CFD Online Discussion Forums

CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   ANSYS Meshing & Geometry (http://www.cfd-online.com/Forums/ansys-meshing/)
-   -   [ANSYS Meshing] Hex Plus Body of Influence (http://www.cfd-online.com/Forums/ansys-meshing/89206-hex-plus-body-influence.html)

slowtype June 7, 2011 15:02

Hex Plus Body of Influence
 
Hello,

I have invested a number of hours (actually days) into ANSYS 13 mesher and have hit a road block.

I am modeling fluid flow, typically in cylindrical vessels. I would like to experiment with hex meshing, but am having difficulty. I am only able to get Hex Dominant to work which I understand is not good for CFD.

Also, I use body of influence to increase the mesh count around areas of interest but it appears this option is unavailable when using any of the meshing "methods". I can not find any documentation about this.

Thanks

PSYMN June 8, 2011 10:41

Yea, body of influence is more of a Tetra/Prism thing... Only tetras can be refined so easily.

For CFD Hexa, you should be looking at the sweep or multizone methods.

slowtype June 8, 2011 11:13

are there alternatives to the body of influence function that can be used with hex?

PSYMN June 8, 2011 11:28

If your solver (like CFX) requires a conformal mesh (no hanging nodes), ICEM CFD has a refinement option that creates lots of mini CGrids to connect the larger mesh with the finer mesh... This requires a 3 to 1 size transition.

For solvers like Fluent that support hanging node configurations, many hex meshers (ICEM CFD, TGrid, Gambit, etc.) allow 2 to 1 refinement.

Otherwise, you can do things like refine a patch on the surface so that the resulting swept mesh is finer, or manually create a topology such that the mesh is finer in certain areas just due to the compressed or converged topology...

What you have hit on here is a primary drawback for hexa mesh. Although it is more efficient at a given size, it can be less efficient if you only need refinement in a few key areas.

slowtype June 9, 2011 15:48

I'm using OpenFOAM solver.

I found that cutcell can get me a decent looking mesh. But I have to use separate bodies with finer meshes in the areas of concern.

I can't figure out what to name the faces on the separate bodies so that they can be ignored by OpenFOAM. When examining the .msh file it appears that ANSYS mesher is declaring the faces as walls even though I try naming them internals or interior. Also all the bodies are set to fluid.

Any ideas how to name faces in ANSYS mesher so that they remain as internals/interiors after getting exported as a .msh file.

slowtype June 20, 2011 09:45

Figured out my own problem.

The separate bodies can not intersect any other face or body.


All times are GMT -4. The time now is 14:29.