Hindawi Publishing Corporation

Advances in Artificial Neural Systems

Volume 2014, Article ID 126317, 8 pages
http://dx.doi.org/10.1155/2014/126317

Research Article
Optimal Design of PID Controller for the Speed Control of
DC Motor by Using Metaheuristic Techniques

Mirza Muhammad Sabir1 and Junaid Ali Khan2

1
University Institute of Information Technology, PMAS-Arid Agriculture University, Rawalpindi 46000, Pakistan
2
Faculty of Engineering and Technology, HUIC, Islamabad 44000, Pakistan

Correspondence should be addressed to Junaid Ali Khan; [email protected]

Received 22 July 2014; Accepted 19 November 2014; Published 10 December 2014

Academic Editor: Paolo Gastaldo

Copyright © 2014 M. M. Sabir and J. A. Khan. 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.

DC motors are used in numerous industrial applications like servo systems and speed control applications. For such systems, the
Proportional+Integral+Derivative (PID) controller is usually the controller of choice due to its ease of implementation, ruggedness,
and easy tuning. All the classical methods for PID controller design and tuning provide initial workable values for 𝐾𝑝 , 𝐾𝑖 , and 𝐾𝑑
which are further manually fine-tuned for achieving desired performance. The manual fine tuning of the PID controller parameters
is an arduous job which demands expertise and comprehensive knowledge of the domain. In this research work, some metaheuristic
algorithms are explored for designing PID controller and a comprehensive comparison is made between these algorithms and
classical techniques as well for the purpose of selecting the best technique for PID controller design and parameters tuning.

1. Introduction that could be incorporated in control applications [6–8].

Among the few mentioned types of controllers, PID con-
In this modern industrial age, there is hardly any industrial troller is one of the earliest and best understood controllers
application in which DC motors are not being used [1, 2]. which is incorporated in almost every industrial control
This is so because of ease of control, low cost maintenance application due to its efficiency and ease of implementa-
especially of brushless DC motor type, low price, and tion [9]. Although there are many classical techniques for
ruggedness of DC motor over a wide range of applications. designing and tuning PID controller parameters (𝐾𝑝 , 𝐾𝑖 , 𝐾𝑑 )
Some industrial applications, which are worth mentioning, in which are widely understood and easily applied, one of the
which DC motors are being used widely are machine tools, main disadvantages of these classical techniques is that, for
paper mills, textile industry, electric traction, and robotics. tuning PID controller through these techniques, expertise
The flexibility in controller design of DC motors is due to and experience are required. This is so because these methods
the fact that armature winding and field winding could be provide a starting point and achieving desired performance
controlled separately [3]. In most of the applications of speed fine tuning of parameters through hit-and-trial method is
control of DC motors, the current in field winding is kept required. However, metaheuristic techniques may be a good
constant and the current in armature winding is varied or choice to its dynamic nature.
vice versa which gives excellent speed control performance Over the years, many metaheuristic and stochastic opti-
over a wide range of desired values. In these applications, the mization techniques have been developed which are being
purpose is to track the speed command by keeping output applied in every discipline of life [10–12]. These techniques are
speed at desired level and to achieve desire speed or position nature inspired depending upon the swarm intelligence, evo-
control in minimum time without having large overshoots lutionary, or foraging behavior of different species. Some of
and settling times [4, 5]. the widely used techniques are genetic algorithm (GA), par-
There are different types of controllers like lead, lag, LQR ticle swarm optimization (PSO), and simulated annealing
(linear quadratic regulator), PID, and sliding-mode control (SA). These metaheuristic algorithms have been successfully
2 Advances in Artificial Neural Systems

Td (s) Table 1: DC motor parameters.

Armature Tm (s) − Load ̇

𝜃(s) Motor parameters Value
Ua (s)
+
1 Km
+
1 𝐾𝑚 0.023 N⋅m/A
Ls + ra Jm s + Bm
− 𝐾𝑡 0.023 N⋅m/A
𝑟𝑎 1.0 Ω
𝐿 0.5 H
Kb
𝐽𝑚 0.01 Kg⋅m2 /rad
𝐵𝑚 0.00003 N⋅m
Figure 1: Block diagram of DC motor.

45
40

Open loop speed (rpm)

applied also in various fields of control systems and results 35
obtained through these techniques have proven their suprem- 30
acy over the classical techniques [13–18]. 25
There have been previously many attempts for the nature 20
inspired PID controller tuning; however, according to the best 15
of our knowledge there has been a little work done in exploit- 10
ing the power of hybrid techniques for the PID controller 5
tuning. In this work, a PID controller design for speed control 0
0 20 40 60 80 100 120
of DC motor is presented. First, the design through classical
Time (s)
techniques like Zeigler-Nichols and Cohen-Coon methods is
presented for establishing a base line. Then, six metaheuristic Figure 2: Open loop step response of DC motor.
optimization algorithms are used to find the best possible
parameters of PID controller subjected to minimization of a
cost function and among these three of the hybrid techniques 𝑈𝑎 (𝑠) − 𝐾𝑏 𝜃̇ (𝑠)
are used to establish the superiority of hybrid metaheuristic 𝐼𝑎 (𝑠) = ,
𝑟𝑎 + 𝐿𝑠
techniques over the others. A comprehensive comparison is
made between the classical techniques and the metaheuristic 𝑇𝑙 (𝑠) = 𝐽𝑚 𝑠2 𝜃 (𝑠) + 𝐵𝑚 𝑠𝜃 (𝑠) ,
techniques to show the strength, stability, and efficiency of
these methods over the classical techniques. Although some 𝑇𝑙 (𝑠) = 𝑇𝑚 (𝑠) − 𝑇𝑑 (𝑠) ,
stochastic algorithms have been used [13, 16] previously (1)
for design of PID controller for different applications, this
study presents some hybrid techniques by combining the where 𝑈𝑎 is armature applied voltage, 𝑈𝑏 is back-emf, 𝐾𝑚 is
global and local search techniques of swarm intelligence motor constant, 𝐾𝑏 is back emf constant, 𝐽𝑚 is inertia of rotor,
and evolutionary algorithms for PID controller design and 𝐵𝑚 is viscous damping, 𝑇𝑚 is developed motor torque, 𝑇𝑙 is
it could provide a framework of PID controller tuning by torque delivered to load, 𝑇𝑑 is disturbance torque (we neglect
considering the hybrid nature of metaheuristic techniques. it), 𝑟𝑎 is armature resistance, 𝐿 is armature inductance, 𝐼𝑎 is
The organization of the rest of the paper is as follows. In armature current, and 𝑠 is s-plane.
Section 2, mathematical model of DC motor is derived. In By using (1) the transfer function of DC motor is
Section 3, overview of classical techniques for PID tuning is
presented. In Section 4, an overview of GA, PSO, SA, and 𝜃̇ (𝑠) 𝐾𝑚
Nelder-Mead (NM) is presented. In Section 5, results and = . (2)
𝑈𝑎 (𝑠) (𝑟𝑎 + 𝐿𝑠) (𝐽𝑚 𝑠 + 𝐵𝑚 𝑠) + 𝐾𝑚 𝐾𝑏
simulations are discussed. In Section 6, conclusions about the
present study are presented.
The parameters selected for DC motor simulation are
listed in Table 1 [20].
2. Mathematical Model of DC Motor The open loop step response of DC motor without PID
controller is shown in Figure 2 and output values of response
The dynamic behavior of DC motor is given by following set are provided in Table 2.
of relations [19] and its block diagram is shown in Figure 1.
A simplified linear model is presented for this work ignoring
the nonlinearities like the backlash and dead zones to simplify 3. Classical Techniques
the application of metaheuristic techniques. Consider The response of DC motor is calculated by Ziegler-Nichols
(Z-N) and Cohen-Coon (C-C) methods, respectively, in
𝑇𝑚 (𝑠) = 𝐾𝑚 𝐼𝑎 (𝑠) , order to validate the proposed scheme.
𝑈𝑎 (𝑠) = (𝑟𝑎 + 𝐿𝑠) 𝐼𝑎 (𝑠) + 𝑈𝑏 (𝑠) ,
3.1. Ziegler-Nichols (Z-N) Method. Ziegler and Nichols pro-
𝑈𝑏 (𝑠) = 𝐾𝑏 𝜃̇ (𝑠) , posed a rule for design and tuning of PID controller. From the
Advances in Artificial Neural Systems 3

Table 2: DC motor output response. Table 3: 𝐾𝑝 , 𝐾𝑖 , and 𝐾𝑑 values by using Z-N method.

Response Value Method 𝐾𝑝 𝐾𝑖 𝐾𝑑

Settling time, 𝑡𝑠 68.6029 Ziegler-Nichols 0.8000 0.6380 0.2508
Rise time, 𝑡𝑟 38.2458
Maximum percent overshoot, 𝑀𝑝 0
Table 4: 𝐾𝑝 , 𝐾𝑖 , and 𝐾𝑑 values by using C-C method.
Steady state error, 𝑒𝑠𝑠 0.0237
Method 𝐾𝑝 𝐾𝑖 𝐾𝑑
open loop step response of system one can find the following Cohen-Coon 0.3522 0.2317 0.0798
set of points (𝑡𝑎 , 𝑦𝑎 ) and (𝑡𝑏 , 𝑦𝑏 ). The first point corresponds
to output of step response at a value of 35.3% and the second
point corresponds to output of step response at a value of 4. Metaheuristic Algorithms: An Overview
85.3% [20]. By using this method, we get open loop step
response values of DC motor given as follows: 4.1. Genetic Algorithm (GA). The GA was invented by John
Holland by inspecting adaptation in natural and artificial
𝑡𝑎 = 8 sec, 𝑦𝑎 = 14.52,
(3) systems in 1975 [21]. Holland proposed GA as a heuristic
𝑡𝑏 = 37.7 sec, 𝑦𝑏 = 35.09, tool which is based on the method of “survival of the fittest.”
The principle of GA is simple: imitate genetic and natural
and dc gain is selection in the form of computer program. The parameters of
the problem are coded as DNA-like linear data structure, a
𝛽 = 41.1449. (4) vector or a string. GA differs from conventional optimization
According to Ziegler-Nichols rule the following relations techniques in a way that
could be used for determining PID controller parameters:
(i) GA works with coding the solution set and the
𝑇 solution is not itself used;
𝐾𝑝 = 1.2 , (5)
𝛽𝛼 (ii) GA uses population of solution for finding the opti-
mum solution;
𝑇𝑖 = 2𝛼, (6)
(iii) it uses the fitness function rather than derivatives for
𝑇𝑑 = 0.5𝛼, (7) evaluation purpose;
(iv) it uses the probabilistic transition operators whereas
𝐾𝑃
𝐾𝑖 = , (8) other techniques use deterministic transition opera-
𝑇𝑖 tors.
𝐾𝑑 = 𝑇𝑑 𝐾𝑃 , (9) The general steps in GA algorithm are as follows.
where
Step 1 (create a random initial population). An initial popula-
𝑇 = 0.67 (𝑡𝑏 − 𝑡𝑎 ) , (10) tion is created from a random selection of solution.

𝛼 = 1.3𝑡𝑎 − 0.29𝑡𝑏 . (11) Step 2 (evaluate fitness). Each candidate solution is evaluated
against a fitness function.
By using (5) to (11) we get the values of 𝐾𝑝 , 𝐾𝑖 , and 𝐾𝑑
listed in Table 3. Step 3 (reproduction). The chromosomes with higher value of
fitness value are more likely to reproduce offspring.
3.2. Cohen-Coon (C-C) Method. Cohen and Coon also pro-
posed a method for design and tuning of PID controller. Step 4 (next generation). If the new generation contains
Based on this method, PID controller parameters could be desired solution, then problem has been solved, and if it is not
determined as follows [20]: the case, then the new population will go through the same
process as described above.
𝑇 4 𝛼
𝐾𝑝 = [ + ],
𝛼𝛽 3 4𝑇 Step 5. Termination based on number of iterations or required
fitness value was achieved.
32 + 6𝛼/𝑇 In our design method PID controller parameters (𝐾𝑝 ,
𝑇𝑖 = 𝛼 [ ], (12)
13 + 8𝛼/𝑇 𝐾𝑖 , 𝐾𝑑 ) are viewed as individuals and each parameter value
is coded by a real number.
4
𝑇𝑑 = 𝛼 [ ].
11 + 2𝛼/𝑇
4.2. Particle Swarm Optimization (PSO). The particle swarm
By using (12) and also from (8) to (11) the following values optimization algorithm was proposed by Kennedy and Eber-
of 𝐾𝑝 , 𝐾𝑖 , and 𝐾𝑑 have been obtained and listed in Table 4. hart in 1995 [22]. This algorithm is based on the social
4 Advances in Artificial Neural Systems

behavior of swarm such as bird flocking and fish schooling. Table 5: Parameters of PID through Z-N and C-C methods.
Here, each individual bird is called the particle, and in our
case each individual particle has three attributes which are Method 𝐾𝑝 𝐾𝑖 𝐾𝑑
𝐾𝑝 , 𝐾𝑖 , and 𝐾𝑑 . Ziegler-Nichols 0.8000 0.6380 0.2508
Each particle in PSO flies in the whole search space with Cohen-Coon 0.3522 0.2317 0.0798
a certain velocity which is dynamically adjusted according
to its own flying experience and the knowledge gain by the
Table 6: Response for Z-N and C-C methods.
velocities of its peers. Each particle is treated as particle in
𝑑-dimensional search space. Method 𝑡𝑠 𝑡𝑟 𝑀𝑝 𝑒𝑠𝑠
The change in velocity and position of each particle is Ziegler-Nichols 6.2222 0.8058 29.2329 0.5
governed by the following relations:
Cohen-Coon 12.6787 1.3370 36.8658 0.5
V (new) = 𝛿V (old) + 𝑎1 [𝑝best − 𝑥 (old)] + 𝑎2 [𝑔best − 𝑥 (old)]
𝑥 (new) = 𝑥 (old) + V (new) , vertex and the search is continued by forming the triangular
(13) patterns. The function values at vertices get smaller and
smaller and the triangle size reduces which ultimately results
where 𝛿 is inertia weight factor, 𝑝best is local best, 𝑔best is in finding the local minimum.
global best, V is velocity of the particle, 𝑥 is the position of
the particle, 𝑎1 [𝑝best − 𝑥(old)] is individual intelligence, and
𝑎2 [𝑔best − 𝑥(old)] is collective intelligence. 5. Results and Discussions
Using the classical methods of PID tuning, Ziegler-Nichols
4.3. Simulated Annealing (SA). The simulated annealing was and Cohen-Coon, PID controller was tuned by keeping
originally inspired by the crystal formation process when the objective that settling time 𝑡𝑠 and rise time 𝑡𝑟 should
solids are cooled down from some high temperature [21]. The be minimum and there should be no overshoots; that is,
slower the cooling is, the more perfect crystal is formed. In maximum percent over shoot 𝑀𝑝 should be zero. The PID
cooling process, we ultimately converge towards minimal and controller parameters obtained by these classical methods are
stable energy state. The system movement towards stability given in Table 5.
is random, but the probability to stay in any particular By using the above values of 𝐾𝑝 , 𝐾𝑖 , and 𝐾𝑑 , PID
configuration directly depends on the energy of the system controller was designed and the closed loop response of this
and on its temperature. The said probability is given by the controller with the plant is shown in Figure 3 and also the
following Gibbs relation: output values of 𝑡𝑠 , 𝑡𝑟 , 𝑀𝑝 and steady state error 𝑒𝑠𝑠 are shown
in Table 6.
𝑝 = 𝑒𝐸/𝑘𝑇 , (14) It is apparent from Figure 3 and Table 6 that the settling
time and maximum percent overshoot are very large and PID
where 𝐸 is the energy in different energy states, 𝑘 is the controller designed by these methods will not provide desired
Boltzmann constant, and 𝑇 is the temperature. performance. So these techniques are not helpful in tuning
In 1970, Kirkpatrick formed the simulated annealing the PID controllers for the systems in which the parameters
algorithm by using the above said analogy. The SA behaves of 𝑡𝑠 , 𝑡𝑟 , and 𝑀𝑝 are critical.
like hill climbing algorithm, but it avoids the local minimum The same problem is also solved by six metaheuristic
and it usually gives good results. As in stochastic hill climbing algorithms, namely, GA, PSO, SA, GA-NM, PSO-NM, and
algorithm, in each iteration of SA a new solution in the SA-NM, for designing and tuning of PID controller. Not only
neighborhood of actual solution is randomly chosen. If the the global search but also the hybrid searching techniques
fitness value of the new solution is better than the previous using the Nelder-Mead algorithm as the local minimization
value, then the new solution is accepted as the new current search technique are also applied. The purpose is to study the
solution. If the fitness function is not improved, then the new optimization capabilities of these metaheuristic algorithms
solution is retained with a probability and to demonstrate that by using these techniques PID con-
troller design and tuning give the more accurate and better
𝑝 = 𝑒−[𝑓(𝑦)−𝑓(𝑥)]/𝑘𝑇 , (15) results. There might be many possible objective functions like
integral of time-absolute-error (ITAE), integral of absolute-
where 𝑓(𝑦)−𝑓(𝑥) is the difference of fitness function between error (IAE), integral of time-weighted-squared-error (ITSE),
the new and the old solutions. and integral of squared-error (ISE), but the objective function
used in these algorithms for the minimization is defined as
4.4. Nelder-Mead (NM) Method. Nelder-Mead method is a follows [24]:
simplex method which is used to find the local minimum
of a function of several variables [23]. For the two variables min 𝐽 = (1 − 𝑒−𝜉 ) (𝑀𝑝 + 𝑒𝑠𝑠 ) + 𝑒−𝜉 (𝑡𝑠 − 𝑡𝑟 ) . (16)
function simplex is a triangle of three vertices and the method
is a pattern search. At the worst vertex the function value One can change the value of 𝜉, which is the weighting
is the largest and this worst vertex is replaced with new factor, in three steps, that is, 𝜉 = 0.5, 𝜉 = 1.0, and 𝜉 = 1.5,
Advances in Artificial Neural Systems 5

Table 7: Parameters range.

Parameter Lower bound Upper bound

𝐾𝑝 0.0001 1.5000 10−0.3
𝐾𝑑 0.0001 1.0000
𝐾𝑖

Fitness value
0.0001 1.0000

10−0.4
1.4
1.2
1
Amplitude

0.8
10−0.5
0.6
0.4 0 20 40 60 80 100
0.2 Number of runs
0
0 2 4 6 8 10 12 14 16 18 NM-GA
NM-PSO
Time (s)
NM-SA
Cohen-Coon
Ziegler-Nichols Figure 4: Fitness (𝐽) for GA-NM, PSO-NM, and SA-NM, 𝜉 = 0.5.

Figure 3: Step response of system tuned by ZN and CC methods.

and see the effect of change of 𝜉 on 𝑡𝑠 , 𝑡𝑟 , and 𝑀𝑝 . The 10−0.3

lower and upper bounds on the values of 𝐾𝑝 , 𝐾𝑖 , and 𝐾𝑑 are
Fitness value

given in Table 7 [25], while the parameters settings or values

used in the simulations of various solvers are provided in
Table 8.
The fitness function 𝐽 values obtained by using these
10−0.4
algorithms are shown in Tables 9–11.
As shown in Tables 9–11 and also in Figures 4, 5, and 6 0 10 20 30 40 50 60 70 80 90 100
the minimum objective function fitness value is found by SA- Number of runs
NM. The objective function contains the exponential factor
NM-GA
and also the SA algorithm is governed by the Boltzmann NM-PSO
distribution of the energies in different energies states which NM-SA
is also in the form of exponential function. So it is the reason
that, for this particular problem, when we combined the Figure 5: Fitness (𝐽) for GA-NM, PSO-NM, and SA-NM, 𝜉 = 1.0.
SA with the local search method of NM it gave the best
results even better than PSO and GA alone. So if the problem
landscape and the algorithm which is being used match, then
best results would be obtained. Also from Tables 8–10, it could a simplified linear model of the DC motor is considered
be seen that the standard deviation for the SA-NM is small, neglecting the nonlinearities like backlash, dead zone, and
which is a measure of convergence speed, and it could be effects of load torque changes.
inferred that SA-NM converges fast as compared to other
algorithms used in this research work. 6. Conclusion
The step response for the mean values of 𝐾𝑝 , 𝐾𝑖 , and 𝐾𝑑
tuned by SA-NM is shown in Figure 7 and also values of 𝑡𝑠 , 𝑡𝑟 , On the basis of results obtained in the previous section the
and 𝑀𝑝 are given in Table 12. The step response in Figure 7 following conclusions could be drawn.
shows marked improvement over the previous step response The problem under discussion could be solved using
of Figure 3 when PID controller was tuned by using classical metaheuristic techniques which give a substantial amount of
techniques. improvement in terms of 𝑡𝑠 , 𝑡𝑟 , and 𝑀𝑝 as can be seen from
A multiple setpoint command is applied to the system Tables 6 and 12. Using metaheuristic techniques 𝑡𝑠 decreased
tuned by SA-NM, Z-N, and C-C. The response to multiple by 83.93%, 𝑡𝑟 decreased by 17.58%, and 𝑀𝑝 decreased by
setpoint is also shown in Figures 8 and 9. It is quite obvious 98.71% as compared to Z-N method. Also the decrease in
from Figure 9 that the multiple setpoint tracking is excellent value for 𝑡𝑠 , 𝑡𝑟 , and 𝑀𝑝 was 92.11%, 50.32%, and 98.97%,
when PID controller is tuned using SA-NM. In this work, respectively, as compared to C-C method. So the objective
6 Advances in Artificial Neural Systems

Table 8: Parameters setting and fitness function.

Fitness function: min 𝐽 = (1 − 𝑒−𝜁 )(𝑀𝑝 + 𝑒𝑠𝑠 ) + 𝑒−𝜁 (𝑡𝑠 − 𝑡𝑟 )

GA parameters PSO parameters SA parameters NM parameters
Population type: double vector Population type: double vector Max. evaluations: 3000 Max. iterations: 200
Population size: 20 Population size: 40 Initial temperature: 100 Max. function evaluations: 600
Temperature update function: Start point = Best parameters of
Selection function: tournament Cognitive attraction: 0.5
exponential SA, GA, PSO
Crossover fraction: 0.8 Social attraction: 1.25 Reannealing interval: 100
Mutation function: Gaussian Stall generations: 50 Annealing function: Boltzmann Others = Default
Stall generations: 50 Function tolerance: 1𝑒 − 6 Stall iterations: 500 𝑋 tolerance: 1𝑒 − 4
Function tolerance: 1𝑒 − 6 Number of epochs: 100 Function tolerance: 1𝑒 − 6 Function tolerance: 1𝑒 − 6
Number of epochs: 100 Others = Default Number of epochs: 100 Number of epochs: 100

Table 9: Fitness (𝐽) case 1, 𝜉 = 0.5. Table 10: Fitness (J) case 2, 𝜉 = 1.0.

Algo. Min. Max. Mean Var. Std. Runs Algo. Min. Max. Mean Var. Std. Runs
GA 0.4303 0.6804 0.4693 0.0019 0.0433 100 GA 0.4563 14.5462 0.6789 2.2793 1.5097 100
PSO 0.3010 0.4402 0.3665 0.0037 0.0612 100 PSO 0.3793 0.4566 0.4206 0.0014 0.0368 100
SA 0.3040 15.9880 1.1158 6.7519 2.5984 100 SA 0.3803 18.3989 NaN NaN NaN 100
NM-GA 0.3011 0.5827 0.4272 0.0016 0.0398 100 NM-GA 0.3795 0.5436 0.4604 0.0005 0.0233 100
NM-PSO 0.3010 0.4311 0.3655 0.0038 0.0618 100 NM-PSO 0.3793 0.4559 0.4203 0.0014 0.0371 100
NM-SA 0.3010 0.4573 0.3108 0.0005 0.0229 100 NM-SA 0.3793 0.4015 0.3839 0 0.0046 100

10−0.28
Table 11: Fitness (J) case 3, 𝜉 = 1.5.

Algo. Min. Max. Mean Var. Std. Runs

10−0.3 GA 0.4739 0.5564 0.4907 0.0003 0.0177 100
PSO 0.4268 0.5093 0.4524 0.0006 0.0236 100
Fitness value

SA 0.4282 43.1661 1.8708 40.1765 6.3385 100

10−0.32
NM-GA 0.4301 0.5259 0.4767 0.0002 0.0149 100
NM-PSO 0.4268 0.4933 0.4520 0.0005 0.0231 100
10−0.34 NM-SA 0.4268 0.4740 0.4306 0 0.0070 100

10−0.36 Table 12: Step response values for SA-NM.

0 20 40 60 80 100 Method 𝑡𝑠 𝑡𝑟 𝑀𝑝 𝑒𝑠𝑠

Number of runs SA-NM 0.9999 0.6641 0.3773 0.5
NM-GA
NM-PSO
NM-SA
𝑀𝑝 . Hence, metaheuristic techniques are far better than the
Figure 6: Fitness (𝐽) for GA-NM, PSO-NM, and SA-NM, 𝜉 = 1.5. classical techniques, and among these hybrid approach is
preferable. However, in future work, these effects could be
considered to form a comprehensive nonlinear model of DC
motor and solve the problem of PID tuning based on the
of minimizing 𝑡𝑠 , 𝑡𝑟 , and 𝑀𝑝 is better achieved through mentioned metaheuristic techniques because according to
metaheuristic algorithms, which validates the capabilities of the best of our knowledge there has been a lot of research
these algorithms of best optimization algorithms for complex potential for the PID controller tuning by considering the
problems like PID controller tuning. nonlinearities of DC motor model. Some other techniques
Among all metaheuristic techniques for optimization the like ant colony optimization (ACO), bacterial foraging algo-
hybrid approach of SA-NM showed the best results and rithm (BFA), and differential evolution (DE) could also be
through this we obtained the PID controller parameters considered by making the hybrid pairs of these algorithms
which showed the excellent results in terms of 𝑡𝑠 , 𝑡𝑟 , and with some local search techniques.
Advances in Artificial Neural Systems 7

1.4 application,” IEEE Transactions on Magnetics, vol. 48, no. 11, pp.
1.2 3348–3351, 2012.
1 [2] A. Sathyan, M. Krishnamurthy, N. Milivojevic, and A. Emadi,
Amplitude

0.8 “A low-cost digital control scheme for brushless DC motor

drives in domestic applications,” in Proceedings of the 2009 IEEE
0.6
International Electric Machines and Drives Conference, pp. 76–
0.4 82, May 2009.
0.2 [3] P. C. Krause, O. Wasynczuk, and S. D. Sudhoff, Analysis of
0 Electric Machinery and Drives, John Wiley & Sons, 2013.
0 0.5 1 1.5 2 2.5 3 3.5 4
[4] A. Andrzejewski, “The time-minimal and without overshoot
Time (s)
speed control of DC motor,” in Proceedings of the International
SA-NM 𝜁 = 0.5 Conference on Computer as a Tool (EUROCON ’07), pp. 1792–
SA-NM 𝜁 = 1.0 1799, September 2007.
SA-NM 𝜁 = 1.5 [5] S. Galijašević, Š. Mašić, S. Smaka, A. Akšamović, and D. Balić,
“Parameter identification and digital control of speed of a
Figure 7: Step response for SA-NM. permanent magnet DC motors,” in Proceedings of the 23rd
International Symposium on Information, Communication and
30 Automation Technologies, October 2011.
[6] F. E. Hoyos, A. Rincón, J. A. Taborda, N. Toro, and F. Angulo,
Speed command (rpm)

20
“Adaptive quasi-sliding mode control for permanent magnet
10 DC motor,” Mathematical Problems in Engineering, vol. 2013,
0 Article ID 693685, 12 pages, 2013.
−10 [7] H. Zhang, L. Ge, M. Shi, and Q. Yang, “Research of compound
control for DC motor system based on global sliding mode
−20 disturbance observer,” Mathematical Problems in Engineering,
−30 vol. 2014, Article ID 759147, 7 pages, 2014.
0 25 50 75 100 125 150
[8] J. Yao, G. Yang, Z. Jiao, and D. Ma, “Adaptive robust motion
Time (s) control of direct-drive DC motors with continuous friction
Multiple setpoint tracking Zn compensation,” Abstract and Applied Analysis, vol. 2013, Article
Speed command (rpm) ID 837548, 14 pages, 2013.
Multiple setpoint tracking CC [9] R. Namba, T. Yamamoto, and M. Kaneda, “Robust PID con-
troller and its application,” in Proceedings of the 1997 IEEE
Figure 8: Multiple setpoint tracking for Z-N and C-C tuned PID. International Conference on Systems, Man, and Cybernetics, pp.
3636–3641, October 1997.
20 [10] Y. Harrath, B. Chebel-Morello, and N. Zerhouni, “A genetic
15 algorithm and data mining based meta-heuristic for job shop
Speed command (rpm)

10 scheduling problem,” in Proceedings of the 2002 IEEE Interna-

tional Conference on Systems, Man and Cybernetics, pp. 280–285,
5
October 2002.
0
[11] X. Li and M. Yin, “Self-adaptive constrained artificial bee colony
−5 for constrained numerical optimization,” Neural Computing
−10 and Applications, vol. 24, no. 3-4, pp. 723–734, 2014.
−15 [12] Z. Cui and X. Gao, “Theory and applications of swarm intel-
−20 ligence,” Neural Computing and Applications, vol. 21, no. 2, pp.
0 25 50 75 100 125 150 205–206, 2012.
Time (s)
[13] R. G. Kanojiya and P. M. Meshram, “Optimal tuning of PI con-
Speed command (rpm) troller for speed control of DC motor drive using particle swarm
Multiple setpoint tracking SA-NM optimization,” in Proceedings of the International Conference on
Advances in Power Conversion and Energy Technologies (APCET
Figure 9: Multiple setpoint tracking for SA-NM tuned PID. ’12), August 2012.
[14] Y. Wang, C. Xia, M. Zhang, and D. Liu, “Adaptive speed
control for brushless DC motors based on genetic algorithm and
Conflict of Interests RBF neural network,” in Proceedings of the IEEE International
Conference on Control and Automation (ICCA ’07), pp. 1219–
The authors declare that there is no conflict of interests 1222, June 2007.
regarding the publication of this paper. [15] H. M. Asifa and S. R. Vaishnav, “Particle swarm optimisation
algorithm based PID controller,” in Proceedings of the 3rd
References International Conference on Emerging Trends in Engineering and
Technology (ICETET ’10), pp. 628–631, November 2010.
[1] W. Cui, Y. Gong, and M. H. Xu, “A permanent magnet brushless [16] C. Cao, X. Guo, and Y. Liu, “Research on ant colony neural
DC motor with bifilar winding for automotive engine cooling network PID controller and application,” in Proceedings of
8 Advances in Artificial Neural Systems

the 8th ACIS International Conference on Software Engineer-

ing, Artificial Intelligence, Networking, and Parallel/Distributed
Computing, pp. 253–258, August 2007.
[17] M. A. Sahib, B. S. Ahmed, and M. Y. Potrus, “Application
of combinatorial interaction design for dc servomotor pid
controller tuning,” Journal of Control Science and Engineering,
vol. 2014, Article ID 576868, 7 pages, 2014.
[18] Y.-C. Luo, Z.-S. Ke, and Y.-P. Kuo, “Sensorless rotor-field
oriented controlled inductiion motor drive with particle swarm
optimization algorithm speed controller design strategy ,”
Controller Design Strategy, vol. 2014, Article ID 861462, 13 pages,
2014.
[19] R. H. B. Richard and C. Dorf, Modern Control Systems, Pearson,
11th edition, 2008.
[20] A. N. A. K. Mishra, “Speed control of DC motor using par-
ticle swarm optimization technique,” International Journal of
Engineering Research and Technology, vol. 2, no. 6, pp. 1643–
1649, 2013.
[21] S. D. S. N. Sivanandam, Principles of Soft Computing, John Wiley
& Sons, New York, NY, USA, 2nd edition, 2013.
[22] J. Kennedy and R. Eberhart, “Particle swarm optimization,” in
Proceedings of the 1995 IEEE International Conference on Neural
Networks, pp. 1942–1948, Perth, Wash, USA, December 1995.
[23] J. H. Mathews and K. K. Fink, Numerical Methods Using
MATLAB, Prentice-Hall, 4th edition, 2004.
[24] Z.-L. Gaing, “A particle swarm optimization approach for
optimum design of PID controller in AVR system,” IEEE
Transactions on Energy Conversion, vol. 19, no. 2, pp. 384–391,
2004.
[25] C.-C. Wong, S.-A. Li, and H.-Y. Wang, “Optimal PID controller
design for AVR system,” Tamkang Journal of Science and
Engineering, vol. 12, no. 3, pp. 259–270, 2009.
 Advances in Journal of
Industrial Engineering
Multimedia
Applied
Computational
Intelligence and Soft
Computing
The Scientific International Journal of
Distributed
Hindawi Publishing Corporation
World Journal
Hindawi Publishing Corporation
Sensor Networks
Hindawi Publishing Corporation Hindawi Publishing Corporation Hindawi Publishing Corporation
http://www.hindawi.com Volume 2014 http://www.hindawi.com Volume 2014 http://www.hindawi.com Volume 2014 http://www.hindawi.com Volume 2014 http://www.hindawi.com Volume 2014

Advances in

Fuzzy
Systems
Modelling &
Simulation
in Engineering
Hindawi Publishing Corporation
Hindawi Publishing Corporation Volume 2014 http://www.hindawi.com Volume 2014
http://www.hindawi.com

Submit your manuscripts at

Journal of
http://www.hindawi.com
Computer Networks
and Communications  Advances in 
Artificial
Intelligence
Hindawi Publishing Corporation
http://www.hindawi.com Volume 2014 Hindawi Publishing Corporation
http://www.hindawi.com Volume 2014

International Journal of Advances in

Biomedical Imaging Artificial
Neural Systems

International Journal of
Advances in Computer Games Advances in
Computer Engineering Technology Software Engineering
Hindawi Publishing Corporation Hindawi Publishing Corporation Hindawi Publishing Corporation Hindawi Publishing Corporation Hindawi Publishing Corporation
http://www.hindawi.com Volume 2014 http://www.hindawi.com Volume 2014 http://www.hindawi.com Volume 2014 http://www.hindawi.com Volume 2014 http://www.hindawi.com Volume 2014

International Journal of
Reconfigurable
Computing

Advances in Computational Journal of

Journal of Human-Computer Intelligence and Electrical and Computer
Robotics
Hindawi Publishing Corporation
Interaction
Hindawi Publishing Corporation
Neuroscience
Hindawi Publishing Corporation Hindawi Publishing Corporation
Engineering
Hindawi Publishing Corporation
http://www.hindawi.com Volume 2014 http://www.hindawi.com Volume 2014 http://www.hindawi.com Volume 2014 http://www.hindawi.com Volume 2014 http://www.hindawi.com Volume 2014