Engineering: Fully Funded EPSRC DTP PhD Scholarship: The effect of length scales and data resolution on image-based simulation
This scholarship is funded by the EPSRC Doctoral Training Partnership.
This PhD opportunity is a joint project with industrial sponsor The Manufacturing Technology Centre (MTC).
Start date: January 2020
Subject areas: Computational modelling, Non-destructive testing, experimental engineering, advanced manufacturing
Conventional simulations like FEA and CFD are the de-facto industry tools for predicting component performance during the engineering design phase. Models are drawn to be geometrically ideal using computer aided design (CAD), therefore assuming that all parts within a batch are identical. Once manufactured, parts are qualified fit for purpose with experimental tests representative of in-service conditions because they rarely perform identically to idealised drawings.
A proposed improved approach is part-specific simulation which has applications in high-value and specialist manufacturing wherever there is non-negligible variability from one component to another e.g. additive manufacturing or composites. One method of creating part-specific models is to construct detailed simulation geometries from micro-resolution 3D images (e.g. X-ray CT or laser scanning). As a contrast to design-based models, these image-based FEM simulations treat each part as unique. The main benefit is that these part-specific models simulate components ‘as-manufactured’ rather than ‘as-designed’ for improved accuracy.
Through advancements in imaging, increasingly higher resolution simulations are possible. However, as resolution increases new challenges emerge. This industrially facing project will investigate how length scale effects and data resolution impact the potential use of image-based modelling for the industrial sector. The candidate will work as part of a research team (involving partners from academia, the MTC, our national institutes (e.g. STFC) and industrial partners in the fusion energy and satellite manufacturing sectors) to address challenges such as:
FEA assumes a material continuum. As finite elements get smaller, they approach the material’s granular scale. The discontinuities at grain interfaces mean that this assumption doesn’t hold true and the behaviour of single grains is different to the homogenised bulk properties.
Also, other length scale effects (e.g. surface tension) may need consideration. For example, the brittle behaviour of a tungsten rod compared with a pliable tungsten filament like in an incandescent light bulb.
Increasing resolutions require additional computational power. By using supercomputing this is now possible within reasonable timeframes, but it comes with an associated financial cost. Due to diminishing returns it is worthwhile considering what level of accuracy gives best added value to current methods.
The successful candidate will have a good undergraduate degree in a relevant subject, e.g. engineering, materials science, physics or computer science. Previous specialisation in image-based simulation is not required as the student will gain this expertise while working closely with the industrial sponsor ‘The MTC’. Through use of some of the most advanced facilities in the world, this project will provide the opportunity to develop transferable skills in industrial processes; advanced manufacturing, non-destructive testing and evaluation; material and component qualification; cutting-edge engineering simulation – thus ensuring the candidate is prepared for a wide range of possible career paths after graduation.
Location: Zienkiewicz Centre for Computational Engineering, Bay Campus, Swansea University
Supervisors: Dr Llion Evans (Swansea/UKAEA), Professor Perumal Nithiarasu (Swansea), Dr Nick Brierley (The MTC)
Candidates should hold a first or upper second class honours degree (or its equivalent) or a Master’s degree in a subject area related to the project (e.g. engineering, materials science, computer science, mathematics and physics).
A strong background in experimental work is required. Experience with simulation is desirable, particularly engineering analysis e.g. FEA/CFD. Good programming skills, Python, CUDA, C/C++ or Fortran is preferred.
Studentships funded by EPSRC are subject to UK/EU residency eligibility.