Paper number 205-4346 (revised) - International Journal of Modelling and Simulation

1
A PC-CLUSTER BASED FULLY DIGITAL REAL-TIME SIMULATION OF A FIELD-
ORIENTED SPEED CONTROLLER FOR AN INDUCTION MOTOR

M. Ouhrouche
(*)
, R. Beguenane
(*)
, A.M. Trzynadlowski
(**)
, J.S. Thongam
(*)
and M. Dube-Dallaire
(***)


(*)
Department of Applied Sciences, University of Quebec at Chicoutimi
555, Bld. De lUniversit, Chicoutimi (Qc), Canada, G7H 2B1
E-mail: [email protected]

(**)
Department of Electrical Engineering, University of Nevada, Reno, USA

(***)
Opal-RT Technologies, 1751 Richardson, suite 2525
Montreal (Qc), Canada, H3K 1G6


Abstract

With the ever-evolving complexity of new adjustable speed drives (ASD) systems, maintaining the
development and the prototyping costs at reasonable level is becoming more and more challenging for the
manufacturers of these equipments. The cost constraint has led design and test engineers to fully
recognize the importance of real-time simulation, and it is now widely used by high-tech industries as an
essential and powerful tool to prototype complex engineering systems in a cost-effective and secure
manner, while reducing the time-to-market.
In this paper, the authors present a fully digital real-time simulation of a high performance indirect
field-oriented controller for an induction motor using RT-Lab
TM1
software package running on a simple
off-the-shell PC. This real-time simulation tool is, for example like dSPACE
TM2
Real-Time Kernel, now
adopted by many high-tech industries, particularly automotive and aeronautics industries, as a real-time

TM1
Trademark of Opal-RT Technologies
TM2
Trademark of dSPACE GmbH
Paper number 205-4346 (revised)

2
laboratory package for rapid prototyping of complex control systems and for hardware-in-the-loop (HIL)
applications.
The proposed speed controller in indirect field-orientation for an induction motor is an adaptive IP
controller. Its parameters are synthesized by considering possible variations in the rotor resistance. This
parameter is estimated online and its actual value is supplied in real-time to the speed controller ensuring
the robustness of the drive. Interacting in real-time with the speed controller to test its robustness against
the rotor resistance variations is not possible without real-time simulations. This achievement is, to the
best of the authors knowledge, reported for the first time in this paper and is believed to be an important
contribution to rapid prototyping of high performance induction machine controllers.

Keywords: Induction Motor, Field-Oriented Control, Online estimation, PC-Cluster, Real-Time
Simulations and Rapid Control Prototyping.


1. Introduction

Presently as a consequence of the important progress achieved in power electronics and with advances of
microelectronics and computer technologies allowing high level identification and control algorithms to
be implemented in real-time, electric drives with induction machines have become the most widely used
in variable-speed applications, for reasons of costs, size and reliability. These ASD are based on the well-
known Field-Oriented Control (FOC), which is now indisputably a standard, or on the Direct Torque
Control (DTC). Active research in this area is still undertaken around the world to develop new
generation of ASD with advanced identification and control algorithms in order to achieve robust and
sensorless operation.
Field-Oriented Control, also known as Vector Control, proposed by Blaschke in 1972 [1], is a
powerful control strategy that allows the achievements of high performance control with induction
Paper number 205-4346 (revised)

3
motors. Its main objective is, as in separately excited DC machine, to independently control the produced
torque and flux. In rotor-flux oriented control, this decoupling is achieved by aligning and linking the d-
axis of a d-q reference frame with the rotor flux space vector. Under this condition and at constant flux
the torque varies linearly with the q-component of the stator current space vector; also known as torque
producing current.
In Direct Field-Oriented Control (DFOC) strategy, both the instantaneous magnitude and position of
the rotor flux are supposed to be available and known with high precision; i.e.: directly measured or
estimated using for example a nonlinear state observer. On the other hand, the position of the rotor flux
space vector is obtained analytically in Indirect Field-Oriented Control (IFOC) strategy. IFOC is much
easier to implement than the DFOC, but the slip-speed calculation involves the rotor time-constant, which
is known to vary with frequency and temperature. To maintain the flux orientation, the variations of this
parameter should then be tracked online, and its actual value supplied in real-time to the speed controller
and to the slip-speed calculation module [2].
This brief introduction gives us a good idea on how complex ASD systems are, and this complexity is
notably increased when new control strategies and new converter topologies are, for example, to be
prototyped for a specific application. In this case and for the manufacturers to be competitive, design and
test engineers face three major problems to solve: reduction of the time-to-market, maintaining the cost at
reasonable level and ensuring zero-risk of damaging equipments involved in the prototyping process.
These constraints have led design and test engineers to fully recognize the importance of real-time
simulation, and is now widely used by high-tech industries, as an essential and powerful tool to prototype
complex engineering systems in a cost-effective and secure manner [3]-[5].
In this paper, the authors present a PC-Cluster-based fully digital real-time simulation of an indirect
field-oriented speed controller for an induction motor using RT-Lab software package [6]. Satisfactory
results are obtained with a fixed time-step of 50 s, which is many times smaller than what could be
achieved using the most recent Digital Signal Processor available on the market. RT-Lab software uses
Matlab/Simulink as a front-end interface for editing graphic models in block-diagram format, which are
Paper number 205-4346 (revised)

4
afterwards used by RT-Lab to generate the necessary code for real-time simulations on a single or more
target processors running QNX [3]-[5]. The proposed speed controller is an adaptive IP-type and its
robustness against the rotor resistance variations is tested in real-time. This achievement is, in authors
opinion, an important contribution.

2. Induction Motor Modeling for Field-Oriented Control

The well-know fifth order (d-q) model of an induction motor established in the synchronous reference
frame rotating at speed
e
is given by the following equation.
( )
( )
2
1
1
1
1
s m m r
sd e sq rd rq
s r s r r s r
s m r m
sd
e sd sq rd rq
s r s r s r r
sq
m
rd
sd rd e r rq
r r
rq
m
r
sq e r rd rq
r r
m
sq r
r
R L L
I I
L L L L L
R L L
I
I I
L L L L L
I
d
L
I
dt
L
I
p L
I
JL

| |
+ + + +
|
\ .
| |
( + +
|
(
\ .
(
( =
+
(
(
(


( )
1
1
0
0
0
sd
s
sq
s
d sd rq r l
V
L
V
L
F p
I T
J J

(
(
(
(
(
(
(
(
(
(
(
(
(
+
(
(
(
(
(
(
(
(
(
(

(

(
(

(1)
In (1),
2
1
m
s r
L
L L

| |
=
|
\ .
is the coefficient of dispersion and
r
r
r
L
R
= is the rotor electrical time-
constant.

Under rotor flux orientation, the d-axis is aligned and linked to the rotor flux space vector so that the
q-component
rq
= 0 and the d-component
rd
=
r
. The induction motor model established in the rotor
flux field coordinate is then given by:

Paper number 205-4346 (revised)

5
( )
2
1
1
1
1
1
0
0
s m
sd e sq r
s r s r r
sd
s sd
s m r
e sd sq r
sq
s r s r r
sq
s r
m
sd r
r
r r
m
sq r r l
r
R L
I I
L L L
V
L I
R L
I I
I
d L L L
V
L
dt
L
I
p L F p
I T
JL J J

(
| |
+ + +
( |
\ . (

(
(
| |

(
+
( |

(
\ . (
= +

(
(

(
(

(
(


(

(

(
(

(
(
(
(
(
(
(
(
(

(2)

m r
e sq
r
pL
T I
L

= (3)

(1 )
m
r sd
r
L
I
S

=
+
(4)

m sq
sl e r
r r
L I


= = (5)
As the control is done in the rotor flux field reference frame, all stator variables (voltages, currents
and fluxes) in the natural abc reference frame are obtained using a simple transformation given by:

( ) ( )
( ) ( )
cos sin
2
2 2
cos sin
3 3
3
2 2
cos sin
3 3
e e
a
d
b e e
q
c
e e
x
x
x
x
x





(

(
(
(
(
(
=
(
(
(

(
(

+ + (

(6)

m sq
e r
r r
L I
dt

| |
= +
|
\ .

(7)
Figure 1 shows the block diagram of the proposed indirect field-oriented control for an induction
motor implemented for digital real time simulations. As shown, the speed controller is designed to take
into account the possible variations in the rotor resistance, which is estimated on-line using a reduced
order Kalman filter. Its actual value is supplied in real-time to the speed controller and to the slip-speed
calculation module to ensure a perfect flux orientation. The control signals for the PWM Voltage Source
Inverter are obtained using a closed-loop hysteresis AC current controller.
Paper number 205-4346 (revised)

6












Figure 1. Block diagram of IFOC for an induction motor with online estimation of the rotor resistance


3. Design of the IP Speed Controller

In Figure 1, the starred variables are the command values obtained under constant flux condition. Using
(3) to (5), the command variables are then given by:

*
* r
sd
m
I
L

= (8)

*
*
*
e r
sq
m r
T L
I
pL
= (9)

*
*
*
sq
sl
r sd
I
I

= (10)
The parameters K
p
and K
i
of the IP speed controller are synthesized on the basis of the rotor electrical
and mechanical equations and by considering possible variations in the rotor resistance. The reference of
IM
*
e
T
*
r

Eq. 3
*
sq
I
*
sd
I
e

dq
ABC
2h
*
sa
I
sa
I
*
sc
I
sc
I
*
sb
I
sb
I
DC
V
*
sa
V
*
sb
V
*
sc
V
*
r

, s abc
I
, s abc
V

ABC
r
R
2h
2h
ABC
Eq. 4
Eq. 7
Reduced Order
Kalman Filter
r
R
Paper number 205-4346 (revised)

7
the flux producing current
*
sd
I is kept constant at its maximum value, and the reference of the torque
producing current
*
sq
I is taken to be a unit step. Figure 2 below shows the speed control system block
diagram where:

2
2
2 *
2
, ,
r r
t n n sl
r r n
R R
K
L L

= = + = (11)






Figure 2. Speed control system block diagram

It is found that the speed tracking specifications are satisfactory achieved by using an IP-controller
with the following parameters [7]:

2
*
2
2 *
1 , 2
sq
r r
i p
r sd r
I
JR JR
K K
L I L
(
| |
(
= + =
|
|
(
\ .

(12)

4. Online Rotor Resistance Estimation

To ensure the robustness of the whole drive, an online identification procedure of the rotor resistance
should be achieved. This identification allows the actual value of this parameter to be tracked online and
in real-time in order to update the parameters of the speed controller at each sampling time.
Extended Kalman filter is one of the most effective methods that have been reported in the literature
for online states and parameter identification of an induction motor drive [8-10]. The standard Kalman
r

i
p
K
K
s
+
r

*
e
T
*
r

l
T
+

e
T
2 2
2
t
n n
K
s s + +
1
Js F +
Paper number 205-4346 (revised)

8
Filter (KF) is a recursive mean-squared state estimator capable of producing optimal estimates of states
that are not measurable [11]. It uses the plant's input and the output measurements, which are noisy data,
together with the state space model of the system. To estimate the rotor resistance, this time-varying
parameter is treated as a new state in addition to the stator currents and the rotor fluxes. A new nonlinear
model of the machine having the following form is then obtained.


( ) ( ) ( ) { } ( )
( ) ( ) { } ( )
1
2
1 , x k f x k u k w k
y k h x k w k
+ = +
= +
(13)

It is to mention that this model is established in the stator-fixed reference frame with
e
in (1) equals
to zero and replacing d and q suffixes by and respectively. In (13), the state, the command and the
output vectors are as follow:
( ) ( ) ( ) ( ) ( ) ( )
t
s s r r r
x k I k I k k k k

( =

(14)
( ) ( ) ( )
t
s s
u k V k V k

( =

(15)
( ) ( ) ( )
t
s s
y k I k I k

( =

(16)
w
1
(k) and w
2
(k) are respectively the process and the measurement noise vectors. These random
vectors are supposed zero-mean uncorrelated and are characterized by:
( ) { } ( ) ( )
{ } 1 1 1
0, 0
t
kj
E w k E w k w j Q Q = = (17)
( ) { } ( ) ( )
{ } 2 2 2
0, 0
t
kj
E w k E w k w j R R = = (18)

In (17) and (18), Q and R are respectively the process and the measurement covariance matrices,
which are positive and semi-definite.

Paper number 205-4346 (revised)

9
The discrete Kalman filter algorithm is as follows:
Step 1: Prediction
( ) ( ) { }
1/ / , ( ) x k k f x k k u k + = (19)
( ) ( ) ( ) ( ) ( ) 1/ /
t
P k k F k P k k F k Q k + = + (20)
where ( )
x k is the state estimate, P(k) is the estimation error covariance matrix and
( ) ( ) ( ) { }
( ) ( ) /
,
x k x k k
F k f x k u k
x
=

(21)
Step 2: Correction
( ) ( ) ( ) ( ) ( ) ( ) { }
1/ 1 1/ 1 1 1/ x k k x k k K k y k h x k k + + = + + + + + (22)
( ) ( ) ( ) ( ) ( ) ( ) ( ) { }
1
1 1/ 1 . 1 1/ 1 1
t t
K k P k k H k H k P k k H k R k

+ = + + + + + + + (23)
( ) ( ) ( ) ( ) ( ) 1/ 1 1/ 1 1 1/ P k k P k k K k H k P k k + + = + + + + (24)
where K is the Kalman gain matrix, (k+1/k) denotes prediction at time (k+1) based on data up to and
including k and
( ) ( ) { }
( ) ( ) 1 1/
1 1
x k x k k
H k h x k
x
+ = +

+ = +

(25)

(20), (23) and (24) are the Riccati equations and their solution requires a lot of CPU time. This
requirement is considered as a major disadvantage in using this estimator for real-time applications.
In the case of the study presented in this paper, since the stator currents are available for measurement
and in order to reduce the computational requirements, a reduced order model of the induction motor is
used for implementing the rotor resistance estimation algorithm. This reduced order model is established
in the stator-fixed reference frame and is given by (26) and (27) hereafter. Note that the rotor flux
components are also estimated but are not used since this study addresses the indirect rotor flux oriented
control.
Paper number 205-4346 (revised)

10

| |
0
r m r
r r r s
r r
r
r m r
r r r r s
r r
r
R L R
I
L L
R L R d d
x I
dt dt L L
R

(
+
(
(
(
(
(
= = +
(
(
(
(

(
(

(26)

| |
2
2 2
2
2 2
r m m r r m
s
r r s
s s s s
r r r
s
m r r m r m
s s s s r r s
r r r
R L L R L
dI
I
V R I L
L L L
dt
y
dI
L R L R L
V R I L I
dt L L L

(
(
+
(
(
(
( = =
(
(
+ (
(

(

(27)
As shown in the state equation given by (26), the rotor resistance has been introduced as a third state.
It is assumed that its variation due to temperature and skin effect is very slow in comparison to the
dynamics of the other states; this is why its derivative is equated to zero. In discrete form, the state and
output equations, respectively (26) and (27), are as follow:

( ) 1
0 1 0
1 0 0
0 0 1 0 0
r m r
r
r r
r r
s
r m r
r r r
s
r r
k
r r
k k
k k
R L R t
t t
L L
I
R L R t
t t
I
L L
R R

+
( (

( (
( (
( (
(
( (
( (
= +
(
( (
( (

( (
( (

( (
( (

(28)
t represents the time-step integration and the components of the stator current
space vector became the input variables.

( )
( )
2
2 2
2
2 2
r m m r m s
r
s s s s s
r r r
r
s s s s s m s
m r r m
k
r
k
r r r
k
R L L L I
V R I L I
L L L
V R I L I L I
L R L
R
L L L

(

(
(
(
(
(
= (
(
(

(

(
(

(

(29)

( ) ( )
( ) ( )
1
1
s s
s
s s
s
I k I k
I
t
I k I k
I
t

(30)
Paper number 205-4346 (revised)

11
5. PC-Cluster Based Real-time Simulation of an Induction Motor Vector Control

PC-Cluster based real-time simulation is now widely used by high-tech industries, particularly automotive
[13]-[16] and aeronautics industries (aircraft flight control, satellite control), as the main tool for rapid
prototyping of complex engineering systems in a cost-effective and secure manner, while reducing the
time-to-market.
A PC-Cluster is a parallel multiprocessor computer system capable of meeting the real-time
performance requirements of the simulation. Fig.3 below shows the concept of digital real-time
simulation of an induction motor drive system. Real-time simulation is achieved by running on separate
processors (targets) and in parallel the speed and decoupling control module, the static converter module
and the induction motor module. These three modules are actually C-code (numerical modules) obtained
by an automatic code generator for real-time execution.












Figure 3. Real-time simulation of an IM-ASD system

QNX
QNX
Host PC with
RT /OS



IM
QNX
Paper number 205-4346 (revised)

12
Real-time simulation is the first step in any rapid control prototyping development project. As soon as
the design requirements are fulfilled, the physical process to be controlled may be integrated, via fast I/O
interfaces and replace its real-time model, with the simulation environment as shown in Fig. 4 below.
The real process is then driven by the real-time model of the controller.







Figure 4. Hardware-in-the-Loop simulation of IM-ASD system

Once the Hardware-in-the-Loop simulation has been completed, the automatically generated code of
the real-time model of the controller may be used as the actual control system by transferring it into the
memory of an embedded controller. This procedure is known as Rapid Control Prototyping.

In this paper, a PC-Cluster Based fully digital real-time simulation of a high performance indirect
field-oriented speed controller for an induction motor is successfully achieved using RT-Lab software
package. Satisfactory results are obtained with a fixed time-step of 50 s, which is many times smaller
than what could be achieved using the most recent Digital Signal Processors available on the market.
As shown in Fig. 5, RT-Lab [6] uses Matlab/Simulink as a front-end interface for editing graphic
models in block-diagram format, which are afterwards used by this real-time simulator to generate the
necessary C-code for real-time simulations on a single or more target processors running QNX.


Real Time Model of
the Controller
Real Time Model of the
Induction Motor and the
Power Converter
D
to
A
A
to
D
IM
Paper number 205-4346 (revised)

13











Figure 5. Principle of real-time simulation using RT-Lab

Fig. 6 shows the real-time model of the proposed indirect field-oriented speed controller for an
induction motor as implemented in RT-Lab environment. This model is distributed over three target CPU
motherboards. The two first target processors are Pentium IV running at 2.4 GHz, installed on a dual-
CPU with shared memory. The third one, connected to the others through a fast Fire Wire real-time link,
is a Pentium III running at 550 MHz with 512 Mbytes of memory.
The first CPU of the dual-CPU unit, acting as slave # 1: SS_CONTROLLERS, computes in real-time
the reduced order Kalman filter algorithm, the speed controller, the rotor flux decoupling unit and the
coordinate transformation unit. The second one, acting as slave # 2: SS_INV_MOTOR, computes in real-
time the induction motor, the PWM signal generator and the voltage source inverter. The third processor,
acting as master: SM_BREAKOUT_BOX, is dedicated to data acquisition. SC_USER_INTERFACE is
the console used for input reference and command signals and for signal visualisation. Details of these
simulation modules as implemented in RT-Lab environment are given in Appendix B.

RT-LAB
SIMULATOR
SIMULINK
REAL-TIME SIMULATION
WITH (I/O) INTERFACE
MODEL
SEPARATION
Separation into
Subsystems to be
Executed on
Individual
Processors
CODE
GENERATION
Use of Automatic
Code Generator
For Real-Time
Execution
Target Processors are
External Boards
Running
QNX - RTOS
PC
Target 1
PC
Target 2
PC
Target n

Paper number 205-4346 (revised)

14














Figure 6. IFOC implementation using RT-Lab simulator

6. Simulation results

The sampling time is one of the major constraints in real-time simulations. In the case of the work
presented in this paper, a value of 50 s, which is many times smaller than what could be achieved by the
latest available Digital Signal Processors, is used and satisfactory results are obtained with no-overrun.
For one time step, it took 6 s to execute in real-time the PWM signal generation unit, the power
converter and the induction motor; and 13 s to execute the reduced order Kalman filter algorithm, the
speed controller and the rotor flux decoupling unit. This implies that a sampling time of 13 s would in
theory result in no calculation constraints. Simulation results are given in Fig. 7 to Fig. 12.
50us
Ias*
Ibs*
Ics*
Rr
TetaPhir_s
Vs_s
Is_s
Phir_s
Wrm_s
Tem_s
siminfo_drv_s
SS_INV_MOTOR
(Slave # 2)
Teta_Phir
Vs
is
Psi_r
Wm
Wm*
Ias*
Ibs*
Ics*
Phir*_s
siminfo_ctrl_s
RrEst
SS_CONTROLLERS
(Slave # 1)
RrEst
siminfo_ctrl_in
Phir*_in
siminfo_drv_in
Tem_in
Wm
Phir_in
Is_in
Vs_in
Rr
Wm*
Rr_out_s
Wm*_out_s
console_data
siminfo
SM_BREAKOUT_BOX
(Master)
console_data
siminfo
Rr
Wm*
SC_USER_INTERFACE
(Console)
Paper number 205-4346 (revised)

15
The performance of the rotor resistance online estimator for a given reference speed is first evaluated.
At the same time, the robustness of the speed controller against the rotor resistance variations is
investigated. As shown in Fig. 7 (a), the reduced order Kalman filter gives an accurate estimate of the
rotor resistance with a good and fast transient response time even for abrupt changes in the actual value.
Fig 7 (b) illustrates the effect of the rotor resistance variations on the performance of speed controller. As
shown, the output of the speed controller keeps tracking the imposed reference after a sudden change of
the rotor resistance. At this point, it is important to mention that this test is very severe since, in practice,
the rotor resistance variation is very slow in comparison with the speed variation.










Fig. 7. (a): Actual and estimated rotor resistance, (b): Actual and given reference speed

On the other hand, the tracking performance of the speed controller is evaluated and, at the same
time, the effect of the speed variation on the rotor resistance estimator is also investigated. As shown in
Fig 8 (a), the actual speed tracks the imposed speed profile even for abrupt changes. In Fig 8 (b), the
reduced order Kalman filter gives an accurate estimate of the rotor resistance with a fast transient
response time.
Fig. 9 (a) and (b) summarize the two above-mentioned scenarios.
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
Actual and Estimated Rotor Resistance
Time (sec)
R
r
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5
-20
0
20
40
60
Actual and Reference Rotor Speed
Time (sec)
W
r
Actual
Estimated
Reference
Actual
(a)
(b)
Paper number 205-4346 (revised)

16










Fig. 8. (a): Actual and given reference speed, (b): Actual and estimated rotor resistance










Fig. 9. (a): Actual and given reference speed, (b): Actual and estimated rotor resistance

Fig. 10 (a) gives the actual and the reference rotor flux. The changes observed in the actual rotor
flux are related to the applied perturbations on the speed and the rotor resistance for the scenario given in
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5
-100
-50
0
50
100
150
200
Actual and reference Rotor Speed
Time (sec)
W
r
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5
0.15
0.2
0.25
0.3
0.35
0.4
0.45
0.5
Actual and Estimated Rotor Resistance
Time (sec)
R
r
Actual
Estimated
Reference
Actual
(a)
(b)
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5
-50
0
50
100
150
200
Actual and Reference Rotor Speed
Time (sec)
W
r
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
Actual and Estimated Rotor Resistance
Time (sec)
R
r
Actual
Estimated
Reference
Actual
(b)
(a)
Paper number 205-4346 (revised)

17
Fig. 9. Fig 10 (a) to 12 give respectively the electromagnetic torque, the stator line currents and the PWM
stator phase voltages.











Fig. 10. (a): Actual and reference flux, (b): Electromagnetic torque











Fig. 11. Stator abc-line currents.
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5
-60
-40
-20
0
20
40
60
Stator Currents: Ias, Ibs and Ics
Time (sec)
I
a
s
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5
-60
-40
-20
0
20
40
60
Time (sec)
I
b
s
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5
-60
-40
-20
0
20
40
60
Time (sec)
I
c
s
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5
0
0.2
0.4
0.6
0.8
1
Actual and Reference Rotor Flux
Time (sec)

0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5

-100
-50
0
50
100
150
Electromagnetic Torque
Time (sec)
T
e
Actual
Reference
(b)
(a)
Paper number 205-4346 (revised)

18










Figure 12. Stator abc-phase voltages.


7. Conclusions

The authors presented in this paper a PC-Cluster based fully digital real-time simulation of an indirect
field-oriented speed controller for an induction motor using RT-Lab software package. Satisfactory
results are obtained with a fixed time-step of 50 s, which is many times smaller than what could be
achieved by the latest available Digital Signal Processors.
The investigation consisted in testing, for different scenarios, during simulation and in real-time the
robustness of the proposed indirect field-oriented adaptive speed controller against rotor resistance
variations. This achievement is, in authors opinion, reported for the first time in this paper and is
believed to be an important contribution to rapid prototyping of high performance induction machine
controllers.

0 0.005 0.01 0.015 0.02 0.025 0.03
-400
-200
0
200
400
Zoomed Stator PWMVoltages: Vas, Vbs and Vcs
Time (sec)
V
a
s
0 0.005 0.01 0.015 0.02 0.025 0.03
-400
-200
0
200
400
Time (sec)
V
b
s
0 0.005 0.01 0.015 0.02 0.025 0.03
-400
-200
0
200
400
Time (sec)
V
c
s
Paper number 205-4346 (revised)

19
Acknowledgement

This work is supported by the Natural Sciences and Engineering Council of Canada and by the
Foundation of the University of Quebec at Chicoutimi.

References

[1] F. Blaschke, The Principal of Field-Orientation as Applied to the New Transvektor Closed-Loop
Control System for Rotating-Field Machines, Siemens Review XXXIX, (5), 1972, 217-219.
[2] M. Ouhrouche, N. Lchevin and S. Abourida, RT-Lab Based Real-Time Simulation of a Direct
Field-Oriented Controller for an Induction Motor, Proceedings of the 7
th
International Conference
on Modeling and Simulation of Electric Machines, Converters and Systems, Montreal, Canada, Aug
18-21, 2002.
[3] M. Ouhrouche, R. Beguenane and M. Dube-Dallaire, Real-time Simulations: A New Powerful Tool
for Rapid Prototyping of High Performance Induction Machine Controllers, Proceeding of the
IASTED International Conference on Applied Modeling and Simulation, Rhodes, Greece, June 28-
30, 2004.
[4] S. Abourida, "Real-Time PC-Based Simulator of Electric Systems and drives, Proceedings of
APEC2002.
[5] C. Dufour, S. Abourida and J. Belanger, "Real-Time Simulation of Induction Motor IGBT Drive on
a PC-Cluster, Proceedings of the International Conference on Power Systems Transients
(ICPST03), Hong Kong, 2003.
[6] Opal-RT Technologies Inc., RT-Lab Users Guide.
[7] Y. Fu, Adaptive and Decoupled Control for Induction Machines, Ph.D. Thesis (in French),
University of Montpellier II, France, 1991.
Paper number 205-4346 (revised)

20
[8] S. Wade, M.W. Dunnigan, & B.W. Williams, Modeling and Simulation of Induction Machine Vector
Control With Rotor Resistance Identification, IEEE Transactions on Power Electronics, 121(3),
1997, 495-506.
[9] L.C. Zai, C.L. Demarco, & T.A. Lipo, An Extended Kalman Filter Approach to Rotor Time Constant
Measurement in PWM Induction Motor Drives, IEEE Transactions on Industry Applications, 28(1),
1992, 96-104.
[10] F.J. Lin, Robust Speed-Controlled Induction Motor Drive Using EKF and RLS Estimators, IEE
Proceedings, Power Applications, 143(3), 1995, 186-192.
[11] R.E. Kalman, R.S. Bucy, New Results in Linear Filtering and Prediction Theory, Journal of Basic
Engineering, 1961, 95-108.
[12] P. Baracos, G. Murere, C.A. Rabbath and W. Jin, Enabling PC-Based HIL Simulation for
Automotive Applications, Proceedings of the IEEE International Electric Machines and Drives
Conference (IEMDC01), Cambridge, MA, USA, 2001.
[13] M. Glesner, A., Kirschbaum, F.M., Renner and B., Voss, State-of-the-Art in Rapid Prototyping for
Mechatronic Systems, Pergamon, Mechatronics (12), 2002, pp. 987-998.
[14] M. Linjama, T. Virvalo, J. Gustafsson, J. Lintula, V. Aaltonen, M. Kivikoski, Hardware-in-the-Loop
Environment for Servo System Controller Design Tuning and Testing, Elsevier, Microprocessors and
Microsystems (24), 2000, pp. 13-21.
[15] J. Chiasson and L. Tolbert, A Library of Simulink Blocks for Real-Time Control of HEV Traction
Drives, Proceedings of the Future Car Congress, june3-5, 2002, Arlington, Virginia, USA.






Paper number 205-4346 (revised)

21
Biographies

Mohand Ouhrouche received the Ingniorat dtat degree with honors in Electrical Engineering from
Bjaia University, Algeria, in 1988, the M.Sc. and Ph.D. degrees in Electrical Engineering from cole
Polytechnique de Montral, Canada, respectively in 1992 and 1998. From 1993 to 1995 he worked as an
assistant researcher at cole Polytechnique de Montral on integration of Non-Utility Generators into the
Hydro-Quebec's distribution system. In 1997, he joined Hydro-Quebec Research Center (LTEE),
Canada, as a scientific researcher to carry out research on power quality issues. Since December 1998, he
has been with the University of Quebec at Chicoutimi, Canada, where he is currently an Associate
Professor of Electrical Engineering and the Director of the Electric Machines Identification and Control
Laboratory. His areas of interest include System Identification, Rapid Control Prototyping, Power
Quality, Electric Machines and Drives. Dr. Ouhrouche is a registered engineer in the province of Quebec.

Rachid Beguenane received the B.Sc degree in Electronic Engineering from Algiers University of
Sciences and technology, Algeria, in 1988. He earned his DEA and Ph.D. degrees, both in Electrical
Engineering from CNAM, Paris, France in 1991 and 1994 respectively. From 1995 to 1997, he hold a
teaching and research position at the Professional Institute of Amiens, France. During 1997/1998, he was
a post-doctoral researcher at the University of Ottawa, Canada. From 1998 to 2002, he has been
employed as ASIC/FPGA design engineer by Telexis inc., Nortel Networks, and Prova Scientific Corp. in
Ottawa and Montreal, Canada. He is presently employed by the University of Quebec at Chicoutimi,
Canada as Assistant Professor in Electrical and Computer Engineering since august 2002.

Andrzej M. Trzynadlowski received the M.S. degree in electrical engineering, the M.S. degree in
electronics, and the Ph.D. degree in electrical engineering from the technical university of Wroclaw,
Wroclaw, Poland, in 1964, 1969 and 1974 respectively. From 1966 to 1979, he was a member of the
faculty of the technical university of Wroclaw. He was later with the university of salahuddin, Iraq,
Paper number 205-4346 (revised)

22
University of Texas, Arlington, and the university of Wyoming. Since 1987, he has been with the
University of Nevada, Reno, where is currently a Professor of Electrical Engineering. He has published
three books , authored or coauthored more than 140 publications on power electronics and electric drives,
and he is the holder of 12 patents. Dr. Trzynadlowski is an Associate Editor of the IEEE Transactions on
Power Electronics and IEEE Transactions on Industrial Electronics, and a member of the Industrial
Drives and Industrial Power Converter Committees of the IEEE Industry Application Society.

Jogendra S. Thongam completed M.Sc.(Eng.) from Kharkov Polytechnic Institute, former USSR with
full soviet government scholarship in 1991 specializing in Electric Drive and Automation of Industrial
Installations. He is a Senior Lecturer in Electrical Engineering with the North Eastern Regional Institute
of Science and Technology, India. He is presently on study leave for pursuing Ph.D. in the University of
Quebec at Chicoutimi, Canada.

Mathieu Dub-Dallaire received Bachelor degree in mechanical engineering from cole Polyechnique de
Montral, Canada. His work experience includes nuclear power plant simulation for operator training and
now at Opal-RT, as an application engineer, real-time simulation of plant and controller models for
electromechanical and automotive applications.

Appendix A List of symbols

V
sd
, V
sq
: stator voltages in the synchronously rotating reference frame.
I
sd
, I
sq
: stator currents in the synchronously rotating reference frame.

rd
,
rq
: rotor fluxes in the synchronously rotating reference frame.
V
s
, V
s
: stator voltages in the stationary reference frame.
I
s
, I
s
: stator currents in the stationary reference frame.
Paper number 205-4346 (revised)

23

r
,
r
: rotor fluxes in the stationary reference frame.

r
: rotor speed (rad/s).

e
: synchronous speed (rad/s).
R
s
, L
s
: stator resistance and self inductance.
R
r
, L
r
: rotor resistance and self inductance.
L
m
: magnetic inductance.
p : number of pole pairs.
J : total rotor inertia constant (kg.m
2
).
F : damping coefficient (N.m.s).
T
l
: load torque (N.m).

Appendix B Real-time models implementation in RT-Lab environment

Fig.B1 to B4 show the details of the four block diagrams, respectively SS_CONTROLLERS,
SS_INV_MOTOR, SM_BREAKOUT_BOX and SC_USER_INTERFACE, used for IFOC
implementation using RT-Lab simulator as shown in Fig. 6. These block diagrams are actually Simulink
models grouped to form a subsystem. Once the models are grouped into console (SC_), data acquisition
(SM_) and computation (SS_) subsystems, special blocks called OpComm blocks must be inserted into
the subsystems [6]. These are simple feed-through blocks that intercept all incoming signals before
sending them to computation blocks and provide information to RT-Lab concerning the type and size of
these intercepted signals. This signal interception is mandatory because when a simulation model runs in
the RT-Lab environment; all connections between the main subsystems are replaced by hardware
communication links.
Memory blocks are added in each subsystem in order to take into account delays that might occur
during information exchange.
Paper number 205-4346 (revised)

24












Fig. B1. Details of the Slave #1 (SS_CONTROLLERS) block diagram










Fig. B2. Details of the Slave #2 (SS_INV_MOTOR) block diagram


6
RrEst
5
siminfo_ctrl_s
4
Phir*_s
3
Ics*
2
Ibs*
1
Ias*
siminfo
Teta_Phir
Ids*
Iqs*
Ias*
Ibs*
Ics*
dq-ABC
Wm*
Wm
RrEst
Phir*
Tem*
Speed Controller
v_s
i_s
w
RrEst
Rr Estimator
(Reduce-Order
Kalman filter)
Psi*
Wm
RrEst
Tem*
Ids*
Iqs*
Teta_Phir
Rotor Flux
Decoupling
50e-6
OpComm
[vs]
[Teta_Phir]
[Psi_r]
[Teta_Phir]
[Wm]
[is]
[Wm_]
[vs]
[vs]
[Wm]
[Wm]
[Teta_Phir]
[Wm]
[Wm_]
6
Wm*
5
Wm
4
Psi_r
3
is
2
Vs
1
Teta_Phir
7
siminfo_drv_s
6
Tem_s
5
Wrm_s
4
Phir_s
3
Is_s
2
Vs_s
1
TetaPhir_s
siminfo
50e-6
OpComm1
50e-6
OpComm
Load
Torque
Vas
Vbs
Vcs
TL
Rr
Ias
Ibs
Ics
Phir
TetaPhir
Tem
Wrm
Induction
Motor
Ias*
Ias
Ibs*
Ibs
Ics*
Ics
Um
Vas
Vbs
Vcs
Hysterisis
Inverter
[Rr]
311
DC Bus
Voltage
[Rr]
Actual Rr
(User defined
profile to test
the estimator's
performance)
4
Rr
3
Ics*
2
Ibs*
1
Ias*
Paper number 205-4346 (revised)

25











Fig. B3. Details of the Master (SM_BREAKOUT_BOX) block diagram










Fig. B4. Details of the Console (SC_USER_INTERFACE) block diagram
4
siminfo
3
console_data
2
Wm*_out_s
1
Rr_out_s
OpWriteFile
acq. group 27
OpWriteFile
acq. group 26
1e-3
From SS
1e-3
From SC
11
Wm*
10
Rr
9
Vs_in
8
Is_in
7
Phir_in
6
Wm
5
Tem_in
4
siminfo_drv_in
3
Phir*_in
2
siminfo_ctrl_in
1
RrEst
2
Wm*
1
Rr
1
vs
siminfo_
1
one
is
Wm
Tem
100
Slider Gai n
Speed reference
0.4
Slider Gain
Rr actual value
Phi_r
A=1
OpComm
em
em
2
siminfo
1
console_data