Tuning of a Minimum Variance Control System Based on the Estimated Process Gain

Abstract

The control of a nonlinear process raises various problems, both in terms of the control law design and the controller tuning. This paper presents a tuning procedure of a minimum variance control system based on the analysis of the estimated gain of the controlled process. In this regard, the performed study shows that the settling-time of the estimated process gain can be used as a tuning criterion for the minimum variance controller, allowing the improvement of the control performance. The procedure involves closed-loop estimation of the process gain based on parameters estimates of a linearized model that approximates the nonlinear process functionality around an operating point. The basic idea is that, although the parameters estimates of the linearized model are different in closed-loop, respectively open-loop, the steady-state gain estimates are similar in both cases. Thus, the time dynamics of this estimated gain can be a useful indicator for the control system tuning. In order to validate the proposed procedure, an induction generator integrated into a wind energy conversion system was considered as a controlled process.