BACKGROUND
- One of the UK’s largest private water utilities.
- 10+ million customers.
- 31k+ km of infrastructure.
- 500k+ sensors and CPPs.
THE CHALLENGE
A large UK water utility sought to consolidate multiple independent data sources used to determine the level of risk in the water system. Operators in the control room had no singular view of the network, instead having to derive the state of risk by looking at various sources aggregated in different ways and by different tools.
THE SOLUTION
ExploreAI helped the utility to implement the Hydraulic Network Risk Tool, or HNRT, platform. The HNRT uses the network’s pressure and flow data from district meters and CPPs. Those inputs are combined with customer contacts – when a customer reports an issue or submits a complaint – to show all relevant assets on a map, including valves, reservoirs, pipes, meters, and digital assets. HNRT also ingests data from the utility’s operational platform, which provides the states of actuated items like whether, for example, a valve is open or closed. Customer property data from yet another system is also included and shows details like the number of properties in a DMA.
PROJECT HIGHLIGHTS
- ~24-month project.
- Six data sources, including Pi.
- 1B+ rows of data ingested annually.
RESULTS
- £7M saving by predicting a large water outage and enabling prevention.
- The prediction helped reduce the number of properties affected and the time impact for those affected.
- 3.4 ML/day saved in leakage, resulting in annual savings of £1.3M+.
- The prediction helped reduce the number of properties affected and the time impact for those affected.
– Head of Data FactoryThe ExploreAI team has worked alongside us, to challenge and steer us and help us develop our own data science skills.
