Event-driven simulations for computational engineering
Dr M N Campbell-Bannerman
Dr J Gomes
Applications accepted all year round
Self-Funded PhD Students Only
This project is to develop new models, techniques, and software for the simulation of physical systems using the particle dynamics approach. These techniques are well-known and are at the heart of a huge range of applications, from molecular simulations of biological proteins, crushing and grinding of ore, to ragdoll physics within computer-game engines. The approach models the physical world by breaking it down into a large number of both rigid and soft dynamic bodies. The forces between these bodies are then calculated and integrated to determine the motion of the system through time.
As part of this project, training in a wide range of particle dynamics techniques will be given; however, this project will focus in particular on a technique known as event-driven particle simulation. This technique is based on impulses rather than forces, and is both extremely efficient and fundamentally stable. This technique is currently under heavy development and research here at Aberdeen, where the leading event-drive simulation package, DynamO, is developed (http://dynamomd.org).
Depending on the particular skills and motivations of the student, this project will focus on one or more of the following areas:
• Generalisation of event-driven models. Most event-driven simulations are performed using simple models such as the hard-sphere model. While this is extremely efficient, more realistic models are required for simulations of granular or molecular systems.
• Parallelisation of event-driven methods. While there are descriptions of parallel approaches within the literature, there are no widely available implementations. This is due, in part, to the inherent serial nature of the approach as events must be processed in order. New techniques must speculatively process events in parallel and provide for a method to either correct errors or restart calculations from a known safe state. Development of these approaches should lead to wide adoption of event-driven techniques. The student will be expected to develop and implement parallel approaches on local and national High Performance Computing (HPC) systems.
• Coupling of event-driven methods with other techniques. While event-driven particle dynamics excels at the simulation of rigid and soft bodies, there are physical effects which are best calculated using other techniques. One example is the interaction of fluids with the dynamic bodies, where the fluid is best simulated using CFD techniques and the two simulations coupled. There are some excellent particle-fluid methods which may provide the coupling (such as DSMC); however, these approaches are relatively new and need further development.
The successful candidate should have, or expect to have an Honours Degree at 2.1 or above (or equivalent) in Engineering, Physics, Computational Science with knowledge of Programming, algorithms, fundamental physics.
There is no funding attached to this project it is for self-funded students only.
Formal applications can be completed online: http://www.abdn.ac.uk/postgraduate/apply. You should apply for PhD in Engineering, to ensure that your application is passed to the correct College for processing. Please ensure that you quote the project title and supervisor on the application form.
Informal inquiries can be made to Dr M N Campbell Bannerman, ([email protected]) with a copy of your curriculum vitae and cover letter. All general enquiries should be directed to the Graduate School Admissions Unit ([email protected]).