Real-Time Response to Contamination Emergencies of Urban Water Networks

In order to develop an emergency response plan for contaminant flushing in drinking water networks in intentional or accidental contamination, a decision tree-based model is suggested for real-time application. The approach is based on searching for the best layout of flushing nodes and searching for a set of simple rules that can be readily used in real-time situations for determining the optimal duration of hydrant flushing. The methodology consists of a hydraulic simulation model (EPANET), a fast robust multi-objective optimizer (NSGA-II), a Monte-Carlo simulation for including the inherent high uncertainty in deliberate contamination and a well-known multi-criteria decision-making technique, namely TOPSIS. In order to overcome bottleneck of time-consuming optimization and represent simple rules for contaminant flushing of drinking water network, the M5P model tree was utilized. To evaluate the applicability and accuracy of the proposed methodology in supporting decision makers in making effective real-time decisions, it was applied to a real water distribution system. The developed decision tree is shown to be an effective tool for real-time application.