Tracking down water leaks
United Utilities was the UK’s first water provider to introduce sniffer dogs and now it’s turning to artificial intelligence (AI) to help track down leaks.
The North West water company has teamed up with FIDO Tech to develop artificial intelligence (AI) capable of hunting leaks both on pipe and in pipe.
As well as a five-year framework contract with United Utilities, the British start-up recently secured £1.5 million in investment from MicroMega Holdings.
The development emerged from United Utilities’ Innovation Lab annual programme designed to help emerging tech into the UK water sector.
Selected projects undergo an intensive 10-week incubation period, working with United Utilities’ teams.
The news follows another AI solution to come out successfully from UU’s Innovation Lab - Canadian start-up Emagin, which was later acquired by Innovyze.
AI as part of a “Systems Thinking” approach
The adoption of AI is part of United Utilities’ “Systems Thinking” approach to manage its water and wastewater networks.
This includes new and emerging technology to monitor interactions between every individual element of the company’s regional system.
The resulting big data will then be analysed in real-time.
The FIDO AI platform and in field device uses machine-learning to ‘listen’ and interpret the unique data trail left by leaks.
It then tracks them down using smart digital triangulation to pinpoint their exact location.
As the AI-library of leak data grows, FIDO said its system learns to become “more effective at finding tricky leaks – for instance, those in noisy areas where traditional listening methods are difficult”.
The hardware-neutral AI platform analyses both the audio files of a utility network’s existing estate of acoustic loggers as well as FIDO device data from in field.
The company boasts an accuracy rate of over 85 per cent.
A game-changer in leakage detection
United Utilities network and capital delivery director Kevin Fowlie said FIDO has the potential to be a game-changer in leakage detection, especially as many leaks never show themselves above ground, so are difficult to find.
“We have thousands of loggers and monitors installed across our network to detect the barely perceptible sounds and vibrations which can be indicators of leaks, but it’s difficult for humans to interpret which data is a leak, and which could be a pump, traffic noise or something else,” he said.
Fowlie added: “FIDO gathers all this data into its AI leaks library and it learns fast. It can even calculate, with increasing accuracy, how severe the leak is, helping us prioritise. It really is revolutionising our response to leaks.”
Victoria Edwards, co-founder and CEO of FIDO, said: “Regardless of whether you’re using your own data loggers in New Zealand, a FIDO listening stick in Sudan, or a FIDO Pro in pipe in Wigan, everyone can access the FIDO AInytime platform and our deep learning, leak library. This means, the world over, we can now all play our part in helping to detect leaks and save water.”
