Sensat algorithm predicts water leaks

Photo: Pupunkkop |

Sensat is piloting an AI-based leak detection service with north-west water company United Utilities to help it meet its 15% leakage reduction target by 2025.

The pilot is using unmanned aerial vehicle (UAV) technology to map thermal data, which is then combined with high-resolution photogrammetry. A probability-based algorithm then assesses vegetation stress, terrain relief, and thermal scans to predict where leaks might occur along trunk mains in rural areas.

The algorithm should remove the need for manual inspections, which will reduce the time and cost of detecting leaks. Over time, the data collected will help improve the algorithm’s accuracy further.

Sensat CEO James Dean said: “When you consider that the reliable supply of water underpins almost every aspect of human life and the threat of climate change will mean we will have to prepare for a drier future, it’s vital that we work together to tackle the challenges that are affecting the water industry.”

This is not the first time the water company has trialled new tech to tackle water leaks – in the past it has used specially-trained dogs to sniff out leaks. United Utilities also teamed up with inventors to develop FIDO, now in use globally, which uses machine-learning and AI to find the size and location of a leak by the noise it makes.

Story for BIM+? Get in touch via email: [email protected]

Latest articles in News