A nonlinear model predictive control strategy based on dynamic fuzzy model using two-step optimization method

The dynamic fuzzy model implements a set of local dynamic models, identified by the least square method, to approximate the dynamics of a nonlinear process. The nonlinear predictive controller consists of a multi-step predictor based on a dynamic fuzzy model, an output optimizer and a robust filter. The output is optimized by two steps, the descent-gradient method first, and then a linear optimization. The robust filter with one adjustable parameter can resist the model mis-match and improve the transient performance. The simulation of pH neutralization process is given to demonstrate the better performance of the proposed control scheme compared with a conventional DMC controller.