Research within the Computational Science and Engineering Group embeds fundamental developments in algorithms and cutting edge computational modelling within a framework that is guided by end-user pull. We primarily focus on fast, efficient and reliable numerical algorithms for the computational solution of partial differential equations (PDEs), and on the challenge of scale in scientific and information visualization. Members of the group also contribute to the School’s research into Applied Computing in Biology, Medicine and Health.
In PDEs we have a long-standing international reputation for work on unstructured mesh adaption and mesh quality. Furthermore, in collaboration with both industrial and academic partners, this research has resulted in computational techniques, and software, that has been widely applied: in areas such as combustion, lubrication, atmospheric dispersion, river and harbour flows, and many more.
Scientific visualization (scivis), deals largely with numeric data, for example from climate simulation, geophysical surveys or medical scanners, whereas information visualization (infovis) works with discrete/symbolic data, for example relational databases, biological networks, or web logs. Our visualization research offers three main contributions:
(i) computational topology and other mathematical tools supporting analysis and abstraction,
(ii) graphical interfaces, including large scale displays, for working with and navigating through data,
(iii) systems and software for exploration and analysis of large datasets.
We run a regular seminar series featuring talks from group members, related researchers across the university and also external speakers. Our facilities include two wall-sized ultra-high resolution displays, and virtual reality peripheral devices and displays. We also make substantial use of the Leeds’ high-performance computing (HPC) resources. The Computational PDEs Unit provides research and consultancy in the form of problem solving expertise and software to industry and to academic researchers alike.
We often have opportunities within the group, especially for highly motivated graduates with a good first degree (or Masters) in a mathematical, computational or engineering discipline, wishing to undertake studies towards a PhD. Please contact one of us for further information.
Examples of the importance and impact of our research are as follows:
Fundamental work on moving finite elements, adaptive meshes and the computational simulation of free-flowing surface phenomena, in collaboration with groups in Engineering, has commercially-important applications in printing on textiles using ink-jet technology and the spin-coating of non-smooth surfaces.
A long-standing collaboration with Shell Global Solutions has led to the development of state-of-the-art software for commercially-important lubrication problems. This research programme has included the development of efficient solvers, parallel computing strategies and the use of Grid technologies.
The Multifield Extension of Topological Analysis (META) project will address a major shortcoming in visualization techniques, the limited tools available for analysing multiple fields of data, and the result will deliver a new technology for understanding data in the physical sciences, engineering and medicine. Academic beneficiaries will include visualization researchers and end-users, and computational science communities.
Our virtual microscope research aims to develop a visualization system that out-performs the conventional microscope for cancer diagnosis. A wall-sized version of our virtual microscope is already in use for specialist histopathology training in St James' Hospital. A by-product of our research is that we have developed user interfaces that allow users to efficiently navigate collections of gigantic digital images. The research was one of "100 ground breaking pieces of research" that were featured in Research Council UK's Big Ideas for the Future report (see p27).
Virtual reality and numerical modelling research into tolerance stack-up has been exploited by Icona Solutions, a spin-out company from the School of Computing. Icona's software allows designers to visualize the impact of manufacturing variation on the perceived quality of products, and Icona's customers read like a who's who of the automotive industry (Porsche, Mercedes-Benz, Bentley, JCB, etc.). The original research was led by Prof Peter Dew.