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.

is free meshing or mapped meshing is best for flow problems

Please login with a confirmed email address before reporting spam

Hello,

I am trying in flow problems in a rectangular channel where incompressible navier stokes and conduction convection systems I am including.

I am in a confusion to choose mapped or free meshing in future. I did both in the same system but ending with huge different results.

I heard that mapped mesh is preferred than free mesh in any field, is it true?

also in free meshing triangular, tetrahedral and for mapped mesh rectangular, hexagonal meshes are obtaining. which pair is best suited for flow condition and heat transfer equations.

I have validate it with experimental results at the end in order to get good results suggest which type of meshes to prefer.

for me mapped meshes are comfortable than free meshes. because it is orderly arranged and easy for me to do post processing.

Also in free meshing oscillations are created in solutions and not in the case of mapped mesh.

My department colleagues are preffereing free mesh and avoiding mapped mesh because it gives wrong results, is it true?


thanks for you precious help in advance,

Regards
V Shanmugam

1 Reply Last Post Feb 19, 2010, 12:09 p.m. EST
Ivar KJELBERG COMSOL Multiphysics(r) fan, retired, former "Senior Expert" at CSEM SA (CH)

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago Feb 19, 2010, 12:09 p.m. EST
Hi

I do not believe its the free or mapped shape that is the true issue, I suspect more that its the final mesh quality and density and how it evolves (particularly with ALE deformation mode) that is the main difficulty.

Probably a free triangular mesh is more robust during ALE mesh deformations but again it depends on the motion and relative motion you consider. In anycase you must ensure that you have enough mesh elements to correctly resolve your field gradients all over the domains.

Then classically, for all my FEM models, I hit the "refine" once or twice, and rerun one case, just to ensure that my solution is "mesh independent" at least to some 5-10%. If the difference is larger my mesh is too coarse, independet if its a tri, teth, quad, or hexa.

Higher order mesh elements have more nodes and are normally of a higher order, so they take also more RAM and time to solve, but should normaly be of higher "quality", except if you do not respect at least the points above.

Not a definitive reply, but everything is not white and black ;)
Good luck
Ivar

Hi I do not believe its the free or mapped shape that is the true issue, I suspect more that its the final mesh quality and density and how it evolves (particularly with ALE deformation mode) that is the main difficulty. Probably a free triangular mesh is more robust during ALE mesh deformations but again it depends on the motion and relative motion you consider. In anycase you must ensure that you have enough mesh elements to correctly resolve your field gradients all over the domains. Then classically, for all my FEM models, I hit the "refine" once or twice, and rerun one case, just to ensure that my solution is "mesh independent" at least to some 5-10%. If the difference is larger my mesh is too coarse, independet if its a tri, teth, quad, or hexa. Higher order mesh elements have more nodes and are normally of a higher order, so they take also more RAM and time to solve, but should normaly be of higher "quality", except if you do not respect at least the points above. Not a definitive reply, but everything is not white and black ;) Good luck Ivar

Note that while COMSOL employees may participate in the discussion forum, COMSOL® software users who are on-subscription should submit their questions via the Support Center for a more comprehensive response from the Technical Support team.