Internati
o
nal
Journal of P
o
wer Elect
roni
cs an
d
Drive
S
y
ste
m
(I
JPE
D
S)
V
o
l.
6, N
o
. 3
,
Sep
t
em
b
e
r
2015
, pp
. 50
9
~
51
5
I
S
SN
: 208
8-8
6
9
4
5
09
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
On the P
erf
ormances
Investi
g
ation of Diff
erent Su
rf
ace
Mounted Perm
anent M
a
gnet
M
a
chines
Mansouri
Ali, Ms
adde
k
Hejra, Tr
abelsi H
a
fe
dh
Laborator
y
of
C
o
mputer,
Electro
n
ics & Smart En
gineer
ing S
y
s
t
ems Design, Engineering
Sc
hool o
f
Sfax, Sfax, Tu
nisia
Article Info
A
B
STRAC
T
Article histo
r
y:
Received
Mar 24, 2015
Rev
i
sed
Ju
l 31
,
20
15
Accepted Aug 11, 2015
In recent
ye
ars
,
perm
anent m
a
gnet m
achines
hav
e
becom
e
a com
m
on choice
in man
y
industrial app
lications. Ther
ef
ore, several structures
have been
develop
e
d, and the choice of a topolog
y
design
ed for a specified
application
requires th
e kn
owledge of th
e advantag
es an
d disadvantages
of differ
e
nt
topologies
. The
present work deals with
th
e ev
aluation of
the p
e
rformances
of differ
e
nt r
a
dial flux surf
ace-
m
ount
ed permanent magnet motors designed
for an e
l
e
c
tri
c
v
e
hic
l
e m
o
tor
ap
plic
ation
.
Th
e o
b
jec
tive
of this
surve
y
is to
s
how the effect
of the rotor p
o
s
ition (inner o
r
outer) and th
e magnets
segmentation on
the mach
ine ou
tput torq
u
e
and
iron losses. In this contex
t,
four m
achines with: (i) inn
e
r r
o
tor, (i
i) inner r
o
tor segm
ented
m
a
gnets, (ii
i
)
outer rotor and (
i
v) outer rotor s
e
gm
ented magn
ets have been d
e
signed and
s
t
udied.
All
the
s
e m
achines
h
a
ve th
e same g
e
ometrical dimensions and
current
load
ing.
The m
a
in
ide
a
is to dev
e
lop
a m
achin
e with
sm
oothness
torque, lower
tor
que ondulation
,
lower ir
on losses, and which
is mechanically
robust. Firstly
,
the output torq
ue of
the differ
e
nt structur
e is computed.
Secondly
,
b
y
means of an improved
analy
t
ical model cou
p
led with
2
dim
e
ns
ional tr
an
s
i
ent fini
te
elem
ent an
al
ysis (FE
A
), the m
ach
ine
s
iron losses
are pr
edi
c
ted
.
Keyword:
Finite elem
ent
analysis
Iron losses
Mag
n
e
ts seg
m
en
tatio
n
Out
put
t
o
r
q
ue
Perm
an
en
t m
a
g
n
e
t m
o
to
r
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
:
Man
s
ou
ri
Ali,
Depa
rt
em
ent
of C
o
m
put
er
En
gi
nee
r
i
n
g,
High
er In
stitu
te of
App
lied
Scien
ce an
d Tech
no
log
y
o
f
Gafsa (ISSAT
Gafsa),
C
a
m
pus Uni
v
e
r
si
t
a
i
r
e
- Si
di
Ahm
e
d
Zar
r
o
u
k
- 21
1
2
G
a
fsa
,
T
uni
si
a.
Em
a
il: man
s
ouriali2
00
2@yaho
o.fr
1.
INTRODUCTION
D
u
e to
t
h
eir
hig
h
e
r
p
e
r
f
o
r
m
a
n
ces, synchr
on
ou
s
perm
anent m
a
gnet
m
a
chines a
r
e bec
o
m
e
a very
attractiv
e so
lu
t
i
o
n
in
sev
e
ral ap
p
lication
s
such
as elect
ri
c and
hy
b
r
i
d
vehi
cl
e t
r
act
i
on [
1
]
,
[2]
an
d
wi
n
d
po
we
r
g
e
n
e
ration
[3
]. In
t
h
e literatu
re,
n
u
m
ero
u
s p
e
rm
an
en
t m
a
g
n
e
t m
ach
in
es were
p
r
o
p
o
s
ed
[4
]-[5
]
. From
th
ese
machines, t
h
e
m
o
st recognized a
r
e
those
with ra
dial flux
and s
u
rface
mounted pe
rm
anent m
a
gnets [6]. The
rad
i
al
flux
co
nfigu
r
ation
is the first
p
e
rm
an
en
t m
a
g
n
e
t m
a
c
h
in
e em
erg
e
d
i
n
t
h
e indu
stry.
Th
is m
ach
in
e i
s
on
e
of the m
o
st classic, easiest t
o
c
o
nstruct
and the
m
o
st
commo
n
l
y u
s
ed to
po
log
y
[
6
]. Dep
e
nd
ing on
t
h
e ro
tor
arran
g
e
m
e
n
t
, two
con
f
i
g
uratio
n
s
m
a
y b
e
enco
un
tered
:
inner ro
tor m
ach
in
es an
d tho
s
e
in
wh
ich
t
h
e
ro
tor is
m
ount
ed o
n
t
h
e o
u
t
s
i
d
e.
W
h
en
desi
g
n
i
n
g
a perm
anent
m
a
gnet
m
a
chi
n
e, se
ver
a
l
param
e
t
e
rs sh
oul
d be
con
s
i
d
ere
d
,
ess
e
nt
i
a
l
l
y
t
h
e coggi
ng t
o
r
q
ue, t
h
e el
ect
rom
a
gnet
i
c
t
o
r
que
ri
ppl
e a
n
d
t
h
e i
r
on l
o
sses
.
Ge
n
e
ral
l
y
,
th
ese p
a
ram
e
t
e
rs d
e
p
e
n
d
o
n
th
e ro
t
o
r
d
i
sp
o
s
ition
and
t
h
e m
a
g
n
e
ts shap
e. In
o
r
d
e
r
to
m
e
e
t
th
e req
u
i
red
perform
a
nces,
it is necessary t
o
st
udy t
h
ese
param
e
ters
In
p
e
rm
an
en
t mag
n
e
t m
ach
i
n
es, a sev
e
re prob
lem
g
e
n
e
rally en
co
un
tered, is th
e iro
n
losses, wh
ich
form
a
larg
er p
o
rtion
o
f
th
e
to
tal lo
sses th
an
in
in
du
ction
mach
in
es. Th
i
s
is, essen
tiall
y, d
u
e
to
th
e ro
tati
ng
mag
n
e
tic field. It
was shown
,
t
h
at th
e iron
losses
u
n
d
e
r ro
tatin
g
m
a
gn
etic field is
g
e
n
e
rally sign
ifican
tly
larger t
h
an t
h
at from
a
lternating
ones. T
h
ere
f
ore, it is
im
perative to eval
uate these losse
s accurately and take
th
em
in
to
con
s
id
eration
d
u
ring
th
e m
o
to
r
d
e
sig
n
pro
c
ess.
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
. 3, Sep
t
em
b
e
r
2
015
:
50
9 – 515
51
0
The prese
n
t work
deals wi
th th
e design
of a pe
rm
anent
m
a
gnet m
o
tor for a
n
electric vehicle
appl
i
cat
i
o
n. T
h
e m
a
i
n
p
u
r
p
o
s
e, i
s
t
o
ha
ve
a desi
gn
wi
t
h
sm
oot
hness
t
o
rq
ue,
l
o
wer t
o
rq
ue
o
n
d
u
l
a
t
i
o
n a
n
d
m
echanically robust. T
o
m
eet th
is goal,
the
perform
a
nces of
fo
ur s
u
rface m
ounte
d
pe
rm
anent m
a
gnet
topologies
are investigate
d
. In
each
ca
se, the electrom
a
gnetic torque and
the iron los
s
e
s
are com
pute
d
. T
h
e
co
m
p
u
t
atio
n
of th
e iron
lo
sses is b
a
sed
on
th
e tran
sien
t fin
ite elem
en
t an
alysis co
up
led
with
ad
eq
u
a
te
analytical m
odel.
2.
IRON LOSSE
S MODELLING
In
perm
anent
m
a
gnet
m
achines, t
h
e m
a
gn
et
i
c
fi
el
d i
s
ge
neral
l
y
di
st
o
r
t
e
d an
d
n
onsi
n
usoi
dal
.
As a
resu
lt th
e estimatio
n
o
f
th
e i
r
on
lo
sses in
these
m
ach
in
es is a v
e
ry co
m
p
licated
task
. One o
f
th
e
first iro
n
l
o
ss
m
o
d
e
ls is th
e Stein
m
etz fo
rm
u
l
a [7
], in
wh
ich
t
h
e time av
erag
e
p
o
wer lo
ss p
e
r
u
n
it
vo
lu
m
e
P
v
(t) v
is
exp
r
esse
d by
:
β
α
v
B
ˆ
f
.
K
=
)
t
(
P
(
1
)
Whe
r
e f,
B
ˆ
, K and
α
are res
p
e
c
tively the freque
ncy, the pe
ak
val
u
e
of m
a
gnet
i
c
fl
u
x
an
d t
h
e l
o
s
s
coefficients
.
It
was
sh
o
w
n
,
t
h
at
un
de
r
si
nus
oi
dal
m
a
gnet
i
c
fi
el
d, t
h
i
s
m
odel
gi
ves acc
urat
e
resul
t
s
[7]
.
Neve
rt
hel
e
ss
, t
h
e m
a
gnet
i
c
fiel
d i
n
perm
anent
m
a
gnet
m
achi
n
e i
s
hi
g
h
l
y
di
st
ort
e
d an
d n
onsi
n
usoi
dal
.
Thi
s
i
s
way th
is m
o
d
e
l was im
p
r
ov
ed
t
o
m
eet th
is n
on-sinu
so
id
al
v
a
riatio
n
.
As
a resu
lt a m
o
d
i
fied
Stein
m
etz
m
o
d
e
l
was p
r
op
ose
d
i
n
[
8
,
9]
. In t
h
i
s
m
odel
,
t
h
e t
i
m
e average
po
wer l
o
ss pe
r u
n
i
t
vol
um
e,
)
t
(
P
v
are expresse
d a
s
fo
llows:
r
β
α
eq
v
f
B
ˆ
f
.
K
=
)
t
(
P
1
(
2
)
f
r
i
s
t
h
e
rem
a
gnet
i
zat
i
o
n f
r
e
que
ncy
a
n
d f
eq
i
s
an
eq
ui
va
l
e
nt
f
r
eq
ue
ncy
de
ri
ve
d
fr
om
t
h
e
rat
e
o
f
chan
ge
o
f
t
h
e
f
l
ux
de
nsi
t
y
.
In sp
ite
of th
i
s
im
p
r
o
v
e
m
e
n
t
, it was sh
own
i
n
[8
,
9
]
th
at th
e
u
s
e
of this m
o
d
e
l in
the iron lo
ss
estim
a
tion in
perm
anent m
a
gnet m
ach
ines
causes
inacc
urate res
u
lts. T
o
ove
rc
om
e t
h
i
s
i
n
co
n
v
eni
e
n
t
, ot
he
r
im
pro
v
em
ent
were i
n
t
r
o
d
u
ce
d i
n
t
h
e m
odel
pr
op
ose
d
by
[
9
]
.
I
n
t
h
at
m
odel
,
t
h
e i
r
o
n
l
o
ss per
u
n
i
t
vol
um
e is
exp
r
esse
d as
a
fu
nct
i
o
n
of t
h
e
fl
u
x
densi
t
y
B
(
t
)
an
d i
t
s
rat
e
o
f
c
h
an
ge
(dB
/
d
t
):
dt
B
dt
dB
.
K
T
=
P
α
β
α
T
v
∫
0
1
1
(
3
)
Whe
r
e K
1
is c
o
efficient
de
duc
ed
from
K,
α
an
d
β
.
Desp
ite all th
e
i
m
p
r
ov
em
en
ts mad
e
on
th
e
Stein
m
e
t
z
m
o
d
e
l,
it r
e
m
a
in
s li
mited
to
th
e sin
u
so
id
al
non
d
i
sto
r
ted
m
a
g
n
e
tic f
i
eld
s
. I
n
t
h
e sam
e
co
n
t
ex
t, Ber
t
o
tti in
1
988
was
f
ound
th
at th
e to
tal
ir
o
n
l
o
sses P
t
can
be
separated into
three term
s
namely: hysteresis losses P
h
, classical eddy current losses P
c
and excess losses P
e
[1
0]
w
h
ich
can
be e
x
pres
sed
a
s
f
o
llows:
e
c
h
t
P
+
P
+
P
=
P
(
4
)
f
B
ˆ
C
=
P
h
2
0
(
5
)
()
2
2
2
6
2
f
B
ˆ
)
d
(
σ
π
=
P
c
(
6
)
2
3
max
1
e
f
B
C
P
(
7
)
Whe
r
e C
0
, 2
d
,
σ
and
C
1
denotes respect
ively: the hysteresis
loss c
o
efficient, t
h
e
steel sheet
thickne
ss, t
h
e e
l
ectrical conductivity
and t
h
e
excess l
o
ss c
o
e
fficient.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
J
PEDS
I
S
SN:
208
8-8
6
9
4
On the
Performances Investi
gation of Different Surf
a
ce
M
o
unted
Permanent
M
a
gnet…
(Mansouri Ali)
51
1
The Bert
otti m
odel was then applied i
n
num
erous su
rveys. Based
on such m
odel, an analytical
expressi
on of t
h
e three iron l
o
ss
com
p
o
n
ents
in the
stator
teeth an
d y
oke
wa
s de
velo
pe
d in
[1
1]
.
In
order t
o
improve the estimati
on of the i
r
on losses in rotating
m
achines, a m
odel considering the
effect of
the m
i
nor hy
stere
s
is
loo
p
s was
de
velo
ped
in
[1
2]
. T
h
e
o
b
t
ained a
n
aly
tical results
give
g
o
o
d
agreem
ents with the
m
easured ones.
The
use
of all t
h
e fl
ux
den
s
ity
harm
onics in t
h
e ir
o
n
losses c
o
m
putation
was p
r
o
p
o
se
d in [
13]
. B
u
t it h
a
s sho
w
n that this ap
pr
oac
h
is not en
o
u
g
h
ac
curate.
An
othe
r ap
pr
oac
h
usi
ng the
two o
r
th
o
g
o
n
al
com
ponents
o
f
the flu
x
de
ns
ity
were app
lie
d to estim
a
te the iro
n
losse
s in [
14]
. T
h
e res
earch
to im
prove t
h
e iron losses
esti
m
a
t
i
on in
electrical
m
a
c
h
ines continues to attract
th
e
attention of
several
auth
ors
.
I
n
o
u
r
wo
rk
, we e
v
aluate the m
achine iro
n
losse
s by
usin
g the F
E
A an
d a thr
e
e
com
pone
nts m
odel.
In this case, the instantaneous
iron l
o
sses are calculated
usi
n
g th
e f
o
llowi
n
g
fo
rm
ulations [1
5]
:
)
t
(
p
+
)
t
(
p
+
)
t
(
p
=
)
t
(
p
e
c
h
v
(
8
)
2
2
y
y
2
x
x
h
dt
dB
H
dt
dB
H
)
t
(
p
(
9
)
2
y
2
x
c
c
dt
dB
dt
dB
K
)
t
(
p
(
1
0
)
4
3
2
y
2
x
e
e
dt
dB
dt
dB
K
)
t
(
p
(
1
1
)
β
, k
c
and
k
e
are respectivel
y the Steinm
e
t
z, the cla
ssic
a
l eddy c
u
rre
n
t and t
h
e e
x
cess losses
coefficients.
B
x
, B
y
, H
x
, H
y
are the two ort
h
o
g
onal co
m
ponents o
f
the flu
x
de
nsity
and the m
a
gn
etic intensity
bot
h in the
stator yoke a
n
d teeth are calculat
e
d in eac
h
fi
nite ele
m
ent from
the 2-D
transi
ent FEA sim
u
l
a
tions.
The com
putation
of the t
o
tal iron lo
sses
pe
r
lengt
h u
n
it f
o
r pe
rio
d
ic va
ri
ab
le m
a
gnetic fields is de
scri
bed i
n
the following fo
ur ste
p
s
[16]:
1.
Meshing the machine
stator into N
e
finite element
m
e
sh
2.
Devisi
ng
the ti
m
e
dom
ain into
N ste
p
s
den
o
t
ed t
j
,
3.
Inte
gratin
g t
h
e
pre
v
io
us e
q
uations
o
v
e
r
one
p
e
rio
d
,
4.
Summing the equations obtained in step
3 over the total num
ber of elem
ents
1
0
2
2
,
,
2
,
,
1
1
)
(
N
j
j
t
t
e
t
ey
t
ey
t
x
e
t
ex
h
j
j
j
j
j
j
dt
A
dt
dB
H
dt
dB
H
T
e
P
(
1
2
)
dt
A
dt
dB
dt
dB
T
K
e
P
e
N
j
j
t
t
t
ey
t
ex
c
c
j
j
j
j
1
0
2
,
2
,
1
1
(
1
3
)
1
0
4
3
2
,
2
,
1
1
)
(
N
j
j
t
t
e
t
ey
t
ex
e
e
j
j
j
j
dt
A
dt
dB
dt
dB
T
K
e
P
(
1
4
)
Here B
ex,tj
, B
e
y
,tj
, an
d H
ex,tj
, H
ey
,
t
j
denote th
e
two
ort
h
o
g
ona
l com
pone
nts
of t
h
e fl
ux
de
nsity
B
and
intensity H in
the finite el
em
ent e at the tim
e
step t
j
. A
e
a
n
d p a
r
e re
spect
ively the finite elm
e
nt area and t
h
e
m
achine
pair pole
num
b
er.
Three c
o
efficients (
β
, k
c
an
d k
e
) used in the previ
o
us iron loss m
odel have to be identified. These
coefficients a
r
e
de
rive
d from
the inform
ati
on su
pplie
d
by
th
e lam
i
nation m
a
nu
factu
r
ers
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
S
SN:
2
088
-86
94
I
J
PEDS
Vo
l.
6,
No
. 3,
Sep
t
em
b
e
r
2
015
:
50
9 – 515
51
2
3.
STUD
IED T
O
POLOGIES
The
stu
d
ied t
o
pol
ogies
are
a
ll of
3
-
p
h
ase
,
4-
p
o
le, a
n
d
48
-slots.
T
h
e
dif
f
ere
n
t c
o
n
f
ig
u
r
ations
are
sho
w
n in Fi
gu
r
e
1. T
h
e stator
is
m
a
de up o
f
a lam
i
nated iron co
re (
(
M
A
80
0-
6
5
A
)
an
d a three
pha
se arm
a
tur
e
win
d
in
gs
fed
b
y
three
ph
ase s
i
nus
oidal c
u
r
r
e
n
t. T
h
e
rot
o
r is
m
a
de up
of
ir
on
co
re
on
w
h
ich are
m
ounte
d
, i
n
opposite direct
ion the m
a
gnet
s
whic
h are
NdFeB rare
-ear
t
h
. The m
a
jor m
a
chines
param
e
ters and feat
ure
s
are
respectively
pr
ovi
ded
in Tabl
es
1
an
d 2.
Figu
re
1.
Stu
d
i
e
d to
p
o
lo
gies;
(a): in
ne
r r
o
to
r,
(
b
):
in
ne
r r
o
to
r se
gm
ented m
a
gnet,
(c
): o
u
te
r r
o
to
r se
gm
ented
m
a
gnet,
(d
):
ou
ter r
o
to
r se
gm
ented m
a
gnet
Table
1. C
h
ara
c
teristics of t
h
e
stu
d
ied t
o
p
o
lo
gy
Characteristic
Value
No
m
i
nal output po
wer
9.
42 kw
No
m
i
nal speed
1500 r
p
m
M
a
xim
u
m
speed
4500 r
p
m
No
m
i
nal tor
que
60 N.m
Table
2. M
a
c
h
i
n
e’s
dim
e
nsion
s
Variable
Value Variable
Value
outer
stator
diam
eter
170
m
m
shaft diam
eter
50
m
m
inner
stator
diam
eter
91.
8
m
m
r
o
tor
cor
e
dia
m
eter
85
m
m
active
m
a
chine length
160
m
m
stator
tooth width
4.
8
m
m
air
g
ap length
1
m
m
slot opening an
gle
30º
half pole angle
60 º
r
o
tor
y
oke height
17.
5
m
m
stator
slot pitch
11.
9
m
m
m
a
gnet thickness
2.
4
m
m
4.
R
E
SU
LTS AN
D ANA
LY
SIS
Owi
ng t
o
the symmetries and the peri
odicities in the
studied structures
,
our st
udies are l
i
m
i
ted to the
fourt
h
of the m
achines. In
o
u
r
su
rvey
we
use
a dy
nam
i
c
m
o
deling;
so
we
need
to e
n
s
u
re
the co
ntin
uity
of
the
FEA m
odel w
h
en t
h
e r
o
to
r
m
oves. Fo
r thi
s
reas
on
, we
h
a
ve u
s
ed a
n
in
terp
olation tec
hni
que c
o
nsisting
o
f
equalizing the
potential vector A in th
e air
g
ap. T
h
is techni
que
req
u
ires t
h
e divisi
on
of
the airga
p
into
two
parts: the stator
half an
d
the rotor one.
Then we de
fi
ne equality constr
aints between coincident
nodes.
4.1. Electr
o
m
a
gnetic T
o
rque
In
order to show the i
n
fl
uence of t
h
e rot
o
r
disposition and th
e m
a
gnet segm
entation on the
output
to
rq
u
e
of
th
e stu
d
i
ed
m
ach
in
es o
n
th
e ou
tpu
t
to
r
q
u
e
o
f
th
e stu
d
i
ed
topo
logy,
sev
e
r
a
l FEA in
v
e
stig
ation
s
wer
e
carried out.
The obtained resu
lts are illustrat
e
d in Figure
2.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
J
PEDS
I
S
SN:
208
8-8
6
9
4
On the
Performances Investi
gation of Different Surf
a
ce
M
o
unted
Permanent
M
a
gnet…
(Mansouri Ali)
51
3
Figu
re
2.
O
u
tp
ut to
rq
ue
versus electrical angle
Based
on the results related to the elect
rom
a
gne
tic torque
variation of the different studied
topologies, we can
deduce
th
e
toque peak val
u
e
and ondu
lation corresponding t
o
each
t
opology. This result is
illustrated in Table 3.
Table
3. T
o
rq
u
e
o
n
d
u
latio
n a
n
d
pea
k
value
T
opology
T
o
r
que ondulation
(
%
)
Peak value
I
nner
r
o
tor
34
39,
552
In
n
e
r ro
to
r seg
m
en
ted
m
a
gnet
24
33,
92
Outer
r
o
tor
33
55,
808
Ou
ter ro
to
r seg
m
e
n
ted
m
a
gnet
14
54,
784
Acco
r
d
in
g
to
the pre
v
io
us res
u
lts, we
ca
n de
duce:
The outer
rot
o
r
m
achine has a
higher
electromagnetic torque than the inne
r rot
o
r
one. This is essentially
d
u
e
to
th
e
h
i
gher
airg
ap d
i
am
eter
in
th
e
o
u
t
er
ro
tor
to
po
logy.
Th
e ou
tpu
t
to
rqu
e
r
each
e
s
5
5
.
8
Nm
f
o
r
the
oute
r
r
o
to
r m
achine t
h
at is a
b
out
1.
5 tim
es
greater than for
the
inner rotor
machine.
The n
o
n
-
se
gm
ented m
a
gnets m
achine has th
e highest to
rq
u
e
on
dulatio
n. T
h
is on
d
u
lation
is caused by
th
e
so called cogging toque.
In t
h
is cont
ext, it
was shown in
[17]
that the m
a
gnets segm
entat
i
on is
one
of the
m
o
st used techniques to c
o
nsider
a
b
ly
red
u
ce
the co
ggi
ng t
o
rq
ue in
perm
anent m
a
gnet
m
a
chines. For the
non-segm
ented
m
a
gnets m
a
c
h
ine, t
h
e torque ondulation e
x
ceeds 33% in bot
h cases
of
oute
r
and inne
r
rot
o
r m
achines.
R
e
ferri
ng t
o
fi
gu
re 2
,
an
d co
m
p
ared to the results o
b
ta
ine
d
in [
16]
,
we can de
d
u
ce that inverting the rotor
position leads t
o
doubling th
e
machine out
p
ut toque.
4.
2. I
r
on
L
o
s
s
e
s C
o
mp
ut
ati
o
n
The iron losse
s in the stator
teeth and y
oke
fo
r the differe
n
t studie
d
pe
rm
anent m
a
gnet
m
achines
were investigat
ed by m
eans of the FE
A and
the analyti
cal
m
odel describe
d in sectio
n 2.
The obtained results
for
no load and load cond
itions are
respectively illustra
ted in Tables
4 and
5.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
S
SN:
2
088
-86
94
I
J
PEDS
Vo
l.
6,
No
. 3,
Sep
t
em
b
e
r
2
015
:
50
9 – 515
51
4
Table
4.
No load iron losses
L
o
ss Co
m
ponent
I
nner
r
o
tor
I
nner
r
o
tor
seg
m
ented
m
a
gnets
Outer
r
o
tor
Outer
r
o
tor
seg
m
ented
m
a
gnets
E
ddy
cur
r
ent (W
)
32
29
43.
6
45
Hy
steresis (W
)
6
5.
5
6.
2
7.
4
E
x
cess (W
)
45
35
85
80
T
o
tal (W
)
83
69,
5
134,
8
132,
4
Table
5. L
o
a
d
i
r
o
n
l
o
sses
L
o
ss Co
m
ponent
I
nner
r
o
tor
I
nner
r
o
tor
seg
m
ented
m
a
gnets
Outer
r
o
tor
Outer
r
o
tor
seg
m
ented
m
a
gnets
E
ddy
cur
r
ent (W
)
55
50
50
60
Hy
steresis (W
)
10
9
7.
5
9.
4
E
x
cess (W
)
194
147
396
378
T
o
tal (W
)
259
206
453,
5
447,
4
As it can be noticed is Tables 4 and 5, the outer
rot
o
r m
achine with
non-se
gm
ented
m
a
gnets has the
highest iron losses,
whereas the in
ne
r r
o
to
r
m
achine wit
h
m
a
gnets se
gm
ented
has
the l
o
west
ones
.
5.
CO
NCL
USI
O
N
This pa
per is
devoted t
o
the
perform
a
nces ev
aluation of four surface m
o
unted
perm
anent m
a
gnet
m
achines. This evaluation is carried
out by
com
puting the
iro
n
losses
and the electro
m
a
gnetic torque. Based
on
the
o
b
taine
d
results,
it can
be
af
firm
ed th
at, o
n
o
n
e
han
d
the
outer
r
o
t
o
r
to
pol
ogy
ha
s the
sm
oother
and
the
hig
h
er
o
u
tp
ut t
o
r
q
ue, in
fact
f
o
r t
h
e in
ne
r r
o
t
o
r m
achine t
h
e
tor
que
is a
b
o
u
t
1.
5 tim
es. On
the
other
ha
n
d
,
the
oute
r
rotor m
achine
has also t
h
e
highest iron losses.
The inner rotor configuration allo
ws a very
com
p
act construction, since
the rot
o
r can
be of sm
all
vol
um
e. Ho
we
ver
,
this
res
u
lts in a l
o
w
to
r
que
at sm
all r
a
dius
o
f
the
air ga
p.
C
o
ncer
nin
g
the
o
u
ter
rot
o
r
top
o
lo
gy
, the a
i
r ga
p ra
dius m
u
st be
greate
r
than a ce
rt
ain
m
i
nim
u
m
value, so t
h
at this con
f
ig
uratio
n
pr
od
uce
s
a hig
h
er t
o
r
q
u
e
com
p
ared to
the inne
r r
o
to
r
one
. As
far a
s
the co
ggi
ng t
o
rq
ue is a seve
r
e
pr
o
b
lem
degradi
n
g
the m
achine pe
rform
a
nces, it
needs to be
m
i
nim
i
zed. A wa
y
to do t
h
is is t
h
e segm
en
tation
of
the m
a
gne
ts.
Based on t
h
e obtained results, an
exterior-segm
e
nted perm
anent m
a
gnets topology will
be selected
in order to br
optim
ized an
d integrate
d
in
an in-wheel vehicle app
lication. This choi
ce is j
u
stified
by the
sm
ooth o
u
t
put
tor
q
ue
of t
h
is
m
achine.
ACKNOWLE
DGE
M
ENTS
This
work is supported
by the
Higher Institute of
Applied
Scie
nces and Technology
of Gafsa.
REFERE
NC
ES
[1]
Z. Q
.
Zhu
and
D. Howe,
“
E
le
ctri
cal
m
achin
es
an
d drives
fo
r e
l
e
c
t
r
ic,
h
y
brid
and
f
u
el
cel
l v
e
hic
l
es
”,
in
Proceeding
s
of th
e I
EEE
, vo
l. 95, no. 4, pp. 74
6–765, 2007
.
[2]
J. Wang, K. Atallah
,
Z.
Q. Zhu
,
and D. Howe,
“Modular 3-phase pe
r-manent
magnet brushles
s machines for in-
wheel appli
c
a
tio
ns
”,
Vehicular T
echnolog
y, IEEE Transactions,
v
o
l. 57
, no
. 5
,
pp
.
2714–2720, 200
8.
[3]
T.D. Strous
, “Design of a permanen
t m
a
gne
t
radi
al f
l
ux
co
ncentr
ated
co
il
generator for
a range ex
tender
application”,
Ma
s
t
er
thes
is
, Delf
t University
of
Technolog
y
,
2010
.
[4]
Y. Li, M. Jian
,
Y. Qiang, and
L. Jiang
y
u
,
"A N
ovel
Dir
ect Torq
ue Control Perm
anent Magnet Sy
nchronous
Motor
Drive us
ed in
El
ectr
i
ca
l Vehi
cl
e",
International
Journal of Power
El
ectronics and
Drive Systems (IJPEDS)
,
vol. 1
,
no. 2
,
pp
.
129-1
38, 2011
.
[5]
K. George, K.S. Shino
y
,
and S.
Gopinathan
, "Finite
El
em
ent M
odeling
of F
i
ve
P
h
as
e P
e
rm
anen
t M
a
gne
t BLDC
Motor for High
Power Density Application",
I
n
ternational Jo
urnal of Pow
e
r Electronics an
d Drive S
y
stem
s
(
I
JPEDS)
,
vol. 3
,
no. 4, pp.
384
-390
, 2013
.
[6]
F. Liber
t
, “
D
esi
gn, Optim
iza
tio
n a
nd Comparison of Permanent Magnet Motor
s
for a Low-Speed Direct-Driven
Mixe
r”
,
PhD th
e
s
is
, Ro
y
a
l Institute of Technolo
g
y
, Department
of El
ectrical En
gineer
ing, El
ectr
ical Machines and
Power Electron
ics, Stockho
lm 2
004.
[7]
K. Venkatach
al
am
, C. S
u
llivan,
T. Abdallah an
d H. Tacca
, “
A
ccura
te predi
c
ti
on of ferrite cor
e
loss with non-
sinusoidal wavef
o
rms u
s
ing
onl
y S
t
einm
etz para
m
e
ters
”,
in
IEEE Workshop on
Compute
rs in Power Electronics
,
pp. 36-41
,
3
-4 J
une 2002.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
J
PEDS
I
S
SN:
208
8-8
6
9
4
On the
Performances Investi
gation of Different Surf
a
ce
M
o
unted
Permanent
M
a
gnet…
(Mansouri Ali)
51
5
[8]
A. Krings, and J. Soulard, “Over
v
iew and com
p
a
r
ison of iron loss m
odels for ele
c
tri
cal m
ach
ines
”, in
ecolog
i
ca
l
vehi
c
les
and
r
e
n
e
wable
enr
e
gi
es
confer
en
ce
,
Monaco, March
,
25
-28, 2010
.
[9]
J. Li, T. Abdallah, and C
.
R. Sullivan, “Improved
calcu
lation of
co
re loss with non-
sinusoidal wavef
o
rms”, in
IEEE
Industry Applica
tions
Society An
nual
Meeting
,
v
o
l. 4
,
pp
. 2203-2
210, 2001
.
[10]
G. Berto
tti, “General properties
of power
losse
s in soft fe
rroma
gne
t
ic
ma
te
ria
l
s”,
I
EEE Transactio
ns on Magneti
cs
,
vol. 24
, no
.1, pp. 621–630, 1988.
[11]
W. Jiabin
, T. Ib
rahim, and
D.
Howe, “Prediction and measurement of ir
on
los
s
in a short-stro
ke, sing
le-ph
a
se,
tubular
permanent
magnet ma-ch
i
ne”,
I
EEE Transactions on Mag
n
etics
, vol. 46
, n
o
. 6
,
pp
. 1315–1
318, 2010
.
[12]
E. Barbisio
, F. Fiorillo ,
and C. Ragus
a, “Predict
i
ng loss in
m
a
gnetic steels
under arbitrar
y
indu
ctio
n waveform
and
with minor h
y
steresis loops”,
IEEE Transactions
on Magnetics
, v
o
l. 40
, no
.4, pp.
1810–1819, 200
4.
[13]
Slemon, G.R. an
d Liu, X, “Core lo
sses in perm
anent m
a
gnet m
o
to
rs”,
IEEE Transactions on Magnetic
., vol. 26
, no.
5, pp
. 1653
-165
5, 1990
.
[14]
C. Mi, G.R. Sle
m
on, and R. Bonert, “
M
odell
ing
of ir
on losses o
f
surface m
ounte
d
perm
anent m
a
gnet s
y
n
c
hronou
s
m
o
tors”, in
T
h
ir
ty Si
x I
A
S Ann
ual
Mee
ting
Confer
ence
r
ecor
d
o
f
I
EEE
,
vol. 4
,
pp
.
2585-2591, 200
1,
[15]
A. Mansouri,
an
d H. Tr
abelsi, “
O
n the Performances Invest
igatio
n and Iron
Losses Computation o
f
an Inset Surface
Mounted Permanent Magnet Mo
tor”,
in
2012 9
th
Signals Systems
Decision and In
fo
rmation Techn
o
logy con
f
eren
ce
,
2012.
[16]
A. M
a
ns
ouri, an
d H. Trabels
i
, “
E
ffec
t
of the Num
b
er
Magnet-Segments on the Output Torque an
d the Iron Losses
of a
SMPM”
,
in
2013 10
th
Signa
ls Systems Decis
i
on and In
fo
rmation Techno
logy
conferen
ce
, 18-2
1
Mars, 2013
,
[17]
A. Ghasem
i, “
C
ogging Torque
Reduction
and
Optim
izatio
n in
Surface m
ounte
d
Perm
anent Magnet Motor Using
Magnet Segmentation
Method”,
Electric Pow
e
r
Components and
Systems
, vol. 42
, no
. 12
, p
.
p 123
9-1248, 2014
.
BIOGRAP
HI
ES OF
AUTH
ORS
Ali Man
s
ou
ri
r
ece
ived th
e B.S
.
degree
in e
l
e
c
t
r
om
echanic
al
en
gineer
ing, th
e
M
.
S
.
degre
e
in
ele
c
tri
cal m
ach
i
n
e anal
ys
is
and
control and th
e
P
h
. D. degree in
ele
c
tri
cal
engine
ering from
S
f
ax
Engineering School (SES), University
of
Sfax, Sfax, Tun
i
sia, in 2002, 2
003 and 2009
res
p
ect
ivel
y.
He
is
working in
the fi
eld of
ele
c
tri
cal m
a
chine
des
i
gn at S
E
S
.
He joined
the
Tunisian
Univer
sity
,
Tunisia,
in
2003 as an
Assist
ant
at Gafsa
En
gineer
ing Institu
t
e. Currently
h
e
is working as A
ssociate Prof
essor in th
e Higher
Institute
of Appl
ied Sci
e
nc
e and
Techno
log
y
of
Gafsa. He is a member of the Labor
ator
y
of
Computer, Electronics
& Smart Engineering
S
y
ste
m
s De
sign
the
Unive
r
sity
of Sfa
x
.
Hejra Msaddek
received
the
B.S
.
degre
e
in ele
c
tri
cal
and a
u
tom
a
tic eng
i
ne
ering, th
e M
.
S
.
degree
in autom
a
tic and intelligent techn
i
c from
Gabes Engineering School
(GES), University
of
Gabes, Gabes,
Tuni
sia, in 2011and 2012 respectively
.
She is
working in the field of electrical
machine d
e
sign
and optimization
at Sfax
Engin
e
er
ing School SES
.
She is
a member of th
e CES
Laborator
y
of C
o
mputer, Electr
onics & Smart
E
ngineering S
y
s
t
ems Design the University
of
Sfa
x
.
Hafe
dh Tr
abe
l
si
received th
e B
.
S. degree from Sfax E
ngineerin
g School (SES), University
of
Sfax, Sfax, Tun
i
sia, in 1989
, the
M.S. degree in
the Centr
a
l Scho
ol of Ly
on, Fran
ce,
in 1990, th
e
Ph.D. degree fr
om
the Universit
y
of Paris XI
Orsay
,
France, i
n
1994 and the
"Habilit
ation à
Diriger d
e
s Recherches" (
a
ccred
itation
to lead
r
e
search)
degr
ee
from Sfax Engineering
School
(S
ES
), Univers
i
t
y
of
S
f
ax, S
f
ax,
Tunis
i
a
,
in
200
9 all
in
ele
c
tr
ic
al eng
i
ne
ering
.
He is
working
toward the R
e
se
arch Manag
e
m
e
nt Abilit
y d
e
gre
e
in the f
i
eld of
el
ectr
i
ca
l m
achin
e
design at SES.
He joined
the
Tunisian Univ
ersity
, Tun
i
sia,
in 1995 as an
Assistant Professor at Gafsa
Engine
ering Ins
titut
e
. He
is cur
r
entl
y
a Professor of Ele
c
tri
c
a
l
Engine
ering a
t
SES. He is a
member of the C
E
S Labor
ator
y
in the University
of Sfax.
Evaluation Warning : The document was created with Spire.PDF for Python.