Discussion Closed This discussion was created more than 6 months ago and has been closed. To start a new discussion with a link back to this one, click here.
Mesh refinement issues
Posted 12 mars 2010, 21:57 UTC−5 1 Reply
Please login with a confirmed email address before reporting spam
I am working on a simple natural convection problem in comsol, weakly compressible navier stokes and heat transfer. I am trying to match results for the same simulation done in another computational tool.
Using the default meshing tools at the third mesh level my model converges nicely and the results are fairly consistent with the benchmark simulation (amplitude and period of the oscillations). When I further refine the mesh the behavior becomes inconsistent and is not expected based on literature trends from multiple authors. This continues for all finer meshes.
Has anyone had similar problems with comsol? Any thoughts on what might be causing this issue? Mesh refinement studies are standard practice for simulation accuracy so I thought the community might have many thoughts ....
Thanks in advance.
Using the default meshing tools at the third mesh level my model converges nicely and the results are fairly consistent with the benchmark simulation (amplitude and period of the oscillations). When I further refine the mesh the behavior becomes inconsistent and is not expected based on literature trends from multiple authors. This continues for all finer meshes.
Has anyone had similar problems with comsol? Any thoughts on what might be causing this issue? Mesh refinement studies are standard practice for simulation accuracy so I thought the community might have many thoughts ....
Thanks in advance.
1 Reply Last Post 13 mars 2010, 02:16 UTC−5
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
Hi
so far for me (but I'm mostly in structural + thermal less often in NS) mesh refinements has always improved results, never diverged.
But you must have something special in your model as you talk about oscillatins with convection, for me convection is non-oscillatory by nature, so you are cenrtainly checking a particular situation with spcific BCs, could that be involved ?
Good luck
Ivar
so far for me (but I'm mostly in structural + thermal less often in NS) mesh refinements has always improved results, never diverged.
But you must have something special in your model as you talk about oscillatins with convection, for me convection is non-oscillatory by nature, so you are cenrtainly checking a particular situation with spcific BCs, could that be involved ?
Good luck
Ivar