Leakage Prediction PhD

PhD Leakage prediction

PhD Leakage prediction

"Leakage prediction and localisation using Artificial Intelligence"

Better management of leakage helps to protect the environment and reduce costs through better use of existing water resources and reductions in energy usage.

SWW has good levels of asset data and continuously collects operational data that is being used to improve the localisation and prediction of leaks. This project aims to address business requirements to improve the detection, localisation and prediction of leaks in the organisation and to reduce penalties or increase rewards from the Ofwat ODI scheme

This PhD project is developing new machine learning and optimisation techniques to improve data-driven leakage detection and to support strategic decision making and crew deployment to localise leaks in the real world.  

Team members

• Matthew Hayslep, PhD Candidate
• Professor Edward Keedwell, Professor of Artificial Intelligence (UoE)
• Professor Razieyh Farmani, Professor of Water Engineering (UoE)
• Josh Pocock, Leakage Reporting Manager (SWW)

Funding

• South West Water

Publications

This work has led to two publications with a third currently in review.

• (Best Paper Nominee) Multi-Objective Multi-Gene Genetic Programming for the Prediction of Leakage in Water Distribution Networks' 
• 'Understanding district metered area level leakage using explainable machine learning'

 In progress: 'An Explainable Machine Learning Approach to the Prediction of Pipe Failure using Minimum Night Flow'