A generalised lattice Boltzmann method with block-structured adaptive mesh refinement
A generalised lattice Boltzmann method with block-structured adaptive mesh refinement
The Lattice Boltzmann method (LBM) is a relatively new development in computational fluid dynamics (CFD). Instead of approximating the Navier-Stokes equations, the approach is based on solving a simplified version of the Boltzmann equation on a specific discrete space. It can be shown via a Chapman Enskog expansion that for vanishing Knudsen number the LBM recovers the Navier-Stokes equations. The standard LBM is limited to equally spaced Cartesian grids, making this approach notably expensive for capturing thin boundary layers, and therefore impractical for most relevant problems in aerodynamics. In this work, conventional numerical methods are implemented to solve the two-dimensional discrete-velocity Boltzmann equation in generalised curvilinear coordinates to simulate fluid flows with non-Cartesian grids. Several test cases are used for verification and validation, and the results have been extensively compared with the available numerical and experimental literature with very favourable outcome confirming the feasibility of the implemented methodology. The generalised LBM resolves large gradients in the wall vicinity with fewer mesh elements than the Cartesian LBM; therefore, it substantially reduces the computational effort. Flows over the 2D circular cylinder and the NACA0012 aerofoil are specifically investigated to assess the accuracy and performance of this approach for external aerodynamics. Moreover, the generalised LBM has been implemented into AMROC (Adaptive Mesh Refinement in Object-oriented C++) software infrastructure, a fully parallelised finite-volume framework that uses a block-structured adaptive mesh refinement. This development is a new class of LBM scheme with block-structured adaptive mesh refinement on curvilinear grids. The latest is implemented with the mapped mesh strategy, and various test cases are solved for validation, including heat transfer flows. Overall, this project enhanced the current capabilities of the standard LBM by combining the generalised LBM with AMROC capabilities.
University of Southampton
Reyes Barraza, Juan
5d754742-de9f-47e5-a5f1-10327f04d437
November 2021
Reyes Barraza, Juan
5d754742-de9f-47e5-a5f1-10327f04d437
Deiterding, Ralf
ce02244b-6651-47e3-8325-2c0a0c9c6314
Reyes Barraza, Juan
(2021)
A generalised lattice Boltzmann method with block-structured adaptive mesh refinement.
University of Southampton, Doctoral Thesis, 208pp.
Record type:
Thesis
(Doctoral)
Abstract
The Lattice Boltzmann method (LBM) is a relatively new development in computational fluid dynamics (CFD). Instead of approximating the Navier-Stokes equations, the approach is based on solving a simplified version of the Boltzmann equation on a specific discrete space. It can be shown via a Chapman Enskog expansion that for vanishing Knudsen number the LBM recovers the Navier-Stokes equations. The standard LBM is limited to equally spaced Cartesian grids, making this approach notably expensive for capturing thin boundary layers, and therefore impractical for most relevant problems in aerodynamics. In this work, conventional numerical methods are implemented to solve the two-dimensional discrete-velocity Boltzmann equation in generalised curvilinear coordinates to simulate fluid flows with non-Cartesian grids. Several test cases are used for verification and validation, and the results have been extensively compared with the available numerical and experimental literature with very favourable outcome confirming the feasibility of the implemented methodology. The generalised LBM resolves large gradients in the wall vicinity with fewer mesh elements than the Cartesian LBM; therefore, it substantially reduces the computational effort. Flows over the 2D circular cylinder and the NACA0012 aerofoil are specifically investigated to assess the accuracy and performance of this approach for external aerodynamics. Moreover, the generalised LBM has been implemented into AMROC (Adaptive Mesh Refinement in Object-oriented C++) software infrastructure, a fully parallelised finite-volume framework that uses a block-structured adaptive mesh refinement. This development is a new class of LBM scheme with block-structured adaptive mesh refinement on curvilinear grids. The latest is implemented with the mapped mesh strategy, and various test cases are solved for validation, including heat transfer flows. Overall, this project enhanced the current capabilities of the standard LBM by combining the generalised LBM with AMROC capabilities.
Text
PermissionToDepositThesis_formReyes
Restricted to Repository staff only
More information
Published date: November 2021
Additional Information:
Parts of this work have been published as:
J. A. Reyes Barraza and R. Deiterding. A Lattice Boltzmann method in generalized curvilinear coordinates, pages 477–488. International Centre for Numerical Methods in Engineering (CIMNE), 2019
J. A. Reyes Barraza and R. Deiterding. Towards a generalised lattice Boltzmann method for aerodynamic simulations. Journal of Computational Science, 45:101182, 2020
Identifiers
Local EPrints ID: 457805
URI: http://eprints.soton.ac.uk/id/eprint/457805
PURE UUID: fc1c54da-384b-4dcb-8f49-6d65c5fac588
Catalogue record
Date deposited: 16 Jun 2022 17:05
Last modified: 17 Mar 2024 03:39
Export record
Contributors
Author:
Juan Reyes Barraza
Download statistics
Downloads from ePrints over the past year. Other digital versions may also be available to download e.g. from the publisher's website.
View more statistics
