Heuristic Dynamic Programming Using Echo State Network For Multivariable Tracking Control Of Wastewater Treatment Process

Asian Journal of Control

Han

et al. 2018

Due to the characteristics of strong coupling and high nonlinearity in the control process, an intelligent decoupling control strategy based on recurrent fuzzy neural network (RFNN) is proposed in this paper to control the wastewater treatment process (WWTP). Firstly, the architecture of the RFNN controller is designed with a mechanism analysis of WWTP. Secondly, a decoupling strategy in combination with a gradient descent search algorithm is used to decouple the control loop of dissolved oxygen (D O ) concentration and nitrate nitrogen (S NO ) concentration. Finally, stability analysis based on a Lyapunov function is investigated. The proposed approach has been applied to the WWTP simulation model. Compared to model predictive control, echo state network-based HDP (E-HDP), conventional RFNN, and neural network on-line modelling and controlling methods, the proposed method has better control performance.

Decoupling control for wastewater treatment process based on recurrent fuzzy neural network

Asian Journal of Control

Han

et al. 2018

Intelligent optimal tracking with asymmetric constraints of a nonlinear wastewater treatment system

Wang

Zhao

Intl J Robust & Nonlinear

2021

The wastewater treatment is an effective method for alleviating the shortage of water resources. In this article, a data‐driven iterative adaptive tracking control approach is developed to improve the control performance of the dissolved oxygen concentration and the nitrate nitrogen concentration in the nonlinear wastewater treatment plant. First, the model network is established to obtain the steady control and evaluate the new system state. Then, a nonquadratic performance functional is provided to handle asymmetric control constraints. Moreover, the new costate function and the tracking control policy are derived by using the dual heuristic dynamic programming algorithm. In the present control scheme, two neural networks are constructed to approximate the costate function and the tracking control law. Finally, the feasibility of the proposed algorithm is confirmed by applying the designed strategy to the wastewater treatment plant.

Optimal trajectory tracking control for a class of nonlinear nonaffine systems via generalized N‐step value gradient learning

Zhao

Wang

Intl J Robust & Nonlinear

et al. 2022

In this paper, the tracking control problem of unknown nonlinear systems is solved by using the generalized N-step value gradient learning algorithm with parameter 𝜆 [GNSVGL(𝜆)]. The GNSVGL(𝜆) algorithm can provide optimal tracking decisions faster than traditional ones. Initialized by different positive semi-definite functions, the monotonicity and convergence properties of the proposed algorithm are proven. Under some conditions, the stability analysis of the value-iteration-based algorithm is provided. The one-return and 𝜆-return critic neural networks are constructed to approximate the gradient of the one-return and 𝜆-return cost functions. The action neural network is employed to approximate the control law of the error system. It is emphasized that one-return and 𝜆-return critic networks are combined to train the action neural network. Finally, via conducting simulation studies and comparisons, the excellent tracking performance of the proposed algorithm is confirmed.