Int
ern
at
i
onal
Journ
al of Ele
ctrical
an
d
Co
mput
er
En
gin
eeri
ng
(IJ
E
C
E)
Vo
l.
10
,
No.
5
,
Octo
be
r
2020
,
pp.
4782
~
4788
IS
S
N: 20
88
-
8708
,
DOI: 10
.11
591/
ijece
.
v10
i
5
.
pp
4782
-
47
88
4782
Journ
al h
om
e
page
:
http:
//
ij
ece.i
aesc
or
e.c
om/i
nd
ex
.ph
p/IJ
ECE
Des
i
gn o
f a
n
adap
tive
st
ate fe
edback c
ont
ro
ll
er
for a ma
gnetic l
evitati
on syst
em
Omar
W
aleed
A
b
dulw
ahha
b
Depa
rtment
o
f
C
om
pute
r
Engi
n
e
eri
ng,
Univer
si
t
y
of
Baghd
ad, I
ra
q
Art
ic
le
In
f
o
ABSTR
A
CT
Art
ic
le
history:
Re
cei
ved
Feb
12
, 20
20
Re
vised
Ma
r 1
0
,
2020
Accepte
d
Ma
r
25
, 202
0
Thi
s
pap
er
pre
s
e
nts
designi
ng
an
ada
p
ti
ve
state
f
ee
dba
ck
cont
rol
l
er
(AS
FC
)
for
a
m
agne
t
ic
l
evi
t
at
ion
s
y
ste
m
(MLS),
which
i
s
an
unstable
s
y
s
te
m
and
has
high
nonli
n
ea
ri
t
y
and
rep
r
ese
n
ts
a
ch
al
l
engi
n
g
cont
rol
probl
em.
First,
a
nonad
apt
iv
e
state
fe
edback
c
ontrol
ler
(SF
C)
is
designe
d
b
y
l
ine
ar
iz
a
ti
o
n
abou
t
a
select
ed
equi
li
brium
poi
nt
and
designi
n
g
a
SF
C
by
pol
e
-
place
m
ent
m
et
hod
to
ac
hieve
m
axi
m
um
ov
ershoot
of
1.
5%
and
sett
li
ng
t
ime
of
1s
(5%
cri
t
eri
on).
W
hen
the
oper
ating
point
cha
ng
es,
th
e
designe
d
cont
r
oll
er
c
an
no
longe
r
a
chieve
the
design
spe
c
ifi
c
at
io
ns
,
sin
ce
it
is
design
ed
base
d
on
a
li
n
ea
r
iz
a
ti
on
a
bout
a
d
iffe
r
ent
oper
ating
point.
Thi
s
give
s
r
ise
t
o
uti
l
iz
in
g
the
ad
aptive
con
trol
sche
m
e
to
p
ara
m
et
er
ize
th
e
stat
e
f
ee
db
ac
k
c
ontrol
ler
in
te
rm
s
of
the
oper
ating
point.
The
resul
ts
of
the
sim
ula
ti
on
show
tha
t
the
oper
a
ti
ng
po
int
has
signifi
c
a
nt
eff
e
ct
on
the
per
form
anc
e
of
nonada
pt
ive
SF
C,
and
thi
s
p
e
rform
anc
e
m
a
y
degr
ade
as
th
e
o
per
ating
poin
t
d
evi
a
te
s
from
the
equ
il
ibr
ium
point
,
while
the
AS
FC
ac
hi
eve
s
th
e
req
u
ir
ed
design
spec
ifica
t
ion
for
an
y
oper
ating
point
and
outperform
s
the
state
fee
db
ac
k
cont
roller
f
rom
t
his po
int of
v
ie
w
.
Ke
yw
or
d
s
:
Ad
a
p
ti
ve st
at
e feedbac
k
con
t
ro
ll
er
Indirect
ly
ap
un
ov
'
s the
or
em
Ma
gn
et
ic
levit
at
ion
syst
em
Un
sta
ble
nonlinear
syst
em
Copyright
©
202
0
Instit
ut
e
o
f Ad
vanc
ed
Engi
n
ee
r
ing
and
S
cienc
e
.
Al
l
rights re
serv
ed
.
Corres
pond
in
g
Aut
h
or
:
Om
ar
W
al
eed
Abd
ulwa
hh
a
b,
Dep
a
rtm
ent o
f C
om
pu
te
r
E
ng
i
neer
i
ng,
Un
i
ver
sit
y o
f B
aghda
d,
Ir
a
q
.
Em
a
il
:
o
m
ar.wal
eedl@coe
ng.
uoba
ghda
d.
e
du.iq
1.
INTROD
U
CTION
Ma
gn
et
ic
le
vitat
ion
te
ch
no
l
ogy
has
re
centl
y
beco
m
e
an
interest
in
g
to
pic
of
stu
dy,
si
nc
e
it
is
a
go
od
so
luti
on
f
or
m
any
m
otion
syst
e
m
s
[1
,
2].
T
he
ad
va
ntages
of
a
MLS
a
re
it
s
abili
ti
es
to
el
i
m
inate
fr
ic
ti
on
by
el
i
m
inati
ng
th
e
con
ta
ct
betw
een
m
ov
in
g
a
nd
sta
ti
on
a
ry
par
ts
[3
]
,
decre
asi
ng
t
he
co
s
t
of
m
ai
ntenance
,
an
d
achievin
g
pr
ec
ise
po
sit
io
n
[
4].
Th
e
MLSs
has
beco
m
e
su
it
a
ble
f
or
tr
ai
ns
,
bear
i
ng
s
,
vibrat
ing
isol
at
ion
syst
e
m
s,
and levit
at
ion
of w
i
nd tu
nnel
[1,
4
]
.
By
m
agn
et
ic
le
vitat
ion
,
a
fe
rrom
agn
et
ic
m
ass
is
su
s
pe
nded
in
t
he
ai
r
by
an
el
ect
ric
m
agn
et
ic
fiel
d
[
5].
T
he
basic
co
ntr
ol
aim
is
to
pr
e
ci
sel
y
po
sit
ion
the
le
vita
ti
ng
obj
ect
[
6].
T
o
sta
bili
ze
the
MLS,
the
m
agn
et
ic
fiel
d
stren
gth
m
us
t
be
var
ie
d
by
cha
ng
i
ng
the
cur
r
ent
of
the
coil
[5,
7].
Since
the
MLS
is
un
sta
ble
an
d
has
high
no
nlinearit
y,
design
i
ng
a
co
ntr
ol
le
r
for
this
sy
stem
with
ade
qu
at
e
s
pecifica
ti
ons
is
no
t
a
trivia
l
ta
sk
;
thus,
t
he
con
t
ro
l
of
this
syst
e
m
has
rec
ei
ved
c
onside
r
able
interest
[
4]
,
and
it
has
be
com
e
a p
la
tf
or
m
to
te
st dif
fer
e
nt con
trol alg
or
it
hm
s
[1, 5].
Seve
ral
con
tr
ol
app
r
oac
hes
wer
e
use
d
to
s
ta
bili
ze
the
M
LS,
su
c
h
as
fe
edb
ac
k
li
near
i
zat
ion
[
8
-
10
]
,
wh
ic
h
requires
an
accurate
m
od
el
of
this
syst
e
m
;
ho
we
ve
r,
ob
ta
ini
ng
a
n
accurate
m
od
el
represe
nts
a
prob
le
m
because
of
the
hi
gh
no
nlinea
rity
of
this
sy
stem
and
the
var
ia
ti
on
of
th
e
gai
n
par
am
et
er
with
the
di
sta
nce
betwee
n
the
le
vitat
ing
obj
ect
and
th
e
m
agn
e
t.
Lineariza
ti
on
-
base
d
m
et
ho
ds
wer
e
al
so
us
e
d,
w
her
e
the
sy
stem
is
li
near
iz
ed
a
bout
a
ce
rtai
n
equ
il
ib
rium
point
an
d
a
c
on
t
r
oller
is
desig
ne
d
to
sta
bili
ze
the
syst
em
,
su
ch
as
PI
D
co
ntr
oller
[1,
2,
5,
6,
7,
11]
,
f
ract
ion
al
order
P
ID
c
ontr
oller
[4,
12
-
15]
,
L
QR
[
1,
2,
16
,
17]
,
le
ad
com
pen
sa
tor
[1
]
,
H
_∞
con
t
ro
ll
er
[
18,
19]
,
f
uzzy
lo
gic
co
ntr
oller
(F
LC)
[
16,
20,
21]
,
a
nd
ada
ptiv
e
FLC
[22];
howev
e
r,
the
perform
ance
of
su
ch
c
on
t
ro
ll
ers
degrad
e
wh
e
n
the
dev
ia
ti
on
betwee
n
the
operati
ng
Evaluation Warning : The document was created with Spire.PDF for Python.
In
t J
Elec
&
C
om
p
En
g
IS
S
N: 20
88
-
8708
Desig
n of
an adapti
ve state
fe
edback c
ontrol
l
er for a
mag
ne
ti
…
(
Om
ar W
aleed A
bd
ulwa
hhab
)
4783
po
i
nt
a
nd
the
equ
il
ib
rium
point
(the
point
that
the
syst
em
was
li
near
iz
e
d
a
bout)
inc
re
ases.
T
o
ha
nd
l
e
this
pro
blem
,
sl
iding
m
od
e
co
nt
ro
ll
er
(S
MC
)
[23
-
25
]
,
a
da
ptive
SMC
[
26
]
,
P
ID
-
no
tc
h
filt
ers
[
27
]
,
and
li
near
iz
at
ion
-
ga
in
sche
duli
ng
con
t
ro
ll
er
PID
con
t
ro
ll
er
[28
]
,
li
near
iz
at
ion
-
gai
n
sche
duli
ng
PI
c
ontr
oller
[
29
]
,
and
li
near
iz
at
ion
-
a
dap
ti
ve
P
D
c
on
t
ro
ll
er
[
30
]
wer
e
desi
gn
e
d
t
o
pro
vide
r
obus
tne
ss
a
gainst
operati
ng
po
i
nt
var
ia
ti
on.
T
his
pa
per
pro
pos
es
an
ASFC
to
sta
bili
ze
the
MLS,
wh
e
re
the
co
ntr
oller
par
am
et
ers
be
com
e
a
functi
on
o
f
t
he
op
e
rati
ng
point,
a
nd
po
le
placem
ent
m
eth
od
is
us
e
d
to
desig
n
t
he
c
ontr
oller.
The
r
es
t
of
this
pap
er
is:
section
2
pre
se
nt
s
the
m
athe
m
ati
cal
m
odel
of
the
MLS,
section
3
present
s
the
desig
n
of
a
n
A
SFC
for
this
syst
em
by
po
le
place
m
ent,
sim
ul
at
ion
res
ults
a
nd
disc
us
si
on
s
are
giv
e
n
i
n
sect
ion
4,
an
d
finall
y
the concl
usi
on
s that ca
n be
drawn f
ro
m
the obtai
ned res
ults
are
giv
e
n
in
se
ct
ion
5.
2.
MA
T
HEM
AT
ICA
L
MODE
L OF THE
M
LS
In
a
MLS,
a
fe
rrom
agn
et
ic
ball
is
le
vitat
ed
by
a
m
agn
et
ic
fiel
d,
and
t
he
ba
ll
po
sit
ion
is
fed
back
t
o
con
t
ro
l t
he
c
urren
t
of the c
oil
[31]. T
he p
os
it
ion
of the
b
al
l i
s
̈
=
−
̇
+
+
(
,
)
(1)
w
h
e
r
e
a
n
d
a
r
e
t
h
e
m
a
s
s
t
h
e
v
e
r
t
i
c
a
l
p
o
s
i
t
i
on
o
f
t
h
e
b
a
l
l
,
k
i
s
a
v
i
s
c
o
u
s
f
r
i
c
t
i
o
n
c
o
e
f
f
i
c
i
e
n
t
,
i
s
t
h
e
g
r
a
vi
t
y
a
c
c
e
l
e
r
a
t
i
o
n
,
(
,
)
i
s
t
h
e
e
l
e
c
t
r
o
m
a
g
n
e
t
f
o
r
c
e
,
a
n
d
i
s
t
h
e
c
o
i
l
c
u
r
r
e
n
t
[
3
1
]
.
T
h
e
i
n
d
u
c
t
a
n
c
e
w
h
i
c
h
i
s
a
f
u
n
c
t
i
o
n
o
f
t
h
e
b
a
l
l
p
o
s
i
t
i
o
n
i
s
a
p
p
r
o
x
i
m
a
t
e
l
y
(
)
=
1
+
0
1
+
(2)
wh
e
re
1
is
the
el
ect
ro
m
agn
et
ic
coil
inducta
nce
with
out
th
e
su
s
pende
d
ba
ll
,
0
is
the
inducta
nce
du
e
t
o
the
ball,
an
d
is
the
ai
r
gap
wh
e
n
t
he
le
vitat
ed
ball
is
in
equ
il
ib
rium
[32].
T
he
i
nduct
ance
has
it
s
hi
gh
e
st
value
1
+
0
as
the
ball
tou
c
hes
th
e
m
agn
et
an
d
decr
ease
s
to
1
wh
e
n
it
is
rem
ov
e
d
a
way
fro
m
the
coil.
If
(
,
)
=
1
2
(
)
2
is t
he
elec
tr
om
agn
et
stored
en
e
rg
y,
the
for
ce
F
is [
31]
(
,
)
=
=
−
0
2
2
(
1
+
)
2
(3)
The
m
agn
et
ic
fl
ux
li
nka
ge
is
=
(
)
(4)
and acco
r
ding
to K
i
rchh
off'
s v
oltage
law,
th
e coil v
oltage
is
=
̇
+
(5)
wh
e
re
is
the
ci
rcu
it
resist
ance
.
Using
1
=
,
2
=
̇
,
an
d
3
=
as
sta
te
var
ia
bl
es,
=
as
co
ntro
l
input
,
and
as the c
on
trolle
d o
utput,
the stat
e m
at
ri
x
e
qu
at
io
n
a
nd
the outp
ut e
qu
at
ion
beco
m
e
:
̇
=
[
2
−
2
−
0
3
2
2
(
+
1
)
2
1
(
1
)
(
−
3
+
0
2
3
(
+
1
)
2
+
)
]
=
(
,
)
(6)
=
1
(7)
The
e
quil
ibrium
po
int
of
sys
tem
(6
)
ca
n
be
fou
nd
by
set
ti
ng
̇
=
0
.
If
t
his
po
i
nt
is
de
sig
nated
by
(
ss
,
ss
)
wh
e
re
ss
=
[
1ss
2
s
s
3ss
]
=
[
,
2ss
,
ss
]
, a
nd
ss
=
ss
, th
en
0
=
2ss
,
(8)
0
=
−
2ss
−
0
ss
2
(
+
)
2
(9)
0
=
1
(
)
(
−
ss
+
0
2
ss
2
(
+
)
2
+
ss
)
(10)
So
lvi
ng (8)
-
(
10)
for
,
2ss
,
a
nd
ss
in term
s o
f
ss
yi
eld
s
Evaluation Warning : The document was created with Spire.PDF for Python.
IS
S
N
:
2088
-
8708
In
t J
Elec
&
C
om
p
En
g,
V
ol.
10
, No
.
5
,
Oct
ob
e
r 2
020
:
47
82
-
47
88
4784
[
,
2ss
,
ss
]
=
[
1
√
0
2
ss
−
,
0
,
ss
]
(11)
The
li
nea
rizat
ion o
f
syst
em
(
6) ab
out t
he
e
quil
ibriu
m
p
oi
nt
(
ss
,
ss
)
is
̇
=
+
(12)
wh
e
re
=
|
(
ss
,
ss
)
an
d
=
|
(
ss
,
ss
)
.
Fo
r
a
ny
e
quil
ibri
um
po
int
(
ss
,
ss
)
,
at
le
ast
one
of
the
t
hr
ee
ei
genvalues
of
m
at
rix
has
pos
it
ive
real
par
t.
Th
us
,
by
i
nd
i
re
ct
Ly
apun
ov
'
s
The
or
em
,
the
s
yst
e
m
is
un
sta
ble.
The val
ues
of t
he param
et
ers
of the ML
S ar
e
g
ive
n
i
n
Ta
ble
1
.
Table
1.
Param
et
ers
of t
he
M
LS
Para
m
eter
Descripti
o
n
Valu
e
Mass o
f
the b
all
0
.1k
g
Visco
u
s
f
riction
coef
f
icien
t
0
.00
1
N/
m
/s
Gravity
acc
el
eratio
n
9
.81
m
/s
2
Air
g
ap
when
the l
ev
itated
ball is in
e
q
u
ilib
riu
m
0
.05
m
0
Ind
u
ctan
ce du
e to lev
itated
ball
0
.01
H
1
Electr
o
m
ag
n
etic c
o
il ind
u
ctan
ce witho
u
t the su
sp
en
d
e
d
ball
0
.02
H
Series
resistan
ce o
f
the circuit
1Ω
3.
CONTR
OLL
ER D
E
SIG
N
To dem
on
strat
e the e
nh
a
nce
d pe
rfor
m
ance of th
e
prop
os
e
d ASFC,
a
nona
da
ptive S
FC
=
−
=
[
1
2
3
]
[
1
2
3
]
(13)
I
s
desig
ne
d
t
o
sta
bili
ze
the
cl
os
e
d
lo
op
syst
e
m
at
=
0
.
04
m
,
wh
ic
h
c
orres
ponds
to
t
he
e
qu
il
ib
rium
po
i
nt
(
ss
,
ss
)
=
(
[
0
.
04
0
5
.
6377
]
,
5
.
6377
)
.
For t
his
equ
il
ib
rium
p
oi
nt,
the li
near
s
yst
e
m
(
12) bec
om
es
̇
=
[
0
1
0
218
−
0
.
1
−
3
.
4801
0
13
.
6178
−
39
.
13304
]
+
[
0
0
39
.
13304
]
(14)
The
gain
m
at
ri
x
is
desig
ned
t
o
locat
e
the
cl
os
e
d
lo
op
pole
s
at
po
sit
io
ns
s
o
that
the
pe
rc
entage
overs
ho
ot
is
1.5% a
nd the
s
et
tl
ing
tim
e is 0.5s (5%
crite
r
ion).
=
−
√
1
−
2
⟹
1
.
5
100
=
−
√
1
−
2
⟹
=
0
.
8
an
d
s
=
3
⟹
0
.
5
=
3
0
.
8
⟹
=
7
.
5
rad
s
Th
us
,
the
tw
o
com
plex
conj
ugat
e
pole
s
are
−
6
±
j4
.
5
.
To
m
ake
the
po
le
s
1
,
2
do
m
inant,
the
thir
d
po
le
is
sel
ect
ed
su
c
h
t
hat
|
Re
(
3
)
|
≥
5
|
Re
(
1
,
2
)
|
; l
et
3
=
−
30
. U
sin
g Ac
ker
m
an'
s f
orm
ula, the
gai
n m
a
trix is
=
[
0
0
1
]
[
2
]
−
1
(
(
−
(
−
6
+
4
.
5
)
)
(
−
(
−
6
−
4
.
5
)
)
(
−
(
−
30
)
)
)
=
[
−
79
.
6114
−
4
.
3064
0
.
0731
]
(15)
and the c
ontr
ol law
beco
m
es
=
+
=
−
79
.
6114
1
−
4
.
3064
2
−
0
.
0731
3
+
5
.
6377
(16)
A
blo
c
k
diag
ra
m
of
the
MLS
with
S
FC
is
s
how
n
i
n
Fi
gure
1.
A
draw
bac
k
of
this
co
ntr
oller
is
that
it
as
su
re
s
the
sta
bili
zat
ion
of
the
sy
stem
and
it
achieves
t
he
require
d
desig
n
s
pecifica
ti
ons
on
ly
in
a
certai
n
neig
hborh
ood
of
the
li
nea
riz
at
ion
-
base
d
po
int,
i
.e.,
the
e
qu
il
ib
rium
po
int
that
corres
ponds
to
=
0
.
04
m
.
To
sta
bili
ze the syst
e
m
at anot
her
posit
ion
,
the c
on
t
ro
ll
er m
ay
f
ai
l t
o
sta
bil
iz
e the syst
em
,
o
r
at le
ast
it
wi
ll
n
ot
achieve t
he req
uire
d desig
n
s
pe
ci
ficat
ion
s.
Evaluation Warning : The document was created with Spire.PDF for Python.
In
t J
Elec
&
C
om
p
En
g
IS
S
N: 20
88
-
8708
Desig
n of
an adapti
ve state
fe
edback c
ontrol
l
er for a
mag
ne
ti
…
(
Om
ar W
aleed A
bd
ulwa
hhab
)
4785
To
overc
om
e
this
pro
blem
,
a
n
a
dap
ti
ve
sta
te
fee
db
a
ck
c
ontrolle
r
is
desi
gn
e
d.
This
ca
n
be
achieve
d
by
pa
ram
et
erizi
ng
the
li
near
syst
e
m
(1
4)
in
te
rm
s
of
it
s
equ
il
ibriu
m
po
in
t,
i.e.,
the
qu
a
ntit
ie
s
ss
=
[
1ss
2
s
s
3
s
s
]
and
ss
=
ss
are
not
gi
ven
c
onsta
nt
values;
rat
her,
they
are
co
ns
i
der
e
d
as
par
a
m
et
ers,
and syst
em
(
14)
ca
n be r
e
w
ritt
en
as
̇
=
(
ss
,
ss
)
+
(
ss
,
ss
)
(17)
ss
=
[
1
0
√
0
2
]
+
[
0
0
√
0
2
]
=
(
)
(18)
ss
=
√
0
2
(
+
)
=
(
)
(19)
and syst
em
(
17)
ca
n be r
e
w
ritt
en
as
̇
=
(
)
+
(
)
(20)
and the
ASFC
is
=
−
(
)
=
[
1
(
)
2
(
)
3
(
)
]
[
1
2
3
]
(21)
wh
e
re
(
)
is gi
ven b
y
(
)
=
[
0
0
1
]
[
(
)
(
)
(
)
(
(
)
)
(
)
]
−
1
(
(
−
(
−
6
+
4
.
5
)
)
(
−
(
−
6
−
4
.
5
)
)
(
−
(
−
30
)
)
)
(22)
The
c
ontr
ol
la
w
(21)
is
a
fa
m
ily
of
co
ntr
ol
le
rs,
i.e
.,
a
n
ad
aptive
sta
te
fee
db
ac
k
co
ntr
oller,
w
ho
s
e
par
a
m
et
ers
1
,
2
,
an
d
3
are
ch
ang
e
d
(
desi
gn
e
d)
acc
ordin
g
t
o
the
val
ue
of
the
ref
e
re
nce
input
.
A
blo
c
k
diagr
am
of
the MLS
with
AS
FC is
sho
w
n
in
Fig
ure
2.
Figure
1. Bl
oc
k diag
ram
o
f
th
e MLS
with S
F
C
Figure
2. Bl
oc
k diag
ram
o
f
th
e MLS
with
A
SFC
4.
SIMULATI
O
N RESULTS
AND DIS
C
USSION
A
sim
ulatio
n
of
the
cl
os
e
d
lo
op
MLS
was
c
arr
ie
d
out
us
in
g
script
MA
T
LAB
pro
gr
am
.
Fo
ur
cases
wer
e
co
ns
i
der
e
d,
re
gardin
g
th
e
op
erati
ng
ra
nge
of
the
syst
e
m
.
The
first
case
is
wh
en
the
syst
e
m
op
erat
es
in
a
ra
nge
t
hat
li
es
relat
ively
c
lose
t
o
t
he
e
quil
ibriu
m
p
oint
that
c
orres
pond
s
to
=
0
.
04
m
;
this
r
ang
e
was
achieve
d
by
ta
king
a
n
init
ia
l
posit
io
n
0
=
0
.
02
m
a
nd
a
desire
d
posi
ti
on
=
0
.
06
m
.
The
sec
ond
case
is
wh
e
n
the
syst
e
m
op
erates
i
n
a
ra
ng
e
t
hat
dev
ia
te
s
f
r
om
the
equ
il
ibr
ium
po
int
by
a
relat
ively
m
od
e
rate
distance;
this ra
ng
e
was
ac
hie
ved
by taki
ng
an
init
ia
l po
sit
i
on
0
=
0
.
06
m
an
d
a
desir
ed
posit
io
n
=
0
.
10
m
.
The
thir
d
case
is
wh
en
the
sy
stem
op
erates
in
a
range
that
dev
ia
te
s
f
ro
m
t
he
eq
uili
br
ium
po
int
by
a
rela
ti
vely
la
rg
e
distanc
e;
this
ra
ng
e
w
as
ac
hiev
e
d
by
ta
king
a
n
in
it
ia
l
po
sit
ion
0
=
0
.
10
m
and
a
desi
red
po
sit
io
n
=
0
.
14
m
.
T
he
f
ourt
h
ca
se
is
w
he
n
t
he
syst
e
m
op
e
rates
in
a
wide
ra
nge;
t
his
range
was
achie
ved
by
ta
king
an
init
ia
l
posit
ion
0
=
0
.
01
m
and
a d
esi
r
ed
po
sit
io
n
=
0
.
10
m
.
Figure 3
s
hows
t
he
ra
nges o
f
t
he
operati
ng
po
i
nts
of
the
f
our
cases
relat
ive
to
the
li
ne
arizat
ion
-
base
d
po
i
nt,
an
d
Ta
ble
2
s
hows
t
he
pe
rfor
m
anc
e
of
the syst
em
w
ith
the
SF
C
and
with the
ASFC
, for al
l case
s.
Evaluation Warning : The document was created with Spire.PDF for Python.
IS
S
N
:
2088
-
8708
In
t J
Elec
&
C
om
p
En
g,
V
ol.
10
, No
.
5
,
Oct
ob
e
r 2
020
:
47
82
-
47
88
4786
Figure
3. Ra
ng
es of
op
e
rati
ng
points
of
the
f
our
ca
ses
Table
2.
Per
for
m
ance of th
e
s
yst
e
m
Ris
e ti
m
e
(s)
Settlin
g
ti
m
e
(s)
Percentag
e
o
v
ersh
o
o
t
Cas
e 1
SFC
0
.26
0
.94
3
.30
%
ASFC
0
.29
0
.58
0
.73
%
Cas
e 2
SFC
0
.28
1
.10
4
.92
%
ASFC
0
.31
0
.54
0
.47
%
Cas
e 3
SFC
0
.30
1
.21
5
.08
%
ASFC
0
.32
0
.51
0
.35
%
Cas
e 4
SFC
Un
stab
le
ASFC
0
.34
0
.69
0
.53
%
The
res
ults
give
n
in
Ta
ble
2
s
hows
that
as
th
e
op
e
rati
ng
point
de
viate
s
fro
m
the
li
near
iz
at
ion
-
base
d
po
i
nt,
the
perf
or
m
ance
of
the
SFC
de
gr
a
de
d
(the
rise
tim
e,
the
set
tl
ing
ti
m
e,
an
d
the
perc
entage
overs
ho
ot
are
increase
d),
an
d
it
becam
e
un
sta
ble
in
t
he
four
t
h
case.
H
ow
e
ve
r,
the
A
SFC
sho
wed
be
tt
er
perform
a
nce
an
d
rob
us
tness
,
sin
ce
the
co
ntr
oller
gai
n
m
at
rix
was
a
dap
te
d
w
it
h
ever
y
new
r
efere
nce
in
put
to
m
ai
ntain
the
sam
e
require
d
desi
gn sp
e
ci
ficat
io
ns of
the syste
m
. Th
e res
ponse
s o
f
both c
ontr
ollers for
the four cases are s
how
n
in
Figure
s
4
-
11.
Figure
4. Step
respo
ns
e
of M
LS w
it
h
S
FC:
case 1
Figure
5. Step
respo
ns
e
of M
LS w
it
h
S
FC:
case 2
Figure
6. Step
respo
ns
e
of M
LS w
it
h
S
FC:
case 3
Figure
7. Step
respo
ns
e
of M
LS w
it
h
S
FC:
case 4,
Evaluation Warning : The document was created with Spire.PDF for Python.
In
t J
Elec
&
C
om
p
En
g
IS
S
N: 20
88
-
8708
Desig
n of
an adapti
ve state
fe
edback c
ontrol
l
er for a
mag
ne
ti
…
(
Om
ar W
aleed A
bd
ulwa
hhab
)
4787
un
sta
ble syst
e
m
Figure
8. Step
respo
ns
e
of M
LS w
it
h ASFC
: case
1
Figure
9. Step
respo
ns
e
of M
LS w
it
h ASFC
: case
2
Figure
10. St
ep
r
es
pons
e
of M
LS w
it
h ASFC
: case
3
Figure
11.
Step
r
es
pons
e
of M
LS w
it
h ASFC
: case
4
5.
CONCL
US
I
O
N
In
this
pa
per
,
the
desig
n
of
a
n
AS
FC
f
or
a
MLS
ha
s
been
pro
posed
.
T
he
SFC
was
de
sign
by
firs
t
li
near
iz
ing
t
he
MLS
ab
ou
t
a
s
el
ect
ed
equ
il
ib
rium
po
int,
the
n
the
cl
os
e
d
lo
op
pole
s
are
posit
ion
e
d
at
lo
cat
ion
s
so
as
to
ac
hie
ve
ce
rtai
n
desi
gn
sp
e
ci
ficat
io
n.
H
ow
e
ve
r,
w
hen
the
ref
e
re
nce
in
put
cha
nged
,
the
no
nadapti
ve
sta
te
feedback
con
t
ro
ll
er
c
ou
ld
no
lo
nger
s
at
isfy
the
cl
ose
d
lo
op
desi
gn
sp
eci
ficat
io
ns
and
it
s
perfor
m
ance
degra
ded,
or
e
ven
it
fail
s
to
s
t
abili
ze
the
M
LS,
w
hile
the
AS
FC
sat
isfie
d
the
cl
os
ed
lo
op
desig
n
s
pecif
ic
a
ti
on
s
for
al
l
ref
e
ren
c
e
inputs.
Se
ve
r
al
con
cl
us
io
ns
can
be
draw
n
from
the
ob
ta
ined
res
ults.
Fi
rst,
the
li
nea
rizat
ion
desig
n
m
et
ho
d
has
a
li
m
itati
on
w
he
n
app
li
ed
t
o
high
ly
no
nlinea
r
s
yst
e
m
,
su
ch
as
the
MLS.
S
econd
,
the
A
SFC
is
a
su
it
able
s
olu
ti
on
to
sta
bili
ze
hi
gh
ly
no
nlinear
syst
e
m
,
an
d
it
outpe
rfor
m
s
the
nona
dap
ti
ve
sta
te
feedbac
k
c
on
t
r
oller.
REFERE
NCE
S
[1]
M.
H.
A.
Yase
en
and
H.
J.
Abd,
“
Modeli
ng
and
cont
ro
l
for
a
m
agne
tic
l
evi
t
at
ion
s
y
s
te
m
ba
sed
on
SIM
LAB
pla
tform i
n
re
al t
ime,
”
Re
sul
ts i
n
Phy
sics
,
vol
.
8
,
pp.
153
-
159
,
20
18.
[2]
M.
H.
A.
Yase
e
n
and
H.
J
.
Abd
,
“
A
new
pla
n
ar
el
e
ct
rom
agneti
c
le
vi
ta
t
ion
s
y
s
tem
improvem
ent
m
et
hod
base
d
o
n
SIM
LAB
pla
tfor
m
in
re
al
ti
m
e
o
per
ation,
”
Progr
ess In
E
l
ectrom
agnet
i
c
R
ese
arch
,
vol. 62, pp. 211
-
221,
2017
.
[3]
W
.
W
iboonj
aro
e
n
and
S.
Sujit
jo
rn,
“
Stabi
lizatio
n
of
a
m
agne
tic
le
vitati
on
cont
r
ol
s
y
stem
via
st
at
e
-
PI
fe
edback,
”
Inte
rnational
Jo
urnal
of
Ma
the
m
ati
cal Models an
d
Me
thods
in Ap
pli
ed
Scienc
e
,
vo
l.
7
,
no
.
7
,
2013
.
[4]
S.
Folea
,
et
a
l
.
,
“
The
ore
tica
l
anal
y
sis
and
exp
erim
ent
al
validatio
n
of
a
sim
pli
fie
d
fra
ct
ion
al
orde
r
cont
roller
for
a
m
agne
ti
c
l
evi
t
at
i
on
s
y
st
em,
”
I
EEE
Tr
ansacti
ons
on
Control
S
ystem
s Tec
hnology
,
vol.
24
,
no
.
2
,
pp
.
756
-
763
,
2016
.
Evaluation Warning : The document was created with Spire.PDF for Python.
IS
S
N
:
2088
-
8708
In
t J
Elec
&
C
om
p
En
g,
V
ol.
10
, No
.
5
,
Oct
ob
e
r 2
020
:
47
82
-
47
88
4788
[5]
A.
Duka,
e
t
a
l
.
,
“
IMC ba
sed
PID
cont
rol
of a
m
ag
net
i
c
l
evi
t
at
io
n
s
y
stem,
”
Proce
di
a
Technol
og
y
,
v
ol.
22
,
2016
.
[6]
P.
Bal
ko
and
D.
Rosinová,
“
Modeli
ng
of
m
agne
tic
le
vitat
ion
s
y
ste
m
,
”
in
21st
Inte
rnationa
l
Confe
r
enc
e
on
Proce
ss
Control
(
PC)
,
Strbske
Pleso,
Slovaki
a
,
2017
.
[7]
P.
Raj
a
la
kshm
y
,
et
al
.
,
“
Deve
lopment
of
m
agne
tic
le
v
it
a
ti
o
n
s
y
stem,
”
Inter
nati
onal
Journ
al
of
Innov
a
tiv
e
Technol
ogy
and
Ex
ploring
Engi
n
ee
ring
,
vol
.
8
,
n
o.
4
,
2019
.
[8]
R.
Morale
s,
et
a
l
.
,
“
Ad
apt
iv
e
co
ntrol
base
d
on
f
ast
onli
ne
al
g
eb
rai
c
ide
nt
ifi
c
at
io
n
and
GP
I
cont
rol
for
m
agne
t
i
c
le
vitati
on
s
y
st
e
m
s
with
ti
m
e
-
va
r
y
ing
input
ga
in
,
”
In
te
rnationa
l
Journal
of
Cont
rol
,
vol
.
87
,
no
.
8,
pp
.
1604
-
21
,
2014.
[9]
L.
Jin
-
hui,
e
t
al
.
,
“
A
pra
ct
ical
ro
bust
nonli
ne
ar
c
ontrol
ler
for
m
agl
ev
le
vi
ta
t
ion
sy
stem,
”
Journal
of
Cent
ral
South
Unive
rs
it
y
,
vo
l.
20,
no
.
11
,
2013
.
[10]
I.
Ahm
ad
and
M
.
A.
Java
id,
“
Nonlinear
m
odel
and
cont
roller
de
sign
for
m
agne
ti
c
le
vitat
ion
s
y
st
em,
”
in
ISPR
A'1
0
Proce
ed
ings o
f
t
he
9th
WSE
AS
in
te
rnational
con
f
ere
nce on
S
ignal proce
ss
ing, roboti
cs
and
autom
ati
on
2010
,
UK
.
[11]
S.
Yada
v,
e
t
a
l
.
,
“
Perform
anc
e
enh
anc
emen
t
of
m
agne
t
ic
l
evi
t
at
ion
s
y
st
e
m
using
te
a
ching
learni
ng
b
ase
d
opti
m
iz
ation,
”
A
le
xandria
Engi
n
ee
ring
Journal
,
vol.
57
,
no
.
4
,
pp
.
2427
-
33
,
2018
.
[12]
A.
Te
pljakov,
et
al
.
,
“
Design
of
ret
uning
fra
ction
al
PID
cont
roll
e
rs
for
a
cl
osed
-
l
oop
m
agne
ti
c
l
e
vit
ation
cont
ro
l
s
y
stem,
”
in
201
4
13th
Int
ernational
Confe
ren
c
e
on
Control
A
utomati
on
Robo
ti
cs
and
Vi
sion
-
2014
ICA
RCV
,
Singapore
,
2014
.
[13]
J.
Zong
and
L.
Li
,
“
Tun
ing
Fra
ct
ion
al
-
Order
PI
D
Control
lers
for
a
s
oli
d
-
Core
Magn
et
i
c
Be
ari
ng
S
y
s
te
m
,
”
I
EEE
Tr
ansacti
ons on Control Sy
st
ems Tec
hnology
,
vo
l. 23, no. 4, pp. 16
48
-
1656,
2015
.
[14]
H.
Gole,
et
al
.
,
“
Inve
stiga
ti
on
o
f
fra
c
ti
ona
l
con
t
rol
per
form
an
c
e
for
m
agne
tic
l
e
vit
ation
exp
eri
m
ent
a
l
set
-
up
,
”
i
n
2012
Inte
rnatio
nal
Confe
renc
e
on
Eme
rging
Tr
ends
in
Sci
enc
e,
Engi
ne
erin
g
and
Technol
ogy
(
INCOSET
)
,
Ti
ruc
h
ira
ppa
ll
i
,
India
,
2012
.
[15]
N.
Sadat
i
,
“
Opt
imum
design
of
fra
ctiona
l
ord
er
PID
for
MIM
O
and
SIS
O
S
y
stems
using
par
ticle
sw
arm
opti
m
iz
ation
tec
hnique
s,
”
in
Pro
ce
ed
ing
of
Int
ernati
onal
Con
fe
re
nce
on
Me
chat
ro
nic
s
,
Kum
amoto,
Japa
n
,
200
7
.
[16]
T.
Sali
m
and
V.
M.
Karsli,
“
Control
of
single
axi
s
m
agnetic
l
evi
t
at
ion
s
ystem
using
fuz
z
y
logi
c
cont
ro
l,
”
Inte
rnational
Jo
urnal
of Adv
an
c
ed
Computer
S
cienc
e
and
Appl
i
c
ati
ons
,
vo
l. 4, no
.
11
,
pp
.
83
-
88
,
2013.
[17]
A.
Suebsom
ran
,
“
Optimal
cont
ro
l
of
e
le
c
tromagn
et
i
c
sus
pension
EMS
sy
st
em,
”
T
he
open
Aut
oma
ti
on
and
Control
Syste
ms
Journal
,
vol. 6, no. 1, pp. 1
-
8,
2014.
[18]
S.
K.
Choudhary
,
“
Robust
fee
d
bac
k
cont
ro
l
an
aly
s
is
of
m
agne
ti
c
le
v
it
a
ti
on
s
y
stem,
”
WSEA
S
Tr
ansacti
ons
on
Syste
ms
,
vol
.
13
,
pp.
285
-
291,
20
14.
[19]
Z.
Yanhong
,
e
t
al
.
,
“
Robustness
of
improving
ac
t
ive
m
ag
le
v
m
otori
z
ed
spindle
equilibr
ium
positi
on,
”
TEL
KOMNIKA
Indone
sian
Jour
nal
of
Elec
tric
al
Engi
ne
ering
an
d
Computer
Sci
e
nce
,
vo
l.
11
,
no.
9,
pp.
5161
-
68
,
2013.
[20]
K.
Ishaque
,
e
t
al
.
,
“
Modeli
ng
and
control
of
m
agne
ti
c
s
y
st
e
m
via
fuz
z
y
lo
gic
con
troller,
”
in
2011
Fou
rt
h
Inte
rnational
Co
nfe
renc
e
on
Mod
el
ing
,
Simula
ti
o
n
and
App
li
ed
O
pti
mization
,
Kua
la
Lumpur,
Mal
a
y
sia
2011.
[21]
D.
S.
Shu
'
ai
bu
,
et
al
.
,
“
Eff
icie
nt
fuz
z
y
log
ic
cont
roller
for
m
agne
t
ic
le
vi
ta
t
io
n
s
y
stems
,
”
Nig
erian
Journal
o
f
Technol
ogi
cal
D
ev
e
lopment
,
vol
.
13,
no.
2,
pp.
50
-
57,
2016
.
[22]
J.
J.
Herná
nd
ez
-
Casaña
s
,
et
a
l
.
,
“
Chara
cteri
z
at
i
on
and
ad
apt
iv
e
fuz
z
y
m
odel
r
efe
ren
ce
cont
rol
for
a
m
agne
t
i
c
le
vitati
on
s
y
s
te
m
,
”
Alex
andria
En
gine
ering
Journ
al
,
vo
l. 55, no. 3, pp. 2597
-
2607,
2016.
[23]
E.
T
.
Moghadd
am
and
J.
Ganj
i,
“
Slidi
ng
m
ode
cont
rol
of
m
a
gnet
i
c
le
v
it
a
ti
on
s
y
stems
using
h
y
brid
ext
e
nded
Kalman
filter
,
”
Ene
rgy
S
cienc
e
and
Technol
og
y
,
vol. 2, no. 2, pp. 35
-
42,
2011.
[24]
B.
Krishna
and
D.
Chandra
n
,
“
Slidi
ng
m
ode
con
trol
for
m
agne
tic
l
evi
t
at
ion
s
y
st
em:
an
ei
ge
n
str
uct
ure
assignm
en
t
appr
oac
h
,
”
in
9t
h
Asian
Therm
al
Spray
Conf
ere
n
ce
,
2018.
[25]
S.
A.
Al
-
Sam
arr
ai
e
,
“
Vari
able
Struct
ure
Con
trol
Design
for
a
Magne
t
ic
L
evi
t
at
ion
S
y
s
tem
,
”
Journal
o
f
Engi
ne
ering
,
vol
.
24
,
no
.
12
,
pp
.
84
-
103,
2018
.
[26]
S.
A.
Al
-
Sam
arr
ai
e
,
et
al
.
,
“
Ada
pti
ve
slid
ing
m
ode
cont
ro
l
for
m
agne
t
ic
l
evi
t
at
io
n
s
y
stem,
”
Al
-
N
ahrain
Journal
f
or
Engi
ne
ering
Sc
i
enc
es
(
NJE
S)
,
v
ol.
21
,
no
.
2
,
pp
.
266
-
74,
2018
.
[27]
A.
Nos
hadi
,
e
t
al
.
,
“
A
Gene
t
ic
a
lgori
thm
-
ba
sed
s
y
stem
identifi
c
ation
of
a
ct
iv
e
m
agne
tic
bea
ring
s
y
st
e
m
:
a
fre
quen
c
y
-
do
m
ai
n
appr
oac
h
,
”
in
11th
IE
E
E
Inte
rnationa
l
Confe
renc
e
on
Control
&
Au
tomati
on
(
ICCA)
,
Ta
i
chung
,
Taiwa
n
2014,
.
[28]
C
.
-
A
.
Boja
n
-
Dragos,
et
al
.
,
“
Proportional
-
integ
ral
-
der
iva
t
ive
g
ai
n
-
sche
du
li
ng
cont
rol
of
a
m
agne
t
ic
l
evi
t
at
io
n
s
y
stem,
”
Int
ernati
onal Journal
o
f
Computers Comm
unic
ati
ons
&
Control
,
vo
l. 12, no. 5, pp. 599
-
6
11,
2017
.
[29]
C
.
-
A
Boja
n
-
Dra
gos,
et
al.,
“
Gai
n
-
sche
duli
ng
co
ntrol
solut
ions
for
m
ag
net
i
c
l
evitat
ion
s
y
st
ems
,
”
Ac
ta
Pol
y
techni
ca
Hungar
ic
a
,
vol
.
15,
no
.
5
,
pp
.
89
-
108,
2018
.
[30]
Z.
Zh
ang
and
X.
Li
,
“
Rea
l
-
tim
e
ada
pti
v
e
co
ntrol
of
a
m
agne
tic
le
v
it
a
ti
on
s
y
stem
with
a
l
a
rge
ran
g
e
of
lo
ad
disturba
nc
e,
”
Se
nsor
s
,
vol.
18
,
n
o.
5
,
2018
.
[31]
H.
K Khalil,
N
o
nli
near
S
yste
ms
,
3rd
ed,
US
A:
Prentice
H
al
l
:
200
2.
[32]
H.
J.
Marqu
ez,
Nonli
near
Contr
ol
systems
.
New
Jerse
y
:
John W
ile
y
&
Sons
,
2003
.
Evaluation Warning : The document was created with Spire.PDF for Python.