J Electr Eng Technol.

2017; 12(3): 1177-1186 ISSN(Print) 1975-0102

http://doi.org/10.5370/JEET.2017.12.3.1177 ISSN(Online) 2093-7423

Torque Ripple Reduction in Direct Torque Control of Five-Phase

Induction Motor Using Fuzzy Controller with Optimized Voltage Vector
Selection Strategy

Hye Ung Shin*, Seong Yun Kang** and Kyo-Beum Lee

Abstract This paper presents a torque ripple reduction method of direct torque control (DTC) using
fuzzy controller with optimal selection strategy of voltage vectors in a five-phase induction motor. The
conventional DTC method has some drawbacks. First, switching frequency changes according to the
hysteresis bands and motors speed. Second, the torque ripple is rapidly increased in long control
period. In order to solve these problems, some/most papers have proposed torque ripple reduction
methods by using the optimal duty ratio of the non-zero voltage vector. However, these methods are
complicated in accordance with the parameter. If this drawback is eliminated, the torque ripple can be
reduced compared with conventional method. In addition, the DTC can be simply controlled without
the use of the parameter. Therefore, the proposed algorithm is changing the voltage vector insertion
time by using the designed fuzzy controller. Also, the optimized voltage vector selection method is
used in accordance with the torque error. Simulation and experimental results show effectiveness of
the proposed control algorithm.

Keywords: Torque ripple reduction, Five-phase induction motor drive, Direct torque control, Fuzzy
controller

1. Introduction ripple reduction method becomes essential to solve these

problems.
Multiphase motor control has more advantages than Recently, many papers are dealing with the torque
the three phase-motor drive system. If the number of phases ripple reduction methods [7-11]. A torque ripple reduction
increases, the motor drive system possesses multiple voltage method that aims to make the instantaneous torque equal
vectors. Also, multiphase motor helps reducing the rotor with the reference value at the end of the cycle is
harmonic currents, lowering the dc link current harmonics, researched. Another method is that the average torque
reducing the current per phase without increasing the equals to the reference value among the entire cycle. In
voltage per phase, and having many voltage vectors by addition, there is a method that aims to the torque ripple
increasing the number of phases [1]. Above all, by RMS value over one cycle to be minimal. In these methods,
increasing the number of voltage vectors, it is possible to insertion time of the voltage vector (Ts) is calculated for the
select an optimal voltage vector for the high performance torque ripple reduction. The equation for calculating Ts is
of the direct torque control (DTC) in five-phase induction required for a number of motor parameters. Therefore, the
motor [2, 3]. motor parameters are significantly affected by the torque
In the conventional DTC scheme, both the torque and ripple reduction. In order to solve this problem, [12]
the stator flux errors between the reference and the proposed that the formula of Ts is fixed as a constant value.
estimated value are directly compared, and the appropriate However, this method has some drawbacks. First, when Ts
voltage vector is produced by a switching table [4, 5]. By has a low value, the dynamic performance is decreased.
optimal selection of the inverter switch states in each Second, Ts is found by tuning. The fixed Ts value needs to
sampling period, the DTC achieves rapid and effective be changed by considering the torque error to solve this
control of the stator flux and torque [6]. However, the problem. The fuzzy controller is applied in the proposed
advantage of the rapid response includes the increased torque ripple reduction algorithm for changing Ts [13]. The
torque ripple. The torque ripple causes increased noise input of the fuzzy controller is the torque error, and its
and unstable operation of the motor. Therefore, the torque output is the Ts [14].
Corresponding Author: Dept. of Electrical and Electronic Engineering,
This paper presents DTC controlled technique for five-
Ajou University, Korea. ([email protected]) phase induction machine. Relation between the switching
* Dept. of Electrical and Electronic Engineering, Ajou University, states and voltage vectors in five-phase induction motor
Korea. ([email protected])
drive is analyzed, and this paper proposes an improved
** Dawonsys Co., Ltd., Ansan, Korea. ([email protected])
Received: August 17, 2016; Accepted: March 12, 2017 torque ripple reduction method using the fuzzy control

1177
Copyright The Korean Institute of Electrical Engineers
This is an Open-Access article distributed under the terms of the Creative Commons Attribution Non-Commercial License (http://creativecommons.org/
licenses/by-nc/3.0/) which permits unrestricted non-commercial use, distribution, and reproduction in any medium, provided the original work is properly cited.
 Torque Ripple Reduction in Direct Torque Control of Five-Phase Induction Motor Using Fuzzy Controller with Optimized Voltage Vector ~

Fig. 1. Circuit of a five-phase induction motor Fig. 2. Voltage vectors of each switching state in five-phase
inverter
which is applied to find the optimal Ts under change of the
motor speed and the torque error. In addition, this paper
suggests a method of considering torque error and motor
speed to select an optimal voltage vector. The proposed
method is verified by simulation and experimental results.

2. Voltage Vector in Five-phase Inverter

The schematic of star winding and VSI in five-phase

induction motor is shown in Fig. 1. Two-level inverter is
used in this paper to drive five-phase induction motor. A
five-phase induction motor uses a full pitch concentrated
winding, instead of the traditional sinusoidal winding
distribution. This winding structure is designed with a Fig. 3. Plane of voltage vectors for five-phase inverter
rectangular distribution to better accommodate the quasi-
rectangular input currents provided by the five-phase. The
five-phase inverter contains two switches in each leg. The function, ph represents the phase angle, a=e j2/5, and Vs
switches of each leg operate in a complementary manner. reflects the voltage magnitude. (3) used as a formula for
[Sa Sb Sc Sd Se] shows the five-phase inverter switching the voltage magnitude calculation determined by each
states. The components of each switching state have 0 voltage vector state [1]. The magnitude of voltage vectors
(off state of the upper switch) or 1 (on state of the upper of the five-phase inverter switching states using the
switch) [15]. Therefore, five-phase inverter has 32 formula (3) is shown in Fig. 2. The calculated voltage
switching states. magnitude has three classes of the voltage value (0.6472
The non-zero voltage vectors of thirty and zero-voltage VDC, 0.4 VDC, and 0.2472 VDC). (1) and (2) are used for the
vectors of [000000], [111111] are compounded in these phase calculation of voltage vector.
switching states. 32 voltage vectors are included in three Each calculated phase of the voltage vectors has one of
class vectors each, have different magnitudes of the voltage the values which are divided by intervals of 36. The five-
vectors. Additionally, 10 sectors with different voltage phase inverters voltage vectors plane is shown in Fig. 3.
vectors phases are components of those three set vectors. The voltage magnitude and the phase are represented by
The formula for the phase and the voltage magnitude the voltage vector plane [16]. The increased number of
values of the voltage vectors is determined by switching voltage vectors of five-phase inverter compared to its
states, which is expressed as: three-phase counterparts helps increasing the choices of the
optimum voltage vector selection [15]. As a result, this
2 helps to reduce the torque ripple and conduct high
f ds1q1s = ( f as + afbs + a 2 f cs + a 3 f ds + a 4 f es )
5 (1) performance drive in the DTC.
= f dss 1 + jf qss 1
f qss 1 3. Proposed Torque Ripple Reduction Method
ph = tan 1 ( ) (2)
f dss 1 in Five-phase Induction Motor
2
Vs = VDC ( S a + aSb + a 2 Sc + a 3 S d + a 4 Se ) (3) 3.1 Direct torque control
5

where, f sd1q1s shows the d-q axis stationary frame complex Without considering the complex field-orientation block
space vector, fabcdes shows the complex space vector of each and inner current regulation loop, the DTC provides a very
phase, Sabcde indicates the five-phase inverter switching quick and precise torque response. This allows the DTC to

1178 J Electr Eng Technol.2017; 12(3): 1177-1186

 Hye Ung Shin, Seong Yun Kang and Kyo-Beum Lee

be a simple, fast response, and effective control method. number of the section along with the torque and the flux
The voltage vector selection in switching table of the DTC comparators outputs make up the lookup table [18]. When
allows the torque and the flux to be directly controlled. OF value becomes 1, the flux increases. In contrast, when
Additionally, for the optimal selecting of the voltage vector, OF value is set to -1, it means that the flux decreases.
the rotor flux position is demanded. The voltage model for Otherwise, the torque increases followed by the change of
estimating the rotor flux position is used in this paper. (4) the OT (+3, +2, +1).
and (5) are equations of the voltage model for the rotor flux For example, when the value of the OT is +3, the torque
position: increases when selecting the large voltage vector. The
s
fastest response time and largest torque value are driven
d dqs
from the large voltage vector. The torque is increased by
s s s s s
vdqs = Rs idqs + dqs = (vdqs Rs idqs ) dt (4)
dt the medium voltage vector when the value of the OT is +2.
s The small voltage vector is selected to increase the torque
Lr s qr
s
dqr = s
(dqs Ls idqs ) e = tan 1 ( s ) (5) value when that of OT is +1. In addition, if the value of the
Lm dr
OT is 0, the torque becomes decreased by the zero voltage
vector.
where, vsdqs , Rs , isdqs , sdqs , sdqr , , Ls , L r , Lm , and e
indicate the d-q axis voltage of the synchronous reference
frame, stator resistance, stator d-q axis current, flux of 3.2 Proposed method
stator stationary reference frame, flux of rotor stationary
reference frame, leakage flux, stator inductance, rotor The conventional torque ripple reduction method in the
inductance, mutual inductance, and flux angle. Without DTC of the three-phase induction motor is shown in Fig.
requirement of the motor speed in the flux estimation, 4 [12]. By considering the real torque and the torque
the voltage model is a convenient flux estimator for reference, the optimal time is calculated at the end of the
sensorless induction motor drives. Moreover, a few motor current period by this method. The slopes of the rising
parameters are needed for the voltage model to estimate the torque and the falling torque are necessary to calculate the
rotor flux position. Therefore, the voltage model is often optimal time.
used in DTC [17]. The slopes of the rising torque and the falling torque are
The estimated rotor flux position is divided by ten determined by using (6) and (7) [19]:
sectors listed in the Table 1. Thus, the rotor flux position is
included in the sector number. Table 2 shows the voltage Te 3P Lm
S1 = + {Vds* qr + Vqs* dr
vector selection lookup table, with the torque comparator Rs Rr 1 4 Ls Lr
( + ) (6)
containing four-levels. The OT is the torque comparator Ls Lr
output and OF is the result of the flux comparator. The m (ds qs + dr qr )}

Table 1. Angle value about sector number Te 3 P Lm

S0 = m (ds qs + dr qr )
Rs Rr 1 4 Ls Lr
Number of sector Angle ( + )
1 0 < e 36 Ls Lr
2 36 < e 72 (7)
3 72 < e 108
4 108 < e 144 where S1, S0, vdqs, dqs, dqr, m, P, and Te are the slope of
5 144 < e 180 the torque by nonzero voltage vector, the slope of the
6 180 < e 216
7 216 < e 252
8 252 < e 288
9 288 < e 314
10 314 < e 360

Table 2. Voltage vector selection lookup table

Number of sector
OF OT
1 2 3 4 5 6 7 8 9 10
+3 VL4 VL5 VL6 VL7 VL8 VL9 VL10 VL1 VL2 VL3
+2 VM4 VM5 VM6 VM7 VM8 VM9 VM10 VM1 VM2 VM3
-1
+1 VS4 VS5 VS6 VS7 VS8 VS9 VS10 VS1 VS2 VS3
0 V31 V0 V31 V0 V31 V0 V31 V0 V31 V0
+3 VL3 VL4 VL5 VL6 VL7 VL8 VL9 VL10 VL1 VL2
+2 VM3 VM4 VM5 VM6 VM7 VM8 VM9 VM10 VM1 VM2
+1
+1 VS3 VS4 VS5 VS6 VS7 VS8 VS9 VS10 VS1 VS2 Fig. 4. Conventional torque ripple reduction method in
0 V0 V31 V0 V31 V0 V31 V0 V31 V0 V31 three-phase induction motor

http://www.jeet.or.kr 1179
 Torque Ripple Reduction in Direct Torque Control of Five-Phase Induction Motor Using Fuzzy Controller with Optimized Voltage Vector ~

torque by zero voltage vector, d-q axis stator voltage, d-q fuzzy controller.
axis stator flux, d-q axis rotor flux, motor angular velocity, (12) and (13) are inputs of the fuzzy controller:
number of poles, and the electromagnetic torque. The
effective vector insertion time can be expressed as: eTe (k ) = Tref (k ) Tcal (k ) (12)
ceTe (k ) = eTe (k ) eTe (k 1) (13)
Tref To S0Tsamp Tref To S0
Ts = = + (8)
( S1 S 0 )Tsamp ( S1 S0 )Tsamp S1 S 0 where eTe is the torque error and ceTe is the torque error
variation. The information of the torque error is included in
where Ts is the effective vector insertion time, Tref is the the input values that are used in the fuzzy controller.
torque reference, T0 is the initial torque, and Tsamp is the Therefore, Ts is optimal time for the torque ripple reduction.
control period. According to the speed these parameters are Fig. 5 presents the proposed fuzzy system block diagram.
changed. Therefore, those values are not determined. If Inputs of fuzzy controller are all normalized in the interval
(6) and (7) are used for calculating the effective time for of [-1, 1]. The maximum value is scaled to 1 by inserting
the vector injection, as (6) and (7) are composed of many the input to fuzzy controller. After that, by using the
motor parameters, the effective vector insertion time does membership function, the output in fuzzification area is
not have an optimal value. (8) can be redefined as follows calculated.
to solve this problem: The membership function of the fuzzy controller is
shown in Fig. 6. The torque error allows the membership
ET E Tref To ref 0 function to be controlled. Through the simulation and
Ts = + F = + (9) testing, the graphs of membership functions are refined.
KT KF KT KF
The linguistic labels used to describe the fuzzy sets were
ET = Tref To (10) Negative Big (NB), Negative Medium (NM), Negative
Small (NS), Zero (Z), Positive Small (PS), Positive
EF = ref 0 (11) Medium (PM), Positive Big (PB) [13]. It is possible to
assign the set of decision rules as shown in Table 3. The
where KT and KF are constant values. By changing the
fuzzy rules are extracted from the simulation and the
values as (9), (9) is independent of the motor parameters.
experiment results about the process. Each control input
However, this conventional method has two drawbacks.
has seven fuzzy sets so that there are at most 49 fuzzy rules
First, the characteristic of this conventional method is that
[20-23].
the torque response time and the ripple are determined by
KT and KF. When the KT and the KF have small values, Ts
is decreased. As a result, the flux and torque ripple are
reduced and the response speed decreases, however when
the KT and the KF have greater value, the injection time of
the non-zero vector increases. As a result, the response
time is improved. Fig. 5. Fuzzy system block diagram
Otherwise, the flux and the torque ripple are increased.
Second, the conventional method cannot use the effective
voltage vector during the whole control period. Because
the denominator of the conventional method equation has a
fixed value. For example, if the control condition needs
the high speed and torque, the speed and torque are not
controlled by the reference when using the conventional
method. Because the conventional method does not have
a high effective voltage vector injection time ratio in the
control period. Fig. 6. Membership function of proposed fuzzy system
In order to solve the drawbacks, this paper proposes a
method of changing the KT by using a fuzzy controller Table 3. Rule base of proposed fuzzy system
with considering the motor speed. KF is rated flux value. eTe
NB NM NS Z PS PM PB
Because, the effect of the KF is much smaller. In addition, ceTe
NB PB PB NM NM PS PS Z
voltage vector injection time formula of conventional
NM PB NM NM PS PS Z NS
method is (9) which simplifies (8). (9) is used in the NS NM NM PS PS Z NS NS
proposed method. However, the difference between the Z NM PS PS Z NS NS PM
conventional method and the proposed method is PS PS PS Z NS NS PM PM
complement the conventional drawback by changing the PM PS Z NS NS PM PM PS
value of KT without using a fixed constant by using the PB Z NS NS PM PM PS PS

1180 J Electr Eng Technol.2017; 12(3): 1177-1186

 Hye Ung Shin, Seong Yun Kang and Kyo-Beum Lee

Fig. 7. Proposed method for torque ripple reduction in five-

phase induction motor

Fig. 9. Normal DTC simulation results without torque

ripple reduction method

Fig. 8. Block diagram of proposed all system

The optimal value of the KT is decided by defuzzi-

ficating of the fuzzy controller. As a result, the proposed
method overcomes the weakness of the conventional
method by using the designed fuzzy system. The proposed
torque ripple reduction method in the five-phase induction Fig. 10. DTC with torque ripple reduction method
motor is shown in Fig. 7. The five-phase inverter can simulation results by using the fixed constant
choose among three voltage vectors of different sizes. value (KT)
The voltage vector section according to the position of
the torque and the speed of the motor helps to select the reference of rated flux is 0.58 Wb. The KF is set as 0.58.
voltage vectors. For instance, the torque is rapidly increased The results of normal DTC are shown in Fig. 9. The torque
when the large voltage vector is selected such as the first and flux comparators are used for the normal DTC. When
section of Fig. 7. On the other hand, the torque is smoothly the torque is produced above the torque reference, it
increased when the small voltage vector is selected such become reduced by the zero voltage vector. On the other
as the second section of Fig. 7. Therefore, by applying this hand, if the produced torque is below the torque reference,
method, the proposed method can reduce torque ripple. the torque is increased by non-zero voltage vector. The
Additionally, only zero voltage vector is used for the torque overall torque value is controlled above the reference of
ripple decrease. Thus the reduced slope of the torque torque by DTC, because if the torque is produced below
equals to three sections. The control block of the proposed the torque reference, the torque increases sharply by non-
method is shown in Fig. 9. The algorithm for the torque zero voltage vector. The output of the torque comparator
ripple reduction method is based on the control block is shown in Fig. 9(b). The torque comparators output is
diagram shown in Fig. 8. controlled by the torque error. The torque has 5.9 Nm value
by the large voltage vector. Also, the torque is increased
about 3.3 Nm, 1.6 Nm by the medium voltage vector and
4. Simulation and Experiment Results small voltage vector. As a result, the maximum torque is
9.3 Nm, the minimum output of the torque is 3.1 Nm and
4.1 Simulation results the torque average is 5.35 Nm. The flux which is controlled
by rated flux value (0.58 Wb) is shown in Fig. 9(c).
Simulations have been performed using the PLECS tool In addition, by using the fixed speed motor as load in
to verify the performance of the proposed method. The this simulation, the speed is fixed at 400 rpm as indicated

http://www.jeet.or.kr 1181
 Torque Ripple Reduction in Direct Torque Control of Five-Phase Induction Motor Using Fuzzy Controller with Optimized Voltage Vector ~

in Fig. 9(d). The current of phase-A is shown in Fig. 9(e). above that the proposed method has good performance in
The DTC with the torque ripple reduction method by using terms of torque control characteristic, particularly, in the
the fixed constant value (KT) is shown in Fig. 10. Fig. 10(a) low speed region. On the other hand, torque ripple
shows the output torque. The maximum output torque increases in the high speed region in Fig. 12 as a rising
value is 8.8 Nm, the minimum output of torque value is 3.0 slope of the torque is increased by motor speed.
Nm and the average torque value is 4.46 Nm. Fig. 13 shows the simulation results of the DTC in
The simulation result of Fig. 10(a) by using the fixed high torque and speed state. Fig. 13(a) is the result of the
constant value (KT) is reducing the torque ripple compare conventional torque ripple reduction method. The Torque is
to Fig. 9(a), and the flux is controlled by rated flux 0.58 not controlled by the reference torque. Average torque of
Wb in 400 rpm. The results of the proposed method are the Fig. 13(a) is 5.8 Nm. The ratio of the effective voltage
shown in Fig. 11. The fuzzy system input has the torque vector injection time is increased in high torque and speed
error. Thus, state. However, the conventional method cannot increase
The KT value is the optimal value in each control period. enough the effective voltage vector injection time ratio
Fig. 11(a) indicates the output torque of proposed torque as this method uses a fixed value for calculating the TS.
ripple reduction method. The maximum torque is 7.15 Nm, Thus, the conventional method always has a zero voltage
the minimum torque is 3.0 Nm. As a result, the average vector insertion time ratio. As a result, the torque is
torque becomes 3.97 Nm. controlled under the torque reference. On the other hand,
Table 4 is listed in the simulation results of the torque the torque is controlled by the reference torque using the
ripple analysis. The proposed torque ripple reduction me proposed torque ripple reduction method in Fig. 13(b).
thod uses fuzzy controller for calculating the voltage Fig. 14 shows torque of the DTC about 5 Nm by
vector injection time. It is possible to reduce the torque changing KF. It can be shown from the simulation results
ripple more than at least 30%, compared with the con- above that the torque average has the similar value by
ventional DTC algorithm. The proposed method possesses changing KF. As a result, KF is able to fix as 0.58 which is
lowest torque ripple in Table 4. the rated flux value.
Fig. 12 shows simulation results of the transient state by
using the proposed torque ripple reduction method. A step
torque command is applied to the five-phase induction
motor. Torque reference is changed from 2 Nm to 7 Nm
in 300 rpm and 600 rpm. The torque is controlled by the
reference torque acquired using the proposed method in
transient state. It can be seen from the simulation results

(a)

Fig. 11. DTC with torque ripple reduction method

simulation results by the proposed method

Table 4. Comparison of the simulation result

Avg.
Max. Min. Avg.
Torque
(Nm) (Nm) (Nm)
Reduction ratio (b)
(%)
Basic DTC 9.3 3.0 5.35 - Fig. 12. Simulation results of the DTC in transient state (2
The fixed KT 8.8 3.0 4.46 65.9 Nm7 Nm) by using the proposed method: (a)
The proposed method 7.15 3.0 3.97 97.7 300 rpm and (b) 600 rpm

1182 J Electr Eng Technol.2017; 12(3): 1177-1186

 Hye Ung Shin, Seong Yun Kang and Kyo-Beum Lee

Table 5 Specification of five-phase induction motor

Motor specification
Capacity 1.5 kW
Rated voltage 220 V
Rated frequency 60 Hz
Rated current 4.9 A
Rated speed 1684 rpm

(a)

(a) (b)
Fig. 15. Experimental set-up: (a) Five-phase induction
motor and load motor and (b) control board and
power board

(b)
Fig. 13. Simulation results of the DTC in high torque and
speed state: (a) conventional torque ripple
reduction method and (b) proposed torque ripple
reduction method

Fig. 16. Experiment result of the DTC with torque ripple

reduction method by using the fixed constant value
(KT)

Fig. 16 and 18 show the experiment results in

accordance with the simulations. The experiment result of
the torque ripple reduction method for the DTC by using
Fig. 14. Simulation results of the DTC (5 Nm) by changing the fixed constant value (KT) is shown in Fig. 16. This
KF figure shows the output torque, output flux, and output
current of phase-A. The reference output torque is 4 Nm
and flux is controlled by the rated flux of 0.58 Wb in this
5. Experiment Results experiment. This figure indicates the output torque, the
output flux, and the output current of phase-A. In
The motor specification is listed in Table 5. The experimental results, the maximum torque ripple is 3.7 Nm
experimental set consists of the load motor, the five-phase due to the maximum torque of 6.6 Nm and the minimum
induction motor, the control board, and the power board as torque of 2.9 Nm. This torque ripple reduction method
shown in Fig. 15. DSP controller TMS320F 28335 from reduces the torque ripple compared to the normal DTC.
Texas Instruments is used for the digital implementation of The current ripple of phase-A is also reduced together
the proposed techniques. The sampling time of the digital with the torque ripple reduction.
controller in the experimental setup is set to 150 s. The The experimental result of the proposed torque ripple
electromagnetic torque is estimated by the deterministic reduction method is shown in Fig. 17. Fig. 17 indicates the
model [24]. output torque, output flux, and output current of phase-A.

http://www.jeet.or.kr 1183
 Torque Ripple Reduction in Direct Torque Control of Five-Phase Induction Motor Using Fuzzy Controller with Optimized Voltage Vector ~

(a)

Fig. 17. Experiment result of the DTC with torque ripple

reduction method by the proposed method

(b)
Fig. 19. Experiment results of the DTC in a speed transient
state by using the proposed method (Torque
reference is 5 Nm): (a) 300 rpm600 rpm and (b)
600 rpm300 rpm

(a)

(a)

(b)
Fig. 18. Experiment results of the DTC in a transient state
(2 Nm7 Nm) (a) conventional method and (b)
(b)
proposed method
Fig. 20. Experiment results of the DTC in high torque and
In this experimental result, the maximum torque ripple is speed state (7 Nm, 1600 rpm): (a) conventional
torque ripple reduction method and (b) proposed
1.4 Nm with the maximum output torque of 5.1 Nm and
torque ripple reduction method
the minimum output torque of 3.7 Nm. As a result, the
output torque ripple and the current ripple of phase-A are
reduced when the proposed torque ripple reduction method comparing to the conventional method.
is compared with the result of torque ripple reduction Fig. 19 shows the experiment results in a speed transient
method by using the fixed constant value and normal DTC state by using the proposed method. The torque reference
method. A quasi-rectangular input current induces the is 5 Nm. Speed reference of Fig. 19 (a) is changed from
nearly rectangular flux linkage in the air-gap and thus 300 rpm to 600 rpm. After that, speed reference of Fig. 19
results in higher power density and more torque output. (b) is changed from 600 rpm to 300 rpm. Two results of
Fig. 18 shows the experiment results in a transient state. Fig. 19 are controlled well according to the torque
The torque reference is changed from 2 Nm to 7 Nm. The reference. The experiment results of the DTC in high
torque is controlled by its reference using both the torque and speed state are shown in Fig. 20. Fig. 20(a)
conventional method and proposed method in a transient shows the conventional torque ripple reduction method.
state. However, torque ripple of Fig. 18(b) is decreased by As explained in Fig. 14, the conventional method for the

1184 J Electr Eng Technol.2017; 12(3): 1177-1186

 Hye Ung Shin, Seong Yun Kang and Kyo-Beum Lee

torque ripple reduction is not controlled by the torque [5] Y. N. Tatte and M. V. Aware, Selection of Voltage
reference. On the other hand, the proposed control method Vectors in Three-Level Five-Phase Direct Torque
for the torque ripple reduction has good performance in Control for Performance Improvement, J. Power
high torque and speed region. Electron., vol. 16, no. 6, pp. 2162-2172, Nov. 2016.
[6] L. Zheng, J. E. Fletcher, B. W. Williams, and X. He,
A Novel Direct Torque Control Scheme for a
6. Conclusion Sensorless Five-Phase Induction Motor Drive, IEEE
Trans. Ind. Electron., vol. 58, no. 2, pp. 503-513, Feb.
An improved torque ripple reduction method in the five- 2011.
phase induction motor drive for DTC with the optimal [7] B. Singh, S. Jain, and S. Dwivedi, Torque Ripple
selection of the voltage vector has been presented. A Reduction Technique with Improved Flux Response
proposed method based on the optimal selection of the for a Direct Torque Control Induction Motor Drive,
voltage vector states is conducted in the control period of IET Power Electron., vol. 6, no. 2, pp. 326-342, Feb.
150 s. Compared to the conventional method of the torque 2013.
ripple reduction, the proposed method is capable of [8] Y. Zhang and J. Zhu, A Novel Duty Cycle Control
reducing the torque ripple more than at least 30%. The Strategy to Reduce Both Torque and Flux Ripples
insertion time of voltage vector is calculated by the fuzzy for DTC of Permanent Magnet Synchronous Motor
system in the proposed algorithm which is designed Drives with Switching Frequency Reduction, IEEE
optimally with using the simulation results and testing in Trans. Power Electron., vol. 26, no. 10, pp. 3055-
this system. Therefore, the proposed method has advantage 3067, Oct. 2011.
in the independent motor parameters. Additionally, the [9] X. Xiao and C. Chen, Reduction of Torque Ripple
optimal voltage vector is used in the proposed algorithm Due to Demagnetization in PMSM Using Current
for the torque ripple reduction acquired by considering the Compensation, IEEE Trans. Appl. Supercon., vol. 20,
torque error in five-phase induction motor. The result of the issue 3, pp. 1068-1071, Jun. 2010.
proposed control schematics providing high performance is [10] Y. S. Cho, K. B. Lee, Y. I. Lee, and J. H. Song,
shown in experiment and the simulation results. Torque-Ripple Minimization and Fast Dynamic
Scheme for Torque Predictive Control of Permanent-
Magnet Synchronous Motors, IEEE Trans. Power
Acknowledgements Electron., vol. 30, no. 4, pp. 2182-2190, Apr. 2015.
[11] Y. Zhang, J. Zhu, W. Xu, and Y. Guo, A Simple
This research was supported by Basic Science Research Method to Reduce Torque Ripple in Direct Torque-
Program through the National Research Foundation of Controlled Permanent-Magnet Synchronous Motor
Korea (NRF) funded by the Ministry of Science, ICT & by Using Vectors with Variable Amplitude and Angle,
Future Planning (2016R1A2B4010636). IEEE Trans. Ind. Electron., vol. 58, no. 7, pp. 2848-
2859, Jul. 2011.
[12] Y. Zhang and J. Zhu, Direct Torque Control of
References Permanent Magnet Synchronous Motor with Reduced
Torque Ripple and Commutation Frequency, IEEE
[1] J. A. Riveros, J. Prieto, F. Barrero, S. Toral, M. Jones, Trans. Power Electron., vol. 26, no. 1, pp. 235-248,
and E. Levi, Predictive Torque Control for Five- Jan. 2011.
Phase Induction Motor Drives, in Proc. 36th Annu. [13] T. Ramesh, A. K. Panda, and S. S. Kumar, MRAS
Conf. IEEE Ind. Electron. Soc., pp. 2467-2472, 2010. Speed Estimator Based on Type-1 and Type-2 Fuzzy
[2] H. A. Toliyat and H. Xu, A Novel Direct Torque Logic Controller for the Speed Sensorless DTFC-
Control (DTC) Method for Five-Phase Induction SVPWM of an Induction Motor Drive, J. Power
Machines, in Proc. IEEE APEC, New Orleans, LA, Electron., vol. 15, no. 3, pp. 730-740, May. 2015.
pp. 162-168, 2000. [14] S. Y. Kang, H. U. Shin, and K. B. Lee, Improved
[3] K. B. Lee and F. Blaabjerg, Performance Improvement Torque Ripple Reduction Method of Five-Phase
of DTC for Induction Motor-Fed by Three-Level Induction Motor Using Fuzzy Controller, in Proc.
Inverter with an Uncertainty Observer Using RBFN, IPEMC-ECCE Asia Conf., pp. 635-640, 2016.
IEEE Trans. Energy Convers., vol. 20, no. 2, pp. 276- [15] Y. N. Tatte and M. V. Aware, Torque Ripple
283, Jun. 2005. Reduction in Five-Phase Direct Torque Controlled
[4] M. O. Kizilkaya and K. G, Feed-Forward Approach Induction Motor, in Proc. PEDES Conf., pp. 1-5,
in Stator-Flux-Oriented Direct Torque Control of 2014.
Induction Motor with Space Vector Pulse-Width [16] L. R. L. V. Raj, A. Jidin, Z. Ibrahim, K. A. Karim, M.
Modulation, J. Power Electron., vol. 16, no. 3, pp. A. Said, and M. H. Jopri, Optimal Torque Control
994-1003, May. 2016. Performance of DTC of 5-Phase Induction Machine,

http://www.jeet.or.kr 1185
 Torque Ripple Reduction in Direct Torque Control of Five-Phase Induction Motor Using Fuzzy Controller with Optimized Voltage Vector ~

in Proc. ICEMS Conf., pp. 2094-2099, 2013. Seong Yun Kang He received the B.S.
[17] G. Foo and M. F. Rahman, Sensorless Direct Torque and M.S. degrees in electrical and
and Flux-Controlled IPM Synchronous Motor Drive electronic engineering from the Ajou
at Very Low Speed without Signal Injection, IEEE University, Suwon, Korea, in 2015 and
Trans. Ind. Electron., vol. 57, no. 1, pp. 395-403, Jan. 2017, respectively. Since 2017, he has
2010. been with the Dawonsys Co., Ltd, from
[18] K. B. Lee, J. H. Song, I. Choy, and J. Y. Yoo, Torque Ansan, Korea in present. His research
Ripple Reduction in DTC of Induction Motor Driven interests include motor drives and grid-
by Three-Level Inverter with Low Switching Fre- connected systems.
quency, IEEE Trans. Power Electron., vol. 17, no. 2,
pp. 255-264, Mar. 2002.
[19] K. K. Shyu, J. K. Lin, V. T. Pham, M. J. Yang, and Kyo-Beum Lee He received the B.S.
T. W. Wang, Global Minimum Torque Ripple Design and M.S. degrees in electrical and
for Direct Torque Control of Induction Motor Drives, electronic engineering from the Ajou
IEEE Trans. Ind. Electron., vol. 57, no. 9, pp. 3148- University, Suwon, Korea, in 1997 and
3156, Sep. 2010. 1999, respectively. He received the
[20] R. Arulmozhiyal, Design and Implementation of Ph.D. degree in electrical engineering
Fuzzy PID Controller for BLDC Motor Using FPGA, from Korea University, Seoul, Korea,
in Proc. PEDES Conf., pp. 1-6., Dec. 2012. in 2003. From 2003 to 2006, he was
[21] R. J. Wai and K. H. Su, Adaptive Enhanced Fuzzy with the Institute of Energy Technology, Aalborg University,
Sliding-Mode Control for Electrical Servo Drive, Aalborg, Denmark. From 2006 to 2007, he was with the
IEEE Trans. Ind. Electron., vol. 53, no. 2, pp. 569- Division of Electronics and Information Engineering,
580, Apr. 2006. Chonbuk National University, Jeonju, Korea. In 2007, he
[22] H. H. Choi, H. M. Yun, and Y. Kim, Implementation joined the Department of Electrical and Computer
of Evolutionary Fuzzy PID Speed Controller for PM Engineering, Ajou University, Suwon, Korea. His research
Synchronous Motor, IEEE Trans. Ind. Informat., vol. interests include electric machine drives, renewable power
11, no. 2, pp. 540-547, Apr. 2015. generations, and electric vehicle applications.
[23] C. L. Tseng, S. Y. Wang, S. C. Chien, and C. Y. Chang,
Development of a Self-Tuning TSK-Fuzzy Speed
Control Strategy for Switched Reluctance Motor,
IEEE Trans. Power Electron., vol. 27, no. 4, pp.
2141-2152, Apr. 2012.
[24] L. Zheng, J. E. Fletcher, B. W. Williams, and X. He,
A Novel Direct Torque Control Scheme for a
Sensorless Five-Phase Induction Motor Drive, IEEE
Trans. Ind. Electron., vol. 58, no. 2, pp. 503-513, Feb.
2011.

Hye-Ung Shin He received the B.S. in

Electrical and Computer Engineering
from Kunsan University, Kunsan,
Korea, in 2012. He received the M.S.
in Department of Electronic Systems
Engineering from Hanyang University,
Ansan, Korea, in 2014. He received the
Ph.D. degree in electrical engineering
from the Ajou University, Suwon, Korea, in 2017. His
research interests include machine design, electric machine
drives and power conversion.

1186 J Electr Eng Technol.2017; 12(3): 1177-1186