Fuzzy model-based predictive control by instantaneous linearization

A linearization technique for product-sum crisp-type fuzzy model and a multistep predictive control strategy for the construction of a model-based predictive fuzzy controller is presented in this paper. A model-based predictive controller is based on a local linear model generated by using the proposed linearization technique that linearizes the product-sum crisp-type fuzzy process model around the current operating point. The control of pH in a continuous stirred tank reactor is chosen as a realistic nonlinear case study for the demonstration of the proposed control algorithm. The controller is shown to be capable of controlling the nonlinear processes that operate over a wide range, and of providing better overall system performance than the optimal PI controller. (C) 2001 Elsevier Science B.V. All rights reserved.