Implementación de un algoritmo de control predictivo en espacio de estados sobre una plataforma de simulación desarrollada en Matlab

Abstract

Model Predictive Control (MPC) is a methodology in the field of modern control engineering. Since its beginnings it has progressively evolved to become a tool capable of largely mitigating the problems facing industry from the control theory point of view. Essentially, the predictive control strategy uses an internal mathematical model and an optimization strategy to predict system outputs within a time period known as the prediction horizon. The formulation supporting the algorithm has a high computational cost. This is why this strategy has found its niche since the start in computers which speed up and support substantial matrix calculations. This paper details the mathematical analysis and procedure for implementation of a predictive control algorithm under Matlab®. Also presented are the results of its application to a singleinput single-output (SISO) plant and a multiple-input multiple-output (MIMO) plant. In both cases, restrictions are incorporated in the process variables, regarded as one of the most attractive features of this strategy.