Efficient predictive control and set-point optimization based on a single fuzzy model

Abstract

The idea proposed in the paper consists in significant simplification of the control structure with a predictive control algorithm and a steady-state target optimization. It is done by application of only one fuzzy (nonlinear) dynamic control plant model for both: predictive control and set-point calculation. The approach exploits possibilities offered by a fuzzy model used by the predictive control algorithm. The fuzzy model is of Takagi-Sugeno type with step responses used as the local models. Such a model can be obtained relatively easy and well tuned using neural networks. The proposed approach, despite simplification of the control system, offers very good control performance. It is demonstrated using an example of a control system of a nonlinear chemical reactor with inverse response. © 2011 Springer-Verlag.