Real time simulation finds important applications in finite element structural analysis within Civil Engineering. On several typical design projects, structural engineers will identify and evaluate a number of possible configurations for the structure they are designing, ranging in complexity from buildings to bridges. These configurations are subtle variations of a single design, for example, adjusting the height of storey columns in a building whilst keeping everything else the same. Each configuration is however an individual numerical model, typically high-dimensional, that must be analysed seamlessly in rapid succession. Each individual analysis informs about the fitness of the configuration allowing to pick the optimal under the underpinning engineering and budgetary constraints.
These numerical models, typically addressed via a finite element method, yield systems of equations with sufficient similarity, a fact that if exploited can expedite the design process beyond what is currently feasible. A promising approach is to exploit the low-dimensional structure of the underpinning elliptic models in order to reduce their dimensionality. The focus of the project is in developing algorithms rooted in numerical randomised linear algebra for efficient model-order reduction, with provable performance guarantees. The key challenge in this is to design the appropriate probability distribution to randomly sample the parameter space. If done optimally, a small percentage of the overall numerical operations suffice to approximate the low-dimensional surrogate models, which in turn speeds up the design workflow by 100 times of more. The approach is also amenable to couple with data-driven (model-agnostic) reduction methods. A faster design workflow leads to better optioneering and hence better designs that make more efficient and sustainable use of materials.
The project is in collaboration with Arup numerical analysis and simulation research.
Minimum entry qualification - an Honours degree at 2:1 or above (or International equivalent) in a relevant science or engineering discipline, possibly supported by an MSc Degree.
The desired candidate will have a background in mathematics, statistics, computer science, or engineering while knowledge of numerical computing will be considered as additional advantage.
Preference will be given to candidates with an MSc in a numerate discipline.
Further information on English language requirements for EU/Overseas applicants.
Applications are welcomed from self-funded students, or students who are applying for scholarships from The University of Edinburgh or elsewhere.