Dynamic Analysis of Synchronous Reluctance Motor Drives Based on

Simulink

and Finite Element Model

Francesco Parasiliti, Marco Villani Alessandro Tassi
Department of Electrical and Information Engineering Spin Applicazioni Magnetiche S.r.l.
University of LAquila
67100 Poggio di Roio, LAquila 29010 Pianello Val Tidone, Piacenza
ITALY ITALY
[email protected] [email protected]


Abstract A fine motor analysis that takes the driving control
into account allows to evaluate with a good accuracy the
dynamic performance. The use of Simulink

together with FEA
has allowed to investigate deeply the dynamic behaviour of the
Synchronous Reluctance Motor in different operating
conditions, giving results closer to the actual motor
performance.

I. INTRODUCTION

Synchronous Reluctance Motors (SRM) could be
considered as alternative to its counterparts, namely
Permanent Magnet, Switched Reluctance and Induction
Motors. Moreover, it has been demonstrated that SRM has
the same or higher power density than the Induction Motors
in the low and medium power level.
The fine analyses of these motors are very difficult because
of their highly saturated operation conditions and their
salient structure. The use of linear models in evaluating the
performances of the considered motors can lead to serious
errors since the SRM present notable nonlinear
characteristics due to the effects of the saturation and cross-
coupling phenomena occurring in the magnetic circuits.
These phenomena can be taken into account only by an
accurate nonlinear analysis which can be performed by Finite
Element programs that allow an accurate prediction of
machine parameters and performances [1].
Moreover, the growing demand of high dynamic
performance motors (e.g. with fast torque response) requires
a fine motor analysis taking into account the driving control.
Usually, to simplify the controller, constant values of axis
inductances (L
d
and L
q
) are considered, but it is well known
that the axis currents and the stator-rotor relative position
have a considerable influence on the inductances, with
significant effects on torque behaviour. For these reasons, in
order to predict with a good accuracy the dynamic
performance, it is suitable to develop a model that allows to
link the drive scheme with a fine motor model.
With this aim, the link between Simulink

code (by
Mathworks) and a Finite Element software Flux2D

by
Cedrat has been employed to simulate the SRM drive. This
approach allows to simulate, with good accuracy, the motor
behaviour in different operating conditions related to
imposed control strategy [2], giving to the designer useful
indications (e.g. the torque ripple) in view of the motor and
control design refinements.


The proposed study concerns with a Synchronous Reluctance
Motor with two flux barriers 4 pole, 200 V, 20 Nm: a view of
the rotor structure is shown in Fig.1
The paper presents a preliminary analysis with constant
values of axis inductances; then, more accurate models have
been carried out tacking into account the parameters
variation.












Fig. 1 - View of the two flux barriers rotor


II. THE CONTROL STRATEGIES OF THE SRM

The considered torque expressions of the SRM, with
reference to the rotating d-q frame synchronized with the
rotor, are:

( ) 2
2
3
2
sin I L L p T
s q d
= , (1)
( )
q d q d
I I L L p T =
2
3
, (2)

where p are the pole pairs, I
s
the space vector of the stator
current, I
d
and I
q
the direct and quadrature axis currents and
the angle between d-axis current and the vector I
s
.
A suitable control strategy is to maximize the Torque-
Current ratio and this can be achieved if is equal 45: it
corresponds to impose the same values for the axis currents.
Since for this type of motors L
d
is usually 78 times higher
than L
q
, the above mentioned strategy cannot be adopted
when I
d
component increases; then, the d-axis current should
be kept constant and only the q-axis component increases.


1516 1-4244-0136-4/06/$20.00 '2006 IEEE

Alternatively, the d-axis current can be always fixed
constant.
Therefore, the control strategies could be (Fig.2):
- d-q current angle control (A, B, C trajectory);
- constant I
d
control (A, B, C trajectory).
In this paper the constant I
d
control only has been
investigated, and the control scheme is shown in Fig.3.

=45
A
D C B
i
d
i
q
i
dmax


Fig. 2 - Locus of the stator current vector

III. THE FE ANALYSIS OF THE SRM

The proposed study concerns the SRM whose main data are
presented in Tab.I. The investigated motor has a flux barriers
rotor that presents, respect to the axially laminated one [2], a
simplicity in mechanical construction, lower manufacturing
cost, and the rotor skewing possibility; on the other hand it
has a quite large level of torque ripple due to the inductances
variation with rotor position that can cause, particularly at
low-speed, inaccuracy in motion control, noise and
vibrations.
TABLE I
MAIN DATA OF THE MOTOR
Number of poles 4
Stack length (mm) 130
Outside stator diameter (mm) 152
Inner stator diameter (mm) 90
Stator slots 36
Number of flux barriers per pole 2
Number of turns per phase 198
Phase voltage (V) 200
Phase current (rms) (A) 7.7
Torque (Nm) 20
Frequency (Hz) 50
















An accurate bi-dimensional FE model has been carried out
by the software Flux2D

ver.9

by Cedrat, [3], introducing a
parametric model of motor in order to modify the geometric
dimensions of stator and rotor shape, the rotor position ()
and the currents.
The magnetostatic analyses allow to evaluate the axis
inductances, whose values could depend on the axis currents
only (if one rotor position only is analyzed):
- L
d
= f(I
d
, I
q
); (3)
- L
q
= f(I
d
, I
q
);

or on currents and rotor position (see Fig.4):
- L
d
= f(, I
d
, I
q
); (4)
- L
q
= f(, I
d
, I
q
).

If a transient analysis is carried out, the calculated
inductances depend also on the variation of currents with
time:
- L
d
= f(, I
d
(t), I
q
(t)); (5)
- L
q
= f(, I
d
(t), I
q
(t)).
A further FE model has been developed, taking into account
the rotor skewing. This design solution allows to drastically
reduce the high torque pulsation in flux barriers type SRM
[1], [4].
Referring to Finite Element analysis, skewing presents
difficulties in computing the magnetic field distribution, as
well as machine parameters and performances evaluation.
These difficulties are overcome considering a simplified
several-blocks equivalent rotor and by using several cross
sectional 2D FEA [5]. As alternative the Skew module of
Flux2D can also be used.
Fig.4 shows the axis inductances profiles for a skewed rotor
(one stator slot pitch): in this case a skewing of one pole
pitch has been considered and the rotor has been divided into
10 blocks. The inductances profiles have been compared with
the no-skewed rotor ones (4).


















A

Fig. 3 - Scheme of the constant I
d
control
1517

Comparison Ld
0,00
0,05
0,10
0,15
0,20
0,25
0 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90
Position [degrees]
I
n
d
u
c
t
a
n
c
e

[
H
]
Ld
Ld Skewed

Comparison Lq
0,000
0,005
0,010
0,015
0,020
0,025
0,030
0,035
0 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90
Position [degrees]
I
n
d
u
c
t
a
n
c
e

[
H
]
Lq
Lq Skewed


Fig. 4 - Axis inductances profiles (I
d
=3.7 A; I
q
=10.2 A)

IV. DINAMIC ANALYSIS OF SRM BY SIMULINK


In this step the SRM drive (Fig.3) is simulated by Simulink
code, solving the dynamic equations of the SRM in d-q
reference frame, where inductances values are those
evaluated by preliminary (out of line) FEA; the SMR is then
simulated by lumped parameters like inductances and
resistances. The analyses concern the no-skewed rotor.
Two different tests have been carried out on the grounds of
two inductance evaluation hypotheses:
a) variable inductances with currents but independent of
stator-rotor relative position (3);
b) variable inductances with currents and stator-rotor
relative position (4).
Simulink solves the SRM model using the preliminary
evaluated inductances values corresponding to the reference
currents imposed by the controller.
Speed responses between 0 and 1500 rpm with constant 22
Nm load torque have been simulated. Steady-state axis
currents were: I
d
=3.7 A and I
q
=10.2 A.
The time torque response for both cases are shown in Fig.5.
In the case (a), since the axis inductances are constant, the
torque pulsation is absent. In the case (b) a significant ripple
appears, that is about 11% (defined like as the ratio between
the difference of the maximum and minimum values of
torque and the average one).

V. DINAMIC ANALYSIS BY SIMULINK AND FEM

With the aim to link the drive scheme with a fine motor
model, the Simulink control scheme has been directly
interfaced with the FE model of the SRM.





Fig. 5 - Torque waveforms (I
d
=3.7 A; I
q
=10.2 A).

This analysis has been carried out thanks to the link between
the FE software Flux2D and Simulink code [3]. By such co-
simulation method, the transitory effects and the cross-
coupling effects are accurately reproduced.
Respect to the analysis in the Section IV, the SMR is not
simulated by lumped parameters but it is accurately modelled
by on-line transient FE analysis where the axis inductances
depend on the rotor position and the currents both variable
with time (5).
The drive scheme of the SRM with constant I
d
control is
shown in Fig.6, where it is evident the block for the FEA of
the Synchronous Reluctance Motor.
This co-simulation between Simulink (for the drive) and
Flux2D (for the motor) needs an external circuit that allows
to link the two codes. Particularly, in order to fed the FE
model of the SRM, three voltage generators have been
introduced, whose values (variable with time) depend on the
controller. Then, starting from the motor phase resistance and
the inductances values (evaluated on-line by FEA), the phase
currents are calculated by Flux2D. From these values, the
currents in each stator slot are imposed according to the
winding distribution.
The time currents, torque and speed responses are presented
in Fig.7 and Fig.8, where a load torque of 22 Nm has been
imposed. After the transient operation, the speed reaches
1500 rpm and the axis currents settle to the imposed values.

case (b)
case (a)
d-axis inductances
q-axis inductances
skewed
no-skewed
skewed
no-skewed
1518

























In this case, the torque ripple is higher than the previous
analysis one (Fig.5b), giving result closer to the actual motor
torque.
It is important to underline that knowledge of the torque
ripple it is very important for the SRM since it can cause,




Current Id=costant
-25
-20
-15
-10
-5
0
5
10
15
20
0,00 0,05 0,10 0,15 0,20 0,25
Time [s]
C
u
r
r
e
n
t

[
A
]
Ia Ib Ic

Current dq Id=costant
0
5
10
15
20
25
0,00 0,05 0,10 0,15 0,20 0,25
Time [s]
C
u
r
r
e
n
t

[
A
]
Id
Iq
Idrif


Fig. 7 Phase currents and axis currents vs. time



























particularly at low-speed, inaccuracy in motion control, noise
and vibrations.
This example demonstrates how an accurate analysis of the
motor by a suitable tool, allows to predict the real motor
performance and verify the dynamic behaviour for different
operating conditions.


Torque Id=costant
0
5
10
15
20
25
30
35
40
45
0,00 0,05 0,10 0,15 0,20 0,25
Time [s]
T
o
r
q
u
e

[
N
m
]
Torque
Drag Torque


Angular Velocity Id=costant
-200
0
200
400
600
800
1000
1200
1400
1600
1800
0,00 0,05 0,10 0,15 0,20 0,25
Time [s]
A
n
g
u
l
a
r

V
e
l
o
c
i
t
y

[
r
p
m
]
Angular Velocity
Reference


Fig. 8 Torque and speed vs. time


Fig. 6 - SRM drive with constant I
d
control
FEA
Phase currents
Speed

speed
Axis currents
1519


VI. CONCLUSIONS

The growing demand of motors with high dynamic
performance requires not only an adequate (and more
efficient) designs procedure but also a fine motor analysis
taking into account the driving control.
The use of Simulink

directly interfaced with Flux FE
software allows to investigate deeply the dynamic behaviour
of the SRM in different operating conditions and this
approach can be defined like as a virtual experimental test.
The goodness of the results depends not only on the proposed
method but also on the accuracy in the construction of the FE
and control scheme models.
This approach could represent an effective tool for the
analysis of motor performance and the refinement of SRM
design in order to improve and optimise [6] its dynamic
performance.
VII. REFERENCES
[1] G.Conti, F.Parasiliti and M.Villani, "Torque Ripple
Analysis in Synchronous Reluctance Motors", Internat.





































Journal ELECTROMOTION, vol. n.3 1996, pp. 188-
193.
[2] I. Boldea, Reluctance Synchronous Machines and
Drives, Clarendon Press: Oxford, 1996.
[3] Cedrat, Flux Users Guide and Flux To Simulink
Technology Manual, Vers. 9, 2005.
[4] E.Chiricozzi, F.Parasiliti and M.Villani, "Design
Solutions to Optimize Torque Ripple in Synchronous
Reluctance Motor", Intern. Conference on Electrical
Machine (ICEM96), Vigo (Spain), Sept. 1996, pp.148-
153.
[5] G.Conti, F.Parasiliti and M.Villani, Analysis of
Synchronous Reluctance Motor with Skewed Rotor by
Finite Element Method, 8
th
Biennal IEEE
Conferenceon Electromagnetic Field Computation
(CEFC'98), Tucson (USA), June 1998.
[6] L.Cirio, S.Lucidi, F. Parasiliti and M. Villani, "A global
optimization approach for the synchronous motors
design by finite element analysis", International
Journal of Applied Electromagnetics and Mechanics,
vol. n. 16, 2002, pp. 13-27.

















1520