Data-driven decision-making for wastewater treatment process

Membrane fouling has become a serious issue for the safe operation of wastewater treatment process (WWTP). To deal with this problem, this paper proposes a data-driven decision-making method to reduce the incidence of membrane fouling in WWTP. The main novelties of this proposed data-driven decision-making method are threefold. First, a long-term prediction method, based on a self-organizing deep belief network (SDBN) and the multi-step prediction strategy, is developed to predict the membrane permeability. Second, a multi-warning method, based on an independent component analysis-principal component analysis (ICA-PCA) algorithm, is proposed to detect and warn membrane fouling with multiple indicators. Third, a multi-category diagnosis method, based on the kernel function, is designed to diagnose membrane fouling for providing the decision support. Finally, an intelligent decision-making system, consisting the above methods and required sensors, is developed for some real wastewater treatment plants. The experimental results demonstrated the efficiency and effectiveness of the proposed data-driven decision-making method.