Sliding Window-based Particle Swarm optimization Algorithm for a Sewer Network

This paper deals with the optimization of a sewer network using a Sliding Window-based Particle Swarm optimization (PSO) algorithm. The sewer network corresponds to a city with a population of 250,000 inhabitants, its model being implemented in BSM Sewer. The optimization of the sewer network is done with respect to the amount of the discharged water minimizing this way the environmental impact of the wastewater collecting and transport systems. In the simulations, an influent containing three components was considered: domestic, pluviometric and industrial influent. Daily weather forecast for influent was considered. The control strategy proposed in this paper is based on the calculus of the optimal control actions for each individual tank with respect to the liquid level in that tank. The optimal controls were determined over a 24-hour sliding time window. The optimization algorithm was run over a time horizon of 28 days, during which two events were considered, one of rain and one of storm. As the daily forecast of the influent is not 100% accurate, a disturbed influent was considered to observe if the performance of the algorithm is maintained in this scenario. The simulation results were compared with those obtained in the operating mode “no control” (all control actions set at 100%).