AI that helps water sensors check themselves.
The Self-Calibrating Sensor Network for Sustainable Water Management (SCSN) project investigates how AI can help water sensors identify drift, support correction and indicate when cleaning or recalibration may be needed. Led by Southern Water, it is supported by the Ofwat Innovation Fund through the Water Breakthrough Challenge 4 Catalyst stream.
- Status
- Research and development
- Lead partner
- Southern Water
- Programme
- Ofwat Breakthrough Challenge 4 Catalyst
The reliability of the input matters
A sensor reading can change because the water has changed, but it can also change because the instrument has drifted, fouled or needs attention.
The programme installs and studies multi-parameter monitoring equipment alongside laboratory sampling so those different causes can be investigated.
Compare the physical sensor with a virtual estimate
ANNA can develop virtual sensors that estimate a parameter from relationships across the rest of the network. Comparing the physical reading with the AI-derived estimate creates a way to investigate drift, correction and maintenance needs.
A collaborative research programme
Southern Water leads the programme, with work involving UnifAI Technology, WATR, Wessex Water and other water-sector and standards partners. It combines sensor deployment and laboratory sampling with AI model development, sensor assurance and evaluation for wider use.