Design and Implementation of Soft Computing-Based Robust PID Controller for CSTR

Abstract

AbstractThis paper presents the design of soft computing-based robust PID controller for a Continuous Stirred Tank Reactor (CSTR) which is widely used in the process industry and facilitates the dilution of reagents through mixing. To have an efficient operation and to prevent the temperature deviation from its set point, parameters like concentration and temperature need to be controlled. Literature shows use of a conventional method like Ziegler–Nichols (Z–N) and soft computing-based methods like Genetic Algorithm (GA), Particle Swarm Optimization (PSO), Artificial Bee Colony (ABC) and Teaching Learning-Based Optimization (TLBO) to tune PID which regulate the temperature of the CSTR. The dynamic behaviour of CSTR is very uncertain and it is observed in the literature that there is scope to design a robust controller that takes care of parametric uncertainty. The work presented in this paper is a sincere attempt to design efficient soft computing-based PID controller for CSTR. The simulation results demonstrate the performance of the designed controller with respect to time response specifications like rise time, peak time, settling time and overshoot, which shows improved response for proposed TLBO-based PID controller than other conventional and metaheuristic methods. Error indication using Mean Square Error (MSE) and Integral Time Absolute Error (ITAE) show that the TLBO-tuned PID is fast and efficient. It provides optimized controller parameters by reducing the error with less control effort. Robustness analysis is carried out by considering the parametric uncertainty in the CSTR model. Finally, stability of proposed TLBO-based PID is proved by using frequency response analysis.KeywordsCSTRRobust PIDGAPSOABCTLBO