RESEARCH

Accelerated aircraft noise prediction for trajectory optimisation

Aircraft noise in the vicinity of airports is both a nuisance and has significant implications for the health of residents. The noise generated by aircraft as they arrive at or depart from airports can be measured but available data is limited to points where measurement equipment can be placed. The distribution of noise over a wider region is almost impossible to measure but offers important data for trajectory optimisation and noise minimisation.

​

Fly-over noise may be modelled and simulated instead. In-house software called FLIGHT can be used to simulate fly-over noise for a given trajectory and aircraft/environment configuration. However, the process is slow due to the complexity of the models involved and the spatial resolution of the process. The faster-than-real-time generation of noise map data would enable real-time trajectory optimisation and hence the minimisation of noise for residents on a aircraft-by-aircraft basis.

​

This project explores the potential of GPU-accelerated simulation in the context of fly-over noise prediction. The project aims to develop software for GPU which accurately predicts aircraft fly-over noise measurement. The accelerated noise map data will then be used to perform automatic trajectory optimisation and confidence analysis.

​

Applicants should have a strong knowledge of geometrical and analytical acoustic prediction techniques. In particular, they should be familiar with ray-tracing. GPU-programming experience is desirable but not essential. Applicants should also have strong experience of fluid mechanics modelling.

Real-time virtual reality simulation for engineering design and analysis

Engineering simulation is a critical activity for design and analysis. However, conducting accurate simulations requires a substantial investment in terms of time for meshing, simulating and post-processing as well as access to suitable computing resources including supercomputing facilities. There is significant industrial interest in the development of cost-effective, simulation tools for rapid analysis which sacrifice accuracy for speed. These tools can be used upstream in design processes for parameter space exploration and can be used to direct more detailed studies later in the design cycle.

​

This project builds on existing work in the research group which has developed a real-time, interactive simulation tool based on the novel combination of GPU-accelerated flow simulation using NVIDIA CUDA and the Unreal Engine 4 game engine. CAD geometry can be imported into a virtual wind tunnel and fluid dynamics predicted and visualised around the geometry in real time. The user can interact with the geometry and the simulation using the HTC Vive Virtual Reality headset and controllers. This project aims to improve the performance of the solver-engine integration, expand the visualisation options and user experience and increase the fidelity of the modelling offered by the solver. Application areas beyond engineering are to be explored. The development of such an interactive tool is unprecedented and will amount to a world-leading contribution to the field of interactive simulation.

​

Ideal applicants should be enthusiastic about both physical simulation as well as computer games and related technologies. Some prior experience with C/C++ programming is essential. Applicants should have a strong knowledge of fluid mechanics including computational fluid dynamics and mathematical modelling. Prior experience of GPU programming is desirable but not essential. Applicants must have a strong desire to learn and be able to work independently and have excellent written and verbal communication skills.

Interactive engineering simulation using mobile device clusters

In this age of pervasive computing, most individuals will carry around with them a mobile phone or tablet which possess the same amount of computing power as a desktop machine less than a decade ago. The ability to harness the computing power of a classroom or office for simulation is an attractive, low-power alternative to traditional simulation hardware.

​

This project will develop an existing P2P communication library along with a simulation kernel written for mobile GPUs using OpenGL Shading Language to enable high-performance, low-power, scalable simulation across local collections of mobile devices running the Android OS. The project will require the development of code in Java and GLSL and applicants should have strong programming skills and a desire and aptitude for learning new languages if required. There is also the option of including real-time interaction capabilities using aspects of the Android API.