ESR12 | Alexandra-Ioanna Liosi

High fidelity simulations of open wheel geometries using high order methods

ESR12 | Alexandra-Ioanna Liosi

National Techincal University of Athens

Host Institution: McLaren Racing Ltd

Phd awarding institution: Imperial College London

Master Title: Integrated Master's Degree in Mechanical Engineering

Research Interest: Numerical modelling, software development and validation for complex, physical systems with primary focus on racing/automotive applications, especially Computational Fluid Dynamics (CFD) using high fidelity methods. Further interests include CFD-based optimization methods for engineering applications.

Some of the complex flow phenomena in open-wheel geometries, like bluff body wakes and multiple vortex interactions, require more advanced turbulence modelling than RANS to be properly simulated. Hence, the need for LES modelling, and high-order spectral/hp methods can provide an efficient way of running LES simulations when compared to more traditional finite volume approaches. High-order methods can resolve many scales across a fluid’s flow.

Additionally, they can be a favourable option for the most common compromise between the computational model’s accuracy and its cost. Particularly, in engineering applications where a certain level of accuracy is required, spectral/hp methods can provide the desired results with reduced computational time as shown in [1]. However, the adoption of high-order methods in the industrial workflow gives rise to various challenges as it is essential to standardize and automate the circular process of design, simulation, and evaluation, to keep up with the fast-paced environment of manufacturing companies. Therefore, this project is focused on developing and suggesting the best practices for obtaining fast and robust results from implicit LES simulations within the framework of spectral hp/methods. These will be achieved by investigating, among others, the numerical schemes and solution algorithms as well the mesh resolution requirements and their relationship with the polynomial order to enhance the efficiency of the computation.

These techniques will be tested on geometries with varying complexity and primarily focusing on the sub-systems of a F1 race-car. Lastly, a database will be generated which will include the results of the examined configurations and can be used for extending the research on industrializing Spectral/hp methods. This project will be performed using the open-source framework Nektar++, which is currently developed by the Imperial College London and the University of Utah, in collaboration with the CFD Methodology Group of McLaren Racing Ltd. [1] Karniadakis, G., & Sherwin, S. (2005-06-02). Spectral/hp Element Methods for Computational Fluid Dynamics. : Oxford University Press. Retrieved 31 Jan. 2022, from https://oxford.universitypressscholarship.com/view/10.1093/acprof:oso/9780198528692.001.0001/acprof-9780198528692

[wpdm_login_form]