[Getdp] unexpected "segmentation violation" at Pre-processing

Patrick Dular Patrick.Dular at ulg.ac.be
Tue Feb 9 16:08:47 CET 2010


Dear Guillaume,

The 'Region[All]' is not supported as a support of a BasisFunction. 
Sorry for that...

You should define it explicitly, with

   Support Region[Omega];

instead of

   Support Region[All];

Thank you for reporting this bug. It will be corrected.

Also, the constraint of type 'LinkCplx' asks for an exact correspondence 
of nodes between the surfaces to be linked. You can specify this in your 
.geo file with:

Transfinite Line {1,4} = (14e-006-10.8e-006)/lc_air + 1;
Transfinite Line {2,8} = (10.8e-006-3.1e-006)/lc_air + 1;
Transfinite Line {3,12} = (3.1e-006-1e-007)/lc_air + 1;

... and do not forget to correct the Function that applies Region 
SurfBlochRight on SurfBlochLeft, i.e.

   Function Vector[$X-period,$Y,$Z] ;

instead of

   Function Vector[$X+period,$Y,$Z] ;


Regards,
Patrick



On 09/02/2010 15:35, gdemesy at physics.utoronto.ca wrote:
> Dear Getdp List,
>
> I am trying to calculate the field diffracted by a grating enlighten 
> by a plane wave in the visible range (for 1 polarization case = 2D pb, 
> in harmonic regime). In short, it amounts to solve a scalar Helmholtz 
> equation together with Bloch conditions, PMLs and an appropriate 
> source term located inside the diffractive element (just like we did 
> in www.opticsinfobase.org/oe/abstract.cfm?uri=oe-15-26-18089).
>
> I have been struggling for a while on the rather simple attached 
> problem. This gmsh/getdp code is a "benchmark problem" for which I 
> already have the solution via another FEM code. I am stuck with this 
> "Segmentation violation (invalid memory reference)" at Pre-processing 
> step.
>
> I have unsuccessfully tried to simplify each step of the model as much 
> as possible:
> - a geometry generated whithout using "Extrude"
> - a rough mesh
> - Neumann instead of Bloch conditions
> - source = 0
> ...
> My conclusion: I'm missing something elementary somewhere.
> Any hint would be greatly appreciated...
>
> Thanks a lot for your time.
>
> Best,
> Guillaume Demésy,
> Postdoc at University of Toronto,
> Sajeev John's Group
>
> ----------------------------------------------------------------
> This message was sent using IMP, the Internet Messaging Program.
>
>
> _______________________________________________
> getdp mailing list
> getdp at geuz.org
> http://www.geuz.org/mailman/listinfo/getdp
>    

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.geuz.org/pipermail/getdp/attachments/20100209/c9289d23/attachment.html>