Active and passive acoustics to detect and localize leaks in water distribution networks

Problem Statement

Estimates put anywhere between 20-30% of water is lost between treatment and delivery in most urban water distribution systems. Such loss is a huge loss of a valuable resource, not to mention the waste of energy that goes into sourcing, treating, and delivering this water. Detecting and locating leaks is notoriously difficult as many leaks remain undetected and underground and could run for days or weeks before they surface.

 

Approach

Our approach senses the acoustic energy resulting from the leak using passive acoustic sensors or combined with an active sonar source to detect and localize leaks. These sensors are located at the network’s fire hydrants and other natural locations. Acoustic data is processed using new signal processing techniques and machine learning methods we are currently developing, e.g., time delay estimation, time-frequency methods, support vector machines, and unsupervised clustering methods.

 

Ongoing Work:

Through funding from the US Navy, we partnered with Digital Water Solutions to pilot this technology at the Naval Base in Ventura County. In this project, we are developing and testing sophisticated machine learning and probabilistic methods to detect and locate leaks in complex pipe networks. Both active and passive techniques are being employed toward this goal.

 

Key Results and Visuals:

Figure 1. (a) Representative water distribution system where this technology is being applied; (b) bespoke acoustic monitoring system being deployed in a fire hydrant (also shown are the sensing, ADC, storage and backhaul modules); (c) leak detection accuracy from machine learning classifiers as a function of size during simulated testing; and (d) time delay estimation using mean shift clustering for the case of active leak detection.

 

Relevant Publications

  • Kafle, M. D., Fong, S., & Narasimhan, S. (2022). Active acoustic leak detection and localization in a plastic pipe using time-delay estimation, Applied Acoustics, 187, 108482.
  • Cody, R. & Narasimhan, S. (2020). A field implementation of linear prediction for leak-monitoring in water distribution networks, Advanced Engineering Infor- matics, 45, 101103.

See Publications for a complete list.

 

Students

Gabriel Earle

Stan Fong