Digital platforms and machine learning models

Project overview

Yorkshire Water was keen to explore how digital platforms and machine learning (ML) models could help it to make more effective operational and investment decisions. Our software solution, Helix, provided a platform that could be implemented rapidly and configured flexibly, could access data from different sources and had the proven ability to scale to an enterprise level.

Our role

We used our Helix digital platform and the machine learning models hosted on it to help Yorkshire Water progress on its digital journey by focusing on a high-profile pumping station asset. The Helix dashboard concisely displays key metrics to help guide decision-making, resulting in improved process compliance, optimisation and troubleshooting. We brought together different real-time datasets from multiple sources, so all the information was available in one location. Powerful ML models were deployed to identify data of interest, helping staff to focus their attention where it mattered and ensuring that important events were not missed. The improved understanding of the asset’s performance enabled longer range evidence-based maintenance window forecasts.

We deployed several ML models to identify anomalous equipment behaviour and to predict the performance of the asset. The ML models detect blockages and rising main bursts earlier, and this forecast of asset overload enables proactive stakeholder engagement. This not only reduces the likelihood of asset failure but also improves the public perception of the asset and improves stakeholder relations.

The team’s expertise and collaborative style of working meant that the users were empowered to discover the insights within their data and share these with their teams. We worked in an agile way with various Yorkshire Water subject matter experts to deliver new insights in each sprint.