2019 IEEE PES/IAS PowerAfrica

Model-based Speed Control of a DC Motor Using a

Combined Control Scheme
Ihechiluru Okoro and Clinton Enwerem
Department of Electrical Engineering, University of Nigeria, Nsukka 410101, Nigeria
ihechiluru.okoro@unn.edu.ng and clintonenwerem@gmail.com

Abstract—The separately-excited DC motor is a high- drive the motor at desired speed in the presence of parameter
performance variable speed drive vital for industrial variations and changing operating conditions [1,5]. It is
applications such as robotics, actuation, control and guided desired that a speed control system for the DC motor should
manipulation because of its precision, simplicity, continuous satisfy the following performance criteria: stability of the
control feature and wide speed range. Hence there is need to closed loop system, robustness i.e. insensitivity to motor
accurately regulate and drive the motor at desired speed. The variations, and excellent reference speed tracking [6].
individual armature voltage and field current methods are
flawed in their inability to control the motor speed over a wider The PID (Proportional Integral and Derivative) controller,
range. Hence this paper is proposing a combined armature the conventional and popular feedback control scheme, used
voltage and field current control method using Internal Model in the industry for control gives a performance that is
Control (IMC) feedback control scheme that will ensure unacceptable because of its oscillatory response and
reference speed tracking, a fast and non-oscillatory response for sensitivity to motor parameter variations. It is also limited in
the DC motor speed. The armature voltage and field current its ability to deal with inverse-response and to provide
control techniques are both modelled, fully capturing the predictive control action to compensate for changing
dynamic effects of the armature and field excitation of the DC operating conditions. The PID controller is thus, not the
motor. The derived model is then used to obtain optimal settings optimal control strategy for variable speed drive applications
for a PID controller that will control the motor speed. This
[2,6,7]. This presents the need to design a robust and high-
model-based controller enhances the performance of the motor.
performance controller that will achieve the desired objectives
Computer simulations are presented to show the effectiveness of
the proposed control scheme.
for the DC motor speed.
Hence, the main contribution of this paper is to propose a
Keywords—DC motor control, Internal Model Control, Speed combined control scheme – that is, the armature voltage
control, Armature voltage control, Field current control control method and field current control method combined
together – that will effectively regulate the speed of a
I. INTRODUCTION separately-excited DC motor; which will be designed using
The DC motor is an example DC machine that converts internal model control techniques. Unlike the individual
electrical energy into mechanical form. It achieves this using control methods, in the combined control scheme the armature
direct current in both stator and rotor. It has widespread voltage, � and the field voltage, � are varied simultaneously
industrial application, supportive of sustainable development, in order to vary the speed. A model of this dynamic
such as in machine tools, robotic manipulators, household relationship is developed for both parts of the motor, stating
appliances, servomotors, paper mills, and propulsion systems desirable control objectives for the speed of the DC motor, and
for electric vehicles. This is due to its precision, simplicity, from these optimal settings for the PID control algorithm is
continuous control features and because its speed can be obtained. Various reference speeds – below the rated speed, at
controlled over a wide range with relative ease. As a the rated speed, and above the rated speed – will be utilised to
distinguishing feature of DC machines, the field circuit and test the performance of the combined armature voltage and
the armature circuit can be interconnected in several ways to field current control method on the DC motor. The results
deliver a wide range of performance characteristics. A variant obtained from the combined control method is compared with
of DC motors is the separately-excited motor where the field the individual control methods. This paper, therefore, has the
circuit and the armature circuit are fed with independent following objectives:
voltage sources. This type of high performance variable speed • To model the dynamic relationship between the
drives is vital for industrial applications such as robotics, armature voltage and field current with the DC motor speed.
actuation, control and guided manipulation where precision
movements are essential [1,2,3,4,5]. • To design and build an optimal IMC-tuned PID
(feedback) controller for both parts of the motor for improved
A major performance feature of the DC motor is the performance of the DC motor speed.
torque-speed characteristic, which is a description of how the
torque produced by the machine varies as a function of the • An evaluation of the performance of the designed
speed of rotation of the motor for steady speeds. Hence, a model-based control scheme and a comparison with the
servo drive system is rated ‘high performance’ when the shaft individual armature voltage control and field control methods.
speed can be made to follow a desired reference at all time. A
major control objective for the machine is to regulate and

978-1-7281-1010-3/19/$31.00 ©2019 IEEE

Authorized licensed use limited to: INDIAN INSTITUTE OF TECHNOLOGY DELHI. Downloaded on February 28,2023 at 17:42:57 UTC from IEEE Xplore. Restrictions apply.
 2019 IEEE PES/IAS PowerAfrica
II. BACKGROUND
Recently considerable effort has been devoted to DC
motor speed control using different approaches, like PID, Ra
fractional-order dynamics, sliding mode control (SMC), but
ia
no attempt has been made to use a combined model of the +
armature and field parameters of the DC motor to obtain La - Va
Field
Lf
optimal tuning settings for a PID controller for speed control. if

An approach based on linear parametric models, derived Armature

using state estimation and system identification techniques, is ωm
used to control the motor, but their displacement response had + -
Te
TL,Jd ω,Bω
oscillations and due to plant-model mismatch the controller Vf
performance is reduced [8]. An approach incorporated Load Inertia,J
Friction,B
fractional-order dynamics to improve the performance and
change the dynamics of the entire DC motor by eliminating
Fig. 1. Equivalent Circuit of the SEDCM.
the original influence of the conventional PID, without
making internal changes into the system. But indices like
settling time, rise time, steady-state error and overshoot were � =� t + + �
not use to evaluate its performance [9]. An effective and
profound approach based on SMC and Adaptive PID with �
SMC gives a good motor speed performance with acceptable � = = �� + +�
robustness. The only flaw with this technique is that the
performance characteristics however are not as optimized as a Where � = , Te = K�a is the development electrical
larger rise time is observed with the SMC method than with torque, � = �Eb = Kω, is the back EMF, the constant K
the conventional, and there is the chattering phenomenon K denotes the torque/back EMF constant, � B denotes the
where the control signal switches between two structures viscous friction coefficient/motor damping, J is the motor
leading to high frequency oscillations [2,10]. shaft inertia, � ω is the motor speed and � TL is the load
An approach using Universal Learning Networks (ULNs), torque: typically, � TL is either constant or some function of
a class of Neural Networks, for the identification and control the speed ω ω in a motor [1]. To derive the input-output
of a separately excited DC motor, loaded with a centrifugal relationship between the armature voltage � �a and the motor
pump, gave an overall satisfactory performance. Firstly, speed �ω, the Laplace transformation of (1) and (2) yields:
offline a Universal Learning Network Identifier (ULNI) is
trained to emulate the dynamic behaviour of the DC motor � =� + + �
system, using the forward propagation algorithm. This = �� + � +�
identifier is then used online to train a Universal Learning
Network Controller (ULNC) that regulates the motor speed to If current in the armature is made the subject from (4) and
track a desired reference signal. Performance indicators, like resubstituted in (3) we have the system transfer function that
settling time, rise time, steady-state error and overshoot were depicts the input-output relationship between the motor speed
not discussed here [11]. The above stated limitations with � and armature voltage � and load torque � :
literature will be handled by this paper.
This paper is organised as follows. First a mathematical
� = �
model of the separately-excited DC motor (SEDCM) is � + � � �+
derived in Section III. The proposed IMC-based PID control + +( )
technique; designed using a model of the motor is introduced +�
in Section IV. Section V presents the simulation results as well − �
� + � � �+
as an evaluation of the performance of the approach. Finally, + +( )
and in Section VI, major conclusions are drawn.
� = � � +� �
III. MODELING OF THE SEDCM Where � Gm is the transfer function that depicts the
The separately excited DC motor is a machine wherein the relationship between the motor speed � and the armature
field and armature circuit have independent voltage sources. voltage � , while � depicts the relationship between the
The equivalent circuit showing the electromechanical motor speed � and load torque, the disturbanceVa . Hence (5)
arrangement of this type of DC motor is shown in Fig. 1. is rewritten to be (assuming that the Va load torque applied to
the DC motor is negligible, � = ):
The motor speed consistent with the rated armature
voltage is its rated speed. Armature voltage control serves to
� = �
smoothly vary the motor speed from zero to its rated speed + � + � + � �+
(base speed) by varying the armature voltage in the constant
torque region. The armature current increases with increase in The expression in (7) can be represented with the block
the armature voltage, which consequently leads to an increase diagram shown in Fig. 2.
in the development torque and, hence the motor speed
[1,4,12]. The dynamics of the motor and its load are
represented by the following set of differential equations:

Authorized licensed use limited to: INDIAN INSTITUTE OF TECHNOLOGY DELHI. Downloaded on February 28,2023 at 17:42:57 UTC from IEEE Xplore. Restrictions apply.
 2019 IEEE PES/IAS PowerAfrica
1 technique, speed control below and above the rated speed is
Va(s) +
K 1
ωm(s)
sLa + Ra Js + B achieved, and hence the limitations and flaws associated with
- the individual speed control methods is addressed.
A. IMC-tuned Feedback Control
The IMC control scheme is a model-based control strategy
K
[13], that is based on an assumed system model which gives
rise to analytical expressions for the controller settings. The
Fig. 2. Block Diagram Configuration of armature voltage control benefit of this approach is that it permits model uncertainty
and trade-offs between performance and robustness to be
With the field control method, the field voltage is varied in the considered in a more systematic manner. Instead of setting a
constant power region while the armature voltage, and hence control structure and then trying to ‘extract’ optimality from
armature current, is constant [12]. The dynamics in the field this controller, as is the case with conventional PID control,
of the motor is represented by the following set of differential the approach of this work is to postulate a model for both the
equations:
field and armature parts of the DC motor with desirable
� � control objectives for the speed of the DC motor specified.
= −( ) − ( )� + ( ) Then, from these a suitable control structure and parameters
will be obtained in a straightforward manner [6,13].
� � The simplified control system configuration of the combined
= −
armature voltage and field current method using IMC
� � approach is shown in Fig. 4. �� is the IMC controller for
= − � − � the armature part, while �� is the IMC controller for the
field part.
Where = . Taking the Laplace transformation of (1) - SEDC Motor
(3) the following expression is obtained:
∗ ωref + GIMC1
Va
G1 ω
� = ∙ � − � − � Reference - Actual
+� � � � Speed
Armature
Speed
Vf
GIMC2 G2
∗
Where = (� ) . From (11), the block diagram
� Field

configuration of the relationship between the field voltage and

the motor speed is shown in Fig. 3 (assuming the Va load
torque, � = ): Fig. 4. Combined Control Structure using IMC approach

+ 1 1 The equivalent feedback control structure derived using the

Vf(s) K* ωm(s)
Lfs+ Rf B Internal Model Control algorithm is shown in Fig. 5. Here,
- the assumed models of the DC motor, controller output and
actual speed output are used to obtain optimal settings for the
PID controller, � .

Js(Lfs+ Rf) GC
Gm
K* ωref PID ω
+ Controller Motor
Fig. 3. Block Diagram Configuration of field current control model
-
IMC
The transfer function that defines the input-output Tuning
relationship between the field coltage, � and motor speed, � Algorithm
is given thus:

∗
� = Fig. 5. Equivalent Feedback Control Structure
+ (� + � ) + ��
The combined armature and field control technique is
Where � relates the motor speed to the field voltage, � . executed in the MATLAB/Simulink environment as shown
in Fig. 6. Here two distinct internal model controllers are
designed and connected to the armature and field circuitry
IV. PROPOSED APPROACH correspondingly. The output speed of the motor is measured
In the following subsections an IMC-based feedback and fed back to the input. The comparison between the
controller, designed using both the armature and field models reference speed and the actual speed of the DC motor
of the motor, is developed. The speed of the motor is generates an error signal which is utilized by the IMC-tuned
controlled by simultaneously varying the voltage applied to PID controllers for control.
the field winding and the armature. With this combined

Authorized licensed use limited to: INDIAN INSTITUTE OF TECHNOLOGY DELHI. Downloaded on February 28,2023 at 17:42:57 UTC from IEEE Xplore. Restrictions apply.
 2019 IEEE PES/IAS PowerAfrica
Figs. 3 and 4, it can be seen from block diagram
transformation that the feedback and IMC controllers are
equivalent if � Gc and �� satisfy the relation:
��
� =
− �� �
Hence the equivalent feedback controller in Fig. 4 is a
function of the DC motor model and IMC tuning algorithm.
The derived IMC-tuned PID controller is in the parallel form
and is expressed thus:
Fig. 6. Simulink Model for combined control of the SEDCM
� �
�� = � + +
The proposed model-based control scheme is designed in two
steps: where � k p is the proportional gain,k i � is the integral gain
Step 1. The DC motor model is factored as while is the derivative gain.
� = � +� −
V. SIMULATION RESULTS AND DISCUSSION
Where � − G ̃m- is the invertible part and � + G ̃m+ is the
non-invertible part containing any time delays and right-half To illustrate the performance of the proposed feedforward –
plane zeros. Since there are no time delays or right-half plane feedback controller, a separately excited DC motor is chosen
zeros, the approximated plant model is the invertible part. with the following parameters:
Step 2. The IMC controller is designed thus = 16.2mH, � = 2 Ω, � = 220 V, = 71.47 H, � = 180
Ω, � = 220 V, = 1.4, = 0.699, = 1.22 A, = 45.95
�� = A, � = 40N.m, � = 0.01 N.ms, = 0.117 kg.m2 � =
� −
��.
Where fis a low-pass filter with a steady-state gain of one The IMC-PID controller parameters where designed to be
and is typically of the form � = 8.363; � = 70.717; = 0.0677 for armature control
and � = 8.240; � = 0.698; = 3.17 for field control.
=
� + MATLAB/Simulink software was used to carry out the
simulation, while the reference speed levels for the SEDCM
� τc is the desired closed-loop time constant – the choice of were chosen as follows:
which is a major decision in the IMC design procedure, ris  1500 rpm was chosen as the reference below the
a positive integer parameter that will be chosen as for this rated speed – DC motors are typically run at speeds
their rated.
design. The low-pass filter f helps to ensure that the
resultant IMC-tuned PID controller is reasonably robust,  1750 rpm which is the rated speed of the motor.
physically realizable and stable [6,14].  2000 rpm – this reference was chosen, just slightly
Since the desired closed-loop response (transfer function) above the rated so as not to overrun and damage the
for the DC motor is a second order system with the general motor.
form Figs 6 to 8 shows the speed response of the motor with the
combined approach compared with the individual methods at
� = various setpoint values. Fig 6 illustrates the transient response
+ +
of the DC motor at 1500 rpm reference speed. It is observed
The tuning parameter � is obtained from the roots of the
that the combined control method has a minimum overshoot
characteristic equation, that is, the poles of the system; which
(0.0026%), smaller settling time (0.058seconds) and, hence,
can either be real, or a complex conjugate pair. The systems
faster response compared to the armature and field control
damping ratio � < , which means there is some overshoot
methods. This means that the combined control method
to the step response. Typically for an underdamped response
tracks the reference speed faster and more accurately.
the poles of the system are a complex conjugate pair (−� ±
�), hence � is obtained from the reciprocal of the real part
of the poles [15].
� =
|� |
The tuning parameter � ensures a reasonable trade-off
between performance and robustness, where a small value is
ideal for performance while a large value favours robustness.
This work utilizes a value of � that will favour good
performance subject to satisfactory robustness [16]. A value
of 0.02 is chosen for � , while a value of 0.05 is chosen for
�� � , for �� and �� respectively. Performance for the
controller, here, is defined as excellent setpoint tracking,
improved robustness, a fast and non-oscillatory response for
the DC motor speed. Comparing the two block diagrams in Fig. 7. Response of the SEDCM at 1500 rpm

Authorized licensed use limited to: INDIAN INSTITUTE OF TECHNOLOGY DELHI. Downloaded on February 28,2023 at 17:42:57 UTC from IEEE Xplore. Restrictions apply.
 2019 IEEE PES/IAS PowerAfrica
A summary of the performance characteristics of the TABLE II. PERFORMANCE SUMMARY AT RATED SPEED
combined control method compared to the individual control
methods, at reference speed 1500 rpm, is given in Table 1; At rated speed (� = 1750 rpm)
where is the settling time, is the rise time, is the
steady-state error, �� the percent overshoot. Parameter Field Control Combined Control

TABLE I. PERFORMANCE SUMMARY AT BELOW RATED SPEED (s) 0.1645 0.0580

Below rated speed (� = 1500 rpm) (s) 0.0880 0.0198

Parameter Armature Control Field Control Combined Control �� (%) 0.0182 0.0028

(s) 0.0905 0.1654 0.0580 0.00 0.00

(s) 0.0331 0.0882 0.0198

�� (%) 0.0532 0.0168 0.0026

0.00 0.00 0.00

At the rated speed, 1750 rpm, which is the that corresponds

with the rated armature voltage, rated armature current, and
rated field current. The field control and combined control
methods are used to achieve and maintain the speed of the DC
motor at the rated speed and above the rated speed. Armature
voltage control is not considered here because it is only used
to achieve speed below the rated [12]. Fig 7 shows the
response of the motor at the rated speed of 1750 rpm. As can
be seen from the response the combined control method has
a faster response and minimal overshoot compared to the
field control method.
The speed of the motor was then increased to just slightly
above the rated speed, 2000 rpm, in order not to damage the
motor. The combined control and field control methods are Fig. 9. Response of the SEDCM at 2000 rpm
used to achieve this by maintaining the armature voltage at
its rated value while varying the field current in the constant TABLE III. PERFORMANCE SUMMARY AT ABOVE RATED SPEED
power region [12]. Fig. 8 shows the response of the motor at
reference speed value, above the rated speed. It can be seen Above rated speed (� = 2000 rpm)
that the combined control method gives a better performance.
Parameter Field Control Combined Control

(s) 0.1637 0.0580

(s) 0.0878 0.0198

�� (%) 0.0191 0.0024

0.00 0.00

VI. CONCLUSION
In this paper, the speed of a SEDCM was successfully
controlled using model-based control techniques. From the
simulation results it is clear that modelling both the armature
Fig. 8. Response of the SEDCM at 1750 rpm voltage and field current control methods is the most effective
in accurately capturing the dynamic effects of the armature
and field excitation of the motor. An accurate and robust
model is essential for a better control performance.

Authorized licensed use limited to: INDIAN INSTITUTE OF TECHNOLOGY DELHI. Downloaded on February 28,2023 at 17:42:57 UTC from IEEE Xplore. Restrictions apply.
 2019 IEEE PES/IAS PowerAfrica
A wider range of speed control – from below the rated speed [8] S. Cong, Y. Liang, PID-Like Neural Network Nonlinear Adaptive
to above the motor’s rated speed – is achieved with the Control for Uncertain Multivariable Motion Control Systems, IEEE
Transactions on Industrial Electronics. 56, 10 (2009) 3872 – 3879.
combined armature voltage and field current control [9] A. Tepljakov, E. A. Gonzalez, E. Petlenkov, J. Belikov, C. A. Monje,
methods, which is an added advantage compared to the I. Petras, Incoporation of fractional-order dynamics into an existing
individual control methods. From the performance PI/PID DC motor control loop, ISA transactions. 60 (2016) 262 – 273.
comparisons it can be seen that the combined control method [10] R. K. Munje, M. R. Roda, B. E. Kushare, Speed Control of DC motor
produces better result with respect to the rise time , settling using PI and SMC in Proceedings of 2010 IEEE IPECON, Singapore,
Singapore, October 27 – 29, 2010.
time , and overshoot.
[11] A. Hussein, K. Hirasawa, J. Hu, Identification and Control of a PV-
These results are indicative of the ability of the model-based supplied separately excited DC motor using Universal Learning
control to suitably select an output voltage in order to achieve Networks, IFAC Proceedings. 36, 16 (2003) 1411 – 1416.
a steady motor reference speed. [12] A. A. Sadiq, G. A. Bakare, E. C. Anene, B. H. Mamman, A fuzzy-based
speed control of DC motor using combined armature voltage and field
REFERENCES current in Proceedings of 3rd IFAC International Conference on
Intelligent Control and Automation Science, Chengdu, China,
[1] G. Rizzoni, Fundamentals of Electrical Engineering, first ed., McGraw September 2 – 4, 2013
Hill, New York, 2009.
[13] D. E. Rivera, M. Morari, S. Skogestad, Internal Model Control 4 PID
[2] T. R. Dil Kumar, S. J. Mija, Design and Performance Evaluation of Controller Design, Industrial & Engineering Chemistry Process Design
Robust SMC schemes for Speed Control of DC Motor in Proceedings and Development. 25, 1 (1986) 252 – 265.
of 2014 IEEE International Conference on Advanced Communication
Control and Computing Technologies (ICACCCT 2014), [14] M. V. Subramanyam, K. S. Prasad, P. V. Gopi Krishna Rao, Robust
Ramanathapuram, India, May 8 – 10, 2014. Control of steam turbine speed using improved IMC tuned PID
controller, Procedia Engineering. 38 (2012) 1450 – 1456.
[3] P. C. Sen, Principles of Electrical Machines and Power Electronics,
third ed., John Wiley & Sons, New York, 2013. [15] N. S. Nise, Control Systems Engineering, Seventh ed., John Wiley &
Sons Inc., California, 2015.
[4] A. M. Harb, I. A. Smadi, Tracking control of DC motors via MIMO
nonlinear fuzzy control, Chaos, Solitons and Fractals. 42 (2009) 702 – [16] C. Grimholt, S. Skogestad, Optimal PID Control of double integrating
710. processes in Proceedings of 11th IFAC Symposium on Dynamics and
Control of Process Systems, including Biosystems, Trondheim,
[5] B. Bhusan, M. Singh, Adaptive Control of DC motor using bacterial Norway, June 6 – 8, 2016.
foraging algorithm, Applied Soft Computing. 11 (2011) 4913 – 4920.
[17] C. Grimholt, S. Skogestad, Optimal PI-Control and Verification of the
[6] D. E. Seborg, T. F. Edgar, D. A. Mellichamp, F. J. Doyle III, Process SIMC Tuning Rule in Proceedings of IFAC Conference on Advances
Dynamics and Control, third ed., John Wiley & Sons, New Jersey, in PID Control PID’12, Brescia, Italy, March 28 – 30, 2012.
2011.
[18] K. Soltesz, C. Grimholt, S. Skogestad, Simultaneous Design of PID
[7] B. Lennox, O. Marjanovic, Process Control & Automation: Lecture Controller and Measurement Filter by Optimization, IET Control
Notes (Academic Year 2010/11), School of Electrical & Electronic Theory and Applications, 11, 3 (2017) 341 – 348.
Engineering, University of Manchester.

1
Authorized licensed use limited to: INDIAN INSTITUTE OF TECHNOLOGY DELHI. Downloaded on February 28,2023 at 17:42:57 UTC from IEEE Xplore. Restrictions apply.