False Negative/Positive Issues in Contaminant Source Identification for Water-Distribution Systems

A contaminant source identification (CSI) methodology for water distribution systems is intended to identify possible events (i.e., intrusion nodes, times, duration, and mass rate). The methodology has to be both rapid and able to incorporate uncertainties when identifying possible intrusion nodes (PINs). Identification of PINs has two major issues: the false-negative rate (failure to identify the true ingress location) and the false-positive issue (falsely identifying a location that is not the true ingress location). A data-mining procedure is described and applied, which involves mining an offline-built database to select PINs that possess first-detection times within $\pm m$ from the online sensor first-detection time, with $m$ selected to address issues of false negatives and positives. This data-mining approach is made possible through the power of parallel computing, which demonstrates huge potential by simulating scenarios simultaneously. In the case studies, scenario simulation times are reduced linearly to the number of processors applied. Results show that increasing the number of scenarios in the database can provide input to compute the $m$ value, always reduces the false-negative rate of each sensor, and usually reduces the number of false PINs. Demonstrated by data-mining online application for two case studies of water distribution systems, the procedure identifies the PINs within less than 4 min, demonstrating that data mining represents a rapid and efficient CSI procedure.