2017 International Conference on Intelligent Computing,Instrumentation and Control Technologies (ICICICT)

Comparative Analysis of Different Control

Approaches of Direct Torque Control Induction
Motor Drive

Prathmesh J. Koratkar Dr. Anupa Sabnis

M.Tech Student, Department of Electrical Engineering Associate Professor, Department of Electrical Engineering
BVB’s Sardar Patel College of Engineering, BVB’s Sardar Patel College of Engineering,
Mumbai, India Mumbai, India
Email: [email protected] Email: [email protected]

Abstract— Direct torque control gives fast and robust response switching modes. The selection is such that flux and torque
in induction motor drives. This paper deals with Direct torque errors are restricted within respective flux and torque hysteresis
control technique of induction motor using different approaches. bands to obtain fast torque response. Equation (1) shows torque
The first is Conventional direct torque control which uses (ܶ௘ ) equation of induction motor (IM) with number of poles ‘P’.
hysteresis controllers. The second one is direct torque control
ଷ ௅೘
using space vector modulation which uses predictive controllers ܶ௘ ൌ ቀ ቁ ܲ ȁ‫ܨ‬௦ ȁȁ‫ܨ‬௥ ȁ •‹ ፽ (1)
and the third is simplified direct torque control space vector ଶ ௅ೞ ௅ೝ
modulation which uses imaginary switching time concept. The where, ‫ܨ‬௦ is stator flux linkage space vector, ‫ܨ‬௥ is rotor flux
MATLAB/SIMULINK platform is used for analysis of these linkage space vector. ፽ is the angle between ‫ܨ‬௦ ܽ݊݀‫ܨ‬௥ . The rotor
techniques. Comparison of all the three approaches is discussed. time constant of standard squirrel cage induction motor is large.
Thus, rotor flux linkage changes slowly as compared to the stator
Keywords— direct torque control, space vector pulse width
flux linkage. If the stator flux linkage and rotor flux linkage are
modulation, predictive controllers, induction motor drives.
assumed to be constant, it follows from equation (1) that the
I. INTRODUCTION electromagnetic torque can be rapidly changed by changing ፽ in
the required direction (decided by the torque command). This is
Direct Torque Control (DTC) is popularly used for an essence of direct torque control. The angle ፽ can be easily
controlling induction motor drives since its introduction in 1986. changed by appropriate selection of stator voltage space vector.
It is not only simple but also produces fast torque and flux
The block diagram of basic DTC IM drive is shown in Fig.1. ܶ௘
control [1], [2]. Basic DTC has drawback of high torque and flux
and ‫ܨ‬௦ are estimated from induction motor dynamic model [9].
ripples. The performance of basic DTC has been improved with
The magnitudes of estimated torque and flux are compared with
much efforts in recent years. The association of DTC with space
reference values and the error signal is obtained. Then hysteresis
vector modulation (SVM) is studied in [3]-[6]. This association
controllers are fed with this error signal. Two level comparator
requires the calculation of conduction times of inverter switches
is used for flux and three level comparator is used for torque.
in each modulation period. This method involves sector
Optimal switching is to be selected for inverter such that the flux
determination and angle calculation of voltage vector which
and torque are within permissible limits of hysteresis bands.
makes it complicated and it consumes more memory. Another
There are six active voltage vectors and two zero voltage
simple approach, namely DTC SVM with imaginary switching
vectors. Inverter output is fed to the induction motor.
time concept is developed which does not require sector
determination, consumes less memory and produces same A. Flux and torque estimation
results. It uses the concept of imaginary switching times [7]. In Flux and torque are estimated from actual inverter output
this paper, these three approaches are simulated and studied voltages and measured motor currents. Following equations are
using MATLAB/SIMULINK platform. The results are implemented in flux and torque estimators.
compared which is beneficial for selecting drive depending on • Flux estimation is performed by integrating stator
application. voltages, namely,
In section II, Basic DTC scheme with hysteresis controllers ‫ܨ‬ௗ௦ ൌ ‫׬‬ሺܸௗ௦ െ ݅ௗ௦ ܴ௦ ሻ݀‫ݐ‬ (2)
is explained. Section III gives details about DTC SVM with
predictive controllers. Section IV explains third approach which ‫ܨ‬௤௦ ൌ ‫׬‬ሺܸ௤௦ െ ݅௤௦ ܴ௦ ሻ݀‫ݐ‬ (3)
is DTC SVM with imaginary switching time concepts.
Simulation results are discussed in section V. ȁ‫ܨ‬௦ ȁ ൌ ට‫ܨ‬ௗ௦ ଶ ൅ ‫ܨ‬௤௦ ଶ (4)

II. BASIC DIRECT TORQUE CONTROL

In DTC, it is possible to directly control stator flux linkage where, ‫ܨ‬ௗ௦ and ‫ܨ‬௤௦ are stator flux vectors. ܸௗ௦ and ܸ௤௦ are d and
and electromagnetic torque by selection of optimum inverter q axis stator voltage vectors respectively.

978-1-5090-6106-8/17/$31.00 ©2017 IEEE 831

Authorized licensed use limited to: NATIONAL INSTITUTE OF TECHNOLOGY CALICUT. Downloaded on February 05,2024 at 10:12:45 UTC from IEEE Xplore. Restrictions apply.
 2017 International Conference on Intelligent Computing,Instrumentation and Control Technologies (ICICI

• From estimated flux and measured currents ݅ௗ௦ and ݅௤௦ , Table 1 SWITCHING TABLE
torque is estimated as follows:
ଷ ௉ ‫ܪ‬ி 1 0
ܶ௘ ൌ ቀ ቁ ቀ ቁ ሺ‫ܨ‬ௗ௦ ݅௤௦ െ ‫ܨ‬௤௦ ݅ௗ௦ ሻ (5)
ଶ ଶ

where, ݅ௗ௦ and ݅௤௦ are d and q axis components of stator ‫்ܪ‬ 1 0 -1 1 0 -1
currents. S1 V2 V7 V6 V3 V0 V5
• Reference values (‫ܨ‬௦ ‫ כ‬and ܶ௘ ‫ ) כ‬and estimated values are S2 V3 V0 V1 V4 V7 V6
now compared to obtain error signals.
ο‫ܨ‬௦ ൌ ‫ܨ‬௦ ‫ כ‬െ ‫ܨ‬௦ (6) S3 V4 V7 V2 V5 V0 V1

οܶ௘ ൌ ܶ௘ ‫ כ‬െ ܶ௘ (7) S4 V5 V0 V3 V6 V7 V2

B. Optimum Switching selection S5 V6 V7 V4 V1 V0 V3
Fig.2 shows inverter output voltage vectors in
S6 V1 V0 V5 V2 V7 V4
complex plane. Table 1 shows switching strategy considering
rotation of flux in anti-clockwise direction. The proper voltage
vectors should be applied to produce desired flux. There are six
sectors based on stator flux angle. The sector number and the
output of hysteresis controllers are required for optimum
switching selection. Inverter output voltage vector is defined as,
మഏ మഏ
ଶ
ܸ௦ ൌ ቀ ቁ ܸௗ௖ ሺܵ௔ ൅ ݁ ௝ య ܵ௕ ൅ ݁ ௝ య ܵ௖ ሻ, (8)
ଷ

where, ܸௗ௖ is DC voltage applied to the inverter. ܵ௔ , ܵ௕ , ܵ௖ are

switching pulses.
Lookup tables are used to implement this switching. ‫ܪ‬ி & ‫்ܪ‬
are outputs of flux and torque hysteresis respectively. S1 to S6
are six sectors. V1 to V6 are six active voltage vectors. V0 and
V7 are zero voltage vectors.
C. Disadvantages of Basic DTC
1. High torque ripple causes acoustic noise and
vibrations in drive.
Fig. 1 Implementation of Basic DTC 2. Switching frequency of the inverter is variable.

III. DIRECT TORQUE CONTROL WITH PREDICTIVE

CONTROLLERS
This section explains DTC with predictive controllers. Basic
DTC is combined with SVM and predictive controller. SVM is
used to apply appropriate voltage vectors to the induction motor.
Thus, working with constant commutation frequency is possible.
The method also has an advantage of elimination of hysteresis
comparators and switching table. Error in reference torque and
estimated torque, flux reference, estimated stator flux and stator
currents are fed to the predictive controller. Calculation of
reference voltages in stationary reference frame for space vector
modulation (SVM) block is aim of predictive controller. SVM
block generates pulses for inverter. Fig.3 shows block diagram
of such configuration. Following are the steps of synthesis of
reference voltages for SVM block. Simulink implementation of
space vector modulation is explained in [8].
A. Reference flux calculation
Torque error is processed in PI controller. The output of
Fig. 2 Voltage vectors of inverter and corresponding flux torque PI controller is proportional to slip speed, ‫ݓ‬௦௟ ‫ כ‬. This

978-1-5090-6106-8/17/$31.00 ©2017 IEEE 832

Authorized licensed use limited to: NATIONAL INSTITUTE OF TECHNOLOGY CALICUT. Downloaded on February 05,2024 at 10:12:45 UTC from IEEE Xplore. Restrictions apply.
 2017 International Conference on Intelligent Computing,Instrumentation and Control Technologies (ICICICT)

calculation is not required in this approach. This control

technique uses flux and torque error to determine imaginary
switching times. These times can be used to generate voltage
vector required for tracking demanded load torque. The
adjacent two active voltage vectors (from V1-V6) and zero
voltage vectors (V0 or V7) are applied for certain time for
constructing reference voltage vector in any sector. Time for
which active voltage vectors are applied is called as ‘effective
time’. During effective time only, there is transfer of power
from inverter to motor. Imaginary switching times are
calculated using equations (19) and (20). The output of torque
PI controller is proportional to slip speed. When it is added with
actual rotor speed, synchronous speed of reference flux vector
is obtained. The error between estimated and reference flux
vectors gives imaginary time vector. The components of this
imaginary time in stationary reference frame are calculated as
follows:
Fig. 3 Direct torque control using predictive controller
Voltage equation of induction motor is given by,
ௗிೞ
reference slip speed is added with actual rotor electrical speed. ܸ௦ ൌ ܴ௦ ݅௦ ൅ (15)
ௗ௧
This will give new reference synchronous speed. Integration of
new reference synchronous speed gives argument of reference where, ܸ௦ is stator voltage vector, ‫ܨ‬௦ is stator flux vector. If small
flux. Then, d and q axis reference flux co-ordinates are stator resistance is considered, ο‫ܨ‬௦ ൌ ܸ௦ ο‫ݐ‬
calculated as follows: Separating real and imaginary parts,
‫ݓ‬௠ ൌ ‫ݓ‬௦௟ ‫ כ‬൅ ‫ݓ‬௥ (9) ο‫ܨ‬ௗ௦ ൅ ݆ο‫ܨ‬௤௦ ൌ ሺܸௗ௦ ൅ ݆ܸ௤௦ ሻο‫ݐ‬ (16)
ߠ ൌ ‫ݓ ׬‬௠ ݀‫ݐ‬ (10)
where, ‫ܨ‬ௗ௦ and ‫ܨ‬௤௦ are d and q axis stator flux vectors
‫כ‬ ‫כ‬
‫ܨ‬ௗ௦ ൌ ȁ‫ ݏܨ‬ȁ ‘• ߠ (11) respectively.
‫ܨ‬௤௦ ‫ כ‬ൌ ȁ‫ כ ݏܨ‬ȁ •‹ ߠ (12) From equation (13),
οி೏ೞ ி೏ೞ ‫ି כ‬ி೏ೞ
where, ‫ݓ‬௠ is new reference synchronous speed, ‫ݓ‬௥ is actual ܸௗ௦ ൌ ൌ (17)
ο௧ ο௧
rotor speed. ߠ is argument of reference stator flux vector.
οி೜ೞ ி೜ೞ ‫ି כ‬ி೜ೞ
B. Reference voltage calculation ܸ௤௦ ൌ ൌ (18)
ο௧ ο௧
The following equations give reference voltages for SVM: Where, ο‫ ݐ‬ൌ ܶ௦ , sampling time.
ி೏ೞ ‫ି כ‬ி೏ೞ
ܸௗ௦ ‫ כ‬ൌ ൅ ܴ௦ ݅ௗ௦ (13) Following equations give imaginary switching times in
்ೞ
ி೜ೞ ‫ି כ‬ி೜ೞ
stationary reference frame.
ܸ௤௦ ‫ כ‬ൌ ൅ ܴ௦ ݅௤௦ (14) ௏
ܶௗ௦ ൌ ೏ೞ ܶ௦ (19)
்ೞ ௏೏೎
‫כ‬ ‫כ‬
ܸௗ௦ and ܸ௤௦ are reference voltages for generation of space
௏೜ೞ
vector modulation. ܴ௦ is stator resistance, ܶ௦ is sampling time. ܶ௤௦ ൌ ܶ (20)
௏೏೎ ௦
This method provides constant switching frequency for
inverter. where, ܸௗ௖ is DC supply of inverter, ܶௗ௦ and ܶ௤௦ are d axis and
q axis components of imaginary switching time respectively.
C. Disadvantages of direct torque control with predictive Therefore, imaginary time vector is given by,
controllers
௏ೞ
1. DTC SVM with predictive controller is complicated ܶ௦௦ ൌ ܶௗ௦ ൅ ݆ܶ௤௦ ൌ ܶ (21)
௏೏೎ ௦
because of space vector modulation which involves
where, ܶ௦௦ is imaginary switching time vector, ܸ௦ is stator
sector determination and angle calculation.
voltage vector.
2. It requires more memory and computational time.
Using two phase to three phase transformation, this switching
time can be converted to three phase switching times. The
IV. SIMPLIFIED DIRECT TORQUE CONTROL SPACE VECTOR procedure for determining actual switching times from
MODULATION WITH IMAGINARY SWITCHING TIME CONCEPT imaginary switching time is given in [7]. In each sampling
interval, switching instants of inverter are calculated according
Due to use of predictive controller and hysteresis to this time. The process of determining pulse width modulation
comparators, we lose the simplicity of DTC in methods (PWM) pulses from imaginary switching times is described in
explained in section III and IV. There is another simplified [7]. In each interval torque and flux are corrected.
method proposed in [7]. Sector determination and angle

978-1-5090-6106-8/17/$31.00 ©2017 IEEE 833

Authorized licensed use limited to: NATIONAL INSTITUTE OF TECHNOLOGY CALICUT. Downloaded on February 05,2024 at 10:12:45 UTC from IEEE Xplore. Restrictions apply.
 2017 International Conference on Intelligent Computing,Instrumentation and Control Technologies (ICICICT)

V. SIMULATION RESULTS
Using MATLAB/SIMULINK software, simulations have been
carried out for three approaches of DTC technique. Fig.4-6 show
results of simulations of these techniques. Induction motor
parameters used for the study are given in Tables 2 and 3.
Parameters used for conventional direct torque control with
hysteresis controllers are shown in Table 2. The parameters used
for other two techniques are shown in Table 3. Fig.4 shows
torque command tracking of conventional DTC. X-axis
represents time as independent variable and Y-axis shows
torque(N-m) response. It is clearly observed that there are
ripples in the torque response. Fig.5 shows torque response of
DTC SVM using predictive controllers. It can be seen that
ripples are reduced considerably. Fig.6 shows result of DTC
SVM with imaginary switching time concept. It produces same
response as that of DTC SVM with predictive controllers but it Fig. 6 DTC SVM using imaginary switching time concept
has very low complexity.
Table 4 shows comparison of results of three control techniques.
Table 2 One can select control approach for drive depending upon
required performance.
ܴ௦ =1.77 ܴ௥ =1.34 ‫ܮ‬௦ =0.01393 ‫ܮ‬௥ =0.01212
‫ܮ‬௠ =0.369 Poles = 4 ‫=ܬ‬0.025 Flux=0.9 Wb
Table 3
VI. CONCLUSION
ܴ௦ =0.029 ܴ௥ =0.022 ‫ܮ‬௦ =0.00059 ‫ܮ‬௥ =0.00059 In this paper, the analysis and simulation results of three control
‫ܮ‬௠ =0.217 Poles = 4 ‫=ܬ‬63.87 Flux=0.9 Wb approaches are presented. First is conventional DTC, which has
advantage of simplicity and robustness but there are torque
ripples and high switching losses. Second approach is DTC
SVM using predictive controllers which reduces torque ripples
and switching losses but it increases the complexity of DTC
technique. Third approach is DTC SVM using the concept of
imaginary switching time, which has all the advantages of
second approach and it also retains the simplicity of
conventional DTC. Its results are same as DTC SVM with
predictive controllers but it is simpler and requires less
computational time and memory.

Table 4 COMPARISON BETWEEN THREE CONTROL

APPROACHES
Fig. 4 Basic DTC
DTC SVM DTC SVM
Control Conventional with with imaginary
approach DTC predictive switching time
controllers concept

Torque
High Low Low
ripples
Switching
Variable Fixed Fixed
frequency
Switching
High Low Low
losses
Complexity Low High Low
Fig. 5 DTC SVM using predictive controllers

978-1-5090-6106-8/17/$31.00 ©2017 IEEE 834

Authorized licensed use limited to: NATIONAL INSTITUTE OF TECHNOLOGY CALICUT. Downloaded on February 05,2024 at 10:12:45 UTC from IEEE Xplore. Restrictions apply.
 2017 International Conference on Intelligent Computing,Instrumentation and Control Technologies (ICICICT)

REFERENCES
[1] Takahashi and T. Noguchi, "A New Quick-Response and High-Efficiency
Control Strategy of an Induction Motor," in IEEE Transactions on
Industry Applications, vol. IA-22, no. 5, pp. 820-827, Sept. 1986.
[2] M. Depenbrock, "Direct self-control (DSC) of inverter-fed induction
machine," in IEEE Transactions on Power Electronics, vol. 3, no. 4, pp.
420-429, Oct 1988.
[3] C. Lascu, I. Boldea and F. Blaabjerg, "A modified direct torque control
(DTC) for induction motor sensorless drive," Conference Record of 1998
IEEE Industry Applications Conference. Thirty-Third IAS Annual
Meeting (Cat. No.98CH36242), St. Louis, MO, USA, pp. 415-422 vol.1.
1998.
[4] T. G. Habetler, F. Profumo, M. Pastorelli and L. M. Tolbert, "Direct
torque control of induction machines using space vector modulation,"
in IEEE Transactions on Industry Applications, vol. 28, no. 5, pp. 1045-
1053, Sep/Oct 1992.
[5] C. Mohamed, A. Golea and M. T. Benchouia, "Implementation of a
predictive DTC-SVM of an induction motor," 4th International
Conference on Electrical Engineering (ICEE), Boumerdes, pp. 1-4, 2015.
[6] A. Ouarda and F. Ben Salem, "Induction machine DTC-SVM: A
comparison between two approaches," 10th International Multi-
Conferences on Systems, Signals & Devices 2013 (SSD13), Hammamet,
pp. 1-7, 2013.
[7] Kumar, B. G. Fernandes and K. Chatterjee, "Simplified SVPWM-DTC of
3-phase induction motor using the concept of imaginary switching times,"
30th Annual Conference of IEEE Industrial Electronics Society, 2004.
IECON, pp. 341-346 Vol. 1, 2004.
[8] B. Lang, M. Miao, W. Liu and G. Luo, "Simulation and experiment study
of space vector pulse width modulation," 2009 9th International
Conference on Electronic Measurement & Instruments, Beijing, pp. 1-
408-1-412, 2009.
[9] B. Ozpineci and L. M. Tolbert, "Simulink implementation of induction
machine model - a modular approach," Electric Machines and Drives
Conference, 2003. IEMDC'03. IEEE International, pp. 728-734 vol.2,
2003.
[10] Dae-Woong Chung, Joohn-Sheok Kim and Seung-Ki Sul, "Unified
voltage modulation technique for real-time three-phase power
conversion," in IEEE Transactions on Industry Applications, vol. 34, no.
2, pp. 374-380, Mar/Apr 1998.
[11] Peter Vas, “Sensor less vector and direct torque control”, Oxford
University Press, London, 1998.

978-1-5090-6106-8/17/$31.00 ©2017 IEEE 835

Authorized licensed use limited to: NATIONAL INSTITUTE OF TECHNOLOGY CALICUT. Downloaded on February 05,2024 at 10:12:45 UTC from IEEE Xplore. Restrictions apply.