Real time optimization of a network of wastewater treatment plants

This master thesis aims to propose a control strategy through the use of convex optimization to optimize the distribution of wastewater in the SIAAP network of Paris, with the objective of minimizing pollutant concentrations in the Seine River. SIAAP (Syndicat Interdépartemental pour l'Assainissement de l'Agglomération Parisienne) is a public utility responsible for the transportation and treatment of waste, rain, and industrial water for 9 million people in Paris and its suburbs. It treats a daily volume of 2.4 million cubic meters of wastewater, which is discharged into the Seine River. Through its operations, SIAAP plays a crucial role in safeguarding water resources and making a significant environmental impact. The objective of this master thesis is to present a control strategy that minimizes the substrates concentration entering the Seine River by optimizing the distribution of wastewater among the wastewater treatment plants within the SIAAP network. To address this problem, convex optimization has been employed, utilizing second-order cones (SOC) to approximate microbial growth in the treatment plants. The alternating direction method of multipliers (ADMM) algorithm has been implemented to solve the biconvexity in the objective function. To test and validate the control strategy, a simplified model of the western part of the Seine River Basin has been developed. The control strategy was tested under different weather conditions, including dry, rainy, and stormy weather, and was compared against real measurements for two distinct periods. Evaluation using non-optimized data demonstrated the effectiveness of the model in all weather conditions and revealed, for example, a 12% improvement in ammonia nitrogen concentration and an 11% improvement in nitrite concentration during rainy conditions. These results indicate that the existing partitioning of wastewater is suboptimal and that the convex optimization control strategy offers an optimized partitioning approach that can improve water quality.