Abstract The choice of the best mesh in terms of cost, time and accuracy of computational solutions in the CFD industry is a challenging topic and a subject of some controversy. Generating meshes based on hexahedral elements requires significant time and effort, however, these meshes are claimed to produce high quality solutions. Meshes that employ tetrahedral elements can be constructed much faster in complex geometries, but may increase the levels of numerical diffusion. The objective of this study is to better establish quantitative assessment of the influence of cell geometry in the computational mesh on the CFD results of pollutant dispersion around buildings in order to help modelers to choose the most effective mesh type for their applications. In order to achieve this objective, two widely used mesh styles, i.e., hexahedral-based and tetrahedral-based meshes, are considered in the simulation of this flow problem. Quantitative grid convergence was calculated based on a grid convergence index (GCI). The mesh style was found to have an observable effect on the calculated pollutant concentrations. For instance, the hexahedral-based mesh was observed to have GCI values that were in an order of magnitude below the tetrahedral-based mesh values for all resolutions considered, even in the very fine tetrahedral-based mesh. Furthermore, the GCI value, and hence the truncation error, remains high compared to conventional hexahedral cases. The study recommends taking special care when employing an unstructured tetrahedral-based mesh to ensure that the mesh is fine enough and any numerical errors should be documented for selected variables reported analogous to experimental uncertainty in order to assess the quality of the numerical solution.
[1]
K. Ghia,et al.
Editorial Policy Statement on the Control of Numerical Accuracy
,
1986
.
[2]
Jiri Blazek,et al.
Computational Fluid Dynamics: Principles and Applications
,
2001
.
[3]
D. Spalding,et al.
A calculation procedure for heat, mass and momentum transfer in three-dimensional parabolic flows
,
1972
.
[4]
Shuzo Murakami.
Computational wind engineering
,
1990
.
[5]
N T Frink,et al.
Recent Progress Toward a Three-Dimensional Unstructured Navier-Stokes Flow Solver
,
1994
.
[6]
P. Roache.
Perspective: A Method for Uniform Reporting of Grid Refinement Studies
,
1994
.
[7]
L. Richardson.
The Approximate Arithmetical Solution by Finite Differences of Physical Problems Involving Differential Equations, with an Application to the Stresses in a Masonry Dam
,
1911
.
[8]
R. Ooka,et al.
Studies on critical Reynolds number indices for wind-tunnel experiments on flow within urban areas
,
2003
.
[9]
T Frink Neal,et al.
A Fast Upwind Solver for the Euler Equations on Three-Dimensional Unstructured Meshes
,
1991
.
[10]
Weeratunge Malalasekera,et al.
An introduction to computational fluid dynamics - the finite volume method
,
2007
.