Research For Water Quality Simulation And Contamination Source Determination Of Water Supply Networks

Tracking the dynamic variation of water quality and identifying contamination source in distribution systems play a significant role in drawing up reasonable strategies and ensuring the security of potable water. In the study, Eulerian and Lagrangian method was applied to the simulation of spatio-temporal variation of residual chlorine in a laboratory small-scaled water distribution system. Based on experimental study, decay constant of residual chlorine was determined. Sodium hypochlorite solution was introduced from different intrusion points to the system for simulating the accidental contamination in the municipal water system, in which the concentration of residual chlorine indicated contaminant amount. On the basis of guaranteeing the accuracy of the measured results, authors collected sets of contaminant concentrations at monitoring sites to be data source. The improved simulation-optimization method based on Lagrangian algorithm was employed to determine the location of contaminant intrusion points; via the comparison of running results, authors analyze the effects of influencing factors such as system topologic structure, simulation time horizon, and water hydraulic conditions. By analysis and collation of 40 sets of monitoring data, the model inputs and outputs of BP network were determined. The advantages and limitation of the model were proposed on the basis of theoretical analysis and results comparison.Results show that on the premise of influencing factors appropriately set, the simulation-optimization model successfully determined source location with 93.3% accuracy and BP network with 100%. The two algorithms have advantages in different aspects; however, for the research as far, the simulation-optimization algorithm demonstrates its stronger robustness and broader application.