Leak Detection and Localization in Water Distribution Networks Using Machine Learning

Water distribution networks, which provide clients drinking water, are prone to leaks. Leaks are a critical area of concern, as they cost billions of dollars each year, and can allow water-born pathogens into drinking water, which is a public health concern. Current leak detection methods are very labor intensive, as many of them require experienced operators, or require surveys to be filled out. In contract, Digital Water Solution’s hydrant.ai is a device that can be installed on fire hydrants, which, listens to these water networks using hydrophones, which are similar to microphones, but are designed to work well underwater. This research project will study the performance of these devices in detecting and localizing leaks across pipe networks that vary in shape. Further, a leak detection and localization algorithm will be developed, which will leverage recent advances in artificial intelligence and machine learning.

Faculty Supervisor:

Chul Min Yeum;Giovanni Cascante

Student:

Partner:

Digital Water Solutions

Discipline:

Engineering

Sector:

Transportation and warehousing

University:

University of Waterloo

Program: