Integration of flow decomposition and discharge testing for customer-side leakage detection in district metered areas

Customer-side leakage, a component of background leakage in water distribution systems, is commonly characterized through inlet flow signal decomposition and mitigated through pressure management. However, decomposition-based methods often lack physical grounding, and the direct identification and spatial localization of such subtle leakages remain insufficiently explored. This study proposes an integrated framework that combines signal decomposition with controlled field testing to enable precise detection of customer-side leakage within district metered areas. High-resolution smart meter data support refined hydraulic simulation and flow signal analysis. Controlled discharge experiments are conducted to induce representative leakage conditions, from which characteristic leakage signatures can be extracted from inlet flow series using variational mode decomposition with experimental support. The decomposed components are compared with consumption signals from all individual household meters via dynamic time warping to quantify similarity and locate leakage sources. The proposed framework is applied to a real district metered area in Danzhou, China, where minor customer-side leakages at household meter connections are successfully identified and verified through field inspection. The results demonstrate that integrating data-driven mode decomposition with experimental calibration provides a robust and transferable approach for proactive customer-side leakage management.