High order resolution and parallel implementation on unstructured grids

Yao, Yufeng (1996) High order resolution and parallel implementation on unstructured grids. PhD thesis, University of Glasgow.

Abstract

The numerical solution of the two-dimensional inviscid Euler flow equations is given. The unstructured mesh is generated by the advancing front technique. A cell-centred upwind finite volume method has been adopted to discretize the Euler equations. Both explicit and point implicit time stepping algorithms are derived. The flux calculation using Roe's and Osher's approximate Riemann solvers are studied. It is shown that both the Roe and Osher's schemes produce an accurate representation of discontinuities (e.g. shock wave). It is also shown that better convergence performance has been achieved by the point implicit scheme than that by the explicit scheme. Validations have been done for subsonic and transonic flow over airfoils, supersonic flow past a compression corner and hypersonic flow past cylinder and blunt body geometries. An adaptive remeshing procedure is also applied to the numerical solution with the objective of getting improved results.

The issue of high order reconstruction on unstructured grids has been discussed. The methodology of the Taylor series expansion is adopted. The calculation of the gradient at a reference point is carried out by the use of either the Green-Gauss integral formula or the least-square methods. Some recently developed limiter construction methods have been used and their performance has been demonstrated using the test example of the transonic flow over a RAE 2822 airfoil. It has been shown that similar pressure distributions are obtained by all limiters except for shock wave regions where the limiter is active. The convergence problem is illustrated by the mid-mod type limiter. It seems only the Venkatakrishnan limiter provides improved convergence. Other limiters do not appear to work as well as that shown in their original publications. Also the convergence history given by the least-square method appears better than that by the Green-Gauss method in the test.

• Item Type: Thesis (PhD)
• Qualification Level: Doctoral
• Subjects: T Technology > TL Motor vehicles. Aeronautics. Astronautics
• Colleges/Schools: College of Science and Engineering > School of Engineering > Autonomous Systems and Connectivity
• Supervisor's Name: Richards, Prof. B.E.
• Date of Award: 1996
• Depositing User: Elaine Ballantyne
• Unique ID: glathesis:1996-1974
• Copyright: Copyright of this thesis is held by the author.
• Date Deposited: 30 Jun 2010
• Last Modified: 10 Dec 2012 13:49
• URI: https://theses.gla.ac.uk/id/eprint/1974