a
t
i
ng i
n
o
p
p
o
si
t
e
di
rect
i
ons at
t
h
e eq
ui
val
e
nt
ang
u
l
a
r s
p
ee
d
of t
h
e i
n
put
si
gnal
.
Wh
en t
h
e fre
que
nci
e
s
are l
o
c
k
ed t
h
e
negat
i
v
e
seq
u
e
n
ce com
p
o
n
e
n
t
i
nput
v
o
ltag
e
app
e
ars as
d
i
stu
r
b
a
n
c
e in
pu
t to th
e PLL oscillatin
g
at twice th
e
fun
d
a
m
e
n
t
al frequ
en
cy.
LPF(s
)
=
p
s
p
(17)
(1
8)
(19)
PI(s
) =
[
K
p
+
K
i
]
/
2
(20)
4.
ARTIF
ICI
A
L
NEU
R
A
L
NETWOR
K BA
S
E
D SP
A
C
E V
E
CTOR
PW
M
PW
M signals
are use
d
for m
o
dulating t
h
e tim
e
dur
ation
of th
e pu
lses to p
r
o
v
i
d
e
a com
p
en
satio
n
cu
rren
t in
oppo
sitio
n
th
e cu
rren
t h
a
rm
o
n
i
cs p
r
esen
t at
PCC [1
0
]
. Th
e AN
NSVPW
M
measu
r
es th
e am
p
l
itu
d
e
and
an
gl
e
of t
h
e re
fer
e
nce
fr
am
e t
o
ge
nerat
e
t
h
e
gat
i
n
g
p
u
l
s
es. T
h
e m
easure
d
harm
on
i
c
com
pone
nt
of t
h
e
cu
rren
t is fed
as in
pu
t to
the ANNSVPWM wh
ich
acts
as a feed
forward neur
al netw
or
k
f
o
r
non
lin
ear
m
a
ppi
n
g
sy
st
em
. The i
n
p
u
t
i
s
con
v
ert
e
d i
n
t
o
di
rect
axi
s
and
qu
ad
rat
u
re
axi
s
com
pone
nt
s V
d
and V
q
. Th
e
m
a
gni
t
ude
an
d
an
gl
e are se
p
a
rat
e
d
by
C
a
rt
esi
a
n t
o
p
o
l
a
r
con
v
e
r
si
o
n
of
t
h
e t
w
o axi
s
c
o
m
pone
nt
s [
1
1
]
. The
ANN is u
s
ed
to
calcu
late th
e ti
m
e
p
e
riod
s fo
r th
e an
gu
lar
p
o
s
ition
of th
e d
i
fferen
t
sect
o
r
s in
t
h
e referen
ce
fram
e
. The com
put
at
i
onal
bu
rde
n
o
f
cust
om
ary
SVP
W
M
t
o
chec
k t
h
e l
ook
up t
a
bl
e f
o
r
cal
cul
a
t
i
ng t
h
e t
i
m
e
del
a
y
s
i
s
red
u
c
e
d by
usi
n
g A
N
N
S
V
P
W
M
[
12]
.
T
h
e neu
r
a
l
n
e
two
r
k is tr
ain
e
d u
s
i
n
g eq
uatio
n
s
15
–
21
an
d the
ANN m
odel is gene
rated a
n
d placed
using s
i
m
u
link. T
h
e tu
rn on tim
e for the three
pha
s
e
inputs are gi
ven
by
th
e fo
llowing
eq
u
a
tion
s
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
J
PED
S
I
S
SN
:
208
8-8
6
9
4
Mod
ified
S
y
n
c
h
r
ono
us Referen
ce Frame
b
a
sed
Ha
rmo
n
i
c Extra
c
tio
n fo
r
Sh
un
t .... (CS
.
Su
ba
sh
K
u
ma
r)
90
1
The t
i
m
e of a
p
pl
i
cat
i
on
of
t
h
e
di
f
f
ere
n
t
sect
o
r
s i
s
fo
u
n
d
usi
n
g
vol
t
a
ge
sec
o
nd
p
r
i
n
ci
pl
e. T
h
e sect
or i
s
detected
by c
o
m
p
aring the
angle with pre
d
ef
i
n
ed val
u
e
s
. Th
e
ne
ural
net
w
o
r
k
t
a
kes
*
as
re
fere
nce
volta
ge
p
o
s
ition
.
Mu
lt
ilayer n
e
uron
s
are
u
s
ed in
t
h
e
first an
d sec
ond
layer
and
t
h
e nu
m
b
er of
n
o
d
e
s, th
e
weigh
t
s are
set in
th
e train
i
n
g
stag
e as shown in
Tab
l
e
1
.
T
A-ON
=
5
;
*)
(
sin
3
sin
*
.
.
4
2
3
;
*)
(
sin
3
sin
*
.
.
4
2
2
;
*)
(
sin
3
sin
*
.
.
4
2
6
,
1
;
*)
(
sin
3
sin
*
.
.
4
4
*
*
*
*
S
V
K
T
t
T
S
V
K
T
t
to
T
S
V
K
T
t
T
S
V
K
T
T
s
a
o
s
b
o
s
b
o
s
o
(21)
T
A-ON
=
*)
(
*)
(
4
A
g
V
f
Ts
(22)
f(
V*
)
→
Volta
ge am
plitude s
cale factor
g(
α
*)
→
Tu
rn
o
n
Sign
al at
Un
it Vo
ltag
e
g
A
(
α
*)
=
5
;
*)
(
sin
3
sin
4
,
3
;
*)
(
sin
3
sin
2
;
*)
(
sin
3
sin
6
,
1
;
*)
(
sin
3
sin
*
*
*
*
S
K
S
K
S
K
S
K
(2
3)
T
B-ON
=
6
,
5
;
*)
(
sin
3
sin
*
.
.
4
2
4
;
*)
(
sin
3
sin
*
.
.
4
2
3
,
2
;
*)
(
sin
3
sin
*
.
.
4
4
6
,
5
;
*)
(
sin
3
sin
*
.
.
4
2
*
*
*
*
S
V
K
T
T
T
T
S
V
K
T
T
T
S
V
K
T
T
S
V
K
T
T
T
s
b
a
o
s
b
o
s
o
s
a
o
(
2
4)
g
B
(
α
*)
=
6
,
5
;
*)
(
sin
3
sin
4
;
*)
(
sin
3
sin
3
,
2
;
*)
(
sin
3
sin
1
;
*)
(
sin
3
sin
*
*
*
*
S
K
S
K
S
K
S
K
(2
5)
T
C-ON
=
6
;
*)
(
sin
3
sin
*
.
.
4
2
5
,
4
;
*)
(
sin
3
sin
*
.
.
4
4
3
;
*)
(
sin
3
sin
*
.
.
4
2
2
,
1
;
*)
(
sin
3
sin
*
.
.
4
2
*
*
*
*
S
V
K
T
T
T
S
V
K
T
T
S
V
K
T
T
T
S
V
K
T
T
T
T
s
b
o
s
o
s
a
o
s
b
a
o
(26)
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
.
4
,
D
ecem
b
er
2
015
:
89
7 – 905
90
2
g
C
(
α
*)
=
6
;
*)
(
sin
3
sin
5
,
4
;
*)
(
sin
3
sin
3
;
*)
(
sin
3
sin
2
,
1
;
*)
(
sin
3
sin
*
*
*
*
S
K
S
K
S
K
S
K
(2
7)
Tabl
e
1.
A
N
N
base
d S
V
P
W
M
T
r
ai
ni
n
g
1
st
Layer
12 Neuron (
L
og Sig)
Input Layer
2
nd
lay
e
r
12 N
e
uron (L
og Si
g)
H
i
dden lay
e
r
3
rd
Layer
3 N
e
uron (Purelin)
O
u
tput lay
e
r
T
r
aining E
r
ror
0.53 %
T
r
aining E
poch
9843
0
5.
RESULT AND DIS
C
USSI
ON
Th
e sim
u
latio
n
of
th
e
pr
oposed
system
is
v
a
lid
ated
b
y
mean
s o
f
Matlab
/ Sim
u
lin
k
f
o
r
a So
ur
ce
v
o
ltag
e
of
30
0 V
,
w
ith
a sou
r
ce in
du
ctor
of
2
0
0
m
i
cr
o
h
e
nr
y D
C
in
k
vo
ltag
e
of
400
V, lo
ad
r
a
ti
n
g
o
f
5
KW
non linea
r l
o
ad. T
h
e input to t
h
e system
is a three
ph
ase
pu
re sinu
so
i
d
al si
g
n
a
l as sh
own
in
Figur
e
3
.
Fi
gu
re
3.
S
o
u
r
ce V
o
l
t
a
ge
Th
e l
o
ad cu
rren
t du
e th
e
no
n lin
ear l
o
ad
is
d
i
sto
r
ted
as sho
w
n
i
n
Figu
re
4
.
Th
e inp
u
t
vo
ltag
e
t
o
the
sy
st
em
i
s
si
nusoi
dal
w
h
ereas
t
h
e cur
r
ent
wa
ve fo
rm
of t
h
e t
h
ree pha
ses are no
n si
n
u
s
o
i
d
al
and c
onse
q
uent
l
y
the reactive power c
o
nsum
ed by
the load is
increase
d
, the
power
factor
of t
h
e sy
st
em
is red
u
ce
d [
13]
.
Thi
s
leads to poor e
fficiency t
o
the
va
rious
loa
d
s
connected to the powe
r syste
m
.
Fi
gu
re
4.
Loa
d
cu
rre
nt
The de
form
ed load c
u
rrent
imp
licates th
e same in
th
e source current.
In
orde
r
t
o
av
oi
d t
h
i
s
d
r
aw
bac
k
shunt active fi
lter is connected at
th
e po
in
t
o
f
co
mm
o
n
co
up
ling
[1
4
]
.
Th
e non
sinu
so
id
al lo
ad
curren
t is
measured a
n
d the
harm
onic c
onte
n
t
of the l
o
ad c
u
rre
n
t
is ex
tracted and the co
m
p
en
sation
cu
rren
t is
g
e
n
e
rated
b
y
u
s
ing
ANNSV
P
W
M
techniq
u
e
.
Th
e co
mp
ensatio
n
cu
rren
t
o
f
th
r
ee
p
h
ases to
b
e
inj
e
cted
in
to
th
e syste
m
to
cancel
t
h
e
ha
r
m
oni
cs out
of
pha
se i
s
s
h
ow
n
i
n
Fi
gu
re
5.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
J
PED
S
I
S
SN
:
208
8-8
6
9
4
Mod
ified
S
y
n
c
h
r
ono
us Referen
ce Frame
b
a
sed
Ha
rmo
n
i
c Extra
c
tio
n fo
r
Sh
un
t .... (CS
.
Su
ba
sh
K
u
ma
r)
90
3
Fig
u
re
5
.
Co
mp
ensatio
n Curren
t
The com
p
ensat
i
on c
u
r
r
ent
i
s
i
n
ject
e
d
at
t
h
e
poi
nt
of c
o
m
m
on c
o
upl
i
n
g a
n
d t
h
e
harm
oni
cs prese
n
t
i
n
th
e lo
ad
cu
rrent are elimin
ated
an
d
t
h
e
so
urc
e
cu
rre
nt
i
s
m
a
de si
nus
oi
dal
i
n
s
h
a
p
e as
sh
o
w
n
i
n
Fi
g
u
re
6
.
Fi
gu
re
6.
S
o
u
r
ce C
u
r
r
e
n
t
In Figure 7
the
com
p
arison of
s
o
u
r
ce curren
t to
lo
ad
cu
rren
t is
s
h
own for
one phase of
t
h
e
three
pha
se sy
st
em
. The l
o
ad
cu
rre
nt
w
h
i
c
h
i
s
no
n
si
n
u
soi
d
i
s
m
a
de si
nus
oi
dal
i
n
s
h
a
p
e.
Fi
gu
re
7.
S
o
u
r
ce C
u
r
r
e
n
t
Vs
Loa
d
C
u
rre
nt
The c
o
m
p
ari
s
o
n
o
f
p
o
w
er
fac
t
ors
of a c
o
n
v
e
nt
i
onal
SR
F PLL wi
t
h
a E
M
R
FPLL (Fi
g
ure
8) s
h
ow
n
t
h
at
by
usi
n
g t
h
e
pr
o
pose
d
t
e
chni
que
t
h
e
o
v
e
ral
l
p
o
we
r
fac
t
or
of
t
h
e
sy
st
em
i
s
im
prove
d.
Fig
u
re
8
.
Co
mp
ariso
n
of
PF
with
co
nv
en
tion
a
l PLL and
EMRFPLL
The
objective
of t
h
e
proposed
work is to re
duce the
THD bel
o
w the st
anda
rds
specified
by
p
r
o
f
ession
al
bo
d
i
es for p
o
wer
qu
ality
stan
dard
s. Co
m
p
ariso
n
of THD
for
co
nv
en
tio
n
a
l
v
s
p
r
o
p
o
s
ed
sy
ste
m
is
sho
w
n i
n
Ta
bl
e 2,
an
d c
o
m
p
ari
s
o
n
of
PF
f
o
r
con
v
e
n
t
i
onal
vs
pr
o
pose
d
sy
st
em
i
s
sho
w
n
i
n
Ta
bl
e 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
.
4
,
D
ecem
b
er
2
015
:
89
7 – 905
90
4
Tab
l
e
2
.
C
o
m
p
ar
ison
o
f
THD
f
o
r
C
o
nv
en
tio
nal v
s
Prop
osed System
Conventio
nal SRF PL
L
Modif
i
ed SRF with EMRFP
L
L
Phase
Before Co
m
p
ensa
tion
After
co
mpensa
ti
o
n
Before Co
m
p
ensa
tion
After
co
mpensa
ti
o
n
R Phase
24.55 %
0.86 %
25.57 %
0.82 %
Y Phase
25.04 %
0.85 %
24.66 %
0.83 %
B Phase
24.78 %
0.85%
24.93 %
0.83%
Tabl
e
3. C
o
m
p
ari
s
o
n
of
PF
f
o
r C
o
n
v
ent
i
o
nal
vs
P
r
o
p
o
sed
Sy
st
em
Conventio
nal SRF PL
L
Modif
i
ed SRF with EMRFP
L
L
PF
Before Co
m
p
ensa
tion
After
co
m
p
ensati
on
Before Co
m
p
ensa
tion
After
co
m
p
ensati
on
PF Va
lues
0.72
0.9857
0.72
0.996
6.
CO
NCL
USI
O
N
A n
o
v
el
co
nt
r
o
l
t
echni
q
u
e f
o
r
det
ect
i
on
o
f
harm
oni
cs by
EM
R
FPLL i
s
pr
o
pose
d
fo
r
a t
h
ree
pha
se
th
ree
wire syste
m
in
th
is p
a
per. Th
e ex
tracted
h
a
rm
o
n
i
cs from
the three
pha
se loa
d
currents a
r
e gi
ven to
a
m
odi
fi
ed A
N
N
S
V
P
W
M
t
o
gene
rat
e
t
h
e c
o
m
p
ensat
i
ng c
u
r
r
ent
s
fo
r t
h
e
di
st
ort
i
o
ns
. T
h
e sy
st
em
i
s
m
odel
l
e
d
and
si
m
u
l
a
t
e
d usi
n
g M
a
t
l
a
b /
Sim
u
l
i
nk a
nd t
h
e m
easured
h
a
rm
oni
c l
e
vel
s
as gi
v
e
n i
n
t
a
bl
e s
h
o
w
s t
h
e TH
D
of
EM
R
FPLL
i
s
su
peri
or
com
p
are
d
t
o
c
o
n
v
e
n
t
i
onal
PLL
[
1
5]
-[
1
7
]
.
T
h
e E
M
R
FPLL i
s
us
ed t
o
t
r
ac
k t
h
e
pha
se
and
fre
q
u
ency
i
n
f
o
rm
at
i
on an
d ge
nerat
e
acc
urat
e val
u
es
of
angl
e f
o
r t
h
e t
r
ans
f
orm
a
t
i
on i
n
SR
F t
h
e
o
ry
.
The
propose
d
res
u
lts
are use
d
t
o
validate
th
e effe
ctiveness
of the control technique.
REFERE
NC
ES
[1]
Naim
ish Zaver
i
,
Ajitsinh Chud
a
s
am
a, “
C
ontrol
strateg
i
es
for h
a
rm
onic m
itiga
t
i
on and power
f
actor
corr
ect
ion
using shunt active filter under v
a
rious source vo
ltag
e
conditions”,
Electrical Po
wer and Energy Systems
, vol. 4
2
,
pp. 661–
671, 20
12.
[2]
R. Zah
i
ra, A. Peer Fathim
a, “A
Techn
i
cal Survey
on C
ontro
l Str
a
teg
i
es of Activ
e Filter for Harm
onic Suppression”,
Pr
ocedia
Engin
e
er
ing
, vo
l. 30, p
p
. 686
– 693
, 20
12.
[3]
Abdelm
adjid Ch
aoui, e
t
al
., “
D
PC controll
ed thr
ee-phase
act
ive
filte
r for power qualit
y im
prove
m
e
nt”,
El
ec
trica
l
Power and
Ener
gy Systems
, vol.
30, pp
. 476–485
, 2008.
[4]
Suresh Mikkili,
AK. Panda, “Real-tim
e
im
plem
entation o
f
PI
an
d fuzzy
log
i
c
co
ntrollers
based shunt active f
ilt
er
control strategies for po
wer quality
improvement”,
Electrical Po
wer and Energy
Systems
, vol. 43, pp. 1114–1126,
2012.
[5]
Lin Xu,
et
al
.,
“A Novel Control Strat
e
g
y
for
D
y
nam
i
c Vol
t
ag
e Restor
er usin
g Decoupl
ed Multipl
e
Ref
e
ren
c
e
Fra
m
e
PL
L (DMRF-PL
L
)
”
,
W
S
EAS Transactio
ns On Systems
, v
o
l/issue: 8
(
2), 20
09.
[6]
Metin Kesler
, Engin Ozdemir
,
“Sy
n
chronous-Refer
en
ce-Frame-Based Control Method for UPQC
Under
Unbalanced and
Distorted Load
Conditions”,
IEEE Transactions On Indu
strial Electronics
, vol/issue: 58(9), pp
.
3967–3975, 201
1.
[7]
CS. Subash Ku
mar, V. Gopalakrishna
n, “Design of sy
n
c
hrono
us referenc
e frame based harmonic detection and
space ve
ctor pul
se-width m
odulation based swit
ching of shunt act
ive fil
t
er
”,
Australian Journal of Electrica
l
&
Electronics Engineering
, vol/issue: 10(3)
, 2013
.
[8]
Saeed Golestan
, et al
., “
D
esign-Oriented Stud
y
of Adva
nced S
y
nchronous refere
nce Fram
e Phase-Lock
ed Loops”
,
IEEE Transactio
ns
on Power
Electronics
, vol/issue: 28(2)
, pp
. 76
5–778, 2013
.
[9]
Xiao-Qiang GU
O, et al., “Phase
locked
loop and s
y
nchronization methods
for
gr
id
interfaced
conv
erters: a
review”,
Prze
gl
ą
d E
l
ek
tr
otechn
i
cz
ny (
E
l
e
c
t
r
i
cal
Re
vi
ew)
, R. 87
NR 4, 2011.
[10]
Tar
ı
k
Erfid
a
n,
Erhan Bu
tun, “
L
ow Cost Implementation
of
Artificial Neural Network B
a
sed Space Vector
Modulation”, Sp
ringer, LNCS, v
o
l. 3611
, pp
. 204
– 209
, 2011
.
[11]
Avik Bhattachar
y
a
, Chand
a
n Ch
akraborty
, “A Sh
unt Ac
tive Power Filter With En
ha
nced Performance Using ANN-
Based Predi
c
tiv
e
and Ad
aptiv
e C
ontrolle
rs”,
IEEE Transactions
On
Industrial Electronics
, vo
l/issue: 58(2)
, 2011
.
[12]
Abdul Ham
i
d Bhat, Pr
am
od Agarwal, “
A
n Artifici
a
l-Neura
l
-Net
work-Based Space Ve
ctor PW
M of a
Three-p
h
as
e
High Power Fac
t
or Convert
er fo
r Power Qualit
y
Im
provem
e
nt
”,
Proceedings of I
ndia Int
e
rna
tion
a
l Confer
enc
e
o
n
Power Electron
ics, 2006
.
[13]
Peng Xiao,
et
al.
,
“
M
ultipl
e
Referen
ce Fram
e-Based Contro
l of Thre
e-Phas
e PW
M Boost
Rect
ifiers und
er
Unbalanced and Distorted
Inpu
t Conditions”,
IEEE Transactions
On
Power
Electronics
, vol/issue: 23(4)
, 2008
.
[14]
Rondineli Rodrigues Pereira, Carlos Henr
ique da Silva, Luiz Edu
a
rdo Borges
da Silva, Germano Lambert, “Contr
o
l
of PWM rectifier under grid v
o
ltag
e
dips”, Bu
lletin Of
The Polish Academ
y
Of Sciences, Technical Sciences
,
vol/issue: 57(4), 2009.
[15]
Safa Ahm
e
d, et al
., “Real
Ti
m
e
Control of
an
Act
i
ve Po
wer Filter
under Distorted Vol
t
age Cond
ition
”
,
International Jo
urnal of
Power
Elec
tronics and
Drive
System (
I
JPEDS)
, vol/issu
e: 2(4)
,
pp. 424-
433, 2012
.
[16]
Bouzidi Mansou
r, e
t
a
l
.
,
“
S
lidin
g m
ode Control using 3D-SVM for Three-pha
s
e
F
our-Leg S
h
u
n
t Activ
e F
i
l
t
er
”
,
International Jo
urnal of
Power
Elec
tronics and
Drive
System (
I
JPEDS)
, vol/issu
e: 3(2)
,
pp. 147-
154, 2013
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
J
PED
S
I
S
SN
:
208
8-8
6
9
4
Mod
ified
S
y
n
c
h
r
ono
us Referen
ce Frame
b
a
sed
Ha
rmo
n
i
c Extra
c
tio
n fo
r
Sh
un
t .... (CS
.
Su
ba
sh
K
u
ma
r)
90
5
[17]
Zhenfeng
Xiao,
et a
l
.
,
“
A Shunt Active Power Filter with Enh
a
nced
D
y
n
a
mic Per
f
ormance using
Dual-Repetitiv
e
Controller
and P
r
edic
tive Com
p
e
n
sation
”,
International Journal of Power El
ectron
i
cs and Drive S
y
stem (
I
JPEDS)
,
vol/issue:
3(2),
pp. 209-217
, 20
13.
BIOGRAP
HI
ES
OF AUTH
ORS
C.S
.
S
ubas
h
Kum
a
r rece
ived h
i
s
B.E degr
ee
in
Electrical and
Electronics Eng
i
neer
ing from
Kumaraguru College of
Techno
lo
g
y
in 2002
. He r
eceived h
i
s M.E
degree
in
Electrical
Machin
es
from PSG College of Technolog
y
in 2008. Currently
,
he is an Assistant Professor (Sl.Gr) in the
Departm
e
nt of
Ele
c
tri
cal
and
Ele
c
troni
cs Eng
i
neer
ing a
t
PSG Institute
of
Techno
log
y
and
Applied Resear
c
h
, Coim
batore,
Tam
ilnadu
,
Indi
a.
His
current res
earch in
ter
e
s
t
s
include power
electronics applications
in power sy
stem and active power f
ilters for power conditioning. He is
the
Life
M
e
m
b
er
of IS
TE
and
m
e
m
b
er of IE
EE
.
Dr.V.Gopalakr
ishnan, Associate
Professor, Government Colleg
e of
Technolog
y
,
Coimbator
e
com
p
leted h
i
s B.E. (
E
l
ectr
i
c
a
l a
nd Ele
c
troni
cs
Engineering) degree in Shanmu
gha Colleg
e
of
Engineering (Bh
a
rath
idasan University
),
Thanjavur
in the
y
ear
1989 and subsequently
jo
ined as
Lecturer
in Alagappa Chettiar
College of
Eng
i
neer
ing and
Technolog
y
,
K
a
raikudi He was
transferred
to Government College of Engin
eerin
g, Tirun
e
lveli in
1996. In the
y
e
ar 2000 he was
transferred
to
Governm
e
nt Colleg
e
of Te
chno
log
y
, Coim
bator
e
and l
a
ter h
e
com
p
leted h
i
s
M.E.(Computer
Science and
Engineering) degr
ee at Govern
ment College
of Technolog
y
,
Coimbatore in the
y
ear 2003. H
e
completed his
PhD in the y
ear
2009 in the field
of Compute
r
Networks
His
res
earch
ar
eas
i
n
clude
Network
S
ecurit
y
and
Com
puter Netw
orks
, He has
publis
hed thre
e
Res
earch ar
ti
cl
es
in reputed I
ndi
an and Inter
n
ation
a
l Journals and fourteen
res
earch
arti
cles
in nationa
l an
d interna
tiona
l conferen
ces
. He is
a life m
e
m
b
er of Indian
S
o
ciet
y for Tech
nica
l
Edu
c
a
tion (IS
TE)
and
Institution of
Engin
e
ers in India (I
EI).
Evaluation Warning : The document was created with Spire.PDF for Python.