On a new methodology for the design of structure reduced multivariable controllers

Abstract

This paper proposes a new methodology for the design of multivariable controllers with a reduced structure. The structure of a multivariable controller may be defined in several ways. In this paper it is defined as the number of the input to output connections of the controllers. Based on this structure definition, a generic cost function is determined where both the structure of the controller and the performance of the resulting closed loop system appear as minimization terms. It has been shown that there exists an optimal weighting such that the optimization of the weighted sum of these two terms will always result in a unique reduced structure controller. The methodology is then successfully applied to the model of the ALSTOM (ALSTOM Power Technology Centre, formerly GEC ALSTHOM) gasifier challenge problem, which is a difficult multivariable system to control.