Dual-objective optimization for energy-saving and fouling mitigation in MBR plants using AI-based influent prediction and an integrated biological-physical model

Abstract Optimal operational strategies for wastewater treatment plants have been employed widely to improve the economic and environmental operations. However, inadequate optimization systems have been applied to membrane bioreactor (MBR) plants, resulting in high energy consumption due to uncertain influent associated with complex biological and physical interactions associated with membrane fouling. Here, we develop a dual-objective optimization system based on harmony search algorithm via process simulation. A bidirectional gated recurrent unit produced the most accurate predictions of fluctuating variations in hourly influent among competing models. And the predicted influent information was used to suggest operational strategies. The optimization system searched predictive operational strategies, including aeration intensities and permeation-cleaning durations, using the integrated biological and physical model. The suggested aeration intensities were then used for scouring air on membrane surfaces when the permeation-cleaning duration was optimized. A pilot MBR plant equipped with the suggested optimization system improved energy efficiency by 4% and mitigated fouling by 39% compared with a manually operated system. Dual-objective optimization also demonstrated feasibility and reliability at a full-scale MBR plant by increasing energy efficiency by 12% and decreasing fouling by 26%.