Analysis of PSO-PID controller for CSTR temperature control

Continuous Stirred Tank Reactor (CSTR) is a pivotal unit in chemical process industries, and finds many applications. However, it is unstable, nonlinear and disturbance prone. Thus, it needs to be robustly controlled. Industrial process control techniques have greatly progressed over recent decades and numerous control methods have been proposed. Here, we have analyzed a Particle Swarm optimization (PSO) based PID, whose gains are optimized for stabilizing temperature control in CSTR plant. This is carried out by selecting parameters to obtain best possible solution for different conditions. Two input disturbances are considered to ensure and verify that the PSO-PID is able to control the plant robustly. Further, the system response is compared against conventional methods such as Ziegler-Nichols tuning, genetic algorithm and classical PSO. It is found that the patterns of PSO-PID gains versus iterations, are able to improve performance much better against the considered counterparts.