Asian Journal of Convergence in Technology Volume V Issue I

ISSN NO: 2350-1146 I.F-5.11

A Review on Tuning Methods for PID Controller

Sangram Bharat, Arunangshu Ganguly, Rohit Chatterjee, Student Member, IEI, Biswajit Basak, Debkumar Sheet and
Anirban Ganguly, Student Member, IEEE

Abstract— PID controller tuning is important for having A. PID controller basic
minimal overshoot in steady state response and lesser settling time. A PID (proportional- integral- derivative) controller has a
Apart from two conventional methods of tuning such as Ziegler
control loop feedback mechanism. It continually calculates an
Nichols closed loop oscillation and Cohen-Coon’s process reaction
curve method there are other methods such as fuzzy set-point weight
error e(t) which is the difference between the process variable
(FSPW) tuning and genetic algorithm (GA) are drawing the interest and set point and then applies a correction based on
of researchers. This reports a survey on works related to PID proportional, integral and derivative action. That means it
controller tuning methods using fuzzy-logic and GA. A comparison automatically applies accurate and responsive correction to a
is made between conventional two methods and fuzzy-logic based control function. For example in cruise control of a motor
tuning methods through simulation using MATLAB Simulink. where external influences such as increase of mechanical load
would decrease the speed of motor, then a PID algorithm
increases the speed of motor in an optimal way by minimizing
Keywords— PID controller; Ziegler Nichol’s (ZN) method; the overshoot and by controlling input current or power of the
process reaction curve; FSPW; GA; Simulink. motor.

I. INTRODUCTION
Feedback control systems have made their strong
establishment in the field of industrial manufacturing, process
control and robotics since the last three decades. Many robust and
adaptive control mechanisms [1]-[3] have been proposed in recent
past. Still the popularity of PID controller in feedback control
system is unmatched. This control scheme can be used in
proportional only (P-mode), proportional – integral (PI mode),
proportional- derivative (PD) and proportional-integral-derivative
(PID mode) as per the process requirement [4]. A number of
works [5]-[7] reported for use of standard PID controller along
with neuro-fuzzy controllers. Fig 1 Block diagram of fundamental operation of PID
The area of controller tuning is an old and important Controller
domain of research. A significant amount of papers these days
reports the use of fuzzy logic and more especially genetic The above block diagram depicts the fundamental operation of
algorithm (GA) in PID controller tuning. Therefore, a PID controller, here r(t) is the reference input signal and e(t) is
literature survey in this field becomes a necessary to proceed the error signal. The control action is produced by
further for in depth research work. proportional, differentiator and integrator operating in parallel
The organization of the paper is such that section II gives a brief on the error signal. The controller attempts to minimize the
overview on PID control algorithm and two conventional PID error over time by adjustment of control variable u(t).
controller tuning. Section III provides a literature survey on PID The mathematical form of PID control action is given by
controller tuning using fuzzy logic and genetic algorithm (GA).
Section IV makes a comparison between three types of controller 1
t
de(t )
tuning mechanism through MATLAB simulation. Finally section
V concludes the paper.
u (t )  K P (e(t )  
Ti 0
e(t )dt  Td
dt
) (1)

II. BACKGROUND THEORY

B. PID controller tuning
Selection of the value of P,I and D parameter is dependent on
process, so the knowledge of plant dynamics is important for

www.asianssr.org 1
Asian Journal of Convergence in Technology Volume V Issue I
ISSN NO: 2350-1146 I.F-5.11

selection of parameters. It is quite difficult to obtain the methodology says the first method is a kind of trial-and-error
mathematical model of the plant. So we have to find method so it is highly dependent on operator’s skill. The
experimentally the optimum settings of controller in which it second method i.e. method of PRC can sometimes be
is possible experimentally find the optimum values of the misleading as disturbances may cause misleading results.
controller parameter which is known as tuning. There are two Alternatively, in recent past two methods of tuning namely
widely used methods for controller tuning: 1. Ziegler Nichol’s fuzzy set-point weight (FSPW) tuning and tuning using
continuous cycle method [8] 2.Cohen Coon’s process reaction genetic algorithm has been reported in some literature. Genetic
curve method [9]. algorithm based tuning has been emerged as more competitive
Ziegler Nichols closed loop oscillation method (Continuous these days.
Cyclic Method) – This method is used to tune the controller in Fuzzy logic based tuning method is proposed by A. Visioli in
[10]. As we know, fuzzy logic converts the user supplied
closed loop mode. The control action is first set to proportional
human rules into their mathematical equivalents. The fuzzy
mode where then gain Kp is set to Ku (large possible value of Kp)
logic contains number of if-then statement. In fuzzy logic, the
for which closed oscillation are allowed to occur around the set- truth of any statement is always a matter of degree. The
point. The time period of oscillation (T u) is taken into account. system depends on a function, which is known as the
The values for optimum Kp, Ti and Td are shown in table I. membership function, which helps to calculate the correct
The method of Process reaction curve is an open loop method value between 0 and 1. In other words, we can say that a
of tuning. Here a closed loop system is first broken and then a membership function represents the degree of truth in the
fuzzy logic. As proposed by Visioli, this method is referred to
step input is given to the process. An overlap of output
as fuzzy set-point weight tuning. The control law can be
response and input is plotted in the graph as shown in Fig 2. written as

u  t   K p (b  t  ysp  t   y  t  )
t
de(t )
 Kd  K i  e( )d (2)
dt 0

Fig 2. Process reaction curve method

Fig 3. Fuzzy set-point weight tuning [10]

In Fig 2. M, L and K is measured; N=slope of the curve, Time Where b (t )  w  f (t)

constant (T) = k/N; Lag ratio=L/T. The optimum settings of
Kp, Kd, Ki, are the gains for the proportional, differential,
PID are mentioned in table I.
integrator respectively. On the expression of b(t), w is a fixed
TABLE I. OPTIMUM SETTINGS FOR PID CONTROLLER positive parameter, whose value is in between 0 and 1. The
PARAMETERS function f(t) is the output of the fuzzy inference system. The
fuzzy inference system for this case consist of five triangular
Controller Ziegler Nichols Cohen-Coon’s membership functions for the two inputs (one is the error input,
settings closed loop process reaction other one is the time derivative of the error input), and nine
oscillation curve
method
triangular membership function for the output.
KP 0.6Ku M/NL(4/3+R/4)
G. Mallesham and A. Rajani described the way by which the
Ti Tu/2 L(32/13+6R/20R) optimum response of a system is obtained by fuzzy logic
controllers by using fuzzy set point weighting method in [11].
Td Tu/8 L(4/11+2R) Here, a comparison has also been made with the fuzzy set
point controllers with the conventional methods and with the
different shapes and numbers of designed membership
function.
III LITURATURE SURVEY Genetic algorithm has extensively been used as a tuning
There has been several works reported for the PID controller methodology. The genetic algorithm is also used in fault
tuning using Ziegler-Nichols method and the process reaction analysis, stability analysis, robot path planning’s etc. It is a
curve methods as discussed in the previous section. As the global, parallel, search and optimization method. In this

www.asianssr.org 2
Asian Journal of Convergence in Technology Volume V Issue I
ISSN NO: 2350-1146 I.F-5.11

method, for a particular problem, a set of potential solution is

made, that is known as the population. Every individual in
these populations represents a potential solution. Every
individual, those are the solution for a particular problem is
assigned a value, which is known as the fitness value, and for
every fitness value in the genetic algorithm domain, there
corresponds a cost value in the problem domain. An explicit
mapping is done between the two domains. There are
operators present in genetic algorithm, which makes the
reproductions of the individual in the population. Two main
operators are single point crossover operator and the binary Fig 4. FSPW subsystem in Simulink
mutation operator.
M. J. Neath, A. K. Swain, U. K. Madawala and D. J.
Thrimawithana described that in a normal traditional PID tuning The performance of the second order system with three types
method, there are many difficulties, and to overcome those, the of PID controller tuning method is depicted in table II.
paper proposes a derivative free optimization technique, which is
TABLE II. PERFORMANCE ANALYSIS OF PID CONTROLLER
based on genetic algorithm to determine the optimal parameters
of PID controllers used in bidirectional inductive power transfer Factors Continuous Process Fuzzy logic
systems in [12]. B. Porter and A.H. Jones, described the technique cycle reaction [10]
of genetic algorithm as an alternative means for tuning of digital method [9] curve [8]
PID controllers in their work [13]. K. Valarmathi, D. Devaraj rise 1.1 1.9 1.4
time
and T. Radhakrishnan [14] described the enhanced genetic
tr
algorithm used in tuning the PI controller for the pH control
process to avoid premature convergence and to reduce delay 1.13 1.47 1.27
computation time. Here a comparison is also made between the time
performances of the proposed genetic algorithm tuning method td
with the traditional Ziegler Nichols tuning. D. Devaraj and B. peak 10 7.5 9
Selvabala described in [15], the real tuning of PID controllers in Overshoot
automatic voltage regulator system. Here, real coded genetic Mp (%)
algorithm and sugeno fuzzy logic approach is used for obtaining
the optimal gains of PID controllers. Here, blend crossover and ITAE 0.16 0.05 0.07
uniform mutation operator is used for the real variables in the ITSE 0.02 0.00 0.00
genetic population.


IV RESULTS AND DISCUSSION V. CONCLUSION

Comparative analysis between Ziegler-Nichol’s method, PRC Proper tuning of PID controller is found to be a necessary in
method and FSPW method for PID controller tuning has been any industry. With incorrect or no tuning a plant cannot run
considered for the second order process efficiently. There are several tuning methodologies which
mentioned below. require significant knowledge about the process can
sometimes be very challenging. Methods of ZN as in [8] and
1
G(s)  PRC [9] are not always efficient as the former is a kind of
s  2s  1
2
trial-and-error method and the later can give misleading
results while the disturbance in a process in high. So, a
demand of more efficient tuning methodology has forced the
researchers to use methods of fuzzy logic based tuning as in
[10] and methods of genetic algorithm. A comparison between
The simulation has been performed using MATLAB Simulink. ZN continuous cycle method, Cohen-Coon’s PRC method and
The Simulink model for FSPW tuning is shown in Fig. 4. FSPW [10] method (with 7 membership function for both
error and error rate) have been made. The comparison shows
The comparative results of step response for three types of
tuning methods are shown in the fig 5. FSPW method has reduced overshoot than ZN method and
least rise time but greater ITAE than PRC method.
The area of GA based tuning is not considered for simulation
at present but it is a part of our literature and future works of
GA based tuning can be associated with fuzzy logic based
methods also.

Fig 5. Step response of second order system

www.asianssr.org 3
Asian Journal of Convergence in Technology Volume V Issue I
ISSN NO: 2350-1146 I.F-5.11

Acknowledgment [14] K. Valarmathi, D. Devaraj & T. Radhakrishnan, “ Adaptive Enhanced

Genetic Algorithm-Based Proportional Integral Controller Tuning for pH
The authors would like to express gratitude to department Process,” in Instrumentation Science & Technology - INSTRUM SCI
TECHNOL, vol. 35, no. 6, pp. 619-635, 2007.
of electrical engineering of Hooghly Engineering &
Technology College, Hooghly, India. [15] D. Devaraj & B. Selvabala, “Real-coded genetic algorithm and fuzzy
logic approach for real-time tuning of proportional-integral - derivative
controller in automatic voltage regulator system,” in IET Generation,
Transmission & Distribution, vol. 3, no. 7, pp. 641 – 649, 2009.
References
[1] J. Kaloust and Z. Qu, "Robust control design for nonlinear uncertain
systems with an unknown time-varying control direction," in IEEE
Transactions on Automatic Control, vol. 42, no. 3, pp. 393-399, March
1997.
[2] G. Grimm, M. J. Messina, S. E. Tuna and A. R. Teel, "Nominally Robust
Model Predictive Control With State Constraints," in IEEE Transactions
on Automatic Control, vol. 52, no. 10, pp. 1856-1870, Oct. 2007.
[3] D. A. Linkens, M. Mahfouf and M. Abbod, "Self-adaptive and self-
organising control applied to nonlinear multivariable anaesthesia: a
comparative model-based study," in IEE Proceedings D - Control
Theory and Applications, vol. 139, no. 4, pp. 381-394, July 1992.
[4] Lipták Béla G., Process control. Boca Raton, FL: CRC Press, 2000.
[5] R. Ramachandran, B. Madasamy, V. Veerasamy and L. Saravanan,
"Load frequency control of a dynamic interconnected power system
using generalised Hopfield neural network based self-adaptive PID
controller," in IET Generation, Transmission & Distribution, vol. 12, no.
21, pp. 5713-5722, 27 11 2018.
[6] E. Oliveira Freire, F. G. Rossomando and C. M. Soria, "Self-tuning of a
Neuro-Adaptive PID Controller for a SCARA Robot Based on Neural
Network," in IEEE Latin America Transactions, vol. 16, no. 5, pp. 1364-
1374, May 2018.
[7] A. Dehghani and H. Khodadadi, "Designing a neuro-fuzzy PID
controller based on smith predictor for heating system," 2017 17th
International Conference on Control, Automation and Systems (ICCAS),
Jeju, 2017, pp. 15-20.
[8] J.G. Ziegler and N. B. Nichols , “Optimum settings for automatic
controllers,” Transactions of the ASME, vol. 64, pp. 759–768, 1942.
[9] G.H. Cohen and G.A. Coon, Theoretical Consideration of Retarded
Control, Trans. ASME, 75, pp. 827-834, 1953.
[10] A. Visioli, "Fuzzy logic based set-point weight tuning of PID
controllers," in IEEE Transactions on Systems, Man, and Cybernetics -
Part A: Systems and Humans, vol. 29, no. 6, pp. 587-592, Nov. 1999.
[11] G. Mallesham and A. Rajani, “Automatic Tuning of PID Controller
using Fuzzy Logic,” in proc. of 8th International Conference on
Development and Application Systems, Suceava, Romania, May , 2006.
[12] M. J. Neath, A. K. Swain, U. K. Madawala and D. J. Thrimawithana,
"An Optimal PID Controller for a Bidirectional Inductive Power
Transfer System Using Multiobjective Genetic Algorithm," in IEEE
Transactions on Power Electronics, vol. 29, no. 3, pp. 1523-1531,
March 2014
[13] Porter, Brenda & A.H. Jones, “Genetic tuning of digital PID
controllers,” in Electronics Letters, vol. 28, no. 9, pp. 843 – 844, 1992.

www.asianssr.org 4