Multi-fidelity aerodynamic modelling of competition cyclists.

点击翻译按钮获取中文摘要

This project focuses on advancing the state of the art of computational aerodynamics in the challenging domain of competition cycling aerodynamics, in close collaboration with aerodynamicists in the British Cycling Team (BCT). We will be provided with 3D scans of the athlete in different positions, as well as wind tunnel data at a range of speeds (NDA signed). Such flows are characterised by complex 3D moving geometries and transitional Reynolds numbers, both of which present significant challenge to turbulence modelling. We will develop a multi-fidelity approach to provide accurate prediction at a reduced computational cost; consisting of the combination of steady state (RANS) modelling at moderate mesh resolution, turbulence scale resolution (LES) at high mesh resolution and experimental data. High-fidelity simulations will be combined with experimental data via field inversion, to fill gaps and ensure key features are correctly captured. Resulting datasets will be used to improve prediction of lower-order models in order to efficiently explore parameters such as rider position. Where modelling is insufficient or prohibitively expensive - such as with regions of transitional flow - we will develop a framework to incorporate flow vis images using data-assimilation to fix transition location. Aside from pushing the state of the art in multifidelity modelling, this project has potential for very high impact via close collaboration with the British Cycling Team in the run up to Olympic Games in Paris and LA. It would also provide exciting publicity for the University to help promote our engineering research and teaching in aerodynamics.