Design of a Performance-Driven Controller with 1-Parameter tuning

Abstract

In process industries, it is important to maintain the desired control performance in steady state because of requirement of high-quality products and energy saving. This is the main motivation in the development of a performance-driven control scheme based on the minimum variance control law. In that scheme, controller parameters are adjusted when the control performance based on the MV-Index (Minimum Variance Index) deteriorates. However, this MV-Index based tuning scheme only takes into account only controller error variance unfortunately with no regard for manipulative variable activity. To circumvent this omission, this paper describes a design of a new performance-driven control system whose control performance assessment scheme takes into account controller error variance as well as the manipulating variable variance. Furthermore, in this newly proposed scheme, controller parameters are calculated using a fictitious reference iterative tuning (FRIT) scheme which is one of the data-driven tuning scheme. The effectiveness of the proposed scheme is demonstrated by evaluation on a numerical example.