[Gmsh] About Scaling

David Colignon David.Colignon at ulg.ac.be
Fri Mar 2 17:47:19 CET 2007


You are right... There is an already known bug with the 3D Delaunay algorithm on the _Windows_ platform. I tested your file mainly under Linux and it worked very fine with scale factors ranging from 0.001 to 1000, and I forgot the Windows problem...

In the meantime, if switching to Linux is not a solution for you ( ... ) , you can combine the use of your "facteur" variable ( with value 0.1 for example in order to get a reasonable number of elements ) with the system variable  Mesh.Scaling :

Mesh.ScalingFactor = 10 ; // Global scaling factor applied to the saved mesh

and use the Netgen 3D algo  in order to obtain a final 3D mesh with the dimension you want...

Cheers,

Dave



Cécile Giorla wrote:
> Hi David,
> 
> Houston, we have a problem! ;-) The 2.0.4 Windows version exits in the 
> middle of the meshing process with my previous example. I tried the 
> Delaunay algorithm with the previous version of GMSH, and it exits the 
> same way (on the same .geo file).
> Cheers,
> Cecile Giorla.
> 
> David Colignon a écrit :
>> Hi Cécile,
>>
>> gmsh-2.0.0 uses by default the netgen 3d meshing algorithm, which as 
>> you found it has a bug when you scale up the geometry and/or work with 
>> big objects...
>>
>> first solution: keep gmsh-2.0.0 but use the 3D Delaunay algorithm: 
>> Tools -> Options -> Mesh -> General -> 3D Algorithm -> Delaunay
>>
>> second (and best) solution: be one of the first to download ( 
>> http://www.geuz.org/gmsh/bin/Windows/gmsh-2.0.4-Windows.zip )and try 
>> the new 2.0.4 version (which uses by default the 3D Delaunay meshing 
>> algorithm !
>>
>> Cheers,
>>
>> Dave
>>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: Thisiscontrol.wav
Type: audio/wav
Size: 33026 bytes
Desc: not available
URL: <http://www.geuz.org/pipermail/gmsh/attachments/20070302/2d49075d/attachment.wav>