Abstract
A multi-stage response procedure for identifying possible ingress nodes (PINs) and quantifying the likelihood that a PIN in a given water distribution system is the actual point of ingress is described. The procedure uses data mining to successively decrease the number of PINs based on a pre-constructed database. In each stage, query sentences are executed to locate the PINs and a Euclidean distance is proposed to estimate the probability, to allow the identification of locations with the highest probabilities of being the true ingress location. As demonstrated in a case study, the ranges of PINs are reduced in the 1st, 2nd, and 3rd stages; except the first sensor alarm, the Euclidean distance metric can identify the true ingress node with the program run-time of less than 2 min; the multi-stage procedure saves roughly 3 h in identifying the true ingress node after the second sensor alarm, instead of waiting for a third sensor alarm to provide the location information. The multi-stage response procedure is shown to be an effective and efficient way for identification and probability quantification of PINs.
Published Version
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call