Fuzzy Neural Network-Based Adaptive Asymmetric Constraint Control in Wastewater Treatment Process

Wastewater treatment process (WWTP) is an important mean to prevent water pollution and improve ecological environment. Dissolved oxygen (DO) and nitrate nitrogen (NO
${}_{3}$
-N) concentrations are the main indicators to affect the effluent quality (EQ). In order to achieve high accuracy control, the artificial intelligence based asymmetric constraint control method with actuator saturation processing technology is proposed in WWTP. First, the fuzzy neural network (FNN) model is used to estimate the unknown situations, in which the maximum correlation entropy criterion is introduced into the adjustment of model structure to deal with the dynamic changes of WWTP. Second, the unified actuator saturation processing model is established to achieve the stable tracking performance. Third, the asymmetric barrier Lyapunov function (BLF) is introduced into the controller design. Not only the DO and NO
${}_{3}$
-N concentrations, but also the tracking error are kept within the asymmetric constraint range to guarantee the control performance. Finally, the effectiveness of the proposed method is verified via the dynamical and constant set-points simulation experiments of benchmark simulation model 1 (BSM1).