PhD Studentship: Solving stormwater overflows through spatial catchment optimisation

Project Description

The aim of this PhD project is to reduce stormwater spills and flooding by developing an AI/ Machine Learning driven optimisation tool. This tool will enable research to identify catchment-scale stormwater management strategies using readily available data.

Storm water inflow to combined sewer systems reduces network capacity and increases the risks of environmental pollution, overflows, treatment costs and flooding. Risk is exacerbated by climate change, urban growth and an increasing regulatory and public spotlight on stormwater management.

To manage this exacerbated risk, we need to significantly boost capacity in the existing pipe network, however the traditional methods for achieving this by replacing/ upsizing underground pipes and storage can be expensive and disruptive. In response to this, contemporary stormwater management has increasingly proposed adding solutions outside of the network using green infrastructure, storage and SuDS. However, despite established technical understanding of how to build these solutions, their application remains fragmented and ad-hoc due to an inability to strategically screen and apply them systematically at a catchment scale.

This project will connect understanding of site focused storm water interventions towards a systematic catchment scale strategy through exploring and developing spatial optimisation methods using machine learning approaches. This will provide a more effective and synergistic implementation of interventions through optimising placement and configuration across a catchment.

The method developed to spatially optimise catchments will also contribute a novel scientific resource, unlocking potential for extensive future application to answer fundamental questions about how we can integrate and apply spatial solutions to manage urban water at scale.

This project will be supervised by Dr James Webber (supporting water engineering) and Professor Ed Keedwell (supporting ML implementation).

Advert information