International Journal of Engineering & Technology, 7 (2.

14) (2018) 267-271

International Journal of Engineering & Technology

Website: www.sciencepubco.com/index.php/IJET

Research Paper

Identification and Optimization Speed Control of BLDC Motor Using

Fuzzy Logic Controller
Izza Anshory1,2*, Imam Robandi1, Wirawan1
1
Department of Electrical Engineering, Institut Teknologi Sepuluh Nopember(ITS), Surabaya, Indonesia
2
Department of Electrical Engineering, Universitas Muhammadiyah Sidoarjo, Sidoarjo, Indonesia
*Corresponding author E-mail: izza.anshory15@mhs.ee.its.ac.id

Abstract
Better performance, such as speed of response, overshoot and steady state error on BLDC motors, needs to be improved continuously
through optimization. The first method, is to design the hardware, take the current data, voltage and motor speed through measurement.
The data obtained is simulated through the existing tools in Matlab to obtain the mathematical model in the form of transfer function.
BLDC motor transfer function obtained, used as the basis for optimizing the fuzzy logic method. The results showed that the stability
performance in BLDC motor speed control system showed an improvement after optimization using fuzzy logic.

Keywords: Instrumentation, fuzzy logic, transfer function, BLDC Motor

1. Introduction 2. Theoretical Background

Brushless Direct Current (BLDC) motors have been widely used
A. BLDC Motor
in various industrial applications, including transportation, home One of the BLDC motor applications in the world of transporta-
appliance and office automation, and home industries[1][2]. One
tion industry is as an electric bicycle drive. BLDC motors in-
form of BLDC motor applications that are used in transportation
clude a permanent type of magnet synchronous motor which
is as an electric bicycle drive. BLDC motor is widely used be-
means the magnetic field is generated by the stator and by the
cause it has many advantages and little loss. The advantages of rotating rotor at the same frequency. BLDC motor configura-
BLDC motor include high efficiency, easy maintenance and
tions have 1 phase, 2 phase and 3 phase. In accordance with its
precision control. However, BLDC motors also have the disad-
type, the stator has the same number of turns and for the 3-phase
vantages of higher cost and heavier weight when compared to
type motor is the most widely used type of motor. There are two
Direct Current (DC) motors at same great power[3]. types of stator winding variant, which is trapezoidal and sinus-
To improve the better performance of the BLDC motor, optimi-
oidal electric motor. This distinction is made on the basis of the
zation is required. The optimization process can be done by
interconnection of the coil within the stator winding to provide
performing mathematical modeling in the form of transfer func-
various types of back emf[6].
tion and simulate it. By analyzing the parameters of motor rota-
The BLDC motor model equation consists of a voltage equation,
tion speed, and electric current, we will get the best control
a torque equation and a motion equation. The stator on a BLDC
strategy to improve system performance[4].
motor generally has three windings as it does on an induction
Motor rotation speed parameter, and electric current signal is
motor or a permanent magnet synchronous motor. The circuit
obtained by designing a BLDC motor drive system connected
diagram for stator windings is shown in Fig.1.
with arduino microcontroller. The generation of pulse width
modulation (pwm) signals from the microcontroller results in
changes in voltage and current values. The change of value will
cause the change of switching speed of the mosfet transistor
which affect the rotation speed of the BLDC motor[5].
BLDC motor used in this study is a motor for electric bicycle
drive with 350 Watt power, which is connected with the current
and speed sensor. Results of instrumentation on the BLDC mo-
tor, then simulated with tools contained in the application pro-
gram Matlab to obtain mathematical modeling in the form of
transfer function.
Fig.1: Stator circuit diagram

The voltage equation at different terminals is shown in the fol-

lowing equation [2]:

Copyright © 2018 Authors. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted
use, distribution, and reproduction in any medium, provided the original work is properly cited.
268 International Journal of Engineering & Technology

𝑑𝑖 𝐴
𝑈𝐴 = 𝑅𝐼𝐴 + (𝐿 − 𝑀) + 𝑒𝐴
𝑑𝑡
𝑑𝑖 𝐵
𝑈𝐵 = 𝑅𝐼𝐵 + (𝐿 − 𝑀) + 𝑒𝑏
𝑑𝑡
𝑑𝑖 𝐶
𝑈𝑐 = 𝑅𝐼𝑐 + (𝐿 − 𝑀) + 𝑒𝑐
𝑑𝑡
With Laplace transform the BLDC motor transfer function can
be written as follows:

Ω(s) 𝐾𝑇 Fig.4: Design Optimization BLDC Motor

𝐺𝑈 𝑠 = =
𝑈𝑑 (𝑠) 𝐿𝑎 𝐽𝑠 2 + 𝑟𝑎 𝐽 + 𝐿𝑎 𝐵𝑣 𝑠 + (𝑟𝑎 𝐵𝑣 + 𝐾𝑒 𝐾𝑡 )
Therefore the method used is to obtain mathematical model of
B. Instrumentation Parameters BLDC motor transfer function to be optimized with fuzzy logic.
Difficulty in knowing the parameters of BLDC motors used as
electric bike drivers, such as resistance value, inductance, elec- 4. Results and Discussion
tronic torque, and mechanical torque, cause mathematical mod-
elling in the form of transfer function not achieved. Therefore,
the stage is done to identify a system through measurement. A. Instrumentation Design
System identification is a process done to get a mathematical The hardware design results to obtain the measurement data of
model of a system based on data input and output parameters[7]. voltage, current, motor load and motor rotation speed are shown
There are three modeling approaches used to identify a system, in Fig. 5.
ie white box model, gray box model and black box model based
on measurement data. Procedures are performed to identify a
dynamic system model, as shown in Fig.2. [8].

Fig.2: System Identification Fig.5: Design Results hardware design

The system identification in a model describes the measured Sensors used in measurement data collection include current
relationship between the input signal and the output signal[9]. sensors, voltage, load and motor rotation speed. The sensor
System identification toolbox (SIT), an application contained in serves to take input and output data. The input data used in this
the Matlab program used to model a system mathematically, study include current, voltage, and load. While data output taken
based on input and output data collection[10]. data in the form of rotational speed of motor. The hardware
design results are shown in Fig. 6
3. Design Instrumentation And Simulation

A. Instrumentation Design
The block diagram used to perform the measurement of the cur-
rent signal and motor rotation speed as shown in Fig. 3.

Fig.3: Block Diagram Design Instrumentation

.
System in Figure 3, can be explained that data retrieval can be Fig.6: Hardware Design Results
done after all hardware is connected properly. The first stage of
data retrieval for signal current, voltage, and motor speed. After After that done taking input and output data from hardware de-
the data collected the analysis is done by simulating the data into sign result, and the result is as follows:
the application program Matlab, namely System Identification
Toolbox (SIT). Table.1: BLDC Motor Identification and Results
PWM I1(A) I2 (A) I3 (A) V1 (V) V2 (V) V3 (V) RPM
B. Design Optimization with Fuzzy Logic 177 0,12 0,07 0,91 0 10 45 0
179 0,78 0,28 0,59 45 0 15 346
Optimization is done with the aim of improving system perfor-
181 0,3 0,25 0,3 0 0 0 273
mance such as speed of response time, overshoot, and steady- 183 0,07 0,81 0,81 20 0 45 273
state error. In Fig.6 it can be explained that for a BLDC motor 185 0,96 0,01 1,12 30 10 0 280
the plant will be optimized with a Proportional Integral Deriva- 187 0,04 0,15 1,44 30 20 0 283
tive (PID) controller. The results will be compared if optimized 189 0,59 1,02 0,59 20 45 0 290
by using fuzzy logic. 191 0,01 0,22 0,3 0 45 20 296
International Journal of Engineering & Technology 269

193 0,3 0,7 0,54 0 45 15 310 After being simulated with step-unit input, with simulation time
195 0,62 0,25 0,7 0 45 25 316 of 10 s, the result shown is shown in Fig. 10.
197 0,2 0,28 1,02 30 0 0 323
199 0,2 0,01 1,18 15 0 0 330
201 0,25 0,49 0,41 10 45 0 330
203 0,07 0,52 1,49 45 30 0 343
205 0,38 0,07 0,07 0 10 25 343
207 0,2 0,28 0,2 0 0 0 353
209 0,46 0,65 1,31 15 45 0 356
211 1,2 0,28 1,04 45 0 0 366
213 0,99 0,41 0,49 0 40 40 370
215 0,01 0,09 0,96 35 35 0 373
In Table 1.Shows the measurement results for BLDC motor
speed control. From these data indicate the instability of the
motor rotation speed value, which adjusts the magnitude of volt-
age and current values.
Furthermore, the measurement data for input and output were
analyzed using toolbox identification software to obtain the Fig.10: Unit Step Response
mathematical model in the form of transfer function of the
BLDC motor plant. The transfer function is a system defined as From the plot above, it can be explained that the system re-
the ratio between output and input. In Fig.7, shows the stages of sponse when in the first second, the position of the moving mo-
data processing input and output of BLDC motor to obtain the tor changes the input signal until the response enters the steady
transfer function. state.

Fig.11: Block Diagram of PID Controller System

Once simulated with close loop system optimized by conven-

tional method that is using PID controller. The result is shown in
Fig. 12.

Fig.7: System Identification Toolbox Process

After the process of processing and analysis then obtained math-

ematical modeling in the form of transfer function, as shown in
Fig.8.’

Fig.12: Close Loop Response Using PID Controller

In Fig. 12, the system uses a PID controller, with setting values
Fig.8: Results identification of transfer function for Kp = 0,51429, Ki = 10,7136 and Kd = 0.0019418. The result
shows the instability of the 1st-second, and then the system is
Fig.8 shows the mathematical modeling in the transfer function stable.
for the BLDC motor plant, which is:
B. Optimization Using Fuzzy Logic Controller
The performance of the controller depends on the value of Pro-
𝐺 𝑠 portional, Integral, and Derivative parameters (P, I and D). In
𝑠
order to parameterize the PID it requires experimentation and
The acquired transfer function is used as a basis for optimization high knowledge (expert knowledge). There are several steps that
with PID controller, and fuzzy logic controller. In Figure 9, need to be considered to get a fuzzy output, such as fuzzification,
shows determining membership functions, rule evaluation, and defuzzi-
fication. Fuzzification process conducted with the aim to change
the input variables that are not fuzzy crisps to the fuzzy set
where in this process crisps which are variable; error and delta
error converted into fuzzy variables using techniques member-
ship functions (MF). The membership function consists of five
Fig.9: Block Diagram of Open Loop System sets which are NB: Negative Big, NS: Negative Small, ZE: Zero,
PS: Positive Small and PB: Positive Big.
270 International Journal of Engineering & Technology

Fig.13: Input variables of fuzzy logic error

Fig.19: Surface view of Kfi fuzzy controller

Fig.14: Input variables of fuzzy logic delta error

Fig.20: Surface view of Kfd fuzzy controller

Fig.18-20, shows the surface generation of the rule of member-

ship fuzzy logic controller, which is a combination of two inputs
and three outputs. Then the result of the rule membership func-
tion is simulated in fuzzy logic controller optimization design as
Fig.15: Output variables Kfp of fuzzy logic delta error shown in Fig. 21.

Fig.21: Design of PID Optimization by Fuzzy Logic

The simulation results of the design optimization fuzzy logic

Fig.16: Output variables Kfi of fuzzy logic delta error controller, shown in Fig.22.

Fig.17: Output variables Kfd of fuzzy logic delta error

Fig.23: Optimization of PID Controller Using Fuzzy Logic

Fig. 23, also shows the optimization process of PID controller

by using fuzzy logic controller. The process of fuzzy logic con-
troller in BLDC motor speed control system can improve stabil-
ity, reduce overshoot, and faster reach steady state.

Fig.18: Surface view of Kfp fuzzy controller

International Journal of Engineering & Technology 271

5. Conclusion
The process of controlling the speed control of the BLDC motor
for the type of electric bike begins with the identification system
to know the mathematical modeling in the form of transfer func-
tion. The transfer function obtained was used as the basis for
optimizing by using PID controller. Fuzzy logic controller to get
the system performance then optimizes the PID controller. The
results show that the transfer function has been found to increase
stability, reduce overshoot, and increase steady state achieve-
ment.

Acknowledgment
The authors would like to thank the Ministry of Technology and
Higher Education who has provided funding assistance for doc-
toral dissertation research.

References
[1] A. Sabir and M. Kassas, “A novel and simple hybrid fuzzy/pi
controller for brushless dc motor drives,” Automatika, vol. 56, no.
4, pp. 424–435, 2015.
[2] I. Anshory, I. Robandi, and Wirawan, “Monitoring and
optimization of speed settings for Brushless Direct Current
(BLDC) using Particle Swarm Optimization (PSO),” in
Proceedings - 2016 IEEE Region 10 Symposium, TENSYMP
2016, 2016.
[3] K. Nam, W. Lee, and J. Hong, “Reducing Torque Ripple of
BLDC Motor by Varying Input Voltage,” IEEE Trans. Magn.,
vol. 42, no. 4, pp. 1307–1310, 2006.
[4] C. Cai, H. Zhang, J. Liu, and Y. Gao, “Modeling and Simulation
of BLDC motor in Electric Power Steering,” 2010 Asia-Pacific
Power Energy Eng. Conf., pp. 1–4, 2010.
[5] S. A. Zulkifli, M. N. Hussin, and A. S. Saad, “MATLAB-
Arduino as a low cost microcontroller for 3 phase inverter,” 2014
IEEE Student Conf. Res. Dev. SCOReD 2014, 2014.
[6] B. Das, S. Chakrabarti, P. R. Kasari, and A. Chakraborti, “Novel
reverse regeneration technique of BLDC motor for capacitor
charging,” in 2014 International Conference on Control,
Instrumentation, Energy and Communication (CIEC), 2014, pp.
246–253.
[7] M. Fruk, G. Vujisić, and T. Špoljarić, “Parameter Identification
of Transfer Functions Using MATLAB,” pp. 571–576.
[8] M. M. Hassan, A. A. Aly, and A. F. Rashwan, “DIFFERENT
IDENTIFICATION METHODS WITH APPLICATION TO A
DC MOTOR,” vol. 35, no. 6, pp. 1481–1493, 2007.
[9] L. Ljung, “Ljung L System Identification Theory for User.pdf,”
PTR Prentice Hall Upper Saddle River NJ, vol. 25. pp. 475–476,
1987.
[10] M. Fruk, G. Vujisić, and T. Špoljarić, “Parameter Identification
of Transfer Functions Using MATLAB,” pp. 571–576, 2013.