Nonlinear Model-Predictive Control for Industrial Processes: An Application to Wastewater Treatment Process

Because of their complex behavior, wastewater treatment processes (WWTPs) are very difficult to control. In this paper, the design and implementation of a nonlinear model-predictive control (NMPC) system are discussed. The proposed NMPC comprises a self-organizing radial basis function neural network (SORBFNN) identifier and a multiobjective optimization method. The SORBFNN with concurrent structure and parameter learning is developed as a model identifier for approximating the online states of dynamic systems. Then, the solution of the multiobjective optimization is obtained by a gradient method which can shorten the solution time of optimal control problems. Moreover, the conditions for the stability analysis of NMPC are presented. Experiments reveal that the proposed control technique gives satisfactory tracking and disturbance rejection performance for WWTPs. Experimental results on a real WWTP show the efficacy of the proposed NMPC for industrial processes in many applications.