Fuzzy Neural Network-Based Model Predictive Control for Dissolved Oxygen Concentration of WWTPs

Dissolved oxygen (DO) concentration is a key variable in the operation of wastewater treatment processes (WWTPs). In this paper, an adaptive fuzzy neural network-based model predictive control (AFNN-MPC) is proposed for the control problem of DO concentration. The main contributions of AFNN-MPC are threefolds: First, an AFNN, based on a novel learning method with adaptive learning rate, is designed to model the unknown nonlinearities of WWTPs with high predicting performance. Second, a gradient method is used to solve the optimal control problem of AFNN-MPC to reduce the computational cost. Third, the convergence of AFNN, as well as the stability analysis of AFNN-MPC, has been given in detail. Finally, the proposed AFNN-MPC is applied to the benchmark simulation model No. 2. The promising results indicate that the proposed AFNN-MPC is a suitable solution to control DO concentration. Moreover, the comprehensive experiments clearly show the superiority and efficacy of the proposed AFNN-MPC.