Learning Model Predictive Control in a Virtual Environment Through a Practical Case: A Continuous Stirred Tank Reactor

Abstract

This article presents a virtual platform to learn Model Predictive Control (MPC) through a brief analysis of the mathematical model of a continuous stirred tank reactor (CSTR). The control algorithm was developed from a linearity process to regulate the CSTR at the operation point. The MPC optimization took the temperature in the jacket and the molar concentration as the control objectives. The final version of the virtual platform presented flexibility to change every parameter of the MPC to see their effect on the control algorithm to learn effectively the MPC regulation.