Data-Driven Digital Twins for Predicting Climate-Driven Deterioration in UK Bridges

About the Project

This PhD project will develop data-driven digital twin methods for predicting climate-driven deterioration in UK bridges. The project will combine civil engineering, machine learning, statistical modelling, and climate resilience to address an important challenge in infrastructure asset management.

Many UK bridges are ageing and are increasingly exposed to changing environmental and climate conditions. More intense rainfall, flooding, temperature variation, freeze-thaw effects, and extreme weather events can accelerate deterioration and increase uncertainty in long-term maintenance planning. Although bridge inspections generate valuable condition, defect, maintenance, and asset data, these records are often underused for predictive modelling and strategic decision support.

The successful candidate will explore how historical bridge inspection records, defect classifications, condition ratings, maintenance histories, structural characteristics, and environmental exposure data can be brought together to understand how bridge condition changes over time. The project will also investigate how climate-related variables may influence future deterioration patterns.

The main research aim is to create dynamic, asset-level digital twin models that can support bridge condition forecasting. These models will be designed to update as new data become available, helping to move bridge assessment from static inspection records towards more predictive and evidence-based asset management.

The project will examine both fully data-driven and hybrid modelling approaches. This may include machine learning, statistical deterioration models, uncertainty quantification, and physics-informed methods that incorporate engineering understanding of bridge behaviour and deterioration mechanisms. A strong focus will be placed on model interpretability, data limitations, uncertainty, and the practical value of the outputs for infrastructure owners and decision-makers.

The project is suitable for applicants with a background in civil engineering, structural engineering, data science, computer science, or a closely related field. The final research direction will be refined with the successful candidate, depending on their skills, experience, and interests.