[Gmsh] step import error

[Mark Smith]

Hi Christophe, Gmshers,I found a work around for my simple geometry with shared internal faces which might work for othersIt concerns cases where STEP (or other OCC probably) geometry won't import correctly into gmsh with the cut and merge faces option enabled i.e. bits disappear or still have duplicate elementary entities.

Work arounds so far.
1) Import STEP file into GiD 11.1.9d (free Evaluation version but may not work with large models?) then do geometry>edit>collapse>model then utilities>repair model. Then export as iges.
Use this iges file in merge"file.igs"; in a .geo file (disable the cut and merge faces option) if you want you could then save it as a  .brep file and in a new .geo file use this .brep file in the merge"file.brep" line. Either way, you should now have geometry without any duplicate geometric entity problems.
2) load the .step file exported from CAD into gmsh using File>Open (not via a .geo file with the merge line)
save it as a .geo file in the GEO options make sure NONE of the check boxes are ticked. I know this is notrecommended but for my geometry it seems to work and lets me proceed.
Load this .geo file and click coherence to remove the duplicate geometric objects then reload.

No guarantee that either will work for all possible geometries ;-( 
HTH.RegardsMark
       From: Christophe Geuzaine <cgeuzaine at ulg.ac.be>
 To: Mark Smith <piedycat at yahoo.com> 
 Sent: Wednesday, February 18, 2015 4:38 PM
 Subject: Re: step import error
   

> On 17 Feb 2015, at 11:46, Mark Smith <piedycat at yahoo.com> wrote:
> 
> Hi Christophe,
> I agree that the file imports OK but the inner surfaces (those that separate the volumes) have duplicate
> surfaces overlying one another so when you 2D mesh you get over lapping meshes?
> The defined physical groups>surfaces numbers do not change position now on reload which is good.
> Using the 3D mesh with Elmer for a CFD simulation produces incorrect unreasonable results
> as would be expected, if I get rid of the duplicate surfaces using GiD which
> for this model can be used in the evaluation mode, exported as iges, read into GMSH and exported as
> brep then all the problems go away.
> The previous example I sent, the cut and merge faces option sorted the duplicate faces/lines/points
> out but not so with the split_pipe.step I sent you?
>  

Yes, it screwed up the topology somehow... I guess for now the only robust solution is to eliminate the duplicates using a "real" CAD kernel, and only import the correct STEP into Gmsh.

Christophe

> Regards
> Mark
> From: Christophe Geuzaine <cgeuzaine at ulg.ac.be>
> To: Mark Smith <piedycat at yahoo.com> 
> Sent: Monday, February 16, 2015 6:56 PM
> Subject: Re: step import error
> 
> 
> > On 16 Feb 2015, at 13:44, Mark Smith <piedycat at yahoo.com> wrote:
> > 
> > Hi Christophe,
> > gmsh still has step file import problems as I discovered with the simple geometry attached.
> > with cut and merge faces checked, I get the following warning & 1 face and volume is missing:
> > OCC STL triangulation of surface 11 failed.
> > 
> > Any ideas?
> 
> Hi Mark - No... "Cut & Merge" is really experimental; the file imports fine without it...
> 
> 
> 
> 
> > Thanks in advance.
> > Regards
> > Mark
> 
> > <split_pipe.STEP>
> 
> -- 
> Prof. Christophe Geuzaine
> University of Liege, Electrical Engineering and Computer Science 
> http://www.montefiore.ulg.ac.be/~geuzaine


> 
> 
> 
> 
> 

-- 
Prof. Christophe Geuzaine
University of Liege, Electrical Engineering and Computer Science 
http://www.montefiore.ulg.ac.be/~geuzaine




  
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.geuz.org/pipermail/gmsh/attachments/20150219/875f6610/attachment.html>