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
. 40
4~
41
4
I
S
SN
: 208
8-8
6
9
4
4
04
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
Convergence P
a
ramet
e
r Analysis
for Diff
erent M
e
t
a
heuri
s
tic
Meth
ods Control Constan
t
Estim
a
tion and it’s Trad
eoff
Inference
R
.
Saga
ya
ra
j *
,
S.
Tha
n
ga
vel
*
*
* Department of
Electrical and
El
ectron
i
cs Eng
i
neering, Pavai College
of
Technolo
g
y
, Namakkal, I
ndia
** Departm
e
n
t
o
f
El
ectr
i
c
a
l
&
El
ectron
i
cs
Eng
i
ne
ering,
K.S.Rang
asam
y Co
lleg
e
of
Tech
no
logy
, Tiru
chengod
e,
India
Article Info
A
B
STRAC
T
Article histo
r
y:
Received
Ja
n 30, 2015
Rev
i
sed
May 15
, 20
15
Accepted
May 28, 2015
This paper
is an extension of
our
previous
work, which discussed the
difficu
lty
in implementing differen
t
methods of resistance emulation
techn
i
ques on the hardware due
to its c
ontrol co
nstant estim
atio
n delay
.
I
n
order to get rid
of the delay
th
is paper
att
e
m
p
ts to include th
e m
e
ta-heur
i
stic
methods for the
control
constants of th
e con
t
roller. To
ach
iev
e
the minimum
Total Harmonic
Disturbance (TH
D
) in the AC si
de of the conver
t
er modern
meta-heuristic methods
are co
mpared
with
th
e tr
aditional methods.
The
convergen
ce par
a
meters, which
are primar
y
for
the earlier estim
ation of th
e
control cons
t
a
nt
s
,
are com
p
ared with the m
eas
ured param
e
ters
, t
a
bula
t
ed and
tradeoff in
feren
c
e is done among the me
thods. This kind of implementation
does not n
eed
th
e mathemat
ical
model of th
e s
y
s
t
em under stud
y for find
ing
the con
t
rol co
ns
tants
.
Th
e p
a
ram
e
ters
cons
i
d
ered for
es
ti
m
a
tion are
population
size,
maximum nu
mb
er of
epoc
hs,
an
d global best solution of
th
e
control constants, be
st THD value and exe
c
utio
n tim
e. Matlab
TM
/Sim
ulink
based sim
u
lation
is optim
ized wi
th the M-fil
e
bas
e
d optim
iz
ation
techn
i
ques
like Par
tic
le Swarm
Optim
izatio
n (PSO), Geneti
c Algorithm
(GA), Cucko
o
S
earch Algorith
m
,
Gravit
y
S
ear
ch Algorithm
,
Harm
ony
S
e
arc
h
Algorithm
and Bat Algor
ith
m.
Keyword:
Cu
ckoo
Sear
ch A
l
gor
ith
m
Grav
ity Search Algorith
m
Opt
i
m
i
zati
on T
echni
que
s
Particle Swarm Op
ti
m
i
zatio
n
Resistance Emulation
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
:
R. Sag
a
yaraj,
Depa
rtem
ent of Electrical a
n
d
El
ect
ro
ni
cs E
n
gi
nee
r
i
n
g,
Pavai
C
o
l
l
e
ge
of
Tech
n
o
l
o
gy
,
Pachal
,
Nam
a
kkal
6
3
7
01
8,
T
a
m
i
l
Nadu,
I
n
d
i
a.
Em
a
il: rsraj
eee@yaho
o
.co.in
1.
INTRODUCTION
M
ode
rn c
o
m
put
ers, c
o
m
m
u
n
i
cat
i
on a
nd e
l
ect
roni
c sy
st
e
m
s get
t
h
ei
r “Li
f
e B
l
oo
d”
fr
om
power
electronics
, which aids all
ene
r
gy
har
v
et
i
n
g
s
y
st
em
s
[1
]
.
Ene
r
gy
har
v
est
i
n
g
has
becom
e
a very
i
m
port
a
nt
fi
el
d
i
n
el
ect
ri
cal
engi
neeri
ng as
every
sm
al
l
am
ount
of e
n
e
r
gy
de
vel
o
ped
can be t
a
p
p
ed
fo
r u
s
e i
n
i
t
s
ow
n
mag
n
itu
d
e
. Apart form
th
e ap
p
lication
s
of
Power Elect
onics in energy
harvesting
syste
m
s, electric
m
o
tors
m
o
t
i
o
n
con
t
ro
l
,
redu
cing
th
e
n
o
i
se
g
e
n
e
ration
i
n
in
m
o
to
rs;
it p
l
ays a v
ital
ro
le in
im
p
r
ov
in
g
t
h
e m
o
to
r stead
y
state and
dyna
mic characteristics
[2]
.
Power
Factor C
o
rrection
(PFC) is a
prim
e factor tha
t
would i
n
creas
e the
p
o
wer lo
ss,
which
m
u
st b
e
in
trodu
ced
in
almo
st all th
e in
dustrial d
r
iv
e
u
n
i
t
.
Resistan
ce em
u
l
a
tio
n
is o
n
e su
ch
energy harvest
i
ng m
e
thod us
ed for
re
ne
wa
ble energy res
o
urces like th
e
wind e
n
ergy syste
m
. Even though
t
h
ere are l
o
w
po
we
r de
vi
ces
t
h
at
are de
vel
ope
d i
n
wi
rel
e
ss sens
or
net
w
or
k n
o
d
es, t
h
e
need
of
hi
g
h
-
d
en
si
t
y
p
o
wer is a
n
e
ed
i
n
th
e field ev
en tod
a
y.
Th
e Max
i
m
u
m
Power Po
in
t Track
i
ng
(MPPT) alg
o
rith
m
fo
r the wi
n
d
gene
rat
o
r ba
se
d co
n
v
ert
e
r i
s
appl
i
e
d
usi
ng t
h
e resi
st
a
n
ce em
ul
at
i
on t
echn
i
que.
The
bo
o
s
t
con
v
ert
e
r,
whi
c
h
wo
ul
d act
as a
n
M
PPT c
ont
r
o
l
l
e
d co
n
v
ert
e
r
[3]
.
T
h
e resi
st
ance em
ul
at
i
on
m
e
t
hod d
eal
s wi
t
h
t
h
e t
h
ree
pha
se
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
:
404
–
4
14
4
05
rect
i
f
i
e
r,
w
h
er
e a swi
t
c
he
d
re
si
st
ance em
ul
at
i
on m
e
t
hod i
s
i
n
t
r
o
d
u
ced
wi
t
h
t
w
o ca
paci
t
o
rs a
n
d
t
h
ree
re
si
st
or
s
are use
d
for s
h
api
ng the input curren
t at th
e AC sid
e
similar to
th
at
o
f
th
e
v
o
ltag
e
[4
]. Th
e resi
stan
ce
em
ul
at
i
on t
ech
ni
q
u
e fo
r ha
r
m
oni
c elim
i
n
at
i
on d
o
es’t
sen
s
es t
h
e i
n
p
u
t
vol
t
a
ge a
nd t
h
e l
o
ad cur
r
e
n
t
[5]
.
A
scalar syste
m
m
o
d
e
l with
th
e PI con
t
ro
ller
h
a
s b
e
en
in
trod
u
c
ed
to
d
e
v
e
l
o
p
a sing
le-phase sh
un
t active filter.
A highe
r
power fact
or
operation
of a three phase
rectifier im
pl
em
ent
a
t
i
on i
s
possi
bl
e on
a DSP
TMS32
0
F240F f
r
o
m
Tex
a
s I
n
stru
m
e
n
t
s w
a
s po
ssi
b
l
e an
d
th
e con
t
r
o
l
alg
o
r
ith
m
h
a
s b
een
g
i
v
e
n
respo
n
s
e
with
in
4
0
sec
.
To o
v
e
r
com
e
the exce
ssi
ve
o
v
ers
h
oot
a
nd
d
a
m
p
i
ng a
uni
v
e
rsal
m
e
t
hod
was u
s
ed
[6]
.
The m
e
t
hod
cal
l
e
d Kari
m
’
s
m
e
t
hod,
w
h
i
c
h u
s
ed t
h
e P
I
cont
rol
l
e
r
fo
r t
h
e o
u
t
e
r l
o
o
p
and t
h
e P
D
c
o
nt
r
o
l
l
e
r f
o
r t
h
e
i
nne
r
lo
op
in
o
r
d
e
r t
o
ach
iev
e
th
e ab
ov
e sai
d
criterion
s.
Also
it is it is d
i
fficu
lt fo
r the PID co
ntro
ller to
resp
ect well
to
ch
ang
e
s in
t
h
e op
erating
po
in
t, and
th
ey ex
h
i
b
it p
o
or
perform
a
nce when the syst
e
m
is su
bj
ected
to larg
e
l
o
ad
va
ri
at
i
ons
[7]
.
A si
m
p
l
e
PSO
(S
PS
O)
w
a
s use
d
t
o
re
je
ct the effect of
external
di
st
u
r
bance
an
d a
ssu
re t
h
e
out
put
. T
h
e P
I-P
D pa
ram
e
ter est
i
m
a
t
i
on was d
o
n
e by
sol
v
i
ng t
h
e S
PSO
pr
obl
em
s. The p
o
w
er
fact
or
co
rrectio
n
stage is b
u
ilt u
s
in
g th
e b
o
o
s
t conv
erter topo
l
ogy, w
h
ich
h
a
s the ad
v
a
n
t
ag
es
of
g
r
ou
nd
ed
tran
sistor,
sm
al
l
i
nput
i
n
d
u
ct
o
r
, si
m
p
l
i
c
ity
and hi
gh e
ffi
ci
ency
(9
5%)
[
8
]
.
The c
ont
r
o
l
l
er used
fo
r t
h
e
PFC
i
s
usual
l
y
a PI
cont
rol
l
e
r
[
8
]
and
f
o
r
eve
r
y
cont
rol
l
e
r t
h
e
m
a
t
h
em
at
i
c
al
m
odel
i
ng o
f
t
h
e ci
rcui
t
a
nd t
h
e c
ont
r
o
l
l
e
r c
onst
a
nt
est
i
m
a
ti
on m
u
st
be do
ne be
f
o
re ha
nd
, w
h
i
c
h
i
s
t
i
m
e
consu
m
i
ng an
d t
r
i
v
i
a
l
proce
ss. T
h
e
resi
st
ance em
ul
at
i
on
t
echni
q
u
e f
o
r t
h
e t
h
ree
-
p
h
ase
i
nduct
i
o
n m
o
tor d
r
i
v
e sy
st
em
i
s
t
a
ken and
im
pl
em
ent
e
d usi
n
g t
h
e t
ech
ni
q
u
e o
f
in
trodu
cing
three sing
le-ph
a
se in
v
e
rter and
t
h
e p
a
ssi
ve
powe
r fact
or correct
i
on ci
rcuit ele
m
ents [9].
Thi
s
pape
r at
t
e
m
p
t
s
t
o
de
v
e
l
op a
co
nt
r
o
l
l
e
r co
nst
a
nt
est
i
m
a
ti
on
usi
n
g
di
f
f
e
r
ent
o
p
t
i
m
i
zat
i
on
t
echni
q
u
e.
T
h
e
pr
o
p
o
s
ed
i
m
plem
ent
a
t
i
on i
s
devel
ope
d
usi
n
g a
DS
P
pr
oce
ssor
co
ul
d g
e
t
a res
p
o
n
se
t
h
at
i
s
o
f
m
i
cro seco
n
d
s
ran
g
e t
h
e o
p
t
i
m
i
zat
i
on t
echn
i
que ca
n
be a
d
ded
i
n
t
h
e est
i
m
at
i
on o
f
t
h
e c
ont
rol
c
o
nst
a
nt
s i
n
t
h
e
PI con
t
ro
ller of th
e resistan
ce e
m
u
l
atio
n
te
ch
n
i
q
u
e
.
Differen
t trad
ition
a
l
an
d
th
e m
o
d
e
rn
op
tim
iza
tio
n
are
t
a
ken f
o
r a
n
a
l
y
z
i
ng whi
c
h
wo
ul
d be c
o
m
put
at
i
onal
l
y
and ec
on
om
i
cal
l
y
eff
ectiv
e. Th
e op
timizatio
n
techniques
use
d
for the c
o
mparative
analy
s
is are
Pa
rticle Swarm
Op
timizatio
n
(PSO),
Gen
e
tic Algo
rith
m
(GA), Cuc
k
oo
Searc
h
Al
gorithm
,
Gravity Se
arch
Al
gorithm
,
Har
m
o
n
y
Search
Algorithm
an
d
Bat Algo
rith
m
.
The
param
e
t
e
rs co
nsi
d
e
r
e
d
f
o
r
est
i
m
a
t
i
on a
r
e
po
p
u
l
a
t
i
on s
i
ze, m
a
xim
u
m
num
ber
of
ep
o
c
hs,
an
d
gl
o
b
a
l
best
sol
u
t
i
o
n o
f
t
h
e cont
r
o
l
const
a
nt
s, best
TH
D val
u
e an
d exec
ut
i
on t
i
m
e. Th
e param
e
t
e
rs consi
d
ere
d
are t
h
o
s
e
,
whi
c
h hel
p
t
o
kn
o
w
whet
her
t
h
i
s
t
ech
ni
q
u
e can be
i
m
pl
em
ent
e
d o
n
t
h
e
ha
rd
ware
.
Th
is Pap
e
r is
o
r
g
a
n
i
zed
as fo
llo
ws.
A
b
r
ief abou
t resistan
ce em
u
l
atio
n
m
e
th
o
d
fills
Sectio
n-II;
Di
ffe
re
nt
opt
i
m
i
zat
i
on t
echn
i
ques a
r
e i
n
t
r
o
duce
d
i
n
Sect
io
n-
II
I.
Section-
IV
d
e
liv
er
s the idea about propose
d
sy
st
em
unde
r
anal
y
s
i
s
. Sect
i
o
n
-
V
deal
s
wi
t
h
t
h
e
resul
t
s
a
nd
di
sc
ussi
o
n
on t
h
e w
o
rk c
a
rri
ed
o
u
t
.
C
o
ncl
u
si
o
n
and the Re
fere
nce
follow i
n
t
h
e last Section.
2.
RESIST
AN
C
E
EMUL
ATI
O
N
METHO
D
The i
d
ea
behi
nd resista
n
ce
e
m
ulation is t
h
at the
ci
rcu
i
t
after t
h
e
b
r
i
d
g
e
rectifier in th
e
AC-DC
co
nv
erter circuit wo
u
l
d
ab
sorb
on
ly
p
u
r
e sinu
so
i
d
al cu
rren
t, wh
ich
is p
r
opo
rtion
a
l to
th
e AC sup
p
l
y v
o
l
tag
e
.
Thi
s
i
d
ea was
pre
v
i
o
usl
y
im
pl
em
ent
e
d usi
ng t
h
e
passi
ve
com
ponent
s
.
The resi
st
a
n
ce
em
ul
at
i
on t
echni
que
b
o
ils
do
wn
to sh
ap
i
n
g th
e i
n
pu
t curren
t
,
supp
ly b
e
ing
o
f
c
o
nst
a
nt
v
o
l
t
a
ge. The Av
era
g
e C
u
rrent M
o
de (ACM)
m
e
thod is a s
u
ccessful m
e
thod im
plem
ented for em
ulating resistance
by th
e use of
powe
r electronic de
vices.
The
boost c
o
nverters are
us
ually us
ed
for PFC in
m
a
n
y
Switch
e
d Mod
e
Power
Supp
ly (SMPS) ap
p
licatio
n
s
an
d th
e
sam
e
h
a
s b
e
en
tak
e
n up
in th
is
p
a
p
e
r
[1
].
T
h
e boost conve
r
ters are
natural ha
rmonic
re
duction devices
, as t
h
e capa
c
itor i
n
their l
o
ad
side
wo
ul
d el
i
m
i
n
at
e t
h
e secon
d
orde
r ha
rm
oni
cs i
n
t
h
e sup
p
l
y
si
de, hen
c
e i
t
i
s
i
n
ferre
d t
h
at
onl
y
t
h
e od
d
harm
oni
cs a
r
e
t
o
be t
a
ke
n
up
seri
ousl
y
a
n
d
t
h
e
P
W
M
t
echni
que
s ar
e de
vel
o
ped
t
o
wa
r
d
s re
d
u
ci
ng
o
r
el
im
i
n
at
i
ng t
h
e od
d
harm
oni
cs. The
P
W
M
cont
r
o
l
i
n
t
h
i
s
al
go
ri
t
h
m
i
s
ai
ded
by
t
h
e u
s
e of a P
I
c
o
n
t
rol
l
e
r
wh
ose c
ont
rol
con
s
t
a
nt
s are t
o
be
pre
d
et
er
m
i
ned i
n
o
r
d
e
r to
attain
th
e lo
west THD. Th
is p
a
p
e
r attem
p
ts
to
d
e
term
in
e th
ese p
a
ram
e
ters o
n
t
h
e ru
n, wh
ich
m
ean
s th
at the control
constants
are
determ
ined when t
h
e
syste
m
is ON can
b
e
tak
e
n
as a ro
ad
th
at this can
b
e
im
p
l
e
m
en
ted
on
the DSP
b
o
a
rd
s
[5
]. Th
e op
timizatio
n
al
go
ri
t
h
m
cons
i
d
ers T
H
D as t
h
e o
b
j
ect
i
v
e
fu
nct
i
o
n
,
w
h
i
c
h
m
u
st
be
m
i
nim
i
zed, a
nd t
h
e c
onst
r
ai
nt
s a
r
e t
a
ken
as the control
constant’s lim
its. This
novel
m
e
thod
of
det
e
rm
ining the c
ont
rol c
onsta
nt
s will have a
good
accuracy le
vel
as com
p
ared to the tra
d
itional
m
e
thods
.
3.
OPTIMIZ
A
T
I
ON TE
CH
NI
QUES
The traditional optim
ization techni
que
s like the gr
a
d
ient de
scent
m
e
thod a
nd
quasi ne
wt
on m
e
thod
wo
ul
d
w
o
r
k
o
n
l
y
on t
h
e di
f
f
e
rent
i
a
bl
e f
u
nc
t
i
ons. B
u
t
t
h
e
bi
o
-
i
n
s
p
i
r
e
d
t
echni
que
s use
d
i
n
t
h
i
s
pa
per a
r
e n
o
t
Evaluation Warning : The document was created with Spire.PDF for Python.
I
J
PED
S
I
S
SN
:
208
8-8
6
9
4
Co
n
verg
e
n
ce Pa
rameter
Ana
l
ysis fo
r
Differen
t
Metah
eu
ristic Metho
d
s
Contro
l Con
s
tan
t
…
(R. Sa
ga
ya
ra
j)
40
6
depe
n
d
ent
on t
h
e f
unct
i
on e
v
en i
f
i
t
i
s
di
ffe
rent
i
a
bl
e
or
no
t
.
The o
r
i
g
i
n
al
i
n
t
e
nt
i
on
resea
r
chi
ng
on
bi
r
d
fl
oc
k
m
ove
m
e
nt was to gra
phicall
y
sim
u
late
the graceful an
d unpredictable chore
o
graphy
of a bird fl
ock which
wh
en
an
alyzed
turn
ed ou
t t
o
b
e
an
op
ti
mizer called
Particle Swarm
Op
tim
izat
io
n
(PSO)
[10
]
. The PSO
m
e
thod sta
r
ts with the i
n
itialization
of
the population within
the
sol
u
tion s
p
ace create
d
.
Objective
functi
on
for th
e in
itial p
opu
latio
n
is created
and
th
e pb
est,
gb
est
valu
es are d
e
termin
ed
[10
]
.
With
th
is as t
h
e
in
itial
so
lu
tion
set th
e iteratio
n
will g
o
on
, wh
ere th
e n
e
w pop
u
l
a
tio
n
set is g
e
n
e
rated
using
th
e v
e
lo
city fun
c
tio
n
as
defi
ned
bel
o
w
,
id
id
id
id
gd
id
id
id
id
v
x
x
x
p
rand
c
x
p
rand
c
v
v
)
(
*
()
*
)
(
*
()
*
2
1
(1
)
Whe
r
e,
x
id
is
th
e cu
rren
t v
a
lu
e
an
d
t
h
e next
value of the i
th
po
pul
at
i
o
n,
c
1
and
c
2
are the c
o
nstants,
p
id
i
s
t
h
e nei
g
hb
or
i
ng
best
val
u
e
,
p
gd
i
s
t
h
e gl
obal
best
val
u
e
.
F
o
r
t
h
e new set
o
f
po
pul
at
i
o
n ge
nerat
e
d usi
n
g
equation
(2) t
h
e obj
ecti
v
e function is
recalcula
ted until the
optim
izat
ion
c
o
ndition is re
a
c
hed.
Gen
e
tic algorith
m
is o
n
e
o
f
th
e earliest ev
olu
tio
n
a
ry
al
g
o
r
i
t
h
m
s
(EA
)
, w
h
i
c
h use
d
t
h
e conce
p
t
o
f
n
a
tural selection
for th
e op
timizatio
n
pro
b
l
em
so
lu
tio
n
s
.
Fo
r i
n
itializatio
n
m
a
n
y
so
lu
tion
s
are tak
e
n
and
tho
s
e
so
lu
tion
s
are
called
th
e in
it
ial p
o
p
u
l
ation
.
Th
e in
itial
po
pu
latio
ns are sp
read
ou
t in th
e who
l
e ran
g
e
of
possible sol
u
tions. The sele
c
tion process succeeds the initializa
tion pr
ocess, where the fitter solution are
tak
e
n fro
m
th
e
in
itialized
v
a
lues b
y
m
ean
s
o
f
find
ing
t
h
e fittest
so
l
u
tio
n
of th
e o
b
j
ectiv
e fu
n
c
tion
or rando
m
l
y
selectin
g
fro
m
th
e in
itial p
o
p
u
l
atio
n
.
Th
e
gen
e
tic o
p
e
rators o
f
m
u
tatio
n
an
d
cro
ssov
e
r
are app
lied
o
n
th
ese
selected sol
u
tion val
u
es.
The
s
e values are
c
onsi
d
ere
d
a
s
t
h
e par
e
nt
a
n
d t
h
e chi
l
d
ren
are
fo
u
nd
by
c
o
m
b
i
n
i
n
g
t
h
ese sel
ect
ed
pare
nt
s
o
l
u
t
i
o
ns.
The
n
ne
w
pare
nt
s a
r
e s
e
lected for e
v
ery child
a
n
d the
above process
of
m
u
ta
tio
n
and
cro
s
sov
e
r con
tin
u
e
s un
til a d
e
sired
n
u
m
b
e
r
o
f
so
lu
tion
s
are o
b
t
ain
e
d
.
The so
lu
tion
s
are ag
ain
checke
d
for fi
tness on
the
ob
j
ective fun
c
tio
n. Th
e term
i
n
atio
n of t
h
e
alg
o
rith
m
o
ccu
rs if th
e num
b
e
r o
f
iteratio
n
is
reach
ed or t
h
e
o
b
j
ectiv
e
fu
n
c
tio
n
is e
ith
er
min
i
mized
or m
a
x
i
mi
zed [11]. C
u
c
k
oo
Searc
h
Al
g
o
ri
t
h
m
(C
SA) i
s
al
s
o
a
p
o
pul
at
i
o
n
base
d
m
e
t
heuri
s
t
i
c
m
e
t
hod
wi
t
h
t
w
o
su
b
o
p
erat
i
ons
,
fi
rst
o
n
e
b
e
i
ng t
h
e
d
i
rect search
based
on
th
e Lev
y
flig
h
t
s and
a rand
o
m
search
b
a
sed
o
n
t
h
e p
r
o
b
a
b
ility o
f
th
e ho
st b
i
rd
to
fi
nd
out
w
h
et
he
r i
t
i
s
an al
i
e
n egg
[1
2]
. Thi
s
i
s
ba
sed o
n
t
h
e fact
t
h
at
C
u
ck
oo
w
oul
d use t
h
e
ne
st
of di
f
f
ere
n
t
bi
r
d
s
to
d
e
v
e
l
o
p its
o
f
f
s
pr
ing
f
r
o
m
th
e
p
e
r
i
o
d
o
f
layin
g
egg
s
.
Th
e al
g
o
r
ith
m is d
e
p
e
n
d
e
n
t
o
n
how
d
o
es
cu
ckoo
strateg
i
ze to gro
w
its offspring
fro
m
th
e
hat
c
hi
n
g
st
a
g
e i
n
t
h
e h
o
st
bi
rd
’s
ne
st
.
Th
e step
s i
n
vo
l
v
ed in
t
h
e CSA m
e
th
o
d
is as
defin
e
d
in th
e
follo
wing
,
As in
ev
ery op
ti
m
i
zatio
n
al
g
o
rith
m
h
e
re th
e in
itial p
o
p
u
l
atio
n
is th
e n
u
m
b
e
r of host n
e
st, wh
ich in
o
u
r
p
r
ob
lem is th
e
p
o
p
u
l
ation
of
co
n
t
ro
l con
s
tan
t
s in
sid
e
its limits. Th
e n
e
st with
h
i
g
h
e
r quality
lev
e
l will
g
o
to
th
e n
e
x
t
g
e
n
e
ratio
n
.
Th
e
p
r
o
b
ab
ility lev
e
l o
f
th
e ho
st b
i
rd
to fin
d
wh
eth
e
r th
ere is an
alien eg
g
is m
easu
r
ed
.
If
the probability is above a des
i
red lim
i
t
then the host bird
woul
d either thro
w the alien e
gg
outsi
de the nest or
i
t
wo
ul
d
m
i
grat
e fr
om
t
h
at
n
e
st
t
o
bui
l
d
a
new
ne
st
.
Wh
en t
h
e
nest
i
s
aban
d
one
d
t
h
e
nest
g
o
es
o
u
t
o
f
t
h
e
solution s
p
ace. In
orde
r to re
place the ne
w
nest instead
of t
h
e rem
ove
d
one, as the
num
b
er of the
nest
m
u
st be
con
s
t
a
nt
, t
h
e L
e
vy
Fl
i
ght
’s al
go
ri
t
h
m
i
s
used t
o
m
ove t
o
a new sol
u
t
i
o
n poi
nt
, w
h
i
c
h w
oul
d bec
o
m
e
the ne
w
nest
ad
de
d i
n
t
h
e
next
ge
nera
t
i
on [
1
2]
.
Gra
v
i
t
y
Search al
go
ri
t
h
m
i
s
devel
o
ped
o
n
t
h
e
bas
i
s of
l
a
w
of
g
r
avi
t
y
an
d m
a
ss in
teractio
n
s
. Th
e
in
teractio
n b
e
t
w
een th
e ag
e
n
t
s
,
w
h
i
c
h a
r
e
o
b
ject
s
ha
vi
n
g
t
h
ei
r
per
f
o
r
m
a
nce
measured
by their m
a
sses, are carried
out
u
s
i
ng t
h
e
fo
rce
of
gra
v
i
t
y
bet
w
een t
h
em
. The fo
ur
param
e
ters t
h
at
d
e
fi
n
e
th
e
GSA are
p
o
sitio
n, in
ertial m
a
ss,
activ
e
g
r
av
itatio
n
a
l m
a
ss, an
d p
a
ssiv
e
grav
itatio
n
a
l m
a
ss. Th
e
p
o
s
ition
of th
e
m
a
ss wo
u
l
d
determin
e th
e so
lu
tion
o
f
th
e
o
b
j
ectiv
e fun
c
tio
n
,
wh
ere as th
e g
r
av
itatio
n
a
l an
d
th
e in
ertial m
a
sses are
d
e
termin
ed
u
s
i
n
g
a
fitn
ess fun
c
tion
.
Th
e m
o
v
e
men
t
o
f
t
h
e m
a
s
s
es, wh
ich
is t
h
e n
e
w
so
lu
tion
p
o
i
n
t
, is co
n
t
ro
lled
b
y
th
e u
s
e of th
e grav
itati
o
n
a
l an
d
th
e i
n
erti
al
m
a
sses. Th
e h
eav
iest m
a
ss is th
e
sol
u
t
i
o
n i
n
t
h
e
searc
h
space
[1
3]
.
Harm
ony
searc
h
al
go
ri
t
h
m
i
s
anot
her
o
p
t
i
m
i
zat
i
on
al
go
ri
t
h
m
,
whi
c
h i
s
deri
ved f
r
om
the co
ncept
o
f
f
i
ndi
n
g
t
h
e best
harm
ony
created from
the
musicians.
T
h
e best harm
ony created
is th
e b
e
st so
lutio
n
,
wh
ile each
m
u
sician
is th
e d
ecisi
o
n
va
riable, the
play they create is the ge
ne
rated
value; a
not
e i
n
t
h
e
pl
ay
i
s
t
h
e
val
u
e
fo
r
fi
n
d
i
n
g t
h
e
best
harm
ony
[
14]
. B
A
T
al
go
ri
t
h
m
i
s
a bi
o-i
n
spi
r
ed
al
g
o
r
i
t
h
m
,
whi
c
h de
ri
ves
t
h
e echol
ocat
i
on
beha
vi
o
r
o
f
t
h
e
m
i
crobat
s
for
vary
i
n
g p
u
l
s
e rat
e
s of l
o
ud
ness a
nd em
i
ssi
on
.
By u
s
in
g
th
ese en
tire d
i
scu
s
sed
algo
rith
m
s
t
h
e op
timiza
tio
n of the T
HD i
n
the boost conve
rter is carried out
with
th
e estim
a
tio
n
o
f
th
e con
t
ro
l con
s
tan
t
s i
n
th
e PI con
t
ro
ll
er
u
s
ed in
t
h
e co
nv
erter.
4.
BOOST CONVERTER
DE
SIGN
B
oost
c
o
nve
rt
er ba
sed
PFC
h
a
s bee
n
a t
r
e
n
d
,
as i
t
has
t
h
e i
nhe
re
nt
desi
gn
, t
h
at
w
o
ul
d el
im
i
n
at
e t
h
e
seco
nd
o
r
der
h
a
rm
oni
cs i
n
t
h
e su
p
p
l
y
si
de.
The
re
duct
i
o
n
of
ha
rm
oni
cs a
n
d
t
h
e
v
o
l
t
a
ge
ri
p
p
l
e
i
s
t
a
ke
n
care
by
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
:
404
–
4
14
4
07
th
e ACM
m
e
t
h
od
. Th
is circu
it is ob
tain
ed b
y
co
m
b
in
in
g
th
e un
con
t
ro
lled
rectifier with
the b
o
o
s
t
con
v
e
rter
to
po
log
y
wh
ich
is th
en
connected
to
th
e Vo
ltag
e
So
ur
ce in
v
e
rter (VSI)
with
th
e th
ree p
h
a
se in
du
ction
m
o
to
r
as gi
ven
i
n
t
h
e
Fi
gu
re
1.
Th
e In
du
ction
m
o
to
r is
mad
e
to
work
with
ou
t an
y co
n
t
ro
l tech
n
i
qu
e, thu
s
ru
nn
ing
at its
rated
sp
eed.
Th
e sp
ecif
i
catio
n
of
th
e indu
ctio
n
m
o
to
r
consid
er
ed
fo
r
t
h
e
r
e
sear
ch
stud
y
is 5
.
4
H
P
, 400V
,
1
430
rp
m
,
5
0
H
z
,
4
po
les on
e [1
]. As th
e m
o
to
r
is a 4
0
0
V three p
h
a
se i
n
du
ctio
n
m
o
to
r, i
n
ord
e
r to
lim
i
t
th
e startin
g
current, we
sho
u
l
d
ha
ve
us
ed t
h
e st
a
r
t
e
r i
n
o
r
der t
o
get
r
i
d o
f
t
h
e st
a
r
t
i
ng c
u
rre
nt
dy
n
a
m
i
cs, but
t
h
e
i
n
d
u
ct
o
r
i
n
t
h
e
bo
ost
conve
r
ter would serve the purpose of
t
h
e
sm
oot
h st
art
i
ng of t
h
e i
n
du
ct
i
on m
o
t
o
r, h
e
nce st
art
e
r can b
e
avoi
ded
.
T
h
e s
c
hem
a
t
i
c
of t
h
e co
nve
rt
er
wi
t
h
t
h
e
i
n
duct
i
on
m
o
t
o
r i
s
as
gi
ven
i
n
Fi
g
u
re
1
.
Fi
gu
re 1.
1
-
B
o
o
s
t
Rectifier with
h
e
3
-
Electric Dri
v
e
Syste
m
Th
e
b
o
o
s
t co
nv
erter is d
e
si
gn
ed fo
r t
h
e
fo
ll
o
w
i
n
g d
e
si
g
n
criteria.
Wh
en
th
e transisto
r
switch
e
s
ON,
th
e equ
a
tion
o
f
th
e cu
rren
t
i
L
(
t
)
i
s
gi
ve
n
by
t
h
e f
o
l
l
o
wi
ng
eq
uat
i
o
n
(
3
)
as
L
s
v
L
L
V
dt
L
di
|
|
(2
)
(a)
(b
)
(c)
Fi
gu
re
2.
Si
n
g
l
e
-P
hase B
oost
R
ect
i
f
i
e
r f
o
r t
h
e El
ect
ri
c D
r
i
v
e Sy
st
em
:
(A)
Po
wer C
i
rcui
t
and
E
qui
val
e
nt
Circuit f
o
r
Tra
n
sisto
r
T i
n
(B)
O
n
-State a
n
d (
C
) O
f
f
-
State
Evaluation Warning : The document was created with Spire.PDF for Python.
I
J
PED
S
I
S
SN
:
208
8-8
6
9
4
Co
n
verg
e
n
ce Pa
rameter
Ana
l
ysis fo
r
Differen
t
Metah
eu
ristic Metho
d
s
Contro
l Con
s
tan
t
…
(R. Sa
ga
ya
ra
j)
40
8
Due t
o
the
fa
ct that |
v
s
| >
0
,
th
e ON
s
t
ate
o
f
tr
an
s
i
s
t
or
T
always
produces a
n
inc
r
ease in t
h
e
inductance
current
i
L
. The
des
i
gn pa
ram
e
t
e
rs
f
o
r
t
h
e desi
g
n
of
t
h
e b
oost
re
ct
i
f
i
e
r
i
s
gi
ven
i
n
eq
uat
i
o
n (3
) as
)
(
5
.
8
2
0
2
0
0
0
min
min
2
approx
mF
L
hld
V
V
t
P
C
(3
)
whe
r
e,
C
0
=
Out
put ca
pacit
a
nce,
P
0
=
Out
put
p
o
we
r of
t
h
e
c
o
nve
rt
e
r
4
K
W
,
t
hld
=
Hol
d
up
t
i
m
e
, no
rm
al
ly
20 m
s
,
V
0min
=
Min
i
m
u
m
v
a
lue
of
th
e
ou
tpu
t
regu
lated
vo
ltag
e
(400
V DC),
V
0Lmin
=
Ran
g
e
of
inpu
t vo
ltag
e
(
230
V A
C
).
The
val
u
e
of t
h
e bo
ost
i
n
d
u
ct
o
r
af
fect
s m
a
ny
ot
he
r de
si
g
n
pa
ram
e
t
e
rs. M
o
st
of t
h
e c
u
r
r
ent
t
h
at
fl
o
w
s
th
ro
ugh
th
is ind
u
c
t
o
r is at low
freq
u
e
n
c
y. Th
is is
p
a
rtic
u
l
arly tru
e
at th
e l
o
west
inp
u
t
voltag
e
wh
ere the in
pu
t
current is the highest. No
rm
ally, the acce
ptable level of ripple cu
rre
nt is between 10 and 20 %.
For a
switching fre
quency of 100 kH
z, t
h
e
followi
ng form
ula will produce
acce
ptable
results.
L
a
=
3
000
/ Po
m
H
L
a
=
30
0 /
2
5
0
=
1m
H (
a
p
p
r
o
x.
)
Th
e cap
acito
r
th
at is d
e
sig
n
e
d
fro
m
th
e b
o
o
s
t con
v
e
rter
co
nfigu
r
ation
w
ill eli
m
in
ate
th
e second
harm
oni
c i
n
t
h
e fi
rst
ha
nd
. T
h
e F
o
uri
e
r
ana
l
y
s
i
s
t
e
l
l
s
t
h
at t
h
e am
ount
o
f
t
h
e
seco
n
d
h
a
rm
oni
cs p
r
es
ent
i
s
ab
ou
t 0.02
% wh
ereas th
e
th
ird
h
a
rm
o
n
i
c is ab
ou
t 63.
9
3
%. C
onsi
d
era
b
l
e
at
t
e
nt
i
on i
s
gi
ve
n t
o
wa
r
d
s
sup
p
r
essi
n
g
t
h
e t
h
i
r
d a
n
d s
u
c
cessi
ve
o
d
d
ha
rm
oni
cs i
n
o
u
r
pr
o
pose
d
sy
st
em
, whi
c
h i
s
o
n
e
o
f
t
h
e
c
ont
ri
b
u
t
i
o
n
s
of
the researc
h
work
[1].
5.
PROP
OSE
D
WORK
As t
h
e e
x
t
e
nsi
o
n
of
o
u
r
p
r
e
v
i
ous
wo
rk
as
i
n
[
1
]
t
h
is
p
a
p
e
r
is m
ean
t to
d
e
v
e
lop
a trad
eoff estim
at
io
n
of t
h
e di
ffe
ren
t
opt
i
m
i
z
at
i
on al
go
ri
t
h
m
s
defi
ned i
n
t
h
e
abo
v
e
sectio
n. Th
e algo
rith
m
is u
s
ed
to
esti
mate th
e
co
n
t
ro
l co
nstan
t
s in
t
h
e PI co
n
t
ro
ller
u
s
ed
i
n
th
e
bo
os
t co
nve
rter.
The
p
a
ram
e
ter for c
o
m
p
arison
f
o
r
all these
algorithm
s
are as m
e
ntioned a
b
ove a
n
d the
s
e
res
u
lts
are tabu
lated
an
d d
i
scu
ssed
in th
e
n
e
x
t
section
.
Th
e fitn
ess functio
n
fo
r th
e op
ti
m
i
zatio
n
tec
h
n
i
q
u
e
is th
e To
tal Harm
o
n
i
c Disto
r
tion
calcu
lated
fro
m
th
e Matlab
TM
/Si
m
u
lin
k
m
o
d
e
l wh
ich
will b
e
calcu
l
ated
b
y
the u
s
e
of th
e m
a
th
em
a
tical fo
rm
u
l
a as g
i
v
e
n in
the
fo
rm
ulae
THD
F
V
2
2
V
3
2
...
.
V
n
2
V
1
2
(4
)
The
population is created
fo
r tw
o con
t
ro
l co
nstan
t
s
K
p
and
K
i
an
d th
e
o
p
tim
izat
i
o
n is
carried out for the
m
i
nim
i
zat
i
on o
f
t
h
e
TH
D a
s
d
e
fi
ne
d i
n
eq
uat
i
on
(
4
).
6.
R
E
SU
LTS AN
D ANA
LY
SIS
The pa
ram
e
ter
s
that are calculated for the
pe
rform
a
nce
measure are po
p
u
l
a
t
i
o
n
s
i
z
e
,
ma
x
i
mu
m
num
ber
of
ep
o
c
hs,
an
d
gl
o
b
al
best
s
o
l
u
t
i
o
n
o
f
t
h
e
co
nt
r
o
l
c
o
nst
a
nt
s,
be
st
T
H
D
val
u
e a
n
d
execut
i
o
n t
i
m
e.
The Si
m
u
l
i
n
k
m
odel
fo
r t
h
e abo
v
e
bo
ost
con
v
ert
e
r wi
t
h
t
h
e re
si
st
an
ce em
ul
at
i
on m
e
t
hod
was
devel
ope
d
wi
t
h
t
h
e
P
I
c
o
nt
ro
l
l
e
r an
d t
h
e c
o
nt
r
o
l
co
nst
a
nt
s o
f
t
h
i
s
co
nt
r
o
l
l
er are
t
h
e
val
u
es t
h
at
a
r
e
opt
i
m
i
zed
by
t
h
e
use
o
f
di
ffe
re
nt
o
p
t
i
m
i
zat
i
on t
ech
ni
q
u
es
di
scu
ssed
abo
v
e.
T
h
e
ob
j
ect
i
v
e f
unct
i
o
n
f
o
r m
i
nim
i
zati
on i
s
t
h
e T
H
D
cal
cu
l
a
t
i
on. T
h
e
res
u
l
t
s
o
b
t
a
i
n
e
d
fr
om
di
ffere
nt
o
p
t
i
m
i
zat
i
on t
echni
que
as
gi
ve
n
bel
o
w
7.
ALGO
RITH
M P
A
RA
ME
TERS
7
.
1
.
PARTICLE
SWARM OPTIMISATION
Obj
ectiv
e
fu
n
c
tio
n
:
Mean (THD) in percen
t.
Num
b
er of
va
r
i
abl
e
s
= 4
(
K
p
1
, Ki
1
,
Kp
2,
K
i
2).
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
:
404
–
4
14
4
09
Ine
r
t
i
a
wei
g
ht
=
0.
9 – 0.
4.
Acceleration c
onsta
nt1 =
2.
Acceleration c
onsta
nt2 =
2.
R
a
nge
o
f
vari
a
b
l
e
s = LB
[
0
0
0
0]
,
U
P
[
0
.1 7 2 3
]
G
l
ob
al b
e
st so
l
u
tio
n
:
kp
1
=
0
.
00
242
1,
k
i
1
=
1
.
78
549
6,
kp
2=
2.
00
0
0
0
0
,
ki
2=
1.
5
6
3
3
9
9
7.
2. GENETI
C ALGO
RIT
H
M
Obj
ectiv
e
fu
n
c
tio
n
:
Mean (THD) in percen
t.
Num
b
er of
va
r
i
abl
e
s
= 4
(
K
p
1
, Ki
1
,
Kp
2,
K
i
2).
Num
b
er
of
m
u
t
a
t
i
on chi
l
dre
n
(Ga
u
ssi
a
n
) =
4
.
Num
b
er
of
m
u
t
a
t
i
on chi
l
dre
n
(ra
nd
om
) = 4
.
Nu
m
b
er
o
f
elitis
m
ch
ild
ren
=
2
.
R
a
nge
o
f
vari
a
b
l
e
s = LB
[
0
0
0
0]
, U
P
[0
.1
7
2
3]
G
l
ob
al b
e
st so
l
u
tio
n
:
kp
1
=
0
.
01
010
5,
k
i
1
=
1
.
96
773
2,
kp
2=
1.
81
5
0
1
1
,
ki
2=
0.
3
3
3
2
6
1
7
.
3
.
CU
CKOO
SEARC
H A
L
GOR
I
THM
Obj
ectiv
e
fu
n
c
tio
n
:
Mean (THD) in percen
t.
Po
pul
at
i
o
n si
ze
=2
0.
Num
b
er of
va
r
i
abl
e
s
= 4
(
K
p
1
, Ki
1
,
Kp
2,
K
i
2).
Prob
ab
ility o
f
ab
ando
n (Pa) = 0
.
25
.
R
a
nge
o
f
vari
a
b
l
e
s = LB
[
0
0
0
0]
, U
P
[0
.1
7
2
3]
M
a
xi
m
u
m
epochs =
1
0
0
.
G
l
ob
al
b
e
st
so
l
u
tio
n
:
kp
1
=
0
.
00
000
0, k
i
1
=
1
.
76
659
4,
kp
2=
2.
00
0
0
0
0
,
ki
2=
0.
0
0
0
0
0
0
G
l
ob
al b
e
st TH
D =
1
.
8
670
32
Ex
ecu
tio
n ti
m
e
=10
254
.0
2 sec.
7.
4.
GR
A
V
IT
Y SE
AR
CH
A
L
GORIT
HM
:
Obj
ectiv
e
fu
n
c
tio
n
:
Mean (THD) in percen
t.
Po
pul
at
i
o
n si
ze
=2
0.
Num
b
er of
va
r
i
abl
e
s
= 4
(
K
p
1
, Ki
1
,
Kp
2,
K
i
2).
In
itial Grav
itatio
n
a
l co
nstan
t
(G0
)
=
1
0
.
Acceleration c
onsta
nt (al
p
ha) = 10.
Ep
silon
=
0
.
000
1.
Eu
clid
ean
length
of
R
(
R
nor
m)
=
1
.
Po
wer
o
f
R
(R
po
we
r) =
1
.
P
e
r
c
en
t of
ag
en
ts
ap
p
l
y for
c
e (
f
i
n
d_p
e
r
)
=
2
.
R
a
nge
o
f
vari
a
b
l
e
s = LB
[
0
0
0
0]
, U
P
[0
.1
7
2
3]
M
a
xi
m
u
m
epochs =
5
0
.
G
l
ob
al b
e
st so
l
u
tio
n
:
kp
1
=
0
.
00
009
0,
k
i
1
=
1
.
93
939
8,
kp
2=
2.
00
0
0
0
0
,
ki
2=
1.
1
1
2
1
6
9
G
l
ob
al b
e
st TH
D =
1
.
8
942
84
Ex
ecu
tio
n ti
m
e
= 952
3.09
secs.
7
.
5
.
HARM
ON
Y
SEARC
H A
L
GOR
I
THM
Obj
ectiv
e
fu
n
c
tio
n
:
Mean (THD) in percen
t.
Po
pul
at
i
o
n si
ze
=2
0.
Num
b
er of
va
r
i
abl
e
s
= 4
(
K
p
1
, Ki
1
,
Kp
2,
K
i
2).
Pi
t
c
h
B
a
n
d
wi
dt
h (
b
w
)
= 0.
9.
Harm
ony
M
e
m
o
ry
c
o
nsi
d
e
r
i
n
g R
a
t
e
(
H
M
C
R
)
=
0.
9
5
Pitch
Adj
u
stmen
t
Rate (PAR
) =
1
R
a
nge
o
f
vari
a
b
l
e
s = LB
[
0
0
0
0]
, U
P
[0
.1
7
2
3]
M
a
xi
m
u
m
epochs =
5
0
.
Gl
o
b
al
be
st
sol
u
t
i
on:
k
p1=
0.
0
0
0
0
0
0
,
ki
1=
2.
4
0
1
0
9
4
,
kp
2=
2.
00
0
0
0
0
,
ki
2=
0.
26
0
7
3
6
G
l
ob
al b
e
st TH
D =
1
.
9
631
94
Evaluation Warning : The document was created with Spire.PDF for Python.
I
J
PED
S
I
S
SN
:
208
8-8
6
9
4
Co
n
verg
e
n
ce Pa
rameter
Ana
l
ysis fo
r
Differen
t
Metah
eu
ristic Metho
d
s
Contro
l Con
s
tan
t
…
(R. Sa
ga
ya
ra
j)
41
0
Ex
ecu
tio
n ti
m
e
= 408
.8
8secs
7.
6. B
A
T
ALGORITHM
Obj
ectiv
e
fu
n
c
tio
n
:
Mean (THD) in percen
t.
Po
pul
at
i
o
n si
ze
=2
0.
Num
b
er of
va
r
i
abl
e
s
= 4
(
K
p
1
, Ki
1
,
Kp
2,
K
i
2).
Pi
t
c
h
B
a
n
d
wi
dt
h (
b
w
)
= 0.
9.
Lo
ud
ness
(
A
)
= 0.
9
Rate o
f
pu
lse emissio
n
(
r)
= 0.1
M
i
n
i
mu
m f
r
e
q
u
e
n
c
y
(
Q
mi
n
)
=
0
M
a
x
i
mu
m f
r
e
q
u
e
n
c
y
(
Q
ma
x
)
=
2
R
a
nge
o
f
vari
a
b
l
e
s = LB
[
0
0
0
0]
, U
P
[0
.1
7
2
3]
M
a
xi
m
u
m
epochs =
5
0
.
Gl
o
b
al
be
st
sol
u
t
i
on:
K
p1=
0
.
00
6
1
5
2
9
, ki
1=
6.
90
7
3
k
p
2
=
2
k
i
2
=
2
.
9
943
G
l
ob
al b
e
st TH
D =
2
.
2
826
Ex
ecu
tio
n ti
m
e
=62
5
3
.
17
secs.
7.
7. CO
N
V
ER
GENCE
G
R
A
P
H
Fi
gu
re
3.
Part
i
c
l
e
Swa
r
m
Opt
i
m
i
zati
on T
H
D
vs
N
o
.
o
f
It
erat
i
ons
Fig
u
re
4
.
Gen
e
tic Alg
o
rith
m
THD vs
No.of
Iteratio
ns
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
:
404
–
4
14
4
11
Fi
gu
re
5.
C
u
c
k
oo
Sea
r
ch
Al
g
o
ri
t
h
m
Figure
6. Gravi
t
y Search Al
gorithm
Fi
gu
re 7.
Ha
rm
ony
Searc
h
Al
go
ri
t
h
m
Evaluation Warning : The document was created with Spire.PDF for Python.
I
J
PED
S
I
S
SN
:
208
8-8
6
9
4
Co
n
verg
e
n
ce Pa
rameter
Ana
l
ysis fo
r
Differen
t
Metah
eu
ristic Metho
d
s
Contro
l Con
s
tan
t
…
(R. Sa
ga
ya
ra
j)
41
2
Fi
gu
re 8.
B
A
T
al
go
ri
t
h
m
7
.
8
.
TR
AD
EOFF ANA
LY
SIS COMPAR
ISION TA
BLE
Tabl
e
1. C
o
m
p
ari
s
o
n
of
TH
D
val
u
es
f
r
om
Di
ffe
rent
C
ont
rol
Tech
ni
q
u
es
Contr
o
l T
echnique
Global Best T
HD
W
ithout ACM
63.
93 %
W
ith AC
M
4.
93 %
FL
C
2.
9
%
ANFI
S
2.
8
%
PSO 1.
9189
16%
GA 2.
0221
01%
CSA 1.
8670
32%
GSA 1.
8942
84%
HSA 1.
9631
94%
BAT 2.
2826%
Tab
l
e
2
.
C
o
m
p
arison
o
f
Parameters fro
m
v
a
riou
s
Op
tim
iza
tio
n
Algo
rith
m
s
Nam
e
Population
Size
M
a
xim
u
m
E
pochs
E
x
ecution T
i
m
e
in sec
PSO 10
100
2531
5.
08
GA 20
50
2078
7.
73
CSA 20
100
1025
4.
02
GSA 20
50
9523.
0
9
HSA 20
50
408.
88
BAT 20
50
6253.
1
7
The t
r
a
d
e
o
f
f
i
n
fere
nce i
s
de
pe
nde
nt
o
n
w
h
et
her t
h
e algorit
h
m
can be im
p
l
e
m
ented on a
process
o
r
or
th
e param
e
ters lik
e th
e pop
u
l
atio
n
size and
ex
ecu
tion ti
m
e
wh
ich is
d
e
p
e
n
d
e
n
t
on
th
e me
m
o
ry and
t
h
e
sp
eed
of the
process
o
r is taken care. Also
the accuracy of THD m
i
nimization
m
u
st
be taken as it is the ultimate aim
of
t
h
e e
x
peri
m
e
nt
.
CSA ex
h
i
b
its th
e op
tim
a
l
THD
v
a
lue and less ex
ecu
tion
tim
e co
m
p
ared
wit
h
o
t
h
e
r
o
p
tim
izat
io
n
alg
o
rith
m
s
. Ho
wev
e
r PSO
also
p
r
ov
id
es
b
e
st THD
v
a
lu
e
b
u
t
it tak
e
s m
o
re ex
ecu
t
io
n ti
m
e
an
d
lesser
p
opu
latio
n
size.
GA tak
e
s lesser ex
ecu
tion
ti
me co
m
p
ares with
PS
O,
but
its favorable
THD val
u
e is
greater tha
n
PSO algo
rith
m
.
GSA g
i
v
e
s
b
e
tter THD
v
a
lu
e co
m
p
ared wit
h
PSO and
GA with
lesser execu
tio
n tim
e.
BAT algo
rithm tak
e
s lesser ex
ecu
tion
tim
e
co
m
p
ared
with
PSO,
GA and
CSA, bu
t its o
p
tim
al THD
v
a
lu
e is poo
rer th
an o
t
h
e
r algo
rith
m
s
.
HSA algo
rithm p
r
ov
id
es
opti
m
u
m
THD valu
e with
v
e
ry
less ex
ecu
tion ti
m
e
co
m
p
ared
with
o
t
h
e
r
alg
o
rith
m
s
, b
u
t
CSA al
go
rithm
o
v
e
rru
l
es all th
e
o
t
h
e
r algorith
m
s
to
ob
tain th
e
b
e
st THD
v
a
lu
e.
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
:
404
–
4
14
4
13
8.
CO
NCL
USI
O
N
Th
e co
nv
erg
e
nce g
r
aph
shows th
at th
e lowest ti
m
e
tak
i
n
g
algo
rith
m
for settlin
g
is the Harm
o
n
i
c
Searc
h
Algorit
h
m
(HSA). CSA
gives
the l
o
west T
H
D cal
cu
lated
am
o
n
g
all th
e algo
rithm
s
u
s
ed
.
GSA
h
a
s t
h
e
secon
d
lowest
THD op
tim
iz
ed
. Th
e PSO
u
s
ed
v
e
ry lesser n
u
m
b
e
r of iteratio
n
co
m
p
ared
to
all th
e o
t
h
e
r
alg
o
rith
m
s
. Fro
m
th
e tab
l
e 1
an
d
2
,
it is o
b
v
i
ou
s th
at th
e PSO is th
e
m
o
st
m
e
m
o
ry effi
cien
t an
d
th
e HSA is
th
e
m
o
st ti
me
efficien
t alg
o
rith
m
to
b
e
im
p
l
e
m
en
ted
o
n
th
e con
t
ro
l co
nstan
t
esti
matio
n
fo
r resistan
ce
em
ul
at
i
on i
n
a
bo
ost
c
o
nve
rt
e
r
. C
S
A
has
p
r
o
v
ed
i
t
s
el
f t
o
be
a m
o
re acc
ura
t
e
m
e
t
hod
i
n
e
s
t
i
m
a
ti
ng t
h
e
c
ont
rol
constants.
Hen
ce tt can
be in
ferred
th
at
, th
e efficien
t alg
o
rith
m
to
u
s
e in
th
is o
p
timizatio
n
can
b
e
th
e HSA,
wh
ich
resu
lted
in
lowest THD v
a
lu
e.
ACKNOWLE
DGE
M
ENTS
The w
o
r
k
desc
ri
be
d i
n
t
h
i
s
paper ha
s bee
n
sup
p
o
rt
e
d
by
t
h
e R
e
searc
h
C
e
nt
re o
f
t
h
e D
e
part
m
e
nt
of
Electr
i
cal an
d Electr
o
n
i
cs
En
g
i
n
eer
i
n
g
o
f
K
.
S. Rang
asa
m
y Co
lleg
e
o
f
Techn
o
l
o
g
y
, Tir
u
ch
en
gode. Th
e
au
tho
r
s wou
l
d
lik
e to
ex
press
th
eir
g
r
atitud
e
for th
e su
ppo
rt
o
f
th
is st
u
d
y
.
REFERE
NC
ES
[1]
R. Sagay
a
r
a
j an
d Dr. S. Thangavel, “I
m
p
lem
e
nt
ation of Int
e
ll
ige
n
t Control
Strategies on Current
Ripple Redu
ctio
n
and Harmonic Analy
s
is at th
e Converter Side of
the Industrial Inverters and Tradeoff Analy
s
is”,
Journal of
Theoretical and
Applied
Informa
tion Technology
, Vol. 65, No. 2
,
July
2014
pp. 34
4 ~ 351.
[2]
ZM
S
.
EI-Barb
a
r
y
,
H.Z
.
Aza
z
i,
M
K
. M
e
twall
y
,
“
T
otal
Harmon
ic DistortionAnaly
s
i
s
of a Four
Switch 3-phase
Inverter fed Sp
eed Sensorless Control of IM
Drives”,
Interna
tional Journal
of Po
wer Electronics and Drives
System
, Vol. 4,
No. 1, March
20
14 pp. 81 ~ 90
.
[3]
Yen Kheng Tan, Student Member, IEEE
, and
S
a
njib Kum
a
r P
a
n
d
a, “
O
ptim
iz
ed
W
i
nd Energ
y
H
a
rves
ting S
y
s
t
e
m
Using Resistance Emulator And Active R
ectif
ier For Wireless Sensor
Nodes”,
IEEE Transa
c
tions On Power
Electronics
, Vol. 26, No. 1
,
Janu
ar
y
2011.
[4]
Predrag Pejovic
́
, “
A
Nove
l
Low-
Harm
onic Thr
e
e
-
P
h
as
e Re
ctif
ier
”
,
IEEE Transactions on
Ci
rc
uits
and Sy
ste
m
s—I:
Fundamental Theory and
Applications
, Vol. 49
,
No. 7, July
2002
.
[5]
P. Srinivasa rao
,
G. Sar
a
van
a
Ilango
and C
.
Nag
a
mani, “Line Cu
rrent Shap
i
ng us
ing Shunt Activ
e Filter with ou
t
Sensing Input V
o
ltag
e
and
Lo
ad
Current”,
TENCO
N 200
8
I
E
EE Reg
i
o
n
10
C
o
n
f
eren
ce
, 2008
.
[6]
Naveen K. Vastrakar, Pr
abin K.
Padh
y
,
“Simplified PSO PI-PD Contro
ller
for
Unstable Proces
ses”, 2013 4
th
International
Co
nference on
Intel
ligent
Syst
ems,
Modelling
and S
i
mulation
.
[7]
G. Seshagiri R
a
o, S. Raghu
, N. Ra
jasek
a
ran
,
“Design of Feedback C
ontro
ller for Boost
Converter using
Optimizatio
n Technique”,
Intern
ational Journal
of Powe
r Electronics and Drives System
,
Vol.
3,
No.
1,
March
2013, pp
. 117~1
28.
[8]
Oscar García,
Member, IEEE,
José A. Cobos,
Member,
IEEE, Roberto
Prieto, Member,
IEEE, Pedro
Alou, an
d
J
a
vier Uc
eda
,
S
e
nior M
e
m
b
er
, I
EEE
“
S
ingle P
h
as
e P
o
wer F
a
cto
r
Correc
tion:
A
S
u
rve
y
”
,
I
E
EE Transactions
On
Power
E
l
ec
tr
oni
cs
, Vol. 18
, No.
3, May
2003.
[9]
R. Carbone A
.
Scappatura, “A Resistan
ce Emulation Technique to
Improve Efficiency
of
a PWM Adjustable Speed
Drive with P
a
s
s
ive P
o
wer F
a
c
t
o
r
Correc
tion”
, P
r
oceed
ings
of th
e
5
th
WS
EAS In
t. Conf. on
Pow
e
r Systems and
Electromagnetic Compatibility
,
Corfu, Greece,
August 23-25, 2
005 pp570 ~ 576
.
[10]
Russel C.
Eberh
a
rt, Yuhui Shi, “
P
article Swarm
Optimis
ation: Developments, Ap
p
lications and
R
e
sources”.
[11]
Srinivas. M an
d Patnaik
.
L,
“
A
daptive prob
a
b
ilit
ies of crossover and m
u
tati
on in genet
i
c a
l
gorithm
s
”,
IEEE
Transactions on
System,
Man an
d Cybernetics
, v
o
l. 24
, no
. 4
,
pp
.
656 ~ 667, 1994.
[12]
Xin-She Yang,
Suash Deb, “Cu
c
koo Sear
ch via Le
́
vy
Fl
i
g
ht
s”,
Internationa
l
Journal of Mod
e
rn Education
an
d
Computer Scien
c
e
, Vol. 5, No. 1
2
, December 20
13.
[13]
Esm
a
t Rashedi
,
Hossein Nezam
a
b
adi-pour
and S
aeid
Sar
y
a
z
di, G
S
A: A Gravit
ati
onal Se
arch
Alg
o
rithm
,
El
se
vi
e
r
Journal on, In
fo
rmation Science,
pp. 2232-2248
Volume 179 Issue 13, June, 200
9.
[14]
X.S
.
Yang, “
H
a
r
m
ony S
ear
ch a
s
a M
e
tah
e
uris
ti
c Algorithm
”
, i
n
: M
u
s
i
c-Ins
p
ir
ed Harm
on
y
S
e
arch Algor
ithm
:
Theor
y
and App
lic
ations (
E
ditor
Z.
W
.
Ge
em
),
Studies in
Com
putation
a
l
Inte
lli
gence
,
Spring
e
r Berlin
, vol. 191
,
pp. 1~14 (2009)
.
Evaluation Warning : The document was created with Spire.PDF for Python.