A nonlinear optimization and fuzzy modelling in predictive control scheme

The model-based predictive control (MBPC) technologies are based on the prediction of future behaviour of the process to be controlled which is obtained by the model of the plant. Using the explicit process model and an optimization approach MBPC can be applied to complex, multivariable, nonminimum-phase, open loop unstable systems or processes with a long delay time. The paper introduces the predictive control scheme with nonlinear optimization and a fuzzy Takagi-Sugeno model which is used to describe the nonlinear properties of the process and to predict the process behaviour. A neural network for model parameter estimation in the predictive control scheme is applied, and simulation results of the process control is presented to illustrate the benefits possible with the given concept.