Author(s): Simona Corrado; Elisabetta La Penna; Giulia Acconciaioco; Oronzo A. Pizzutolo; Gianfredi Mazzolani; Orazio Giustolisi
Linked Author(s): Orazio Giustolisi
Keywords: Asset deterioration; Leak detection; Noise loggers; Water losses
Abstract: Water loss remains one of the most pressing challenges for utilities, demanding smarter and more sustainable management strategies. Moving beyond reactive practices, this work introduces a proactive, data-driven approach to predict asset deterioration and optimize rehabilitation planning. The methodology combines continuous acoustic monitoring from IoT noise loggers with GIS data and Big Data analytics, enabling real-time observation of leak evolution and advancing from conventional detection to predictive modelling. Network segments are classified by criticality, and Evolutionary Polynomial Regression is applied to derive symbolic models that estimate failure propensity and reveal key drivers of risk. The approach has been validated on Acquedotto Pugliese’s networks in Italy, where since 2021 approximately 20,000 noise loggers have been deployed across 3,000 km of pipelines in 26 municipalities. This large-scale implementation demonstrates the feasibility and scalability of the solution, paving the way for smarter, proactive water asset management.
Year: 2026