Hi Mark,

in the meantime, I have discovered two (at least partial) solutions on the 
Elmer side:

1) At mesh import, Elmer can merge nodes which are near to each other. I order 
to use this, both surfaces have to be meshed with the same characteristic 
length. The threshold distance for node merging is a parameter you can set. 
Drawback: If there are several interfaces with different mesh widths, this 
method might fail.
2) By "misusung" periodic boundary conditions, Elmer can be told that certain 
variables are continuous at the interface even if the meshes do not match. I 
did not use this yet, and it seems that care has to be taken in some cases, 
e.g. with direct solvers.

Other FEM programs might have similar possibilities.

But still, a proper cleanup of double interfaces on gmsh side would be the best 
solution since it tackles the cause of the problem directly.

So if there is such a solution someday, I would still be most interested (and I 
might not be the only one).

Best regards,

Matthias

----------------------------------------------------------------
ERBE Elektromedizin GmbH
Dr.rer.nat. Matthias Zenker
Dipl. Phys.
Grundlagenentwicklung / Fundamental Development
Waldhoernlestrasse 17
D-72072 Tuebingen
Phone + 49 (0) 7071 - 755 226
Fax + 49 (0) 7071 - 755 5226
E-Mail: <mailto:[email protected]>
URL: http://www.ERBE-med.de
----------------------------------------------------------------

-----Ursprüngliche Nachricht-----
Von: Mark van Doesburg [mailto:[email protected]] 
Gesendet: Montag, 13. September 2010 09:58
An: Zenker, Dr. Matthias; [email protected]
Betreff: Re: [Gmsh] Double surfaces after STEP import

Hello Matthias,

I worked on that code and consider it a complete faillure. In many cases the 
OpenCASCADE algorithms would not find the intersections. The initial version 
was quite simple but at some point I started to stack one workaround on top of 
the other.

Currently I do not have the time to work on another method, but some algorithms 
available in CGAL look promising. But in that case I would probably use BRL-CAD 
to generate the geometry, and importing STEP into BRL-CAD might be sub-optimal.

regards,

Mark

"Zenker, Dr. Matthias" <[email protected]> wrote:

        Hi,

        I have not done FEM calculations for some time (a year or so). The last
        time I did them (using gmsh and Elmer), I had the problem that in
        geometries imported as STEP files (generated by Solidworks), adjacent
        bodies have their own external surface, resulting in double surfaces at
        the interfaces. When I use the mesh generated by gmsh in a FEM
        calculation (I use Elmer), I get no thermal and electrical connection
        because the body meshes are not connected at the interface. There have
        been attempts to clean up double surfaces in gmsh, but I don't know if
        they have come to a success. Could someone of the developpers update me
        on this point?

        Thank you!

        Matthias 

        ----------------------------------------------------------------
        ERBE Elektromedizin GmbH
        Dr.rer.nat. Matthias Zenker
        Dipl. Phys.
        Grundlagenentwicklung / Fundamental Development
        Waldhoernlestrasse 17
        D-72072 Tuebingen
        Phone + 49 (0) 7071 - 755 226
        Fax + 49 (0) 7071 - 755 5226
        E-Mail: <mailto:[email protected]>
        URL: http://www.ERBE-med.de
        ----------------------------------------------------------------


        _____________________________________________________________________
        ERBE Elektromedizin GmbH
        Firmensitz: 72072 Tuebingen
        Geschaeftsfuehrer: Christian O. Erbe, Reiner Thede
        Registergericht: Stuttgart HRB 380137


_____________________________________________________________________
ERBE Elektromedizin GmbH
Firmensitz: 72072 Tuebingen
Geschaeftsfuehrer: Christian O. Erbe, Reiner Thede
Registergericht: Stuttgart HRB 380137


_______________________________________________
gmsh mailing list
[email protected]
http://www.geuz.org/mailman/listinfo/gmsh

Reply via email to