Int
ern
at
i
onal
Journ
al of
P
ower E
le
ctr
on
i
cs a
n
d
Drive
S
ystem
s
(
IJ
PEDS
)
Vo
l.
1
2
,
No.
3
,
Septem
be
r
202
1
, pp.
163
2
~
164
3
IS
S
N:
20
88
-
8694
,
DOI: 10
.11
591/
ij
peds
.
v
1
2
.i
3
.
pp
163
2
-
164
3
1632
Journ
al h
om
e
page
:
http:
//
ij
pe
ds
.i
aescore.c
om
LCL
filt
er
design fo
r grid
-
connect
ed si
ngle
-
phase fly
back
microin
verte
r: a s
tep by st
ep gui
de
Muham
ad
Fai
zal Ya
ak
u
b
1
, Mo
hd Amr
an
Mohd
R
ad
zi
2
, Ma
as
p
aliza
Azr
i
3
,
Faridah
Ha
nim
Moh
d
Noh
4
1
Fakult
i Te
kno
lo
gi
Kejur
u
te
r
aa
n
El
ektrik
d
an
El
e
ktroni
k,
Univer
s
it
i Te
kn
ika
l
Mal
aysia
Me
la
k
a, Ma
la
ysi
a
1,2
Facul
ty
of Eng
ine
er
ing
,
Univer
siti
Putra
Mal
aysia
,
Mal
aysia
3
Fakult
i
K
ej
urut
era
an
E
le
ktr
ik,
Univer
siti
Te
kn
i
kal
Ma
la
ysia
Me
la
ka
,
Ma
la
ysia
4
Facul
ty
of Engi
nee
ring
T
ec
hnol
ogy
,
Univer
si
ti
Tun
Hus
sei
n
On
n
Mala
ysi
a, Ma
l
aysia
1,2
Advanc
ed Li
g
htni
ng,
Pow
er
an
d
Ene
rgy
Rese
ar
ch
Cen
tre,
Univ
e
rsiti
Putr
a
Mal
ay
sia
Art
ic
le
In
f
o
ABSTR
A
CT
Art
ic
le
history:
Re
cei
ved
J
un
7
, 20
21
Re
vised
Ju
l
12
,
20
21
Accepte
d
Auf
11
, 20
21
Rec
en
tl
y,
LC
L
h
as
be
com
e
am
o
ngst
th
e
most
attra
c
ti
ve
fi
lter
us
ed
for
gr
id
-
conne
c
te
d
f
lyback
inve
r
te
r
s
.
No
net
he
le
ss
,
th
e
sw
it
chi
ng
of
pow
er
device
s
in
the
inve
r
te
r
con
figura
t
ion
creat
e
s
har
mon
ic
s
that
aff
e
ct
th
e
end
appl
i
ca
t
ion
beha
vior
and
mi
ght
shorten
it
s
l
ife
t
ime.
Furtherm
or
e,
th
e
resona
n
ce
fre
quencie
s
pro
duce
d
by
the
LC
L
net
work
cont
ribu
te
to
the
sys
te
m
insta
bility
.
Thi
s
pap
er
proposes
a
st
ep
-
by
-
step
guide
to
design
i
ng
an
LC
L
fil
ter
by
consid
e
ring
seve
r
al
k
ey
aspe
ct
s
such
as
t
he
resona
n
ce
fr
e
quenc
y
and
ma
xim
u
m
cur
r
ent
r
ippl
e
.
A
single
-
phas
e
grid
-
connect
ed
flyb
ac
k
mi
cro
inv
erter
w
it
h
an
LC
L
fi
lter
was
d
esigned
the
n
constru
c
te
d
in
the
MA
TL
AB
/Sim
uli
nk
env
ironm
en
t
.
Seve
ral
d
iffe
r
e
nt
par
a
meter
v
ar
ia
ti
ons
and
dam
ping
soluti
o
ns
were
used
to
ana
ly
ze
th
e
p
erf
orma
nc
e
of
the
ci
rcu
it
.
The
s
im
ula
t
ion
result
show
s
a
prom
i
sing
total
har
mo
nic
distortion
(
T
HD
)
val
ue
bel
ow 5% and
h
arm
oni
c
suppres
sion u
p
to
14%.
Ke
yw
or
d
s
:
LCL
filt
er
Gr
i
d
c
onnected
Flyb
ac
k
micr
oin
ve
rter
This
is an
open
acc
ess arti
cl
e
un
der
the
CC
BY
-
SA
l
ic
ense
.
Corres
pond
in
g
Aut
h
or
:
M
uha
mad
Faiz
al
Y
aak
ub
Faculty
of E
ngineerin
g
Un
i
ver
sit
i P
utr
a M
al
a
ys
ia
Ser
dang, Sel
an
gor, Mala
ys
ia
Emai
l:
muham
adf
ai
zal
@
utem
.edu.m
y
1.
INTROD
U
CTION
In
rece
nt
dec
ades,
the
de
m
and
f
or
e
nerg
y
has
e
scal
at
ed
e
xpone
ntial
l
y
dr
i
ven
by
the
gro
wing
popula
ti
on
a
nd
increasi
ng
we
al
th
across
the
globe.
He
nce,
g
ree
n
e
nerg
y
harvested
f
rom
an
i
nexha
ust
ible
so
urce
s
uc
h
as
so
la
r
,
wi
nd,
a
nd
ti
de
hav
e
at
tract
ed
e
norm
ous
a
tt
entio
n
[1],
[2]
.
Since
t
he
distrib
uted
powe
r
gen
e
rati
on
s
yst
ems
became
mainstrea
m,
gri
d
-
c
onnected
microi
nv
e
rters
tur
ned
t
o
be
a
researc
h
ho
t
sp
ot
.
I
n
mo
st
cases,
t
he
gr
id
-
c
onnect
ed
micr
oinvert
er
is
co
ntro
ll
e
d
by
a
pulse
width
mod
ula
ti
on
(
PWM
)
s
witc
hing
te
chn
i
que
w
hich
gen
e
rates
harmo
nics
an
d
re
quires
a
lo
w
pa
s
s
filt
er
to
el
imi
nate
the
unwa
nt
ed
hi
gh
-
f
re
quency
sign
al
s
f
rom
ge
tt
ing
th
r
ough
into
t
he
distrib
ution
s
ys
te
m.
Numer
ous
a
pp
ro
ac
hes
of
l
ow
pas
s
filt
ers
ha
ve
bee
n
pro
po
se
d
a
nd
impleme
nted
s
uch
a
s
L
-
,
LC
-
,
a
nd
LCL
-
f
il
te
r.
A
mon
g
th
ese
topolo
gies,
the
LCL
-
t
ype
ha
s
sh
ow
n
treme
ndous
at
te
nuat
ion
of
ha
rm
on
i
c
su
pp
ressio
n
com
par
e
d
to
ot
her
s
[3]
.
Unf
ort
un
at
el
y,
t
he
phys
ic
al
filt
er
,
as
s
hown
in
Fi
gure
1
,
is
im
possib
le
to
be
c
on
st
ru
ct
e
d
us
in
g
t
he
t
heoreti
cal
ly
cal
c
ulate
d
va
lues.
Ther
e
f
or
e,
se
ver
al
pa
ramet
ers
ha
ve
t
o
be
reconsi
dered,
especial
ly
in
re
gards
to
t
he
i
nducta
nce
a
nd
capaci
ta
nce
val
ue
that
va
ry du
e to
numer
ous
factors s
uch as
com
pone
nts a
gi
ng
.
Evaluation Warning : The document was created with Spire.PDF for Python.
In
t J
P
ow Elec
& Dri S
ys
t
IS
S
N:
20
88
-
8
694
LCL fi
lt
er d
esi
gn for
g
ri
d
-
c
on
nected
sin
gle
-
phase
fl
yb
ack
m
ic
ro
invert
er
… (
M
uhama
d Fa
izal Y
aa
k
ub
)
1633
(a)
(
b)
Figure
1:
LCL
filt
er (
a)
w
it
hout d
a
mp
i
ng r
es
ist
or
a
nd
(
b) w
i
th a
dam
ping
r
esi
stor
Since
t
he
gri
d
side
i
nducto
r
L
2
of
a
filt
er
is
c
onnected
in
se
ries
to
the
gr
i
d,
the
in
duct
ive
impeda
nce
of
a
l
ong
gri
d
c
able
a
nd
lo
w
powe
r
t
ran
s
f
ormers
ma
y
ha
ve
infl
ue
nced
pa
r
amet
er v
ariat
io
n
i
n
L
2
.
A
dd
it
ion
al
ly
,
t
he
inducta
nce m
ay
al
so
be
va
ried
with
it
s
rat
e
values
up
t
o
75%
due
to
th
e
cor
e
’s
ma
gn
et
ic
per
mea
bili
ty
on
a
current
flo
wing
in
the
net
work
[
4]
.
T
he
ca
pa
ci
ta
nce
tolera
nce
is
i
n
t
he
ra
ng
e
of
±
5%
t
o
±20%
,
de
pe
ndin
g
o
n
their
diele
ct
ric
pro
per
ti
es.
T
he
se
var
ia
ti
ons
cause
r
eso
na
nc
e
fr
e
quenc
y
s
hi
fting
a
nd
subs
equ
e
ntly
res
ulti
ng
in
instabil
it
y
to
t
he
c
ontrol
s
ys
te
m
a
nd
the
filt
er
beh
a
vi
or
.
A
great
numb
e
r
of
w
orks
hav
e
bee
n
car
ried
ou
t
t
o
deal
with
t
he
r
eso
nan
ce
a
nd
inst
abili
ty
suc
h
as
the
pas
sive d
am
ping
ap
pro
ach
[
5]
–
[7]
,
ac
ti
ve
dam
ping metho
d
[8]
–
[
10]
,
an
d
e
ven
dam
pingle
ss
met
hod
[
4],
[
11]
to
achie
ve
the
sta
bili
ty
crit
erion
base
d
on
va
rio
us
ty
pes
of
con
t
ro
l
te
ch
niques.
N
onet
hele
ss,
t
her
e
is
ve
r
y
li
mit
ed
li
te
ratur
e
that
el
a
borates
a
cl
ear
st
ep
-
by
-
ste
p
guide
t
o
choose
ap
pro
pri
at
e
passi
ve
el
ement
values
f
or
the
filt
er
,
e
s
pecial
ly
of
the
op
e
n
-
lo
op
ty
pe
.
T
his
pap
e
r
presents
the
ste
pw
ise
de
sign
procedu
re
of
a
n
open
-
lo
op
LCL
filt
er
for
a
gr
i
d
-
t
ie
d
mic
ro
i
nv
e
r
te
r
[
12]
,
[13
]
.
T
he
pro
po
se
d
design
ste
ps
are
ba
sed
on
the
c
urre
nt
rip
ple
minimi
zat
ion
a
ppro
ac
h.
It
is
extrem
el
y
im
port
ant
t
o
regulat
e
the
ou
tpu
t
c
urre
nt
an
d
vo
lt
a
ge
with the
ri
pp
le
val
ue
as
mi
nim
um
as
possi
ble
w
hi
ch
c
ompli
es w
it
h
the
value
set
by
t
he
sta
ndar
ds
[
14]
,
[15]
.
I
n
a
nutshell
,
this
pa
per
c
on
tri
bu
te
s
to
a
sim
ple
a
nd
orga
nized
s
te
p
of
desig
ning
t
he
LCL
fil
te
r.
O
nl
y
the
esse
ntial
mathemat
ic
al
e
qu
at
io
n
f
or
t
he
init
ia
l
theor
et
i
cal
cal
culat
ion
with
numerical
desi
gn
e
xam
ple
s
a
r
e
prese
nted.
T
he
e
quat
ions
a
nd
the
us
age
a
r
e
sta
te
d
i
n
t
he
desig
n
flo
wc
ha
rt.
An
impro
vised
de
sign
flo
w
with
an
ad
diti
on
al
resona
nce
che
ck
-
l
oop
an
d
pa
rameter
c
ompli
ance
te
st
is
propose
d
in
this
pa
per.
The
key
par
a
mete
rs
obta
ine
d
f
rom
the
th
e
or
et
ic
al
cal
culat
ion
an
d
par
a
mete
r
co
mp
li
a
nce
te
st
wer
e
simulat
e
d i
n
the
Sim
ulin
k plat
form
a
nd
the r
es
ults are
discusse
d.
The
pa
per
is
orga
nized
as
f
ollow
s;
t
he
filt
er’
s
ess
entia
l
paramet
ers
us
e
d
i
n
t
he
desig
n
ar
e
desc
ribe
d
and
pr
e
sente
d
in
Sect
ion
2,
t
og
et
her
with
the
co
r
res
pondi
ng
mathemat
ic
al
equ
at
io
ns
.
T
he
desi
gn
flo
w
/st
eps,
numerical
desi
gn
e
xam
ple,
a
nd
value
sel
ect
ion
rati
onal
it
y
are
viv
i
dly
ex
plained
i
n
Sect
ion
3.
Subse
qu
ently,
simulat
ion
res
ults
are
e
xtens
ively
discuss
e
d
in
Sect
io
n
4,
i
nclu
ding
t
he
ef
fect
on
pa
ramet
er
va
riat
ion
to
emulat
e real
sc
enar
i
o
c
onditi
ons. Fi
nally
, a
c
on
cl
us
io
n
is
pr
ov
i
ded at t
he
e
nd of t
he
arti
cl
e.
2.
THE
FILT
ER
D
ESI
GN
2.1.
Cu
rren
t
ri
p
ple
The
hi
gh
-
fr
e
quenc
y
c
urren
t
rip
ple
at
te
nu
at
ion
of
the
filt
er
is
pr
ima
rily
i
nf
l
uen
ce
d
by
t
he
LC
pa
rt.
The
i
nducta
nc
e
pla
ys
a
sig
nificant
r
ole
in
con
t
ro
ll
in
g
t
he
cu
rr
e
nt
rip
ple
an
d
s
uppresse
s
the
lo
w
-
f
requen
c
y
harmo
nics
sp
e
ct
ru
m
[
5],
[
16]
.
As
su
c
h,
the
i
nducto
r
has
to
be
pro
per
l
y
de
sign
e
d
by
co
nsi
der
in
g
the
al
lo
wab
le
current
rip
ple
tolerance
.
T
he
pea
k
-
to
-
pea
k
cu
rr
e
nt
rip
ple
of
one
par
ti
c
ular
filt
er
f
or
the
un
i
po
la
r
P
W
M
inv
e
rter
ca
n
be
cal
culat
ed
ba
sed
on
E
quat
ion
(1).
T
he
mag
nitud
e
dis
tribu
ti
on
is
il
lustrate
d
i
n
Fi
gure
2
wh
e
rein
the
va
lue is
var
ie
d
de
pendin
g o
n
the
curre
nt phase
and the
mod
ul
at
ion
in
de
x
.
∆
(
)
=
2
(
1
−
sin
)
sin
(1)
w
ith
V
i
is
t
he
i
nput
DC
volt
age,
L
is
the
in
duct
ance,
f
sw
is
t
he
s
witc
hi
ng
f
r
equ
e
nc
y,
an
d
m
x
is
the
m
odul
at
ion
ind
e
x.
The
ma
ximum
fluctuat
ion
,
1
,
of
the
L
1
current a
ppears
when
sin
ω
o
t =
1
/(
2m
x
)
,
and
de
fine
d
as
1
=
8
1
(2)
Evaluation Warning : The document was created with Spire.PDF for Python.
IS
S
N
:
2088
-
8
694
In
t J
P
ow
Ele
c
&
Dr
i
S
ys
t
,
V
ol
.
12
, N
o.
3
,
Se
ptembe
r
202
1
:
163
2
–
164
3
1634
Figure
2:
Cu
rrent ri
pp
le
mag
nitud
e
d
ist
ri
buti
on
w
he
re
L
1
is
th
e
in
ver
te
r’
s
gr
i
d
side
in
du
ct
a
nce
,
I
1
is
t
he
r
at
ed
c
urre
nt
,
a
nd
β
c
is
the
in
du
ct
or
c
urren
t
rip
ple
coeffic
ie
nt
w
hi
ch
is
def
i
ned
a
s
I
rpmax
/I
1
.
T
yp
i
cal
ly
,
β
c
is
set
in
the
ra
ng
e
of
5
%
to
30%
[
5],
[
17]
–
[
20]
a
nd
with
so
me
p
a
ramete
rs
s
ubsti
tuti
on
,
the
in
du
ct
or m
aximum c
urre
nt
r
ip
ple
is
e
xpr
essed
a
s
:
=
8
1
(3)
Althou
gh
the
main object
ive
is
to
pro
duce
a
small
er r
ip
ple cu
r
ren
t
,
a smal
le
r
ri
pp
le
c
urre
nt
will
le
ad
to
slow
e
r
powe
r
dev
ic
e
switc
hing
duri
ng
co
nducti
on.
Subse
quently
,
it
would
inc
r
ease
the
i
nduc
tor
siz
e
an
d
ge
ner
at
e
higher
core lo
s
ses
[
21]
.
T
hese
are
t
he
c
halle
nges a
nd tra
des off to
b
e
d
eci
de
d upo
n desig
ni
ng
t
he fil
te
r.
2.2.
Inv
er
ter Side
Induc
tance
The
in
duct
anc
e
of
the
LC
L
filt
er
is
cr
ucial
and
i
mpose
s
a
sig
nificant
impact
on
t
he
harmo
nic
su
pp
ressio
n
an
d
c
urre
nt
ri
pp
l
e
co
ntr
ol.
It
li
mit
s
the
c
urre
nt
ri
pp
le
pro
duced
by
the
in
ver
te
r
s
witc
hing.
At
a
par
ti
cula
r
s
witc
hing
fr
e
quenc
y,
the
volt
age
-
current
rati
o
is
tracea
ble
onl
y
by
t
he
i
nv
e
rte
r
in
duct
ance
.
Durin
g
switc
hing
ti
me
T
sw
,
the
i
n
ve
rter
ave
ra
ge
vo
lt
age
is
c
on
si
dered
c
on
sta
nt
w
he
nev
e
r
th
e
f
un
dame
ntal
fr
e
qu
ency
is
way
to
o
s
m
al
l
comp
a
red
t
o
the
s
witc
hing
f
reque
ncy.
The
mi
nimum
inv
erte
r
side
inducta
nce
val
ue
is
ob
ta
ine
d
from
(1),
(2),
a
nd
(
3)
us
in
g
a
sim
ple
mathe
mati
cal
manipula
ti
on
wi
th
a
n
ass
umpti
on
that
the
gri
d
sy
ste
m is
sti
ff
and h
a
s a
unit
y p
ow
e
r fact
or,
1
_
=
8
(4)
2.3.
Capa
ci
to
r
The
ha
rm
onic
s
wh
ic
h
can
ne
gativel
y
a
ff
ect
the
sy
ste
m
ca
n
be
s
uppresse
d
us
i
ng
a
ca
pa
ci
tor
at
the
gr
i
d
sect
io
n
of
the
LCL
filt
er
.
H
ow
e
ve
r,
from
a
c
on
t
ro
l
sy
ste
m
po
i
nt
of
view
,
the
capaci
tor
provi
des
a
diff
e
re
nt
be
ha
vi
or
al
in
flue
nce
dep
e
ndin
g
on
wh
e
re
t
he
volt
age
a
nd
c
urre
nt
are
bein
g
me
asur
e
d
[
6],
[22
]
–
[24
]
.
Her
e
,
duri
ng
t
he
desi
gn
proc
ess,
the
ca
pacit
or
’
s
volt
age
i
s
measu
re
d
an
d
the
gr
id
c
urr
ent
is
con
t
ro
ll
ed
as
il
lustrate
d
in
F
igure
3.
T
he
s
el
ect
ion
is
ma
de
by
t
he
fact
that
the
filt
er
i
s
integr
at
ed
int
o
the
i
nv
e
rter
sy
ste
m
with
de
dicat
ed
cu
rr
e
nt/v
oltag
e
se
ns
ors
an
d
the
c
ontr
ol
stra
te
gy
that
is
use
d
f
or
the
fl
yback
-
base
d
c
on
ver
te
r
.
T
he maxim
um
capaci
ta
nce
val
ue
ca
n be calc
ul
at
ed
us
i
ng
E
quat
ion
(
5) b
el
ow;
=
2
(5)
w
he
re
V
o
is
the
rated
ou
t
pu
t
volt
age,
P
o
is
t
he
rated
powe
r
in
total
,
a
nd
ω
o
is
a
gr
i
d
an
gula
r
f
reque
ncy.
W
hile,
α
is
the
po
wer
rati
o
an
d
α
P
o
is
a
qu
a
ntit
y
of
reacti
ve
po
wer
absor
bed
by
the
filt
er
w
hich
is
typ
ic
al
ly
le
ss
tha
n
5%
[
5],
[
11],
[
18]
.
E
quat
ion
(
5
),
it
can
be
s
een
that
the
re
act
ive
powe
r
a
bs
or
ption
is
di
rectl
y
pro
porti
on
al
to
the
ca
pacit
anc
e
val
ue,
w
hic
h
means
that
the
higher
the
capaci
t
ance,
th
e
m
or
e
po
wer
is
a
bs
or
bed
by
t
he
capaci
tor
.
This
le
ads
to
a
hi
gher
c
urre
nt
de
man
d
f
rom
th
e
inducto
r
an
d
switc
hes
ca
usi
ng
a
higher
c
urren
t
rip
ple
an
d
l
ower
filt
erin
g
e
ff
i
ci
ency
.
Ne
ve
rtheless,
t
he
val
ue
of
the
ca
pa
ci
tor
m
us
t
not
be
to
o
l
ow
as
it
will
increase i
nduct
or size
and i
ncrea
se the
volt
ag
e dro
p
ac
ross t
he
in
ve
rter
si
de
inducto
r.
Evaluation Warning : The document was created with Spire.PDF for Python.
In
t J
P
ow Elec
& Dri S
ys
t
IS
S
N:
20
88
-
8
694
LCL fi
lt
er d
esi
gn for
g
ri
d
-
c
on
nected
sin
gle
-
phase
fl
yb
ack
m
ic
ro
invert
er
… (
M
uhama
d Fa
izal Y
aa
k
ub
)
1635
Figure
3:
Vo
lt
a
ge
a
nd curre
nt
measu
reme
nt a
t
LCL
filt
er
of
a micr
oinverter
2.4.
Grid Side
Ind
uct
an
ce
The
gri
d
in
duc
ta
nce
a
ff
ect
s
th
e
ove
rall
pe
rfo
rma
nce
a
nd
e
ffec
ti
ven
ess
of
the
gri
d
-
side
in
du
ct
a
nce
of
the
LCL
filt
er.
I
n
mo
st
li
te
ratur
e,
the
gri
d
i
nductance
factor
is
om
it
te
d
f
rom
th
e
tra
nsf
er
functi
on
a
nd
t
he
resona
nce
f
requen
c
y
der
iv
at
ion.
This
om
iss
ion
is
acce
pta
ble
in
sti
ff
gri
d
ci
rcumst
anc
es,
howe
ve
r,
the
gri
d
inducta
nce
is
a
lways
fi
nite
in
distri
bu
ti
on
gri
ds
as
repo
rted
in
I
EC
6072
5:2
01
2
[
4]
,
[9],
[
25],
[
26]
.
A
ty
pical
var
ia
ti
on
value
of
gri
d
in
du
ct
ance
is
in
the
r
ang
e
of
±
20%
[10],
[14
]
.
Th
e
gri
d
in
duct
anc
e
va
riat
ion
c
ha
ng
e
s
the
resona
nce
f
reque
ncy.
Ha
vi
ng
this
i
n
min
d,
the
gri
d
-
side
in
du
ct
or
L
2
ha
s
to
be
desig
ne
d
i
n
s
uc
h
a
wa
y
for
the
filt
e
r
t
o
be
able
t
o
a
bsor
b
a
ny
c
hanges
at
gri
d
-
in
du
ct
ance
value
.
T
he
res
on
a
nce
f
r
equ
e
nc
y,
,
of
the
filt
er by
ta
king
into
acc
ount
the
gr
i
d
in
duct
ance
var
ia
ti
on
is
d
e
fine
d
in
E
quat
ion
(6) bel
ow
;
=
√
1
+
2
+
1
(
2
+
)
(6)
The
res
on
a
nce
fr
e
quenc
y
,
,
is
directl
y
relat
e
d
to
the
i
nduct
ance
rati
o
,
r
,
wh
ic
h
is
depen
den
t
on
the
vo
lt
ag
e
a
nd
c
urre
nt
se
nsor
po
sit
io
n
at
t
he
filt
er.
Acc
ordi
ng
to
the
posit
ion
of
the
se
ns
ors
as
in
dicat
ed
in
Figure
3, the
re
so
na
nce
freq
ue
ncy
,
, is
giv
e
n
i
n
E
quat
ion (
7)
;
=
1
√
1
+
(7)
w
he
re
Z
b
is
th
e
base
impe
da
nce
def
i
ned
as
V
o
2
/P
o
.
T
o
si
mp
li
f
y
the
L
2
desig
n,
only
st
iff
gr
i
d
c
ondit
ion
is
consi
der
e
d,
i
gnori
ng
the
gri
d
inducta
nce
va
riat
ion
.
The
refor
e
,
the
a
ngula
r
res
onance
fr
e
qu
e
nc
y
,
,
of
th
e
LCL
filt
er is
fi
nally
e
xpresse
d
as
;
=
√
1
+
2
1
2
(8)
With
t
he
pres
ence
of
a
filt
er
ca
pacit
or
in
the
filt
er
net
w
ork,
the
gr
id
-
s
ide
in
duct
ance
L
2
ca
n
be
expresse
d wit
h res
pect to t
he
i
ndu
ct
ance
rati
o
,
r
;
2
=
1
(9)
T
he
relat
ion
s
hi
p
betwee
n
the
resona
nce
fr
e
quenc
y
a
nd
the
inducta
nce
rati
o
r
is
il
lustrate
d
in
Fig
ure
4
.
The
resona
nce
fr
e
quenc
y
of
the
L
CL
filt
er
m
us
t
com
ply
with
t
he
fo
ll
owin
g
c
on
st
raint
wh
e
r
e
it
shou
l
d
be
in
the
range
betwee
n
te
n
ti
mes
of
gri
d
fr
e
quenc
y
and
hal
f
of
t
he
switc
hi
ng
f
re
qu
e
nc
y
[1],
[
3],
[5],
[
6]
,
[
11]
,
[
27],
[28]
.
10
≤
≤
0
.
5
(10)
Evaluation Warning : The document was created with Spire.PDF for Python.
IS
S
N
:
2088
-
8
694
In
t J
P
ow
Ele
c
&
Dr
i
S
ys
t
,
V
ol
.
12
, N
o.
3
,
Se
ptembe
r
202
1
:
163
2
–
164
3
1636
Figure
4: R
es
onance
freq
ue
nc
y
,
,
a
nd in
du
ct
a
nce
rati
o
,
r
,
relat
ion
s
hip
ge
nerat
ed
f
rom E
qu
at
ion
(8)
Figure
5: T
he
i
mp
act
of a
da
mp
e
ning
resist
or
,
R
d
,
on t
he
st
abili
ty of a syst
em
2.5.
Da
m
ping
The
LC
L
filt
er
will
gen
e
rate
a
res
on
a
nce
fr
e
qu
e
nc
y
that
ca
us
es
i
ns
ta
bili
ty
to
the
overall
sy
ste
m.
The
eff
ect
ive
ness
of
the
at
te
nuat
ion
intr
oduce
d
by
the
LCL
filt
er
will
be
ac
hi
eved
only
if
dam
ping
is
pr
op
e
rly
impleme
nted
i
n
the
filt
er.
Ot
herwise,
a
n
ev
en
higher
cu
rrent
ri
pp
l
e
will
be
i
ntr
oduce
d
.
T
his
is
a
co
mmo
n
issue
re
ported
in
var
i
ou
s
li
te
r
at
ur
es
acc
omp
anied
wit
h
t
he
da
mp
i
ng
te
c
hniq
ues
us
e
d
to
recti
f
y
the
pr
ob
le
m
[3],
[
5], [7]
,
[
10]
,
[
16],
[
22],
[29],
[
30]
.
A
pa
ssive d
am
ping method
is
c
hos
en
as
it
is wide
ly
ad
opte
d
due
to
it
s
ease
of
im
ple
mentat
io
n.
T
he
meth
od
re
quir
es
ad
ding
a
res
ist
or
in
series
with
the
filt
er
capaci
ta
nce
.
A
tt
ention
sh
oul
d
be
giv
e
n
to
the
res
onance
f
reque
nc
y,
,
an
d
sam
pl
ing
fr
e
qu
e
nc
y
to
av
oid
high
losses.
Fi
gure
5
il
lustrate
s
the
e
ff
ect
of
va
rio
us
dam
ping
resi
stors
values
on
the
sta
bili
ty
of
the
s
ys
te
m
gi
ve
n
by
the
ope
n
-
lo
op
trans
fer
f
un
ct
i
on
[
31]
.
A
hi
gher
da
mp
i
ng
re
sist
or
value
yie
lds
a
m
ore
sta
bl
e
sy
ste
m
but
one
that
s
uffe
rs
from
higher
losse
s.
The
refo
re,
an
opti
mize
d
da
mp
in
g
resist
or
val
ue
mu
st
be
co
ns
ide
re
d
t
o
kee
p
t
he
ba
la
nce
betwee
n
the
s
yst
em
’s
sta
bili
ty
an
d
the
l
os
s
es
reducti
on.
Gen
e
rall
y,
t
he
dam
ping
resist
or
of
t
he
LCL
filt
er
is
cal
culat
ed by u
sing t
he follo
w
ing
e
quat
io
n
w
hich
is
one
-
t
hir
d of t
he
im
pe
da
nce
of the
filt
er ca
pacit
or
[
5]
.
≈
1
3
(11)
3.
DESIG
N
P
R
OCEDU
RE
A
ND EX
AMPL
E
The
LCL
filt
er
desi
gn
procedure
is
descr
i
be
d
acc
ordin
g
to
the
propose
d
de
sign
ste
p
s
s
ummari
ze
d
i
n
Figure
6
:
i.
The
in
it
ia
l
pa
r
amet
ers
of
the
s
ys
te
m
su
c
h
as
the
rated
powe
r
(
P
),
ou
t
pu
t
volt
age
(
v
o
),
DC
-
li
nk
vo
lt
age
(
v
i
)
,
s
ys
te
m
fr
e
qu
e
nc
y
(
f
g
)
,
s
witc
hi
ng
fr
e
quenc
y
(
f
sw
)
are
first
to
be
de
fine
d.
Ta
ble
1
su
m
marizes t
he
v
al
ue
of t
he varia
ble of t
he system
f
or
t
he
pro
po
se
d desi
gn ste
p.
Evaluation Warning : The document was created with Spire.PDF for Python.
In
t J
P
ow Elec
& Dri S
ys
t
IS
S
N:
20
88
-
8
694
LCL fi
lt
er d
esi
gn for
g
ri
d
-
c
on
nected
sin
gle
-
phase
fl
yb
ack
m
ic
ro
invert
er
… (
M
uhama
d Fa
izal Y
aa
k
ub
)
1637
ii.
Fo
r
a
wide
r
ipp
le
mar
gin
in
the
prel
imi
nary
desig
n
c
al
culat
ion
,
the
induct
or
c
urr
ent
ri
pp
le
coeffic
ie
nt,
β
c
is
sel
ect
ed
as
30%
an
d
th
us
,
gi
ves
t
he
minim
um
in
ver
te
r
-
si
de
i
nducta
nce
L
1min
=
1.6
8
mH. T
his is
obta
ined usin
g E
qu
at
io
n (
4)
.
Fi
nally, t
he neare
st p
os
sible
val
ue
L
1
=
1.7 mH
is sele
ct
ed.
iii.
The
ad
opte
d
powe
r
rati
o,
α
=
3%,
the
a
ntici
pated
qu
a
ntit
y
of
reacti
ve
po
w
er
a
bsor
be
d
by
the
filt
er
is
appr
ox
imat
el
y
60
V
AR.
S
ubs
ti
tuti
on
of
the
par
a
mete
rs'
values
to
Eq
uatio
n
(
5)
yields
t
he
maxim
um
filt
er
capaci
ta
nc
e
value,
C
fmax
=
3.95
µ
F.
T
he
refor
e
,
C
f
=
3
µF
is
chosen
c
on
si
der
i
n
g
t
he
near
est
off
-
the
-
s
helf
valu
e
an
d
cl
os
e
t
o
the
ma
xim
um
val
ue
as
possible
t
o
e
nsure
the
small
siz
e
of
filt
er
inducto
rs.
iv.
The
value
of
t
he
in
duct
ance
rati
o,
r
,
is
sel
ect
ed
to
co
mp
l
y
with
t
he
res
onance
fr
e
quenc
y
c
onditi
on.
As
f
or
the
ec
onomi
cal
po
i
nt
of
v
ie
w
mea
nt
f
or
fu
t
ur
e
hardw
a
re
de
velo
pme
nt,
r
=
1
is
c
ho
se
n
at
init
ia
l
f
res
= 3
.2 kHz.
The
refore
,
L
2
is s
et
to b
e eq
ual to
L
1
=
1.7 mH.
v.
Using
E
quat
io
n
(8),
t
he
res
onance
f
re
qu
e
nc
y
is
once
m
ore
ve
rified
us
i
ng
sel
ect
ed
value
s
in
ste
ps
ii
,
iii,
and
iv
t
o
guara
ntee
al
l
c
hose
n
passive
c
ompone
nts
paramet
ers
a
re
c
omplyi
ng
with
(
10)
since
the
values
are sl
ig
htly
diff
e
re
nt from the
preli
m
inar
y
cal
culat
ion.
vi.
Finall
y,
the
da
mp
in
g
resist
or
is
obta
ine
d
fro
m
E
quat
ion
(
11)
w
hich
is
on
e
-
thir
d
of
t
he
i
mp
e
dan
ce
of
the f
il
te
r
ca
pac
it
or
, yie
lds
R
d
= 5
Ω
.
vii.
In
a
real
case
sce
nar
i
o,
a
va
riat
ion
c
ompli
ance
te
st
is
r
equ
i
red
to
be
performe
d
to
validat
e
t
he
performa
nce
a
nd
sta
bili
ty
a
gainst
t
he
paramet
er
var
ia
ti
on
that
mig
ht
occ
ur.
M
ini
mu
m
six
te
st
conditi
ons are
pro
po
se
d f
or
t
he
test
.
F
ur
the
r e
la
borati
on is
discusse
d i
n
t
he
lat
er s
ect
io
n.
Table
1.
Sy
ste
m p
a
ramete
rs
Para
m
eter
Sy
m
b
o
l
Valu
e
Un
it
Rated
ou
tp
u
t Power
P
o
2
kW
Grid v
o
ltag
e
v
o
220
V
DC
-
Link
vo
ltag
e
v
i
350
V
Grid f
requ
en
cy
f
g
50
Hz
Switch
in
g
f
requ
en
cy
f
sw
10
k
Hz
Table
2.
Cal
c
ulate
d
filt
er
par
a
mete
rs
Para
m
eter
Sy
m
b
o
l
Valu
e
Un
it
Inv
erter
Side
I
n
d
u
ctan
ce
L
1
1
.7
mH
Grid Sid
e I
n
d
u
ctan
ce
L
2
1
.7
mH
Cap
acito
r
C
f
3
µF
Dam
p
in
g
Resis
to
r
R
d
5
Ω
Ind
u
ctan
ce r
atio
r
1
Res
o
n
an
ce Fr
eq
u
en
cy
f
r
es
3
.2
k
Hz
Figure
6: Fl
owchar
t
of
LCL
pa
rameter
f
il
te
r desig
n
Evaluation Warning : The document was created with Spire.PDF for Python.
IS
S
N
:
2088
-
8
694
In
t J
P
ow
Ele
c
&
Dr
i
S
ys
t
,
V
ol
.
12
, N
o.
3
,
Se
ptembe
r
202
1
:
163
2
–
164
3
1638
4.
SIMULATI
O
N
Simulat
io
n
of
the
desig
ne
d
LCL
filt
er
i
s
a
pp
li
ed
t
o
a
sin
gle
-
phase
gri
d
-
c
onnect
ed
fly
bac
k
microi
nv
e
rter
was
co
nducte
d
in
M
at
la
b/Si
mu
li
nk
softwa
re
base
d
on
t
he
ci
rc
uit
sh
own
i
n
Fig
ur
e
3
.
T
he
ob
je
ct
ive
is
t
o
ve
rif
y
t
he
pr
opos
e
d
cal
cula
te
d
passive
el
ements
v
al
ue
for
t
he
filt
er
a
nd
a
nalyze
the
filt
er
beh
a
vior
with
key
co
mpo
nent
par
a
mete
rs
va
riat
ion
.
T
he
s
pecifica
ti
on
of
the
desig
ne
d
filt
er
an
d
s
ys
te
m
are
su
m
marized
i
n
Ta
ble
1
an
d
T
able
2.
T
h
e
Bo
de
pl
ot
of
the
de
sign
e
d
filt
er
tr
ansf
e
r
f
unct
ion
is
pl
otted
in
Fi
gure
7.
F
rom
Fig
ure
7
,
the
syst
em
ca
n
be
see
n
as
sta
ble
with
90
de
grees
an
d
26.
8
dB
of
phase
ma
rg
i
n
a
nd
ga
i
n
mar
gin
,
r
es
pec
ti
vely.
Figure
7
: B
ode
p
lot
of the
d
es
ign
e
d
LC
L
filt
er
bas
ed
on a
p
a
rameter i
n
Ta
bl
e 2
The
sim
ulati
o
n
re
su
lt
of
the
desig
ne
d
filt
e
r
is
s
hown
in
Figure
8
.
Fig
ure
8
(a)
sho
ws
the
in
ver
te
r
ou
t
pu
t
volt
age
and
it
s
fr
e
quen
cy
sp
ect
r
um
upon
ente
rin
g
t
he
LCL
filt
er.
Fi
gure
8
(b)
an
d
(
c)
s
how
the
c
urre
nt
sign
al
an
d
fr
e
qu
e
nc
y
sp
ect
r
um
s
for
the
i
nv
e
rter
sid
e
inducto
r
an
d
gri
d
side
i
nduc
tor,
resp
ect
i
ve
ly.
The
maxim
um
c
urr
ent
rip
ple
rec
orde
d
at
the
in
ve
rter
side
,
i
L1
i
s
2A
a
nd
the
RMS
value
of
the
rated
c
urr
ent
is
16.6A
res
ulti
ng
in
the
maxi
mu
m
cu
rr
e
n
t
r
ipp
le
to
rate
d
current
rati
o
e
qu
al
t
o
12
%.
As
dep
ic
te
d
in
Fig
ur
e
8
(
b),
the
do
minant
ha
rm
on
ic
mag
nitu
des
of
c
urren
t
i
L1
a
pp
ea
r
at
20
kHz
a
nd
40
kH
z
w
hich
are
1.8%
a
nd
0.4%
of
the
f
unda
mental
harmo
nic
,
res
pecti
vely,
ge
ner
at
e
d
from
the
s
wi
tc
hin
g
dev
ic
es.
T
hro
ugh
the
f
il
te
r,
the domi
na
nt c
urren
t
ha
rm
on
i
c sp
ect
r
ums ar
e su
cces
sf
ully
su
pp
resse
d
at
i
L2
as sho
wn in Fi
gure
8
(c
).
In
Fig
ur
e
9(
a
),
the
res
on
a
nce
fr
e
que
ncy
ap
pe
ars
at
63
rd
harmo
nic
order
w
hich
is
3.2
kHz
as
per
th
e
init
ia
l
theor
et
ic
al
cal
culat
ion
.
The
sim
ulati
on
re
veals
that
t
he
domi
na
nt
r
eso
nan
ce
s
pect
rum
is
14.
74%
from
it
s
fun
dame
ntal
com
pone
nt
a
nd
produce
d
THD
at
15.
2%
at
the
ou
t
put
filt
er
point.
Eve
ntua
ll
y,
with
t
he
impleme
ntati
on
of
a
dam
ping
resist
or
c
on
nected
i
n
serie
s
with
a
filt
er
capaci
tor
,
th
e
resona
nce
f
re
quenc
y
gen
e
rated
f
r
om
the
pas
sive
el
ement
c
omp
on
e
nt
at
t
he
63
rd
ha
rm
on
ic
is
s
uc
cessf
ully
s
uppr
essed
to
0.1
4%
with
resp
ect
to
it
s
f
undame
ntal
c
omp
on
e
nt.
T
he
sign
ific
a
nt
res
ult
ca
n
be
see
n
f
rom
the
pl
otted
fr
e
que
ncy
s
pectr
um
in Figu
re
9(b).
I
m
pact
on p
arameter
va
ri
ation
Du
e
to
the
fac
tors
discu
ssed
earli
er,
it
is
di
ff
ic
ult
to
buil
d
a
physi
cal
filt
er
that
pr
eci
se
ly
fits
the
theo
reti
cal
des
ign
.
The
se
as
pects
m
us
t
be
ta
ken
i
nto
ac
count
durin
g
t
he
desig
n
pro
cess.
An
i
nduc
ta
nce
var
ia
ti
on
in
a
gri
d
-
ti
ed
filt
er
c
ou
l
d
ste
m
fro
m
seve
ral
fact
or
s
t
hat
co
uld
be
ca
us
ed
by
mate
rial
pro
pe
rtie
s
su
c
h
as
cor
e
pe
rme
abili
ty
coeffic
ie
nt
an
d
ma
nuf
act
ur
in
g
tole
ra
nce.
It
co
uld
a
lso
be
t
he
res
ul
t
of
an
i
nc
on
s
ist
ent
gr
i
d
inducta
nc
e
connecte
d
i
n
series
with
L
2
.
The
ca
pa
ci
ta
nce
var
ia
ti
on
in
the
c
apacit
or
due
to
the
man
uf
act
ur
e
r
's
tolera
nce
or
t
he
a
ging
co
ul
d
al
s
o
be
a
co
ntributi
ng
fact
or
.
T
he
refor
e
,
a
sim
ulati
on
on
t
he
rob
us
tness
a
ga
inst
the
var
ia
ti
on
is
pe
rforme
d
t
o
e
mu
la
te
t
he
real
case
sc
enar
i
o.
Six
c
onditi
ons
of
pa
r
amet
er
var
ia
ti
ons a
re c
on
si
der
e
d wit
h t
he
ass
umpti
on that t
he
e
ve
nt h
a
pp
e
ns o
n
e a
t a ti
me
•
30% inc
reme
nt
in
L
1
in
duct
an
ce value
.
•
30%
decr
eme
nt
in
L
1
i
nductan
ce value
.
•
30% inc
reme
nt
in
L
2
in
du
ct
a
nc
e v
al
ue
.
•
30%
decr
eme
nt
in
L
2
i
nductan
ce value
.
•
20% inc
reme
nt
in
C
f
ca
pacit
ance
value.
•
20%
decr
eme
nt
in
C
f
ca
pacit
ance
value.
Evaluation Warning : The document was created with Spire.PDF for Python.
In
t J
P
ow Elec
& Dri S
ys
t
IS
S
N:
20
88
-
8
694
LCL fi
lt
er d
esi
gn for
g
ri
d
-
c
on
nected
sin
gle
-
phase
fl
yb
ack
m
ic
ro
invert
er
… (
M
uhama
d Fa
izal Y
aa
k
ub
)
1639
The
i
nductanc
e
va
riat
ion
in
t
his
pap
e
r
is
set
to
be
10%
higher
co
mp
a
re
d
t
o
oth
e
r
li
te
ratu
re
[
1],
[
10]
,
[14]
to
sec
ure
a
highe
r
s
ys
te
m
sta
bili
ty
ma
rg
i
n.
T
he
20%
var
ia
ti
on
of
c
apacit
ance
is
a
dopted
ba
sed
on
th
e
maxim
um
t
oler
ance set
by m
ost
manuf
act
ur
e
r
s
.
(a)
(b)
(c)
Figure
8
: Sim
ul
at
ion
r
es
ult
of
LCL
filt
er
(a)
i
nv
e
rter
volt
age
(b)
i
nv
e
rter
side in
duct
or
i
L1
current
,
and
(c)
g
rid
side in
duct
or
i
L2
cur
re
nt
(a)
(
b)
Figure
9: Si
gnal
an
d s
pectrum f
or cu
rr
e
nt
:
(
a)w
it
ho
ut d
a
m
ping
resist
or a
nd (b) wit
h a
da
mp
in
g resist
or
Evaluation Warning : The document was created with Spire.PDF for Python.
IS
S
N
:
2088
-
8
694
In
t J
P
ow
Ele
c
&
Dr
i
S
ys
t
,
V
ol
.
12
, N
o.
3
,
Se
ptembe
r
202
1
:
163
2
–
164
3
1640
The
sim
ulati
on
res
ult
for
t
he
first
co
ndit
ion
is
show
n
i
n
Fi
gure
10(a).
T
he
filt
er
sho
we
d
e
xcell
ent
sta
bili
ty
an
d
th
e
abili
ty
to
a
bsor
b
the
30%
i
nc
reme
nt
of
L
1
. Th
e b
ode p
lot shows
t
he
gain
mar
gin
is
at
26.9
dB
wh
il
e
t
he
phas
e
ma
rg
i
n
is
at
90
de
grees
w
it
h
a
sli
gh
t
s
hift
at
t
he
sy
ste
m
respo
ns
e.
F
ur
t
he
rm
ore,
th
e
T
HD
value
re
su
lt
ed
from
the
e
ven
t
is
3.7
6%.
On
the
oth
e
r
ha
nd,
in
the
sec
ond
conditi
on
,
the
decr
e
ment
of
L
1
by
30%
ca
us
es
t
he
TH
D
t
o
i
ncr
e
ase
an
d
sli
gh
tl
y
a
bove
the
sta
nd
a
r
d
th
res
ho
l
d
li
mit
to
5.0
1%
as
s
how
n
i
n
Fig
ur
e
10(
b)
.
H
ow
e
ve
r,
the
sy
ste
m
r
emai
ns
sta
ble
with
a
la
r
ger
ga
in
mar
gin,
a
nd
with
a
posit
ive
s
hift
f
rom
26.
8
dB
to 27.3
dB.
The
respo
ns
e
f
rom
the
t
hir
d
c
onditi
on
is
sho
wn
in
Fi
gure
11(a
)
.
T
he
rise
of
L
2
by
30%
s
ti
ll
ho
lds
the
sy
ste
m
at
a
sta
ble
sta
te
with
26.9
dB
a
nd
90
de
grees
gain
and
phase
ma
rg
i
n,
re
sp
ect
iv
el
y.
T
he
T
HD
valu
e
ob
ta
ine
d
is
3.7
5%,
wh
ic
h
is
be
low
the
sta
nd
ard
t
hr
es
hold
li
mit
.
In
c
ontras
t,
the
fou
rth
c
onditi
on
s
hows
wh
e
n
L
2
dec
reases,
THD
i
nc
reases
0.0
2%
higher
than
the
sta
ndard
t
hr
es
hold
l
imi
t
to
5.0
2%,
as
s
hown
in
Figur
e
11(
b)
.
H
owev
er,
t
he
s
ys
te
m
sta
bili
ty
is
m
ai
ntained
at
a
value
27.
3
dB
gain
ma
rg
i
n
and
90
de
gr
ee
s
phase
mar
gin
.
The
ef
fect
of
the
filt
er
ca
paci
tor
va
riat
ion
is
il
lustrate
d
in
Figure
12.
Fi
gure
12(a)
s
hows
t
he
bode
plo
t
an
d
f
re
qu
e
ncy
s
pectr
um
harmo
nics
f
or
the
fifth
c
ondit
ion
.
T
H
D
val
ue
du
e
t
o
the
20
%
capaci
ta
nce
rise
is
measu
red
at
4.2
9%
a
nd
the
recorde
d
sta
bili
ty
mar
gin
is
27
dB
an
d
90
de
gr
ees
f
or
the
gain
a
nd
phase,
resp
ec
ti
vel
y.
S
imi
la
r
res
ults
wer
e
ob
se
r
ved
in
the
sixt
h
c
onditi
on
when
the
ca
pacit
an
ce
is
dec
rease
d.
T
he
simulat
ion
r
es
ult
is
il
lustrate
d
i
n
Fig
ur
e
12(
b).
The
T
H
D
is
meas
ured
at
4.28%
a
nd
the
syst
em
sta
bili
ty
remaine
d
at
a
26.9
dB
gain ma
rg
i
n
a
nd 90 de
gr
ees
pha
se
ma
rg
i
n.
Numerical
res
ults
f
rom
par
a
mete
r
var
ia
ti
on
te
st
sim
ulati
on
as
de
picte
d
in
Fi
gure
10
–
Fi
gure
12
unde
r
se
ver
al
di
ff
ere
nt cases
a
re summa
rized
in
Ta
ble
3.
Table
3
.
T
he
stabil
it
y
mar
gin
and
T
HD
of th
e filt
er w
it
h
c
omp
on
e
nt
var
ia
t
ion
s
Initial
co
m
p
o
n
en
t
v
alu
e
L
1
in
crea
se
30%
L
1
d
ecre
ase
30%
L
2
in
crea
se
30%
L
2
d
ecre
ase
30%
C
f
in
crea
se
20%
C
f
d
ecre
ase
20%
Ph
ase
M
argin
(deg
)
90
90
90
90
90
90
90
Gain
M
argin
(dB
)
2
6
.8
2
6
.9
2
7
.3
2
6
.9
2
7
.3
27
2
6
.9
THD
(%)
4
.28
3
.76
5
.01
3
.75
5
.02
4
.29
4
.28
(a)
(b)
Figure
10: Fre
qu
e
nc
y
s
pectr
um o
f o
utput fil
te
r
c
urren
t a
nd fi
lt
er b
ode
plo
t
for (a) i
nducta
nce
L
1
increa
se
d by
30% a
nd (b) i
nducta
nce
L
1
de
creased
by 3
0%
Evaluation Warning : The document was created with Spire.PDF for Python.
In
t J
P
ow Elec
& Dri S
ys
t
IS
S
N:
20
88
-
8
694
LCL fi
lt
er d
esi
gn for
g
ri
d
-
c
on
nected
sin
gle
-
phase
fl
yb
ack
m
ic
ro
invert
er
… (
M
uhama
d Fa
izal Y
aa
k
ub
)
1641
(a)
(b)
Figure
11: Fre
qu
e
nc
y
s
pectr
um o
f o
utput fil
te
r
c
urren
t a
nd fi
lt
er b
ode
plo
t
for (a) i
nducta
nce
L
2
increa
se
d by
30% a
nd (b) i
nducta
nce
L
2
de
creased
by 3
0%
(a)
(
b)
Figure
12: Fre
qu
e
nc
y
s
pectr
um o
f o
utput fil
te
r
c
urren
t a
nd fi
lt
er b
ode
plo
t
for (a)
ca
pacit
a
nce
C
f
i
ncr
ease
d by
20% a
nd (b) ca
pacit
ance
C
f
de
creased
by 2
0%
5.
CONCL
US
I
O
N
This
paper
ha
s
pro
pose
d
a
ste
p
-
by
-
ste
p
de
sign
guideli
ne
f
or
the
LC
L
filt
er
us
e
d
in
the
gri
d
-
connecte
d
a
pp
li
cat
ion
,
s
uc
h
as
fl
yback
mi
cro
i
nv
e
rter.
T
he
LCL
filt
er
was
desi
gned
base
d
on
seve
ral
key
aspects
to
mai
ntain
the
filt
er’s
sta
bili
ty
and
eff
ect
ive
ness
s
uch
a
s
ma
xim
um
c
urren
t
rip
ple,
in
du
ct
a
nce
rati
o,
and res
onance
fr
e
qu
e
nc
y dam
ping.
E
xten
siv
e simulat
io
n
w
ork
was
done
t
o vali
date the t
heoreti
cal
calc
ulati
on
desig
n
via
M
at
la
b/Simulin
k.
I
n
ad
diti
on
,
a
de
ta
il
ed
analysi
s
on
the
passive
el
ements’
pa
rameter
var
ia
ti
on
is
pro
vid
e
d
i
n
t
his
pa
pe
r
t
o
e
xhibit
it
s
in
flue
nc
e
on
the
f
il
te
r
char
act
e
risti
c
and
performa
nc
e
i.e.
sta
bili
ty
mar
gi
n
and
t
otal
ha
rm
on
ic
distor
ti
on.
T
he
pro
pose
d
de
sig
n
ste
ps
pro
duced
a
filt
e
r
t
hat
remai
ns
sta
ble
a
nd
t
he
THD
value
is
within
the
inter
natio
nal
sta
ndar
d
s
(
i.e.
IEE
E1
547.
1
,
IEC
60725
)
des
pite
so
me
par
a
mete
r
var
i
at
ion
.
Howe
ver,
the
analysis
rev
eal
s
that
wh
e
n
r
is
un
it
y,
in
du
c
tor
sel
ect
ion
mu
st
be
crit
ic
al
ly
desig
ne
d
because
THD
value
will
increa
se
a
s
i
nducta
nce
dec
re
ase
a
nd
it
c
ould
be
worse
ning
i
f
i
nducta
nce
dr
op
s
be
yond
30%
of it
s d
esi
gn
at
e
d value.
Evaluation Warning : The document was created with Spire.PDF for Python.