-
Dear Community, In Moose I have problems reading the MSH file attached. Because Moose uses libmesh and it looks that the effect is somehow related to libmesh I'd like to ask here for help. The corresponding discussion with Moose can be found here: idaholab/moose#30220 I'm using a
The MSH file is a reduced sample from a larger file. This is the smallest file I was able to create showing my problems. Unfortunately, its still very large. If I import that file, I get 2 (actually 3) problems:
There is one more problem not shown in the file provided below: If the physical name IDs of volumes do not start with 1 but there are lower-dimensional sidesets starting with id=1, the volumes will be ignored (up to my understanding, MSH does not require for each dimension-category to start at 1). Gmsh is reading the MSH file without problems and shows it correctly. How can I resolve the problems described above? It is not clear to me, if my MSH file is faulty or there are some issues in libmesh. The GEO file used
MSH file smpl2.msh
|
Beta Was this translation helpful? Give feedback.
Replies: 1 comment 9 replies
-
I can replicate this; looking at it now.
And I can replicate this. Looks like there's 3 nodes getting a -2 id. |
Beta Was this translation helpful? Give feedback.
What an odd encoding. I would have stored "foo bar 2" as "foo bar 2" rather than "bar foo -2", myself.
Their alternative is at least well-defined for line elements, where the 2-element symmetry group is isomorphic to <{1,-1},×>, but is there an interpretation for other element types as well? The symmetry group of a triangle has 6 elements, 3 of which are of the opposite orientation to the identity and the other two. If we see a negative there, which of those three flips do we use? Keep the first node as the first node, I guess? The docs at https://gmsh.info/doc/texinfo/gmsh.html say that negative tags have this "reverse orientation" interpretation for surfaces as well as curves, but don't…