Tracking Control of a Continuous Stirred Tank Reactor Using Advanced Control Algorithms

Abstract

The Continuous Stirred Tank Reactor (CSTR) is a common type of industrial equipment used in chemical processes with second-order nonlinear dynamics. CSTR is a nonlinear and linked nature makes it challenging to develop a robust control with a bigger operating zone. In the process industry, a CSTR is a crucial component. It serves as a foundation for studying and controlling other chemical reactors. Good state estimation and disturbance rejection are required in industrial processes. The most important characteristic for CSTR operation is temperature. The behavior is obtained by steady-state and dynamic analysis of the model which is usually represented by a set of differential equations. An event-based Neural network, Model Predictive Control (MPC), and Model Reference Adaptive Control (MRAC) controller are presented in this work to provide robustness to the system with the added benefit of conserving energy expenditure under parameter variations and fast changing dynamics. Numerical simulations were used to confirm the controller's robustness and efficacy. In comparison to the ANN and MPC, the simulation results clearly show that the MRAC technique delivers appropriate performance in terms of process functional improvements, more flexibility, and improved system-tracking precision in control action.