Pseudo-Spectral Model Predictive Control of Continuous Stirred-Tank Reactor

The goal of this paper is to develop a model predictive control based on pseudo-spectral (PS) discretization to control the reactor temperature and concentration of species in a continuous stirred-tank reactor (CSTR). CSTR is a highly nonlinear system which is an important process in chemical engineering. To apply online optimal control to the CSTR, nonlinear model predictive control (NMPC) is the best option. In the online application of model predictive control, in order to have a computationally feasible control algorithm, the low discretization nodes are necessary to decrease the computation time of the NMPC. Motivated by the successful implementation of PS optimal control in aerospace and military industry, the Legendre-Gauss PS optimal control in a receding horizon manner is implemented to CSTR. The simulation results prove the better performance of the proposed method over the equal distance discretization of NMPC in computation time.