Diseño y simulación de un método de sintonización de un controlador predictivo basado en modelo mediante un filtro extendido de Kalman para un reactor continuamente agitado

Abstract

In this work a method of tuning a model predictive controller (MPC) through an extended Kalman filter is presented. The system on which the algorithm is tested is a continuous stirred tank reactor known as CSTR. The development of the work begins with the mathematical model of the CSTR through a process of identification of a linear system of reduced order. Subsequently, the predictive controller is formulated from a state space system. Finally, the extended Kalman filter algorithm is implemented to tune the error weights and control effort from the MPC.