A high-resolution pressure-driven method for leakage identification and localization in water distribution networks Ivo Daniel 1,2 , Jorge Pesantez 3 , Simon Letzgus 4 , Mohammad Ali Khaksar Fasaee 3 , Faisal Alghamdi 3 , Kumar Mahinthakumar 3 , Emily Berglund 3 , Andrea Cominola 1,2,* 1 Chair of Smart Water Networks - Technische Universität Berlin, Berlin, Germany 2 Einstein Center Digital Future, Berlin, Germany 3 Department of Civil, Construction, and Environmental Engineering - North Carolina State University, Raleigh, USA 4 Machine Learning Group - Technische Universität Berlin, Berlin, Germany * andrea.cominola@tu-berlin.de ABSTRACT Water losses are one of the main consequences of infrastructure failures in water distribution networks (WDNs). While background leakages and pipe bursts in well maintained systems generally amount to only 3-7% of the total water supplied, they can account for more than 50% for poorly maintained WDNs worldwide (Puust et al., 2010). Methods that support prompt detection and accurate localization of leakages are crucial to help water utilities implement timely mitigation measures and avoid unnecessary loss of water. Beyond the direct effect on reducing water losses, effective leakage management strategies can avoid revenue losses and other undesired effects, including contaminant infiltration, property damage, and WDN inefficiencies. Several works in the literature proposed approaches for leakage detection and localization. However, the ongoing digitalization of urban water systems (Makropoulos and Savic, 2019; Stewart et al., 2018), along with the development of distributed sensor networks and improved real-time communication, are fostering the development of a new generation of on-line, data-driven leakage identification methods that process data stored in the supervisory control and data acquisition (SCADA) system in real-time. A lack of studies that comparatively analyze and benchmark different leak identification and localization methods has culminated in the organization of the Battle of the Leakage Detection and Isolation Methods (BattLeDIM 2020; Vrachimis et al., 2019). The BattLeDIM is an international competition organized for the purpose of comparing the performance of different leakage detection and localization methods based on time-to-detection and location accuracy. This research develops a high-resolution pressure-driven method for leakage identification and localization in WDNs and tests the approach using the benchmark dataset provided as part of the BattLeDIM. Our method is composed of two modules that operate sequentially. The first module performs leakage event identification. Our leakage identification algorithm processes the pressure from SCADA data observed at different sensor nodes in a WDN and identifies time history of leakage events by analyzing pressure differences between pairs of nodes. The model is