You are on page 1of 7

G.Glan Devadhas et. al.

/ International Journal of Engineering Science and Technology


Vol. 2(10), 2010, 5831-5837

ROBUST TEMPERATURE
CONTROLLER DESIGN FOR A
CHEMICAL PROCESS
G.Glan Devadhas1
1
Research scholar
PRIST University
Thanjavur

Dr.S.Pushpa Kumar2
2
Former Principal,
Govt. College of Engineering
Kannur, Trivandrum

Abstract—This paper attempts to tuning out a new PID control strategy to provide Robust Control for a
Chemical process. Chemical process control is a challenging problem due to the strong on-line non-linearity and
extreme sensitivity to disturbances of the process. The proposed method has the advantage that it takes into
account all the parameters variations associated with the process. The variations in the process parameters are
modeled as a gaussian noise and an adaptive gaussian filter is placed in the feedback path. The adaptive
gaussian filter in the feedback path adapts its filter coefficients based on a kalman estimation algorithm. This
adaptive filter adapts so as to maintain the mean square error a minimum. The LQG (Linear Quadratic
Gaussian) in Robust Control is used in designing of the proposed strategy. The analysis of a PID tuning [7]
strategy and the necessity of such an adaptive strategy is also explored in this paper. The proposed strategy of
Robust Control has been designed for a First Order Lag Plus Delay (FOLPD) process. The proposed strategy of
Robust Control has been simulated for an FOLPD process in SIMULINK.

Key words—Robust controller, Nonlinear Process, LQG, PID, Industrial Process

1. Introduction.
PID controllers have dominated the process control industry over the decades owing to its associated
simplicity and easiness in implementation. The needs for better control strategies for process control in order to
achieve better performance have always motivated research interests. The design of PID controllers, tuning
involves selecting the amounts of Proportional, Integral and Derivative components required at the output of the
controller. Since the design of PID controllers involves obtaining the P, I and D components there always occur a
compromise in the design. The design of the optimum values for the PID controller parameters [7,8] has always
been challenging. Many new tuning techniques have been developed for the design of PID controllers, however
then still exists a scope for better tuning method.
Alternatives for PID control have led to better-advanced control strategies. The fuzzy controllers and neural
controllers are some of the result of this [10,11,12]. However the combination of these different control strategies
in process control is still to be explored. The combination of the control strategies in controlling a DC motor is
explored in literature. In the present work a combination of the control strategies in the control of temperature in a
CSTR is explored. A comparative study of the PID control strategies are tried out to control temperature is
studied. The different control strategies are studied individually and also in combination, in the control process.
The robust control of the CSTR process is explored in this work. The PID controllers used in the control are tuned
using the conventional tuning methods of Zeigler- Nichols tuning and Cohen- Coon tuning[7,8]. A mathematical
model of the CSTR is developed and this is used in the study of the control strategy of the process.
2.Conventional Methods.

Yun Li et al [2006] describes the three-term functionality offering treatment of both transient and steady state
responses, proportional-integral-derivative (PID) control provides a generic and efficient solution to real world
control problems. It also presents remedies for problems involving the integral and derivative terms. PID design
objectives, methods, and future directions are discussed. Subsequently, a computerized, simulation-based
approach is presented, together with illustrative design results for first-order, higher order, and nonlinear plants.

ISSN: 0975-5462 5831


G.Glan Devadhas et. al. / International Journal of Engineering Science and Technology
Vol. 2(10), 2010, 5831-5837

Oscar Montiel et al [2007] presenting the usefulness of an innovative method called Simple Tuning Algorithm
(STA) for tuning fuzzy controllers, it has only one variable to adjust to achieve the tuning goal, this in counterpart
to other methods like the Proportional Integral Derivative (PID) controller wish has three variables to adjust for
the same goal. Comparative examples of the STA and the PID methods are presented in a speed control of a real
DC gear motor application. The PID controller was tuned using the Ziegler-Nichols method.
Xiao-Fen et al [2007] proposed a method to improve the control quality of conventional PID controller used
in complex control system, a new design strategy of fuzzy-PID controller is proposed by utilizing the advantages
of both fuzzy and auto tuning PID controls and their mutual compensation.
V.Tipsuwanporn et al [2004] designed an industrial controller with the neuro-fuzzy model, based on Sugeno-
type fuzzy inference. The designed controller is a nonlinear system, which uses the relation between input and
output data. controls, are used to evaluate the efficiency of our controller.
Toru Yamamoto et al [2007] discuss the challenge to manufacture high quality products it is necessary to
regularly monitor performance of control loops that regulate the quality variables of interest. It also describes a
scheme of performance-adaptive PID controllers, which are based on the above control strategy.
For the further development APC algorithm of DCS the design of Neural-Net Controller for Nonlinear Plant
is discussed by Zhao Yingkai et al [6]. First, a neural-networks model is obtained for the nonlinear plants. Next,
the GPC algorithm based on the liberalized model is used to synthesize a linear controller for each operating
point.
3.PID Control System Analysis and Design
The transfer function of a PID controller is often expressed in the ideal form
GPID(s) = U(s)/E(s) = KP (1 + 1/ TIs + TDs )------------------------------- (1)
Where U(s) is the control signal acting on the error signal E(s), KP is the proportional gain, TI is the integral
time constant, TD is the derivative time constant, and s is the argument of the Laplace transform.
A PID controller can be considered as an extreme form of a phase lead-lag compensator with one pole at the
origin and the other at infinity. Similarly, its cousins, the PI and the PD controllers, can also be regarded as
extreme forms of phase lag and phase-lead compensators, respectively. However, the message that the derivative
term improves transient response and stability is often wrongly expounded. Practitioners have found that the
derivative term can degrade stability when there exists a transport delay.
While matters concerning commissioning and maintenance (such as pre- and post processing as well as fault
tolerance) also need to be considered in a complete PID design, controller parameters are usually tuned so that the
closed-loop system meets the following five objectives:
1) Stability and stability robustness, usually measured in the frequency domain
2) Transient response, including rise time, overshoot, and settling time
3) Steady-state accuracy
4) Disturbance attenuation and robustness against environmental uncertainty, often at steady state
5) Robustness against plant modeling uncertainty, usually measured in the frequency domain.
Most methods target one objective or a weighted composite of the objectives listed above.
PID is a generally applicable control technique that derives its success from simple and easy-to-understand
operation. However, because of limited information exchange and problem analysis, there remain
misunderstandings between academia and industry concerning PID control. For example, the message that
increasing the derivative gain leads to improved transient response and stability is often wrongly expounded.
These misconceptions may explain why the argument exists that academically proposed PID tuning rules
sometimes do not work well on industrial controllers. In practice, therefore, switching between different
structures and functional modes is used to optimize transient response and meet multiple objectives.
In tracking PID problems, it is desirable to use standard PID structures for a reasonable range of plant types
and operations. Modularization around standard PID structures should also help improve the cost
effectiveness of PID control and maintenance. This way, robustly optimal design methods such as PID easy
can be developed. By including system identification techniques, the entire PID design and tuning process
can be automated and modular code blocks can be made available for timely application and real-time
adaptation.

ISSN: 0975-5462 5832


G.Glan Devadhas et. al. / International Journal of Engineering Science and Technology
Vol. 2(10), 2010, 5831-5837

4. System Description
A chemical system common to many chemical processing plants, known as a continuous stirred tank reactor
(CSTR), was utilised as a suitable test for, TSK Fuzzy control, ANFIS control and PID control. It suffices to
know that within the CSTR two chemicals are mixed, and react to produce a product compound with
concentration Ca(t). The temperature of the mixture is Ԃ(t). A schematic representation of the system is
shown in Fig. 1.The reaction is exothermic, producing heat which acts to slow the reaction down. By
introducing a coolant flow rate qc(t), the temperature can be varied and hence the product concentration
controlled. This system can be described by the following nonlinear simultaneous differential equations1
which effectively combine the laws of chemical reaction and thermodynamics:
˙Ca(t) =Q(Ca0− Ca(t))/V− k0Ca(t)e−E/RԂ(t) …………………………………….(2)
˙Ԃ(t) = Q(Ԃ0 − Ԃ(t))/V+ k1Ca(t)e−E/RԂ(t)+k2qc(t) (1 − e−k3/qc (t)) (Ԃc0− Ԃ(t))…….. (3)

Fig. 1. CSTRprocess

Table:1
The CSTR parameters

Parameter Description Nominal


value
Q Process flow rate 100 l/min
V Reactor volume 100 l
k0 Reaction rate 7.2 × 1010
constant 1/min
E/R Activation energy 1 × 104 K
Ԃ0 Feed temperature 350K
ԂC0 Inlet coolant 350K
temperature
ΔH Heat of reaction −2 × 105
cal/mol
Cp, Cpc Specific heats 1 cal/gK
ρ,ρc Liquid densities 1 × 103 g/l
Ca0 Inlet feed 1mol/l
concentration
ha Heat transfer 7 × 105 cal
coefficient

ISSN: 0975-5462 5833


G.Glan Devadhas et. al. / International Journal of Engineering Science and Technology
Vol. 2(10), 2010, 5831-5837

Consider the flow rate qc(t) as the input and product concentration Ca(t) as the output of the system. As
seen from Fig. 2, the gain and damping of the system, vary widely over the whole operating region,
from 0.08 mol/l to 0.13mol/l

Fig. 2. Open-loop step response of the CSTR

5.Design and performance of a robust PID Controller.

The section describes a robust PID controller design for a first order process with dead time(FODT process).
Most over damped processes can be sufficiently well approximated by a first-order system with a time-lag
element as follows:
G(s) = K e-Ls
1+T
where K, T and L denotes the system gain, the time-constant and the time-lag respectively.
Equation(3) is the model of temperature variations of CSTR for the change in concentration of the solution.It
can be approximated in to a FODT process by the following transfer function
G(s) = 0.5 e-45s
1 + 100s
The discrete-time model corresponding to the above equation is considered, where the sampling interval is Ts
= 10.0[s], and this system is disturbed by Gaussian white noise with mean and variance 0.001.

6.Scheme Used for the Implementation of Proposed Method

The method proposed in this paper is experimented on a first order process with dead time (FODT process).
This is justifiable since over 80% of the industrial processes can be approximated as an FODT process. The
process performance parameters have also been shown in the result. In the present work we also try our possible
control

ISSN: 0975-5462 5834


G.Glan Devadhas et. al. / International Journal of Engineering Science and Technology
Vol. 2(10), 2010, 5831-5837

R(s)
Y(s)
CONTROLLER PROCESS
1

C(s)

CONTROLLER
2
N(s)

Figure. 3. Schematic Diagram for the proposed method

strategy on the same process and compare the results thus obtained with the results outlined in the paper. The
combination of PID and fuzzy controller is expected to give better results.

1.8

1.6

1.4

1.2

0.8

0.6

0.4

0.2

0
0 50 100 150 200 250 300 350 400 450 500
Figure. 4. Simulated result of an FODT process with LQG filter in the feedback path
1.6

1.4

1.2

0.8

0.6

0.4

0.2

-0.2
0 50 100 150 200 250 300 350 400 450 500

Figure. 5. Simulated result of an FODT process without LQG filter in the feedback path
1.5

0.5

-0.5

-1
0 50 100 150 200 250 300 350 400 450 500

Figure. 6. Error signal to the PID controller with noise


0.5

0.45

0.4

0.35

0.3

0.25

0.2

0.15

0.1

0.05

0
0 50 100 150 200 250 300 350 400 450 500

Figure . 7. Open loop response of the FODT for Cohen-coon tuning the PID controller

ISSN: 0975-5462 5835


G.Glan Devadhas et. al. / International Journal of Engineering Science and Technology
Vol. 2(10), 2010, 5831-5837

0.4

0.3

0.2

0.1

-0.1

-0.2

-0.3

-0.4
0 50 100 150 200 250 300 350 400 450 500

Figure. 8. Noise modeled as a result of the parameter variations

1.5

0.5

-0.5
0 50 100 150 200 250 300 350 400 450 500

Figure. 9. Feedback signal with noise added(noise occurring as a result of parameter variations)
Pxx - X Power Spectral Density
-1
10

-2
10

-3
10

-4
10

-5
10

-6
10

-7
10

-8
10
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Frequency

Figure. 10. Spectrum output with LQG, which reveals the reduction in harmonics
Pxx - X Power Spectral Density
-1
10

-2
10

-3
10

-4
10

-5
10

-6
10
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Frequency

Figure 11.Spectrum of output without LQG

ISSN: 0975-5462 5836


G.Glan Devadhas et. al. / International Journal of Engineering Science and Technology
Vol. 2(10), 2010, 5831-5837

7. Results and Discussions.

The optimum values of kp,Ki&Kd is established through the cohen-coon tuning method is obtained from the
sigmoid curve shown in fig.7
Fig.8 shows the band limited white noise modeled as a measure of parameter variations. Apart to noise drifts
,band limited white noise is a symbolic of probable noise inference also.
Disturbance at the output end of process is usually caused as perturbations are getting feed back to. means
noise is overlapping the feedback signal. The feedback signal thus obtained ,with noise signal added when
examined is as given in Fig.9
Fig .5 shows the simulated result of FODT process with PID controller in the feedback path.It is lucid in close
examination of response that noise is prevalent, dominating and cannot be suppressed.
Similarly as on the feedback signal the noise imparts it’s impact on the error signal also. This happens because
the error signal feedback of the error detected is a mixed function of feedback signal as well as the set point.S
imply error signal is the difference of process outcome and signal feedback. The influence of noise is clean in
result shown in Fig.6.
In the spectral analysis of response with PID controller it is noticeable that the ripples have acquired
comparatively significant magnitudes and the attenuation fails beyond an extent. The spectral response is
picturised as in Fig.11.
The LQG when introduced in the feedback signal proves it’s identity, nullifying variations in the form of
guassian noise. To validate the performance of this robust controller, hardwired with a PID and LQG,a band
limited white noise is probably employed. The robustness in all the sense is easy to be examined from the
response shown in Fig 5.

8. Conclusion.
This paper proposes an advanced control strategy to provide robust control for an FOLPD process. The
strategy developed involves a LQG (Linear Quadratic Gaussian), which adapts its coefficients. The adaptation is
done by a Kalman Estimation algorithm which serves to minimize the Mean Square Error (MSE). The proposed
strategy is validated for an FOLPD and all the parameter variations of the process are modeled as a Gaussian
noise. Although the present work involves a first order process, the proposed strategy can however be extended to
higher order processes also. The simulation results are presented for a FOLPD, which is simulated using
SIMULINK.
References.

[1] Astrom.K.J, & T. Hagglund, The future of PID control Control Engineering Practice, 2001, pp.1163 – 1175..
[2].Lee.C.C, Fuzzy logic in control systems: Fuzzy logic controller-Part I. IEEE Trans. Syst., Man, Cybern., vol. 20, pp. 404 -418, 1990
[3] Loh.A.P, K.O. Looi, & K.F. Fong, Neural network modeling and control strategies for a CSTR neutralization process, Journal of
Proc. Control, 6, 1995, pp. 355-362[11]
[4] Oscar Montiel, Roberto Sepúlveda, Patricia Melin, Oscar Castillo, Miguel Ángel Porta, Iliana Marlen Meza, “Performance of a Simple
Tuned Fuzzy Controller and a PID controller on a DC motor”, IEEE 2007.pp. 531 - 537
[5].Rajani K. Mudi, Chanchal Dey, & Tsu – Tian Lee, An improved auto – tuning scheme for PI controllers, Journal of science Direct ISA
Transactions 47 pp. 45 – 52,2008[9]
[6] Shinskey.F.G, process control system: application, design and tuning ( McGraw-Hill, 4th Ed., 1996)
[7] Tipsuwanporn.V, S.Intajag, K.Witheephanich, N.Koetsam-ang and S.Samiamag, “Neuro-Fuzzy Controller Design for Industrial Process
Control”s, SICE 2004. PP 1656-1661
[8] Toru Yamamoto and Sirish L. Shah, “Design of a Performance-Adaptive PID controller”,.International conference on Networking,Sensing
and Control IEEE 2007 pp547-552.
[9] Xiao-Feng Li, Jian Sun, Hui-Yan Wu, Wei-Dong Zong, “Application of the fuzzy PID to the Power Plant”, IEEE 2007
[10] Yun Li,Kiam Heong Ang and Gregory C.Y.Chong, “PID Control System Analysis and Design – Problems, Remedies, and Future
Directions”. IEEE control system magazine, February 2006 pp. 32-41
[11] Yamada.K, & K. Watanabe, A State Space Design Method of Stable filtered Inverse Systems and Its Application to H2 Suboptimal
Internal Model Control, Proc. IFAC96,1996, pp.379-382.
[12]Zhao Yingkai, Lin Jinguo, Shu Zhibing, “The Design of Neural Network Controller for Nonlinear Plants”, IEEE 1997. International
Conference on Intelligent Processing Systems IEEE 1997 Vol.1 of 2

ISSN: 0975-5462 5837