Yorkshire Water has launched a two-year research project with UnifAI Technology to develop an AI system capable of predicting water quality at inland bathing sites in near real time.
The project, funded through Ofwat’s Water Breakthrough Challenge, will monitor 20 bathing water locations across Yorkshire, including Ilkley, Wetherby and Knaresborough. Partners include The Rivers Trust, BSI, Southern Water and environmental monitoring company SOCOTEC UK & Ireland.
Sensors installed at each site will record water quality data every 15 minutes, measuring factors such as dissolved oxygen, pH, ammonia, temperature and turbidity. The data will be combined with laboratory testing of water samples taken four times a week to detect bacteria including E.coli and Enterococci.
Researchers hope the project will generate more than 7,800 datasets over 24 months, allowing machine learning models to predict harmful bacteria levels faster than current testing methods, which rely on laboratory analysis and can take time to produce results.
Yorkshire Water said the aim is to provide swimmers and other river users with quicker access to water quality information through a future public web app, if the technology proves successful.
The company said the project could also help support future environmental monitoring requirements under the Environment Act, which will require continuous monitoring around combined sewer overflows by 2035.
According to the press release, earlier AI trials at Warleigh Weir and Bournemouth Boscombe achieved accuracy levels of 87%, although Yorkshire Water said the new model will require two years of testing before its performance can be fully assessed.
