52 IEEE TRANSACTIONS ON TRANSPORTATION ELECTRIFICATION, VOL. 6, NO.

1, MARCH 2020

Magnetic Field Analysis of SynRel and PMASynRel

Machines With Hyperbolic Flux Barriers
Using Conformal Mapping
Maedeh Sadat Mirazimi and Arash Kiyoumarsi

Abstract— Synchronous reluctance (SynRel) and permanent

magnet (PM)-assisted SynRel (PMASynRel) machines with
hyperbolic-shaped flux barriers are gaining more interest in
commercialized products of known companies, because of their
high saliency ratio. This article presents a faster approach for
calculation of the air-gap flux density distribution and the electro-
magnetic torque of these machines than the finite-element method
(FEM). The method is based on the conformal mapping (CM)
techniques and magnetic equivalent circuit (MEC) solutions.
A suitable CM is introduced to compute the magnetic reluctance
of hyperbolic-shaped rotor flux barriers. The analytical method
is then implemented on both SynRel and PMASynRel machines
with the same stator structure as the Prius interior PM motor
and five hyperbolic flux barriers per pole in their rotors. The Fig. 1. Different geometries of rotor flux barriers for a SynRel motor
accuracy of analytical results is assessed by the finite-element including (a) circular shaped [20] and (b) Zhokovski [11] flux barriers.
analysis (FEA).
Index Terms— Conformal mapping (CM), hyperbolic flux
barriers, optimization, permanent magnet-assisted synchronous lines to define the flux barriers. The circular flux barriers are
reluctance (PMASynRel) machines.
also employed in some cases to make the flux passing the rotor,
more smoothly [1] [see Fig. 1(a)]. In other cases, the barrier is
I. I NTRODUCTION composed of the conic-shaped segments [9], including hyper-

T HE optimally designed synchronous reluctance (SynRel)

motors with advanced controls are potential candidates
in many fields, such as electric and hybrid electric vehi-
bolic, parabolic, or semielliptical shape or segments, which are
expressed by Zhokovski’s function [10], [11] [see Fig. 1(b)].
Recently, the conformal mapping (CM)-based methods
cles [1], [2] and household applications [3], [4]. Their known are applied for the modeling and magnetic field analy-
benefits consist of rugged construction, the absence of exci- sis of the electrical machines; some of which are based
tation winding, less need for permanent magnets (PMs), low on the well-known Schwarz–Christoffel (SC) transformation
production cost, and high-speed capabilities [5]. It is obvious techniques, which allow them to transform field analysis
that these motors are not without their drawbacks that consist from a complex geometry into an equivalent problem, and
of low power factor, low torque density, and their far-reaching are defined on a simpler domain with an easier analytical
developments in industrial applications [6]. These demerits can solution [12]–[15].
be decreased by applying the proper PM materials into the In addition, analytical models, based on magnetic equivalent
rotor flux barriers. Thus, these motors are named PM-assisted circuit (MEC) methods, are presented in [16] and [17] for fast
SynRel (PMASynRel) motors [7]. and accurate analysis of the SynRel and PMASynRel motors to
The torque density in SynRel and PMASynRel motors be applied in the design and optimization procedures. In their
can be improved by increasing the saliency ratio, which is studies, the reluctance of rotor flux barriers is computed based
defined by different reluctances that the rotor exhibits along on the assumption that the thickness of the barriers is constant
the orthogonal d-axis and q-axis [8]. This ratio is highly along their length. These models may be hardly applied in
influenced by the geometry of rotor flux barriers. The most cases, such as circular-shaped, conic-shaped, or Zhokovski flux
common one is the C-shaped, which is associated with straight barriers, which may be of nonuniform widths. If the MEC is
composed of finer flux tube reluctance calculations, the more
Manuscript received July 29, 2019; revised October 14, 2019; accepted
November 13, 2019. Date of publication December 13, 2019; date of current accurate solutions can, therefore, be seriously achieved. Subse-
version March 20, 2020. (Corresponding author: Arash Kiyoumarsi.) quently, a number of articles, in which the combination of both
The authors are with the Department of Electrical Engineering, Faculty the CM and MEC methods are applied for field analysis of the
of Engineering, University of Isfahan, Isfahan 81746-73441, Iran (e-mail:
[email protected]; [email protected]). electrical machines, has been published by some authors. Thus,
Digital Object Identifier 10.1109/TTE.2019.2959400 the best characteristics of both methods are utilized. In [18],
2332-7782 © 2019 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission.
See https://www.ieee.org/publications/rights/index.html for more information.

Authorized licensed use limited to: Chung Yuan Christian University. Downloaded on February 02,2021 at 01:01:19 UTC from IEEE Xplore. Restrictions apply.
MIRAZIMI AND KIYOUMARSI: MAGNETIC FIELD ANALYSIS OF SynRel AND PMASynRel MACHINES 53

Fig. 3. GH at flux barrier which is considered as segment GH’ that

orthogonally intersects flux barrier hyperbolic borders at points G and H’.

Fig. 2. Characteristic geometric quantities for a single hyperbolic flux barrier. radius R − δ and center O at points A, B and G, H. Barrier
contours AB and GH are approximated as elliptic segments
that are perpendicular to the hyperbola AG at points A and G
the CM is applied to determine analytically the magnetic (see Fig. 3). Hence, each flux barrier is described by the
reluctances used in the MEC for no-load calculation of the geometric parameters of , h, δ, ε, α, where α is the angle
salient-pole wound-field synchronous generator. An iterative between the horizontal line passing through points O and D
method based on the CM and the MEC is presented in [19] and the line connecting O to the central point of segment GH
for the on-load analysis of the saturated surface-mounted and angle γ is defined as a deviation angle around α, which is
PM machines with integer-slot and fractional-slot windings. approximated from (1). According to the above assumptions
An appropriate CM is introduced and applied in [20] in order and introducing a Cartesian reference frame having the axes x
to calculate the reluctance of the circular-shaped flux barriers, and y and its origin O, the coordinates of any generic points
which are then used in the MEC calculations of the SynRel of the plane, i.e., A, B, D, and E are expressed as
motors with the circular-shaped flux barriers.  
In this article, a new conformal transformation is presented ε/2 ε
γ = sin−1 ∼ (1)
and applied to compute the reluctance of hyperbolic-shaped R 2R
flux barriers with nonuniform widths, which are considered x A = (R−δ) cos(π −α+γ ), y A = (R − δ) sin(π − α+γ )
here as an approximation of the Zhokovski flux barriers. (2)
Based on these computations and through the extension of the
x B = (R−δ) cos(π −α−γ ), y B = (R − δ) sin(π − α − γ )
presented analytical method in [20], the air-gap flux density
of both the SynRel and PMASynRel motors with hyperbolic (3)
flux barriers is analytically determined. Finally, the analytical x D = −R + +h, y D = 0, x E = −R + , y E = 0. (4)
method is evaluated through the finite-element analysis (FEA)
simulations. Note that the assumption (ε  R) holds true in (1).
This article is, therefore, an extension of [20] and is The coordinate of the foci (F1 ) of the hyperbola 1 passes
organized as follows. The mathematical modeling of the rotor through points A and E can be determined as
structure is presented in Section II. The new extended ana- 
lytical method for hyperbolic-shaped flux barrier calculations x F1 = −c1 , y F1 = 0, c1 = a12 + b12
is proposed in Section III. Then, in Sections IV and V, the y A (R − )
proposed analytical method is applied to predict the air-gap a1 = R − , b1 =  . (5)
(x A )2 − (R − )2
field distribution of two case studies, i.e., a SynRel and a
PMASynRel, respectively. Moreover, the results of this ana- The Cartesian equation of the hyperbola 1 with F1 as the
lytical method are evaluated by comparison against the FEA foci and O as the center is yield
simulations. Finally, the conclusions are given in Section VI.
x2 y2
This article also has two appendices about the design process − = 1. (6)
of the rotor hyperbolic-shaped flux barriers of the SynRel a12 b12
machine and the PM design of the PMASynRel motor. The Cartesian equation of 2 passes through
points B and D with F1 as its foci and O1 as its center is
II. G EOMETRIC M ODEL D ESCRIPTION yield
The rotor structure of a SynRel motor with a generic (x − x 0 )2 y2
represented hyperbolic flux barrier is shown in Fig. 2. − = 1. (7)
a22 c12 − a22
According to this modeling, the machine rotor is of center O
and radius R, and the barrier borders AG and BH lie on The x-coordinate of the point O1 and the unknown
hyperbolas 1 (with center O and foci F1 ) and 2 (with center parameter a2 can be determined by replacing the coordi-
O1 and foci F1 ), which intersect the circumference having nates of the points B and D in (3) and (4), which would

Authorized licensed use limited to: Chung Yuan Christian University. Downloaded on February 02,2021 at 01:01:19 UTC from IEEE Xplore. Restrictions apply.
54 IEEE TRANSACTIONS ON TRANSPORTATION ELECTRIFICATION, VOL. 6, NO. 1, MARCH 2020

yield (8). By solving (8), the unknown parameters of 2

Cartesian equation are determined as follows:
(x A − x 0 )3 + (τ1 )(x A − x 0 )2 + (τ2 )(x A − x 0 ) + τ3 = 0 (8)
y 2 + (x B − x A )2
x A − x 0 = a 2 , τ1 = B , τ2 = −c21
2(x B − x A )
c2 (x B − x A )
τ3 = − 1 . (9)
2
As mentioned, the segments AB and GH are approximately
assumed to belong to an ellipse that orthogonally intersects
the confocal hyperbolas 1 and 2 . By naming F1 as its foci,
its Cartesian equation is yield as
x2 y2
+ = 1. (10)
c12 (cosh(β))2 c12 (sinh(β))2
Fig. 4. Conformal transformation of the flux barrier geometry (z-plane) into
The sum of the focal distances of point A is equal to the a rectangle (w-plane).
focal length of the ellipse, so the unknown parameter β can
be determined as follows:
  B. Implementation of the Selected Map
2c1 cosh(β) = (x A +c1 )2 + y2A + (x A −c1 )2 + y2A . (11)
By applying the complex function w = f (z) expressed
III. F LUX BARRIER R ELUCTANCE C OMPUTATION in (12), the geometry of a flux barrier in the original
z-plane is transformed into a rectangle in the new w-plane,
Contrary to C-shape and rectangular flux barriers, the reluc-
in a conformal manner (see Fig. 4). As to a geometric
tance computation for the hyperbolic flux barriers is com-
context, the selected map transforms confocal hyperbolas
plicated, due to their nonuniform thicknesses. An analytical
of z plane into the vertical straight lines of w-plane (see
method is proposed, here, to compute the magnetic reluctance
Fig. 4). Furthermore, by applying the selected map, the elliptic
of a hyperbolic-shaped flux barrier through CM.
curves having the same foci as the mentioned hyperbolas and
orthogonally intersecting them in z-plane are mapped into
A. Analytical Properties of the Selected Map horizontal straight lines of w plane.
The physical complex plane z = x + j y identified by origin The hyperbolic curve 1 passing through points A and E
O and axes x and y (see Fig. 2) is mapped into a new with the Cartesian form of (3) is transformed into line u 1
complex plane w = u + j v through the following conformal through the following equation:
transformation [27]: ⎛  ⎞
  (x E + c1 )2 + y E2 − (x E − c1 )2 + y E2
−1 z
w = sin . (12) u 1 = sin−1 ⎝ ⎠
c1 2c1
The inverse of which is  
− R
= arcsin . (18)
z = c1 sin(w). (13) c1
Equation (12) can be written in an equivalent form as The hyperbolic curve 2 passing through points B and D
  with the Cartesian (2) and (4) is transformed into line u 2
−1 x + j y
w = u + j v = sin through the following equation:
c1
    ⎛  ⎞
S1 − S2 S1 + S2
= sin−1 + j sgn(y)arccosh (14) (x D + c1 )2 + y 2D − (x D − c1 )2 + y 2D

2c1

2c1 u 2 = sin−1 ⎝ ⎠
2c1
S1 = (x + c1 )2 + y 2 , S2 = (x − c1 )2 + y 2 . (15)  
− R+h
Equation (14) naturally leads to define the real and imagi- = arcsin . (19)
c1
nary parts (u, v) of variable w in terms of real and imaginary
parts (x, y) of variable z as As shown in Fig. 4, the segments AB’ and GH’ belong-
   ing to an ellipse that orthogonally intersects the hyperbolas
−1 (x + c1 )2 + y 2 − (x − c1 )2 + y 2
u(x, y) = sin 1 and 2 are mapped on the lines v 1 and v 2 through the
2c1 following equation:
(16) ⎛  ⎞
v(x, y) = sgn(y)arccosh (x A + c1 )2 + y 2A + (x A − c1 )2 + y 2A
   v 1 = arccosh ⎝ ⎠
(x + c1 )2 + y 2 + (x − c1 )2 + y 2 2c1
× . (17)
2c1
v 2 = −v 1 . (20)

Authorized licensed use limited to: Chung Yuan Christian University. Downloaded on February 02,2021 at 01:01:19 UTC from IEEE Xplore. Restrictions apply.
MIRAZIMI AND KIYOUMARSI: MAGNETIC FIELD ANALYSIS OF SynRel AND PMASynRel MACHINES 55

C. Determining the Magnetic Field in the Flux Barrier as follows:

⎡ ⎤
Both the vector potential and the magnetic field inside the  ∂ A(x, y)
Bx (x, y) ⎢ ∂y ⎥
generic hyperbolic flux barrier in the z plane (see Fig. 4) are B(x, y) = =⎣
determined by fixing the proper boundary conditions along its B y (x, y) ∂ A(x, y) ⎦
−
borders. By assuming the infinitely permeable iron core except ⎡ ∂x y y ⎤
the iron bridges, where some magnetic saturation occurs, it is
−B0 sgn (y) +
⎢ S S2 ⎥
reasonable to deduce that Neumann boundary conditions are =  2 ⎣ x + c11 x − c1 ⎦ . (26)
held along the barrier borders AG and B’H’ and impose the S1 +S2 +
c1 2c1 −1 S1 S2
Dirichlet boundary conditions on segments AB’ and GH’.
Imposing the Dirichlet boundary conditions on segments AB’
and GH’ is equivalent to assume that no flux lines pass through D. Flux Barrier Reluctance Calculation
these segments. Such an assumption is reasonable, because The magnetic reluctance of the generic hyperbolic flux
the amount of the flux crossing the borders AB’ and GH’ is barrier (Rb ) is computed as
assumed to be negligible compared with the flux crossing the
borders AG and B’H’ [20]. Rb = FED /AG (27)
Once the applied boundary conditions are fixed in the com-
plex plane z, they are automatically transferred to the rectangle where AG is the flux flowing through the barrier border AG
borders in the complex plane w. By imposing the mentioned and FED is the magneto-motive force (MMF) drop across the
boundary conditions of the flux barrier borders, in the w plane, barrier between points E and D (see Fig. 4). Parameter AG
the field solution inside the rectangle is expressed through a is computed through the vector potential as follows [20]:
uniform flux density with a constant magnitude B0 oriented AG = L[ A(x G , yG ) − A(x A , y A )] (28)
along the u-axis as follows:
  where L is the core length and A (x G , yG ) and A (x A , y A ) are
 Bu (u, v) B0 the magnetic vector potentials at points A and G, respectively.
B (u, v) = = . (21)
Bv (u, v) 0 As observed in Fig. 4, yG = −y A ; thus, A (x G ,yG ) = −A
(x A , y A ) according to (25). The final explicit expression for
where B0 is a parameter, which depends on the operating point AG is expressed as
of the SynRel motor, and is defined by the stator currents and
rotor position [20]. AG = −2L A(x A , y A ) = 2L B0 arccosh
⎛  ⎞
The magnetic vector potential in barrier domain in the w (x A +c1 )2 + y 2A + (x A − c1 )2 + y 2A
plane is obtained through the following equation [20], [21]: ×⎝ ⎠. (29)
2c1

∂ A (u, v)
Bu = − . (22) The MMF drop FED is derived by integrating the magnetic
∂v field (26) along a linear path linking points E and D as follows:
By inserting (21) in (22), (23) is yield    R−−h
1
FED = B (x, 0) d x
 μ
   0   R−    
A (u, v) = − B0 dv = −B0 v + A0 (23) B0 − R+h − R
= arcsin − arcsin .
μ0 c1 c1
where A0 is an arbitrary constant. According to the CM (30)
theory [22]–[30], the vector potential solutions in z-plane By inserting (29) and (30) into (27), the explicit analytical
( A (u, v)) and w-plane ( A(x, y)) are similar as follows: expression for the reluctance of a generic hyperbolic flux

barrier can be determined
A(x, y) = A (u(x, y), v(x, y)). (24) ⎡ ⎤
   
  ⎢ arcsin −R+h − arcsin −R ⎥
By inserting (17) into (23), where A0 = 0, the vector potential B0 ⎢⎢ c1 c1 ⎥
Rb = ⎢   ⎥
⎥. (31)
solution in w-plane is expressed as μ0 ⎣ (x A +c1 )2 +y 2A + (x A −c1 )2 +y 2A ⎦
arccosh 2c1
A(x, y) = −B0 × sgn(y) × arccosh
By inserting (18) and (19) into (31), it can be rewritten as
   follows:
(x + c1 )2 + y 2 + (x − c1 )2 + y 2
× . (25) 1 u2 − u1 1 X
2c1 Rb = = . (32)
2Lμ0 v 1 μ0 Y L
The flux density components inside the flux barrier in the z This expression is equivalent to the barrier reluctance in the
plane are generated by derivations of vector potential A(x, y) w plane.

Authorized licensed use limited to: Chung Yuan Christian University. Downloaded on February 02,2021 at 01:01:19 UTC from IEEE Xplore. Restrictions apply.
56 IEEE TRANSACTIONS ON TRANSPORTATION ELECTRIFICATION, VOL. 6, NO. 1, MARCH 2020

Fig. 5. Cross section of the studied SynRel motor.

TABLE I
S PECIFICATIONS OF THE S YN R EL M OTOR

Fig. 6. Schematic MEC model for this SynRel motor.

air-gap width. The flux barriers are numbered from 1 to 5.

The whole air-gap of the motor is subdivided into ten regions,
marked as 11, 12, 21, 32, 23, 34, 43, 45, 54, and 55 by the
points of the intersection of the end barriers and the outer
TABLE II rotor periphery. For each of the regions, an MMF source
G EOMETRICAL PARAMETERS OF THE M OTOR F LUX BARRIERS (Fi j ) and an air-gap reluctance (Rgi j ) are defined. Moreover,
the reluctances Rb1 , Rb2 , Rb3 , Rb4 , and Rb5 of the five flux
barriers are defined in the MEC and analytically computed
by the proposed method, as discussed in Section III. The flux
flowing through these reluctances are 1 , 2 , 3 , 4 , and 5 .
Nodes S, A, B, C, D, E, and R of the MEC represent core
regions, where no MMF drop to occur with an assumption
of the fact that the permeability of the iron cores, except
the bridges, is infinite. It is notable that the fluxes flowing
through the bridges are disregarded, because they are small
and saturated [16], [20].

B. Computation of Reluctance Network

IV. A PPLICATION E XAMPLE I
1) Calculation of MMF Sources: The overall air-gap MMF
The methodology presented for analytical computations of field is first defined due to the stator phase currents as
the reluctance of a hyperbolic flux barrier is illustrated and a function of both time and angular position through the
assessed by applying it to an eight-pole five-layer rotor SynRel following equation [20]:
motor, the cross section of which is shown in Fig. 5. The 
motor flux barrier design is discussed briefly in Appendix I. F(t, θ ) = wn (θ ) · i n (t) (33)
Its specifications and some geometrical parameters of the flux n=0,1,2
barriers are shown in Tables I and II, respectively. The stator where θ is the angular position (in electrical radians) measured
yoke and winding configuration of the motor are the same along the mean air-gap circumference with respect to the MMF
as the one in the 2004 Prius IPM motor. An analytical axis of phase a winding, i n is the current of each phase
solution based on the MEC method is applied to determine (n = 0, 1, 2), and wn is the winding function of phase n,
the air-gap flux density of the studied SynRel motor for any which is computed according to [20].
rotor position and the stator phase currents. Accuracy of the Once the total air-gap MMF is computed, the MMF sources
analytical results is evaluated by comparing them through the of the various regions, shown in Fig. 6, are computed as
FEA simulations [21]–[32]. follows:
 θr +α1 p
1
A. Modeling the SynRel Motor Through Reluctance Network F11 (t, θr ) = F(t, θ )dθ (34)
2α1 p θr −α1 p
The MEC modeling of a two-pole SynRel motor is shown in  θr +α j p
Fig. 6. For the sake of clarity and simplicity, the case of a two- 1
Fi j (t, θr ) = F(t, θ )dθ,
pole machine is of concern, here. As shown in Fig. 6, the sta- (α j − αi ) p θr +αi p
tor slotting effect is neglected while allowing the uniform i = 1, 2, . . . , 4, j = i + 1 (35)

Authorized licensed use limited to: Chung Yuan Christian University. Downloaded on February 02,2021 at 01:01:19 UTC from IEEE Xplore. Restrictions apply.
MIRAZIMI AND KIYOUMARSI: MAGNETIC FIELD ANALYSIS OF SynRel AND PMASynRel MACHINES 57

TABLE III
R ELUCTANCE VALUES × 105 (A/W B ) O BTAINED BY THE FE AND
A NALYTICAL M ETHODS FOR THE BARRIERS OF THE
D ESIGNED S YN R EL M OTOR

 θr −αi p
1
F j i (t, θr ) = F(t, θ )dθ (36)
(α j − αi ) p θr −α j p

where, α1 , α2 , α3 , α4 , and α5 are the mechanical angles of

the barrier ends with respect to the rotor q-axis. Fig. 7. MEC of (a) SynRel motor of Fig. 5 and (b) PMASynRel motor
2) Computation of Air-Gap Reluctances: Reluctances of the of Fig. 13.
MEC are computed as follows: SC − F (t, θ ) + R  (t, θ )
  FthSC Rth eq r eq 5 r
1 g 3 (t, θr ) = SC + R
Rgi j = Rg j i =   , i = 1, 2, . . . , 4, Rth b3
μ0 α j − αi RL F4 (t, θr )
j =i +1 (37) − SC (41)
    Rth + Rb3
1 g 1 g
Rg11 = , Rg55 =   (38) SD − F (t, θ ) + R  (t, θ )
FthSD Rth
μ0 2α1 RL μ0 πp − 2α 5 RL 4 (t, θr ) =
eq r eq 5 r
(42)
SD + R
Rth b4
where g is the air-gap length, p is the pole-pair number, R is
FthSE
the mean air-gap radius, and L is the core length. 5 (t, θr ) = SE + R
(43)
3) Computation of the Flux Barrier Reluctances: The Rth b5
reluctances of the hyperbolic flux barriers of this motor where
are computed analytically via (31). To assess the accuracy
F54 FthSD Rth
SD
F45
of the proposed mapping application in magnetic reluctance FthSE (t, θr ) = + SD + (44)
prediction, the reluctance obtained by the finite element and Rg54 Rth + Rb4 Rg45
 −1  
SD −1 −1 −1
analytical methods are compared in Table III. As it is shown SE
Rth (t, θr ) = Rg54 + Rb4 + Rth + Rg45 (45)
the percent error between the values of the reluctances of the
F43 FthSC Rth
SC
F34
1st, 2nd, 3rd, 4th, and 5th flux barrier are 31, 7.2, 5.44, and 4.5. FthSD (t, θr ) = + SC + (46)
It is obvious that the proposed mapping involves a certain Rg43 Rth + Rb3 Rg34
 −1  
SC −1 −1 −1
degree of approximation, since the barrier segments AB and Rth (t, θr ) = Rg43 + Rb3 + Rth
SD
+ Rg34 (47)
GH (see Figs. 3 and 4) are modeled and approximated as
F32 FthSB Rth
SB
F23
elliptic segments that are not perpendicular to the hyperbolic FthSC (t, θr ) = + SB + (48)
segments of the barrier. Rg32 Rth + Rb2 Rg23
   −1  
SB −1 −1 −1
4) Reluctance Network Solution: The MEC model of SC
Rth t, θr = Rg32 + Rb2 + Rth + Rg23 (49)
Fig. 6 for each one of the poles of this motor and the
F12 F11 F21
reluctance network shown in Fig. 7(a) are equivalent. The FthSB (t, θr ) = + + (50)
MEC can be solved through electric circuit theory analysis Rg12 Rg11 + Rb1 Rg21
 −1  −1 −1 −1
methods [21]–[23]. The fluxes 1 , 2 , 3 , 4, and 5 SB
Rth (t, θr ) = Rg21 + Rb1 + Rg11 + Rg12 (51)
passing through each one of the barriers are determined Feq (t, θr ) = FthSE , Req = Rth
SE
. (52)
as [20], [31], [32]
F11 − Feq (t, θr ) + Req 5 (t, θr ) The terms F1 , F2 , F3 , F4 , and F5 in (39) and (40),
1 (t, θr ) = which are the MMF drops due to the flux flowing across each
Rg11 + Rb1
one of the flux barriers, are computed as follows:
F2 (t, θr ) + F3 (t, θr ) + F4 (t, θr )
− (39)
Rg11 + Rb1 F 1 (t, θr ) = Rb1 1 (t, θr ), F 2 (t, θr ) = Rb2 2 (t, θr )
SB − F (t, θ ) + R  (t, θ )
FthSB Rth (53)
eq r eq 5 r
2 (t, θr ) = SB + R
Rth b2 F 3 (t, θr ) = Rb3 3 (t, θr ), F 4 (t, θr ) = Rb4 4 (t, θr )
F3 (t, θr ) + F4 (t, θr ) (54)
− SB + R
(40)
Rth b2 F 5 (t, θr ) = Rb5 5 (t, θr ). (55)

Authorized licensed use limited to: Chung Yuan Christian University. Downloaded on February 02,2021 at 01:01:19 UTC from IEEE Xplore. Restrictions apply.
58 IEEE TRANSACTIONS ON TRANSPORTATION ELECTRIFICATION, VOL. 6, NO. 1, MARCH 2020

The MMF drops expressed in (53)–(55) can be applied to

any path crossing each one of the flux barriers, because the
MMF crossing each one of the barriers is independent of the
path along which it is computed.

C. Air-Gap Flux Density Computation

The MEC solution of this SynRel motor can be applied
in computing the air-gap flux density distribution in various
air-gap regions of the motor. To accomplish this, the air-gap Fig. 8. Comparison of calculated air-gap flux density distribution (radial
flux density due to the stator phase currents is modified by component) using FEM and analytical method for rotor q-axis aligned with
including the flux barriers effects through the MMF drops stator MMF axis. (It is mentionable that the magnetic saturation is not taken
into consideration in this analysis.)
across each one of them. The flux density is computed at an
instant t (whenever the rotor is at θr ) at a point with the a fair comparison, the permeability of iron core, except the
angular position θ P , which is placed in the air-gap region bridges, is assumed to be infinite. Moreover, here, the slotting
spanning between θr -pα1 and θr + pα1 ; by defining the closed effects are disregarded. The sinusoidal currents, having a peak
path  P which crosses the air-gap at positions θ P and θ P + π. value of 250 A, are assigned to all phases. The time instant
As this defined path crosses all flux barriers, for the purpose is set to zero, phase “a” carries 250 A and the other two
of the air-gap flux density derivation, five MMF drops have phases carry −125 A, each. With the given stator currents,
to be of concern, each occurring across each one of the five the rotor q-axis aligned with stator MMF axis and in the other
flux barriers (i.e., along segments P2 P3 , P4 P5 , P6 P7 , P8 P9 , state, the rotor q-axis is shifted by 30 mechanical degrees
and P10 P11 ) and are equivalent to F1 , F2 , F3 , F4 , with respect to stator MMF axis, and for each rotor position,
and F5 , respectively, computed through (53)–(55) [20]. the radial flux density along the mean air-gap circumference
As a result, the air-gap flux density at position θ P, whenever is obtained through the FEA. A comparison between the FEA
rotor is located at θr , is computed as follows: and analytical simulations is shown in Fig. 8, which meets our
expectation. Based on the fast Fourier transform (FFT) analy-
Bg (t, θ p , θr )
  sis, the fundamental amplitudes of the air-gap flux density,
μ0 computed by the FE and analytical methods, are 0.7798 and
= [F(t, θ P ) − F1 (t, θr ) − F2 (t, θr )
g 0.6818 T, respectively. The computation time of the air-gap
− F3 (t, θr ) − F4 (t, θr ) − F5 (t, θr )]. (56) flux density predictions through the analytical and the FE
The above procedure can be extended to extract the resul- methods is 14.9 and 911 s., respectively. An FEM software
tant MMF in different air-gap regions, including the stator was considered for the FEM calculations, which was run on
current contribution and the MMF drops across rotor bar- an Intel Core i5 processor 2.66 GHz, with 4 GB of RAM
riers. The general form of the air-gap flux density can be (a laptop).
expressed by (57), shown at the bottom of this page.
This analytical method for the air-gap flux density determi- D. Torque Calculation
nation is applied to the SynRel motor. The obtained results The electromagnetic torque Te (t, θr ) can be computed
are compared with the ones obtained by the FEA. To make through the derivative of total stored magnetic energy of the

B g (t, θ , θ r )
⎧ ⎫
⎪
⎪ F(t, θ )−F T ot al5 (t, θ r ) i f (0 ≤ |mod(θ −θ r , 2π )| ≤ pα 1 ) ∨ (2π − pα 1 ≤ |mod(θ −θ r , 2π )| ≤ 2π ) ⎪
⎪
⎪
⎪ ⎪
⎪
⎪
⎪ F(t, θ )+F T ot al5 (t, θ r ) i f (π − pα 1 ≤ |mod(θ − θ r , 2π )| ≤ π + pα 1 ) ⎪
⎪
⎪
⎪ ⎪
⎪
⎪
⎪ F T ot al5 (t, θ r ) = F 1 (t, θ r )+F 2 (t, θ r )+F 3 (t, θ r )+F 4 (t, θ r ) + F 5 (t, θ r ) ⎪
⎪
⎪
⎪ ⎪
⎪
⎪
⎪ F(t, θ )−F T ot al4 (t, θ r ) i f ( pα 1 ≤ |mod(θ − θ r , 2π)| ≤ pα 2 ) ∨ (2π − pα 2 ≤ |mod(θ − θ r , 2π)| ≤ 2π − pα 1 ) ⎪
⎪
⎪
⎪ ⎪
⎪
⎪ F(t, θ )+F T ot al4 (t, θ r ) i f (π + pα 1 ≤ |mod(θ − θ r , 2π )| ≤ π + pα 2 ) ∨ (π − pα 2 ≤ |mod(θ −θ r , 2π )| ≤ π − pα 1 )⎪ ⎪
⎪
⎪
⎪ ⎪
⎪
⎪
⎪ F T ot al4 (t, θ r ) = F 2 (t, θ r ) + F 3 (t, θ r ) + F 4 (t, θ r ) + F 5 (t, θ r ) ⎪
⎪
⎪
⎪ ⎪
⎪
⎨F(t, θ )−F T ot al3 (t, θ r ) i f ( pα 2 ≤ |mod(θ − θ r , 2π )| ≤ pα 3 ) ∨ (2π − pα 3 ≤ |mod(θ − θ r , 2π )| ≤ 2π − pα 2 ) ⎬
= F(t, θ )+F T ot al3 (t, θ r ) i f (π + pα 2 ≤ |mod(θ − θ r , 2π)| ≤ π + pα 3 ) ∨ (π − pα 3 ≤ |mod(θ −θ r , 2π )| ≤ π − pα 2 )
⎪
⎪ ⎪
⎪
⎪
⎪ F T ot al3 (t, θ r ) = F 3 (t, θ r ) + F 4 (t, θ r ) + F 5 (t, θ r ) ⎪
⎪
⎪
⎪ ⎪
⎪
⎪
⎪ F(t, θ )−F T ot al2 (t, θ r ) i f ( pα 3 ≤ |mod(θ − θ r , 2π)| ≤ pα 4 ) ∨ (2π − pα 4 ≤ |mod(θ − θ r , 2π )| ≤ 2π − pα 3 ) ⎪
⎪
⎪
⎪ ⎪
⎪
⎪ F(t, θ )+F T ot al2 (t, θ r ) i f (π + pα 3 ≤ |mod(θ − θ r , 2π )| ≤ π + pα 4 ) ∨ (π − pα 4 ≤ |mod(θ −θ r , 2π )| ≤ π − pα 3 ) ⎪ ⎪
⎪
⎪
⎪ ⎪
⎪
⎪
⎪ F T ot al3 (t, θ r ) = F 4 (t, θ r ) + F 5 (t, θ r ) ⎪
⎪
⎪
⎪ ⎪
⎪
⎪
⎪ F(t, θ )−F 5 (t, θ r ) i f ( pα 4 ≤ |mod(θ − θ r , 2π )| ≤ pα 5 ) ∨ (2π − pα 5 ≤ |mod(θ − θ r , 2π )| ≤ 2π − pα 4 ) ⎪
⎪
⎪
⎪ ⎪
⎪
⎪ F(t, θ )+F 5 (t, θ r ) i f (π + pα 4 ≤ |mod(θ − θ r , 2π )| ≤ π + pα 5 ) ∨ (π − pα 5 ≤ |mod(θ −θ r , 2π)| ≤ π − pα 4 ) ⎪ ⎪
⎪
⎩ ⎭
F(t, θ ) ot herwi se
(57)

Authorized licensed use limited to: Chung Yuan Christian University. Downloaded on February 02,2021 at 01:01:19 UTC from IEEE Xplore. Restrictions apply.
MIRAZIMI AND KIYOUMARSI: MAGNETIC FIELD ANALYSIS OF SynRel AND PMASynRel MACHINES 59

Fig. 11. Analytical and the FEA results for the no-load air-gap flux density
distribution of the PMASynRel motor.

Fig. 9. Comparison between the FEA and the analytical results for the
electromagnetic torque as a function of the rotor position with fixed stator
currents.

Fig. 12. Initial rotor design of the SynRel motor and the simplified modified
rotor design based on the solid rotor flux lines.

V. A PPLICATION E XAMPLE II
A. No-Load Flux Density Prediction in
the PMASynRel Motor
By applying the PMs into the SynRel motor flux barriers,
a PMASynRel motor is achieved, the cross section of which
Fig. 10. Electromagnetic torque as a function of time during rated operation is shown in Fig. 12. The PM dimensions of the PMASynRel
obtained by the analytical and the FE methods. motor are determined according to [33]–[35]. The thickness
and the location of the PMs are limited to the shape of the
machine (Wm (t, θr )) with respect to the rotor position(θr ) as rotor barriers of the existing SynRel motor [36], [37].
follows [34]: The MEC model of the PMASynRel motor is shown in
Fig. 7(b). The analytical method based on the CM techniques,
∂ Wm (t, θr ) Wm (t, θr + θr )
Te (t, θr ) = − =− (58) combined with magnetic network analysis, which is proposed
∂θr θr in this article, is extended and applied to predict the no-load
where Wm (t, θr ) is the sum of the stored energy in the flux density of this PMASynRel motor. The air-gap flux
g
air gap (Wm (t, θr )) and the stored energy in the flux bar- density can now be expressed in the same way as shown
riers (Wmb (t, θr )), which are approximated by the following in (57), which is not brought in this article, because of the lack
equations [34]: of enough space. The obtained results are compared with the
 one calculated through the FEA (see Fig. 11). The remanent
g L Rg 2π flux density of the used Ferrite PMs is 0.39 T. It is mentionable
Wm (t, θr ) = Bg (t, θ, θr )2 dθ (59)
2μ0 0 that the rotors of the SynRel and the PMASynRel machines,
5 used in this article, are completely designed by the authors
Wmb (t, θr ) = p Rbi i (t, θr )2 (60) based on the stator configuration of the Prius interior PM
i=1 synchronous motor, which is briefly explained in Appendix II.
where R is the mean air-gap radius and Bg (t, θ, θr ) and VI. C ONCLUSION
i (t, θr ), i = 1, .., 5 are given by (57) and (39)–(43),
respectively. Here, in this article, an analytical method for both the pre-
Torque values computed with the energy-based method with diction of the air-gap flux density distribution of SynRel and
an assumption of neglecting magnetic saturation are assessed PM-assisted SynRel (PMASynRel) motors with hyperbolic
by the ones obtained by the FEA. A comparison between the flux barriers, and the reluctance calculation of these barriers
FEA and analytical results for the electromagnetic torque as a is proposed. This method is based on the CM techniques and
function of the rotor position with fixed stator currents (with a MEC solutions. The reluctances associated with flux barriers
peak value of 250 A) is shown in Fig. 9. The electromagnetic are computed by means of the presented CM. As a result of the
torque as a function of time during synchronous operation, FEA-based assessment, a satisfactory matching exists between
at nominal conditions, is obtained by the analytical and the FE the results of this proposed method and the FEA predictions.
methods and compared in Fig. 10. The results show that the
A PPENDIX I
torque ripple and the average torques obtained by the analytical
P RELIMINARY ROTOR D ESIGN OF THE SynRel M OTOR
method are not in very good accordance with those obtained
by the FEA. It is due to neglecting the tangential component The preliminary rotor design of the SynRel motor is
of the air-gap flux density in its energy estimation. based on the pattern of flux lines in a solid rotor that

Authorized licensed use limited to: Chung Yuan Christian University. Downloaded on February 02,2021 at 01:01:19 UTC from IEEE Xplore. Restrictions apply.
60 IEEE TRANSACTIONS ON TRANSPORTATION ELECTRIFICATION, VOL. 6, NO. 1, MARCH 2020

represents the minimum reluctance in the d-axis while a

high reluctance in the q-axis [11]. The barrier curves are
drawn based on the analytical equations of the flux lines as
follows [11], [29], [30]:
 
2
rshaft r 2 − rshaft2
r− sin( pθ ) = C, 0 < C < rotor (61)
r rrotor
where C is the constant for each barrier curve and p is Fig. 13. Flux density distribution contours for the initial and final designs
the pole-pair number. The parameters rrotor , rshaft , r , and θ of the PMASynRel motor.
represent the rotor radius, the shaft radius and the radial
distance of a given point from the center of the shaft and
the mechanical angle of the same point with respect to the
phase “a” MMF axis, respectively.
The thickness and the position of flux barriers are
determined, so that an optimized uniform flux distribution
in the rotor is achieved [11], [36], [37]. Based on the
hyperbolic-shaped flux barriers in the analytical method, using
the CM techniques, the initial rotor design is modified into a Fig. 14. Selected electromagnet torque at rated condition for the initial and
rotor design with hyperbolic flux barriers (see Fig. 12), then final designs of the PMASynRel motor.
the new analytical method is used.

A PPENDIX II optimization process, which is governed through reducing

P RELIMINARY ROTOR D ESIGN OF torque pulsations at the load angle related to maximum-torque-
THE PMASynRel M OTOR per-ampere (MTPA) operating point. The widths of the PMs
are chosen as the factors or design variables. These design
Actually, as an application example, this new method can
variables are supposed to change in four different levels. Given
be effectively applied hereafter to determine the final size of
the factors and their levels, an L 16 (45 ) orthogonal array is here
the magnets through the optimization procedure as explained
chosen, which needs 16 experimental trials. Each trial is an
in the following.
MEC calculation and solution to obtain torque characteristics.
Whenever the predesigned SynRel motor with hyperbolic
The reluctances of the MEC are computed by the CM method
flux barriers is converted into a PMSynRel motor, the size of
through (24). The optimum level for each of the factors is
the magnets should then be carefully determined. The design
determined, based on the analysis of mean and the analysis of
of the magnet size is followed by two steps in this example:
variance [38].
1) The preliminary PM design based on the simplified
The flux density distribution calculated through the FEM,
presented MEC.
for the initial and final rotor designs, are shown in Fig. 13.
2) The final PM design through the Taguchi-based opti-
As it is shown, there is less magnetic flux in some areas
mization process.
As the thickness of the PMs are limited to the shape of the of the flux guides in the initial design, which are denoted
hyperbolic barriers of the predesigned SynRel motor, the PM by the rectangles, compared with the saturated areas near
widths determination are of concern. With an assumption that the iron bridges. However, by increasing the magnet widths
the PM widths are similar and they are placed in the center located in layers 2, 4, and 5, the magnetic flux has been more
of the rotor barriers of the SynRel motor, at first, the MEC balanced in the rotor of the final design. The electromagnetic
shown in Fig. 7(b) is converted into the simplified MEC. As it torque profiles, obtained by the time-stepping FEM for the
is observed, gi is the air-gap flux flowing through the i th flux preliminary and the final designs of the PMASynRel motors,
barrier. Based on this simplified MEC, the PM widths are are shown in Fig. 14. As it is shown, the torque pulsation is
computed through suppressed in the final design.


5
R EFERENCES
Hc lmi = (Rg11 + 5Rm1 )g1 , g1
[1] S. Taghavi and P. Pillay, “A sizing methodology of the synchronous
i=1
 g reluctance motor for traction applications,” IEEE J. Emerg. Sel. Topics
= Bg1avg A g11 , A g11 = 2α1 Rin + L stk sin( pθ ) Power Electron., vol. 2, no. 2, pp. 329–340, Jun. 2014.
2 [2] A. A. Arkadan, N. Al-Aawar, and A. A. Hanbali, “Design optimization
(62) of SynRM drives for HEV power train applications,” in Proc. IEEE Int.
Electr. Mach. Drives Conf., vol. 1, May 2007, pp. 810–814.
where Bg1avg is the average air-gap flux density in front [3] H. Nam, S. B. Park, G. H. Kang, J. P. Hong, J. B. Eom, and
T. U. Jung, “Design to improve starting performance of line-start
of the first flux barrier, which depends on the adopted PM synchronous reluctance motor for household appliances,” in Proc. IEEE
material [37], and Rm1 is the reluctance of the first PM. Ind. Appl. Conf., 39th IAS Annu. Meeting, vol. 1, Oct. 2004, p. 85.
By applying the ferrite PMs with a residual flux density of [4] H. Kiriyama, S. Kawano, Y. Honda, T. Higaki, S. Morimoto, and
Y. Takeda, “High performance synchronous reluctance motor with multi-
0.39 T, and using (62), the PM width is determined. The flux barrier for the appliance industry,” in Proc. IEEE Ind. Appl. Conf.
final design of the PMs are based on the Taguchi-based 33rd IAS Annu. Meeting, vol. 1, Oct. 1998, pp. 111–117.

Authorized licensed use limited to: Chung Yuan Christian University. Downloaded on February 02,2021 at 01:01:19 UTC from IEEE Xplore. Restrictions apply.
MIRAZIMI AND KIYOUMARSI: MAGNETIC FIELD ANALYSIS OF SynRel AND PMASynRel MACHINES 61

[5] P. Niazi, H. A. Toliyat, D. H. Cheong, and J. C. Kim, “A low-cost [27] E. Kreyszig, Advanced Engineering Mathematics, 6th ed. Hoboken, NJ,
and efficient permanent-magnet-assisted synchronous reluctance motor USA: Wiley 1988.
drive,” IEEE Trans. Ind. Appl., vol. 43, no. 2, pp. 542–550, Mar. 2007. [28] R. V. Churchill, J. W. Brown, and R. F. Verhey, Complex Variables and
[6] Y. S. Kim and I. H. Park, “Topology optimization of rotor in synchronous Applications, 3rd ed. New York, NY, USA: McGraw-Hill, 1974.
reluctance motor using level set method and shape design sensitivity,” [29] J. D. Jackson, Classical Electrodynamics, 3rd ed. New York, NY, USA:
IEEE Trans. Appl. Supercond., vol. 20, no. 3, pp. 1093–1096, Jun. 2010. Wiley, 1999.
[7] P. Guglielmi, B. Boazzo, E. Armando, G. Pellegrino, and A. Vagati, [30] R. Schinzinger and P. A. Laura, Conformal Mapping: Methods and
“Magnet minimization in IPM-PMASR motor design for wide speed Applications. Mineola, NY, USA: Courier Corporation, 2003.
range application,” in Proc. IEEE Energy Convers. Congr. Expo., [31] J. W. Nilsson, S. A. Riedel, Electric Circuits, 9th ed. Upper Saddle
Sep. 2011, pp. 4201–4207. River, NJ, USA: Prentice-Hall, 2011.
[8] T. A. Lipo, “Synchronous reluctance machines-a viable alternative for [32] A. Vagati, M. Pastorelli, G. Francheschini, and S. C. Petrache, “Design
AC drives,” Electr. Mach. Power Syst., vol. 19, no. 6, pp. 659–671, 1991. of low torque ripple synchronous reluctance motors,” IEEE Trans. Ind.
[9] S. Jurkovic, K. M. Rahman, Q. Niu, and P. J. Savagian, “Interior Appl., vol. 34, no. 4, pp. 758–765, July/Aug. 1998.
permanent magnet machine,” U.S. Patent 9 118 230, Jul. 1, 2015. [33] N. Bianchi and H. Mahmoud, “An analytical approach to design the PM
in PMAREL motors robust toward the demagnetization,” IEEE Trans.
[10] E. Howard, M. J. Kamper, and S. Gerber, “Flux barrier and skew design
Energy Convers., vol. 31, no. 2, pp. 800–809, Jun. 2016.
optimisation of reluctance synchronous machines,” in Proc. Int. Conf.
[34] D. Wang, X. Wang, and S.-Y. Jung, “Cogging torque minimization
Elect. Mach. (ICEM), Sep. 2014, pp. 1186–1192.
and torque ripple suppression in surface-mounted permanent magnet
[11] S. Yammine, C. Henaux, M. Fadel, S. Desharnais, and L. Calégari, synchronous machines using different magnet widths,” IEEE Trans.
“Synchronous reluctance machine flux barrier design based on the flux Magn., vol. 49, no. 5, pp. 2295–2298, May 2013.
line patterns in a solid rotor,” in Proc. Int. Conf. Elect. Mach. (ICEM), [35] N. Bianchi, H. Mahmoud, and S. Bolognani, “Fast synthesis of per-
Sep. 2014, pp. 297–302. manent magnet assisted synchronous reluctance motors,” IET Electric
[12] E. E. Costamagna and P. A. Di Barba Savini, “Conformal mapping Power Appl., vol. 10, no. 5, pp. 312–318, 2016.
of doubly connected domains: An application to the modelling of [36] R.-R. Moghaddam, F. Magnussen, and C. Sadarangani, “Novel rotor
an electrostatic micromotor,” IET Sci., Meas. Technol., vol. 3, no. 5, design optimization of synchronous reluctance machine for low torque
pp. 334–342, 2009. ripple,” in Proc. XXth Int. Conf. Elect. Mach., Sep. 2012, pp. 720–724.
[13] A. T. Driscoll and N. T. Lloyd, Schwarz-Christoffel Mapping, Vol. 8. [37] T. Matsuo and T. A. Lipo, “Rotor design optimization of synchronous
Cambridge, U.K.: Cambridge Univ. Press, 2002. reluctance machine,” IEEE Trans. Energy Convers., vol. 9, no. 2,
[14] K. Abbaszadeh and F. R. Alam, “On-load field component separation pp. 359–365, Jun. 1994.
in surface-mounted permanent-magnet motors using an improved con- [38] R. K. Roy, Design of Experiments Using the Taguchi Approach: 16 Steps
formal mapping method,” IEEE Trans. Magn., vol. 52, no. 2, pp. 1–12, to Product and Process Improvement, Hoboken, NJ, USA: Wiley, 2001.
Feb. 2016.
[15] K. Ramakrishnan, “Improved method for field analysis of surface
permanent magnet machines using Schwarz-Christoffel transformation,”
IET Electr. Power Appl., vol. 11, no. 6, pp. 1067–1075, 2017.
[16] N. Bianchi, S. Bolognani, D. Bon, and M. D. Pre, “Rotor flux-barrier
design for torque ripple reduction in synchronous reluctance and PM- Maedeh Sadat Mirazimi was born in Isfahan, Iran,
assisted synchronous reluctance motors,” IEEE Trans. Ind. Appl., vol. 45, in 1987. She received the Ph.D. degree in electrical
no. 3, pp. 921–928, May/Jun. 2009. power engineering from the University of Isfahan,
[17] R.-R. Moghaddam and F. Gyllensten, “Novel high-performance SynRM Isfahan, Iran, in 2018.
design method: An easy approach for a complicated rotor topol- Her current research interests include the applica-
ogy,” IEEE Trans. Ind. Electron., vol. 61, no. 9, pp. 5058–5065, tion of time-stepping finite-element analysis (FEA)
Sep. 2014. in electromagnetic and electrical machines, and
[18] A. Hanic, Z. Hanic, M. Vrazic, and D. Zarko, “Calculation of no- design and optimization of permanent-magnet syn-
load saturation curve for wound field synchronous machines based chronous machines.
on magnetic equivalent circuits and conformal mapping,” in Proc. Int.
Aegean Conf. Elect. Mach. Power Electron. (ACEMP), Int. Conf. Optim.
Elect. Electron. Equip. (OPTIM) Int. Symp. Adv. Electromech. Motion
Syst., Sep. 2015, pp. 513–518.
[19] A. Hanic, D. Zarko, D. Kuhinek, and Z. Hanic, “On-load analysis of
saturated surface permanent magnet machines using conformal mapping Arash Kiyoumarsi was born in Shahr-e-Kord, Iran,
and magnetic equivalent circuits,” IEEE Trans. Energy Convers., vol. 33, in September 1972. He received the B.Sc. degree
no. 3, pp. 915–924, Sep. 2018. (Hons.) in electronics engineering from the Petro-
[20] A. Tessarolo, “Modeling and analysis of synchronous reluctance leum University of Technology (PUT), Abadan, Iran,
machines with circular flux barriers through conformal mapping,” IEEE in 1995, and the M.Sc. and Ph.D. degrees in electri-
Trans. Magn., vol. 51, no. 4, pp. 1–11, Apr. 2015. cal power engineering from the Isfahan University of
[21] Z. Q. Zhu, D. Howe, E. Bolte, and B. Ackermann, “Instantaneous Technology (IUT), Isfahan, Iran, in 1998 and 2004,
magnetic field distribution in brushless permanent magnet DC motors. respectively.
Part I: Open-circuit field,” IEEE Trans. Magn., vol. 29, no. 1, In March 2005, he joined the Department of Elec-
pp. 124–135, Jan. 1993. trical Engineering, Faculty of Engineering, Univer-
[22] Z. Q. Zhu and D. Howe, “Instantaneous magnetic field distribution in sity of Isfahan, as an Assistant Professor of electrical
permanent magnet brushless DC motors. Part IV: Magnetic field on machines. He was a Post-Doctoral Research Fellow of the Alexander-von-
load,” IEEE Trans. Magn., vol. 29, no. 1, pp. 152–158, Jan. 1993. Humboldt Foundation, Institute of Electrical Machines, Technical University
[23] D. Zarko, “A systematic approach to optimized design of permanent of Berlin, Berlin, Germany, from February to October 2006 and July to August
magnet motors with reduced torque pulsations,” Ph.D. dissertation, Dept. 2007. He was also a Visiting Guest Professor with the Institute of Electrical
Elect. Comput. Eng., Univ. Wisconsin-Madison, Madison, Wisconsin, Machines, RWTH Aachen University, Aachen, Germany, in July 2014; the
2004. Department of Electrical and Computer Engineering, University of Manitoba,
[24] D. Zarko, D. Ban, and T. A. Lipo, “Analytical calculation of magnetic Winnipeg, MB, Canada, from April to September 2017; and the Institute
field distribution in the slotted air gap of a surface permanent-magnet of Electrical Energy Conversion (IEW), University of Stuttgart, Stuttgart,
motor using complex relative air-gap permeance,” IEEE Trans. Magn., Germany; from July to September 2019. In March 2012, he became an
vol. 42, no. 7, pp. 1828–1837, Jul. 2006. Associate Professor of electrical machines. His research interests include the
[25] D. Žarko, D. Ban, and T. A. Lipo, “Analytical solution for cogging application of time-stepping finite-element analysis (FEA) in electromagnetic
torque in surface permanent-magnet motors using conformal mapping,” and electrical machines, and interior permanent-magnet synchronous motor
IEEE Trans. Magn., vol. 44, no. 1, pp. 52–65, Jan. 2008. drive.
[26] I. Boldea, Reluctance Synchronous Machines and Drives, New York, Dr. Kiyoumarsi received the Best Teachers’ Award from the University of
NY, USA: Oxford, 1996. Isfahan in 2013.

Authorized licensed use limited to: Chung Yuan Christian University. Downloaded on February 02,2021 at 01:01:19 UTC from IEEE Xplore. Restrictions apply.