Identification of the inflow source in a foul sewer system through techniques of inverse modelling

Abstract Infiltration and illegal inflow into foul sewer systems can cause different problems such as a decrease in the performance of treatment plants, surcharge pipelines more frequent overflows, which negative impacts on environment. Water companies are increasingly been driven to address these by reducing infiltrations identifying sources inflows. Overall, traditional techniques applied cases expensive time consuming many times only partially efficient. Examples use CCTV inspections, smoke tests installation large set sensors collect continuous data flow rates, water levels, temperature or concentrations pollutants. The aim this study is apply two types inverse numerical identify source location inflows wastewater based information collected at outlet drained basin calibrated model network. In work, developed using Storm Management Model (SWMM) software distributed Environmental Protection Agency (USA). We considered realistic system with known dimensional hydrological characteristics. Synthetic case studies up test approaches. Assuming hypothetical rainfall event an released certain system, run forward obtain hydrograph network outlet. This then used available observations perform modelling. first investigated technique artificial neural (ANN) feed-forward type. It will be trained recover simulation results SWMM events located positions Once trained, ANN observed flood wave. second procedure derives from Kalman filter techniques: Ensemble Smoother Multiple Data Assimilation (ES-MDA). Also case, method, starting hydrograph, locate source. addition synthetic obtained means procedures, field applicability real discussed.