Optimal Location Of Water Quality Monitoring Sites And Comtamination Source Identification In Water Distribution Networks

In terms of solving the problems of drinking water contamination caused byaccidental contaminant event, we used a series of techniques which could identify thesource location to isolate the contaminated area and minimized its hazards. It wassignificant to the drinking water security emergency handing.Firstly, we proposed a new method for ensuring optimal locations of waterquality monitoring sites in distribution networks, under multi-coverage. A greedyalgorithm had been used to get monitoring sites and two water distribution networksserved as examples for analysis. When the ratio coefficient α was small, the entirepipe network monitoring system coverage ratio was low and the usage effectivenessof the selected monitoring points was low as well. When the ratio coefficient α gotbigger, the coverage ratio of selected monitoring sites became higher, but theproportion of the nodes whose coverage times were equal or greater than2was lower.When the level of service LS was low, the monitoring system had low coverageperformance, with the increasing of monitoring points, the coverage ratio hadobviously increased, but the coverage ratio is no high in total. When the level ofservice LS was high, fewer monitoring points was needed and the system has a highcoverage performance, but response to pollution incidents through the monitoringsystem took a long time. Compared with Cover-Set method, multiple cover methodalmost had the same number of covering nodes in the network, but the number ofcovering nodes whose coverage times were equal or greater than2were superior thancover-set method.Secondly, positioning pollution sources in contaminate emergency of water pipenetworks was investigated. Determined the sources of candidate regions through pathanalysis, and this region was relatively much smaller compared with the entire pipenetwork. Genetic algorithms (GA) and multiple population genetic algorithms (MPGA)were used respectively to solve the presented simulation-optimization model, and theinfluencing factors and the stability of the algorithm were also discussed. In thecalculation process, delaying time, time-divided numbers and evolutionarygenerations needed to be selected for appropriate values. Using GA algorithm to solvethe model, the objective function could quickly converge to the optimal solution withgood calculating results. According to the relationship between the occurred time and the location of source in contamination event, the time range were divided and eachtime-span corresponded to a node set, then using MPGA algorithm to solve the model.MPGA algorithm had higher accuracy and greater stability compared with GAalgorithm.