Leakage localization using pressure sensors and spatial clustering in water distribution systems

Leakages in water distribution systems (WDSs) are a worldwide problem, which can result in an intolerable burden in satisfying the water demands of the consumers. There is an urgent demand to develop technologies that can detect and localize the leakage in a timely and efficient manner. The monitoring data of the WDS is a typical time series, and there is a certain spatiotemporal correlation between the data provided by the devices distributed at different locations of the WDS. This paper proposes a novel model-based method for WDS leakage localization. The method is characterized by (1) developing the dominant sensor sequence for each candidate leakage node to improve the localization accuracy based on the spatial correlation analysis; (2) utilizing multiple time steps of the measurements which are temporal varying correlated; (3) ranking leakage regions and nodes by their possibility to contain the true leakage. A realistic WDS is used to evaluate the performance of the method. Results show that the method can accurately and efficiently localize the leakage.