FUZZY GAIN SCHEDULED CSTR WITH GA-BASED PID

In this article, a new control scheme, the gain scheduled genetic algorithm (GA)-based PID is proposed for a continuous stirred tank reactor (CSTR). A CSTR is a highly nonlinear process that exhibits stability in certain regions and instability in other regions. The proposed control scheme implements the characteristics of the genetic algorithm's (GA) global optimization to optimize the PID's three control parameters, k_p, k_i, k_d, to obtain the best control effect by minimizing the integral square error online. The PID controller parameters tuned by the GA for each region are gain scheduled by a fuzzy logic scheduler. Fuzzy gain scheduling is a special form of fuzzy control that uses linguistic rules and fuzzy reasoning to determine the controller parameter transition policy for the dynamic plant subject to large changes in its operating state. Simulation results show the feasibility of using the proposed controller for the control of the dynamical nonlinear CSTR.