International Journal of Electrical and Computer Engineering (IJECE)

Vol. 14, No. 6, December 2024, pp. 6237~6247

ISSN: 2088-8708, DOI: 10.11591/ijece.v14i6.pp6237-6247  6237

High-performance speed control for three-phase induction

motor based on reverse direction algorithm and artificial neural
network

Mustafa A. Al-Khawaldeh1, Jasim A. Ghaeb1, Samer Z. Salah2, Mohammad S. Alrawajfeh3

1
Department of Mechatronics Engineering, Faculty of Engineering and Technology, Philadelphia University, Amman, Jordan
2
Department of Mechatronics Engineering, Faculty of Engineering, The University of Jordan, Amman, Jordan
3
Al-Warkaa Food Industries Company, Amman, Jordan

Article Info ABSTRACT

Article history: This research proposes two approaches for determining the required
frequency and modulation index for a pulse-width-modulation (PWM)
Received Apr 15, 2024 system in a variable frequency drive (VFD) to control the speed of the three-
Revised Jul 7, 2024 phase induction motor. The first approach which is the reverse direction
Accepted Aug 6, 2024 algorithm (RDA), uses a set of equations to calculate the necessary
frequency and voltage for maintaining a constant motor speed under varying
load conditions. The second one involves training a neural network (NN) on
Keywords: data collected by the RDA, which can then be used to continuously adjust
the motor speed in real time to adapt to changing load torque requirements.
Induction motors Simulation and laboratory models for the three-phase induction motor are
Load torque built and the proposed RDA-NN controller is examined. Results have proved
Neural network that the proposed controller is effective in providing a stable and responsive
Pulse-width-modulation motor speed control system.
Speed control
This is an open access article under the CC BY-SA license.
Variable frequency drive

Corresponding Author:
Mustafa A. Al-Khawaldeh
Department of Mechatronics Engineering, Faculty of Engineering and Technology, Philadelphia University
Amman, Jordan
Email: malkhawaldeh@philadelphia.edu.jo

1. INTRODUCTION
Induction motors are increasingly used in the electro-mechanical energy conversion field thanks to
their high performance, low cost and reliability [1]. Most of these applications require not only quick
response but also speed control to maximize torque and achieve high efficiency. That is because when the
load on an induction motor changes, the speed of the motor also changes. However, mechanical loads should
be driven at a desired specific speed. The industry must keep the motor speed constant to achieve reliable and
robust machine operation efficiency at different loads. In the past, the three-phase induction motors which
used to be operated at a constant speed were difficult to control due to the inefficient conventional methods.
Thus, a need arises to propose new speed control methods for the three-phase induction motors which the
present study aims to address.
New control methods have allowed the development of advanced methods to control the speed of
induction motors in the industry [2]. For example, Mobarra et al. [3] have developed a control algorithm for
the alternating current (AC) compensator which drives the stator of the variable speed diesel generator. The
speed is modified using a variable frequency drive in response to the load. The generator rotating stator is
adjusted by the controller to enhance the operational efficiency taking into account load variation. The
induction motor's performance could be improved by choosing the right control [4]. The speed control of a
three-phase induction motor can be achieved by changing the supply frequency, the number of stator poles,

Journal homepage: http://ijece.iaescore.com

6238  ISSN: 2088-8708

or the rotor slip. The developed electromagnetic torque is proportional to the ratio of supply voltage and
frequency. For maintaining a constant electromagnetic torque, this ratio has to be kept constant throughout
the speed range [5].
Variable frequency drive (VFD) is used for this purpose [6], [7]. VFDs control the motor speed and
motor torque by controlling the frequency and magnitude of voltages and currents supplied to the motor,
through the pulse-width-modulation (PWM) [8]. Scalar and vector methods are used in variable frequency
control for induction motors. They provide an acceptable response to variable speed and load situations [9].
In [10], the frequency control for induction motors is performed using curve fitting methods based on
Polynomial, Fourier, and Gaussian models. It is expected that these methods could improve motor
performance; the Polynomial curve fitting method has in particular shown the best performance. In [11], a
multilevel inverter fed medium voltage VFDs is used to demonstrate the three control methods: scalar control
(SC), indirect field oriented control (IFOC), and direct torque control (DTC). The DTC has shown a better
dynamic response in comparison to SC and IFOC.
Khudier et al. [12] propose a programming logic controller (PLC) to regulate the VFD for a
constant-speed induction motor subjected to load changes. The frequency varies from 50 to 59.2 Hz when the
load current is changed from 0.095 to 0.774 A for a constant average speed of 1402.45 rpm. Advanced speed
control for a five-leg voltage source inverter (FLVSI) that drives a dual three-phase induction motor system
used in industry is proposed by Lim et al. [13], and the results guarantee stability.
In [14], a simple and elegant control algorithm has been proposed for realizing the DTC-space
vector sinusoidal pulse width modulation (SVSPWM) that relies less on motor parameters. Because there is
just one proportional integral (PI) controller required, there is a significant reduction in the computational
work. In [15], various VFDs operating under induced voltage magnitude variation and voltage imbalance are
compared for their efficiency outcomes using the same motor. In general, when the voltage imbalance
increases, the efficiency of the VFD system decreases.
Azizipanah-Abarghooee and Malekpour [16] suggest an induction motor variable frequency drive to
test its suitability for managing the smart grid's frequency. This controller allows the driver to reduce its
power in response to a drop in grid frequency. Through a motor's speed rate limiter, the dynamic limitation of
the drive caused by the inertia of the load is taken into account.
Intelligent control techniques are used to improve the performance of the three-phase induction
motors subjected to load changes. To reduce the steady-state error and transient response of an induction
motor's speed under varying load conditions, the fuzzy logic controller is proposed by Firdaus et al. [17].
Four families of machine learning algorithms (i.e. decision trees, support vector machines, k-nearest
neighbors, and ensemble) were utilized in MATLAB for fault diagnosis of induction motors fed by VFDs.
For each test, the motors' stator current and vibration signals are recorded concurrently under steady-state
conditions, and the usefulness of both signals for defect identification is assessed. In the lab, two identical
0.25 HP induction motors are tested with various VFD output frequencies and motor loading under healthy,
single and multi-fault circumstances.
Zemmit et al. [18] and Metha and Karvekar [19] have developed an improved DTC of a three-phase
induction motor utilizing a PI controller modified using a genetic algorithm (GA) and fuzzy logic controller
(FLC) to address the problems with conventional DTC that are characterized by high torque and flux ripples.
Their results show improvements in torque and flux ripple. It has reduced the torque and flux ripple by more
than 64.44% and 50%, respectively. Habib [20] present DTC using a traditional PI speed regulator based on
fuzzy logic (FL) and an adaptive neuro-fuzzy inference system (ANFIS). Comparisons between the outcomes
of the simulations using the ANFIS and FL intelligent regulators and a conventional DTC with 36 sectors are
made. The proposed method is valid and reduces the electromagnetic torque as well as the total harmonic
distortion (THD) value of the stator current and stator flux ripple, according to a comparison between DTC
command and intelligent regulators.
An artificial neural network (ANN) control is proposed in [21] for a smooth start of a three-phase
induction motor. The results of the proposed acceleration validate the effectiveness of the ANN controller
under different loading conditions. To get the desired stability and the best response for getting the
appropriate speed of the induction motor, a nonlinear model predictive control is applied [22]. An intelligent
control algorithm using a PID controller-based back-propagation neural network (BP-NN) is proposed for the
AC motor speed control. The results demonstrate the system’s adaptability, robustness and ability to
intelligently regulate the speed of the AC motor [23]. Other meta-heuristic control approaches could also be
used to enhance the smooth start of a three-phase induction motor [24], [25].
Considering the previous related research, this research contributes to knowledge on speed control
of the three-phase induction motor through presenting two approaches: reveres direction algorithm (RDA)
and neural network (NN) to determine both the supply frequency and supply voltage for the three-phase
induction motor for a constant speed at variable load torques. When a load change is applied to the motor

Int J Elec & Comp Eng, Vol. 14, No. 6, December 2024: 6237-6247
Int J Elec & Comp Eng ISSN: 2088-8708  6239

shaft, the RDA calculates the necessary frequency and the modulation index of the PWM inside the VFD to
retrieve the rated speed. To calculate the required frequency and modulation index by RDA, the motor speed
is recorded and recycled through a series of equations in the reverse direction of the motor’s power flow. In
the second stage, a real-time NN controller is proposed to retrieve the rated speed of the three-phase
induction motor subjected to the load change. The NN is trained for a constant speed of the three-phase
induction motor speed using the data set obtained from the RDA control stage.
The rest of the paper is organized as follows; a representation of the three-phase induction motor is
presented in section 2. Section 3 describes the method; the proposed RDA and NN controllers. The results
and discussions are given in section 4, and section 5 concludes the work.

2. REPRESENTATION OF THE THREE-PHASE INDUCTION MOTOR

Although the three-phase induction motors are fixed speed, they are widely used in industry due to
their simplicity and low cost. The VFD provides the variable frequency for a variable speed induction motor
[26]. The three-phase induction motor has two parts; the stator and the rotor. The per-phase equivalent circuit
of the three-phase induction motor is shown in Figure 1 [27]. When the three-phase power supply voltage is
applied to the stator winding, a rotating magnetic flux is produced in the air gap. Due to this rotating
magnetic flux, an induced voltage is produced in the rotor circuit. The interactivity between the stator flux
and rotor flux rotates the rotor at a speed less than the synchronous speed, causing a differential speed
between the synchronous speed and rotor speed called rotor slip speed.

Is rs jXs Ir jXr
+ Io

Vs
rc jXm rr/s
(f)

Figure 1. The per-phase equivalent circuit of the 3-phase induction motor

To analyze the three-phase induction motor, there are two methods to represent the motor: the actual
variables model and the 𝑑 − 𝑞 model. The advantage of the 𝑑 − 𝑞 model is that it has fewer variables in the
representative equations and the inductances are independent of the position of the rotor [28]. In the 𝑑 − 𝑞
modelling of the motor, the three 𝑎, 𝑏, and 𝑐 variables are transformed into two-axis variables to get the
speed and torque expressions. The Park transformation for stator 𝑃(𝜃𝑆 ) is given in (1) [29].

𝐼𝑑 𝐼𝑎
𝐼
[ 𝑞 ] = 𝑃(𝜃𝑆 [𝐼𝑏 ]
) (1)
𝐼0 𝐼𝑐

where

2𝜋 2𝜋
𝑐𝑜𝑠(−𝜃𝑆 ) cos (−𝜃𝑆 + ) cos (−𝜃𝑆 − )
3 3
2 2𝜋 2𝜋
𝑃(𝜃𝑆 ) = sin(−𝜃𝑆 ) sin (−𝜃𝑆 + ) sin (−𝜃𝑆 − ) (2)
3 3 3
1 1 1
[ 2 2 2 ]

Similarly, the transformation for the rotor can be obtained by replacing θS by (θS − θr ). Also, the electro-
magnetic torque can be represented in 𝑑 − 𝑞 form as in (3):
3
𝑇𝑒 = 𝑃𝐿𝑚 (𝑖𝑞𝑠 𝑖𝑑𝑟 − 𝑖𝑑𝑠 𝑖𝑞𝑟 ) (3)
2

where 𝜃𝑟 : Rotor angle, 𝜃𝑆 : Stator angle, 𝑃: No of poles, and 𝐿𝑚 : Mutal inductance.

High-performance speed control for three-phase induction motor based on … (Mustafa A. Al-Khawaldeh)
6240  ISSN: 2088-8708

3. METHOD: SPEED CONTROL OF THE THREE-PHASE INDUCTION MOTOR

This section presents a thorough description of the proposed methodologies and data analysis
procedures. Two phases are proposed to control the speed of a three-phase induction motor by restoring it to
its rated speed following a torque change. In phase one, an RDA is proposed and explained in subsection 3.1.
This algorithm determines the motor's necessary supply frequency depending on speed, torque, power, and
slip. The required voltage for the motor is calculated using the constant relationship between supply
frequency and voltage. The RDA is proven as a high-performance control for the induction motor as
demonstrated in subsection 3.1.
In the second stage of this research, an NN technique is proposed as an online mode controller to
restore the rated speed of the three-phase induction motor after the load torque change. This technique
provides a quick response to load torque changes made to the motor shaft. The data set is acquired from the
first phase and is utilized to train the ANN for real-time control of motor speed subjected to load fluctuation
since the results achieved in the first phase utilizing RDA are accurate and highly perfect.

3.1. Reverse direction algorithm for a constant speed at variable load conditions
A change in the motor load causes a speed variation on the motor. The speed of the three-phase
induction motor depends on the stator frequency, the number of stator poles and rotor slip. When the load
torque is changed, the rotor speed, air-gap power (𝑃2 ), and mechanical power (𝑃𝑚 ) are changed.
In this work, RDA for a constant speed at variable load conditions is used to determine the supply
frequency required for a constant speed at variable load conditions. When a load change is applied to the
motor shaft, the RDA calculates the necessary frequency and the modulation index of the PWM inside the
VFD to retrieve the rated speed. In order to calculate the required frequency and modulation index by RDA,
the motor speed is recorded and recycled through a series of equations in the reverse direction of the motor’s
power flow. This RDA employs the equations related to the speed, torque, power, slip, and supply frequency
of the motor. The steps below explain how the speed of the motor is controlled by RDA after being subjected
to a load change.
− Record the rotor speed (𝜔𝑟 ) and electromagnetic torque (𝑇𝑒 ), when the load torque is changed.
− Determine the air-gap power (𝑃2 ) and mechanical power (𝑃𝑚 ) of the induction motor according to (4) and
(5):

𝑃2 = 𝑇𝑒 × 𝜔𝑠 (4)

where 𝜔𝑠 is the synchronous speed at a rated frequency.

𝑃𝑚 = 𝑇𝑒 × 𝜔𝑟 (5)

− Use the air-gap power (𝑃2 ) and mechanical power (𝑃𝑚 ) to determine the rotor slip, as demonstrated in (6).
𝑃𝑚
𝑆 = 1− (6)
𝑃2

− Substitute the rotor slip in (7) to calculate the required synchronous speed for retrieving the rated speed of
the motor. When the slip in (5) is changed, the synchronous speed in (7) is adjusted to new values if the
load torque changes while maintaining a constant rotor speed.
𝑁𝑟
𝑁𝑠 = (7)
1−𝑆

− Substitute the new value of the synchronous speed in (8) to find the new supply frequency required for
the desired motor speed.
𝑝×𝑁 𝑠
𝑓= (8)
60
where 𝑝 is pole-pairs.
The air-gap flux of the motor is proportional to the ratio of the supply voltage (𝑉𝑠 ) to the frequency
(𝑓). For smooth speed control, constant air-gap flux is required. Also, a high torque may be obtained for all
levels of speed. This can be accomplished by maintaining a constant ratio of (𝑉𝑠/𝑓). Therefore, if there is a
change in the load torque, it is important to change 𝑉𝑠 and f simultaneously. To produce this inside VFD, the
frequency of the reference signal and modulation index are controlled. The RDA control is applied on the
three-phase induction motor for a constant speed at variable load torque conditions. Table 1 shows the RDA

Int J Elec & Comp Eng, Vol. 14, No. 6, December 2024: 6237-6247
Int J Elec & Comp Eng ISSN: 2088-8708  6241

response for load torque changes made to the motor shaft. The rated motor speed to be maintained is
1,467 rpm at a rated torque of 9N.m. The proposed RDA controller demonstrates its superior capability in
maintaining the rated motor speed in the face of the load torque changes as shown by the results.

Table 1. RDA results for load torque changes made to the three-phase induction motor
Case Load Before Control Action After Control Action
No. torque Input Electro- Rotor Air-gap Mech. New New Input Electro- Rotor Air-gap Mech.
TL (Nm) current magnetic speed power, power, stator supply current magnetic speed power, power,
(A) torque, (rpm) P2 (W) Pm (W) freq. voltage (A) torque, (rpm) P2 (W) Pm (W)
Te (Nm) (Hz) (V) Te (Nm)
1 0.495 1.4777 0.9628 1497 151.242 150.937 48.9988 440.9889 1.4624 0.9535 1467 149.768 146.470
2 2.330 1.592 2.796 1491 439.2 436.5 49.1944 442.7493 1.595 2.788 1467 438 428.3
3 3.165 1.679 3.63 1488 570.2 565.7 49.2856 443.5702 1.669 3.623 1467 569.1 556.5
4 4.495 1.852 4.958 1484 778.8 770.4 49.4344 444.9095 1.848 4.958 1467 778 760.8
5 6.495 2.18 6.956 1477 1093 1076 49.6683 447.0145 2.179 6.953 1467 1092 1068
6 7 2.2728 7.4603 1475 1171.9 1152.3 49.7297 447.5670 2.2766 7.4578 1467 1171.5 1145.6
7 12.330 3.442 12.78 1454 2008 1946 50.4642 454.1776 3.456 12.79 1467 2009 1965
8 14 3.8757 14.4504 1445 2269.9 2187.4 50.7442 453.1457 3.9039 14.457 1467 2270.9 2220.4
9 16 4.4488 16.4467 1434 2583.4 2470.5 51.1347 456.6325 4.5334 16.456 1467 2585.0 2528.1
10 17 4.764 17.44 1428 2740 2609 51.3628 458.6699 4.75 17.46 1467 2742 2682

3.2. Neural network control

As a second phase in this work, the NN is proposed as a real-time controller to retrieve the rated
speed of the three-phase induction motor subjected to the load change. It is trained to control the voltage and
frequency of the VFD to keep the motor speed constant. The NN is trained via MATLAB-Simulink
environment. The NN is trained using the RDA data set that was obtained in subsection 3.1. The motor rotor
speed (rpm), mechanical power (W), and motor input current (A), for 100 cases of load torque, are the input
data to the NN. The outputs of the trained ANN are the frequency of the reference signal and the modulation
index that are needed for the VFD to restore the motor speed.
A feed-forward NN is used in this work as shown in Figure 2. The hidden layer contains ten
neurons, the input layer includes three neurons, and the output layer includes two neurons. The training
algorithm used is Levenberg-Marquardt. It requires more memory but less time for training.

Figure 2. ANN structure used in this work

Figure 3 shows the ANN's performance during the training, validation and test stages. The model's
capability to generate the desired output is determined by the regression factor (R). It ranges from 0 to 1. If it
reaches 1, it means that the ANN's training produces the best results. As shown in Figure 3, the regression
factor obtained throughout this training procedure is one for training, validation, testing, and for all, leading
to the finest training. Additionally, Figure 4 displays the mean squared error (MSE) against the periods. It is
shown that the best validation performance is 0.081722 of MSE at epoch 21.

High-performance speed control for three-phase induction motor based on … (Mustafa A. Al-Khawaldeh)
6242  ISSN: 2088-8708

Figure 3. Performance of ANN training in terms of regression factor (R)

Figure 4. Performance of NN training in terms of MSE

4. RESULTS AND DISCUSSION

A 450 V, 3-Phase, 2-pole, 50 Hz induction motor is used. It has a nominal power of 1.38 kW. The
rated speed and the rated torque of this motor are 1467 rpm and 9 N.m respectively. The values of other
parameters of this motor are listed in Table 2.

Int J Elec & Comp Eng, Vol. 14, No. 6, December 2024: 6237-6247
Int J Elec & Comp Eng ISSN: 2088-8708  6243

Table 2. The template values of the three-phase induction motor

The parameter Value
Rated voltage 450 V
Supply frequency 50 Hz
Nominal power 1.38 kW
Stator resistance (R1) 2.45 and
Rotor resistance (R2) 2.7222
Leakage inductance for stator (Ls) 1.56 mH
Leakage inductance for rotor (Lr) 1.56 mH
Mutual inductance (Lm) 0.5634 H
Number of pole pairs 2
Moment of inertia 0.0131 kg.m2
Friction factor 0.002985 N.m.s
Rated speed 1467 rpm
Rated torque 9 N.m.

4.1. Simulation results

In this work, the NN technique is proposed as an online mode controller to restore the rated speed of
the three-phase induction motor after the load torque changes. The block diagram simulation is shown in
Figure 5. Different load torques are applied to evaluate the effectiveness of the proposed NN controller for
the three-phase induction motor's real-time speed control. The proposed NN controller determines the
necessary frequency and the modulation index of the PWM inside the VFD when a load change is applied to
the motor shaft, varying the speed. The 3-phase induction motor receives the output of the VFD at the new
frequency and voltage to restore to the rated speed. The proposed NN controller's and RDA control's actions
for various load changes are shown in Table 3. The results demonstrate the higher capability of the proposed
NN to recover the rated speed as well as the closeness between the two proposed controllers; NN and RDA.
Furthermore, the results show that the load torque variations could recover up and down by roughly twice as
much as the rated torque.

Figure 5. The block diagram simulation

To test the dynamic response of the three-phase induction motor, a simulation model of the motor is
built. Several case studies have been conducted to test the response of the motor to load changes based on the
NN controller. The test case scenario is as follows: starting the induction motor with the rated condition
(rated speed of 1467 rpm at rated torque of 9 N.m) during the time interval from 0 to 1.5 sec, then making a
load change at 1.5 sec, and finally initiating the NN controller at 2.5 sec. Figure 6 shows the motor speed
response when the load torque is changed to 2.33 N.m. It is lower than the rated torque and the motor speed
is increased. When the NN controller is activated at 2.5 sec, new frequency and the modulation index are
applied to the PWM and the rated speed is recovered. Figure 7 shows the motor speed response when the
load torque is increased to 12.33 N.m, which is greater than the motor's rated torque, as well as how the NN
controller retrieves the motor's rated speed.

High-performance speed control for three-phase induction motor based on … (Mustafa A. Al-Khawaldeh)
6244  ISSN: 2088-8708

To further enhance the motor speed response during load changes and the NN controller actions, a
damper with a natural frequency of 5 Hz and a damping ratio of 0.55 is proposed. The motor speed response
for a load torque of 17 N.m, with and without a damper, is shown in Figures 8 and 9, respectively. The results
show an improved speed response of a very small overshoot and few oscillations.
The results in Figures 6 to 9 further demonstrate that the NN controller, based on oscillation and low
steady-state error, provides high performance for the three-phase induction motor speed control when
compared to the methods in the literature. Also, the results show the effectiveness of the NN controller,
which is trained using RDA equations, as opposed to the majority of approaches described in the literature
[30]–[32] which rely on heuristic algorithms for controller training.

Table 3. Comparison between NN controller and RDA controller for induction motor speed control
Case Load torque change and
Control action using Control action using
No rotor speed before
NN controller RDA controller
correction
TL Rotor speed New stator New supply Rotor speed New stator New supply Rotor speed
(Nm) (pm) Freq. (Hz) voltage (V) (rpm) freq. (Hz) voltage (V) (rpm)
1 0.495 1497 48.9800 441.2000 1466 48.9988 440.9889 1467
2 2.330 1491 49.1900 442.8000 1466 49.1944 442.7493 1467
3 3.165 1488 49.2700 443.7000 1466 49.2856 443.5702 1467
4 4.495 1484 49.4300 445.0000 1467 49.4344 444.9095 1467
5 6.495 1477 49.6800 447.1000 1467 49.6683 447.0145 1467
6 7 1475 49.7400 447.6000 1467 49.7297 447.5670 1467
7 12.330 1454 50.4700 452.6000 1467 50.4642 454.1776 1467
8 14 1445 50.8400 452.8000 1468 50.7442 453.1457 1467
9 16 1434 51.1300 456.5000 1467 51.1347 456.6325 1467
10 17 1428 51.2100 458.4000 1466 51.3628 458.6699 1467

Figure 6. Motor speed response to a load torque of Figure 7. Motor speed response to a load torque of
2.33 N.m, employing the NN controller 12.33 N.m, employing the NN controller

Figure 8. Motor speed response to a load torque of Figure 9. Motor speed response to a load torque of
17 N.m, employing the NN controller without a 17 N.m, employing the NN controller with a damper
damper

Int J Elec & Comp Eng, Vol. 14, No. 6, December 2024: 6237-6247
Int J Elec & Comp Eng ISSN: 2088-8708  6245

4.2. Laboratory model of the three-phase induction motor

This section demonstrates the experiments conducted. A three-phase induction motor laboratory
model was built to investigate and verify the suggested RDA-NN technique's ability to control the motor
speed at varying load torques. The laboratory setup is shown in Figure 10 and the ratings of the squirrel
cage three-phase induction motor are; power: 300 W, voltage: 220/380 V, D/Y current: 1,38/0.8 A, Δ/Y
frequency: 50 Hz 𝐶𝑜𝑠𝜑: 0.74, and speed: 2.800 rpm. For this laboratory test, the desired speed is fixed at
2,400 rpm. The oscilloscope measures for the motor speed under varying load torque are displayed in
Figure 11. The load torque varies between 0 and 1.1 N.m., and the speed remains constant at the desired
value of 2,400 rpm. Results have shown that when a varying torque is applied to the motor, the RDA-NN
controller can keep the motor running at the desired speed.

Figure 10. The three-phase induction motor laboratory setup

Figure 11. Motor speed response to a varying load torque

5. CONCLUSION
In this work, two controllers are proposed to determine the required supply frequency for the three-
phase induction motor in order to maintain a constant speed at variable load torques. Using the RDA, a set of
equations is proposed to calculate the required supply frequency and voltage for a constant motor speed at
different load changes. Moreover, an NN-based real-time controller is employed in online mode to control
the motor speed. The NN is trained using the RDA data set. The training, validation, and test stages show the
best performance for the NN based on the regression factor and mean squared error compared to the
previously reported control approaches. An acceptable range of supply frequency is determined by RDA and
NN controllers and applied to the three-phase induction motor to control its speed for a wide range of load
High-performance speed control for three-phase induction motor based on … (Mustafa A. Al-Khawaldeh)
6246  ISSN: 2088-8708

torque changes. In addition, simulation and laboratory models are built for the three-phase induction motor to
verify the effectiveness of the proposed RDA-NN controller.

ACKNOWLEDGEMENTS
The authors express their gratitude to Philadelphia University, Jordan for providing support.

REFERENCES
[1] G. Rajamany, S. Srinivasan, K. Rajamany, and R. K. Natarajan, “Induction motor stator interturn short circuit fault detection in
accordance with line current sequence components using artificial neural network,” Journal of Electrical and Computer
Engineering, vol. 2019, pp. 1–11, Dec. 2019, doi: 10.1155/2019/4825787.
[2] N. Farah et al., “Analysis and investigation of different advanced control strategies for high-performance induction motor drives,”
TELKOMNIKA (Telecommunication Computing Electronics and Control), vol. 18, no. 6, pp. 3303–3314, Dec. 2020, doi:
10.12928/telkomnika.v18i6.15342.
[3] M. Mobarra, B. Tremblay, M. Rezkallah, and A. Ilinca, “Advanced control of a compensator motor driving a variable speed diesel
generator with rotating stator,” Energies, vol. 13, no. 9, May 2020, doi: 10.3390/en13092224.
[4] M. Dems, K. Komeza, J. Szulakowski, and W. Kubiak, “Increase the efficiency of an induction motor feed from inverter for low
frequencies by combining design and control improvements,” Energies, vol. 15, no. 2, p. 530, Jan. 2022, doi:
10.3390/en15020530.
[5] M. A. Awdaa, A. A. Obed, and S. J. Yaqoob, “A comparative study between V/F and IFOC control for three-phase induction
motor drives,” IOP Conference Series: Materials Science and Engineering, vol. 1105, no. 1, Jun. 2021, doi: 10.1088/1757-
899x/1105/1/012006.
[6] P. V. Patil and S. A. Naveed, “Implementation of VFD application for speed control of induction motor,” 2020 International
Conference on Smart Innovations in Design, Environment, Management, Planning and Computing (ICSIDEMPC), Aurangabad,
India, 2020, pp. 168-170, doi: 10.1109/icsidempc49020.2020.9299636.
[7] P. V Hunagund, “Three phase induction motor drive using IGBTs and constant v/f method,” International Journal of Advanced
Research in Electrical, Electronics and Instrumentation Engineering, vol. 1, no. 5, pp. 463–469, 2012.
[8] S. Mishra, A. B. Palazzolo, X. Han, Y. Li, and C. Kulhanek, “Torsional vibrations in open loop volts hertz variable frequency
drive induction motor driven mechanical systems,” 2020 IEEE Texas Power and Energy Conference (TPEC), College Station,
TX, USA, 2020, pp. 1-6, doi: 10.1109/tpec48276.2020.9042586.
[9] A. Saghafinia, H. W. Ping, and M. A. Rahman, “High performance induction motor drive using hybrid fuzzy-PI and PI
controllers: A review,” International Review of Electrical Engineering, vol. 5, no. 5, pp. 2000–2011, 2010.
[10] Ö. Otkun, F. Demir, and S. Otkun, “Scalar speed control of induction motor with curve-fitting method,” Automatika, vol. 63,
no. 4, pp. 618–626, Apr. 2022, doi: 10.1080/00051144.2022.2060657.
[11] R. Sadhwani and R. K., “A comparative study of speed control methods for induction motor fed by three level inverter,” 2016
IEEE 1st International Conference on Power Electronics, Intelligent Control and Energy Systems (ICPEICES), Delhi, India,
2016, pp. 1-6, doi: 10.1109/icpeices.2016.7853158.
[12] K. H. Khudier, K. G. Mohammed, and M. S. Ibrahim, “Design and implementation of constant speed control system for the
induction motors using programmable logic controller (PLC) and variable frequency derive (VFD),” IOP Conference Series:
Materials Science and Engineering, vol. 1076, no. 1, p. 12007, Feb. 2021, doi: 10.1088/1757-899x/1076/1/012007.
[13] Y.-S. Lim, J.-S. Lee, and K.-B. Lee, “Advanced speed control for a five-leg inverter driving a dual-induction motor system,”
IEEE Transactions on Industrial Electronics, vol. 66, no. 1, pp. 707–716, Jan. 2019, doi: 10.1109/tie.2018.2831172.
[14] A. K. Peter, J. Mathew, and K. Gopakumar, “A simplified DTC-SVPWM scheme for induction motor drives using a single PI
controller,” IEEE Transactions on Power Electronics, vol. 38, no. 1, pp. 750–761, Jan. 2023, doi: 10.1109/tpel.2022.3197362.
[15] E. B. Agamloh, S. Peele, and J. Grappe, “Operation of variable frequency drive motor systems with source voltage unbalance,”
Jun. 2017, doi: 10.1109/ppic.2017.8003869.
[16] R. Azizipanah-Abarghooee and M. Malekpour, “Smart induction motor variable frequency drives for primary frequency
regulation,” IEEE Transactions on Energy Conversion, vol. 35, no. 1, pp. 1–10, Mar. 2020, doi: 10.1109/tec.2019.2952318.
[17] A. Z. A. Firdaus, “Design and simulation of fuzzy logic controller for variable frequency drive of three phase AC induction
motor,” Journal of Advanced Research in Dynamical and Control Systems, vol. 11, no. 12-SPECIAL ISSUE, pp. 834–842, Dec.
2019, doi: 10.5373/jardcs/v11sp12/20193283.
[18] A. Zemmit, S. Messalti, and A. Harrag, “A new improved DTC of doubly fed induction machine using GA-based PI controller,”
Ain Shams Engineering Journal, vol. 9, no. 4, pp. 1877–1885, Dec. 2018, doi: 10.1016/j.asej.2016.10.011.
[19] V. R. Metha and S. S. Karvekar, “Speed control of induction motor using a fuzzy logic controller and direct torque controller,”
2018 4th International Conference for Convergence in Technology (I2CT), Mangalore, India, 2018, pp. 1-5, doi:
10.1109/i2ct42659.2018.9058085.
[20] B. Habib, “Fuzzy speed controller of induction machine with 36 sectors DTC based ANFIS hysteresis comparator,” Majlesi
Journal of Mechatronic Systems, vol. 8, no. 3, pp. 1–8, 2019.
[21] A. Abdel Menaem, M. Elgamal, A.-H. Abdel-Aty, E. E. Mahmoud, Z. Chen, and M. A. Hassan, “A proposed ANN-based
acceleration control scheme for soft starting induction motor,” IEEE Access, vol. 9, pp. 4253–4265, 2021, doi:
10.1109/access.2020.3046848.
[22] J. Tamimi, “Simulation of three-phase induction motor using nonlinear model predictive control technique,” Cogent Engineering,
vol. 5, no. 1, Jan. 2018, doi: 10.1080/23311916.2018.1516489.
[23] X. Liu and G. Bai, “Research on intelligent algorithm of the AC motor speed regulation system based on the neural network,”
Mobile Information Systems, vol. 2022, pp. 1–8, Jun. 2022, doi: 10.1155/2022/1590779.
[24] D. Izci, S. Ekinci, C. Budak, and V. Gider, “PID controller design for DFIG-based wind turbine via reptile search algorithm,”
2022 Global Energy Conference (GEC), Batman, Turkey, 2022, pp. 154-158, doi: 10.1109/gec55014.2022.9986617.
[25] S. Ekinci, D. Izci, and M. Kayri, “An effective controller design approach for magnetic levitation system using novel improved
manta ray foraging optimization,” Arabian Journal for Science and Engineering, vol. 47, no. 8, pp. 9673–9694, Nov. 2021, doi:
10.1007/s13369-021-06321-z.

Int J Elec & Comp Eng, Vol. 14, No. 6, December 2024: 6237-6247
Int J Elec & Comp Eng ISSN: 2088-8708  6247

[26] A. Shaltout and O. E. M. Youssef, “Speed control of induction motors using proposed closed loop Volts/hertz control scheme,”
2017 Nineteenth International Middle East Power Systems Conference (MEPCON), Cairo, Egypt, 2017, pp. 533-537, doi:
10.1109/mepcon.2017.8301232.
[27] S. J. Chapman, Electric machinery fundamentals, 5 Ed. The McGraw-Hill Companies, 2012.
[28] K. L. Shi, T. F. Chan, Y. K. Wong, and S. L. Ho, “Modelling and simulation of the three-phase induction motor using Simulink,”
The International Journal of Electrical Engineering & Education, vol. 36, no. 2, pp. 163–172, Apr. 1999, doi:
10.7227/ijeee.36.2.6.
[29] I. Boldea, “Induction machines an introduction,” Induction Machines Handbook, pp. 1–14, 2020, doi: 10.1201/9781003033417-1.
[30] B. L. G. Costa, C. L. Graciola, B. A. Angélico, A. Goedtel, and M. F. Castoldi, “Metaheuristics optimization applied to PI
controllers tuning of a DTC-SVM drive for three-phase induction motors,” Applied Soft Computing, vol. 62, pp. 776–788, Jan.
2018, doi: 10.1016/j.asoc.2017.09.007.
[31] M. A. Niazi, Q. Hayat, B. Khan, and M. Afaq, “Speed control of three phase induction motor using variable frequency derive
control system,” International Journal of Current Engineering and Technology, vol. 10, no. 1, pp. 5–10, Oct. 2021, doi:
10.14741/ijcet/v.10.1.2.
[32] S. V John, K. S. R. Rao, and R. S. Rao, “Analysis of VFD induction motor current disturbances by Prony analysis and motor
redesign by intelligent optimization techniques,” Jan. 2021, doi: 10.1109/icpee50452.2021.9358625.

BIOGRAPHIES OF AUTHORS

Mustafa A. Al-Khawaldeh obtained his B.Sc. degree in electro-mechanical

engineering in 1996 and an MSc degree in mechatronics engineering in 2005 from the
Al-Balqa’ Applied University in Jordan. He completed his PhD in mechatronics engineering at
De Montfort University (United Kingdom) in 2014. He was the Chairman of the Department of
Mechatronics Engineering at Philadelphia University in Jordan from 2020-2022. His research
interests include modeling and simulation of mechatronics systems, control, fuzzy logic, neural
networks, smart homes, ubiquitous robotics, and the internet of things. He participated in different
related conferences and has many publications in these fields. He currently works at the
Department of Mechatronics Engineering, Faculty of Engineering and Technology, Philadelphia
University, Amman, Jordan. He can be contacted at email: malkhawaldeh@philadelphia.edu.jo.

Jasim A. Ghaeb is a professor of power systems and he is currently with the

Department of Alternative Energy Technology, Faculty of Engineering and Technology,
Philadelphia University, Jordan. He received a Ph.D. degree from the University of Bradford,
the UK, in 1989 in electrical power systems-control and power electronics. He has 30 years of
experience in teaching, research, and administration. Research areas include reactive power
control, artificial intelligence, heuristic algorithms-based power system control, embedded
systems, and data integrity. He can be contacted at email: jghaeb@philadelphia.edu.jo.

Samer Z. Salah holds an M.Sc. degree in electrical power and control engineering
from Tafila Technical University, Jordan, conferred in 2021, and a B.Sc. degree in
mechatronics engineering from the Hashemite University Jordan, completed in 2008. He has
accumulated extensive experience, having worked at Philadelphia University for 14 years as a
laboratory supervisor. During his tenure, he was responsible for teaching various mechatronics
labs including control, automation, electrical machines, power electronics, and drives.
Currently, he serves as a lecturer at the Department of Mechatronics Engineering within the
School of Engineering at The University of Jordan, Amman, Jordan. His research interests
encompass smart grids, renewable energy systems, power system stability, artificial
intelligence, and machine learning. For further inquiries, he can be reached via email at email:
samer.salah@ju.edu.jo.

Mohammad S. Alrawajfeh received a master’s degree in mechatronics

engineering from Philadelphia University 2022 (Amman-Jordan), a bachelor’s degree in
mechatronics engineering from Al-Balqa Applied University 2019 (Amman-Jordan). He has
practical experience as an operation and maintenance engineer for three years in the industrial
field and accurate experience in the field of modification and development in inverter
applications. He is currently working in Al-Warkaa Food Industries Company. He can be
contacted at mohu200@gmail.com.

High-performance speed control for three-phase induction motor based on … (Mustafa A. Al-Khawaldeh)