Pressure gauge network design for leakage detection in water distribution systems using a combination of the pressure residual vector method and the entropy concept

Detecting leaks in Water Distribution Systems (WDS) is a challenging task for maintenance managers. One widely used method for leak detection is the Pressure Residual Vector method, which involves comparing pressure values with computed values. In this paper, a new approach using probability distribution function and entropy content is proposed to expand the applicability of the previous models. The study began by collecting real-time information on water consumption using automatic electromechanical water meters over a one-year period. Hourly parameters were calculated to determine the temporal pattern of water consumption per capita. Hypothetical leakage was then imposed on each pipe, and the average entropy of water head variations in the nodes was determined. A combinatorial search process was used to suggest a Pressure Gauge Network (PGN) in a WDS. The study found that after a certain number of pressure gauges, the increase in entropy content is insignificant. This quantity may be deemed as the number of pressure gauges required in the PGN. Furthermore, the study showed that the impact of leak parameters can be disregarded when designing a PGN for leak detection purposes. Overall, the study's findings can be used to design a PGN for leak detection in a WDS. HIGHLIGHTS Per-capita hourly consumption of water was obtained by field measurement.; The entropy concept is used to identify nodes that are most affected by failures in the distribution network.; These nodes are used as candidate nodes for installing pressure gauge sensors to detect leaks.; The effect of leakage parameters on the monitoring network is evaluated.; Information content reflected by pressure variation in nodes is reported.