E3S Web of Conferences 387, 01001 (2023) https://doi.org/10.

1051/e3sconf/202338701001
ICSERET-2023

Performance analysis of controllers in BLDC

motor
S.Muralidharan1, Balasubramaniyam.T2, Prethesh Kumar.K.S2, and Aswin.C2
1
Professor, EEE Department, Mepco Schlenk Engineering College, India
2
UG students, EEE Department, Mepco Schlenk Engineering College, India

Abstract. Due to their better electrical properties and mechanical

features, the usage of BLDC motors is increasing and they have replaced
conventional DC and asynchronous motors in many applications. This
project provides a sensor-based permanent magnet brushless DC motor for
commercial and electric vehicles. To improve the reliability of BLDC
motor speed regulation, a PID controller is employed. The three-phase
voltage inverter delivers a constant DC voltage and is synced with the
BLDC motor. Closed-loop speed control is done utilizing the Ziegler-
Nichols technique, Particle Swarm optimization and PID controller
improved using Firefly algorithm. The recommended system is simulated
using MATLAB simulation and the performance of the BLDC motor is
tested by comparing attributes such as Rise time, Settling time and Peak
overshoot.

1 Introduction
For BLDC, the rotor's DC field winding is replaced by a permanent magnet to produce
variable airflow. Thanks to the magnet of the rotor, the electrical loss in the winding of the
motor is reduced and the thermal performance and performance of the PM motor are
improved. Without mechanical components such as brushes or slip rings, the motor is
lighter and has a higher power-to-weight ratio, increasing efficiency and reliability.
Permanent magnets have disadvantages such as the manufacturing process, high
temperature demagnetization of the magnet, and expensive magnet material. PMDC motors
and PMAC motors are two types of PM motors. Permanent magnets are used to switch
field windings in PMDC motors, similar to a DC commutator. PMACs do not use slip rings,
brushes and commutators as Permanent magnets installed on the rotor create the magnetic
field. Therefore, PMAC is easier to use than BLDC. According to the type of back
electromotive force (EMF), PMACs are divided into two types: sinusoidal and trapezoidal.
Surface-mounted PMSMs and internal PMSMs are additional categories PMAC machines
of the sinusoidal kind. The brushless DC motors (BLDC), often known as trapezoidal
BLDC machines, It generates trapezoidal back EMF waves with the following properties:
• Magnet flux distribution in the air gap is rectangular.
• Concentrated stator windings and a rectangular current waveform

*Corresponding author:[email protected]

© The Authors, published by EDP Sciences. This is an open access article distributed under the terms of the Creative Commons Attribution
License 4.0 (https://creativecommons.org/licenses/by/4.0/).
E3S Web of Conferences 387, 01001 (2023) https://doi.org/10.1051/e3sconf/202338701001
ICSERET-2023

1.1 Literature Survey

Speed control of a BLDC motor is important to keep the motor running at the desired
speed. The speed of the BLDC motor can be controlled by controlling the input DC voltage
and current. The higher the voltage, the faster it will be. The speed control can be achieved
by open loop and closed loop control. Open loop control are not much preferred due to their
poor efficiency, whereas closed loop control provide smooth operation and higher
efficiency than open loop control. Closed loop control provides a feedback path, and a
controller is placed in that path. The paper [1] proposed the speed control of BLDC motors
by PWM technique using FPGA; studies on microcontrollers and DSP controllers are done
to control the speed of the motor [2, 3]. There are various studies and researches done to
control the speed of BLDC motors using PI and PID controllers [4, 5],The majority of PI
controllers used in industry today are traditional ones. Because of the simplicity and
convenience of online tweaking, traditional PID controllers are also employed in industries.
To provide optimized output, tuning must be done by various algorithms. The paper [6, 7]
demonstrates the tuning of a controller by the Ziegler-Nichols method, Ziegler-Nichols (Z-
N) method is a common tuning method for PID controllers [8, 9]. Using this strategy, the
integral and derivative gains are set to zero before raising the proportional gain until the
system becomes unstable, due to the high time consumption nature Z-N methods are not
used now a days; instead, fuzzy logic controllers are being designed and tuned for PI,PID
speed control applications [10, 11]. Firefly algorithms are commonly used algorithms
nowadays. The various studies and researches [12, 13] describe the effectiveness of the
firefly algorithm. To enhance the effectiveness and efficiency of the firefly algorithm, the
continuous firefly algorithm and the Continuous firefly algorithm are used [14, 15]. The
above extract explains the performance of BLDC motors with various algorithms.

1.2 Brushless DC motor

Fig 1. BLDC motor's cross-sectional view

A BLDC motor driver uses a generator to convert alternating current to power the
permanent magnet synchronous motor. It does not seem possible to develop controllers for
these machines using industrial simulation tools, possibly due to the high cost of software
development. In low-cost applications for example, Numerical Control machine tools and
robot drivers, the rotor magnet can be demagnetized during commissioning or modification.
However, special drive topologies for high-end applications may require reconfiguration of
motors and inverters, resulting in higher drive costs. The cost of using new magnetic
materials is superior . The invention has the potential to increase the market for BLDC

2
E3S Web of Conferences 387, 01001 (2023) https://doi.org/10.1051/e3sconf/202338701001
ICSERET-2023

motors to tens of kilowatt applications where debugging mistakes are costly. Modelling is
essential for achieving cost reductions.

1.3 BLDC drives operation with inverter

As it is an essential part of the generator, the front end should have a three-phase inverter as
shown in Fig 2. The inverter acts as an electric commutator in autonomous mode and

Fig 2. BLDC Motor drive system with Hall sensor

receives change logic pulses from the correct position sensors.

The inverter may essentially function in two modes.
• 120-degree switch-on mode
• 180-degree switch-on mode
• PWM voltage and current control mode
Table 1. Motor ratings
Parameters Symbol Ratings
Torque constant Kt 0.06675
Back EMF Kb 48
Moment of Inertia J 0.142e-3
Friction coefficient B 0.04
Electrical Inductance L 0.389e-3
Electrical Resistance R 0.323
Speed RPM 3000

1.4 Transfer function of motor

A common sort of mathematical model is the transfer function. First-order and second-order
transfer functions with voltage input and motor speed output are often used in BLDC motor
models. Given that there is uncertainty and nonlinear processes in the real-time
environment, experiments have been both simulated and carried out in order to compare
their results.
The BLDC motor’s Transfer function is given by:

3
E3S Web of Conferences 387, 01001 (2023) https://doi.org/10.1051/e3sconf/202338701001
ICSERET-2023

(1)

Where,

• s: Laplace variable
• Kt: Torque constant
• Kb: Back EMF
• J: Moment of Inertia
• b: Friction Coefficient
• L: Electrical Inductance
• R: Electrical Resistance
From this the BLDC motor’s Transfer function is given by:

(2)

1.5 Without controller

A controller is necessary to manage the speed and torque of a BLDC (Brushless DC) motor.
This is as a result of the motor's intricate electronic commutation system, which regulates
the current flow to the motor coils in accordance with the rotor position. The motor cannot
operate effectively or smoothly without a controller.
Otherwise, a PWM circuit may be used to regulate a BLDC motor's speed even without
a controller. A PWM circuit is an electrical device that alternately turns on and off the
voltage at a high frequency, usually between tens and hundreds of kilohertz. You may
modify the PWM signal's duty cycle to alter the average voltage delivered to the motor.
However, it is very difficult to reach the desired speed.

1.6 PID controller

This PID speed control system is used because of its advantages such as 0% steady state
error, no oscillation, better reaction, and greater stability. This controller is used to reduce
oscillation and overshoot in system output. The PID controller's output is the total of the
proportional, integral, and derivative gains.
Mathematically:

• Kp: Proportional gain constant

• Ki: Integral gain constant
• Kd: Derivative gain constant
The PID (Proportional-Integral-Derivative) controller calculates the system output using
three constants: Kp, Ki, and Kd .The disparity between the predicted and actual values used
to determine the output is referred to as signal error.

2 Algorithms
Algorithms automate and enhance the tuning process, they are essential for tuning PID
controllers for BLDC motors. Both model-based and model-free techniques are available.
Model-based algorithms employ the system's mathematical models to identify the best PID

4
E3S Web of Conferences 387, 01001 (2023) https://doi.org/10.1051/e3sconf/202338701001
ICSERET-2023

gains, whereas model-free algorithms change the gains depending on the system's reaction
through trial and error.

2.1 Ziegler - Nichols method

The Ziegler-Nichols technique is a popular PID controller tuning method. The integral and
derivative gains are set to zero in this technique before increasing the proportional gain
until the system becomes unstable.

To implement this method in MATLAB by the following steps:

Step 1: Define your system and obtain its transfer function. For instance, if you have a
system that has a transfer function,

(4)
Step 2: To determine the Ultimate gain and period, use the system's step response. To do
this, you can use the ‘stepinfo’ function in MATLAB, It returns a structure containing
different information on a system's step response.
Step 3: Calculate the PID parameters using the Ziegler-Nichols tuning method. There are
two variants of the Ziegler-Nichols method: the classic method and the modified method.
Below are the formulas for both methods:
Traditional Method:

Modified Method:

Here,
 Ku: Constant of Ultimate Gain
 Tu: Constant of Ultimate Time
Step 4: Display the results

2.2 Particle swarm optimization

Particle swarm optimization, or PSO, is a well-liked optimization method that is widely

applied in many disciplines, including physics, computer science, engineering, and
economics. The PSO algorithm takes its cues from social behaviours like fish scooling or
bird flocking, where each individual moves in accordance with its neighbours to discover
the best solution to a particular issue.
Procedure for PSO:
Step 1: The purpose of PSO is to define the problem that must be addressed. This includes
identifying the target function to be optimized as well as any search space constraints.
Step 2: In PSO, start the swarm, and a swarm of particles will explore the search space.
Each particle considered will be responsible for the solution. By dispersing particles at
random over the search space, the swarm is created.
Step 3: Assess fitness: Using the objective function, the fitness of each particle is assessed.
In terms of locating the ideal answer, this establishes how effective or ineffective the
particle is.

5
E3S Web of Conferences 387, 01001 (2023) https://doi.org/10.1051/e3sconf/202338701001
ICSERET-2023

Step 4: Personal best is an updated record of each particle's best location (or solution) to
date. The personal best is updated if the fitness of the current position is higher than the
fitness of the previous position.
Step 5: Update global best: Each particle also maintains track of the best location (or
solution) discovered by the whole swarm, known as the global best, in addition to their own
personal best. If the fitness of the current position is greater than the fitness of the global
best, the global best is updated.
Step 6: Each particle's velocity is updated in accordance with its current location, personal
best, and world best. The velocity of the particle controls its migration over the search
space, as well as its direction and speed.
Step 7: location update: Taking into account the new velocity, each particle's location is
revised. This establishes the particle's new position within the search space.
Step 8: Repeat: The steps from 3 to 7 are repeated a predetermined number of times. (or
until convergence is reached). Until the ideal solution is identified, the swarm advances
around the search area while updating individual and collective best locations and
velocities.
Step 9: Return of the solution: After the PSO algorithm converges, the best location
obtained globally by the swarm is the ideal solution.
Table 2. PSO initial parameters
Number of Particles 20
Maximum Generation 50
Inertia Weight 0.7
Cognitive Weight 1.5
Social Weight 1.5
Number of Iterations 1000

2.3 Firefly algorithm

The flashing activity of fireflies served as the inspiration for the development of the Firefly
algorithm, a unique metaheuristic technique for optimization problems Randomly produced
solutions are represented as fireflies in this manner, and their brightness is governed by how
well they perform on the objective function. Yang devised this swarm-based metaheuristic
algorithm that mimics how fireflies communicate by flashing their lights.
Procedure for Firefly algorithm:
Step 1: Set the algorithm parameters, such as the number of fireflies (n), the maximum
number of generations (MaxGen), and the attraction coefficient ().
Step 2: Define the goal function f(x), where x represents a d-dimensional vector (x1,...,
xd)T.
Step 3: Create the initial firefly population, X = [x1, x2,..., xn].
Step 4: Using f(xi), calculate the light intensity of each firefly i at xi.
Step 5:While the halting requirement (e.g., t MaxGen) is not met:
For each firefly i ranging from 1 to n:
For each individual firefly j = 1 to n:
If the light intensity of firefly j is greater than that of firefly i, move firefly i
closer to firefly j.
j using the following formula:
xi = xi + * exp(-r2) * (xj - xi) + * (rand - 0.5)

6
E3S Web of Conferences 387, 01001 (2023) https://doi.org/10.1051/e3sconf/202338701001
ICSERET-2023

Attractiveness varies with r through Exp[-r2].

Examine new ideas and adjust light intensity
end of j
end of i
Step 6: Sort the firefly by light intensity to discover the best answer right now.
End while
Step 7: Visualisation of post-process outcomes
Step 8: Conclude the operation.

Table 3. Firefly algorithm initial parameters

Number of Fireflies 20
Maximum Generation 50
Alpha 0.5
Gamma 0.2
Number of Iterations 1000

+2.4 Objective function

The objective function is a real-valued function used to either minimize or maximize its
value while adhering to the specified constraints. It is a crucial component in solving
optimization problems.
ISE stands for Integral Square Error, which is a common objective function used in
control systems engineering. The ISE objective function is used to measure control system
performance by evaluating the error between the system's actual output and the intended
setpoint over time.
The mathematical equation for ISE can be expressed as:

(6)

Where,

• e(t) : error between the setpoint and the actual output at time t
• T : final time of the interval.

3 Simulation and results

The Mathematical mode BLDC motor with PID was implemented in MATLAB. In
MATLAB the simulation was done by Simulink

7
E3S Web of Conferences 387, 01001 (2023) https://doi.org/10.1051/e3sconf/202338701001
ICSERET-2023

model.

Fig 3. Simulation model in MATLAB

The following PID parameters are generated by Ziegler-Nichols method, Particle

Swarm Optimization and Firefly Algorithm.
Table 4. PID parameters

Controller PID with ZN PID with PSO PID with FA

0.008898 0.9048 1.3039

0.097949 1.4775 2.9315

7.2749e-5 0.0750 0.0790

The generated PID parameters ( , and ) are substituted into the PID controller,
which is available in the Simulink model. Then, the Simulink model is simulated for 10
seconds to obtain the speed waveform using the scope.
By using the To Workspace block in Simulink, the waveform characteristics are saved to
the workspace of MATLAB. Using the 'stepinfo' command, the parameters of that speed
waveform can be determined.
The control parameters are compared and tabulated below:
Table 5. Waveform parameters
Controller Parameters

Without Controller Rise Time 0.0554

Settling Time 9.9942
Peak Overshoot 6.5394
PID with ZN Rise Time 0.3205
Settling Time 0.5760
Peak Overshoot 0.7631
PID with PSO Rise Time 0.1375
Settling Time 0.7234
Peak Overshoot 0.3388

8
E3S Web of Conferences 387, 01001 (2023) https://doi.org/10.1051/e3sconf/202338701001
ICSERET-2023

PID with FA Rise Time 0.0975

Settling Time 0.6836
Peak Overshoot 0.0975
The waveforms displayed below are obtained from simulation with PID controller tuned
with Ziegler-Nichols method, Particle Swarm Optimization and Firefly Algorithm. And the
speed waveform also obtained without PID controller.

Fig 4. Speed waveform of BLDC motor with PID controller tuned with FA, PSO, ZN and without
controller

Fig 5. Stator current and Back EMF waveform

This simulation was also conducted at a different set speed. In Fig. 6, the set speed is
changed from 2000 rpm to 3000 rpm. And for this waveform, the control parameters are
calculated and tabulated in Table 6.

Fig 6. Speed waveform with variable set speed

For this speed waveform with a variable set speed, the parameters like rise time, settling
time, and peak overshoot are tabulated below.
Table 6. Different set speed waveform parameters

9
E3S Web of Conferences 387, 01001 (2023) https://doi.org/10.1051/e3sconf/202338701001
ICSERET-2023

Controller Parameters
PID with ZN Rise Time 0.8382
Settling Time 0.1661
Peak Overshoot 1.3265
PID with PSO Rise Time 0.0381
Settling Time 0.1540
Peak Overshoot 0.0238
PID with FA Rise Time 0.0288
Settling Time 0.1147
Peak Overshoot 0.0258
The speed waveform was also monitored for different loaded conditions. The obtained speed
waveform as shown in Fig 7.

Fig 7. Speed waveform different loaded condition

5 Conclusion
This study aimed to develop and construct a PID controller in MATLAB utilizing the
Ziegler-Nichols (ZN), Particle Swarm Optimization (PSO), and Firefly Algorithm (FA)
tuning techniques for BLDC motor speed control. The primary goal of this initiative was to
assess the performance of PID controllers adjusted using various optimization strategies
and to choose the most effective one for BLDC motor speed control.
According to the simulation results, all three tuning techniques were successful in rising
the PID controller's performance over that of the original guess tuning. But in terms of
rising time, settling time, and overshoot, the performance of the controller tuned by the FA
approach was discovered to be superior to that of the other two methods. In comparison to
the ZN and PSO-tuned controllers, the FA-tuned PID controller had a quicker rising time, a
lower settling time, and less overshoot.
Overall, the study was successful in demonstrating the PID controller's usefulness for
controlling the speed of BLDC motors when used with various tuning techniques. Electric
cars, drones, and robots are just a few examples of applications that might benefit from
using the suggested ways to precisely regulate the speed of BLDC motors. The FA-tuned
PID controller may be implemented in a physical system, and its performance can be
assessed in real-world circumstances as part of future development.

10
E3S Web of Conferences 387, 01001 (2023) https://doi.org/10.1051/e3sconf/202338701001
ICSERET-2023

References

1. R. M. Pindoriya , S. Rajendran , P. J. Chauhan International J of Advance Eng and

Research Development (IJAERD), 1-6, (2014).
2. Juhi Nishat Ansari , Sapna L International J of Eng Research & Technology (IJERT), 3
,1666-1671, (2014).
3. G.MadhusudhanaRao, B.V.SankkerRam, B.Smapath Kumar, K.Vijay Kumar,
International J of Eng Science and Technology,2,143-147 (2010).

4. Y.Narendra Kumar, P.Eswara Rao, P. Vijay Varma, V. V. Ram Vikas, P. Kasi Naidu, Int.
J of Eng Research and Applications ,4, (2014).
5. C K Karthika and Peter Abraham , AIP Conf Proceedings,1-7,(2020).
6. Vishakha Vijay Patel, Resonance,25, 1385-1397, (2020).
7. Mohd Sazli Saad, Hishamuddin Jamaluddin and Intan Zaurah Mat Darus, ICIC
International, 8,7761-7779, (2012).
8. Mohammed Abdelbar Shamseldin, Adel A. EL-Samahy 15th International Workshop on
Research and Education in Mechatronics (REM),1-6, 9-11 September 2014 , Elgouna,
Egypt (2014).
9. Zemmour N, Ishak Boushaki S, Bendjeghaba O,Firefly 4th International Conf on Power
Eng, Energy and Electrical Drives,1293-1296,13-17 , (2013).
10. Neethu U. and Jisha V. R. 2012 IEEE International Conference on Power Electronics,
Drives and Energy Systems ,1-5,16-19 , (2012).
11. Christina Mathew, Jebin Francis, Prathibha PK, IJSRD - International Journal for
Scientific Research & Development, 5,636-639 (2017).
12. D. Kumanan & B. Nagaraj, An Open Access Journal,52-56 (2013).
13. Pikaso Pal , Rajeeb Dey , Raj Kumar Biswas , Shubhashish Bhakta, International
Journal of Soft Computing, Mathematics and Control (IJSCMC), 4,39-48, (2015).
14. Arman Hadi Azaha , Adam Samsudin , Yep Kow Wai , Amar Faiz Zainal Abidin , Rozi
Rifin , Mohammad Haniff Harun , Mohd. Safirin Karis , Ezzatul Farhain Azmi , and Ili
Najaa Aimi Mohd. Nordin, ARPN J of Eng and Applied Sciences,16, 757-761, (2021).
15. Omar Bendjeghaba, J of Electrical Eng, 65, 44-49 (2014).

11