In past posts, various aspects of CFD meshing were examined. The three main categories of mesh cell types—tetrahedral, hexahedral or polyhedral—were reviewed and related to hybrid meshes, which use a combination of these. Next, specialized schemes, such as sliding and dynamic meshes, were considered and finally adaptive meshing was introduced as a powerful technique to quickly achieve accurate CFD simulations. However, a very important aspect of any CFD solution is whether it is independent of the mesh resolution.
Obviously, the entire mesh can be refined, but this can be a time-consuming process, especially for large 3D grids and systems involving several physical phenomena (turbulence, reactions, heat transfer, etc.). Instead, adaptive meshing can be used since the refinement of the mesh in critical regions will determine if the solution is changing as the mesh is refined.
Finally, it is also important to realize that even if a CFD solution is mesh independent, it still might not be accurate, as the underlying physics may not be captured by the numerical models used and thus comparison with actual experimental results is always a preferred final validation of any simulations.