[Gmsh] Sharp corners and physicals in .stl

[F. A. Portela]

Hey Martin,

Thanks for the thorough explanation! If I manage to get a good mesh with
Gmsh, I won't really need the .stl anyway as I would use them to use
snappyHexMesh for the meshing.

I also get the warnings though I don't know what they mean... I attached
the .geo I used for the structured mesh, I managed to improve a bit the
quality of the cells near the corner by creating a region with co-planar
faces, though I believe I just moved/changed the place where I will run
into issues... I also tried having unstructured elements near the
problematic region but it didn't really help...

Cheers,
Felipe


--
Felipe Alves Portela
MSc student in Aerospace Eng. at TU Delft
http://www.linkedin.com/in/felipealvesportela


On 21 March 2013 16:41, Martin Vymazal <martin.vymazal at vki.ac.be> wrote:

> **
>
> Hi Felipe,
>
>
>
> I didn't manage to mesh your geometry. I'm getting this (after 2 hours of
> running gmsh):
>
>
>
> Volume do not match 1.588845063218419E-04 1.600196016669183E-04
> 7.093476881907630E-03
>
> Volume do not match 1.588932327064256E-04 1.600300053984631E-04
> 7.103497179838322E-03
>
> Volume do not match 1.589864934028117E-04 1.601523686854313E-04
> 7.279787943128194E-03
>
> Volume do not match 1.589450067262323E-04 1.601000560689481E-04
> 7.214546772040652E-03
>
> Volume do not match 1.590410579932190E-04 1.601855161516023E-04
> 7.144579521784781E-03
>
> Volume do not match 1.590582035727257E-04 1.602001491852048E-04
> 7.128243127657754E-03
>
> Volume do not match 1.588969703174746E-04 1.600347350034682E-04
> 7.109485862359648E-03
>
> Volume do not match 1.590485099915781E-04 1.601891519519827E-04
> 7.120594288097972E-03
>
> Volume do not match 1.590046382981625E-04 1.601481475152607E-04
> 7.140321226564688E-03
>
> Volume do not match 1.589345926016806E-04 1.601045520535501E-04
> 7.307471504484324E-03
>
> Volume do not match 1.589321060424286E-04 1.600865990877137E-04
> 7.211678253297149E-03
>
> Volume do not match 1.590316228003549E-04 1.601823800873328E-04
> 7.184044127390844E-03
>
>
>
> Maybe you don't have the same error, though. I have svn revision 14842.
>
>
>
> About stl and physicals: if you need to save only a part of your geometry,
> make only that part physical. If you want to save all surfaces, but in
> different files, repeating the above with setting different surfaces as
> 'physical' will probably not work because the mesh algorithm is not
> deterministic - if you run gmsh twice, the output mesh is not necessarily
> exactly identical. So if you mark one surface in your geo as physical, save
> the stl, then mark another surface as physical and re-run gmsh to save
> another stl, you might end up with nonconforming output surfaces.
>
> Maybe you could write the whole mesh as *.msh, and then see if in the gui
> you cannot remove all physical groups but one. By that I mean that you
> would just remove the physical tag, not the surface itself. Then you could
> export to stl. After that, you reopen your msh and repeat with another
> surface. Not very automatic but I don't have a better idea for the moment.
>
>
>
> Best regards,
>
>
>
> Martin
>
>
>
>
>
>
>
> _______________________________________________
> gmsh mailing list
> gmsh at geuz.org
> http://www.geuz.org/mailman/listinfo/gmsh
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.geuz.org/pipermail/gmsh/attachments/20130321/5e6e251e/attachment.html>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: micro.geo
Type: application/octet-stream
Size: 14003 bytes
Desc: not available
URL: <http://www.geuz.org/pipermail/gmsh/attachments/20130321/5e6e251e/attachment.geo>