Assessment of Direct Torque Control

Strategies for induction Machine

Hamid CHAIKHY*, Mohamed KHAFALLAH*, Abdallah SAAD*, Khalid CHIKH* and Mouna ES-SAADI*
*
Hassan II University-ENSEM
BP 8118, Oasis, Casablanca, Morocco.
chaikhy.hamid@gmail.com; m_khafallah@yahoo.fr; saad.abdal@gmail.com; genielectrique@gmail.com;
mounaessaadi2@gmail.com

Abstract— Direct Torque Control (DTC) has becoming the DTC-based induction machine drive [14-15]. As shown in [14-
industrial standard for Induction Machine (IM) torque control 15] fuzzy logic controller controls the duty ratio of switching
by means of very simple control schemes. This paper presents a states such that the torque ripple can be significantly reduced.
contribution for detailed comparison of Direct Torque Control
strategies (Classical DTC, SVM-DTC and Fuzzy DTC) for The above DTC strategies present reciprocally advantages
Induction Machine. The performances of those various and disadvantages. The aim of this work is to compare various
methodologies are evaluated and compared by simulation in direct torque control methodologies (Classical DTC, SVM-
terms of torque and current ripples, and transient response. DTC and FUZZY-DTC) when applied to IM drives and point
out the strengths and weaknesses that can help to make a
Keywords- Induction Machine; Classical DTC; SVM-DTC; choice between them for a particular application.
FUZZY- DTC.
We will begin this paper by presenting the principle of
I. INTRODUCTION Classical DTC scheme and it will be followed by the principle
of SVM-DTC and FUZZY-DTC. This will be followed by an
INDUCTION machines (IM) have been widely used in explanation of the comparative analysis and simulation tests
industry because of their advantages: simple construction, performed, the main characteristics, advantages and
ruggedness, reliability, low cost, and a minimum of disadvantages of both methods will be showed and discussed in
maintenance. However, due to their highly-coupled nonlinear the results of the simulation tests. Finally, this work will give a
structure, a high-performance control of IM is a challenging summary of the comparison between those strategies not only
problem. Several control strategies have been used for regarding the performance but also considering the control
induction machine drives as: Field Oriented Control (FOC) [1], structure and requirements of both systems.
Direct Torque Control (DTC) [2], and direct self control (DSC)
[3].
Direct torque control (DTC) was introduced in the mid II. PRINCIPLE OF DTC
1980s, it has been widely used for IM drives. It offers high
performances in terms of simplicity in control and fast Since M. Depenbrock and I. Takahashi proposed direct
electromagnetic torque response. In this method, Stator voltage torque control (DTC) for IM in the middle of 1980’s [2], more
vectors is selected according to the differences between the than decade has passed. It is getting more and more popular
reference and actual torque and stator flux linkage. nowadays. The basic idea of DTC for IM is to control the
Furthermore, the inverter is directly controlled by the voltage space vectors properly, which is based on the
algorithm, i.e. a modulation technique for the inverter is not relationship between the slip frequency and torque.
needed [2]-[5]. However, the presence of hysteresis controllers The basic DTC scheme is shown in Figure.1.
leads to a non-constant switching frequency operation. The
time discretization, due to the digital implementation, and the Udc
limited number of available voltage vectors determine the
presence of current and torque ripples [6]. CТe
Т* INVERTER
Recently, much effort has been made to improve the Switching IM

performance of DTC, especially by reducing the ripples, and table

Ψs* Cψs
by fixing the inverter switching frequency [7]-[12]. In order to
apply this principle, the control system should be able to θs
generate the desired voltage vector, using the Space Vector Ψs
Flux and
Modulation technique (SVM). These methods require more Т
torque
complex control schemes than basic DTC scheme and present
estimator
motor parameters dependence.
Recently, Fuzzy Logic Control has been successfully
applied to many control problems, as they need no accurate
Figure.1: block diagram of the DTC control technique
mathematical models of the uncertain nonlinear systems under for IM
control [12-13]. Fuzzy Logic Control has also been applied to

978-1-4673-1520-3/12/$31.00 ©2012 IEEE

 By using a α-β stationary stator reference frame, the III. SVM-DTC
stator flux linkage ψs and electromagnetic torque T are
calculated by using: In this paragraph a SVM-DTC strategy is analyzed in
order to overcome the problems of classical DTC based on
ψ ψα ψβ
switching table strategy. The main idea of a SVM-DTC
(1) strategy is to force torque and stator flux magnitude to achieve
Where: their reference values in one sampling period by applying a
ψ V R i dt
suitable stator voltage vector generated by SVM. It includes
(2)
the PI regulators for the torque (deducted by the usual
ψ V R i dt (3) method), the stator flux and torque estimator block and stator
voltage calculation block.
The angle θ is equal to:
The stator flux angle θ desired is derived from:
artg
Ψβ
θ ω Ω dt
θ (4)
Ψα - (6)
p ψα iβ ψβ iα
The reference stator voltages V and V
Т (5)
in the α-β frame
The error between the estimated torque T and the are calculated by the following equations:
reference torque T* is the input of a three level hysteresis
/0 1 /^0
comparator, whereas the error between the estimated stator V R i (7)
3
flux magnitude ψ and the reference stator flux magnitude ψ
/4 1 /^4
V R i
is the input of a two level hysteresis comparator.
(8)
3
The selection of the appropriate voltage vector is
based on the switching table given in Table I. The input With :
quantities are the flux sector and the outputs of the two ψ ψ 567 cos θ (9)
hysteresis comparators.
ψ ψ 567 sin θ (10)
To determine stator vector voltage to be applied, we begin
by dividing the circular trajectory of the stator flux into six And
symmetrical sectors referred as the inverter voltage vectors.
Then, we study the effect of each stator vector voltage on the ψ^ V R i dt (11)
flux and torque: When the stator flux is in sector Si, the vectors ψ ^
V R i dt (12)
Vi+1or Vi-1 are selected to increase its amplitude, and Vi+2 or Vi-
2 to decrease it. However Vi+1 or Vi+2 increase the torque and
Vi-1 or Vi-2 decreases it. Figure. 2 shows the effect of those IV. FUZZY-DTC
different choices in sector S1.
In the recent years, the Fuzzy Logic Control (FLC) have
θs
been successfully applied to many control problems, as they
2
3
Sector
V3 (FD,TI)
V2 (FI,TI)
need no accurate mathematical models of the uncertain
3
Sector
nonlinear systems under control [12-15].Therefore FLC is
Δψs2
2 Δψs3 relatively easy to implement. This is achieved by converting
Sector
Δψs3
Δψs1
V6 (FI,TD)
the linguistic control strategy of human experience and
4
V5 (FD,TD)
Δψs5 Δψs6
knowledge into an automatic control strategy. FLC has become
4 1 very popular in the field of industrial control applications.
Sector
1 By analyzing the structure of the switching table, we note
Sector
that it can be printed as fuzzy rules. Therefore, a first FLS can
5 replace the switching table and the hysteresis controllers,
Sector
6
whose inputs are the errors on the flux and torque, and the
6 argument θs of the stator flux. Whereas its outputs are the
5
Figure 2. Effect of different Vs choices on the torque states of the voltage inverter switches.
and stator flux in sector S1. The following figure.4 shows the block diagram of the
FUZZY-DTC.
TABLE I. BASIC SWITCHING
Udc
Outputs of hysteresis Sector FLC
comparators 1 2 3 4 5 6
!$ =-1 %
V %&
V %'
V %(
V %)
V %*
V Т* CТe INVERTER
!"# = -1 !$ =0 %+
V %
V %+
V %
V %+
V %
V
!$ =+1 %)
V %*
V %
V %&
V %'
V %(
V
IM

!$ =-1 %&
V %'
V %(
V %)
V %*
V %
V C* Cψs
!"# = +1 !$ =0 %
V %+
V %
V %+
V %
V %+
V
!$ =+1 %(
V %)
V %*
V %
V %&
V %'
V
Ψs^ θs (7)
Flux and
Flux and
Т^ Torque
torque
Estimator
estimator

Figure 4. Block diagram of the FUZZY-DTC.

 Ψs* Vsα
Stator
INVERTER
SVM IM
Voltage
Т* ωsl* θs*
PI Calculation
Vsβ
^
Т Ω
Ψs^
Flux ^
Flux and
and
Torque
Torque
Estimator
Estimator

Figure 3. Block diagram of the SVM-DTC

Six linguistic values, θ1 to θ6 are used to fuzzify the • Time response to a step variation of the torque command, at
domain of stator flux linkage angle. Two linguistic values, different rotor speeds.
negative, and positive denoted as N and P respectively are used
to fuzzify flux linkage error domain. Three linguistic values, The simulation has been carried out by using MATALAB
negative, zero, and positive as N, Z, and P respectively are SIMULINK on 1.5KW induction machine.
used to fuzzify torque error domain.
The fuzzy rules base with 42 rules can be obtained from the
following Table II.

TABLE II. FUZZY –SWITCHER RULES

θ1 and θ7 θ2 θ3
Cψs Cψs Cψs
P N P N P N
CT CT CT
P V5 V6 P V6 V1 P V1 V2 Torque error membership functions
Z V0 V7 Z V7 V0 Z V0 V7
N V3 V2 N V4 V3 N V5 V4

θ4 θ5 θ6
Cψs Cψs Cψs
P N P N P N
CT CT CT
P V2 V3 P V3 V4 P V4 V5
Z V7 V0 Z V0 V7 Z V7 V0 Stator flux error membership functions
N V6 V5 N V1 V6 N V2 V1

The corresponding inputs and output variables membership

functions graphs are shown in Figure 5.
In this paper, Mamdani’s fuzzy rule base model is
developed to perform the function of proposed fuzzy switcher.
Stator flux angle membership functions

V. SIMULATION RESULTS AND DISCUSSION

This section is aimed to give a contribution for a comparison

between direct torque control methodologies (Classical DTC,
SVM-DTC and FUZZY-DTC). The quantities used to evaluate the
performance of those strategies, in steady state and transient
conditions, are: State switches membership functions

• Current and torque ripple values in steady-state operating Figure 5. Membership functions of the FUZZY-switcher
conditions.
A. transient performance B. transient performance
In steady-state operating conditions. The torque, the stator The transient performance of the twose schemes has been
current waveform, the stator flux and the stator current compared analyzing the response to a step variation of the
harmonic spectrum obtained with Classical DTC, SVM-DTC torque command from 0 Nm to 10 Nm (rated torque), at
and FUZZY-DTC are shown respectively in Figure 6. The different rotor speeds.
rotor speed is 1440 rpm and the reference torque is 10 Nm
(rated torque). Figure.7 show the results obtained for a step variation of the
torque command from 0 to 10 Nm, at different rotor speeds
As it is possible to see, in all the operating conditions the (1440 rpm, 720 rpm and 100 rpm).
behaviors of SVM-DTC and FUZZY-DTC are characterized
by a good performances. It can be observed that there is a These results show that different methods provide high
significant reduction in torque and flux ripples. Hence the performances response with quicker torque dynamics. These
harmonic distortion can also reduce; this can be observed in results show that using the classical DTC scheme, a better
stator current waveform and in stator current harmonic torque response can be achieved in terms of settling time. The
spectrum. different dynamic behavior is due to presence of PI regulators
in SVM-DTC scheme, which delay the torque response.
For FUZZY-DTC it can be noticed that stator flux vector
describes a trajectory almost circular, and stator current
waveform is nearly sinusoidal.
For SVM-DTC the stator flux waveform and its circular
trajectory, line-to-line SVM machine voltage with 8 kHz
switching frequency. Nearly sinusoidal stator current
waveform is also achieved.

12
12 12
10
10 10
8
Torque (Nm)

8 8
Torque (Nm)

Torque (Nm)
6 6 6

4 4 4

2 2 2

0 0 0
0.5 0.505 0.51 0.515 0.52 0.525 0.53 0.535 0.54 0.5 0.505 0.51 0.515 0.52 0.525 0.53 0.535 0.54 0.5 0.505 0.51 0.515 0.52 0.525 0.53 0.535 0.54
Time (s) Time (s) Time (s)

6 6 6

4 4 4
Stator current (A)

Stator current (A)

Stator current (A)

2 2 2

0 0 0

-2 -2 -2

-4 -4 -4

-6 -6 -6
0.5 0.505 0.51 0.515 0.52 0.525 0.53 0.535 0.54 0.5 0.505 0.51 0.515 0.52 0.525 0.53 0.535 0.54 0.5 0.505 0.51 0.515 0.52 0.525 0.53 0.535 0.54
Time (s) Time (s) Time (s)

Classical DTC SVM-DTC FUZZY-DTC

SVM-DTC
Figure.6.(a) Stator current. (b) Torque. (c) Stator flux. (d) stator current harmonic spectrum
 12 12 12

10 10 10

8 8

Torque (Nm)
Torque (Nm)
8
Torque (Nm)

6 (1440 rpm) 6 (1440 rpm) 6 (1440 rpm)

4 4 4

2 2 2

0 0
0 0.49 0.495 0.5 0.505 0.51 0.49 0.495 0.5 0.505 0.51
0.49 0.495 0.5 0.505 0.51
Time (s) Time (s)
Time (Nm)
12 12
12
10 10
10
8 8

Torque (Nm)
Torque (Nm)
8
Torque (Nm)

6 (720 rpm) 6
(720 rpm) 6
(720 rpm)
4 4
4
2 2
2
0 0
0 0.49 0.495 0.5 0.505 0.51 0.49 0.495 0.5 0.505 0.51
0.49 0.495 0.5 0.505 0.51 Time (s) Time (s)
Time (Nm)
12
12
10 12
10
8 10
Torque (Nm)

8
(100 rpm) (100 rpm)
Torque (Nm)

6 8

Torque (Nm)
6 (100 rpm) 4 6
4
2 4
2
2
0
0 0.49 0.495 0.5 0.505 0.51
0.49 0.495 0.5 0.505 0.51 Time (s) 0
0.49 0.495 0.5 0.505 0.51

Classical DTC SVM-DTC FUZZY-DTC

Figure.7. Evolution of torque response at different rotor speeds

Finally, Table.V presents a Comparison of all different VI. CONCLUSION

methods not only regarding the performance but also
This paper has presented a contribution for a detailed
considering the control structure and requirements of both comparison of Direct Torque Control strategies (Classical
systems. DTC, SVM-DTC and Fuzzy DTC) for Induction Machine. The
TABLE III
SUMMARY OF THE COMPARISON BETWEEN DTC STRATEGIES performances of those various methodologies are evaluated and
compared by simulation in terms of torque and current ripples,
Classical DTC SVM-DTC FUZZY-DTC and transient response. Both methods (SVM-DTC and
Dynamic FUZZY-DTC) provide a high performance response with
response for Quicker Quicker Quicker quicker torque dynamics and better steady-state behavior by
torque reducing torque and current ripple. It can be noted that SVM-
Steady-state DTC requires more complex control scheme and the
Low ripple
behavior for High ripple and Low ripple and
torque, and distortion
and
distortion
knowledge of some machine parameters.
distortion
stator currents
Behavior at
Not good good good
Low Speed TABLE VI
Requirement
PARAMETERS OF INDUCTION MOTOR
of rotor No Yes No
position Stator resistance Rs=5,63 Ω
Variable, depending Variable, Rotor resistance Rr =2,62 Ω
Switching on the operating depending on the Rotor/ Stator inductances Ls= Lr=0,382H
Constant Magnetizing Inductance M=0,364 H
frequency point and operating point and
during transients during transients Moment of inertia J=0,010Kg.m2
Spread spectrum, Low noise Viscous inertia F=0,015 N.m.s.rad-1
Spread spectrum,
Audible noise high noise especially at a fixed Rated power 1,5 kw
low noise
at low speed frequency
Control tuning Hysteresis bands PI gains Fuzzy Table
Complexity VII. REFERENCES
and processing Lower Higher Lower
requirements
[1] F. BLASHKE, The principle of fiels-orientation as applied to the
Transvector closed-loop control system for rotating-field machines ’ in
Siemens Reviev 34, 1972, pp. 217–220.
[2] I. TAKAHASHI, T. NOGUCHI, A new quick-response and high
efficiency control strategy of an induction machine ’ IEEE Trans. On
Industrial Application, Vol. IA-22, no.5, Sept./Oct. 1986, pp. 820–827.
[3] M. DEPENBROCK, Direct Self Control of Inverter-Fed Induction VIII. BIOGRAPHIES
Machines ‘ IEEE Trans. on Power Electronics, Vol. PE-3, no.4, Oct.
1988, pp. 420–429.
[4] JIA-QIANG YANG, JIN HUANG ″ A New Full-Speed Range Direct Hamid Chaikhy was born in Morocco in 1975.
Torque Control Strategy for Induction Machine″, Proceedings of the He received B.Sc., M. Sc. degrees at ENSET from
Third International Conference on Machine Learning and Cybernetics, the University of Hassan II, Mohammedia in 1998,
Guangzhou, 26-29 August 2004. from University of Ibno Zohr, Agadir in 2005
[5] A. Abbou. H. Mahmoudi “ Real Time Implementation of a Sensorless respectively, all in Electrical Engineering. He is
Speed Control of Induction Motor using DTFC Strategy,” International currently working towards the Doctorate degree in
conference on multimedia computing and systems proceedings, pp. 327- Electrical Engineering at the institute of electrical and mechanical
333. Ourzazzate 2-4 April 2009.
engineering (ENSEM) of the University of Hassan II, Casablanca I.
[6] Y . Li, J. Shao, B. Si, “Direct Torque Control of Induction Motors for
His current research interests are the areas of field oriented
Low Speed Drives Considering Discete Effects of Control and Dead-
Time of Invcrtcrs”, CO!$ Rec. IEEE-/AS Ann. Meeting, pages 78 1-788, controllers and motor drives.
1997. Chaikhy.hamid@gmail.com
[7] P.Z. Grabowski, M.P. Kazmierkowski, B.K. Bose, F. Blaabjerg, A
simple directtoque neuro-fuzzy control of PWM-inverter-fed induction
motor drive, IEEE Trans. Ind. Electron. 47 (4) (2000) 863–870. Mohamed Khafallah was born in Morocco in
[8] t torque control for induction motor sensorless drive, IEEE Trans. Ind. 1964. He received B.Sc., M.Sc. and Doctorate
Appl. 36 (1) (2000) 122–130. degrees from Hassan II University, Casablanca II,
[9] Y.S. Lai, J.H. Chan, A new approach to direct toque control of induction in 1989, 1991 and 1995 respectively, all in
motor drive for constant inverter switching frequency and torque ripple Electrical Engineering. In 1995 he joined the
reduction, IEEE Trans. Energy Convers. 16 (3) (2001) 220–227. Department of Electrical Engineering at the
[10] Yuttana Kumsuwana, Suttichai Premrudeepreechacharna, Hamid A. National School of Electricity and Mechanics
Toliyat , “Modified direct torque control method for induction motor (ENSEM), Hassan II University, Casablanca. His current research
drives based on amplitude and angle control of stator flux“Electric interests are in the application of power electronics converts and
Power Systems Research 78 (2008) 1712–1718.
motor drives.
[11] L. Tang, L. Zhong,A.F. Rahman, Y. Hu, An investigation of a modified
direct torque control strategy for flux and torque ripple reduction for m_khafallah@yahoo.fr
inductionmachine drive systemwith fixed switching frequency, in:
International Conference on IAS, vol. 2, Pittsburgh, Pennsylvania, 2002,
pp. 837–844. Abdallah Saad was born in Morocco in 1956. He
[12] R.Toufouti S.Meziane ,H. Benalla, “Direct Torque Control for Induction received the Engineer and Doctor of Engineering
Motor Using Fuzzy Logic” ICGST Trans. on ACSE, Vol.6, Issue 2, pp. degrees from National Polytechnic Institute of
17-24, June, 2006. Grenoble – France – respectively in 1980 and 1982.
[13] Jia-Qiang Yang, Jin Huang, ″ Direct Torque Control System for From 1982 to 1986, he was Researcher at French
Induction Motors With Fuzzy Speed Pi Regulator″ Proceedings of the National Center for Scientific Research (CNRS) -
Fourth International Conference on Machine Learning and Cybernetics,
Guangzhou, 18-21 August 2005.
Electrostatics and Dielectric Materials Laboratory –
Grenoble. After receiving the Doctor of Physical
[14] I. G. Bird and H. Zelaya De La Parra, “Fuzzy logic torque ripple
reduction for DTC based ac drives,” IEE Electron. Lett., vol. 33, no. 17,
Sciences degree in 1986, he joined Hassan II University of Morocco.
pp. 1501–1502, 1997. Professor of electrical engineering, he has several scientific and
[15] A. Arias, J. L. Romeral, E. Aldabas, and M. G. Jayne, “Fuzzy logic
educational responsibilities. His main fields of interest are High
direct torque control,” Proc. IEEE Int. Symp. Ind. Electron., pp. 253– Voltage and Electrical Insulations, modeling and control, renewable
258, 2000. energy integration.
[16] SANDA VICTORINNE, PATURCA, MIRCEA COVRIG, saad.abdal@gmail.com
COSTIN CEPISCA, SORIN DAN GRIGORESCU, STEFAN
GHEORGHE, “Direct Torque Control with Variable Voltage Duty Ratio
for Inverter Fed Induction Motor Drives”
10th WSEAS Int. Conf. on AUTOMATIC CONTROL, MODELLING Khalid Chikh was born in Morocco in 1982. He
& SIMULATION (ACMOS'08), Istanbul, Turkey, May 27-30, 2008. received the Bachelor’s and Master’s degrees in
[17] Lokriti, A. Zidani, Y. Doubabi, “Fuzzy logic control contribution to the electrical engineering from the Faculty of Science
direct torque and flux control of an induction machine,” Proceedings of and Technology in 2005 and the Faculty of Science
the International Conference on Multimedia Computing and Systems Semlalia in 2007, Marrakech, Morocco,
(ICMCS), Ouarzazate, pp. 1-6, 7-9 April 2011. respectively. He prepares, currently, a Doctorate
[18] SALIH BARIS OZTURK, “Modelling, Simulation and Analysis of Thesis in Electrical Engineering Department at the
Low-Cost Direct Torque Control of PMSM using Hall-Effect Sensors,” National School of Electricity and Mechanics
MASTER OF SCIENCE, December 2005. (ENSEM), Hassan II University, Casablanca, Morocco. His current
research interests are in the application of power electronics converts,
motor drives, renewable energy and energy storage.
genielectrique@gmail.com