- This event has passed.
Data Week 2021: Engineering with Digital Twins
June 16, 2021 -10:00 am - 11:00 amFree
This session explores the fundamentals of digital twins in an environment where equipment integrity, operation, and safety are traditionally supported by engineering simulations.
Backed by a wealth of experience across industries, we will cover how it works, benefits, challenges, and tips on doing it successfully, with Q&A. We particularly welcome questions about challenges from your industry. No expertise required to attend.
Engineering simulations help us turn information we have into information we want, to help us make decisions. In this context, these decisions might be about when to carry out maintenance, or when to replace equipment. This reduces cost, as equipment can be run longer or in more extreme conditions; it can improve safety, reducing the likelihood of unexpected failure.
High degree-of-freedom models can provide accurate simulations of systems from first principles, such as FEA, CFD and EM solvers. With the right validation, these provide effective ways of turning selected input data into more useful data for life predictions. The issue with these is usually one of speed, limiting the use of simulations to single case analyses or short selections.
An engineering simulation can be seen as an expensive and complex mapping function, taking a small set of inputs to deterministically produce a large set of outputs. Treating the complex simulation as a ‘grey-box’ control system, pulling together techniques of system classification, machine learning, data-driven methods and classical engineering physics, a minimal alternative to the mapping function can be found: our Reduced Order Model or Digital Twin. These can drastically reduce simulation times to generate new results from new inputs, from hours to milliseconds.
This physics-based approach to system classification is key to providing confidence in safety, particularly in a regulatory environment.
The speed of these methods opens the possibilities of new outcomes, such as real-time monitoring and probabilistic assessments. Counter-intuitively these simple models, when used right, can increase the fidelity of predictions.
About the speaker
Danny Thomas (Norton Straw): Is a multi-disciplinary consulting engineer with wide-ranging experience in simulations for structural integrity and creating engineering software.
Norton Straw, Innovation members of the South West Nuclear Hub, are consultants with deep technical expertise in engineering, applied mathematics and technical software, based in Bristol and Derby.
Registration for this event is via Eventbrite. It will take place online.
Data Week Online 2021
This event is part of Data Week Online 2021, organised by the Jean Golding Institute. Running from Monday 14 June - Friday 18 June 2021, this will be the 4th annual week-long series of workshops, talks and events in the world of data science and data-intensive research.
Find out more about the Jean Golding Institute by watching their short video below!