Report Open Access

A high-resolution pressure-driven method for leakage identification and localization in water distribution networks

DANIEL, Ivo; PESANTEZ, Jorge; LETZGUS, Simon; KHAKSAR FASAEE, Mohammad Ali; ALGHAMDI, Faisal; MAHINTHAKUMAR, Kumar; BERGLUND, Emily; COMINOLA, Andrea

Water losses are one of the main consequences of infrastructure failures in water distribution networks (WDNs), accounting for more than 50% in some WDNs worldwide. Methods that support prompt detection and accurate localization of leakages are crucial to help water utilities implement timely mitigation measures and avoid unnecessary losses of water and revenues. This research develops a high-resolution pressure-driven method for leakage identification and localization in WDNs and tests it using the benchmark dataset provided as part of the "BattLeDIM", an international competition on leakage detection and localization. Our method is composed of two modules that operate sequentially. The first module performs leakage identification by processing pressure data observed at different sensor nodes in a WDN and by analyzing pressure differences between pairs of nodes. The model is trained using pressure data observed for a “normal” time period, i.e., without leak events occurring in the WDN. The reconstruction error from the model is then analyzed to detect the anomalies on a time series that potentially includes one or more simultaneous leak events. The leakage annotations from the leakage identification module are then used by the second module, which performs leakage localization. A simulation-optimization framework is adapted to locate leakages by iterative mixed-integer linear programming. The framework relies on the pressure response of a set of candidate pipes when a fixed leak value is inserted there. The candidate pipes are selected based on the pressure drop reported by the sensors, and candidate pipes are included in a search area located around the most affected sensor. Our proposed method is tested on the SCADA data provided with a 5-minute sampling frequency for the benchmark medium-sized WDN of L-Town. Preliminary experiments show that our pressure-driven method can promptly detect and localize most of the labeled leakages and it is suitable for real-time applications.

Files (355.7 kB)
Name Size
Leakbusters.pdf
md5:547adf603d0e05219403358f60b87960
355.7 kB Download
530
403
views
downloads
All versions This version
Views 530530
Downloads 403403
Data volume 143.4 MB143.4 MB
Unique views 457457
Unique downloads 371371

Share

Cite as