Discretization error with Richardson extrapolation.
I am a student and for my seminar work I have to do some simulations in CFX. Besides other requests I have to compute discretization error of the mesh with Richardson extrapolation. I have found and studied some theory about Richardson extrapolation however I still don't know how to approach and how to perform this in CFX?
Any hint or some example would be greatly appreciated.
The textbook "Computational Fluid Dynamics" by Roache is a key work in this area. Also articles like this http://journaltool.asme.org/Template...umAccuracy.pdf contain useful advice.
Ideally you take a mesh, then sub divide it (halving the element edge length) and halve it again to get three different meshes. If possible you go further. But often this results in meshes too fine to simulate so it is frequently approximated by remeshing with mesh refinement less than a global 2. Also another approximation which is often done is to refine only the area of interest and leave the rest of the simulation domain with the existing mesh. This makes the mesh refinement study much more tractable, but then introduces approximations and requires some skill and knowledge to make sure it is done appropriately.
|All times are GMT -4. The time now is 14:54.|