A COMPARATIVE STUDY OF PID CONTROLLER TUNING

TECHNIQUES FOR TIME DELAY PROCESSES

The Proportional-Integral-Derivative (PID) controllers are used in process/plant

for controlling their parameters such as thermal or, electrical conductivity. By adjusting
three parameters of PID controller, both transient and steady response can be improved,
and better output can be obtained. There are many PID controller tuning techniques
available in the literature and designing PID controllers for small delay processes with
specified gain and phase margin is a well-known design technique. If the gain margin and
phase margin are not specified, the system may not be optimum. A system with large gain
and phase margins is more robust and gives better performance. When the system is robust,
there will be no effect of slight changes in system parameters on the system performance.
This paper describes a comparative analysis, among different types of tuning techniques
available for first order plus delay time systems (FOPDT) on the basis of the various time
integral performance criteria such as ISE, IAE, ITAE and gain margin and phase margin.
This review concludes that ‘Chien-Hrones-Reswick (CHR) no overshoot set point’ method
provides the best results.

Keywords: PID controller, tuning method, performance index, robustness

1. Introduction

Proportional-Integral-Derivative (PID) type controllers are widely used in

many industrial processes like temperature, pressure and flow control processes.
The PID controllers are very simple in structure and easy to tune. By adjusting
three parameters of a PID controller, both transient and steady response can be
improved, and better output can be obtained using it. These controllers are also
used in a wide range of applications such as flight control, motor drive,
instrumentation, etc.
If the parameters of the PID controller are tuned in the proper way, it
provides robust and reliable performance. Most of the industrial processes such as
chemical processes have a different kind of non-linearities, and inherent dead time
involved. The existence of nonlinearity in the process makes the controller tuning
difficult.
There are many tuning formulas available for the design of PID controllers
[1]–[3]. The first PID controller was developed by Minorsky [4] by which ship
steering was controlled and in 1934 he also developed the first tuning technique
for the PID controller of an integrating plus dead time process. In 1940
mechanical controls like hydraulic and pneumatic were used in industry [5]. The
advantage of these controllers is their ruggedness, but they are lower in response.
Introducing the electronic controllers, the systems became much faster
than in the case of mechanical control. The size of the electronics controller is less
because of the small size of the electronic components but, it has the disadvantage
that it is very sensitive to temperature. To overcome this problem, in 1980 a
flexible digital PID controller was integrated into a control system [6]. In present
days the complex control logic of electronic controllers is replaced by
programming [7]. The researchers are still trying to find out the best tuning
techniques for the PID controller. Several advancements have been done for PID
controller design such as adaptive PID control [8] automatic tuning PID control
[9] and intelligent PID control [10]. In an industrial application such as a medium
chemical plant, more than 98% of the controllers are of PID type because of its
previous successful track record [11]–[13].
In conventional PID control, the manual tuning technique is time-
consuming and it also depends upon process knowledge of the operator. The
comparison of several PID tuning rules is discussed in [14] and the survey of a
different patent for PID controller is presented in [15].
The design and process analysis of a model such as a first-order process
with a delay time (FOPDT) representing widely thermal and chemical processes
are given in [16]. The delay time also known as dead time occurs due to the
involvement of transport lag. The delay causes a decrease in the phase margin and
a more oscillatory closed-loop response. If the gain margin is decreased the
system will move towards instability [17]. Recently, numerous PID tuning
techniques have been floated in the literature in the diverse field [18]–[21].
The different types of tuning techniques of optimal design of PID
controllers applied to a FOPDT process have been considered in the present work
and the techniques are compared on performance basis i.e. rise time, overshoot,
settling time, Integral Square Error (ISE), Integral Absolute Error (IAE), Integral
Time Absolute Error (ITAE), etc. The effect of disturbance rejection and
instability introduced by the delay time has also been compared for different
tuning techniques. System robustness is compared on the basis of gain margin and
phase margin for different methods.
The contents of this paper have been organized as follows. Some well-
known, classical tuning techniques of the PID controller are discussed in Section
2. The result and discussion about the time domain specification by minimizing
performance index criteria are given in Section 3 and the final conclusion is given
in Section 4.

2. Classical Tuning Methods of PID Controller:

There are two methods of PID tuning: (i) open loop (ii) closed loop. The
block diagram of closed loop feedback PID control is shown in Fig 1. The open
loop technique is being used when controllers have manual state and plant works
in the open loop condition. The closed-loop method is used to tune the controller
during an automatic state when the plant operates in the closed loop condition.
There are several types of tuning techniques of the PID controller in present
control literature and most of the tuning technique can be applied to the first order
plus delay time (FOPDT) process. The FOPDT plant most commonly used in the
chemical process [22]. In this paper close loop method with following general
FOPDT process transfer function has been considered.

d(t)
e(t)
+ u(t)+ + y(t)
r(t) Controller Process
_ Gc(s) Gp(s)

Fig. 1. Classical feedback control structure

KesL
Gp (s) 
Ts 1
(1)
where K is the process gain, L is the delay time and T is the time constant
of the system. In the frequency domain the FOPDT system can be represented as:
Ke( jw)L
Gp ( jw)  (2)
T ( jw) 1
The transfer function of the general PID controller is:
 1 
GC (s)  K P 1   Td s 
 Ti s  (3)
The above transfer function can also be written is:
K
GC (s)  K P  i  K ds
s (4)
 Ziegler-Nichols Method (Z-N)
A very useful tuning method was presented by Ziegler and Nichols[23] in
1942. The Z-N method is of two types - one is based on the reaction curve of the
system and another one depends upon the ultimate gain Ku and the ultimate
period Tu. When the step response of the plant exhibits an S-shape curve with
zero overshoot then the reaction curve method is applied. The reaction curve is
characterized by two constant parameters - the delay time L and the time constant
T. The second method considers trial and error tuning which is based on
proportional gain KP; by increasing the value of KP output exhibit oscillation. This
rule works well only when the delay time is less than half the length of the time
constant. The disadvantage of this technique is that it is time-consuming and not
applicable for first and second order without time delay processes because some
process does not have an ultimate gain. This method performs well in disturbance
rejection, but it is poor in tracking reference change.
Cohen-Coon Method
A few sets of tuning rules for the PID controller were proposed by Cohen
and Coon [24]. This method is based on a process reaction curve and can be
applied to the first order plus delay time (FOPDT) system. It is well suited to a
process where the delay time is less than two times the length of the time constant.
Internal Model Control (IMC)
There are many methods of tuning IMC-PID but the first method was
developed by Morari et al. [25] which is called as an internal model controller.
This method works on model-based control technique. The advantage of this
approach is a good setpoint tracking, but sluggish disturbance response especially
when the process has small “delay time”/”time constant” ratio. This is not
desirable for industrial processes because, for many control process applications,
disturbance rejection is more important than set point tracking [26]. The general
block diagram of the IMC based approach is shown in Fig 2.
D(s)

+ U(s)
R(s) Controller Process ++ C(s)
_ Gc(s) Gp(s)

Process _+
Model
Gm(s)

Fig. 2: Basic block diagram of an internal model control

 In this diagram, the controller Gc(s) is used to control the process Gp(s)
and the process model Gm(s), D(s) is the unknown disturbance. The settings of
IMC- PID are given in [14] where λ is a closed loop time constant and the
controller works differently for different value of λ. Here λ=0.25 is selected for
the controller tuning.
Gain, Phase Margin Method (GP)
Gain margin and phase margin specifications are a well-known design
technique of PID controller. The gain and phase margin exhibit the robustness of
the system. This method was firstly proposed by Astrom [27] but it may not
produce a satisfactory performance with large time delay, which results in
oscillatory close loop response. Further Hangm et al. [28] proposed integral
performance criteria, which gives better-closed loop response but the
disadvantage of this method was that the transfer function has to be available. To
overcome this problem a new tuning formula was given by Zhuang and Atherton
[29] where both gain and phase margins are optimized.
Optimum PID Tuning for SetPoint Changes and Disturbance
Rejection
The shape of the system closed loop response from initial state t=0 to the
final state could be used to find the exact controller setting. These rules are used
to derive an optimum tuning value of PID controller parameters to minimize
Integral Square Error (ISE), Integral Absolute Error (IAE), Integral Time
Absolute Error (ITAE), Integral Square Time Error (ISTE), Integral Square Time
Square Error (IST2E) performance criteria. These tuning formulas were found by
Zhuang and Atherton [29]. IAE is good to eliminate small errors, but it produces a
slower response and it does not add weight to any of the errors in system
response. ISE will tolerate small errors for a long period of time, low amplitude
and oscillation. To overcome this disadvantage ITAE and ISTE performance
criteria are used which settle faster as compared to the previous two methods
because they integrate the absolute value of error multiplied by the time [30]. The
different performance criteria are given as
t
IAE   e(t) dt
0 (5)
t
ISE   e2 (t)dt
0 (6)
 t

ITAE   t e(t) dt
0
(7)
t
ISTE  t2e(t)dt
0 (8)
t
` IST 2E  t 2 e 2 (t)dt
0 (9)
Chien-Hrones-Reswick (CHR) PID Tuning
This method was proposed by Chien et al. [31], which focuses on the
important observation on set point (S.P) response and disturbance rejection. It is a
modified form of the open loop Z-N method and uses the quickest response with
0% overshoot or quickest response with 20% overshoot as a design criterion. The
CHR tuning formula based on 20% overshoot design criteria is quite same as the
Ziegler-Nichols tuning method. But when designing on the criteria of 0%
overshoots, the proportional gain KP and the derivative time Td both were smaller,
the integral time Ti was larger as compared to the Z-N method. It means all these
three parameters are smaller. Table 1 is used to get the controller parameter,
where T is the time constant, L is the delay time, K is the gain and the value of the
constant a=KL/T.
Wang-Juang-Chan (W.J.C) Tuning
This tuning formula is based on optimum ITAE Criterion. It was proposed
by Wang et al.[32]. It is a very simple and efficient method to find out PID
parameters. If the parameters of the plant are known i.e. K the plant gain, L the
time delay and T the time constant, the parameters of the controller are

0.5307T
(0.7303  )(T  0.5L)
Kp  L (10)
k(T  L)
Ti  T  0.5L (11)
0.5LT
Td  (12)
(T  0.5L)
2.8 Robust PID Controller
A robust controller deals with plant uncertainty. If there is a slight change
in gain K, time constant T and delay time L, the robust controller will provide
uncertainty and achieves robustness and stability. Parada et al. [33] proposed a
robust algorithm for designing PID parameter for the FOPDT system but this
technique is complex itself. A simple and effective technique based on the H∞
control theory for PID controller design was given by Liu and Wang [34]. The
performance of PID is determined by taking two different values of the tunable
parameter λ as 0.1 and 0.3 respectively.
1
Gc (s)  K p (1  Td s )


Ti s T f s 1
(13)
The PID parameters are
2
T f (s) 


L


2 
2 (14)
Ti
K p (s) 
k (2  L 2)
(15)
L
Ti  T   Tf
2
(16)
TL
Td   Tf
2Ti
(17)

3. Simulation Result Discussions

The following FOPDT process has been considered here [14], [35], [36]
1 0.5s
GP (s)  e
s 1 (18)
The different types of tuning techniques of the PID controller have been
implemented to FOPDT process with a unit step input. The step responses of the
Wang-Juang-Chan (WJC) Tuning, Chien-Hrones-Reswick (CHR) no overshoot
set point, robust H∞, optimum tuning technique, i.e. IAE set point and ISE set
point are shown in fig 3. It is clear, that the WJC tuning technique provides lower
rise time and lower settling time with zero overshoot.
136

Fig. 3: Step response with different PID controllers of the original system, when delay L=0.5

A comparison of controllers in time domain specifications taking rise time,

settling time and maximum overshoot using different PID tuning techniques, has
been shown in table 2. The values of KP, Ki & Kd using different PID tuning
methods, are also mentioned in table 2.
When the delay is increased by 50%, Table 2 provides the transient
behavior of the plant for the same tuning parameter of PID. It can be observed
from Fig 4 that CHR tuning technique for set point (S.P) with zero overshoot
provides higher rise time, but lower settling time and minimum overshoot as
compared to other techniques.

Fig. 4: Step response with different PID controllers when delay increased by 50% (L=0.75)
 The WJC technique is close to the robust controller, but it has lower rise
time, lower overshoot, as well as lower settling time as compared to optimum
tuning technique i.e. IAE, set point and ISE set point.
When the delay is decreased by 50%, Fig. 5 shows zero overshoot, but the
WJC technique is one of the best tuning techniques because it gives lower rise and
settling time as compared to other tuning techniques like robust controller, CHR
set point regulation with 0% and 20% overshoot and IAE set point.

Fig. 5: Step response with different PID controllers when delay decreased by 50% (L=0.25)

The comparative analysis of performance index criteria such as Integral

Square Error (ISE), Integral Absolute Value of the Error (IAE) and Integral of
Time-Weighted Absolute Error (ITAE) is shown in Table 3. For the original
process and when the delay time is increased and decreased by 20%, it can be
observed that by using CHR set point 0% overshoot design technique
performance index is minimum.
The comparison of controller tuning techniques in the frequency domain is
done in Table 4. where gain margin and phase margin play a vital role in
measuring the robustness and performance of the system. The higher the gain and
phase margin, the higher will be the robustness of the system [35]. The oscillatory
magnitude response of the controllers is shown in Fig. 6 and 7; it can be observed
that all controllers are stable.
As shown in fig 8, with 50% deviation in delay time L from their value
because of external disturbances and model uncertainty, the closed loop response
of CHR set point (S.P.) has no overshoot, is stable, whereas WJC, robust
controller, IAE set point has more overshoot resulting in an unstable system. It
also observed that CHR set point 0% overshoot has larger robustness.
Fig. 6 Bode plot of the original system when delay L=0.5
 Fig. 7 Bode plot when a delay is increased by 50% (L=0.75)

Fig. 8 Bode plot when a delay is decreased by 50% (L=0.25)

4. Conclusion

Numerous controllers are available in the literature for FOPDT processes

such as conventional PID controller and IMC-PID controller etc., with a number
of tuning techniques. In this paper, a comprehensive study of well-known tuning
techniques for the PID controller to control the FOPDT process has been carried
out.
out. Five tuning techniques have been applied to the FOPDT process with
different values of the delay process, and in order to identify the best tuning
technique among them, time and frequency responses are compared. By rigorous
review analysis, it is revealed that the CHR set point (0% overshoots) method
exhibit better performance comparatively. It is also observed that CHR set point
(0% overshoot) method minimizes Integral Absolute Error (IAE), Integral Square
Error (ISE) and Integral Time Absolute Error (ITAE).

REFERENCES

[1] K. J. Åström and T. Hägglund, PID controllers : theory, design, and tuning / Karl J.
Astrom and Tore Hagglund. 2018.
[2] A. Karimi, D. Garcia, and R. Longchamp, “PID controller tuning using Bode’s integrals,”
IEEE Trans. Control Syst. Technol., vol. 11, no. 6, pp. 812–821, 2003.
[3] D. J. Wang, “Synthesis of PID controllers for high-order plants with time-delay,” J.
Process Control, vol. 19, no. 10, pp. 1763–1768, 2009.
[4] N. Minorsky., “Directional stability of automatically steered bodies,” J. Am. Soc. Nav.
Eng., vol. 34, no. 2, pp. 280–309, 1922.
[5] S. Bennett, “Development of the PID Controller,” IEEE Control Syst., vol. 13, no. 6, pp.
58–62, 1993.
[6] I. D. Landau and G. Zito, Digital Control Systems, 1st–2006th ed. Springer-Verlag
London, 2006.
[7] P. Rusia and S. Bhongade, “Design and implementation of digital PID controller using
FPGA for precision temperature control,” in 2014 6th IEEE Power India International
Conference (PIICON), 2014, pp. 1–4.
[8] H. Zhang, Y. Shi, and A. Saadat Mehr, “Robust static output feedback control and remote
PID design for networked motor systems,” in IEEE Transactions on Industrial Electronics,
2011, vol. 58, no. 12, pp. 5396–5405.
[9] J. G. Williams, G. Liu, S. Chai, and D. Rees, “Intelligent control for improvements in
PEM fuel cell flow performance,” Int. J. Autom. Comput., vol. 5, no. 2, pp. 145–151,
2008.
[10] J. Ahn, S. Cho, and D. H. Chung, “Analysis of energy and control efficiencies of fuzzy
logic and artificial neural network technologies in the heating energy supply system
responding to the changes of user demands,” Appl. Energy, vol. 190, pp. 222–231, 2017.
[11] L. Desborough and R. Miller, “Increasing customer value of industrial control
performance monitoring-Honeywell’s experience,” AIChE Symp. Ser., no. Figure 1, pp.
169--189, 2002.
[12] A. Vilanova, Ramon, Visioli, PID Control in the Third Millennium, 2012th ed. Springer,
London, 2012.
[13] N. A. Selamat, N. A. Wahab, and S. Sahlan, “Particle Swarm Optimization for
Multivariable PID Controller Tuning,” 2013 Ieee 9th Int. Colloq. Signal Process. Its Appl.,
pp. 170–175, 2013.
[14] W. Tan, J. Liu, T. Chen, and H. J. Marquez, “Comparison of some well-known PID tuning
formulas,” Comput. Chem. Eng., vol. 30, no. 9, pp. 1416–1423, 2006.
[15] Y. Li, G. C. . Chong, and K. H. Ang, “Patents, Software, and Hardware for PID Control:
 A comparative study of PID controller tuning techniques for time delay processes 141

An Overview and Analysis of the Current Art,” IEEE Control Syst. Mag., vol. 26, no. 1,
pp. 42–54, 2006.
[16] P. Chalupa, M. B. A. Ř. Inová, J. Novák, and M. Beneš, “Control System for Chemical
Thermal Processes and Its Usage for Measurement of Collagen Shrinkage Temperature,”
vol. 10, pp. 445–452, 2015.
[17] S. S. Mikhalevich, S. A. Baydali, and F. Manenti, “Development of a tunable method for
PID controllers to achieve the desired phase margin,” J. Process Control, vol. 25, pp. 28–
34, 2015.
[18] R. Bhatt, G. Parmar, R. Gupta, and A. Sikander, “Application of stochastic fractal search
in approximation and control of LTI systems,” Microsyst. Technol., vol. 25, no. 1, pp.
105–114, 2019.
[19] S. Uniyal and A. Sikander, “A Novel Design Technique for Brushless DC Motor in
Wireless Medical Applications,” Wirel. Pers. Commun., vol. 102, no. 1, pp. 369–381,
2018.
[20] J. Agarwal, G. Parmar, R. Gupta, and A. Sikander, “Analysis of grey wolf optimizer based
fractional order PID controller in speed control of DC motor,” Microsyst. Technol., pp. 1–
10, 2018.
[21] I. Uniyal and A. Sikander, “A Comparative Analysis of PID Controller Design for AVR
Based on Optimization Techniques,” in Intelligent Communication, Control and Devices,
Springer, 2018, pp. 1315–1323.
[22] M. Shamsuzzoha and M. Lee, “IMC-PID Controller Design for Improved Disturbance
Rejection of Time-Delayed Processes,” Ind. Eng. Chem. Res., vol. 46, no. 7, pp. 2077–
2091, 2007.
[23] J. G. Ziegler and N. B. Nichols, “Optimum settings for automatic controllers,” InTech, vol.
42, no. 6, pp. 94–100, 1995.
[24] G. H. COHEN, “Theoretical Consideration of Retarded Control,” Trans. ASME, vol. 75,
pp. 827–834, 1953.
[25] D. Rivera, M. Morari, and S. Skogestad, “Internal model control: PID controller design,”
… Chem. Process Des. …, pp. 252–265, 1986.
[26] Y. Lee, J. Lee, and S. Park, “PID controller tuning for integrating and unstable processes
with time delay,” Chem. Eng. Sci., vol. 55, no. 17, pp. 3481–3493, 2000.
[27] K. J. Åström and T. Hägglund, “Automatic tuning of simple regulators with specifications
on phase and amplitude margins,” Automatica, vol. 20, no. 5, pp. 645–651, 1984.
[28] C. C. Hang, K. J. Åström, and W. K. Ho, “Refinements of the Ziegler–Nichols tuning
formula,” IEE Proc. D Control Theory Appl., vol. 138, no. 2, p. 111, 1991.
[29] M. Zhuang and D. P. Atherton, “Automatic tuning of optimum PID controllers,” Control
Theory Appl. IEE Proc. D, vol. 140, no. 3, pp. 216–224, 1993.
[30] K. Rahimi and P. Famouri, “Assessment of Automatic Generation Control performance
index criteria,” in 2014 IEEE PES T&D Conference and Exposition, 2014, pp. 1–5.
[31] J. A. H. and J. B. R. K. L. Chien, “On the Automatic Control of Generalized Passive
Systems,” Trans. Am. Soc. Mech. Engineeing, vol. 74, pp. 175–185, 1972.
[32] F. S. Wang, W. S. Juang, and C. T. Chan, “Optimal tuning of pid controllers for single and
cascade control loops,” Chem. Eng. Commun., vol. 132, no. 1, pp. 15–34, 1995.
[33] M. Parada, D. Sbarbaro, R. A. Borges, and P. L. D. Peres, “Robust PI and PID design for
first- and second-order processes with zeros, time-delay and structured uncertainties,” Int.
J. Syst. Sci., vol. 48, no. 1, pp. 95–106, 2017.
[34] L. Liu, Q. Wang, X. Fu, and X. Liu, “Water temperature control of an electrically heated
boiler based on robust PID,” in Proceedings - 2013 International Conference on
Mechatronic Sciences, Electric Engineering and Computer, MEC 2013, 2013, pp. 201–
205.
142 Nafees Ahamad, Shailesh Uniyal, Afzal Sikander, Gagan Singh

[35] W. K. Ho, K. W. Lim, C. C. Hang, and L. Y. Ni, “Getting more phase margin and
performance out of PID controllers,” Automatica, vol. 35, no. 9, pp. 1579–1585, 1999.
[36] Q. B. Jin and L. Y. Liu, “Design of a robust internal model control PID controller based on
linear quadratic gaussian tuning strategy,” Can. J. Chem. Eng., vol. 92, no. 7, pp. 1260–
1270, 2014.