Anglian Water extends AI monitoring across more than 250 assets
A pump's first warning can be read from the cabinet that powers it, rather than from a sensor mounted on the machine. Anglian Water has signed a two-year extension to its Samotics framework, covering more than 250 clean water assets already monitored by the Sam4 predictive-maintenance system. The deal will also bring new pumping stations being commissioned across the Strategic Pipeline Alliance into the programme.
Sam4 uses electrical signature analysis, or ESA — reading current and voltage signals — from the motor control cabinet. Its industrial AI analyses those signals from pumps and motors continuously and remotely, without sensors installed on the equipment itself. Anglian Water says the system can identify emerging cavitation, bearing wear and misaligned components before faults escalate.
The reported numbers give the programme its scale: 97% detection accuracy and fewer than 1% false positives. Since adopting Sam4 in 2020, Anglian Water says early fault detection and avoided failures have prevented £886,800 in costs. Those figures come from the utility's account of its own deployment, not an independent comparison across the water sector.
The concrete gain is earlier, better-planned maintenance. Anglian Water's teams can use the warnings to decide when to intervene, increase asset uptime, make better use of engineers' time and reduce the risk of disruption. For customers, the intended effect is less exposure to failures at pumping stations and a more dependable water service as the Strategic Pipeline Alliance adds new sites.
Andy Smith, Anglian Water's smart water systems senior manager, said the company's principle was “right maintenance, right time”: understanding what assets are reporting and acting before performance or service is affected. Tom Swain, Samotics' head of UK and Ireland, said the partnership began in 2020 and is now entering its sixth year. Sam4 is therefore a deployed maintenance tool at Anglian Water.
Comments
Loading the thread…
Sign in to leave a comment. Sign in