Improvise 3-Level DTC of Induction Machine using Constant Switching Frequency Method by Utilizing Multiband Carrier

International Journal of Power Electronics and Drive System (IJPEDS)
Vol. 7, No. 3, September 2016, pp. 638~647
ISSN: 2088-8694, DOI: 10.11591/ijpeds.v7i3.10995  638
Journal homepage: http://iaesjournal.com/online/index.php/IJPEDS
Improvise 3-Level DTC of Induction Machine using Constant
Switching Frequency Method by Utilizing Multiband Carrier
Sundram a/l Ramahlingam*
, Auzani Bin Jidin*
, Tole Sutikno**, L. Logan Raj*
*Department of Power Electronics and Drives, Universiti Teknikal Malaysia Melaka (UTeM), Malaysia
**Department of Electrical Engineering, Universitas Ahmad Dahlan (UAD), Indonesia
Article Info ABSTRACT
Article history:
Received Nov 12, 2015
Revised Mar 6, 2016
Accepted Apr 7, 2016
This paper presents the advantage of using optimal PI parameter tuning
strategy of constant switching method in the three phase Direct torque
control (DTC) scheme. The DTC system is known to offer fast decoupled
control of torque and flux via a simple control structure. Nevertheless, DTC
system has two major drawbacks, which are the variable inverter switching
frequency and high torque output ripple. The major factorthat contributes to
these problems the usage of hysteresis based comparators to control the
output torque. The implementation of PI based constant switching method in
DTC able to solve these problems while retaining the simple control structure
of conventional DTC. The combination usage of 3-level CHMI in this system
can further minimize the output torque ripple by providing greater number of
vectors. This paper presents detail explanation and calculationof optimal PI
parameter tuning strategyconsecutively to enhance the performance of 3-
level DTC system. In order to validate the feasibility, the proposed method
compared with convention DTC system via simulation and experiment
results.
Keyword:
3-level voltage vector
3-phase induction machine
CHMI
Constant switching
DTC
Copyright © 2016 Institute of Advanced Engineering and Science.
All rights reserved.
Corresponding Author:
R. Sundram,
Department of Electrical Engineering,
Universiti Teknikal Malaysia Melaka (UTeM),
76100 Durian Tunggal, Melaka, Malaysia.
Email: sundram02@student.utem.edu.my
1. INTRODUCTION
There are two most common ac drives control schemes that are being widely researched. One of it is
Field Oriented Control (FOC) which was proposed by F. Blaschke. Second scheme is Direct Torque Control
(DTC) which was proposed by I. Takahashi and T. Noguchi [1, 2]. There are two major drawbacks of FOC
compared to the DTC which are torque is controlled indirectly and requirement of the pulse encoder [3].
FOC method controls the torque indirectly because its control priority is flux vector. FOC needs the pulse
encoder in order to obtain the speed and position of the rotor. This makes the DTC system as an alternative
and gained the attention of many researchers lately due to its simple structure by elimination of pulse encoder
and simple algorithm with lesser dependency on motor parameters (only requires value of stator resistance Rs
and phase current) [1,4]. Over the past years, the utilization of multilevel inverter topology in the DTC
system has gained popularity for the medium and high voltage applications. Typically, there are three types
of multilevel inverters topologies. Those types are the Neutral Point Clamped (NPC), Flying Capacitor (FC)
and Cascaded H-bridge Multilevel Inverters (CHMI) [6]. The benefit of using multilevel inverter is its
availability of greater number of voltage vectors that will contribute in selecting suitable and optimal voltage
vector to control flux and torque by reducing the slope magnitude of the torque and flux. Moreover, this
features also contributes towards optimum switching for high-efficiency by reducing the switching frequency
and also improve the output voltage quality (by reducing the rate of change of phase voltage, dV/dt) [6].

 ISSN: 2088-8694
IJPEDS Vol. 7, No. 3, September 2016 : 638 – 647
639
In conventional DTC, the variable switching frequency is inevitable due to the nonlinear effects of
applied voltage vectors on torque and flux variations, in restricting the variations within hysteresis bands. As
the stator flux space vector forms a circular locus, one of the two suitable voltage vectors that has the most
tangential to the stator flux gives higher rate of torque change compare to less tangential voltage vector[3]. It
can be shown that the application of this vector is dominant (i.e. longer time application) when the flux
moves closer to the sector border in the stator flux plane, on the other hands, the two suitable vectors become
less tangential and switch more often when the flux vector travels around the middle of sectors. So the
torque switching (or flux switching) is more frequent in the border (or middle) of the sector compare to the
middle (or border) of the sector. These phenomena majorly contribute to the variable switching frequency in
DTC. The implementation of constant switching method can overcome the irregular switching problems in
DTC [3, 4, 5]. Previously the constant frequency torque controller (CFTC) method was proposed by [3] in
order to reduce the output torque ripple with a constant switching frequency. In this approach, two triangular
carrier waves were injected at the torque error node and then two comparators were used to generate torque
status. Although this method reduced the torque ripple but still has low frequency torque error oscillation.
This is due to the stator flux hysteresis based controller in which used to regulate flux around its reference
value. This error is less significant and negligible if the PI parameter is calculated correctly.
2. IMPLEMENTATION OF CONSTANT FREQUENCY TORQUE CONTROLLER (CFTC) IN 2-
LEVEL DTC SYSTEM
The constant frequency torque controller (CFTC) was implemented in 2-level DTC system in order
to reduce the output torque ripple while maintaining a constant switching frequency compared to the
conventional DTC method [3]. The switching frequency of conventional DTC system varies based on
operating condition of the system. This is due to the hysteresis based controller is used to control the flux and
torque of the conventional DTC system. The hysteresis based controller controls the magnitude of torque (or
flux) within the prefixed bandwidth (error tolerance). The rate of change of the torque (or flux) is directly
proportional to the frequency of torque (or flux) error signal. Higher rate of change of thetorque produce
higher frequency torque error status frequency and vice versa. It these phenomena occur when the flux locus
enters and leaves a sector of d-q plane flux locus. The conventional DTC system of 3-phase induction
machine has the total of 6 sectors with 30o
angle difference between each sector. The 3-phase conventional
DTC system has 6 active and 2 zero voltage vectors (as shown in Figure 1). For the decrement of the torque
status, a zero voltage vector will be selectedalways in which the rate of change of torque is constant
throughout every operating condition. But for the increment of the torque status, there will be an active
vector selected between two most tangential vectors in the direction of flux locus in which rotate in circular
motion. The selection of the suitable voltage vector effected by the flux error signal produce by the flux
hysteresis controller either to increase or decrease flux locus magnitude along the flux reference signal.
For example in sector 2,the voltage vector V3 and V2 will be chosen for the anti-clockwise stator
flux rotation in d-q plane. At the beginning of sector 2 for anti-clockwise rotation, vector V2 will be most
tangential to the flux locus compare to V3. The direction of flux locus vector is contributed by the direction
of applied voltage vector. The V2 (most tangential to flux locus) will be selected dominantly at the beginning
of sector 2 because its less contribute to the increment of the stator flux. In the other hand V2 will give
highest torque increment since it has highest tangential component. As a result, the rate of changes of torque
error status (frequency) will be higher at the beginning of sector. When the flux locus travel pass the middle
of sector, the both vector V2 and V3 have equal tangential component in where both vectors will be selected
more frequently to maintain the flux regulation. But the both vectors have lower tangential component
compare to the V2’s tangential component at the beginning of the sector. This makes the both vector will to
produce lower rate of change of torque. Consequently, the rate of changes of torque error status (frequency)
will be lesser after the middle of sector. This scenario is clearly describe by the Figure .2(a) in which it
shows the relationships between torque, hysteresis band, torque status error, andthe angle of stator flux locus
in d-q plane. For the anti-clockwise rotation, angle -30o
is the beginning of the sector 2 and 30o
is the middle.
Even the increment of the torque slope differs according to the angle of stator flux locus, but the decrement
of the torque slope is same (because only zero vectors are used). This alternation of torque (and flux)
switching frequency contributes to the alternation of inverter (IGBT) switching frequency.

IJPEDS ISSN: 2088-8694 
Improvise 3-Level DTC of Induction Machine using Constant Switching Frequency Method … (R. Sundram)
640
Figure 1. Basic sector definition and 2-level voltage vector of 3-phase DTC system
The theoretical behavior of DTC system if 3 phase induction machines can be described in terms of space
vectors by the equations that are written in stator stationary reference frame as below:
(1)
(2)
(3)
| || | (4)
∫( ) (5)
∫( ) (6)
where P is the number of pole pairs; Ls (stator inductance), Lr (rotor inductance), and Lm (mutual inductance)
are the inductances of the motor, and are the stator and rotor flux and δ is the angular difference
between stator flux linkage and is stator current space vector . As in equation 5 and 6, the stator flux vector is
written in d-q axis components. The electromagnetic torque given in equation 4 can be rewritten in d-q
coordinates shown in equation 7. This equation also shows the relationship of tangential component over the
torque output.
( ) (7)
Even though variation of switching status of torque and flux effect the inverter switching frequency,
thus the torque status (Tstat) frequency give a high impact compare to the flux. This is because averagely the
torque status switching is higher than compare to the flux status switching. As a simple solution, the constant
frequency torque controller (CFTC) can be used to reduce the output torque ripple while maintaining a
constant switching frequency compared to the conventional DTC method. This is done by injecting two high
frequency triangular carrier waves at the torque error node and then the resultant signal were passed through
two comparators in order to generate the torque error status signal (Tstat) equivalent to the conventional 2-
level DTC. The CFTC block (shown in Figure .4) was used to replace the torque hysteresis controller as
shown in Figure 3 (the area marked with red dotted line). The Tstat signal generated can be described by the
following equation 8.
Sector 3
Sector 2
Sector 1
Sector 4
Sector 5
Sector 6
V2V3
V2V3
d- axis
V2V3
V4
V5 V6
V1
q-axis

 ISSN: 2088-8694
641
{ (8)
(a) (b)
Figure 2. Illustration comparison of (a) basic DTC system and (b) the CFTC system in term of the torque,
controller, torque error status, and stator flux angle (degree).
Figure 3. Basic structure of DTC-hysteresis based motor drive system
The absolute slope value of Tpi signal must be lesser than the absolute slope value of the carrier in order to
regulate the torque output along the reference value. This is possible by the proportional gain of the PI
controller. To archive the above condition, the following equation must be fulfilled:
( )
(9)
and also
Stator
flux
angle
Torque
Carrier
Tpi
Torque
Torque error
status ( )
Stator
flux angle
Torque
error
status
( )

642
| |
(10)
The is positive slope equation and the is negative slope equation. The selection of optimal Ktp value
is based on the smallest value produced the equation (9) and (10). The supporting equations in order to solve
the above equation are given by the (11) until (15).
(11)
( ) (12)
( ) (13)
(14)
( ) (15)
The minimum sampling period can be obtain in the experimental setup by using DSPACE interface
card DS1104 is 20µs. The sampling time of the Matlab simulation is set to 20µs. The minimum sampling
required to make low resolution triangle wave carrier is 8. The 8 step per-cycle with 20µs sample will
produce carrier frequency of 6250Hz. The Table 1 shows the parameter of 3-phase induction machine used in
simulation and experiment.
Figure 4. The Constant Frequency Torque Controller (CFTC)
Table 1. Parameters of 3-Phase Induction Machine Parameters and Experimental Set-up Data
Induction motor parameters
Rated power, P 1.1 kW
Rated voltage, Vs 380 V
Rated current, is,rated 2.7 A
Rated speed, ωm 2800 rpm
Stator resistance, Rs 6.1 Ω
Rotor resistance, Rr 4.51 Ω
Stator self inductance, Ls 306.5mH
Rotor self inductance, Lr 306.5mH
Mutual inductance, Lm 291.9 mH
Combined inertia, J 0.0565 kg-m2
Combined viscous friction, B 0.0245 N.m.s
Number of pole pairs, P 1
Inverter’s Voltage Supply 240V

 ISSN: 2088-8694
643
(a) (b)
Figure 5. FFT result for output torque of (a) basic DTC system and (b) the CFTC system in which simulated
in same operating condition
Figure 2(a) and Figure 2(b) show the theoretical illustration comparison of the conventional DTC
system versus the CFTC implemented DTC system. From the comparison, the variation of the torque
frequency for the both cases can be observed clearly. For the conventional DTC system the torque status
switching frequency varies along the sector 2. However for the CFTC implemented DTC system, the torque
status switching frequency remain constant but only changes in the duty ratio corresponding to the torque
demand needed to regulate the torque around the reference. In order to verify the feasibility of the CFTC
method over the conventional method, the simulation testing was done using Matlab software with the given
parameter. The FFT analysis is done on the torque output to verify the frequency spectrum. The result
analysis is shown in Figure 5. The DTC have wide rage switching frequency involved. Whereas in CFTC
method, the entire frequency spectrum were eliminated except for the frequency of carrier wave in which
6250Hz.
3. PROPOSED OPTIMAL PI PARAMETER TUNING STRATEGY
For the proposed method, the CFTC controller was implemented in 3-level DTC system of 3-phase
induction machine. The principle of applying the CFTC controller in multilevel DTC is to increase efficiency
of the DTC system by selecting most optimal voltage vector at every operating condition. This is possible
because the 3-level inverter provide 18 active voltage vectors in which more than conventional system. As
result, the optimal voltage vector selection willcontribute to more the reduction of torque ripple and lower
harmonic. Figure . 6 shows the constant frequency torque controller (CFTC)which was modified for the
purpose to be implemented in conventional 3-level DTC system. The modified CFTC consists of six
triangular carrier wave, six hysteresis comparators and single proportional-integral (PI) controller. In theory,
the torque error status signal (Tstat) produced from the modifiedCFTC is similar to a 7-level hysteresis
controller, which can be in one of seven states; −3, -2, -1, 0, 1, 2 or 3. There is no modification of the original
(3-level) look-up Table is required. As a result, the simple control structure of hysteresis-based DTC can be
maintained in which the decouple control structure of flux and torque output. The submission of Tstat signal
generated by the comparators in Figure 6 can be described by the following equation:
{
(2)

644
Figure 6. The modified Constant Frequency Torque Controller (CFTC)
Table 2. Parameters of CFTC Controller for 3-Level DTC System
Induction motor parameters
A 382660
38004
0.0672
K 33257
(low speed) 0.0031
(low speed) 1181.5
(medium speed) 0.00052794
(medium speed) 202.0211
(high speed) 0.00039613
(high speed) 151.581
Figure 7. Simulation result using 6-level carrier wave with corresponding output signal. (a) reference torque
with estimated torque output, (b) 6-level carrier wave with the PI compensated torque error signal (Epi), (c) 7-
levelequivalent torque demand status (Tstat) generated by the modified CFTC block.
1
Torquestatus
2
CarrierwavewithEpi
-3
-2
-1
0
1
2
3
Torque(p.u)
0.7
(a)
(b)
(c)
Tstatus
Testimated
Trefference
C6
C5
C4
C3
C2
C1
C1

 ISSN: 2088-8694
645
Previously there was only one set of kp and ki is used for 2-level DTC system because there is only
single magnitude of voltage vector was used throughout the operation. Since the 3-level DTC system
provides 3-level magnitude of voltage vector, there should be 3 set ok kp and ki to be used in 3 different
operating condition in order to provide better torque regulation at every point of operating condition. For
every operation condition there will be 3 different optimal torque error statuses will be selected to reduce the
torque ripple. First, which is low-speed operation where torque 1-0 will be selected. Second, which is
Medium-speed operation where torque 2-1 will be selected. Last will be high-speed operation where torque
3-1 will be selected. The Table 2 shows the parameter of CFTC controller for 3-level DTC system using the
parameters and equation given previously.
4. SIMULATION AND EXPERIMENTAL RESULT
The simulation of proposed method had been performed by using Matlab/Simulink 2011. Figure .8
shows the simulation result of the proposed method. The simulation was done in using the 3-phase induction
machine model. As result of the simulation the CFTC controller was able establish in 3-level DTC system
with a proper regulation of torque along the reference value. The switching frequency of the torque error
status was constant (same frequency as the carrier) regardless the operation condition. The 3-level CFTC
controller able to generate optimal torque status signal according to the torque demands (high, medium, and
low speed).
A complete experimental set-up has been realized to verify the feasibility of the proposed method in
terms of hardware implementation. The experimental set-up consists of DSPACE 1104 controller, hardware
setup of 3-level CHMI and a 2 HP, two-pole induction motor. The Table 1 shows the induction motor
parameter that used for both simulation and experimental setup. The control algorithm is implemented on
DSPACE 1104 with sampling period of 20 µs. Figure 8 shows the experimental result of proposed method in
two different operating conditions which is the high speed and low speed.
(a) (b)
Figure 8. Experimental result of proposed method two different operating conditions which is (a) the high
speed and (b) low speed. (Te) is the estimated torque ripple waveform, (Vtri) is the triangle shape carrier
waveform, (Vcomp) is the compensated signal before feed in to the comparator, (Tstat) is the torque status
demand signal produce by PI controller.
The lower section waveform shows zoomed view of the red square from the upper waveform with
200 times zoom factor. The torque waveform (Te) regulate along the torque reference. It is noticeable that the
gradient for both increase and decrease torque not abrupteven at high and low speed operating condition. The
lower gradient of torque slopes is contributed by the 3-level CHMI topology by choosing optimal voltage
vector magnitude according to the torque demand status (as shown by waveform Tstat). For example, the
higher magnitude voltage vector is chosen for high speed operation (in Figure 8(b)) and vice versa for(in
Figure 8(a)). The compensated signal (Vcomp) is compare with triangular carrier (Vtri) to produce appropriate
torque demand signal (Tstat). The selection of 3-level voltage vector is proportional to themagnitude of
Tstat.The torque demand status (Tstat) in both results clearly shows the characteristic of constant in frequency
(on/off period) regardless to the magnitude of torque status. As conclusion, the experimental and simulation
results of proposed method able to optimized DTC performance by producing constant switching.

646
5. CONCLUSION
This paper has presented the implementation of constant switching method in conventional the 3-
level DTC system. This system has total number of 18 active and 2 zero voltage vectors. The proposed
scheme was simulated and compared with CFTC and conventional method using Matlab/Simulink. The
proposed method was also tested and verified its feasibility using a complete experimental set-up.
ACKNOWLEDGMENT
The authors would like to thank the Ministry of Education (MOE) and Universiti Teknikal Malaysia
Melaka (UTeM) for providing the research grant FRGS/2013/FKE/TK02/02/1/F00159 for this research.
REFERENCES
[1] Takahashi and T. Noguchi, "A New Quick-Response and High-Efficiency Control Strategy of an Induction Motor,"
Industry Applications, IEEE Transactions on, vol. IA-22, pp. 820-827, 1986.
[2] T. Noguchi, M. Yamamoto, S. Kondo, and I. Takahashi, "Enlarging switching frequency in direct torque-controlled
inverter by means of dithering," Industry Applications, IEEE Transactions on, vol. 35, pp. 1358-1366, 1999.
[3] N.R.N. Idris and A.H.M. Yatim, "Direct torque control of induction machines with constant switching frequency
and reduced torque ripple," Industrial Electronics, IEEE Transactions on, vol. 51, pp. 758-767, 2004.
[4] Jidin, N. Idris, N. Rumzi, M. Yatim, A. Halim, T. Sutikno, and M. Elbuluk, "Extending switching frequency for
torque ripple reduction utilizing a constant frequency torque controller in dtc of induction motors," Journal of
Power Electronics, vol. 11, pp. 148-155, 2011.
[5] M. Sanila, "Direct Torque Control of induction motor with constant switching frequency," in Power Electronics,
Drives and Energy Systems (PEDES), 2012 IEEE International Conference on, 2012, pp. 1-6.
[6] Z. Yu, J. Zhenhua, and Y. Xunwei, "Indirect field-oriented control of induction machines based on synergetic
control theory," in Power and Energy Society General Meeting-Conversion and Delivery of Electrical Energy in
the 21st Century, 2008 IEEE, 2008, pp. 1-7.
[7] Casadei, D., Serra, G., and Tani, A. Analytical investigation of torque and flux ripple in DTC schemes for induction
motors. 23rd International Conference on Industrial Electronics, Control and Instrumentation, 1997. IECON 97.
1997.
[8] Rodriguez, J.; Bernet, S.; Steimer, P.K.; Lizama, I.E., "A Survey on Neutral-Point-Clamped Inverters," Industrial
Electronics, IEEE Transactions on , vol.57, no.7, pp.2219,2230, July 2010
[9] Yen, G.W., et al. (2014). "Proposed Voltage Vector to Optimize Efficiency of Direct Torque Control."
International Journal of Power Electronics and Drive Systems 4(4): 578.
[10] Malinowski, M., et al., A Survey on Cascaded Multilevel Inverters. Industrial Electronics, IEEE Transactions on,
2010. 57(7): p. 2197-2206.
[11] Zuber, M.Z.R., et al. (2015). "Improved Torque Control Performance in Direct Torque Control using Optimal
Switching Vectors." International Journal of Power Electronics and Drive Systems 5(3): 441.
BIOGRAPHIES OF AUTHORS
R. Sundram was born in 1989 in Penang, Malaysia. He received the B.Eng. degree (Hons) in
Electrical Engineering from Universiti Teknikal Malaysia Melaka, Malaysia in 2012 and now he
is currently pursuing the M.Sc degree in Power Electronics and Drive. His areas of research
interest include Direct Torque Control of Multi Phase System and Power Electronics.
Auzani Jidin received the B.Eng. degrees, M.Eng.degrees and Ph.D. degree in power
electronics and drives from Universiti Teknologi Malaysia, Johor Bahru, Malaysia, in 2002,
2004 and 2011, respectively. He is currently a Lecturer with the Department of Power
Electronics and Drives, Faculty of Electrical Engineering, Universiti Teknikal Malaysia Melaka,
Durian Tunggal, Malaysia. His research interests include the field of power electronics, motor
drive systems, field-programmable gate array, and DSP applications.

 ISSN: 2088-8694
647
Tole Sutikno is an expert in the field of power electronics, Industrial Electronics & Informatics,
Embedded Systems and Electric Drives. Since 2001 he has been a lecturer in Electrical
Engineering Department, Universitas Ahmad Dahlan (UAD), Indonesia. He is an Associate
Professor at the above University since 2008. He received his B.Eng. and M.Eng. degree in
Electrical Engineering from Diponegoro University, Indonesia and Gadjah Mada University,
Indonesia, in 1999 and 2004, respectively. Currently, he is pursuing PhD degree at the Universiti
Teknologi Malaysia (UTM), Malaysia. He is an Editor-in-Chief of TELKOMNIKA Indonesian
Journal of Electrical Engineering, and also acts as an Editor in some international journals in
electrical and computer engineering, power electronics, motor drive systems and FPGA
applications areas.
Logan Raj Lourdes Victor Raj was born in 1988 in Selangor, Malaysia. He received the
B.Eng. degree (Hons) from the Universiti Teknikal Malaysia Melaka, Malaysia in 2012, where
he is currently pursuing the M.Sc degree. His areas of research interest include vector control of
ac machines and power electronics.

Improvise 3-Level DTC of Induction Machine using Constant Switching Frequency Method by Utilizing Multiband Carrier

More Related Content

What's hot (19)

Similar to Improvise 3-Level DTC of Induction Machine using Constant Switching Frequency Method by Utilizing Multiband Carrier (20)

More from IJPEDS-IAES (20)

Recently uploaded (20)

Improvise 3-Level DTC of Induction Machine using Constant Switching Frequency Method by Utilizing Multiband Carrier