Seeing the unseen: StormHarvester and Anglian Water advance proactive sewer management

The background to this collaboration is one of long-standing challenges in the wastewater sector. Sewer networks are vast, hidden, and complex. In the case of Anglian Water, the network extends across the East of England, covering thousands of kilometres and serving millions of customers. Blockages are one of the most common threats to reliable performance, and they are also among the most damaging. When a blockage occurs, sewage can back up and cause flooding of properties or escape into rivers and streams, creating environmental harm and reputational damage. Historically, utilities detected blockages only after a problem had already manifested, such as a customer report of flooding or a pollution alarm. This reactive model left companies “blind” to the early stages of restriction build-up and meant that many incidents were only managed after negative impacts had occurred.

Real-time data from water level sensors across the sewer network was analysed by StormHarvester’s platform using machine learning algorithms

Sewer blockages are a leading cause of pollution incidents across the United Kingdom, typically accounting for around 40% of recorded events. Many of these are preventable, caused by everyday issues such as fats, oils and greases disposed of down drains, or wet wipes and other unflushable products entering the network. These materials accumulate in pipes, creating narrowing points and eventually full blockages. In many cases, a blockage begins as a minor restriction that, if cleared early, can be quickly resolved. Left undetected, however, it can rapidly escalate into a major incident, especially when heavy rainfall adds pressure to the system. Climate change and more frequent, intense storms are increasing this risk, making proactive management more urgent than ever.

Anglian Water’s challenge, therefore, was threefold. First, the company needed to detect developing blockages before they turned into incidents, creating the opportunity for crews to intervene while problems were still small. Second, it needed to prioritise high-risk areas—locations with a history of pollution or flooding, or where the consequences of an escape would be most severe. Third, it needed to move away from the traditional reactive model and build a solution that could scale across tens of thousands of kilometres of sewers, providing continuous visibility and actionable insight.

StormHarvester’s technology offered the capabilities to address these needs. At the heart of the approach was the deployment of tens of thousands of level sensors across Anglian Water’s sewer network. Each sensor provided continuous data on how water levels behaved in real time. StormHarvester’s cloud-based platform then analysed this information using machine learning algorithms. A critical element of the system was its ability to establish what “normal” behaviour looked like at each location. No two manholes are alike; flows vary depending on local population, industrial activity, and even daily routines. By learning the expected patterns for each site, the system was able to distinguish between normal fluctuations and abnormal rises that might indicate a forming blockage.

What made this analysis especially powerful was the integration of hyperlocal rainfall prediction. Sewer levels naturally rise during wet weather, and without a rainfall context, sensors alone can generate a flood of alarms that overwhelm operators. StormHarvester’s system combines machine learning with hyperlocal rainfall forecasting to establish normal behaviour at each site and to account for local weather. This allows the platform to silence alarms during expected wet-weather events and instead generate alerts only when unusual rises occur that cannot be explained by rainfall. The result is fewer false positives and a stream of concise, credible alerts that operators can trust and act upon.

When the system detected an anomaly, an alert was raised within Anglian Water’s operational systems. These alerts enabled teams to investigate early, often finding minor obstructions that could be cleared before they developed into major blockages. By acting on these insights, Anglian Water was able to restore normal service quickly and reduce the risk of escalation.

To highlight the scale of this achievement, Emily Timmins, Director of Water Recycling at Anglian Water, reflected on the milestone: “Reaching 5,000 proactive blockages identified on 23rd June marks a major milestone for our Dynamic Sewer Visualisation programme. This innovation is transforming how we prevent pollution and flooding-safeguarding both communities and the environment. It's also reducing the need for reactive responses, which is a huge win for the well-being of our teams. This is a proud moment that reflects our commitment to smarter, more sustainable water recycling."

By intervening earlier, Anglian Water not only reduced the likelihood of pollution and flooding but also saved time and resources

The impact of this work is particularly evident in high-risk locations. Of the blockages identified, hundreds were in areas prone to pollution and thousands in flood-risk zones. In the past, these could easily have resulted in incidents with serious consequences for customers and the environment. Instead, proactive detection meant they were managed in time. Encouraged by these outcomes, Anglian Water expanded the DSV programme from its initial deployment to cover 42,000 monitors, creating one of the largest proactive sewer monitoring estates in the UK. This expansion has driven a more than fourfold increase in proactive blockage clearance, firmly embedding the preventive approach across the company’s operations.

The programme’s success has relied not only on technology but also on collaboration and culture. StormHarvester worked closely with Anglian Water’s operational teams to integrate alerts into existing systems, avoiding the need for parallel processes. Crews on the ground adapted to a more data-driven model, where visits were guided by predictive analytics rather than routine scheduling or reactive call-outs. This has helped establish a sustainable, repeatable process that is now part of business-as-usual operations.

Examples from the field illustrate how this system works in practice. In one case in mid-July, a sensor detected a level breach despite no rainfall. StormHarvester’s system flagged the anomaly, and Anglian Water dispatched a crew the following day. They discovered rags causing an obstruction, cleared the blockage, and restored normal service. By the next day, sewer levels had returned to their expected behaviour. In another instance that same month, the system again detected an abnormal rise. Crews attended within two days, removed the blockage, and verified that levels had normalised. In both cases, potential pollution or flooding was prevented, and customers were unaffected.

Beyond operational outcomes, the Dynamic Sewer Visualisation programme has broader significance in the water sector. Utilities across the UK face increasing regulatory and public scrutiny regarding pollution incidents. Regulators such as Ofwat and the Environment Agency have set ambitious targets for reducing sewer overflows and improving environmental performance. Anglian Water’s experience demonstrates how data-driven monitoring and machine learning can help achieve these goals. By embedding predictive insight into daily operations, the company has created a scalable model for proactive sewer management that can serve as an example for others.

The benefits are environmental, operational, and financial. Early intervention reduces pollution risk and customer disruption. It also saves money by replacing costly emergency responses with quicker, targeted maintenance. Just as importantly, it builds a culture of prevention rather than reaction, where teams share a common view of network health and take pride in avoiding incidents before they occur. This cultural change is as valuable as the technology itself, ensuring the programme’s success endures.

Anglian Water’s Dynamic Sewer Visualisation programme, powered by StormHarvester’s analytics, has demonstrated what is possible when large-scale sensing is combined with intelligent data interpretation. With over 50,000 sensors installed, the company has moved decisively away from reactive response and towards proactive management. Thousands of potential blockages have been resolved before they could cause harm, with particularly strong results in high-risk areas. The outcome is a safer, cleaner network, better protection for customers and the environment, and a strong foundation for further innovation in wastewater management.