Finalist award: £49,750
Winner award: £449,844
Led by: Mounce Hydrosmart Ltd and partners (University of Sheffield and RBMTS Ltd)
The ACQUIRE project will utilise and exploit Drinking Water Inspectorate (DWI) reports regarding water quality incidents submitted by water companies to mine the complex knowledge and understanding buried within them. These reports form a valuable resource to explore for understanding and evaluation of alternative solutions. The outcomes will include an interactive tool and open source portal enabling rapid identification of previously effective interventions for a given unfolding incident, and exposing those to be avoided. The solution is based on the latest ‘generative AI’, which enables transformation of lengthy and complex technical reports submitted for compliance to achieve proactive best management.
ACQUIRE was announced as one of 10 winners in the Water Discovery Challenge on 20 February 2024 – read more in the announcement blog.
The quality of drinking water is paramount to protecting public health and well-being. In drinking water distribution systems, when a hazard occurs from catchment to consumers’ taps, it is considered an incident. In more recent years the proportion of incidents occurring has increased and includes microbiological contamination as well as discoloured water. Whilst it is imperative that the risk and occurrence of these is minimised the range of causes and actions needed to mitigate them are extremely complex and interactive. Water companies must submit detailed records of water quality incidents to the regulator – the Drinking Water Inspectorate (DWI).
Our Innovation is in the application of AI to acquire actionable insight and analysis by unlocking the vast potential of DWI water quality incident reports. Since the DWI reports are a trusted source originating from all companies, they are effectively ‘open data’ thus promoting collaboration and alleviating confidentiality concerns. This solution provides a decision support tool using AI that enables water companies to deal more effectively with water quality incidents by yielding deeper insight and allowing the evaluation of solutions. It could also help with longer term strategic planning about the best course of action, by informing response plans.
During the finalist period, we will be working on an updated and expanded Case Based Reasoning (CBR) tool to incorporate ten years of DWI reporting and adding new features. We will use the latest Generative AI to help evaluate DWI outcomes (rather than needing an expert to manually perform this). A more user friendly chat based interface to the CBR will be explored. By the end of the project we aim to have a prototype demonstrator and provision of a (restricted) prototype portal to water companies.
