TELKOM
NIKA Indonesia
n
Journal of
Electrical En
gineering
Vol. 16, No. 3, Dece
mbe
r
2
015, pp. 509
~ 519
DOI: 10.115
9
1
/telkomni
ka.
v
16i3.894
2
509
Re
cei
v
ed
Jul
y
22, 201
5; Revi
sed O
c
tob
e
r 17, 201
5; Acce
pted No
vem
ber 1
3
, 2015
Biogeography Based Optimization Tuned Fuzzy Logic
Controller to Adjust Speed of Electric Vehicle
Salam Waley
*
1
, Chen
gxi
ong M
a
o
2
, Nass
ee
r
K. Ba
cha
c
h
e
3
1,2
State Key
L
a
borator
y of Adv
ance
d
Electr
o
m
agn
etic Engi
neer
ing a
nd T
e
chno
log
y
,
W
uha
n, C
h
i
n
a
3
School of Elec
trical & Electro
n
ic Eng
i
ne
eri
n
g,
Huazh
o
n
g
U
n
iversit
y
of Sci
ence a
nd T
e
chnol
og
y,
W
uhan, Ch
ina
3
Colle
ge U
n
ive
s
it
y
of Hum
anit
y
Stud
ies, Naj
a
f, Iraq
*Corres
p
o
ndi
n
g
author, e-ma
i
l
: salam_
w
a
l
e
y73@
ya
ho
o.co
m
1
, cxmao@h
u
st.edu.cn
2
,
tech_n
20
08@
yaho
o.com
3
A
b
st
r
a
ct
T
here ar
e
ma
ny pow
er
el
ec
tronic co
nverte
rs and
motor
drives c
o
n
nect
ed to
geth
e
r to
form t
h
e
electric
al system
of an Electr
ic Vehic
l
e. In this
paper, we have pr
esented a mo
deling tool that has the
adva
n
tag
e
s of
utili
z
i
ng
cap
a
b
iliti
e
s of th
e
PMSM softw
are in
deta
ile
d s
i
mulati
ons
of c
onverters,
mot
o
r
drives, and
electric machines.
In ad
dition, equiv
a
lent electr
ical
models of
Electric V
ehic
l
e dr
ive system.
T
h
is p
aper
als
o
g
i
ves
a br
ief
id
ea
of PMS
M
vali
dity
as a
n
El
ectric Ve
hi
cle si
mul
a
tion
tool. PMSM
dr
ive
system
is desc
ribed and analy
z
e
d due to
its im
portanc
e in
m
a
ny
applic
ations especially
in Electric Vehicle
app
licati
ons. T
h
is ap
plic
atio
n is hig
h
efficien
cy, low
iner
tia and h
i
g
h
torqu
e
to volu
me rat
i
o. In this pap
e
r
w
e
embo
dy the si
mu
lati
on of F
u
zz
y
Lo
gic Co
n
t
roller. T
he co
ntroll
er gover
n
the s
peed co
ntrol of Electri
c
al
Vehic
l
e EV
usi
ng p
e
r
m
a
nent
ma
gn
et synchr
ono
us
moto
r
PMSM. T
h
is w
o
rk char
acteri
zes to o
b
tain
th
e
opti
m
a
l
para
m
eters of F
L
C. Biog
eogr
ap
hy Based Opti
mi
zation (BBO) is
a new
intell
ig
ent techni
qu
e fo
r
opti
m
i
z
at
ion; it
can be
use
d
to tune
th
e par
ameters in
different fie
l
ds. T
h
e main c
ontri
b
u
tion
of this w
o
rk
efforts the ab
i
lity of BBO to
desi
gn th
e p
a
ra
mete
rs of F
L
C
by deter
mi
nin
g
th
e
sh
apes of
trian
g
l
e
me
mbers
h
ips
of the inp
u
ts a
nd out
put. T
he results
of op
tima
l contro
ller
(BBO-F
LC) compar
ed w
i
th th
e
other contro
ller
s
desig
ned by
Genetic Alg
o
rit
h
m GA w
h
ich it is a
pow
erful meth
od h
a
s be
en foun
d to sol
v
e
the opti
m
i
z
at
io
n prob
le
m. T
h
e impl
e
m
ent
ati
on of BBO al
gorith
m
h
a
s b
een d
o
n
e
by
M-file/Matla
b, this
progr
a
m
link
e
d
w
i
th SIMULINK to calculate
the fine
sses fu
nction w
h
ich h
a
s the compl
e
te math
e
m
atic
a
l
system
m
o
del
has im
plem
ented using. The r
e
sults show
the excellent performanc
e of BBO-FLC com
p
ar
ed
w
i
th GA-F
LC
and PI control
l
er, also the p
r
opos
ed
meth
od w
a
s very fast and ne
ed
a few
numb
e
r of
iteratio
ns. T
hese results als
o
conf
irmed that
the transient torqu
e
an
d current never exc
e
ed the
maxi
mu
m
per
missi
bl
e val
ue.
Ke
y
w
or
ds
:
bio
geo
gra
phy-
base
d
opti
m
i
z
ation
(BBO), fu
zz
y
l
ogic
co
ntroller
(F
LC),
el
ectrical
veh
i
cle
(EV)
,
per
ma
nent
ma
gnet synchr
o
n
ous motor (PM
S
M)
Copy
right
©
2015 In
stitu
t
e o
f
Ad
van
ced
En
g
i
n
eerin
g and
Scien
ce. All
rig
h
t
s reser
ve
d
.
1. Introduction
We thin
k the con
c
ern abo
ut the enviro
n
ment
and e
nergy securit
y
such a
s
in
cre
a
si
ng
gasoline p
r
i
c
es o
r
de
pletion of fossil fuels i
s
chan
ging, due to
emerged el
e
c
tri
c
vehicl
es are
use
d
, anoth
e
r
word th
e pa
sseng
er vehi
cle
s
will
be e
nable
d
on th
e
grid, ma
ny rese
arche
r
s h
a
ve
recogni
ze
d that electri
c
dri
v
e vehicl
e
s
are critical to the future of
the indu
stry [1]. However, so
me
chall
enge
s e
x
ist to g
r
eate
r
adoptio
n,
in
United
States has pl
edg
e
d
to reduce gree
nho
use gas
(G
HG)
emi
ssi
ons
by app
ro
ximately 17 p
e
rcent
befo
r
e
the year
202
0 then fo
re
ca
st 7 million
EV
will be
sol
d
, the pe
rceptio
n of EV co
st
also
ch
argi
ng
infra
s
tru
c
ture may devel
o
ped, we beli
e
ve
that nume
r
ou
s adva
n
tage
s for Electri
c
V
ehicl
es EV
s compa
r
ed
with
Internal
Com
bustio
n
Engin
e
ICE vehi
cle
s
, espe
cially th
e EVs have
a hig
h
e
n
e
r
g
y
efficien
cy,
while
the
ene
rgy effici
en
cy of
ICE vehicl
es i
s
ab
out 30%,
the ene
rgy eff
i
cien
cy
of EVs is over th
an
80%, [2]. PMSM became
at
the top of ac
motors in the
medium ran
ge of
po
wer
and it be
cam
e
very popul
a
r
ch
oice in drive
techn
o
logy
o
v
er the
la
st f
e
w ye
ars
due
to
some
of it
s in
he
rent
ad
vantage
s. Th
ese
adva
n
tag
e
s
inclu
de hi
gh
torqu
e
to
curre
n
t ratio,
larg
e p
o
we
r to
weig
ht
ratio, hig
her efficien
cy a
nd
robu
stne
ss. T
here
a
r
e m
a
n
y
appli
c
ation
of PMSM
in
Elevators,
Wi
nd Ene
r
gy, E
V
drive
and
e
t
c.
becau
se it all
o
ws an
enl
arged
spe
ed
ra
nge
with inve
rter
si
ze lo
we
r than
in a
co
nventional flu
x
-
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN: 23
02-4
046
TELKOM
NI
KA
Vol. 16, No. 3, Dece
mb
er 201
5 : 509 – 519
510
oriente
d
in
du
ction m
o
tor
drive [3, 4].
In [5],
the a
daptive dyna
mic
su
rf
ace control (DS
C
)
of
PMSM ha
s b
een p
r
e
s
e
n
te
d. In [6, 7], the auth
o
rs h
ad de
rived
some fee
dba
ck
control de
sign
methods for
stability of PMSM in
their
results. Som
e
control m
e
t
hods had studied to stabili
ze the
PMSM syste
m
s, such a
s
slidin
g mo
d
e
co
nt
rol
(S
MC)[8],differe
ntial geo
metry method [9
],
passivity con
t
rol [10, 1
1
]. The ta
ngibl
e ben
efit
of ch
oosi
ng
controlle
r i
s
i
t
s sim
p
licity
to
impleme
n
t. It
is not e
a
sy to
find anoth
e
r
controlle
r
wit
h
su
ch
a si
m
p
le st
ru
cture t
o
be
com
para
b
le
in perfo
rman
ce. Fu
zzy rul
e
-ba
s
e
d
mod
e
ls are ea
sy to comp
reh
e
nd be
cau
s
e i
t
uses lin
gui
stic
terms
and th
e structu
r
e of
if-then rule
s [12, 13]. A ve
ry important
step in the
use of
cont
rolle
rs is
the controller pa
ramete
rs
and tu
ning
p
r
oce
s
s [14].
Unfortunately, i
n
spite of
this large
ran
g
e
of
tuning te
chni
que
s, the o
p
timum p
e
rfo
r
m
ance
cann
ot
be a
c
hieve
d
.
In re
cent ye
ar there
are m
a
ny
intelligent o
p
timization
techniqu
es
have
been
em
er
g
ed an
d g
e
t a
great
attentio
n of research
ers
like Ge
netic Algorithm (GA), Particl
e
Swarm
Opt
i
mization
(P
SO) techniq
ues b
ee
col
ony
optimizatio
n (BCO), Ant
Colony O
p
timization
(A
CO), Simulated
Annealin
g (S
A), and
Bact
eria
l
Fora
ging (BF
)
[15]. Usu
a
lly GA has a most algo
rith
ms found
ed i
n
the control
field, like the
search for optimal param
e
ters of
F
L
C controller. B
u
t it still
r
equires enorm
o
us
computational
effort. In this pap
er we
sugge
st a
ne
w
co
mp
utational th
eory
named
(Bio
g
eography
-Ba
s
ed
Optimizatio
n
BBO) to tune
para
m
eters o
f
FLC cont
roll
er. Thi
s
co
ntroller
can g
o
vern a
non
-lin
ear
sy
st
em.
2. Model for a PMSM Dri
v
e
The co
mplete
nonline
a
r mo
del of a PMSM without da
mper
windi
ng
s is a
s
follows:
)
+
i
(L
+
i
pL
+
Ri
=
v
af
d
d
s
q
q
q
q
(1)
i
L
-
p
+
Ri
=
v
q
q
s
d
d
d
(2)
v
d
and v
q
are
the d, q axis voltages, id a
nd iq are t
he
d,q axis state
r
cu
rre
nts, Ld
and Lq are the
d,q axis ind
u
c
tan
c
e,
R a
n
d
s
are
the
st
ater
re
sista
n
c
e
and
invert
er frequ
en
cy
respe
c
tively.
af
is the flux linkage du
e to the
rotor m
agn
ets linki
ng the
stator.
The elec
tric
torque:
)/2
i
)i
L
-
(L
i
3P(
T
q
d
q
d
q
af
e
(3)
The moto
r dynamics:
r
r
Jp
B
T
-
T
L
e
(4)
P is the
num
ber
of pole
p
a
irs,
TL i
s
th
e load
torq
ue
, B is the
da
mping
co
efficient,
r
is
the rotor
spe
ed and
J the moment of in
ertia . The
inverter fre
que
n
c
y is relate
d to the rotor
sp
eed
as
follows
:
r
s
p
(5)
The ma
chine
model is n
onl
inear a
s
it co
ntai
ns p
r
od
uct terms su
ch
as speed
with id and
iq. Note that
r
, i
d
and i
q
are
state varia
b
l
e
s. Durin
g
ve
ctor
cont
rol, i
d
is normally
forced to be
zer
o
.
q
t
q
e
i
K
/2
i
3P
T
af
(6)
Evaluation Warning : The document was created with Spire.PDF for Python.
TELKOM
NIKA
ISSN:
2302-4
046
Biogeog
rap
h
y
Base
d Opti
m
i
zation Tun
ed Fuzzy Log
ic Co
ntrolle
r to Adjust… (Sa
l
a
m
Wa
ley)
511
Figure 1. Block di
agram of
a PMSM
Figur
e 2. Block di
agram of
a PMSM Dri
v
e
3. Speed Control of PMSM Motor
The PMSM is usi
ng control to supp
re
ss ha
rmoni
c n
o
ise to a lev
e
l then, noise
to a level
belo
w
and vibration tra
n
sl
ates into a more
comfortable ride for passeng
ers.
IGBT SPWM
inverters ma
ke the
ride
more
smo
o
th
er with
pre
c
i
s
ely adju
s
tin
g
sp
eed
cont
rol with frequ
ency
and voltage regulation.It has the la
test low-noi
se
power units to
make the
ri
de even qui
eter.
Elevator has
dire
cted hig
h
-spe
ed used (1500 rpm)
P
M
SM. Energy reform in the
elevator gea
red
for small
rise
becau
se trav
el extremely small an
d fast.
Figure 3. Block
Diag
ram o
f
Speed Cont
rol of PMSM
3.1. Biogeog
raph
y
-
Based
Optimizatio
n
Inspired of biogeo
gra
phy Simon develo
ped
a ne
w ap
proa
ch
called
Biogeograph
y-Base
d
Optimiz
a
tion (BBO)
in (2008).
Thi
s
al
g
o
rithm i
s
an
example of h
o
w a
natural
pro
c
e
ss
ca
n
be
modele
d
to solve optimiza
t
ion [18]. In n BBO, each
possible
sol
u
tion is a
n
i
s
lan
d
and th
eir
features that descri
be habi
tab
ility are named
Habitat
Suitabilit
y In
dex (HSI). T
he goodness of
each solutio
n
is named S
u
itability Index Variables
(SIV). For example of
the
natural process,
why some i
s
l
and
s may le
a
n
towa
rd
s to
accumul
a
te
many mo
re
speci
e
s th
an o
t
hers? Be
cau
s
e
of po
ssess certain
enviro
n
mental
feat
ure
s
th
at
a
r
e
more
suitabl
e to
su
staini
ng that
ki
nd
than
other isl
and
s with
fewe
r sp
eci
e
s.
It is axiomati
c the
h
abita
ts with
hig
h
HSI h
a
ve l
a
rge
popul
ations,
also hig
h
im
migratio
n rate
and by feat
ure of a large
numbe
r of sp
ecie
s that migrate
to other h
abit
a
ts. The rate
of immigrati
on will
b
e
lo
wer if the
s
e
habitats a
r
e
alrea
d
y satu
rated
with spe
c
ie
s. On the
oth
e
r h
and, h
a
b
itats
with l
o
w
HSI hav
e high i
mmi
gration
and
low
immigratio
n rate, beca
u
se of the spa
r
se
popul
ation.
The fitne
s
s
function
FF
is a
s
so
ciate
d
wi
th
ea
ch
sol
u
tion
of Biogeo
grap
hy-Based
Optimizatio
n
BBO, which is anal
ogo
us
to HSI of a habitat. A good solutio
n
is
analo
gou
s to
a
habitat h
a
ving hig
h
HSI
a
nd a
poo
r
so
lution r
epresents
a h
abita
t having a
lo
w
HSI. The
best
solutio
n
s
sh
a
r
e thei
r g
e
o
g
rap
h
ie
s of t
he lo
we
st solution
s thro
w mig
r
ation
(emig
r
ation
a
n
d
immigratio
n).
Best solutio
n
s have mo
re resi
stan
ce
to chang
e than lowe
st sol
u
tions. While
the
lowe
st
solutio
n
s
have
more chan
ge fro
m
time to
tim
e
an
d a
c
cept
many
new fe
ature
s
from
b
e
st
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN: 23
02-4
046
TELKOM
NI
KA
Vol. 16, No. 3, Dece
mb
er 201
5 : 509 – 519
512
solutio
n
s. T
h
e immig
r
atio
n rate
an
d e
m
igratio
n
rate of the jth
i
s
lan
d
may
b
e
form
ulated
as
follows in Equation (7
), (8
) [17].
(
7
)
(
8
)
Whe
r
e µ,
λ
j
are th
e immi
gration
rate a
nd the
emig
ration rate of j
individual; I i
s
the
maxim
u
m
possibl
e immi
gration
rate;
E is th
e maxi
mum p
o
ssibl
e
e
m
igr
a
tion
ra
te
; j is th
e
nu
mb
er
o
f
s
pec
ies
of jth individual; and n is th
e maximum n
u
mbe
r
of spe
c
ie
s.
Jth In BBO, t
he mutatio
n
i
s
u
s
ed
to in
crea
se th
e div
e
rsity of th
e
popul
ation to
get the
best solution
s.
Mutation
ope
rator mo
difie
s
a
h
abitat’s SIV ran
dom
ly based
on
mutation
rate. The
mutation rate
mj is expre
ssed in (9
).
(
9
)
Whe
r
e mj is
the mutation
rate for the jth habitat ha
ving a j number of spe
c
ie
s; mmax is the
maximum mu
tation rate; Pmax is the maximum sp
ecies count pro
bability; Pj th
e spe
c
ie
s co
unt
prob
ability for the jth habitat and is given
by Equation (10
)
:
(
1
0
)
Whe
r
e µj
+1,
λ
j+1 are th
e immigratio
n and e
m
igration rate fo
r the jth ha
b
i
tat contain
s
j+1
spe
c
ie
s;
µj-1
,
λ
j-1 are th
e immigratio
n and emig
ration rate fo
r the jth habitat contain
s
j-1
spe
c
ie
s.
3.2. Fuzzy
Logic Con
t
rol
l
er
Fuzzy logi
c
controlle
rs hav
e the follo
win
g
adv
anta
g
e
s
over th
e
con
v
entional
con
t
rollers
that they are
che
ape
r to d
e
velop, they cover
a
wi
de
rang
e of op
erating c
onditio
n
s, an
d they are
more
readily
cu
stomizable
in n
a
tural la
ngua
ge te
rm
s. In M
a
md
a
n
i type FIS t
he
cri
s
p
re
su
lt is
obtaine
d by
defuzzificatio
n
[16], in th
e Mamd
ani F
I
S can b
e
u
s
ed for
both
multiple inp
u
t
and
singl
e output
and multiple i
nputs multipl
e
outputs
system as shown in Figure 4.
Figure 4. Arra
ngeme
n
t of fuzzy logi
c cont
rolle
r
The u
s
efuln
e
ss
of fuzzy logic
co
ntroll
er is ado
pte
d
espe
cially
in a compl
e
x and
nonlin
ear
system. The rules of conventional
FL
C are produ
ced dep
end
on the ope
rator'
s
Evaluation Warning : The document was created with Spire.PDF for Python.
TELKOM
NIKA
ISSN:
2302-4
046
Biogeog
rap
h
y
Base
d Opti
m
i
zation Tun
ed Fuzzy Log
ic Co
ntrolle
r to Adjust… (Sa
l
a
m
Wa
ley)
513
experie
nce or gene
ral
kno
w
led
ge of the
system in
a
heuri
s
tic
way.
The thresh
ol
ds of the fu
zzy
lingui
stic vari
able
s
are usually cho
s
e
n
arbitrarily
in the de
sign
proce
s
s. An improp
er
cont
ro
ller
value lea
d
s t
o
an a
d
verse co
nsequ
en
ce, un
stabl
e
mode,
collap
s
e a
nd
sep
a
ration [1
1]. This
work p
r
o
p
o
s
e
BBO to desi
gn an
Optima
l Fuzzy Logi
c Controlle
r O
F
LC, the o
p
timized
criteria
is
how to minimi
zing the tra
n
sient state.
3.3. Genetic
Algorithm O
p
timization
Many optimi
z
ation probl
ems
have
come to
be
sol
v
ed using Artificial Intelligent (AI).
Relatively Genetic Algorithm (GA
)
i
s
the most
wide
ly
use
d
in th
ese
tech
nique
s, i
t
is a
po
we
rful
tool even
recently there a
r
e
ma
ny
ne
w approa
che
s
h
a
ve
p
r
opo
se
d
for optimum
sea
r
ching. Th
e
variable
s
are
re
pre
s
e
n
ted
as
gen
es on
a chromo
som
e
.
GAs
featu
r
es a gro
up of
pop
ulation
o
n
the resp
on
se
su
rfa
c
e. Th
roug
h
Natu
ra
l Select
io
n
NS an
d g
e
n
e
tic o
perators, mutatio
n
and
cro
s
sove
r, chrom
o
some
s with better fitness fun
c
tion
s FF are found. NS suretie
s
the
recombi
natio
n ope
rato
r, the Gen
e
tic Algo
rithm
com
b
ine
g
ene
s from
best two p
a
rent
chromo
som
e
s to generate
two children
that at l
east one of them have a better fitness from
his
pare
n
ts. Muta
tion allows ne
w are
a
s of th
e respon
se
surface to be e
x
plored.
3.4. SVPWM In
v
e
rter Simulation
The Volta
ge
Source Sp
ace Vecto
r
P
u
lse Wi
dth Mo
d
u
lation
(SPWM) i
s
the
mo
st popul
ar
usa
ge i
n
A.C
drives. S
o
; its perfo
rma
n
ce
sh
oul
d
be
a
Voltage So
urce Inve
rter (V
SI) and
have
a
stiff so
urce
at the i
nput [6].
A pra
c
tical
(V
SI) cons
i
s
t
s
o
f
power b
r
idg
e
devi
c
e
s
wit
h
three
outpu
ts;
each o
ne
co
n
s
ist
s
of t
w
o
p
o
we
r
swit
che
s
a
nd t
w
o f
r
e
e
wh
eeling
di
ode
s. Th
e inv
e
rter is suppli
ed
from D.C.
vol
t
age source via
LC
filter. In
SVPW
M, t
he three out
put legs
considered
as three
indep
ende
nt push-p
u
ll am
plifiers a
s
sho
w
n in Figu
re
5.
Figure 5. A three ph
ase (V
SI) with three
phase re
ctifier
SVPWM inverter can be
s
i
mulated by
MATLAB/
SIMULINK. The
output of the s
w
it
c
h
es
gives (Vao,V
bo,Vco
)
then t
he thre
e pha
se
s to
load n
eutral (V
an,Vbn,Vc
n
)
ca
n be achieved
by
impleme
n
ting
Equation (1
1
)
.
(
1
1
)
3.5. Electric Vehicle (EV) Sy
stem
The dia
g
ra
m
of an Electri
c
Vehi
cle (E
V) System u
s
ing a
n
PMSM sup
p
lied b
y
voltage
inverter:
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN: 23
02-4
046
TELKOM
NI
KA
Vol. 16, No. 3, Dece
mb
er 201
5 : 509 – 519
514
Figure 6. Electri
c
Vehicl
e (EV) System
3.5.1. The Vehicle Load
The vehicl
e inertia torque
defined by th
e followin
g
rel
a
tionship:
(
1
2
)
3.5.2. Aerod
y
namics Force
The force is d
ue to the fricti
on of the vehicle bo
dy, moving throug
h the air.
(
1
3
)
The ae
rodyn
a
mics torq
ue
is:
(
1
4
)
3.5.3. Rolling Force
The rolling
resi
stan
ce i
s
prima
r
ily due
to the tra
c
tion of the tire on the
ro
d
e
. It is
prop
ortio
nal to vehicle
wei
ght, the equat
ion is:
(
1
5
)
(
1
6
)
3.5.4. Hill Cli
m
bing Force
The force ne
eded to drive
the vehicle u
p
a sl
op
e is the most
strai
ghtforward to find.
(
1
7
)
The slo
pe torque is:
(
1
8
)
Figure 7. The
vehicle up a
slop
e
Evaluation Warning : The document was created with Spire.PDF for Python.
TELKOM
NIKA
ISSN:
2302-4
046
Biogeog
rap
h
y
Base
d Opti
m
i
zation Tun
ed Fuzzy Log
ic Co
ntrolle
r to Adjust… (Sa
l
a
m
Wa
ley)
515
3.6. Elecric Vehicle Con
t
rol Sy
stem Application Fl
o
w
Cha
r
t
The Steps a
s
follows (flo
w
cha
r
t as Fig
u
re 8):
Step 1: Creat
e the para
m
e
t
ers of t
he Elecri
c Vehi
cle
Drive sy
stem.
Step 2: Accordin
g to the call sign
al and t
he cu
rrent situation
El
ecri
c Vehi
cle Drive
sy
st
em.
Step 3: By Reload
sq
uad
ron by u
s
in
g u
pdate
val
ue-weig
hted in
ertial, with the
spe
ed
of
conve
r
ge
nce.
Step 4: Use the a
cce
ptan
ce crite
r
ia, to
deci
de wheth
e
r to a
c
cept t
hese ne
w pa
rticles
or
not, and to incre
a
se the di
versity of the particl
es, with
avoid trappin
g
in local O
p
timization.
Step 5: The end of the iterati
on, then the global sea
r
ch the optima
l
solution. If not, step
loop (3
).
Step 6: by using Loo
p to step (2
) until the end
time si
mulation, the
n
output the result.
Figure 8. Electri
c
Vehicl
e (EV) control system appli
c
a
t
ion flow ch
art
4. Si
mulation Results
By using Si
mulation m
o
del PMSM&
S
imulati
on of
EV Drive
sy
stem by Impl
ementing
BBO Tuning f
o
r FL
C Para
meters.
4.1. Implementing BBO
Tuning for F
L
C Parame
te
rs
The imple
m
entation of
BBO in this work
i
s
sa
me wh
at co
mplex, beca
u
se the
perfo
rman
ce
of the system
must be exa
m
ined in
ea
ch iteration an
d particl
es p
o
s
ition du
ring t
he
optimizatio
n
algorith
m
. Th
erefo
r
e, the
o
p
timizati
on
al
gorithm
is im
plemente
d
by usi
n
g
MATL
AB
m-file p
r
o
g
ra
m an
d lin
ke
d
with the
syst
em
simulatio
n
p
r
og
ram
in
MATLAB SIMULING, to
ch
eck
the system p
e
rform
a
n
c
e i
n
each
iterati
on. In this pa
per, the p
r
obl
em sum
m
ari
z
ed in optimi
z
i
ng
three
varia
b
le
s, they
are:
o
ne o
u
tput a
n
d
two
inp
u
ts (spe
ed
and
th
e chan
ge i
n
speed
), ea
ch
one
has th
ree
dim
ensi
onal
spa
c
es, re
present
ed a
s
the
p
r
a
m
s of the tri
a
ngle me
mbe
r
ship
s of FL
C.
A
rand
om of 10
0, Habitats
were a
s
sume
d
and optim
i
z
a
t
ion algorith
m
of 100 iterati
ons i
s
used to
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN: 23
02-4
046
TELKOM
NI
KA
Vol. 16, No. 3, Dece
mb
er 201
5 : 509 – 519
516
estimate the
optimal value
s
of the FL
C cont
roller
parameters
. The fitnes
s
func
tion FF
whic
h
illustrate
d in Equation (19) can calcul
ate
by SIMULINK shown in Fi
gure 9.
(
1
9
)
Figure 9. Model of the system
Figure 10. Th
e conve
r
ge
nce of Fitness
Func
tion in 100 Iterations
Figure 11. Step re
spo
n
se OF PMSM sp
eed in
different co
ntrollers, GA-FL
C
and BBO
-F
LC
and PI contro
ller
Figure 12. An arbitra
r
y spe
ed between (1.1pu
and 0.7p
u)
Figure 13. FL
C memb
ershi
p
s de
sig
ned
by BBO
Evaluation Warning : The document was created with Spire.PDF for Python.
TELKOM
NIKA
ISSN:
2302-4
046
Biogeog
rap
h
y
Base
d Opti
m
i
zation Tun
ed Fuzzy Log
ic Co
ntrolle
r to Adjust… (Sa
l
a
m
Wa
ley)
517
Figure 10
shows the
co
nverge
nce of
Fitness Fu
nction in
10
0 iteration
s
and the
comp
ari
s
o
n
b
e
twee
n GA a
nd BBO. Figu
re 11, 1
2
the
step respon
se with loa
d
a
nd no lo
ad u
s
ing
prop
osed con
t
roller and G
A
-FLC
and
PI
-co
n
troll
e
r
tu
ned by
conve
n
tional meth
o
d
trial an
d e
r
ror.
Figure 13 sho
w
s F
L
C de
sig
ned by BBO and Figu
re 1
4
sho
w
s the surface of FLC
4.2. Simulati
on of Elecric Vehcile Drive S
y
stem
Electri
c
vehicl
e has
submitt
ed a numb
e
r
of tests du
rin
g
the variou
s route
s
:
This test
cla
r
i
f
y the effect
of the d
e
sce
n
t of
vehi
cle moving
o
n
straight ro
ad, This
test
explain the
effect of the
slo
pe on th
e EV, EV torque
i
n
cre
a
se, the E
V
are d
r
iving
in straight roa
d
with
con
s
tant
sp
eed, the
speed
incre
a
se, the v
ehi
cl
e is drivin
g o
n
a
cu
rved
road
on the
ri
ght
side, th
e EV t
o
rqu
e
ju
mp
s
down, the ve
hicle
is dr
ivin
g on
a
cu
rve
d
ro
ad
on th
e
left sid
e
a
nd
EV
torque ju
mp
s back.
4.2.1. Speed, Torque &
Current
Res
p
onse
w
i
th
O
n
e Side at Speed (500,60
0,400,50
0)
In this
step th
ere
are
three
figure
s
(a.sp
eed
re
spo
n
se
, b.Torq
ue
c.
curre
n
t ), a
s
sho
w
in
figure
s
belo
w
:
a) sp
eed
re
sp
onse
b) To
rque
re
spon
se
c
)
Current respons
e
Figure 14. Simulation respon
se of
(a.spe
ed, b.To
rque
c.cu
rren
t) with one
side
4.2.2. Speed, Torque & Current
Res
p
onse
w
i
th T
w
o
Side
s at
Speed (5
00,
1000,1
500,-5
00,-
1000,-150
0)
In this step there are three
figures (a.sp
eed
re
sp
on
se
, b.Torque, c.
curre
n
t), as show in
figure
s
belo
w
:
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN: 23
02-4
046
TELKOM
NI
KA
Vol. 16, No. 3, Dece
mb
er 201
5 : 509 – 519
518
a) sp
eed
re
sp
onse
b) To
rque
re
spon
se
c) Current
respon
se
Figure 14. Simulation resp
onse of (a.sp
eed,
b.Torqu
e
, c.cu
rre
nt)
with two si
de
s
5. Conclusio
n
The sy
stem responses
of different tuni
ng methods
are
illustrated in simul
a
tion result
and a
compa
r
able p
e
rfo
r
m
ance bet
wee
n
the th
ree
cont
rolle
rs in t
h
is
re
sea
r
ch (PI controller,
GA-
FLC
and BB
O-FL
C) a
s
sh
own i
n
Fig
u
re
11, 12.
We f
i
nd the
optimi
z
ed BB
O-F
L
C is cl
osed
with
desi
r
ed
spe
e
d
and its pe
rforma
nce is th
e best compa
r
ed with
GA-FLC an
d PI controlle
r.
We ca
n obtai
n the followin
g
con
c
lu
sio
n
s through
simu
lation analy
s
is:
1) Thi
s
pape
r design fu
zzy
logic co
ntrol
by co
mputati
onal algo
rith
m, it interject Control
con
c
e
p
ts of t
r
ial a
nd e
r
ror in fuzzy cont
rol a
nd
conv
entional
GA-FLC m
e
thod
and the
n
con
t
rol
velocity modu
lation of elect
r
ical vehi
cle
EV in different speed.
2)
Obviou
sly, the BBO tu
n
i
ng of th
e FL
C i
s
the
be
st
intelligent
m
e
thod
whi
c
h
gives
an
excelle
nt sy
stem pe
rforma
nce,
and
the
GA give
s a
g
ood
re
spo
n
se
with
re
spe
c
t
to the traditio
nal
trial and e
rro
r method.
3) In ad
dition
to the impro
v
ing of syste
m
re
spo
n
se, the BBO and
GA can
use
a highe
r
orde
r
system
in the tuning
pro
c
e
ss
whi
c
h avoid
s
the
error of
syst
em order
red
u
ction. It gave a
satisfa
c
to
ry solution du
ring
the first 30 iteration
s
a
s
shown in Figu
re 11.
4) T
he
proposed method
makes
control system have st
rong
flexi
b
ility, instantaneity and
reliability because of the
advanced prediction of FL
C predi
cting controller.
5) It makes control sy
stem
have stronger Re
al-time
controll abilit
y because of
optimal
fuzzy
param
eters have
a
head
predi
ct
for a
po
ssi
bl
e interfe
r
e
source. T
he l
o
we
r inte
rference
freque
ncy, BBO algorithm
is more co
ntrollable.
Referen
ces
[1]
Basu M
a
la
bika
, Gaugh
an K
e
v
i
n, Co
yl
e E
u
g
e
ne.
H
a
rmon
i
c d
i
stortion c
aus
e
d
by EV
b
a
ttery charg
e
rs i
n
the distribution system
s
ne
twork and
its remedy.
In
39th
Internati
ona
l U
n
iv
ersities
Po
w
e
r
Engi
neer
in
g
Confer
ence, U
PEC. 2004.
Evaluation Warning : The document was created with Spire.PDF for Python.