Int
ern
at
i
onal
Journ
al of
P
ower E
le
ctr
on
i
cs a
n
d
Drive
S
ystem
(I
J
PE
D
S
)
Vo
l.
11
,
No.
3
,
Septem
be
r 2020
, pp.
1406
~
1414
IS
S
N:
20
88
-
8694
,
DOI: 10
.11
591/
ij
peds
.
v11.i
3
.
pp
1406
-
1414
1406
Journ
al h
om
e
page
:
http:
//
ij
pe
ds
.i
aescore.c
om
Harmon
ic minim
izat
i
on o
f a singl
e
-
phase a
symmetr
ical TC
HB
multile
vel i
nvert
er base
d on ne
arest le
vel
contr
ol m
ethod
Wail Ali
Ali S
aleh, Nur
ul Ai
n Moh
d Said,
Wah
id
ah A
bd
Ha
li
m
Facul
ty
of Electr
ic
a
l
Eng
ineeri
ng
,
Univer
si
ti T
ekn
ika
l
Mala
ysi
a
M
el
ak
a, Ma
la
ysi
a
Cent
er
for
Robo
t
ic
s &
Industri
al
Automa
ti
on
,
Uni
ver
siti
Te
kn
ika
l
Mala
ysia
Mel
ak
a,
Ma
la
ysi
a
Art
ic
le
In
f
o
ABSTR
A
CT
Art
ic
le
history:
Re
cei
ved
Feb
7
, 2
0
20
Re
vised
M
a
r
6
,
20
20
Accepte
d
A
pr
19
, 20
20
Multi
le
v
el
inv
erters
ar
e
g
ai
ning
s
pec
i
al
intere
st
a
mong
res
ea
rch
er
s
and
in
the
industry
due
t
o
the
ir
wid
espre
ad
app
licatio
ns
and
nume
r
ous
me
rit
s.
Obtai
ning
high
q
ual
it
y
,
mor
e
reli
abl
e
outpu
t
whil
e
using
a
red
u
ced
numb
er
of
el
e
ct
roni
c
co
mp
onent
s
is
th
e
m
ain
purpose
of
mo
st
of
the
rese
a
rc
h
conduc
t
ed
in
th
is
are
a
of
st
udy.
The
purpos
e
of
th
is
study
i
s
to
appl
y
th
e
n
ea
rest
l
evel
cont
rol
(NLC)
me
thod
to
a
13
-
le
ve
l
tr
ansistor
-
cl
a
mpe
d
H
-
br
id
ge
(TCHB)
inve
rt
er
with
un
equa
l
DC
voltag
e
suppli
es.
Th
e
NLC
me
thod
op
era
t
es
at
th
e
funda
me
n
ta
l
fr
e
quenc
y,
thus
re
duci
ng
sw
it
chi
n
g
losses,
and
c
a
n
red
u
ce
th
e
har
monic
con
tent
significan
tl
y.
The
adopted
m
ult
ilevel
inv
ert
er
consists
o
f
two
TCHB
c
el
ls
supplie
d
with
t
wo
asymm
e
tri
c
a
l
DC
input
sour
ce
s
with
a
volt
ag
e
ra
ti
o
o
f
1:
2.
Thi
s
str
uct
ure
red
u
ce
s
the
numbe
r
o
f
elec
tron
ic
com
ponen
ts,
an
d
the
asymmetry
in
th
e
DC
inpu
t
volt
ag
es
resul
ts
in
a
high
er
numbe
r
of
le
ve
l
s.
The
adopt
ed
topol
ogy
and
its
proposed
control
method
were
si
mul
a
te
d
i
n
Matlab/Sim
u
link,
and
the
sim
ula
ti
on
r
esult
s
w
er
e
ver
if
ie
d
through
exp
eri
m
ent
s
using
an
Al
te
ra
field
-
progra
mm
able
g
at
e
arr
ay
(FP
GA
)
boar
d.
The
r
esu
lt
s
show
ed
that
the
topol
ogy
a
nd
it
s
con
trol
me
thod
are
eff
icient
in
ob
tai
ning
a
high
-
qua
l
it
y
output
with
a
n
i
mprove
d
to
ta
l
har
moni
c
distort
ion (THD).
Ke
yw
o
rd
s
:
Asymmetric
al
topolo
gies
M
ulti
le
vel in
ve
rter
Near
est
le
vel c
on
t
ro
l
Total
h
a
rm
onic
d
ist
ort
ion
Transi
stor
-
cl
ampe
d H
-
br
i
dg
e
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
:
Nuru
l
A
i
n Mohd Said
,
Faculty
of Elec
tric
al
Engineer
ing
,
Un
i
ver
sit
i Te
knikal
M
al
aysia
M
el
a
ka,
Hang T
ua
h
Ja
ya
, 76100
D
ur
ia
n
T
unggal
, Me
la
ka,
M
al
aysia
.
Emai
l:
n
ur
ulain@
utem.e
du.
my
1.
INTROD
U
CTION
M
ulti
le
vel
in
ve
rters
(MLIs
)
hav
e
bec
om
e
widely
us
e
d
i
n
the
in
du
st
ry,
especial
ly
in
medium/
hi
gh
vo
lt
age
high
powe
r
ap
plica
ti
on
s.
Ele
ct
ric
and
hybri
d
veh
ic
le
s,
phot
ovoltai
c
energ
y
co
nversi
on
and
un
i
nterru
ptable
power
sup
ply
(U
P
S)
a
re
jus
t
a
few
of
the
man
y
a
ppli
ca
ti
on
s
a
vaila
ble
in
the
in
dustr
y
[
1]
.
Re
searche
rs
ar
e
exe
rtin
g
grea
t
effo
rts
to
de
velo
p
bette
r
to
po
l
og
ie
s
of
MLIs
an
d
bette
r
modu
la
ti
on
me
thods
and
/
or
co
ntr
ol
strat
egies
in
te
rms
of
c
os
t,
si
ze,
ef
fici
enc
y
and
reli
abili
ty.
The
re
a
re
th
re
e
main
to
po
l
ogie
s
of
ML
Is:
Ne
utral
po
i
nt
cl
amp
ed
(N
PC
),
flyi
ng
capaci
tor
(F
C
),
an
d
casca
de
d
H
-
br
i
dg
e
(C
H
B)
[1
-
3]
.
M
a
ny
oth
e
r
topolo
gies
ha
ve
be
en
th
oro
ughl
y
i
nvest
igate
d
i
n
the
li
te
ratur
e,
s
uc
h
as
the
m
odula
r
mu
lt
il
evel
c
onver
te
r
(MMC)
[4]
a
nd
t
he
tra
ns
ist
or
-
cl
ampe
d
H
-
br
idg
e
(
TC
HB)
M
L
I
[5
,
6]
.
T
he
TCHB
MLI
us
es
a
reduce
d
numb
e
r
of
DC
powe
r
s
upplies
a
nd
sw
it
ches
to
pro
duce
the
same
nu
mb
e
r
of
le
vels
as
c
ompare
d
w
it
h
the
c
onve
ntion
a
l
CHB
M
LI
.
More
over,
the
T
CHB
in
ver
te
r
respo
nd
s
m
ore
effe
ct
ively
i
n
re
du
ci
ng
the
harmo
nic
c
on
t
ent
of
ou
t
pu
t
volt
age
an
d
lo
ad
c
urre
nt.
Acc
ordin
g
to
t
he
val
ues
of
DC
power
s
up
plies,
s
ome
t
opologies
of
M
L
I
s
are
cl
assifi
ed
i
nto
ei
ther
s
ymm
et
rical
if
t
hes
e
val
ues
are
eq
ual,
or
as
ym
met
rical
if
the
y
are
unequ
al
.
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
Ha
r
m
on
ic
mi
nimiza
ti
on
of
a
si
ng
le
-
ph
as
e
as
y
mm
e
tri
c
al TC
HB mu
lt
il
evel
inverte
r
…
(
Wai
l Al
i Al
i Sa
le
h
)
1407
Asymmetric
al
topolo
gies
of
M
LI
s
[7
,
8]
pro
du
ce
a
higher
numbe
r
of
outp
ut
le
vel
s
wh
il
e
usi
ng
few
e
r
com
pone
nts as
co
m
pa
red wit
h
the
s
ym
met
rical
o
ne
s.
High
-
s
witc
hing
fr
e
qu
e
nc
y
mod
ulati
on
te
chn
i
ques,
su
c
h
as
mu
lt
ic
ar
rier
/mult
irefere
nce
PWM
[9
-
11]
a
nd
s
pace
ve
ct
or
PWM
[
12]
a
r
e
ef
fici
ent
i
n
obta
inin
g
a
high
-
qual
it
y
outp
ut;
howe
ve
r,
s
witc
hin
g
losses
are
do
minant
in
s
uc
h
te
c
hn
i
ques.
Lo
w
-
fr
e
qu
e
nc
y
mod
ulati
on
s
are
m
ore
e
ffi
ci
ent
f
or
high
-
po
we
r
app
li
cat
io
ns
si
nce
t
hey
mini
mize
switc
hi
ng
lo
sses.
Sele
c
ti
ve
ha
rm
on
ic
el
imi
nation
(SHE)
[13
-
16]
,
near
est
vecto
r
c
ontr
ol
(
NV
C
)
[
17]
,
and
near
est
le
vel
c
on
t
ro
l
(
N
LC)
[18
-
20]
a
re
t
he
mo
st
c
om
m
only
use
d
lo
w
-
fr
e
qu
e
nc
y
m
odulati
ons
in
MLIs.
T
he
S
HE
method
requi
res
the
s
olu
ti
on
of
s
om
e
nonl
inear
tra
ns
cen
den
ta
l
equ
at
io
ns
.
H
oweve
r,
as
the
numb
e
r
of
le
ve
ls
incre
ases,
more
e
quat
ions
e
xist
w
hose
so
luti
ons
bec
ome
m
ore
com
plica
te
d
a
nd
ti
me
-
c
onsumi
ng.
T
he
w
orki
ng
pri
nci
ples
of
NV
C
an
d
NLC
a
re
a
pp
roximat
el
y
t
he
same,
excep
t
that
the
NV
C
c
onsider
s
the
volt
age
ve
ct
or
near
est
t
o
the
re
fere
nce
instea
d
of
the
near
est
vo
lt
age
le
ve
l
in
the
ca
se
of
t
he
NLC
[3]
.
A
lt
ho
ug
h
the
N
VC
is
sim
ple,
i
t
requires
ad
diti
on
al
c
o
m
pu
ta
t
ion
al
resou
rces
[21
]
.
The
NLC
met
hod
is
sim
ple
and
m
or
e
s
uita
ble
f
or
M
L
Is
with
a
high
numb
e
r
of
le
ve
ls.
The
NLC
met
hod
is
app
li
ed
to
asy
mmetric
al
CH
B
M
L
Is
i
n
[7]
and
[
21]
.
T
he
resu
lt
s
s
how
th
at
the
T
HD
is
reduce
d
dr
a
ma
ti
cal
ly
witho
ut
t
he
ne
ed
f
or
filt
rati
on.
S
witc
hing
l
os
ses
are
si
gn
i
ficantl
y
re
duce
d
us
i
ng
the
N
LC
meth
od,
es
pecial
ly
in
as
ymmet
rical
topolo
gies
with
a
high
nu
mb
e
r
of
le
vels
,
sinc
e
m
os
t
of
the
po
wer
is
delivere
d
by
the
high
powe
r
cel
l
wh
i
ch
c
ommutat
es
at
the
f
undam
ental
f
reque
nc
y
[21]
.
T
he
N
LC
meth
od
is
al
so
a
pp
li
ed
to
a
21
le
vel
asy
mmet
rical
TCHB
(
A
-
TC
HB)
i
nv
e
rter
in
[
8]
with
a
DC
li
nk
rati
o
of
1:
4
bet
w
een
tw
o
TC
H
B
cel
ls.
The
r
esults
s
how
a
sig
nifica
ntly
imp
rove
d
T
HD,
ho
wev
e
r,
the
stu
dy
does
no
t
discuss
ap
ply
in
g
t
he
met
hod
t
o
a
reduce
d
numb
e
r
of
le
vels
(less
tha
n
21)
.
A
fter
a
pr
eci
se
rev
ie
w
of
t
he
li
te
ratur
e
,
it
was
f
ound
th
at
fe
w
stud
ie
s
ha
ve
be
en
c
ondu
ct
ed
on
the
NLC
method
a
ppli
ed
to
t
he
TCHB
inv
e
rter
c
omp
ared
t
o
M
M
C
[22
-
25]
and
as
ymmet
ri
cal
CHB
M
L
I
[7
,
21]
;
The
refor
e
,
t
his
pa
per
fo
c
us
es
on
a
pplying
the
NLC
method
to
a
13
le
vel
A
-
TC
HB
i
nv
e
rter
with
a
D
C
volt
age
rati
o
1:
2
bet
we
en
t
wo
TC
H
B
cel
ls.
T
he
method
ope
rates
at
the
f
un
dame
ntal
fr
e
quenc
y
a
nd
ca
n
be
ext
end
e
d
to
an
y
numb
e
r
of
le
ve
ls.
T
he
pro
posed
NLC
met
hod
is
performe
d
t
hro
ugh
simulat
io
n
in M
at
la
b/Sim
ulink
a
nd
the
s
imulat
ion
res
ul
ts
are v
e
rified
thr
ough
e
xperi
ments
us
in
g
a
n Alt
era
FPGA
boar
d.
2.
THE
ADOPT
ED 13
-
LE
VE
L A
-
T
CHB I
NV
E
RTER
The
ci
rcu
it
diagr
a
m
of
the
13
-
le
vel
A
-
TC
HB
in
ver
te
r
is
sho
wn
in
Fig
ur
e
1.
T
he
c
onfi
gurati
on
consi
sts
of
t
w
o
TC
HB
cel
ls
sup
plied
with
une
qu
al
DC
input
vo
lt
age
s.
The
lo
w
-
volt
age
cel
l
(
cel
l
1
)
is
su
ppli
ed
with
,
an
d
the
hi
gh
-
volt
age
cel
l
(
cel
l
2
)
is
s
uppl
ie
d
with
2
.Th
e
switc
hing
sta
t
es
of
each
TCHB
cel
l
ind
ivid
ually
a
re
sh
ow
n
in
Ta
bl
e
1
.
T
he
1:
2
vo
lt
age
rati
o
was
ch
os
e
n
si
nce
it
is
the
small
est
integer
rati
o
f
or
t
he
tw
o
-
cel
l
config
ur
at
io
n;
hen
ce
a
ny
hi
gh
e
r
inte
ger
r
at
io
can
be
as
su
me
d.
T
he
13
-
le
vel
ou
t
pu
t
is
s
yn
t
he
siz
ed
by
ad
di
ng
the
volt
age
s
of
t
he
tw
o
ca
scade
d
cel
ls.
D
et
ai
le
d
volt
age
wa
veforms
f
or
eac
h
cel
l
as
well
as
for
t
he
i
nv
e
rter
outp
ut
a
re
il
lu
strat
ed
i
n
Fig
ure
2.
U
nlike
th
e
hi
gh
-
vo
lt
age
cel
l,
the
l
ow
-
volt
age
cel
l
has
ma
ny
com
mu
ta
ti
ons
per
cycle.
H
oweve
r,
most
of
the
powe
r
is
delive
red
by
the
hi
gh
-
vo
lt
a
ge
cel
l
wh
ic
h
c
om
m
ut
at
es at
the
fun
da
mental
fre
qu
e
ncy
; t
hus,
t
he
s
witc
hing lo
sses
are
si
gn
i
ficant
ly r
e
duced
.
Table
1.
Sw
it
c
hing stat
es
of the TC
HB cell
s
No
ON switch
es
Vo
ltag
e level
ON switch
es
Vo
ltag
e level
cell 1 (
)
cell 2 (
)
1
11
,
41
+
12
,
42
+
2
2
41
,
51
+
1
2
⁄
42
,
52
+
3
11
,
21
(
31
,
41
)
0
12
,
22
(
32
,
42
)
0
4
21
,
51
−
1
2
⁄
22
,
52
−
5
21
,
31
−
22
,
32
−
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
&
D
ri
S
ys
t,
V
ol
.
11
, N
o.
3
,
Se
ptembe
r
2020
:
14
06
–
14
14
1408
C
1
C
2
C
3
c
e
l
l
1
c
e
l
l
2
S
52
S
42
S
32
S
22
S
12
C
4
S
11
S
21
S
31
S
41
S
51
v
1
v
i
nv
L
R
i
o
v
o
V
dc
2
V
dc
v
2
Figure
1. The
13
-
le
vel
as
ym
metri
cal
TCH
B i
nv
e
rter
v
1
v
dc
-
v
dc
v
dc
-
v
dc
v
dc
-
v
dc
2
v
dc
-
2
v
dc
-
2
v
dc
2
v
dc
-
3
v
dc
3
v
dc
v
i
nv
v
2
wt
wt
wt
Figure
2.
Lo
w
-
vo
lt
age
cell
, high
-
vo
lt
age
cell
, and res
ultant
13
-
le
vel
outp
ut volt
ages,
r
es
pe
ct
ively
3.
THE
PROPO
SED
NLC
ME
THOD
The
pro
posed
con
t
ro
l
st
rateg
y
a
pp
li
ed
to
t
he
a
dopted
in
ver
te
r
is
cal
le
d
the
nea
rest
l
evel
co
ntr
ol
(N
LC
)
meth
od
.
The
met
hod
i
s
simple
in
it
s
con
ce
pt
in
wh
i
ch
the
ou
t
pu
t
volt
age
le
vel
ne
arest
to
the
re
f
eren
ce
is
sel
ect
ed
an
d
gatin
g
sig
na
ls
are
ge
ne
rated
t
o
t
he
co
rresp
onding
swi
tc
hes
[7]
.
F
or
A
-
TC
HB
MLI,
th
e
modu
la
ti
on i
ndex,
is
gen
e
rall
y defi
ned as:
=
⁄
(1)
wh
e
re
is
the
a
mp
li
tud
e
of
t
he
ref
e
re
nce,
a
nd
is
the
t
otal
DC
li
nk
vo
lt
ag
e
f
or
cel
ls.
is
cal
culat
ed by
:
=
∑
,
=
1
(2)
Fo
r
t
he
13
-
le
ve
l
A
-
TCHB
in
ve
rter,
since
the
volt
age
rati
o
be
tween
the
tw
o
cel
ls
is
1:
2,
a
ccordin
g
to
(2),
=
dc
,
1
+
dc
,
2
=
+
2
=
3
.
T
he
refore
is
directl
y
pro
portio
nal
to
.
A
s
a
n
e
xam
pl
e:
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
Ha
r
m
on
ic
mi
nimiza
ti
on
of
a
si
ng
le
-
ph
as
e
as
y
mm
e
tri
c
al TC
HB mu
lt
il
evel
inverte
r
…
(
Wai
l Al
i Al
i Sa
le
h
)
1409
To
set
to
1,
t
he
value
of
is
c
hose
n
to
be
1
3
⁄
.
S
o,
,
1
=
1
3
⁄
and
,
2
=
2
3
⁄
.
The
impleme
ntati
on
of
t
he
NLC
method
f
or
t
he
A
-
TC
HB
i
nv
e
rter
by
a
c
omp
ariso
n
al
gorith
m
is
s
how
n
i
n
Figur
e
3.
A
sin
usoidal
ref
e
ren
ce
,
∗
is
first
fe
d
to
t
he
high
-
volt
age
c
el
l
(
cel
l
2
)
.
W
hen
t
he
re
fer
e
nc
e
reache
s
cert
ai
n
vo
lt
age
le
vels,
±
ℎ
,
2
,
cel
l
2
pro
duces
±
1
2
⁄
,
2
.
Wh
e
re
as,
w
hen
t
he
ref
e
ren
ce
reac
hes
±
ℎ
,
2
,
cel
l
2
pro
du
ces
±
,
2
.
Th
e
unm
odulate
d
portio
n
le
ft
by
the
five
-
le
ve
l
wav
e
form
of
cel
l
2
is
the
n
us
e
d
as
the
ref
e
ren
ce
a
nd
f
ed
to
t
he
lo
wer
vo
lt
a
ge
cel
l
(
c
el
l
1
).
Again,
ba
sed
on
t
he
vo
l
ta
ge
le
vels
±
ℎ
,
1
an
d
±
ℎ
,
1
,
cel
l
1
pro
du
ces
±
1
2
⁄
,
1
and
±
,
1
, r
es
pecti
vely
. T
he values
of
ℎ
,
and
ℎ
,
are
deter
mined b
y:
ℎ
,
=
1
4
,
,
ℎ
,
=
3
4
,
(3)
Fo
r
mat
hemati
cal
an
al
ys
is,
th
e ref
e
re
nce
for t
he
hi
gh
-
vo
lt
ag
e cel
l i
s d
e
fine
d
as:
∗
=
2
∗
=
sin
(
)
(4)
This
volt
age
re
fer
e
nce
is
co
m
par
e
d
with
the
two
co
ns
ta
nts
ℎ
,
2
and
ℎ
,
2
.
T
hen
the
gating
sig
nals
are
gen
e
rated
to
t
he
sw
it
ches
of t
he hig
h
-
volt
ag
e cel
l usin
g
l
ogic
g
at
es.
13
-
l
e
v
e
l
o
u
t
p
u
t
G
a
t
i
n
g
S
i
g
n
a
l
s
t
o
c
e
l
l
1
G
a
t
i
n
g
s
i
g
n
a
l
s
t
o
c
e
l
l
2
h
H
,
2
h
L
,
2
h
L
,
1
h
H
,
1
L
2
L
1
v
1
v
2
v
dc
,
2
-
v
dc
,
2
-
v
dc
,
1
v
dc
,
1
2
h
L
,
2
Figure
3. Im
pl
ementat
io
n of
t
he nearest l
e
ve
l con
t
ro
l
meth
od
The refe
rence
for t
he
l
ow
-
volt
age cell
is
giv
e
n by
:
1
∗
=
2
∗
−
2
ℎ
,
2
∗
2
(5)
wh
e
re
2
is
the
switc
hing
pa
tt
ern
res
ulti
ng
from
the
high
-
volt
age
cel
l
after
co
mp
a
r
ison
a
nd
is give
n b
y:
2
=
[
(
2
∗
>
ℎ
,
2
)
+
(
2
∗
>
ℎ
,
2
)
]
−
[
(
2
∗
<
−
ℎ
,
2
)
+
(
2
∗
<
−
ℎ
,
2
)
]
(6)
wh
e
re
the
res
ul
t
of
eac
h
c
omparis
on
is
1
if
true,
or
0
if
not.
I
n
the
same
way,
t
he
refe
re
nce
f
or
the
low
-
volt
age ce
ll
is com
pa
red
with
ℎ
,
1
and
ℎ
,
1
. T
he
n
the
gatin
g
si
gn
al
s a
re
gen
e
r
at
ed
us
i
ng lo
gic g
at
es to t
he
corres
pondin
g
switc
hes
of
the
lo
w
-
volt
a
ge
c
el
l.
The
s
witc
hi
ng
patte
r
n
res
ulti
ng
f
rom
t
he
lo
w
-
volt
age
c
el
l
is
1
and
is
cal
c
ulate
d
in
t
he
same
wa
y
as
(6),
c
onsideri
ng
t
he
re
fer
e
nce
a
nd
const
ant
val
ue
s
for
the lo
w
-
volt
ag
e cel
l.
4.
ANALY
SIS
O
F THE I
NV
E
RTER
OUTP
UT VOL
TA
G
E
The
outp
ut
w
avefor
m
of
th
e
TCHB
in
vert
er
is
a
quarte
r
-
wav
e
s
ym
me
try;
the
refo
re,
the
outp
ut
vo
lt
age
ca
n be
represe
nted usi
ng the
Fou
rier
series as:
(
)
=
∑
sin
(
)
∞
n
=
1
(7)
wh
e
re
is t
he
F
ourier
co
e
ff
ic
ie
nt
. I
t e
qu
al
s z
er
o for e
ven
; howev
e
r, f
or od
d
it
is g
ive
n
by:
=
2
∑
=
1
,
3
,
5
,
…
(8)
Evaluation Warning : The document was created with Spire.PDF for Python.
IS
S
N
:
2088
-
8
694
In
t J
P
ow
Ele
c
&
D
ri
S
ys
t,
V
ol
.
11
, N
o.
3
,
Se
ptembe
r
2020
:
14
06
–
14
14
1410
wh
e
re
is
the
num
ber
of
switc
hing
a
ng
le
s
i
n
a
quarter
-
wa
ve
and
are
the
s
witc
hing
a
ng
le
s
w
hich
sh
oul
d
be
within
[0,
2
⁄
].
Tab
le
2
sho
ws
th
e
numb
e
r
of
ang
le
s
a
nd
vo
lt
age
le
vels
f
or
a
ra
ng
e
of
th
e
modu
la
ti
on
in
dex.
T
he
num
ber
of
le
vels
reduces
si
gn
i
ficantl
y
as
t
he
modu
la
ti
on
in
dex
dec
reases.
Th
e
maxim
um
nu
mb
e
r
of
le
vels
achieve
d
us
in
g
the
a
dopte
d
config
ur
at
io
n
i
s
13
le
vels.
Th
e
switc
hing
a
ngle
s
f
or
the
13
-
le
vel
A
-
TCHB
in
ver
te
r
at
≥
0
.
917
dep
e
nd
only
on
ℎ
,
1
,
since
th
e
ou
t
pu
t
vo
lt
a
ge
remains
t
he
s
am
e
wh
e
n
t
he
i
nv
e
rter
s
witc
hes
a
t
ℎ
,
1
,
as
ca
n
be
seen
i
n
Fi
gure
2.
T
her
e
fore,
The
switc
hing
a
ng
le
s
can
be
cal
culat
ed
as
f
ollow
s:
1
=
−
1
(
ℎ
,
1
)
,
2
=
−
1
(
3
ℎ
,
1
)
,
3
=
−
1
(
5
ℎ
,
1
)
4
=
−
1
(
7
ℎ
,
1
)
,
5
=
−
1
(
9
ℎ
,
1
)
,
6
=
−
1
(
11
ℎ
,
1
)
(9)
The
t
otal ha
rm
on
ic
disto
rtio
n (THD)
is
giv
e
n b
y:
THD
=
1
1
√
∑
2
∞
=
2
(10)
wh
e
re
co
uld
be
the
vo
lt
a
ge
(
)
or
t
he
c
urre
nt
(
).
T
he
f
unda
mental
vo
lt
age
1
,
wh
ic
h
is
e
qu
al
to
1
, can be
calc
ul
at
ed
at
a
ny v
al
ue of
us
in
g (8)
and
(9).
5.
RESU
LT
S
AND DI
SCUS
S
ION
5.1.
Simul
at
i
on
re
sults
The
sim
ulati
on
of
the
13
-
le
vel
A
-
TCHB
i
nv
e
rter
a
nd
th
e
a
pp
li
e
d
NL
C
method
is
c
arr
ie
d
out
in
M
at
la
b/Sim
ulink.
T
he
sim
ulati
on
pa
ramete
rs
wer
e
sel
ect
ed
t
o
matc
h
tho
s
e
of
the
e
xp
e
rime
ntal
se
tup
as
sh
ow
n
i
n
Ta
ble
3
t
o
facil
it
at
e
analy
sis
an
d
c
omparis
on
of
t
he
resu
lt
s.
The
volt
age
T
H
D
evaluate
d
th
rough
ou
t
a
range
of
is
sh
ow
n
in
Fi
gur
e
4.
It
is
obser
ved
t
hat
the
mi
nimum
volt
age
TH
D
is
achie
ved
at
=
1
.
044
and
is
5.18
%
.
It
ca
n
al
so
be
no
ti
ced
t
hat
as
inc
reases,
the
T
HD
decr
ea
se
s.
H
ow
e
ver,
af
te
r
=
1
.
044
,
t
he
THD
inc
reases
with
the
i
ncr
e
ase
in
.
It
ca
n
be
i
nf
e
rr
e
d
fro
m
Fig
ure
4
t
ha
t
the
NLC
m
et
hod
presents
hi
gh
THD
f
or
in
ve
rters
with
a
re
duced
num
ber
of
le
vels
an
d
with
low
m
odulati
on
in
dices
[
3]
,
a
nd
this
is
the
mai
n
dr
a
w
back o
f
th
e N
LC
meth
od.
Table
2.
N
umb
er
of
a
ngle
s a
nd
vo
lt
age
level
s acco
rd
i
ng to
t
he
m
odulati
on
ind
e
x
Mod
u
latio
n
ind
ex
(
)
Nu
m
b
er
o
f
An
g
les
Vo
ltag
e
lev
els
≥
0
.
917
6
13
0
.
75
≤
≤
0
.
916
5
11
0
.
584
≤
≤
0
.
749
4
9
0
.
417
≤
≤
0
.
583
3
7
0
.
25
≤
≤
0
.
416
2
5
0
5
10
15
20
25
30
35
0
.2
0
.3
0
.4
0
.5
0
.6
0
.7
0
.8
0
.9
1
1
.1
1
.2
1
.3
V
ol
t
a
g
e
T
H
D
(
%
)
M
odul
a
t
i
on
i
nde
x
13
-
le
v
e
l
9
-
le
v
e
l
11
-
le
v
e
l
7
-
le
v
e
l
5
-
le
v
e
l
=
5
.
18%
Figure
4. TH
D
v
e
rsu
s
m
odula
ti
on
in
de
x usin
g NLC
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
Ha
r
m
on
ic
mi
nimiza
ti
on
of
a
si
ng
le
-
ph
as
e
as
y
mm
e
tri
c
al TC
HB mu
lt
il
evel
inverte
r
…
(
Wai
l Al
i Al
i Sa
le
h
)
1411
The
sim
ulati
on
res
ults
of
the
inv
e
rter
outp
ut
vo
lt
age
an
d
cu
rr
e
nt
usi
ng
RL
load
at
=
1
are
show
n
in
Fi
gures
5
(a
)
a
nd
(
b)
,
res
pe
ct
ively.
T
he
THDs
f
or
bo
t
h
volt
age
a
nd
c
urren
t
are
s
ho
wn
in
Fig
ur
es
5
(c)
a
nd
(d),
re
sp
ect
ivel
y.
T
he
volt
age
THD
is
5.5
3%,
and
t
he
cu
rr
e
nt
THD
is
3.6
9%
at
=
100
Ω
and
=
1
5
mH
.
If
a b
ig
ge
r
in
duct
ance is
us
e
d,
t
he
c
urre
nt T
HD
will
r
ed
uce,
and its
wav
e
form
w
il
l be
more c
lose to
sin
usoidal.
(a)
(b)
(c)
(d)
Figure
5. Sim
ul
at
ion
r
es
ults
usi
ng
RL
l
oad at
=
1
(a
) ou
t
pu
t
volt
age w
a
ve
form
, (b) loa
d
c
urre
nt w
a
ve
form,
(c)
outp
ut volt
age T
H
D
a
nd (d) cu
rr
e
nt T
H
D
By
co
mp
a
rin
g
the
res
ults
wit
h
th
os
e
ob
ta
i
ne
d
in
[26
]
usi
ng
the
S
HE
met
hod
f
or
t
he
sam
e
num
ber
of
le
vels,
it
was
note
d
that
th
e
m
inimum
volt
ag
e
THD
us
i
ng
t
he
S
HE
met
hod
is
6.7
0%,
w
hi
le
in
the
case
of
t
he
pro
po
se
d
NLC
meth
od,
t
he
m
inimum
ac
hiev
ed
vo
lt
a
ge
T
H
D
is
5.1
8%,
as
sh
ow
n
i
n
Fi
gur
e
5,
w
hic
h
in
di
cat
es
a
sig
nifica
nt
i
mpro
veme
nt
i
n
t
he
volt
age
THD.
A
nothe
r
c
omparis
on
is
ma
de
with
[19]
f
or
t
he
sam
e
c
on
tr
ol
method,
t
he
sa
me
num
ber
of
le
vels,
but
with
dif
fer
e
n
t
as
ymmet
rical
M
L
I
topolo
gy.
It
was
note
d
t
hat
at
=
1
,
the
TH
D
val
ue
s
of
vo
lt
ag
e
and
c
urre
nt
in
[
19]
are
6.30%
and
0.3
5%,
re
s
pecti
vely
,
w
hich
are
very
cl
ose
to
the
res
ults
ob
t
ai
ned
i
n
t
his
study
(5.51%
and
0.7
0%
f
or
the
volt
age
and
cu
rr
e
nt
T
HD
s
,
respec
ti
vely
),
consi
der
i
ng si
mil
ar v
al
ues of
R
a
nd
L
i
n bo
t
h
st
ud
ie
s.
Ho
w
ever, the
num
be
r of
DC s
ourc
es is hi
gh
e
r
i
n
[19]
.
5.2.
Ex
peri
menta
l
va
li
d
at
i
on
To
te
st
t
he
functi
on
al
it
y
of
the
in
ver
te
r
a
nd
it
s
c
on
tr
ol
meth
od,
a
prototype
was
buil
t
in
the
la
boratory.
T
he
co
ntr
ol
al
go
r
it
hm
a
nd
the
ge
ner
at
io
n
of
ga
ti
ng
si
gn
al
s
we
re
ca
rr
ie
d
ou
t
us
in
g
a
n
FP
G
A
boar
d
(Cyclo
ne
I
V
EP4
CE
22F
17C6N).
The
e
xperime
ntal
set
up
is
s
how
n
in
Fig
ur
e
6,
a
nd
the
detai
le
d
s
ys
te
m
par
a
mete
rs
are
li
ste
d
in
Ta
bl
e
3.
A
no
t
her
low
-
c
os
t
F
PGA
(C
yclo
ne
IV
EP4
CE
6E
22
C
8)
is
us
e
d
t
o
pro
vid
e
a
dead
ti
me
of
2
µs
f
or
the
swi
tc
hin
g
sig
nals
to
pre
ven
t
t
he
switc
hes
from
hav
i
ng
a
s
hoot
-
thr
ough
fa
ult.
The
FPGAs
wer
e
processe
d
us
in
g
Qu
a
rtus
I
I
s
of
t
war
e
a
nd
pro
gr
a
mme
d
in
the
V
H
DL
la
ngua
ge.
T
he
outp
ut
vo
lt
age
a
nd
c
urre
nt
of
the
in
ve
rter
a
re
dis
played
on
the
Te
kt
ronix
TD
S
20
24
C
osc
il
losco
pe.
Th
e
vo
lt
a
ge
an
d
current
TH
Ds are meas
ured
usi
ng the
Flu
ke 435
powe
r q
ua
li
ty an
al
yze
r.
Evaluation Warning : The document was created with Spire.PDF for Python.
IS
S
N
:
2088
-
8
694
In
t J
P
ow
Ele
c
&
D
ri
S
ys
t,
V
ol
.
11
, N
o.
3
,
Se
ptembe
r
2020
:
14
06
–
14
14
1412
T
C
H
B
i
n
v
er
t
e
r
L
o
w
-
v
o
l
t
a
g
e
s
u
p
p
ly
O
s
il
lo
s
c
o
p
e
F
lu
k
e
435
H
ig
h
-
v
o
l
t
a
g
e
s
u
p
p
l
y
L
l
o
a
d
R
l
o
a
d
V
o
lt
a
g
e
p
r
o
b
e
F
P
G
A
s
G
a
t
e
d
r
iv
e
s
Table
3.
Sy
ste
m p
a
ramete
rs f
or ex
per
ime
nta
l t
est
s
Para
m
eter
Valu
e
DC v
o
ltag
e f
o
r
ce
l
l 1
6
0
V
DC v
o
ltag
e f
o
r
ce
l
l 2
1
2
0
V
IGBT
IRG4
PC5
0
UD
Po
wer
d
io
d
es
RHR
G3
0
1
2
0
DC
-
lin
k
capacito
rs
2
2
0
0
µF
Load
r
esis
tan
ce
1
0
0
Ω
Load
ind
u
ctan
ce
1
5
m
H
Fu
n
d
am
en
tal f
requ
en
cy
5
0
Hz
Figure
6. The
e
xp
e
rime
ntal set
up
The
ex
per
i
men
ta
l
resu
lt
s
f
or
t
he
outp
ut
vo
lt
a
ge
a
nd
c
urre
nt
of
the
in
ver
te
r
at
=
1
usi
ng
RL
l
oad
are
s
how
n
i
n
F
igure
7.
It
ca
n
be
see
n
t
hat
13
volt
age
le
vels
app
ea
r
at
t
he
outp
ut
an
d
the
c
urren
t
a
ppr
oac
hes
a
sinu
s
oid
al
s
ha
pe.
T
he
TH
Ds
of
bo
t
h
volt
ag
e
and
c
urre
nt
are
5.7%
a
nd
3.3%,
as
s
how
n
in
Fig
ur
es
8
(a
)
an
d
(b),
resp
ect
iv
el
y.
It
was
obse
rv
e
d
f
r
om
the
exp
e
rime
nts
th
at
the
minim
um
volt
age
a
nd
current
T
H
Ds
wer
e
recorde
d
at
=
1
.
044
to
be 5.
2%
a
nd
2.7%,
r
es
pecti
ve
ly.
Figure
7. Ex
pe
rimental
resu
lt
s of
ou
t
pu
t
volt
age a
nd curre
nt
at
=
1
us
i
ng
RL
l
oad
(a)
(b)
Figure
8. M
ea
s
ur
e
d res
ults usi
ng
RL
loa
d
at
=
1
(a)
V
oltage T
H
D
a
nd (b) C
urr
ent T
HD
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
Ha
r
m
on
ic
mi
nimiza
ti
on
of
a
si
ng
le
-
ph
as
e
as
y
mm
e
tri
c
al TC
HB mu
lt
il
evel
inverte
r
…
(
Wai
l Al
i Al
i Sa
le
h
)
1413
It
shou
l
d
be
note
d
that
the
NL
C
method
does
no
t
el
imi
na
te
sp
eci
fic
harmo
nics
li
ke
the
S
HE
met
hod,
wh
ic
h
el
imi
nates
s
om
e
lo
w
-
orde
r
ha
rm
on
ic
s;
rat
her
,
it
minimi
zes
t
he
ov
erall
T
HD
of
t
he
i
nverter
ou
t
pu
t.
The
sim
ulati
on
and
ex
per
ime
ntal
resu
lt
s
of
t
he
volt
ag
e
T
H
D
we
re
c
ompa
red
at
diff
e
re
nt
values
of
M
usi
ng
R
load
as
s
how
n
in
Fi
gure
9.
It
is
no
te
d
that
t
he
sim
ulati
on
a
nd
ex
per
ime
ntal
res
ults
are
in
c
lose
a
greeme
nt
wit
h
each
oth
e
r.
Figure
9. Total
h
a
rm
on
ic
dist
or
ti
on
versus
modu
la
ti
on i
ndex
6.
CONCL
US
I
O
N
In
this
pap
e
r,
t
he
NLC
met
ho
d
was
a
ppli
ed
to
a
13
-
le
vel
A
-
TC
HB
in
vert
er.
The
as
ymmet
ry
in
t
he
DC
s
upplies
of
the
TC
HB
cel
ls
res
ults
i
n
a
n
inc
rease
d
nu
mb
e
r
of
outp
ut
le
vels
a
nd
the
refor
e
le
ss
T
H
D.
T
he
asym
metri
cal
t
opologies
util
iz
e
a
re
du
ce
d
num
ber
of
DC
p
owe
r
s
upplies
an
d
s
witc
hes
to
pro
du
ce
the
same
numb
e
r
of
le
vels
c
ompare
d
to
symmet
ri
cal
ones.
T
he
NLC
met
hod
is
co
nce
ptua
l
an
d
sim
ple
to
be
impleme
nted.
The
13
-
le
vel
outp
ut
sta
rts
at
=
0
.
917
.
A
minim
um
vo
lt
age
T
H
D
of
5.1
8%
is
ac
hieve
d
at
=
1
.
044
.
T
he
sim
ulati
on
re
su
lt
s
i
nd
i
cat
e
the
e
ff
ic
ie
ncy
of
this
me
thod
for
obta
in
ing
hi
gh
-
qual
it
y
outp
ut.
The
T
H
D
is
ca
lc
ulate
d
an
d
a
na
lyzed
at
dif
fere
nt
val
ues
of
t
he
m
odulati
on
ind
e
x.
T
he
e
xp
erimental
re
su
l
ts
are
in
cl
os
e
a
gr
ee
ment
with
the
simulat
ion
res
ults.
T
he
N
LC
meth
od
obta
in
s
a
lowe
r
vo
lt
a
ge
T
HD
c
omp
ared
t
o
oth
e
r
l
ow
-
s
witc
hing
freq
ue
nc
y
mod
ulati
on
methods
,
e
ve
n
th
ough
the
re
qu
i
red
m
odula
ti
on
i
ndex
to
pro
duce
the
13
-
le
vel
ou
tpu
t
is
c
onside
red
hi
gh
(
M
≥
0.917
).
D
ue
to
that,
t
he
NLC
meth
od
is
bette
r
s
uited
for
a
high
numb
e
r
of
le
ve
ls
and
high
m
odulati
on
i
nd
ic
es.
Furthe
r
stu
dies
are
need
e
d
to
in
vestigat
e
the
perf
or
ma
nce
o
f
the meth
od i
n cl
os
ed
-
lo
op appli
cat
ion
s.
ACKN
OWLE
DGME
NTS
This
wor
k
was
sup
ported
u
nd
er
re
searc
h
gr
a
nt
PJ
P/2
019/F
KE(1B)/S
0166
5
from
U
niv
e
r
sit
i
Tekn
i
kal
M
al
aysia
M
el
a
ka.
REFERE
NCE
S
[1]
J.
Rodrigu
ez,
L.
G.
Franque
lo
,
S.
Kouro,
J.
I.
Leon,
R
.
C
.
Port
il
l
o,
M.
A.
M.
Pra
ts,
and
M.
A.
Pe
rez
,
“Mult
ileve
l
conve
rt
ers:
An
ena
bl
ing
te
chno
l
ogy
for
high
-
po
wer
app
li
c
ations,”
Proc
.
I
EEE
,
vol.
97,
no.
11
,
pp.
1786
–
1817
,
Nov.
2009.
[2]
H.
Akagi
,
“Multile
v
el
Converter
s:
Fundam
ent
a
l
Circ
uit
s
and
Sys
te
ms,
”
Proc.
I
E
EE
,
vol
.
105,
n
o.
11
,
pp.
2048
–
2065,
2017
.
[3]
L.
G
.
Franque
lo
,
J.
Rodriguez,
J.
I.
Le
on
,
S.
K
ouro,
R.
Porti
llo,
and
M.
A
.
M.
Prats,
“T
h
e
age
of
multil
ev
el
conve
rt
ers
arr
ives
,
”
IE
EE Ind. El
ec
tron.
Mag
.
,
vo
l.
2
,
no
.
2
,
pp
.
28
–
39,
Jun.
2008.
[4]
P.
M.
Meshra
m
and
V.
B.
Bor
ghat
e
,
“A
simpl
ifi
ed
nea
r
est
le
v
el
cont
rol
(NLC
)
voltage
b
alanc
ing
m
et
hod
for
m
odula
r
multil
e
vel
conve
rt
er (MM
C),
”
IE
EE Tr
ans.
Powe
r
Elec
tr
on.
,
vol. 30, no.
1,
pp
.
450
–
462
,
2015.
[5]
R.
J.
Satpu
taley,
V.
B.
Borgha
te,
V.
Kum
ar
and
T
.
Kum
ar,
"Ex
per
imental
inve
s
ti
gation
of
n
ew
thre
e
phase
f
ive
-
le
ve
l
tr
ansistor c
la
mp
ed
H
-
bridg
e
inve
r
te
r
,
"
E
PE J.,
vol. 27, no. 1,
pp.
31
-
42
,
2017
.
[6]
N.
A.
Rahim,
M.
F.
M
.
El
i
as,
and
W
.
P.
Hew
,
“T
r
ansistor
-
c
lamped
H
-
bridg
e
base
d
ca
sca
d
ed
mul
tilevel
inve
r
t
er
with
new me
tho
d
of
c
apaci
tor
vo
lt
ag
e
ba
la
n
ci
ng,
”
IEEE
Tr
ans.
In
d.
E
lectron.
,
vol
.
60,
no.
8,
pp.
29
4
3
–
2956,
2013
.
[7]
J.
P.
and
S.
K.
M.
Per
ez,
J.
Ro
drigue
z
,
“Power
Distribu
ti
on
in
Hybrid
Mult
i
-
c
e
ll
Converter
wit
h
Ne
are
st
L
evel
Modulat
ion
,
”
20
07
IEEE
In
t. Sym
p.
Ind. Elec
tron.
,
pp.
736
–
741,
2007.
[8]
M.
F.
M.
E
li
as
,
N.
A.
Rahim,
H.
W.
Ping
,
and
M.
N.
Udd
in
,
“As
ymm
et
r
ical
ca
sc
ade
d
mu
lt
i
le
v
el
i
nver
te
r
b
ase
d
on
tra
nsistor
-
c
la
mp
ed
H
-
bridge pow
er
c
el
l
,
”
IE
EE Tr
ans.
Ind. Appl.
,
vol.
50
,
no
.
6
,
pp
.
4281
–
4288
,
No
v.
2014
.
[9]
S.
P.
Gaut
am
,
S.
Gupta
,
and
L
.
Kumar
,
“Re
l
ia
bi
li
ty
im
prove
me
n
t
of
tr
ansistor
c
l
am
ped
H
-
br
idge
-
base
d
c
asc
ad
ed
mul
tilevel
inve
rt
er,
”
IET
Powe
r
El
e
ct
ron.
,
vol
.
1
0,
no
.
7
,
pp
.
770
–
781,
2017
.
[10]
Y.
Zh
ang,
J.
He
,
S.
Padm
ana
ban
and
D.
M.
Ion
el,
"Tra
nsistor
-
C
lamped
Multi
le
ve
l
H
-
Bridge
Inv
erter
in
Si
and
SiC
Hybrid
Configu
rat
ion
for
High
-
Eff
icienc
y
P
hotovol
ta
i
c
Appli
cations,
"
2018
I
E
EE
En
ergy
Con
ve
rs
ion
Congress
and
Ex
posi
ti
on
(
ECCE
)
,
Portla
nd
,
OR,
2018,
pp.
2536
-
2542.
Evaluation Warning : The document was created with Spire.PDF for Python.
IS
S
N
:
2088
-
8
694
In
t J
P
ow
Ele
c
&
D
ri
S
ys
t,
V
ol
.
11
, N
o.
3
,
Se
ptembe
r
2020
:
14
06
–
14
14
1414
[11]
Y.
Babkra
n
i,
A.
Nadda
mi
and
M.
Hil
al
,
"A
smart
ca
sca
d
ed
H
-
bridge
mu
lt
i
l
eve
l
inve
r
te
r
wi
th
An
optimize
d
modul
ation
t
ec
h
nique
s
inc
r
ea
sin
g
the
qu
al
i
ty
and
red
uc
ing
har
mo
nic
s,"
Int
ernati
o
nal
Journal
o
f
P
ower
El
e
ct
roni
c
s
and
Dr
iv
e
S
ystem
s (IJ
PE
DS),
vo
l.
10
,
no
.
4
,
pp
.
1
852
-
1862,
2019
.
[12]
S.
Sanusi,
A
.
Ji
din,
T.
Sutikno,
K.
A.
Kari
m
,
M.
L.
M.
Ja
mi
l
,
and
S.
A.
A
.
T
aru
san,
“Im
p
le
m
ent
a
ti
on
of
spac
e
vec
tor
modulato
r
for
ca
sc
ade
d
h
-
bridge
mul
t
il
ev
e
l
inv
ert
e
rs,”
Inter
nati
onal
Journal
of
Pow
er
E
le
c
t
ronics
and
Dr
ive
Syste
ms
(IJ
PE
D
S),
vol
.
6
,
no
.
4
,
pp.
906
–
918
,
20
15.
[13]
M.
S.
A.
Da
hidah,
G.
Kons
ta
n
ti
n
ou,
and
V.
G.
Ageli
dis
,
“A
R
eview
of
Multi
l
ev
el
Selecti
v
e
Harm
onic
E
li
m
ination
PWM:
Formula
t
ions,
Solving
Al
gorit
hms,
I
mple
me
nt
at
ion
and
A
ppli
c
at
ions,
”
in
I
EE
E
Tr
ans.
Pow
er
El
e
ct
ron.
,
vol
.
30,
no
.
8
,
pp
.
40
91
–
4106,
2015
.
[14]
M.
Srndovic
,
A.
Zhete
ss
ov,
T.
Aliz
ad
eh,
Y.
L.
Fami
liant
,
G.
G
ran
di
and
A
.
Ru
der
ma
n
,
"S
im
u
ltaneous
Sel
ective
Harm
onic
El
i
minati
on
and
THD
Minim
izati
on
f
or
a
Single
-
Phas
e
Multi
l
evel
Inv
ert
er
wi
th
Stai
r
c
ase
Modulation,
"
IEE
E
Tr
ans.
Ind
.
App
l.
,
vo
l. 54,
no.
2
,
pp
.
1532
-
1541,
2018
.
[15]
W.
A
.
Ha
li
m
,
N.
A.
T.
Te
ngku
,
K.
Applasa
m
y,
an
d
A.
Jidin,
“S
el
e
ct
iv
e
Har
mon
ic
El
i
mi
na
ti
on
Based
on
N
ewton
-
rap
hson
Method
for
Casca
d
ed
H
-
bridge
Mult
il
e
ve
l
Inve
r
te
r,”
In
te
r
nati
onal
Journal
of
Pow
er
Elec
tronic
s
and
Dr
ive
Syste
ms
(IJ
PE
D
S
)
,
vol
.
8
,
no
.
3
,
pp.
1193
–
1202
,
2017.
[16]
M.
Rashe
ed,
R
.
Omar
,
M.
S
ula
i
ma
n
and
W
.
A.
Hali
m
,
"A
modified
ca
s
ca
ded
h
-
bridge
multil
evel
inv
erter
base
d
on
par
ticle
sw
arm
opti
mi
sa
ti
on
(
P
SO
)
te
chn
ique,
"
Indone
sian
Jo
urnal
of
El
e
ct
ri
cal
Engi
ne
ering
and
Computer
Sci
en
ce
(IJ
E
ECS
)
,
vol
.
16
,
no
.
1
,
pp.
41
-
51
,
2019
.
[17]
J.
Rodrígu
ez,
L
.
Morán
,
P.
Co
rre
a,
and
C
.
Sil
va,
“A
vector
c
ontrol
technique
for
me
d
ium
-
vo
lt
ag
e
mul
t
il
ev
el
inve
rt
ers,
”
IEEE
Tr
ans.
Ind.
Elec
tron.
,
vo
l. 49, no
.
4
,
pp
.
882
–
888
,
2002.
[18]
R.
Kumar
and
D.
Tha
ng
avelusam
y
,
“A
mod
ified
ne
are
st
le
v
el
modul
a
ti
on
s
ch
em
e
for
a
sym
me
tric
ca
sc
ade
d
H
-
bridge
inve
rt
er,”
Gaz
i
Univ
.
J. Sci.
,
vol
.
32
,
no
.
2
,
pp.
471
–
481
,
20
19.
[19]
M.
D.
Sidd
ique
,
S.
Mekhi
le
f
,
N.
M.
Shah
,
A.
Sarwar,
A.
Iqb
al,
M.
T
ayya
b
,
an
d
M.
K.
Ans
ari,
"Low
Sw
it
chi
n
g
Freque
ncy
Base
d
As
ymm
et
ri
ca
l
Multi
le
v
el
Inv
er
te
r
Topo
logy
With
Reduced
Sw
itch
Count,
"
in
IE
EE
A
cc
ess
,
vo
l.
7,
pp
.
86374
-
86
383,
2019
.
[20]
R.
R.
Kar
asa
ni,
V.
B.
Borghate,
P.
M.
Meshram,
and
H.
M.
Sury
awa
nshi
,
“A
Mo
difi
ed
Sw
itched
-
Diode
Topol
ogy
for
Casca
d
ed
M
ult
ilevel Inverter
s,”
J
.
Powe
r
Ele
ct
ron.
,
vol
.
16
,
n
o.
5
,
pp
.
1706
–
1
715,
2016
.
[21]
S.
Kouro,
R
.
B
ern
al,
H.
Miran
da,
C.
a
Silva,
and
J.
Rodríguez,
“High
-
p
erf
or
ma
nc
e
torque
a
nd
flux
cont
ro
l
for
mul
tilevel
inve
rt
er
fed
indu
ct
ion
mot
ors,”
IE
EE
T
rans
.
Powe
r
El
e
ct
r.
,
vol
.
22
,
no
.
6,
pp
.
2116
–
212
3,
2007
.
[22]
G.
Si,
J.
Zhu,
Y
.
Lei,
L.
Jia,
an
d
Y.
Zha
ng
,
“A
n
enha
nc
ed
l
evel
-
inc
r
ea
sed
n
ea
r
est
le
v
el
modul
a
ti
on
for
modu
la
r
mul
tilevel
converte
r,”
In
t. Tr
ans. Electr. E
nergy
S
yst.
,
vol
.
29
,
no
.
1,
pp
.
1
–
17
,
201
9.
[23]
P.
Hu
and
D.
Jia
ng,
"A
le
ve
l
-
increa
sed
ne
are
st
level
modulat
ion
me
thod
for
mod
ula
r
multil
evel
c
onver
te
rs,"
I
EEE
Tr
ans.
Powe
r E
l
ec
tr.
,
vo
l. 30, no
.
4
,
pp
.
1836
-
18
42,
2015
.
[24]
D.
Wu
and
L.
Peng,
"Characteri
stic
s
of
ne
are
s
t
le
v
el
modu
la
t
i
on
me
thod
wi
th
ci
rcu
la
t
ing
cur
ren
t
con
trol
for
modul
ar
multil
e
vel
conve
rt
er,
"
I
ET
Powe
r
El
e
ct
r
on.
,
vol. 9, no. 2, pp. 155
-
164,
20
15.
[25]
J.
W
ang,
X.
Ha
n,
H.
Ma
,
and
Z
.
B
ai
,
“A
hybrid
modular
multil
e
vel
conve
r
te
r
to
pology
wit
h
an
i
mprove
d
nea
r
est
le
ve
l
modu
la
t
ion
method,
”
J
.
Po
wer
Elec
tron
.
,
v
ol.
17
,
no
.
1
,
pp
.
96
–
105,
2017
.
[26]
W.
A.
Ha
li
m
,
N
.
A.
Rah
im
,
and
M.
Azri
,
“Ge
ne
ral
i
ze
d
sel
ective
har
monic
e
li
minati
on
modul
a
tion
for
tra
nsistor
-
cl
a
mpe
d
H
-
br
idg
e
mu
lt
i
le
ve
l inve
rte
r,”
J
.
Powe
r
E
le
c
tron.
,
vol. 15, no. 4, pp. 964
–
9
73,
2015
.
Evaluation Warning : The document was created with Spire.PDF for Python.