Estudo e implementação de um algoritmo de avaliação de desempenho de malhas de controle sob uma referência DMC.

Abstract

In this work the problem of Model Predictive Control (MPC) performance assessment is treated. Two assessment algorithms are implemented. The rst is the most used in the tools of the genre, this is the Harris Index which its reference is the Minimum Variance Control (MVC). The second is the DMC Index whose reference is the most common MPC controller, the Dynamic Matrix Control (DMC). The calculation of the indices depend on the data collection system in closed loop. The results of these two routines are analyzed by applying some simulated control systems. In these experiments it's used both the DMC controller as PID controller. Still, it was added real problems situations, such as adding noise to the output of the system, imperfect model of the process and variations in the tuning of controllers.