[Dune] Boundary conditions, segments and indexing

[Firmbach, Max]

Hello,


ok this makes sense. In the meantime i tried the procedure with

a tet-mesh and this works as intended.


So thanks again Carsten and Andreas.


Best regards,

Max

________________________________
Von: Carsten Gräser <graeser at mi.fu-berlin.de>
Gesendet: Montag, 24. August 2020 16:24:04
An: dune at lists.dune-project.org
Cc: Firmbach, Max
Betreff: Re: [Dune] Boundary conditions, segments and indexing

[This accidentally went into a private reply. I'm sending
it to the list, too, for completeness.]

Hi Max,

Am 24.08.20 um 14:29 schrieb Firmbach, Max:
[...]> I don't insert the boundary segments "specifically by myself", the
> gmsh reader does that. I only use the factory that is given back,
> in terms of the insertion index of the intersections
> (like i only call the read function from the gmsh parser in DUNE,
> everything there is a blackbox for me). I just specified everything
> in the .msh file, I dont know in which order DUNE inserts them.this
iswhat I was looking for. Now it's clear that the
ordering of boundary segments in the .msh is not consistent
with the generated mesh, according to the output of your
program.

Looking at the gmsh reader code it seems that it simply does not
insert boundary segments of type 3, a.k. "4-node quadrilateral".
Unfortunately there's also no error message or warning printed.
Maybe it's easy to fix this - I'll have a look. I opened a bug
report to document the issue and track the discussion:

  https://gitlab.dune-project.org/core/dune-grid/-/issues/115

Best,
Carsten




-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.dune-project.org/pipermail/dune/attachments/20200824/ec6224d6/attachment.htm>