Tuning of Robust PID Controller with Filter for SISO System Using Evolutionary Algorithms

Abstract

This paper discusses the performance comparison of Evolutionary Algorithm techniques with a view to tuning the desired design parameters of robust PID controller with filter. The design parameters of robust PID controller such as proportional gain K P , integral time constant T i , derivative time constant T d and filter time constant t d /N for Single Input Single Output (SISO) system are computed using Real coded Genetic Algorithm (RGA), Differential Evolution (DE) algorithm and Particle Swarm Optimization (PSO) algorithm. The design parameters are optimized using the statistical measures in twenty independent simulation runs. The computed design parameters obtained using Evolutionary Algorithms are used to determine the performance specifications of the Robust PID controller. The performance specifications determined are robustness with respect to model uncertainties and disturbance attenuation, set point tracking, load disturbance rejection and control energy. The test systems used to ascertain the performance specifications are Phase Locked Loop (PLL) system with motor control and Magnetic Levitation system (MLS). For PLL system, the output response obtained using the computed design parameters by RGA algorithm has proved to be better than DE and PSO. For MLS system, the output response obtained using the computed design parameters by DE algorithm has been better than PSO and RGA. meet out the performance specifications of the given system mainly depends on the tuning of the controller .Since the system has nonlinear plant and various uncertainties the tuning of the robust PID controller becomes difficult. This leads to poor tuning. The reasons of the poor tuning is due to a) lack of knowledge among operators and commissioning personnel b) variety of PID structures that leads to error during tuning c) in some cases the nature of tuning method do not meet the requirement of the process involved [3]. Out of the installed PID controllers in industry, nearly one third controllers are tuned manually and remaining two third are tuned automatically. But the automatically tuned PID controllers are not effective [7].To overcome these limitations optimization algorithms are used for the tuning of robust PID controller. In the past many researchers have used different optimization algorithm for tuning the parameters of robust PID controller. But the parameter tuned to meet out the performance specifications of the system is different in each case. Many researchers have considered set In this paper, Evolutionary algorithm is used to determine the optimal parameters for the robust optimal controller. Evolutionary Algorithms used are Particle Swarm Optimization (PSO), Real coded Genetic Algorithm (RGA) and Differential Evolution (DE). The main focus of this paper is to determine the design parameters of robust PID controller using evolutionary algorithm such that it meet out the system performance specifications like robustness with respect to model uncertainties and disturbance attenuation, set point tracking,...

Evaluation of various benchmark processes with appropriate controller design in LabVIEW platform

“…Among these combinations, PID controller is the most popular one, due to its improved response time in both transient and steady state responses of the system [13]. The stability of the system is maintained by the PID controller and proper tuning makes improved system performance [14]. PID controller consists of proportional, integral and derivate controls as in figure 2 The transfer function for standard PID controller in parallel form is given by,…”