Researchers from UCLA, led by Pranav Agrawal and Sriram Narasimhan, developed an innovative acoustic solution for leak detection and localization in complex water distribution networks. Using hydrophones mounted on fire hydrants, the team has created advanced machine learning and probabilistic algorithms that analyze acoustic signals produced by leaks. With U.S. utilities losing approximately 6 billion gallons of water daily due to leaks, this technology aims to enhance water conservation efforts, particularly in regions facing significant water shortages.
The methodology involves collecting acoustic data, extracting features from it, and implementing sophisticated models for accurate leak identification. Although detecting leaks in simple pipes is relatively straightforward, urban networks present a challenging environment due to their complexity. This system facilitates monitoring in hard-to-access areas without necessitating expensive excavation, allowing sensors to be repositioned as needed.
Their device is commercially available through Digital Water Solutions and has already been deployed in multiple locations across Canada and the U.S. This solution not only promises to save millions of gallons of water but also reduces the energy and resources spent on treating and delivering water. The innovative research will be presented at the 184th Meeting of the Acoustical Society of America in Chicago.
