Real-Time Pumping Optimization of Combined Sewer Overflow Systems

Abstract Submission: Combined sewer overflow systems (CSOS) play a crucial role in sustaining urban environments by managing sewer and urban runoff flows. These systems collect sewer and urban runoff flows in gravitational systems and then pump them into pressurized systems leading to treatment facilities. Given the limited capacity of CSOS, optimizing their operation in real-time is essential to minimize flooding and pollution from overflows. This task poses a multifaceted challenge due to the computational burden of generating an optimal policy in short run times, conflict objectives, and several uncertainty factors regarding the inflow rates from rain runoff and domestic effluents. This study introduces a real-time decision-making framework that integrates data-driven forecasting with multi-objective optimization. Three objectives are targeted: minimizing total overflow volume, reducing energy consumption, and prolonging pump life. We explore the trade-offs among these objectives to illuminate the costs and risks associated with CSOS pump operations. The efficacy of this approach is demonstrated through its application to a real CSOS system, implemented within an online predictive framework to evaluate the optimality and robustness of the solutions.

Learning Objectives/Expected Outcome (Optional) :