Leakages in drinking water networks cause massive global water losses of 120 million cubic metres every year. These leakages have serious consequences in terms of operational failures, health risks, and damages to the surrounding infrastructure. Our project on intelligent online leak detection “iOLE” aims to identify and localise leakages as quickly, automatically, and robustly as possible to sustainably reduce their negative impact. The main focus is on optimal user experience (UX) both when applying data science algorithms and when being operated by drinking water utility staff.

In the iOLE project, we combine two award-winning leakage detection algorithms to provide greater accuracy in locating leakages by addressing different requirements. The LILA algorithm from the Technical University of Berlin uses AI and pressure sensors to locate leakages in the vicinity of the sensors. The Dual Model, co-developed by KWB Hydroinformatics Group Lead David Steffelbauer, uses information from hydraulic network models to locate leakages down to the level of a single pipe. By combining both algorithms in iOLE, a broader spectrum of requirements is covered at the technological level, and a high level of automation and reliability is guaranteed thanks to holistic software development.

The operation of these data-driven algorithms usually requires expertise in programming and data analysis. However, for widespread implementation of our leakage detection technology in drinking water utilities, usability by operators and maintenance staff is crucial. Therefore, we have developed a comprehensive and intuitive visualisation with GIS data to support decision making and ensure confidence in our algorithms.

The iOLE project represents a significant step forward in water supply technology by combining state-of-the-art leakage detection algorithms with a user-friendly application. Through the intuitive application and automation that our system offers, drinking water utilities can detect and repair leakages more efficiently and accurately. This leads to improved water supply security and reduces the environmental impact of water leakages. By translating the complexity of data science into a user-friendly format, we’re contributing to making drinking water supplies more sustainable and future-proof.