A Comparative Analysis of Adaptive Predictive Control Methods Applied in a Heat Exchanger

Abstract

Most industrial processes are nonlinear, which complicates the application of conventional Model-based Predictive Control (MPC) algorithms. Consequently, in this article, the formulations of MPC methods for nonlinear processes represented through polytopic Linear Parameter-Varying models are analysed. The compared methods are adaptive algorithm, synthesised with a prediction model based on a scheduling polytope. At each discrete sampling instant, they determine a model, used for prediction purposes; and optimise the process performances over a finite prediction horizon. These methods are applied to control of a Heat Exchanger system, from which the performance and effectiveness of each technique are discussed. The simulation results are thoroughly analyzed, and the advantages and disadvantages of each strategy are discussed.