The transient wave effect and real losses correlation-frequency, temporal, and spatial analysis

Water distribution systems are prone to transients since pumps need to be started and stopped, pumps may have sudden flow changes, human blunders can occur, equipment may fail, or other unavoidable natural events may ensue. Surge modeling techniques are available to calculate variations in pressure, under extreme or normal operating conditions, serving as a tool to predict extreme events in order to design a suitable system and, or, to aid in the implementation of proper measures to mitigate transients. However, modeling an entire water network system may not be cost effective and may require extensive research and time especially in large distribution systems. Spatial analysis may offer an efficient alternative method of uncovering areas with potential problems. In this research, spatial statistic methods were employed to find whether clustering of leak events is occurring, the distance at which the leak clustering is most prevalent, and the location of the leak concentrations. Specifically, the following spatial statistic methods were utilized: nearest neighbor, Ripley’s K-function, and the Gi* local statistic. The objective for this research was to locate stations with high concentrations of leaks and whether these leak clusters were related to pump trips due to power failures. Although the concentration of leaks was found near two stations, a correlation between the cluster of leaks and the pump trips was not established