|
[Sponsors] |
[ICEM] Recommened order for building volume mesh |
|
LinkBack | Thread Tools | Search this Thread | Display Modes |
August 26, 2009, 07:04 |
[ICEM] Recommended order for building volume mesh
|
#1 |
Senior Member
Stuart
Join Date: Jul 2009
Location: Portsmouth, England
Posts: 740
Rep Power: 26 |
Hi,
I've made a mesh of a wing using the Octree method and applied smoothing to get a Quality of over 0.3. 8 elements are between 0.25 and 0.3, but I guess that's not too bad out of about 14 million. Anyway, which of these two approaches would be best? 1). Run a new volume mesh using the TGrid option from the Octreee surface mesh, smooth that (if needed) and then apply the prism layers. 2) Apply the prism layer, then re-generate the volume mesh using TGrid and smooth if needed? Thanks Last edited by siw; August 26, 2009 at 11:52. |
|
August 27, 2009, 10:16 |
|
#2 |
New Member
Join Date: Aug 2009
Posts: 10
Rep Power: 17 |
I turned to hexa meshes long time ago, but I would choose first variant
|
|
August 27, 2009, 12:59 |
First one is better, but for the tetra, not the prisms.
|
#3 |
Senior Member
Simon Pereira
Join Date: Mar 2009
Location: Ann Arbor, MI
Posts: 2,663
Blog Entries: 1
Rep Power: 47 |
I would agree with CBR, for a simple geometry like a wing, and especially such a large mesh, Hexa would be much better than tetra prism and not very much work (might even be faster start to finish). On a complex model, Tetra/prism might have the speed advantage and might even outweigh the quality advantage of hexas.
But if you are stuck with Tetra/Prism, I would recommend the first approach. Actually, that TGrid hookup to 11.0 was not done very well (doesn’t keep prisms, doesn’t respect density boxes, etc.). For 12.0, we added a TGLib option to the Delaunay approach, which is much better. At 12.1, we have an option for AFT also (but that doesn’t help you until November ;^). Delaunay Tetra fill (TGrid is a brand of Delaunay) is better than Octree for CFD because it more efficiently fills the volume and because its cell to cell volume transitions are better. I don’t think it affects prism, but prism can affect the Delaunay. As prism grows in, the inside triangles may be of lower quality than the outside surface was. They can be stretched or squished in one direction by convex or concave situations. So, if you planned to fill with Delaunay tetra anyway, it is better to do it from the Laplace smoothed outer envelope than after prism reduces the quality. |
|
August 28, 2009, 03:37 |
|
#4 |
Senior Member
Stuart
Join Date: Jul 2009
Location: Portsmouth, England
Posts: 740
Rep Power: 26 |
PSYMN, thanks again.
Unfortunately, I only have the tetra/prism licences - not hexa licence so I cannot make any these or multi-block meshes. In addition, maybe you can help with this add-on question. I have 2 dual-core processors in my PC so I have set Settings/General/Number of Processors to 4. When I use the Octree volume mesh I can see from Windows Task Manager that all are being used. But when I then use T-Grid only 1 processor is being used. How can I make ICEM use all the processors with T-Grid? Cheers. |
|
August 28, 2009, 09:23 |
not yet
|
#5 |
Senior Member
Simon Pereira
Join Date: Mar 2009
Location: Ann Arbor, MI
Posts: 2,663
Blog Entries: 1
Rep Power: 47 |
I understand, you can only use what you have...
As for parallel, the refinement portion of ICEM CFD Octree Tetra is parallel, but TGrid is not parallel. They are working on it, maybe next year. Simon |
|
Thread Tools | Search this Thread |
Display Modes | |
|
|
Similar Threads | ||||
Thread | Thread Starter | Forum | Replies | Last Post |
Moving mesh | Niklas Wikstrom (Wikstrom) | OpenFOAM Running, Solving & CFD | 122 | June 15, 2014 07:20 |
Ratio between bl mesh and volume mesh | nazario | ANSYS Meshing & Geometry | 0 | July 7, 2009 07:33 |
negative volume found in es-ice mesh | lizhihua | Siemens | 1 | August 4, 2007 05:39 |
different mesh types in one volume | mp | FLUENT | 2 | December 20, 2002 19:11 |
unstructured vs. structured grids | Frank Muldoon | Main CFD Forum | 1 | January 5, 1999 11:09 |