Hi all,
I used SALOME to generate mesh. Because of the complex structure, I defined 15 boundaries (boundary groups); actually there will be more in the future. However I found after using ElmerGrid, there are only 9 boundaries. I used '-autoclean' flag for renumbering the entities and removal of excess entities. I removed -autoclean option and the numbers became unordered but I didn't see other more boundaries either.
The model is a 3D model. I am interested in surface boundaries. As requested, there are overlapping between the boundary groups. For example, I defined group 1, 2, 3, and then group 4 is a combination of the groups 1, 2 and 3.
I look forward to your hints about how to enable other boundaries. I went through the ElmerGrid documentation and tried but didn't find any solution yet. Thanks a lot for your time.
Best regards,
http://code-saturne.blogspot.com/
Lost some boundary groups when using ElmerGrid
-
- Site Admin
- Posts: 4832
- Joined: 22 Aug 2009, 11:57
- Antispam: Yes
- Location: Espoo, Finland
- Contact:
Re: Lost some boundary groups when using ElmerGrid
Hi
From what format did you do the conversion from? Also, how are the dual definitions seen in that format. I could think many ways to give the info that could fool the parser. I don't remember coding to any format the possibility of a boundary element to belonging to more than one boundary set. And ElmerSolver certainly does not handle that either. So the only way how this kind of information could even in theory be maintained is to have dummy copies of the same boundary elements with different boundary index.
So to conclude, I would advice againts having dual definitions, and cannot much prioritize that debugging. If you remove the dual definitions, does the import then work without problems?
-Peter
From what format did you do the conversion from? Also, how are the dual definitions seen in that format. I could think many ways to give the info that could fool the parser. I don't remember coding to any format the possibility of a boundary element to belonging to more than one boundary set. And ElmerSolver certainly does not handle that either. So the only way how this kind of information could even in theory be maintained is to have dummy copies of the same boundary elements with different boundary index.
So to conclude, I would advice againts having dual definitions, and cannot much prioritize that debugging. If you remove the dual definitions, does the import then work without problems?
-Peter
Re: Lost some boundary groups when using ElmerGrid
Thanks a lot, Peter,
I used SALOME and unv format. It is okay to have a boundary belong to more than one boundary groups. I used this in Elmer because it is much easier to extract boundary summation / integration at a series of boundaries which should be in the same characteristics, and also want to examine one or two particular boundaries in the series.
If I remove dual definitions, it works very well.
Best regards,
I used SALOME and unv format. It is okay to have a boundary belong to more than one boundary groups. I used this in Elmer because it is much easier to extract boundary summation / integration at a series of boundaries which should be in the same characteristics, and also want to examine one or two particular boundaries in the series.
If I remove dual definitions, it works very well.
Best regards,