This PhD studentship in Computer Science will combine modelling and simulation techniques with machine learning to develop a novel Digital Twin approach for predicting failures in key infrastructure assets. It is located in the School of Computing, host of the National Innovation Centre for Data, in collaboration with the School of Engineering, which coordinates the ACHILLES programme on infrastructure resilience.
There are thousands of kilometres of roads, railways and flood defences in Great Britain; Network Rail alone manages almost 200,000 earthworks. Failures in these assets are common (twice per week) and costly (failure costs dwarf maintenance costs), but the timing of failures is currently unpredictable. Increasingly extreme weather patterns will cause ever more failures in aging infrastructure and new methods of prediction are required for management of assets. Existing computer models are used to produce weather projections, predict water flows, and simulate the effects of water pressure on individual assets, however these models are disconnected from each other and real-time data. A promising area of research is in Digital Twins —digital counterparts of real systems— that combine computer models with real-time data and machine learning to produce calibrated simulations for reliable, real-time predictions about maintenance needs and possible failures. In this project you will develop a novel Digital Twin for resilient geo-infrastructure and demonstrate the approach on test assets owned and monitored by the University.
You will develop the knowledge and expertise necessary to succeed in an interdisciplinary research environment. In partnership with your supervisory team, you will develop a Personal Development Plan, identifying areas for development and training. You will have the opportunity to attend international summer schools, workshops and courses, and to participate in Masters-level training, including MSc modules from Geotechnical Engineering and Computing. These will cover slope stability theory, modelling techniques, machine learning and data science.