Using digitalisation for a real-word implementation of an early warning system for water leakages

Abstract

A timely detection and localisation of leakages is of highest interest to reduce water losses. As stated in the literature, most leakage detection and localisation methods consider unrealistic assumptions and are rarely assessed under real-world conditions. To overcome these limitations, an early warning system for new leakages is implemented in a real-world case study. Therefore, the case study is equipped with digital water meters at customer sites to monitor household water demands at 15-min resolution. These data are subject to the European General Data Protection Regulation, which requires the active customers consent (and therefore some refusals), and are also subject to data loss during transmission due to the use of wireless data communication. The incomplete data (costumer refusals and transmission failures) were then used for the early warning system, including data-based detection and model-based localisation, and tested on 10 engineered leakage events. Thereby, small leakages up to a size of 0.4 l/s could be detected reliably even under high demand fluctuations. Additionally, the search area for the fine search on site could be reduced to 10–40% of the entire district metered area for leakage sizes between 0.1 and 2.0 l/s also under these real-world boundary conditions.