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 - DC Mesh Quality (http://www.cfd-online.com/Forums/ansys-meshing/78315-ansys-meshing-dc-mesh-quality.html)

alven299 July 18, 2010 09:15

ANSYS Meshing - DC Mesh Quality
 
5 Attachment(s)
Hi all,

I have a problem in getting a good mesh quality for my geometry. I am dealing with a half section of a distillation column (DC) tray. The tray as shown in "Base geo" has vast number of sieve holes. I separated the volume into two section (as indicated by the different colors) for the ease of sizing along the vertical axis. I managed to mesh the geometry, however, with poor quality especially in the core region. The meshes look alright at first but as it departs further from the base, instead of structured hexahedral meshes, the core region is dominated by unstructured tetrahedral meshes. I have attached several pics.

Following are my global sizing specs:

Advanced Size Function = Fixed
Relevance centre = fine
Smoothing = high
Transition = slow
Max face size = 0.012 mm
Max tet size = 0.012 mm
Growth rate = 1

I first suspected the max face size and max tet size as the problems but altering these also resulted in the same observations.

Does anyone know how what might be the cause to this problem and perhaps how to rectify this problem?

Any inputs are very much appreciated.

alven299.

sac July 19, 2010 07:21

Quote:

Originally Posted by alven299 (Post 267883)
Hi all,

I have a problem in getting a good mesh quality for my geometry. I am dealing with a half section of a distillation column (DC) tray. The tray as shown in "Base geo" has vast number of sieve holes. I separated the volume into two section (as indicated by the different colors) for the ease of sizing along the vertical axis. I managed to mesh the geometry, however, with poor quality especially in the core region. The meshes look alright at first but as it departs further from the base, instead of structured hexahedral meshes, the core region is dominated by unstructured tetrahedral meshes. I have attached several pics.

Following are my global sizing specs:

Advanced Size Function = Fixed
Relevance centre = fine
Smoothing = high
Transition = slow
Max face size = 0.012 mm
Max tet size = 0.012 mm
Growth rate = 1

I first suspected the max face size and max tet size as the problems but altering these also resulted in the same observations.

Does anyone know how what might be the cause to this problem and perhaps how to rectify this problem?

Any inputs are very much appreciated.

alven299.

What method are you using?

alven299 July 19, 2010 08:36

Hi sac, I'm using sweeping method for the top half of the geo and hex dominant for the bottom half of the geo.

PSYMN July 19, 2010 09:36

Methods...
 
It looks the other way around... It looks like sweep in the thin bottom part with the holes but hexa dominant for the top part (where the problems are)...

Sweep starts by quad meshing the one side and then "sweeping" that thru to give you a hexa mesh.

Hexa Dominant starts with quads on the 3D surfaces and then tries to march those quads in toward the middle. For Convex shapes like this, it must eliminate Hexas as it moves inward and it tries to fit things together with tets and pyramids closer to the middle... I suspect you are using an older version. Hexa Dominant is much better at 12.1 than it was at 10.0.

You could perhaps try sweep for both parts?

You might also try MulitZone for both parts (combine them into a multibody part). It can do a multiple source and target sweep (like the gambit cooper tool).

alven299 July 19, 2010 10:33

Thanks simon for the input. Initially, I tried to sweep the geo (with holes in it), but it did not allow me to use this method. I think perhaps the holes make the geo not sweepable. Therefore I have no choice but to use hex dominant method to mesh. At least now I know what's going on with the centre part. As for multizone method, will it be fine if I choose all of the holes as multiple sources? or should I just leave it as automatic? One thing I observed with multizone method is that it takes unusually long time to mesh. Is it normal or my specifications (min size) are too small?

PSYMN July 19, 2010 11:34

They have worked on Multizone a lot for 13.0... (Such as speeding it up for large numbers of blocks, improving automatic source and target selection, etc.)

I think it would work with 13.0, but I am not sure about 12.1 (it was still pretty new then).

You would have to give it a try and see what happens.


All times are GMT -4. The time now is 14:08.