Engineering, Technology & Applied Science Research Vol. 5, No.

1, 2015, 748-752 748

Study of Direct Torque Controlled 3-phase SCIM

with Two and Three-level Inverters using
ST-DTC and FR-DTC scheme
Pradeep Ranjan Tripathy Bibhu Prasad Panigrahi
Department of Electrical Engg, Department of Electrical Engg
KIIT Polytechnic, KIIT, 751024 Indira Gandhi Institute of Technology, Sarang. 759146
Bhubaneswar, India Dhenkanal, India
[email protected] [email protected]

Abstract— DTC control strategy helps to obtain fast and smooth switching table, based on the flux and torque errors and the
control in 3–phase induction motors. It does not depend upon the angular position of the stator linked flux in space.
machine parameters and transformation of reference frames.
Various control strategies on the Direct Torque Control (DTC) Implementation of this method is normally performed using
scheme have been carried out in the present study. The suggested high speed processors, [4-6, 8-10] which is a costly affair. Low
scheme is meant for designing the low cost control circuit for cost analog circuits can also be implemented for the DTC
DTC implementation on 3-phase SCIM. Three schemes with two scheme, being economical for small industrial sectors [11, 12].
types of inverters have been simulated in MATLAB (Simulink) Torque control is possible in two quadrant operation using low
environment and the THD ratio in each case has been studied cost analog circuits [13-15]. The common drawback of this
along with Fast Fourier Transform (FFT) tools to compare the scheme is the ripple in current and torque. In present context, a
advantages. descriptive study has been carried out for the comparative
study of current, flux and torque response referred to a varying
Keywords-FOC; ST-DTC; FR-DTC; 2LI; 3LI; six sectors; stepped torque command. The study is elaborated with the use
twelve sectors. of two level VSI and diode clamped three level inverter
following a Switching Table based DTC (ST-DTC) scheme
I. INTRODUCTION and Fuzzy Rule based DTC (FR-DTC) scheme. All simulations
Induction motors with conventional control schemes used have been carried out in MATLAB Simulink environment. The
in common industrial drives are not capable of providing fast switching look-up table and sector determination table have
and smooth torque variation, due to the inherent characteristics been realised using the state flow chart of the ST-DTC scheme
of the induction motor along with the complex, non linear and used for two and three level VSI. The Fuzzy logic based look-
time varying dynamics. The inaccessibility of some states and up table uses twelve sectors to obtain better performance in
outputs makes the input voltage prediction more difficult. torque and current. Overall, the present study emphasizes on
Though more effective control strategies in DC motors have the use of the control scheme that can be implemented
been established, safety concerns related to flammability render practically in various environments. An improvement of
these impracticable in many areas. The use of induction motor response is noticed through high level inverter use.
has thus become a popular practice, only after the invention of
Field Oriented Control (FOC) in the 1970s. The FOC scheme II. PRINCIPLE OF DTC
decouples the stator current into torque and flux producing
components so as to control the induction motor like a DC A. Using two level inverter
motor. FOC suffers from limitations arising from The stator terminal voltage and line currents are sensed and
computationally intensive design procedures. The dependence transformed to d-q stationary frame [14, 15]. Two axis stator
of this scheme on most machine parameters and reference linked flux ψds and ψqs are estimated by integrating the emfs,
frame transformations contribute to these difficulties [1, 7]. The which are obtained after compensating stator resistance voltage
Direct Torque Control (DTC) scheme postulated in 1980’s [2, drop from two axes stator voltages. Torque is calculated using
3] simplified the dynamic and steady state torque control. d-q axis flux and current components. The instantaneous
Transformation of the reference frame is not required in this angular position of stator linked flux is determined digitally
scheme. The stator resistance is enough to calculate the stator within six sectors in space [12, 14]. The absolute value of flux
flux and torque. The torque control is performed keeping stator and the estimated torque are compared with their respective
flux intact and changing the torque angle by acceleration or command values to obtain error signals. Both error signals are
deceleration of the stator linked flux, by selecting the optimum inspected by two prefixed band limits to yield control signals.
voltage vector. The optimum voltage vector is derived from a The torque control signal is of three levels (1, 0,-1) and the flux
Authors are thankful to ‘AICTE, New Delhi’, for support of research work.
www.etasr.com Tripathy and Panigrahi: Study of Direct Torque Controlled 3-phase SCIM with Two and Three-level …
 Engineering, Technology & Applied Science Research Vol. 5, No. 1, 2015, 748-752 749

control signal is of two levels (1, 0). The look-up table 30o, the voltage phasor corresponding to sector VI of sextant
(switching table) decides the switching pattern for 2-level provides the advancement of the flux vector. Hence, 180 rules
VSI[14, 15]. It uses three inputs, torque and flux control signal are defined to decide the optimum switching vector. Mamdani
along with the dwelling position of stator flux in sextant. method is used for defuzzification. The fuzzy controller gives
three crisp outputs corresponding to the three legs of a 2-level
The switching table decides the optimum switching voltage VSI.
vector, based on acceleration or deceleration of torque and flux
corresponding to command value. In the FR-DTC scheme, a
fuzzy controller replaces the hysteresis bands used in ST-DTC B. Using Three Level Inverter
scheme [15]. The flux and torque errors obtained comparing The 3-level inverter needs a lower switching frequency,
the respective command values are used as input state variables compared to the 2-level inverter, to achieve the desired
for the fuzzy controller. The fuzzy controller uses the angular sinusoidal voltage output. Hence, current harmonic decreases
dwelling position of stator flux as the third input state variable and efficient utilisation of energy is obtained. In each phase
(Figure 1). there are two complementary switching pairs. While one of the
switch pair is turned on, the other complementary switch pair
remains turned off. In 3-level inverter a set of two switches
(a) turns on at any given time. In a 3-level inverter there are 3-evel
output voltage and 5-level output line voltage. The possible
switching state of any leg is S1-S2, S2-S1’ and S1’-S2’ providing
output voltages 2Vdc, Vdc, 0 respectively. Hence, the possible
switching states in three phase 3-level inverter are 33 = 27. Out
of these 27 states, 24 states are active and 3 states are zero state
vectors. The active vectors are of three types, such as 6 long
(b) vectors, 6 medium vectors and 12 short vectors. The long
vectors are (002), (020), (022), (200), (220) and (202) having
magnitude of √(Vq2+Vd2)=4Vdc/3 . The medium vectors (012),
(021), (120), (210), (102) and (201) have magnitude of 2Vdc/√3
and the rest 12 short vectors have magnitude of 2Vdc/3.
The air gap space is divided into twelve sectors (Figure 2),
according to adjacent medium and long vectors each spreads
(c) with 300. Instantaneous angular position of the stator flux is
determined digitally as in Table I. Signal A1, A2 and A3 denote
the angular position of stator flux. The signals are true if the
absolute value of the ratio ψqs/ ψds is greater than or equal to tan
(pi/2), tan(pi/4) and tan(5.pi/2) respectively. Positive value of
ψqs and ψds denote 1, otherwise 0. Thus the dwelling position of
stator linked flux is identified by the generating signal Sθ. The
absolute value of flux and the calculated torque are compared
Fig. 1. (a) Flux error membership function (b) Torque error membership with their respective commands and generate the errors, which
function (c) Twelve angular dwelling position of stator flux are further inspected by prefixed hysteresis bands to yield the
flux control signal (Sψ) and the torque control signal (ST).
The stator flux error variable is represented by three
membership functions N, Z and P. The three membership
functions denote a decrease of the stator flux, no change of flux
and an increase of the flux. The torque error is calculated by
comparing the calculated electromagnetic torque with the
desired command torque. The torque error variable is denoted
by five membership functions: NL, NS, Z, PS and PL [15].
Corresponding functions represent a large or small decrease in
torque, no change in torque and a small or large increase in
torque. The angular position of the stator flux is defined in
twelve sectors in the sextant. Each sector of the ST-DTC
scheme is divided into two parts to facilitate the accurate
judgement of call for voltage vector. Considering the stator flux
to dwell in the first 60o in the air gap space and assuming the
torque and flux errors to be positive, the demanding voltage
phasor will tend to increase the torque. The voltage demand
can be satisfied with the voltage phasor in sector I or VI. If the
flux is in the first 30o of sector I, then any one of the above
voltage vectors can meet the purpose. When it is in the second
Fig. 2. Twelve sectors with switching vectors

www.etasr.com Tripathy and Panigrahi: Study of Direct Torque Controlled 3-phase SCIM with Two and Three-level …
 Engineering, Technology & Applied Science Research Vol. 5, No. 1, 2015, 748-752 750
TABLE I. DTC SWITCHING TABLE FOR 3-LEVEL INVERTER

Sψ ST Sθ
<1> <2> <3> <4> <5> <6> <7> <8> <9> <10> <11> <12>
V2(220) V9(120) V3(020) V10(021) V4(022) V11(012) V5(002) V12(102) V6(202) V7(201) V1(200) V8(210)
1 1
V8(210) V2(220) V9(120) V3(020) V10(021) V4(022) V11(012) V5(002) V12(102) V6(202) V7(201) V1(200)
1 0 V26(111) V27(222) V25(000) V26(111) V27(222) V25(000) V26(111) V27(222) V25(000) V26(111) V27(222) V25(000)
V6(202) V7(201) V1(200) V8(210) V2(220) V9(120) V3(020) V10(021) V4(022) V11(012) V5(002) V12(102)
1 -1
V7(201) V1(200) V8(210) V2(220) V9(120) V3(020) V10(021) V4(022) V11(012) V5(002) V12(102) V6(202)
V3(020) V10(021) V4(022) V11(012) V5(002) V12(102) V6(202) V7(201) V1(200) V8(210) V2(220) V9(120)
0 1
V10(021) V4(022) V11(012) V5(002) V12(102) V6(202) V7(201) V1(200) V8(210) V2(220) V9(120) V3(020)
0 0 V25(000) V26(111) V27(222) V25(000) V26(111) V27(222) V25(000) V26(111) V27(222) V25(000) V26(111) V27(222)
V5(002) V12(102) V6(202) V7(201) V1(200) V8(210) V2(220) V9(120) V3(020) V10(021) V4(022) V11(012)
0 -1
V11(012) V5(002) V12(102) V6(202) V7(201) V1(200) V8(210) V2(220) V9(120) V3(020) V10(021) V4(022)

The switching table (Table I) is formulated based on the V, Ra=16.575 Ω, Rr=5.2183 Ω, Lls=Llr=0.0404558 H,
angular position of the stator linked flux in space and the Lm=0.5393595 H, P=4, J=0.0048742 kg-m2, B=0.000001Nm-s.
requirement of acceleration or deceleration of flux and torque, The line value of stator voltage and current are sensed to
maintaining a constant magnitude of stator flux. Hence, the transform into d-q component of voltage and current. Using
lookup table decides the optimum switching vector for the flux and torque estimator, flux and torque are calculated after
three-level VSI with the help of the three inputs Sψ, ST, and Sθ. compensating the stator resistance drop from the d-q stator
Based on the dwelling position of the stator flux in a particular voltage [13]. The absolute values of flux and torque are
sector the digital control signals of flux and torque indicate the compared with their respective command values to obtain the
required voltage state for switching. For example, while the corresponding errors. The flux command has been considered
stator flux dwells in sector <1> and both the flux and torque as the rated flux (1.00184 Wb) and the rated torque as 5.0618
signal are 1, then four possible of switching states may be N-m. The simulation has been carried out with a variable two
selected. Those are long vector V2 (220), medium vector V8 quadrant torque command with an equal time interval of 0.06 s.
(210), the short vector V15 (110) and its redundancies for The initial step command is 30% for the first interval, which
increasing in both torque and flux. The switching vectors lie in changes to 60% and then to 30% of rated torque in succeeding
sector <2> and sector <3>. Similarly if flux and torque control time intervals, and finally falls to zero. The negative torque
signals are 0 and 1, voltage vectors V3 (020), V10 (021), V16 command is set to 0.24 second.
(010) and its redundancy may be selected. The redundant
vectors are the switching states from V13 to V24 respectively. The errors obtained are compared in their respective
When the torque control signal is zero, no change in torque is hysteresis bands and control signals Sψ and ST are generated.
needed and a zero vector, V25 (000), V26 (111) or V27 (222), is Both of these control signals along with the position of the
selected. The possible look up table is shown in Table I. Logic stator linked flux Sθ are judged by the distinguished lookup
2 stands for on condition of both upper arm switch, logic 1 for table meant for two-level and three-level inverters to obtain the
on condition of middle two switches and logic 0 for on optimum switching vectors. The flux and torque errors are
condition of lower switches. The optimum switching vector for directly used as the state variable inputs in the fuzzy controller
the corresponding inverter is decided by the switching table to of the FR-DTC scheme. The angular position of the stator
achieve the required torque control. The switching states are linked flux is used as the third state variable. The scheme has
depicted in Figure 3. been simulated for two and three level inverters. All the three
state variables are processed by the fuzzy controller, using 180
fuzzy rules. The crisp outputs of the fuzzy controller are
utilised for switching the inverters after generating its
complement for lower switch of any inverter leg. The
simulation has been performed for all three cases i.e. ST-DTC
for two and 3-level inverters and FR-DTC for 2-level inverter.
The x-y plotting of the d-q flux (Figure 4) is of circular
shape for all schemes proving the spatial relation, and the
magnitude versus time shows a constant magnitude of
1.00184629 Wb, (Figure 5a) at variable torque command in
two quadrants. It is found that the flux ripple reduces with the
Fig. 3. Switching Logic States in 3-Level Inverter use of high level inverter. The plot in Figure 5c shows the
dwelling position of the stator flux. It varies in the sextant
III. SIMULATION STUDY within –π/2 to 3π/2 radians and its direction changes with a
change in torque command. It is shown that the magnitude of
Both ST-DTC and FR-DTC schemes using two and three stator flux remains intact with a variation of torque command
level inverters have been simulated in the MATLAB Simulink either in forward or reverse direction. The two axis flux ψqs and
environment. The machine parameters of the 3-phase, 0.75 kW, ψds maintain π/2 radian phase difference as shown in Figure 5.
50 Hz, star connected induction motor are as follows: VLL=415

www.etasr.com Tripathy and Panigrahi: Study of Direct Torque Controlled 3-phase SCIM with Two and Three-level …
 Engineering, Technology & Applied Science Research Vol. 5, No. 1, 2015, 748-752 751

The variation of current in each scheme has been depicted the Fuzzy Rule based DTC scheme provides faster response but
in Figures 7 and 8. The current settles to steady state ripples in torque increase. Total Harmonic Distortion (THD) is
comparatively quickly in the 3 level inverter based STDTC measured using the FFT tool. It is observed that THD in the
(3LISTDTC) scheme with least dynamic amplitude. The two STDTC scheme, applied with 3 level inverter, is lowest among
axes current resemble the circular shape in x-y plotting and the the three schemes studied but torque response to command is
frequency and magnitude changes along with the change of quite slow. The air gap has been divided into six and twelve
torque command. When motor changes its motion from sectors for different approaches. Figure 11 shows both
forward to reverse the π/2 lagging property of the two axes postulations. The dwelling position of air gap flux in different
current reverses. sectors has been traced out in forward and reverse mode
operation. Accordingly the development of sectors has also
been modified corresponding to the mode of motor operation.

(a) (b)
(a)

(b)

(c)

(c)

Fig. 6. (a) dq flux vs time (2LI STDTC) (b) dq flux vs time (2LI
FRDTC) (c) dq flux vs time (3LI STDTC)

Fig. 4. (a) xy plot of d-q flux (2LI STDTC) (b) xy plot of d-q flux (2LI (a) (b)
FRDTC) (c) xy plot of d-q flux in (3LI STDTC)

(a)

(c)
(b)

Fig. 7. (a) xy plot of dq current in 2LI STDTC (b) xy plot of dq current in

2LI FRDTC (c). xy plot of dq current in 3LI STDTC

(c)

(a)

Fig. 5. (a) abs flux and angle vs time (2LI STDTC) (b) abs flux and angle
vs time (2LI FRDTC) (c) abs flux and angle vs time (3LI STDTC) (b)

Figure 9 shows the superimposed torque with its respective

command value. The fast response of torque to the command
value is noted in all three cases. Rising torque in forward and (c)
reverse motoring during 30% to 60% rise mode has been
considered as instance in Figures 9 and 10. As seen in Table II,
a faster response of change in torque has been observed in
FRDTC based 2-level inverter (0.12 milli-secs). In comparison, Fig. 8. (a) dq current vs time in 2LI STDTC (b) dq current vs time in 2LI
FRDTC (c) dq current vs time in 3LI STDTC

www.etasr.com Tripathy and Panigrahi: Study of Direct Torque Controlled 3-phase SCIM with Two and Three-level …
 Engineering, Technology & Applied Science Research Vol. 5, No. 1, 2015, 748-752 752

comparison to the ST-DTC scheme) but the THD values are

(a) (b) also larger. In conclusion, to implement a low cost DTC
scheme with the least harmonic distortion, the implementation
with a diode clamped 3-level inverter seems preferable.

TABLE II. TORQUE PERFORMANCE & SWITCHING STRATEGIES

2 LI ST- 2LI FR- 3LI ST-

DTC DTC DTC
(c) 0.3T to 0.6T rise in Forward 0.17 0.16 0.22
motoring (m-s)
0.3T to 0.6T rise in Reverse 0.15 0.12 0.15
motoring (m-s)
% THD in Torque 106.52 117.61 46.13

REFERENCES
Fig. 9. (a) rising torque vs time at forward motoring (2LISTDTC) (b) [1] S F. Blaschke, “The principle of field-orientation as applied to the trans
rising torque vs time at forward motoring (2LIFRDTC) (c) rising torque vs vector closed-loop control system for rotating-field machines”, Siemens
time at forward motoring (3LISTDTC) Review, Vol. 34, pp. 217-220, 1972
[2] M. Depenbrock, “Direct self-control (DSC) of inverter–fed induction
Machine”, IEEE Transactions on Power Electronics, Vol. 3, No. 4, pp.
420-429, 1988
[3] I. Takahashi, Y. Ohmori, “High performance direct torque control of an
induction motor”, IEEE Transactions on Industry Applications, Vol. 25,
No. 2, pp. 257-264, 1989
[4] T. G. Habetler, F. Profumo, M. Pastorelli, L. M. Tolbert, “Direct torque
(a) (b)
control of induction machines using space vector modulation”, IEEE
Transactions on Industry Applications, Vol. 28, No. 5, pp. 1045–1053,
1992
[5] U. Baader, M. Depenbrock, G. Gierse, “Direct self control (DSC) of
inverter fed induction machine: a basis for speed control without speed
measurement”, IEEE Transactions on Industry Applications, Vol. 28,
No. 3, pp. 581-588, 1992
[6] D. Casadei, G. Serra, A. Tani, “Implementation of a direct torque control
(c) algorithm for induction motors based on discrete space vector
modulation”, IEEE Transactions on Power Electronics, Vol. 15, No. 4,
pp. 769–777, 2000
Fig. 10. (a) rising torque vs time at reverse motoring (2LISTDTC ) (b) [7] D. Casadei, F. Profumo, G. Serra, A. Tani, “FOC and DTC: two viable
schemes for induction motors torque control”, IEEE Transactions on
rising torque vs time at reverse motoring (2LIFRDTC) (c) rising torque vs
Power Electronics, Vol.17, No. 5, pp. 779-787, 2002
time at reverse motoring (3LISTDTC)
[8] F. Zidani, R. Nait Said, “Direct torque control of induction motor with
fuzzy minimization torque ripple”, Journal of Electrical Engineering,
(a) Vol. 56, No. 7-8, pp. 183–188, 2005
[9] F. Zidani, D. Diallo, M. E. H. Benbouzid, R. Nait Said, “Direct torque
control of induction motor with fuzzy stator resistance adaptation”, IEEE
Transaction on Energy Conversion, Vol. 21, No. 2, pp. 619-621, 2006
[10] M. Vasudevan, R. Arumugam, “High performance adaptive intelligent
direct torque control schemes for induction motor drives”, KMITL
Science and Technology Journal, Vol. 5, No. 3, pp. 559-576, 2005
[11] B. P. Panigrahi, D. Prasad, S. SenGupta, “Digital simulation and PC
based implementation of switching table based direct torque control of
induction motor drive”, IIT Conference, Kharagpur, pp. 1–6, 2005
(a) [12] B. P. Panigrahi, D. Prasad, S. SenGupta, “A simple hardware realization
of switching table based direct torque control of induction motor”,
Fig. 11. (a) six sector in two quadrant operation (b) twelve sector in two Electric Power Systems Research, Vol. 77, No. 2, pp. 181–190, 2007
quadrant operation [13] D. Prasad, B. P. Panigrahi, S. SenGupta, “Digital simulation and
hardware implementation of a simple scheme for direct torque control of
induction motor”, Energy Conversion and Management, Vol. 49, No. 4,
IV. CONCLUSION pp. 687-697, 2008
The DTC scheme has been studied using three strategies in [14] P. R. Tripathy, B. P. Panigrahi, “Simulation study on DTC based
induction motor drives using MATLAB”, IEEE Proceedings of
this paper. The classical ST-DTC scheme has been INDICON, India, Vol. 1, pp. 65-68, 2008
implemented with the logical simulation of low cost analog [15] P. R. Tripathy, B. P. Panigrahi, “Simulation studies on switching table
circuits for 2-level and 3-level inverters. The same inverters based DTC and fuzzy rule based DTC for three-phase squirrel cage
have been simulated using fuzzy logic. The faster torque induction motor”, Engineering, Technology & Applied Science
response is documented in the FR-DTC schemes (in Research, Vol. 2, No. 1, pp. 162-166, 2012

www.etasr.com Tripathy and Panigrahi: Study of Direct Torque Controlled 3-phase SCIM with Two and Three-level …