Design Methods of Transversally Laminated Synchronous ... · (the ratio of the d-axis inductance...
Transcript of Design Methods of Transversally Laminated Synchronous ... · (the ratio of the d-axis inductance...
![Page 1: Design Methods of Transversally Laminated Synchronous ... · (the ratio of the d-axis inductance and the q-axis inductance), and the rotor design has been the main issue. Generally](https://reader034.fdocuments.us/reader034/viewer/2022042016/5e74d1acb480d5292756b348/html5/thumbnails/1.jpg)
164 CES TRANSACTIONS ON ELECTRICAL MACHINES AND SYSTEMS, VOL. 1, NO. 2, JUNE 2017
Abstract—Transversally laminated synchronous reluctance
machine (SynRM) are usually designed with multiple-layer flux
barriers to achieve high electromagnetic performance. This paper
summarizes three design methods to optimize the machine.
Related implementation procedures are detailed. Besides,
advantages and disadvantages of these methods are discussed.
Based on these conventional techniques, a comprehensive
optimization method is proposed, with which a prototype SynRM
is designed. The performances of this prototype are discussed to
verify the optimal design method.
Index Terms—Design method, flux barrier, synchronous
reluctance machine, torque ripple.
I. INTRODUCTION
YNCHRONOUS reluctance machine (SynRM) has attracted
much attention recently, thanks to the advantages of low
cost, robust structure and high torque density [1-3]. Compared
with induction machines (IM), the rotor of SynRM is simpler to
manufacture and produces less losses. In [4], testing with the
same speed and load torque, the temperature rise of an IM is
much higher than a SynRM. As for permanent magnet
synchronous machine (PMSM), the cost of SynRM is much
lower, and, problems such as the magnets demagnetization and
the performance variation with temperature rise do not exist. In
[5], a SynRM is optimized with the same volume of a ferrite
PMSM, and their electromagnetic torque is similar. Besides,
torque ripple of the SynRM is usually lower than that of the
switched reluctance machine (SRM), therefore, producing less
vibration and noise [6].
Define d-axis as the main flux path, namely the smaller
reluctance direction, and q-axis as the larger reluctance
direction. The SynRM is working based on the rotor saliency
(the ratio of the d-axis inductance and the q-axis inductance),
and the rotor design has been the main issue. Generally
speaking, the SynRM has three rotor topologies, namely the
physically salient pole (SP) structure, the transversally
laminated anisotropy (TLA) structure, and the axially
laminated anisotropy (ALA) structure [7]. The SP rotor in Fig.
1(a) has been perceived to be the simplest to manufacture. The
saliency is gained by cutting q-axis magnetic lamination, but
the pole arc of d-axis is shrunk meanwhile. Therefore, saliency
Shun Cai, Jian-Xin Shen, He Hao, and Meng-Jia Jin are with the Department
of Electrical Engineering, Zhejiang University, Hangzhou, China(e-mail:
[email protected], [email protected], [email protected], [email protected]).
of the SP rotor is the lowest among these three structures [8].
The saliency of the TLA rotor in Fig. 1(b) is produced by
punching several layers of air flux barrier in the q-axis. To
maintain rotor mechanical stress, ribs are preserved to connect
the adjacent iron segments, but the ribs cause flux leakage. It is
not difficult to manufacture the TLA rotor, and the
electromagnetic performances are better compared with the SP
rotor. The ALA rotor in Fig. 1 (c) is assembled by stacking
axial lamination layers and axial insulation layers alternatively
[9]. The saliency is usually considered to be the highest, for
example, in [10] the saliency ratio (i.e., Ld/Lq) of 10~20 is
achieved. However, there are mainly two problems about the
ALA rotor. First, the rotor structure is very complicated and the
cost of manufacturing and assembling is high. Second,
mechanically the rotor is not strong and can not operate at high
speed [11]. Therefore, the most promising structure for mass
production in industry application is the TLA rotor.
When the number of flux barrier layers for the TLA rotor is
small, the thickness of the iron segments is large and flux lines
in the q-axis can still pass along the iron segments. The flux
barriers can not separate the q-axis flux line adequately, and
thus the rotor saliency is decreased. It has been demonstrated
that the performance of SynRM with the TLA rotor can be
improved when adopting multiple-layer flux barriers [1], [11].
However, the parameters to be optimized become enormous
and are dependent on each other, making the optimization
procedure complex.
Now, finite element method (FEM) has been widely used to
optimize the electrical machines for accurate calculation and
accepTABLE time consumption [3]. If FEM is utilized to
optimize the motor parameters one by one, the final results may
be not optimal and sometimes even poor. If FEM is applied to
optimize the parameters globally, the parameters form a huge
matrix and the optimization time is not accepTABLE.
Therefore, how to optimize the SynRM both accurately and
rapidly is a key issue to investigate.
The purpose of this paper is to introduce various methods
and propose a comprehensive method to design the
transversally laminated SynRM. First of all, this paper
summarizes the conventional methods to optimize the SynRM.
Implement procedures are stated, and both the advantages and
disadvantages of these method are compared. Then, a
comprehensive optimization method is proposed to design the
SynRM parameters preliminarily. This method is applied to
optimize a 5kW prototype SynRM. The design procedure is
detailed and the performances of the prototype are discussed to
Design Methods of Transversally Laminated
Synchronous Reluctance Machines
Shun Cai, Jianxin Shen, Senior Member, IEEE, He Hao, and Mengjia Jin
S
![Page 2: Design Methods of Transversally Laminated Synchronous ... · (the ratio of the d-axis inductance and the q-axis inductance), and the rotor design has been the main issue. Generally](https://reader034.fdocuments.us/reader034/viewer/2022042016/5e74d1acb480d5292756b348/html5/thumbnails/2.jpg)
CAI et al. : DESIGN METHODS OF TRANSVERSALLY LAMINATED SYNCHRONOUS RELUCTANCE MACHINES 165
validate this method.
(a)
(b)
(c)
Insulation Layer
Lamination Layer
Fig. 1. Three different rotor structures of SynRM. (a) SP Rotor. (b) TLA Rotor.
(c) ALA Rotor.
II. CONVENTIONAL DESIGN METHODS
Much research has been done to improve the calculation
speed and achieve high performances [3], [12-29]. In this
section, the conventional methods to optimize SynRM are
classified as three types.
A. Combination of FEM with Optimization Algorithm
One of the solutions to solve this problem is to combine the
FEM with various optimization algorithms [12-19]. By
utilizing a certain algorithm to search the potentially optimal
parameters according to the proceeding calculation results, the
parameter matrix to be solved can be simplified and the
calculation efficiency can be improved significantly.
In the first step, a parametric model is established, and the
parameters to be optimized are listed. The initial parameters are
given to calculate in the beginning. Take a rotor with two
U-shape (2U) flux barriers [14] as an example to analyze, the
rotor section is shown in Fig. 2. There are 8 parameters in the
rotor, and obviously, they are dependent on each other.
In the second step, the constraint conditions of the
parameters are set, including the parameter range conditions,
the geometry constraint conditions, and the electromagnetic
constraint conditions. The parameter range conditions include
the variation range of each parameter to be calculated, and these
1h2h4h
3h
4t3t
2t1t
Flux Barriers
q axis
d axisIron Segments
Fig. 2. Parametric modelling of rotor with 2U flux barriers.
determine the solving time. If the variation range is too large, it
may take a long time to converge. If the variation range is too
small, the final result may be not optimal. So it is important to
set a reasonable variation range. The geometry constraint
conditions mean the possible combinations of the geometric
structure. For the 2U flux barriers rotor of Fig. 2, the total
thickness of the two layers of flux barriers and three layers of
iron segments along the q-axis is equal to the difference
between the rotor outer radius and inner radius. Therefore, the
summary of h1~h4 is less than the difference of rotor outer
radius and inner radius. Similar constraint condition is also
suiTABLE for the t1~t4. When the above conditions are
satisfied, each combination of parameters can be corresponding
to a rotor structure. The rotor can be optimized but the
optimization time is very long. To reduce the optimization
matrix, the electromagnetic constraint conditions can be
attached. For the 2U flux barriers rotor of Fig. 2, the main flux
lines pass through the segments along the d-axis. Although the
leakage flux of the q-axis may shift the flux lines, the flux
linkage of the d-axis is much larger than that of the q-axis. To
distribute the flux density uniformly, the radial thickness of the
second iron segment (h3) should be similar to the tangential
thickness (t3). Similar constraint condition is also suiTABLE
for the other layers of flux barriers. By limiting the difference
and the ratio, the matrix to be optimized can be reduced
significantly.
In the third step, the objective functions, maximum iterations,
and the minimum convergence limit are set. The conventional
objective to design a machine is to ensure the torque density
and power density. Since the SynRM has inherent drawbacks of
poor power factor, and large torque ripple, the objective
function should be set with comprehensive weights of these
performances. The maximum iteration and minimum
convergence limit determine the calculating time and accuracy,
and ought to be coordinated properly.
In the last step, the computer calculates and outputs the
optimal parameter design. Firstly, the objective function of the
initial model is solved. Then the computer searches among the
parameter matrix to be optimized within the constraint
conditions, and calculates the objective of the new parameters
corresponding prototype. The objective function is compared
with the proceeding result until the maximum iteration or the
![Page 3: Design Methods of Transversally Laminated Synchronous ... · (the ratio of the d-axis inductance and the q-axis inductance), and the rotor design has been the main issue. Generally](https://reader034.fdocuments.us/reader034/viewer/2022042016/5e74d1acb480d5292756b348/html5/thumbnails/3.jpg)
166 CES TRANSACTIONS ON ELECTRICAL MACHINES AND SYSTEMS, VOL. 1, NO. 2, JUNE 2017
minimum convergence limit is reached. In the subsequent
calculation, the computer can utilize various optimization
algorithms to search potentially optimal parameters according
to the results calculated before, such as the genetic algorithm
[12-14], response surface methodology [15], [16], sequential
unconstrained minimization technique [17], Taguchi method
[18], particle swarm optimization technique [19] and so on. At
present, most commercial softwares have embedded these
optimization algorithms, and are capable to take multi-core
parallel calculating to reduce the calculation time.
According to the analysis above, the diagram of the
combination of FEM with optimization algorithms can be
described in Fig. 3.
As can be seen from Fig. 3, most of the work are undertaken
by the computer. The tasks for the designer are the parametric
modelling, and setting the objective functions, initial
parameters and constraint conditions. The professional
requirements for the designers are relatively low, and the
optimal design can be achieved as long as the conditions are set
reasonably.
The drawbacks of this method are the huge calculation task
for the computer and the long solving time. For the 2U flux
barriers rotor of Fig. 2, the number of parameters to be
optimized increases by 4 when the flux barrier increases by one
layer. At the same time, the dimension of the parameter matrix
increases, and the calculation time boosts. Therefore, with the
increases of flux barrier layers, the calculation efficiency
significantly drops, and it becomes much harder to achieve the
optimal design.
B. Analytical Method
Since the FEM is time-consuming, much research has been
done to seek for the replacement of FEM. In [24] and [30], an
analytical method is proposed based on the magnetic circuit
model to evaluate the SynRM with single- and double-layer
flux barriers, respectively. In [23], an analytical model is raised
based on the synthesis of d-axis and q-axis MMF effect. In [29],
a detailed model is discussed to solve the SynRM with
asymmetric flux barriers. The analytical calculation is much
faster compared with FEM. In [20], a saturating lumped
parameter model is presented to take the saturation into account.
In [21], a nonlinear reluctance network model is established to
evaluated the SynRM with a complex rotor structure. FEM is
applied in the beginning to divide the reluctance network
properly. Then the machine can be calculated and optimized
with the fast reluctance network method. Since the saturation
and iron losses are considered, the results can be satisfactory
even with high load [22].
In this part, an analytical method based on the magnetic
circuit model is introduced to evaluate the SynRM with
multiple-layer flux barriers. The calculation speed is very
advantageous, and the result coincide with the FEM when the
saturation is not severe [28].
The rotor topology discussed in this part is shown in Fig. 4.
Each flux barrier has an even thickness and so is with the iron
segments, in order to keep the flux density uniformly
distributed and to avoid local saturation. The number of flux
barrier layers is nb, the thickness of the kth layer flux barrier is
tbk, the circumferential length of the kth layer flux barrier is lbk,
and the position of the kth layer flux barrier end is bk. The flux
barrier from the outermost to the shaft is called as the 1st layer
flux barrier, the 2nd layer flux barrier, …, and the nbth layer
flux barrier.
Parameters to
Be Optimized
Reaching Maximum
Iterations or Minimum
Convergence Limits?
FEM
Calculation
Computer
Solving
Yes
No
Design
Requirements
Optimization
Finshed
Prototype with
Certain Parameters
Solving Objective
Functions
Setting Constraint
Conditions
Setting Initial
Parameters
Parametric
Modelling
Setting Objective
Functions
Searching
Parameters with
Algorithms
Fig. 3. Diagram of combination of FEM with optimization algorithm.
If the saturation in both the stator core and rotor core is
neglected, the magnetic circuit of one pole can be presented as
in Fig. 5, where Rbk is the reluctance of the kth flux barrier, Rgk
is the reluctance of the air-gap adjacent to the kth layer iron
segment, bk is the flux going through the kth layer flux barrier,
gk is the flux through the air-gap adjacent to the kth layer iron
segment, fsk is the average MMF of the stator adjacent to the kth
iron segment, and frk is the MMF of the kth rotor iron segment.
Take the kth layer flux barrier as an example, the flux
conservation can be expressed as
( 1)bk b k gk (1)
where bk and gk can be expressed as
sk rkgk
gk
f f
R
(2)
( 1)rk r k
bk
bk
f f
R
(3)
Define the angle between two flux barrier ends as
1k k k (4)
Then Rbk and Rgk can be expressed as
0
bkbk
bk ef
tR
l l (5)
0 22
ef
gkg
k ef
RD
l
(6)
where 0 is the permeability of vacuum, ef is the effective
air-gap length, lef is the effective axial length and Dg is the
diameter of air gap (the average of the rotor outer diameter and
stator inner diameter).
The three phase winding MMF in the stator frame can be
expressed as
![Page 4: Design Methods of Transversally Laminated Synchronous ... · (the ratio of the d-axis inductance and the q-axis inductance), and the rotor design has been the main issue. Generally](https://reader034.fdocuments.us/reader034/viewer/2022042016/5e74d1acb480d5292756b348/html5/thumbnails/4.jpg)
CAI et al. : DESIGN METHODS OF TRANSVERSALLY LAMINATED SYNCHRONOUS RELUCTANCE MACHINES 167
0, 1, 2,
6 1
3( ) cos( )
ip m w
s s s
i
N I kf p t
p
(7)
bkl
bkt
Iron Segments
Flux Barriers
bk
1st
kth
nbth
d-axis
q-axis
Fig. 4. Rotor parameters in one pole.
1sf
2sf
bsnf
1g
2g
bgn
1gR
2gR
bgnR
1bR2bRbbnR
1rf2rfbrnf1b2bbbn
Fig. 5. Magnetic circuit in one pole.
where Np is the number of coil turns per phase, Im is the
amplitude of stator current, p is the number of pole pairs, is
the order of spatial harmonics in stator current, kw is the
winding coefficient of the th harmonic, s is the position in the
stator frame, is the electrical angular speed and is the phase
angle of current.
By solving equation (1)~(7), the rotor MMF in each iron
segments is shown as (8), and the coefficients are listed as
(9)~(13).
0, 1, 2,
26 1
(cos )( )
iw
rk g k
i
kf D
p
(8)
( 1)2
pmt
(9)
1
, , 1 ,
1
( )sin( ) sin( )b
b
n
k k i k i bi k n bn
i
m m p m p
(10)
1
, ,
1
,1
,
; if
; if
b
b
n q
i j l q j
q i l i
i jn q
l q j
q j l i
b a b i j
m
a b i j
(11)
1 1
1
1
1 [ (( ) ) ]i i i
g bik j i
k jjef bi
aD t
al
(12)
, ;2
ig bi
i j k
k jef bi
D tb a i j
l
(13)
Since the stator MMF and rotor MMF have been expressed
in (7) and (8), the output torque can be calculated according to
[31], and expressed as (14).
2
0
0
( )( )
2
g ef s rr r r
ef r
pD l dfT f d
d
(14)
The advantages of analytical method are the fast speed. The
dimension of the coefficient matrix of (9)~(11) is nb×nb, which
is much smaller compared with the huge node matrix number of
the FEM. So the analytical method can be used to set the initial
parameters and make preparation for the later FEM precise
calculation.
However, the iron core saturation and stator slot opening
effect are neglected in the analytical model. Since the saturation
may reduce the inductance and induce the d-q axis inductance
variation with rotor position, the average torque and torque
ripple are deteriorated under saturation. The SynRM
performance is strongly sensitive to the saturation, therefore,
the analytical method can not provide very accurate result.
C. Combination of FEM with Optimization Algorithm
As has been stated, the method to combine the FEM with
optimization algorithm is very time-consuming whereas the
analytical method can not guarantee the accuracy. So some
researches have been focused on a compromised method to
achieve high performance with high optimization speed. Some
constraint conditions are attached to reduce the number of
parameters to be optimized, decouple the parameters, and
optimize them either one by one or globally [13].
One widely used constraint method is to regard the rotor flux
barriers as uniform slots on the rotor, and therefore the angles
between every two adjacent flux barrier ends are the same [32],
as shown in Fig. 6. The parameters to be optimized become
independent on each other and it is much easier to illustrate the
effect individually. There are generally two types of constraint
conditions as following.
One type is to propose a parameter called as the number of
pseudo barrier ends [31-33], as shown in Fig. 6(a). The number
of flux barrier layer is defined as nb, the angle of adjacent flux
barrier ends is in electrical rad. Then the pseudo flux barrier
ends in one pole-pair nr can be expressed as (15). Note that,
although there is no real flux barrier end beside the q-axis, the
rotor is still divided as several pseudo uniformly distributed
ends. Therefore, nr>2× (2nb-1). The pole arc angle of the first
real flux barrier 1 can be expressed as (16). The consequence
in Fig. 6(a) is nb=3, nr=18, otherwise the relationship is not
fixed as .
2
rnp
(15)
1
1
4 2
rb
nn
(16)
The other refinement technique is to define by the angles
between the adjacent flux barrier ends [27], as shown in Fig.
6(b). The angle between adjacent flux barrier ends is m, the
controlled angle of the first pseudo flux barrier is , and the
![Page 5: Design Methods of Transversally Laminated Synchronous ... · (the ratio of the d-axis inductance and the q-axis inductance), and the rotor design has been the main issue. Generally](https://reader034.fdocuments.us/reader034/viewer/2022042016/5e74d1acb480d5292756b348/html5/thumbnails/5.jpg)
168 CES TRANSACTIONS ON ELECTRICAL MACHINES AND SYSTEMS, VOL. 1, NO. 2, JUNE 2017
parameters satisfy the equation (17).
1
( )2 2
b mnp
(17)
(b) Defined by the angle of
adjacent flux barrier ends
(nb=3)
1
2m
1
2m
m
m
m
m
m
m
Iron Segments
Flux Barriers
d axis
q axis
1
2
1
2
1
Iron Segments
Flux Barriers
d axis
q axis
(a) Defined by the number of
pseudo barrier ends
(nb=3,nr=18)
Fig. 6. Uniformly distributed flux barrier rotor.
The former constraint condition defines the pseudo flux
barrier ends in one pole-pair (nr) as an integer. Therefore the
pole arc angle of the first real flux barrier 1 must be the
multiplier of the angle of adjacent flux barrier ends , and the
constraint conditions are more strict. When the nr is extended to
the decimal, the two constraint conditions, in Fig. 6 (a) and (b)
are actually the same.
It is generally believed that there are three main parameters
for the SynRM, viz, the number of flux barrier layers, the
position of flux barrier ends, and the thickness of flux barriers
[1], [11], [34]. The number of flux barrier layers is defined by
nb, the position of flux barrier ends is defined by the pseudo flux
barrier ends in one pole-pair nr or the angle between adjacent
flux barrier ends (m), the thickness of flux barriers is defined
by the insulation ratio [1], [27], as shown in (18).
ins
Widthof fluxbarrierk
Widthof ironsegment (18)
According to the parameters defined above, the SynRM can
be optimized with these three main decoupled parameters. Note
that, the insulation ratio is corresponding to the the real flux
barrier. For the pseudo flux barrier in Fig. 6(a), the insulation
ratio kins is zero.
In [32], the influences of these parameters are analyzed
theoretically in accordance with a magnetic circuit model. In
[26] and [25], the parameters are optimized for the maximum
torque density and minimum torque ripple, respectively, based
on ideal sinusoidal-wave MMF. In [2], the design principle of
nr to achieve low torque ripple is proposed. In [35], the design
principle of nr is illustrated further with FEM. In [36], the
optimization rule of nr is concluded with the combination of the
number of flux barrier layers and the stator winding
configuration. In [37] and [38], the effect of the three main
parameters are investigated on the basis of mass finite element
calculation.
As can be seen from above analysis, the parameters of the
original SynRM model are dependent on each other and the
matrix to be optimized is huge. With the constraint conditions,
the number of parameters to be optimized decreases
significantly. It becomes much simpler to investigate the effect
of each parameter and more efficient to design the SynRM.
Nevertheless, the theoretical constraint conditions cannot
take the nonlinear problem into consideration. Therefore, the
optimized prototype with this method may be not optimal.
Besides, with the increase of constraint conditions, the
calculation speed can be improved at the sacrifice of
comprehensive performance.
III. A COMPREHENSIVE OPTIMIZATION METHOD
On the basis of the three conventional design methods, this
section proposes a comprehensive optimization method. This
method can help predict the performance rapidly, and set the
initial value of the parameters. The influences of each
parameter is illustrated to achieve fast design procedure.
When the air-gap length is relatively small and the diameter
of air-gap (namely the average of the stator inner diameter and
the rotor outer diameter), (19) is satisfied and the MMF drop on
the air-gap can be neglected compared with that on the rotor
flux barriers.
2
g bk
ef bk
D l
t (19)
Since only the fundamental of the stator MMF rotates at the
same speed as the rotor and produces steady MMF drop on the
rotor flux barriers, only the fundamental of stator MMF is
considered hereafter when analyzing the rotor sTABLE MMF.
For the constraint conditions in Fig. 6(a), the rotor MMF and
the fundamental of stator MMF are shown in Fig. 7, where i is
the angle between the stator current vector and the d-axis. Note
that the consequence in Fig. 7 is when the current vector aligns
with the q-axis. Otherwise the fundamental of stator MMF may
shift whereas the rotor MMF is always symmetric along the
d-axis and q-axis.
The rotor stair MMF can be obtained from the fundamental
stator MMF, and the electromagnetic torque can be calculated
in (14). The average torque is shown in (20), where kr is only
dependent on the rotor parameter and called as rotor coefficient,
as shown in (21).
2
0 19sin 2
g ef p m
ave r
ef
D l N I kT k
p
(20)
1
1rf
brnf
2rf
r
f
rf
0 ( )2
s if
2 p
0
.
.
.
q axis d axis
Fig. 7. Ideal MMF distribution in rotor coordinate.
![Page 6: Design Methods of Transversally Laminated Synchronous ... · (the ratio of the d-axis inductance and the q-axis inductance), and the rotor design has been the main issue. Generally](https://reader034.fdocuments.us/reader034/viewer/2022042016/5e74d1acb480d5292756b348/html5/thumbnails/6.jpg)
CAI et al. : DESIGN METHODS OF TRANSVERSALLY LAMINATED SYNCHRONOUS RELUCTANCE MACHINES 169
2
2
( 4 2)2 sin
2
( 4 2)
2( 1)sin sin
( 2 )1 cos
2sin
r br
rr
r b
br
r br rb
r
r
n nn
nk
n n
nn
n nn nn
n
n
(21)
As can be seen from (20), the ideal average torque is
proportional to 1/p. However, when the number of pole-pairs is
small, the yoke thickness is large, thus the slot area to contain
the stator windings is reduced. Besides, a small number of
pole-pairs means relatively long end-winding, and both the
volume and weight are increased. So the number of pole-pairs
for the SynRM is usually selected as 2 or 3.
Define split ratio as the ratio of the stator inner diameter to
the outer diameter. As can be seen from (20), the average torque
is proportional to the square of the current Ampere-turns.
Therefore, the SynRM needs to have sufficient slot space to
contain the stator winding, thus the split ratio is usually smaller
than that of the PMSM [39]. However, if the split ratio is too
small, the rotor can not produce sufficient saliency. Besides, the
split ratio is also determined by the number of pole-pairs.
Generally speaking, a small number of pole-pairs means a
thicker yoke and a smaller split ratio [36].
As for the winding configuration, the average torque is
proportional to the square of fundamental winding coefficient,
as shown in (20), therefore, the winding configuration ought to
exhibit high fundamental coefficient. Besides, the
electromagnetic torque originates from the interaction of the
stator MMF and rotor MMF, as shown in (14). So when the
harmonic order of the stator MMF coincides with the rotor
MMF, certain harmonic torque ripple is produced. As can be
seen from Fig. 7, when the rotor flux barrier is symmetric, all
the harmonic orders of the rotor MMF are odd. So, the low odd
harmonics in the stator MMF should be avoided.
The main harmonic order for the stair rotor MMF in Fig. 7 is
k1×nr1, where nr1 is the odd number that is nearest to nr, and k1
is an integer. The main odd harmonic order for the stator MMF
is k2×Ns/p±1, where Ns is the number of stator slots and k2 is
the integer that ensures the harmonic order is odd. When the
winding configuration is chosen, selection of nr should avoid
the main harmonic order of stator MMF to achieve low torque
ripple.
Relationship of the rotor coefficient kr with nb and nr is
shown in Fig. 8 according to (21). With the increase of flux
barrier layer number, the rotor coefficient boosts. But when the
number of flux barriers is nb>4, the rotor coefficient hardly
increases with the number of flux barriers. Considered the
manufacturing and processing technique, the number of flux
barriers layer should not be too large. And the selection of nb
should also be combined with that of the nr to achieve a large kr.
With the above analysis, the number of pole-pairs, split ratio,
winding configuration, number of flux barrier layers, and flux
barrier distribution can all be determined preliminarily. Then,
the initial model of the SynRM can be established, and the
parameters can be further optimized with FEM.
The comprehensive optimization procedures are described in
Fig. 9. The preliminary design is largely based on the designer's
experience. The computer calculation includes the optimization
of stator core and rotor flux barriers, and the calculating time
can be reduced significantly. The computation can make up for
the limitation of the designer's experience.
The advantage lays on the high efficiency to design the
SynRM. The parameters are decoupled and can be optimized
one by one. The optimization procedures do not take long time,
and, generally high performance of the SynRM can be
achieved.
However, the optimization result relies on professional
knowledge of the designer and require abundant experience.
Besides, since the constraint conditions are proposed based on
the ideal model, the final prototype may be not optimal.
IV. DESIGN EXAMPLE
In this section, a prototype is optimized with the
comprehensive method proposed above. The design procedures
are detailed according to the diagram shown in Fig. 9. The
performances of this machine are discussed to verify this design
method.
A. Motor Specification
The specification and design requirements of the prototype
SynRM are listed in TABLE I. To satisfy the power density, the
average torque should be maximized. To reduce the vibration
and noise during operation, the torque ripple should be
minimized.
B. Design Procedure
To improve the rotor saliency and reduce the end-winding
length, the number of pole-pairs is selected as 3. Besides, the
split ratio of SynRM is smaller than that of PMSM, and the split
ratio is selected as 0.55 preliminarily when the number of
pole-pairs is 3.
Since the number of pole-pairs is small, modular
concentrated winding of which the numbers of slots and poles
satisfying Ns=2p±1 or Ns=2p±2 do not exist. To improve the
fundamental winding coefficient, the conventional
concentrated winding of which the numbers of slots and poles
satisfying Ns/2p=3/2 is not considered. Therefore, distributed
winding is rather common for the SynRM. For integer-slot
distributed winding, the order of the main harmonic k2×
Ns/p±1 is odd when k2 is an integer. For fractional-slot
distributed winding, the main odd harmonic order is generally
larger. For example, for the Ns/2p=9/2 configuration, the main
harmonic k2×Ns/p±1 is odd only when k2 is an even number.
Therefore, 6-poles 27-slots configuration is selected for the
prototype.
The main odd harmonic order in the stator MMF is 18k±1,
where k is an integer. So the selection of nr should avoid the
multiplier of 18. In accordance with Fig. 8 to gain high rotor
coefficient, the rotor main parameters are selected as nb=5,
nr=29.
![Page 7: Design Methods of Transversally Laminated Synchronous ... · (the ratio of the d-axis inductance and the q-axis inductance), and the rotor design has been the main issue. Generally](https://reader034.fdocuments.us/reader034/viewer/2022042016/5e74d1acb480d5292756b348/html5/thumbnails/7.jpg)
170 CES TRANSACTIONS ON ELECTRICAL MACHINES AND SYSTEMS, VOL. 1, NO. 2, JUNE 2017
After the basic design, the initial model of the prototype
SynRM is established. The other parameters such as the stator
tooth thickness, stator yoke thickness and rotor barrier
thickness can be optimized with FEM. Besides, the mechanical
strength should be enhanced to guarantee the tensile strength at
the maximum speed. The optimized prototype is shown in Fig.
10.
Fig. 8. Relationship of rotor coefficient kr with nb and nr.
Number of
pole-pairs
(2~3)
Initial Split Ratio
(0.5~0.6)
Winding Configuration
(Large Fundamental Coefficient
and High Harmonic Order)
Stator Core
Optimization
Selection of nr
(Avoiding the Harmonics
of Stator MMF)
Selection of nb
(Combination with nr to
Produce Large kr)
Rotor Flux Barrier
Optimization
Computer FEM OptimizationDesign by Experience
Optimization
Finished
Fig. 9. Diagram of the comprehensive optimization method.
TABLE I MOTOR SPECIFICATION AND DESIGN REQUIREMENT
Rated Power 5kW
Rated Speed 3000rpm
Maximum Speed 5000rpm
Winding Connection Y-Connected
Rated Line-Line Voltage 380V
Rated Current Density 5.5A/mm2
Stator Outer Diameter 181mm
Shaft Diameter 55mm
Core Axial Length 75mm
Slot Filling Factor 65%
C. Performance Discussion
The flux density distribution under the rated operation
condition is shown in Fig. 11, and the flux density variation
with rotor position are detailed in Fig. 12.
A+
A-
A-
A+
A-
A-A+
A+
A+
A-
A+
A-
A+
A+A+
A-
A-A-
B+
B+
B+
B+B+
B+
B+
B+B+
B-
B-B-
B- B-
B-
B-
B-
B-
C-
C+
C+C+
C+
C+C+
C+
C+C+
C-
C-
C-C-
C-
C-
C-C-
Fig. 10. Cross section of optimized prototype.
The maximum flux density in the stator tooth, stator yoke,
and rotor iron segment is about 1.6T, close to the knee point of
the lamination BH-Curve. It can help ensure the material
utilization but meanwhile avoid saturation. The radial flux
density in the stator yoke is nearly zero, whereas the tangential
flux density almost only contains the fundamental. The
tangential flux density of stator tooth is nearly zero while the
rial flux density almost only contains the fundamental. The
tangential flux density outweighs the radial flux density in both
the rotor iron segment and rotor bridge, besides, the tangential
flux density contains abundant 9th order harmonic.
The excitation field of SynRM is established by the stator
armature current. When the rotor iron segment is aligned with
the stator tooth, the reluctance is small and the flux lines pass
through that layer of iron segment. When the rotor iron segment
is aligned with the stator slot, the reluctance increases and the
flux lines pass through the adjacent layer of iron segment.
Therefore, the stator slots influence the flux density distribution
in the rotor and presents the tooth harmonic in the iron segment.
The numbers of pole-pairs and stator slots are 3 and 27. So the
main harmonic order in the rotor iron segment is 27/3=9th.
The SynRM has its rotor flux bridges saturated in order to
reduce the inductance of q-axis and produce the rotor saliency.
But the saturation point in the rotor flux bridge is not fixed. The
saturation point may shift because of the stator slots, and thus,
the 9th order harmonic in the flux bridge increases.
The d-axis inductance and q-axis inductance variations with
the d-axis current and q-axis current are shown in Fig. 13,
where the current is normalized by the rated current. The q-axis
magnetic circuit is saturated to reduce the q-axis linkage flux.
When the current amplitude is small, the rotor flux bridge is not
saturated and the q-axis inductance is relatively large. With the
increase of current, the rotor flux bridges become saturated.
![Page 8: Design Methods of Transversally Laminated Synchronous ... · (the ratio of the d-axis inductance and the q-axis inductance), and the rotor design has been the main issue. Generally](https://reader034.fdocuments.us/reader034/viewer/2022042016/5e74d1acb480d5292756b348/html5/thumbnails/8.jpg)
CAI et al. : DESIGN METHODS OF TRANSVERSALLY LAMINATED SYNCHRONOUS RELUCTANCE MACHINES 171
Stator Yoke
B (T)
0.00.20.40.60.81.01.21.41.61.82.02.22.4
Stator Tooth
Rotor
Flux Bridge
Rotor
Iron Segment
Fig. 11. Flux density distribution under rated operation condition.
0 60 120 180 240 300 360-2
-1
0
1
2
Rotor Position (Ele-deg)
Flu
x D
ensi
ty (
T) Stator Yoke
Radial
Tangential
0 60 120 180 240 300 360-2
-1
0
1
2
Rotor Position (Ele-deg)
Flu
x D
ensi
ty (
T) Stator Tooth
0 60 120 180 240 300 360-2
-1
0
1
2
Rotor Position (Ele-deg)
Flu
x D
ensi
ty (
T) Rotor Iron Segment
0 60 120 180 240 300 360-3-2-10123
Rotor Position (Ele-deg)
Flu
x D
ensi
ty (
T) Rotor Flux Bridge
Fig. 12. Stator and rotor core flux density variation with rotor position under
rated operation condition.
The q-axis inductance is then reduced and hardly varies with
the current. Comparatively, the d-axis magnetic circuit is less
saturated and the inductance varies with the current amplitude.
The saturation decreases the d-axis inductance as well as the
rotor saliency. Under the rated operation condition, the d-axis
inductance and q-axis inductance are 19.5mH and 6.2mH,
respectively, hence, the rotor saliency is 3.1.
The torque waveform with the maximum torque per ampere
(MTPA) control is shown in Fig. 14. The average torque and
torque ripple are listed in TABLE II, where the torque ripple is
defined as the peak-to-peak value of the electromagnetic torque.
The ratio of torque ripple to average torque at rated working
point is about 2%. The fluctuating torque is thus small, and the
prototype can operate steadily.
00.2
0.40.6
0.81
00.2
0.40.6
0.815
10152025
id
iq
Ld (
mH
)
00.2
0.40.6
0.81
00.2
0.40.6
0.81
5
10
id
iq
Lq (
mH
)
Fig. 13. d-axis inductance and q-axis inductance variation with d-axis current
and q-axis current.
0 60 120 180 240 300 3600
5
10
15
20
I=0.2Im
I=Im
Rotor Position (Ele-deg)
To
rqu
e (N
.m)
Fig. 14. Electromagnetic torque with MTPA control.
TABLE II TORQUE PERFORMANCE UNDER MTPA CONTROL WITH DIFFERENT CURRENT
AMPLITUDE
Current Amplitude Average Torque Torque Ripple
0.2 0.8Nm 0.08Nm
0.4 3.5Nm 0.28Nm
0.6 7.9Nm 0.34Nm
0.8 12.7Nm 0.31Nm
1 17,2Nm 0.36Nm
V. CONCLUSION
SynRM has been attractive alternative in some traditional
![Page 9: Design Methods of Transversally Laminated Synchronous ... · (the ratio of the d-axis inductance and the q-axis inductance), and the rotor design has been the main issue. Generally](https://reader034.fdocuments.us/reader034/viewer/2022042016/5e74d1acb480d5292756b348/html5/thumbnails/9.jpg)
172 CES TRANSACTIONS ON ELECTRICAL MACHINES AND SYSTEMS, VOL. 1, NO. 2, JUNE 2017
AC drive applications. Especially the TLA rotor structure, with
the combination of excellent electromagnetic performance as
well as simple manufacturing process, is perceived to be the
most promising in mass industrial production. The
electromagnetic performance of the TLA SynRM is high under
the condition that the rotor has multiple-layer flux barriers.
However, the parameters to be optimized become enormous
and are dependent on each other when the number of flux
barriers is large, making the optimization procedure rather
complex.
This paper summarized three conventional methods to
optimize the SynRM. The first method is to combine FEM with
the optimization algorithm. This method can achieve optimal
design on the sacrifice of large calculating time. The second
method is to utilize the fast analytical method to replace the
precise FEM. This method can improve the calculating speed
significantly, whereas the calculation result is not accurate with
the core saturation. The last method is to attach constraint
conditions to refine the parameter relationship. The constraint
condition can help reduce the number of parameters and save
the calculation time. However, the final prototype may be not
optimal under these constraint conditions.
Based on the conventional techniques, this paper proposed a
comprehensive optimization method to design the SynRM. The
initial parameters are selected according to the ideal model.
Then FEM is applied to make precise calculation and
optimization. Since the parameters are decoupled and
preliminarily optimized, it is easy and fast to obtain high
performance of the motor.
A prototype is designed with this comprehensive method for
the high torque density and low torque ripple. The
implementation procedures are detailed to demonstrate the
method. Finally, the machine performances are discussed to
ensure that the prototype can operate steadily.
REFERENCES
[1] T. Matsuo and T. Lipo, "Rotor design optimization of synchronous reluctance machine," Energy Conversion, IEEE Transactions on, vol.
9, pp. 359-365, 1994.
[2] A. Vagati, M. Pastorelli, G. Franceschini, and C. Petrache, "Design of low-torque-ripple synchronous reluctance motors," in Industry
Applications Conference, 1997. Thirty-Second IAS Annual Meeting,
IAS'97., Conference Record of the 1997 IEEE, 1997, pp. 286-293. [3] A. Vagati, A. Canova, M. Chiampi, M. Pastorelli, and M. Repetto,
"Design refinement of synchronous reluctance motors through
finite-element analysis," Industry Applications, IEEE Transactions on, vol. 36, pp. 1094-1102, 2000.
[4] A. Boglietti, A. Cavagnino, M. Pastorelli, and A. Vagati,
"Experimental comparison of induction and synchronous reluctance motors performance," in Industry Applications Conference, 2005.
Fourtieth IAS Annual Meeting. Conference Record of the 2005, 2005,
pp. 474-479. [5] S. Cai, M.-J. Jin, H. Hao, and J.-X. Shen, "Comparative study on
synchronous reluctance and PM machines," COMPEL: The
International Journal for Computation and Mathematics in Electrical and Electronic Engineering, vol. 35, pp. 607-623, 2016.
[6] X. Liu, Z. Pan, and Z. Zhu, "Analysis of average torque in switched
reluctance motor with unipolar and bipolar excitations based on an improved Fourier series model," in 2010 IEEE Vehicle Power and
Propulsion Conference, 2010, pp. 1-6.
[7] J. Kolehmainen, "Synchronous reluctance motor with form blocked rotor," Energy Conversion, IEEE Transactions on, vol. 25, pp.
450-456, 2010.
[8] S. Tahi, R. Ibtiouen, and S. Mekhtoub, "Performance optimization of
synchronous reluctance machines with two rotor structures," in
Electrical Machines (ICEM), 2014 International Conference on, 2014,
pp. 250-255. [9] G. Chenglin and L. Langru, "Calculation of magnetic field, parameters
and performance for axially-laminated anisotropic rotor reluctance
machines," Proceedings of the CSEE, vol. 14, pp. 14-20, 1994. [10] I. Boldea, Z. Fu, and S. Nasar, "Performance evaluation of
axially-laminated anisotropic (ALA) rotor reluctance synchronous
motors," Industry Applications, IEEE Transactions on, vol. 30, pp. 977-985, 1994.
[11] K. Wang, Z. Zhu, G. Ombach, M. Koch, S. Zhang, and J. Xu, "Optimal
slot/pole and flux-barrier layer number combinations for synchronous reluctance machines," in Ecological Vehicles and Renewable Energies
(EVER), 2013 8th International Conference and Exhibition on, 2013,
pp. 1-8. [12] M. Gamba, G. Pellegrino, and F. Cupertino, "Optimal number of rotor
parameters for the automatic design of Synchronous Reluctance
machines," in Electrical Machines (ICEM), 2014 International Conference on, 2014, pp. 1334-1340.
[13] Y. Guan, Z. Zhu, I. Afinowi, J. Mipo, and P. Farah, "Design of
synchronous reluctance and permanent magnet synchronous reluctance machines for electric vehicle application," in Electrical
Machines and Systems (ICEMS), 2014 17th International Conference
on, 2014, pp. 1853-1859. [14] G. Pellegrino, F. Cupertino, and C. Gerada, "Automatic Design of
Synchronous Reluctance Motors Focusing on Barrier Shape Optimization," Industry Applications, IEEE Transactions on, vol. 51,
pp. 1465-1474, 2015.
[15] J. H. Lee, I. K. Lee, Y. H. Cho, and T. W. Yun, "Characteristics analysis and optimum design of anisotropy rotor synchronous
reluctance motor using coupled finite element method and response
surface methodology," Magnetics, IEEE Transactions on, vol. 45, pp.
4696-4699, 2009.
[16] J. M. Park, S. I. Kim, J. P. Hong, and J. H. Lee, "Rotor design on torque
ripple reduction for a synchronous reluctance motor with concentrated winding using response surface methodology," Magnetics, IEEE
Transactions on, vol. 42, pp. 3479-3481, 2006.
[17] S. B. Kwon, S. J. Park, and J. H. Lee, "Optimum design criteria based on the rated Watt of a synchronous reluctance motor using a coupled
FEM and SUMT," Magnetics, IEEE Transactions on, vol. 41, pp.
3970-3972, 2005. [18] H. Azizi and A. VAHEDI, "Sensitivity analysis and optimum design
for the stator of synchronous reluctance machines using the coupled
finite element and Taguchi methods," Turkish Journal of Electrical Engineering and Computer Science, vol. 23, pp. 38-51, 2015.
[19] E. Aydin, Y. Li, I. Aydin, M. Timur Aydemir, and B. Sarlioglu,
"Minimization of torque ripples of interior permanent magnet synchronous motors by particle swarm optimization technique," in
Transportation Electrification Conference and Expo (ITEC), 2015
IEEE, 2015, pp. 1-6. [20] E. C. Lovelace, T. M. Jahns, and J. H. Lang, "A saturating
lumped-parameter model for an interior PM synchronous machine,"
IEEE Transactions on Industry Applications, vol. 38, pp. 645-650, 2002.
[21] T. Raminosoa, I. Rasoanarivo, and F.-M. Sargos, "Reluctance network
analysis of high power synchronous reluctance motor with saturation and iron losses considerations," in Power Electronics and Motion
Control Conference, 2006. EPE-PEMC 2006. 12th International,
2006, pp. 1052-1057. [22] T. Raminosoa, I. Rasoanarivo, F.-M. Sargos, and R. Andriamalala,
"Constrained optimization of high power synchronous reluctance
motor using non linear reluctance network modeling," in Industry Applications Conference, 2006. 41st IAS Annual Meeting. Conference
Record of the 2006 IEEE, 2006, pp. 1201-1208.
[23] S. Talebi, P. Niazi, and H. Toliyat, "Design of Permanent
Magnet-Assisted Synchronous Reluctance Motors Made Easy," in
Industry Applications Conference, 2007. 42nd IAS Annual Meeting.
Conference Record of the 2007 IEEE, 2007, pp. 2242-2248. [24] N. Bianchi, S. Bolognani, D. Bon, and M. Dai Pre, "Torque harmonic
compensation in a synchronous reluctance motor," Energy
Conversion, IEEE Transactions on, vol. 23, pp. 466-473, 2008. [25] R.-R. Moghaddam, F. Magnussen, and C. Sadarangani, "Novel rotor
design optimization of Synchronous Reluctance Machine for low
![Page 10: Design Methods of Transversally Laminated Synchronous ... · (the ratio of the d-axis inductance and the q-axis inductance), and the rotor design has been the main issue. Generally](https://reader034.fdocuments.us/reader034/viewer/2022042016/5e74d1acb480d5292756b348/html5/thumbnails/10.jpg)
CAI et al. : DESIGN METHODS OF TRANSVERSALLY LAMINATED SYNCHRONOUS RELUCTANCE MACHINES 173
torque ripple," in Electrical Machines (ICEM), 2012 XXth
International Conference on, 2012, pp. 720-724.
[26] R. R. Moghaddam, F. Magnussen, and C. Sadarangani, "Novel rotor
design optimization of synchronous reluctance machine for high
torque density," 2012. [27] R.-R. Moghaddam and F. Gyllensten, "Novel high-performance
SynRM design method: An easy approach for a complicated rotor
topology," Industrial Electronics, IEEE Transactions on, vol. 61, pp. 5058-5065, 2014.
[28] S. Cai, H. Hao, M.-J. Jin, and J.-X. Shen, "A Simplified Method to
Analyze Synchronous Reluctance Machine," in Vehicle Power and Propulsion Conference (VPPC), 2016 IEEE, 2016, pp. 1-6.
[29] A. J. Pina, H. Cai, Y. Alsmadi, and L. Xu, "Analytical model for the
minimization of torque ripple in permanent magnets assisted synchronous reluctance motors through asymmetric rotor poles," in
Energy Conversion Congress and Exposition (ECCE), 2015 IEEE,
2015, pp. 5609-5615. [30] N. Bianchi, S. Bolognani, D. Bon, and M. D. Pré, "Rotor flux-barrier
design for torque ripple reduction in synchronous reluctance and
PM-assisted synchronous reluctance motors," Industry Applications, IEEE Transactions on, vol. 45, pp. 921-928, 2009.
[31] S.-H. Han, T. M. Jahns, and Z. Zhu, "Analysis of rotor core
eddy-current losses in interior permanent-magnet synchronous machines," Industry Applications, IEEE Transactions on, vol. 46, pp.
196-205, 2010.
[32] B. Boazzo, A. Vagati, G. Pellegrino, E. Armando, and P. Guglielmi, "Multipolar Ferrite-Assisted Synchronous Reluctance Machines: A
General Design Approach," Industrial Electronics, IEEE Transactions on, vol. 62, pp. 832-845, 2015.
[33] S. Han, T. Jahns, W. Soong, M. Guven, and M. Illindala, "Torque
Ripple Reduction in Interior Permanent Magnet Synchronous Machines Using Stators With Odd Number of Slots Per Pole Pair,"
IEEE Transactions on Energy Conversion, vol. 25, pp. 118-127, 2010.
[34] M. Ferrari, N. Bianchi, and E. Fornasiero, "Rotor saturation impact in
synchronous reluctance and PM assisted reluctance motors," in Energy
Conversion Congress and Exposition (ECCE), 2013 IEEE, 2013, pp.
1235-1242. [35] M. Palmieri, M. Perta, F. Cupertino, and G. Pellegrino, "Effect of the
numbers of slots and barriers on the optimal design of synchronous
reluctance machines," in Optimization of Electrical and Electronic Equipment (OPTIM), 2014 International Conference on, 2014, pp.
260-267.
[36] J.-X. Shen, S. Cai, H. Hao, and M.-J. Jin, "Comprehensive parameter design for transversally laminated synchronous reluctance machines,"
in Electrical Machines and Systems (ICEMS), 2016 19th International
Conference on, 2016, pp. 1-9. [37] S. de Pancorbo, G. Ugalde, J. Poza, and A. Egea, "Comparative study
between induction motor and Synchronous Reluctance Motor for
electrical railway traction applications," in Electric Drives Production Conference (EDPC), 2015 5th International, 2015, pp. 1-5.
[38] S. Armentia, G. Ugalde, J. Poza, and A. J. Escalada, "Numerical 2D
methodology for synchronous reluctance motor rotor design," in Electric Drives Production Conference (EDPC), 2015 5th
International, 2015, pp. 1-4.
[39] L. Wu, Z. Zhu, J. Chen, Z. Xia, and G. W. Jewell, "Optimal Split Ratio in Fractional-Slot Interior Permanent-Magnet Machines With
Non-Overlapping Windings," Magnetics, IEEE Transactions on, vol.
46, pp. 1235-1242, 2010.
Shun Cai received the B.S. and M.S.
degree in electrical engineering from
Zhejiang University, China in 2014 and
2017, respectively. He is currently
pursuing the PHD degree at the
University of Sheffield, UK. His
research interests include synchronous
machine topologies and optimization.
Jian-Xin Shen received the B.S. and M.S.
degrees from the Xi’an Jiaotong University,
China in 1991 and 1994, respectively, and
the PhD degree from the Zhejiang
University, China in 1997. He was with the
Nanyang Technological University,
Singapore (1997-1999), the University of
Sheffield, UK (1999-2002), and IMRA
Europe SAS, UK Research Centre, UK
(2002-2004). Since 2004 he has been a Professor of Electrical
Engineering at the Zhejiang University. He has published more
than 220 technical papers, including 5 prize papers, and he
holds over 30 patents. He is an IET Fellow and an IEEE Senior
Member. His main research interests include design, control
and applications of electrical machines and drives, and
renewable energies.
He Hao received the B.S. and PhD degrees
from the Zhejiang University, China in
2008 and in 2013, respectively. He was an
Electrical Engineer at Hangzhou Wahaha
Group Co., Ltd (2013-2014). Since 2014
he has been a Post-doctor at the Zhejiang
University. His main research interests
include design, control and applications of
electrical machines and drives.
Mengjia Jin received the B.S. and PhD
degree from Zhejiang University, China, in
2001 and 2006 respectively. Now he is
working with Electrical Engineering
Department of Zhejiang University as an
associate professor. His works focus on
electrical machine design and drives.