On Tuesday 5 September, Turing researchers led a session on ‘Grand challenges in data-centric engineering’at the Royal Statistical Society Annual Conference, held at the University of Strathclyde in Glasgow.
The session, supported by the Lloyd’s Register Foundation and the Alan Turing Institute, presented fundamental research and ground-breaking innovation in the applications of data science to engineering. The projects aim to improve infrastructure resilience and society’s safety, while generating scientific advances that will have an international impact.
Programme Director, Mark Girolami gave an overview of the latest activities on the data-centric engineering programme, and made the case for why statisticians should be interested and get involved.
Data-centric engineering group leaders, Catalina Vallejos, Ricardo Silva and Din-Houn Lau all gave talks at the session.
Catalina Vallejos discussed her work on fault prediction for gas turbines in partnership with Siemens. The turbines are fitted with numerous sensors and the statistical methods she is developing (inspired by survival analysis in medical statistics) have implications for reduction of maintenance costs, as “routine” checks might not be needed if faults can be predicted. Improved detection of faults would also lead to enhanced safety.
Ricardo Silva described how Oyster card data can be used to model passenger movement on the London Underground. This can be used to predict passenger behaviour in the event of unplanned disruption and allow TfL to optimise their response to never-before-seen incidents (e.g. closure of a section of track at a specific time of day, which hasn’t previously happened).
Din-Houn Lau introduced novel statistical methods to monitor the condition of railway bridges, by embedding fibre optic sensors into the structure that detect tiny deformations. He is developing novel statistical methods to detect changes to the bridge’s behaviour over time, with a view to better scheduling maintenance checks.
At the RSS evening award ceremony, Turing Fellow Rajen Shah and Group Leader on the data-centric engineering programme Chris Oates, each received the biennial Research Prize, presented by RSS President (and data-centric engineering advisory board member) Sir David Spiegelhalter.
Rajen was recognised for “ground breaking methodology in high-dimensional inference, interaction search and goodness-of-fit tests”, while Chris was awarded for his “outstanding methodological contributions to computational statistics and his innovative contributions to bioinformatics and machine learning”.