Direct and Indirect Self-Tuning Generalized Minimum Variance Control

Abstract

Theoretically, several self-tuning control (STC) algorithms have been developed and many simulation results have proved their feasibility in the past years, but applications of STC are hardly seen. This paper proposes direct and indirect STC plans that can be applied to a chemical process system with a time delay auto-regressive and exogenous model of a varied time constant. The plan controller is a combination of a generalized minimum variance control (GMVC) strategy and an identification algorithm (simplified parameter) called recursive least squares to estimate controller parameters for the direct method and plan parameters for the indirect method. The experimental results show that GMVC is able to track the desired input or set point.