Flower Pollination Optimization Based PID Controller Tuning Scheme With Robust Stabilization of Gain Scheduled CSTR

Abstract

In the area of chemical and control engineering, Continuous Stirred Tank Reactor (CSTR) has been deliberated as a most important topic in temperature process control and it is a extremely nonlinear process, which displays stability only in definite regions and unsteadiness in other regions. In this CSTR, Proportional Integral Derivative (PID) controller has been used to minimize the Integral SquareError (ISE) by adjustment of control parameters like k_p,k_i and k_d.But, this conventional controller has high computational complexity, so to reduce this, the optimization based tool has been designed for PID and it was conceptually simple.Earlier, the Genetic Algorithm (GA) optimization has been focused, but it was not guarantee for good fitness and global optimal solution. So to overcome this, here, Flower Pollination based Global Optimization Algorithm (FPGOA) with Gain scheduling based PID controller is proposed for CSTR. When a plant is exposing to big changes in its operational state, the Gain scheduling has been used. The FPGOA is used for tuning the parameters of PID controller for each region to reduce the ISE. The fuzzy gain scheduler has been combined with multiple local linear PID controllers to check the stability of loop for entire regions with various levels of temperatures. This scheduling is an exceptional form of fuzzy control, which customs linguistic rules and fuzzy perceptive to establish the controller factor transition strategy for the dynamic plant dependent on high changes in its operational state to improve the optimization. MATLAB results demonstrate that the feasibility of using the proposed PID controller compared than existing controller, and it shows the FPGOA attained better results, due to the effectual fitness and global optimal solution for the dynamical nonlinear control of CSTR. The response curves of these systems has been generated and compared with existing schemes.