Leak Detection Methods in Water Distribution Networks: A Comparative Survey on Artificial Intelligence Applications

Essentially, water is an extremely vital resource for human beings. However, each year, a significant amount of water is lost because of leakages in multiple water distribution systems. From this perspective, much ink has been spilled upon the issue of water leakage detection and location. Indeed, since the emergence of data and interest in the development of artificial intelligence (AI) techniques, leak detection and location solutions have been optimized. This survey aims to present a comprehensive review of leak detection and location techniques in water distribution networks (WDNs). The different categories of leak detection and location solutions are set forward, in particular the intelligent ones. A comparative study between AI algorithms is performed using scenarios from the LeakDB data set. To our knowledge, this is the first work that uses a common benchmark data set to offer a comparative experimental study of the most used algorithms in leak detection. The selective choices of scenarios and experiments grant a deep understanding of the leak detection works, as well as a support for future research to develop artificial intelligence methods in this area.