Leak Detection and Localization in Pipelines with Volume Rate Estimation Based on Pressure Signals

Leaks in water distribution networks (WDN) pose serious challenges, leading to water losses, a deterioration in quality through the penetration of contaminants causing waterborne diseases harmful to public health, and significant financial and environmental impacts. Timely and accurate leak detection is essential for efficient water management and infrastructure sustainability. This study presents a robust methodology for leak detection, localization, and volume rate estimation in pipelines, primarily leveraging analysis of internal pressure signals. The approach utilizes real-time pressure monitoring to identify anomalies indicative of a leak. Leak detection is achieved by observing characteristic pressure drops, which are clearly distinguishable from normal operating conditions. Furthermore, the precise location of a leak is determined by analyzing the transient pressure wave propagation. For volume rate estimation, the Area Under the Curve (AUC) of the pressure-time signal is employed as a key metric. Experimental results demonstrate that the AUC provides a quantifiable measure highly correlated with leak size; specifically, a defined baseline AUC for "no-leak"conditions 497.10 PSI.S significantly deviates from AUC values observed during leak events ranging from 451.03 PSI.S for smaller leaks to 430.20 PSI.S for larger leaks. All leak scenarios consistently fall below a pre-established detection threshold 492.10 PSI.S, confirming the effectiveness of the method. This integrated framework offers a practical and accurate solution for proactive pipeline management, enabling rapid identification and characterization of leaks to minimize their impact.