Data-Driven Iterative Adaptive Critic Control Toward an Urban Wastewater Treatment Plant

The wastewater treatment is an important avenue of resources cyclic utilization when coping with the modern urban diseases. However, there always exist obvious nonlinearities and uncertainties within wastewater treatment systems, such that it is difficult to accomplish proper optimization objectives toward these complex unknown platforms. In this article, a data-driven iterative adaptive critic (IAC) strategy is developed to address the nonlinear optimal control problem. The iterative algorithm is constructed with a general framework, followed by convergence analysis and neural network implementation. Remarkably, the derived IAC control policy with an additional steady control input is also applied to a typical wastewater treatment plant, rendering that the dissolved oxygen concentration and the nitrate level are maintained at desired setting points. When compared with the incremental proportional-integral-derivative method, it is found that faster response and less oscillation can be obtained during the IAC control process.