Statistical Challenges in Data-Centric Engineering
Dr Victoria Stephenson (University of Bath)
Seeking Resilience for Historic Engineered Structures; Working with Vague and Uncertain Information
As data-driven tools and techniques become more commonplace in the management of the built environment, so resilient built systems are promoted through the improved performance evaluation offered by digital technologies and automated, real-time assessment. Application of these data-driven techniques to heritage structures is often not a viable solution, and as such other statistical and analytical methods need to be employed to better understand and manage these ageing systems. Limited and vague information is often available about the design, construction and current condition of historic engineered structures. As such a high degree of uncertainty underpins assessment of these physical systems. Added to this, current and future uncertainty about the environmental conditions in which these structures are needed to perform, compounds the issue of predicting and managing them, and through this hinders resilience strategies. This presentation explores this problem via a number of case studies involving historic masonry structures in the UK, through work carried out with Dr. Chris Oates (Newcastle University).
Dr Amy Wilson (University of Edinburgh)
Statistical modelling for assessing the risk of electricity shortfalls in Great Britain
A reliable electricity supply is a key consideration for energy system planners. In Great Britain, a long-term planning study known as the electricity capacity assessment is performed annually to assess the long-term risk of insufficient electricity generation to meet demand. The need to meet climate targets has resulted in a rise in wind and solar generation in Great Britain. As wind and solar generation are weather dependent, the effect of this rise has been an increase in variability in the availability of electricity generation. An additional difficulty is that wind, solar and demand are correlated. New statistical methods that account for this correlation and increased variability are therefore needed to assess the long-term risk of electricity shortfall. This presentation will discuss statistical modelling for the electricity capacity assessment study. Focus will be on models for demand and wind. A new model which makes use of extreme value theory to capture the behaviour of demand-net-of-wind at the extremes will be presented and compared to existing approaches.
Dr Matthew Revie (University of Strathclyde)
Reflections on Data-Driven Decision Making within Engineering Organisations
Based on recent work with Scottish Water, Ministry of Defence, BAE SYSTEMS, Scottish Power, SSE and Bruce Power, Matthew will reflect on some of the existing challenges facing many Engineering organisations that are moving towards a data-centric approach to decision-making. A high level overview of these projects and the modelling involved will be provided, providing the basis for drawing out similar challenges across different sectors. These challenges will focus across a broad spectrum of organisational, data and modelling challenges.
This event is sponsored by the Alan Turing Institute - Lloyd's Register Foundation programme on data-centric engineering.
Attendance is free but for catering and room booking purposes we ask you to register here.
Tea, coffee and biscuits will be available between the talks
Organiser Name Kevin Wilson
Organising Group(s) RSS North Eastern Local Group and RSS Emerging Applications Section