Determining the Optimal Location of Sensors in the Wat

In order to detect pollutants in water distribution systems, sensors are used which, in addition to detecting pollution, also provide useful information to identify the source of pollutants. This information is useful for finding a suitable solution for quality management of the distribution network. The purpose of this study is to determine four characteristics of the pollution source to the system by a simulator-optimizer approach with optimally allocating two sensors in EPANET Example 2. These characteristics are: the contaminant injection node, injection mass rate, injection starting time and injection duration time. EPANET and genetic algorithm were linked as the simulator and optimizer tools to achieve these goals, simultaneously. For this purpose, first ten different pollution events were generated randomly, then the best place to install sensors in the network for identifying the source of pollution was found. Moreover, the shortest time of detection was selected among fourteen different combinations with a contamination source identification likelihood equal to 90%. The results showed that by placing sensors in two nodes 15 and 27 of the benchmark problem, the highest contamination source identification likelihood equal to 90% and the minimum detection time equal to 58 minutes were achieved.