Greater insights on existing data is helping to reduce the excess volumes of water that can affect communities in Gloucester. Arup worked with a tech start-up to create a greater picture of Inflow and Infiltration (I&I) into the wastewater network. The aim was to reduce the need for costly, carbon-intensive storage tanks and additional sensors, while reducing both environmental impact and capital costs. The project gives Severn Trent Water a holistic understanding of the catchment enabling prioritisation of interventions. This supports a targeted approach to investment that accommodates the highly changeable nature of I&I, allowing STW to adapt and pivot its strategy.

A hidden problem

Annual report
Annual Report FY 2024/25
A high angle view of a building.

Inflow and Infiltration can reduce the capacity of sewers, increasing the likelihood that they could overflow. One solution – storage tanks – are costly and carbon-intensive. Severn Trent Water faced an 80,000m³ storage target in Gloucester, but had limited space for the required tanks across the catchment. An innovative, fast, data-led approach was needed.

A 2D visual showing how AI detects and interprets Inflow and Infiltration (I&I) in wastewater networks.
A 2D visual showing how AI detects and interprets Inflow and Infiltration (I&I) in wastewater networks.

Getting the most out of data

The overall aim of the project was to understand where I&I is present to prioritise where to invest and where not to. While technologies such as machine learning or AI could help identify patterns, we first needed to ensure we had good-quality data. We partnered with InfoTiles, using PipeFusion, a machine learning algorithm, to review and update Severn Trent Water’s Geographic Information System data. We also used  Sewer Intelligence to quantify I&I volumes, and the associated economic and carbon costs. Existing records were combined with additional datasets, including telemetry, rainfall, river gauge data and pump activations. This allowed us to estimate I&I by zone and produce data-enabled insights.

By becoming data-powered people, we’re turning insight into action, accelerating outcomes and raising the standard of what we can achieve together

Rebecca Sweeney

Waste Infra Design Team Leader, Severn Trent Water

Scaled up and future-proofed

In just 16 weeks, we modelled I&I across this section of Severn Trent Water’s network. This gave strategy and engineering teams a holistic understanding of the catchment, the risk associated with each pipe and the identification of I&I within each zone. In addition to providing better insights, having the data within one platform means that Severn Trent Water engineers can easily detect patterns such as the interaction with rainfall and river levels. Ongoing monitoring will enable Severn Trent Water to monitor changes within the catchment and pivot their strategy for I&I reduction accordingly.

The success in Gloucester has laid the groundwork for scaling this approach across Severn Trent’s entire network. We are already onboarding another UK water company to benchmark and refine our models, ensuring adaptability to diverse network conditions. Our ambition is to establish a national standard for AI-driven I&I management, which promises to reduce storm overflow spills and enhance resilience across the UK water sector. By embracing digital engineering, we are not just solving today’s challenges – we are shaping the future of water infrastructure.