AI-enabled fire hydrants could help detect water leaks

Researchers in Canada have developed smart acoustic sensors that can pinpoint small leaks in water mains and can easily be retrofitted onto fire hydrants

Spotted: Water systems are a necessary but often messy and costly part of city infrastructure. We’ve covered a wide range of innovations aimed at keeping city water flowing, such as porous concrete that reduces dangerous runoff. Now, researchers at the University of Waterloo in Canada have developed a rubber robot designed to detect leaks in pipes.

The team have developed a system that combines AI algorithms with an acoustic sensor as an alternative, affordable method of detecting leaking pipes. The sensors are trained to listen to the sound of water flowing through a pipe. The AI algorithms are then able to distinguish between a fully sealed pipe and one that is leaking even a small amount of water.

The sensors are easily retrofitted onto existing fire hydrants. As fire hydrants are part of the existing water network that sits above ground, no excavation or shutting off of the water supply is required.

Currently, most leaks aren’t detected until it’s obvious that the water is where it isn’t meant to be. This can have serious implications; water supplies can become contaminated and building foundations rotten. Moreover, water waste is a significant financial burden on cities. This new acoustic system detects very small leaks at up to seventeen millimetres a minute, but researchers are still working on refining the sensors to pinpoint exactly where the leaks have occurred. The research, led by Roya Cody, is published in the Urban Water Journal.