[CIG-SEISMO] CPML Boundary conditions, SPECFEM2D

[Anton Biryukov]

Good morning CIG Computational Seismology members,



I am writing to you with a question about the CPML boundary condition implementation in the current version of SPECFEM2D code, available at geodynamics.org .


Would you be so kind as to shed some light on how to implement the said conditions in the model?

>From the manual I extracted two important bits of information :


1) Additional layers need to be created around the existing area-of-interest mesh, that should inherit the layered structure (if any) and be perfectly matching. This part is clear, and depends on my skill in creating a mesh with Gmsh.


2) Accordingly, changes need to be done in the Par File, to tell SPECFEM2D to use CPML instead of Stacey's conditions.


Now, moving on to the parts that are not clear to me :


a) The regular procedure of adapting an externally created mesh (in Gmsh) is to:

- create a geometry

- mesh it

- assign specific identifiers to the physical surfaces that belong to different materials through commands like Physical Surface ("M1") = Surface{1}
- assign specific identifiers to the boundaries ( e.g.  Physical Line("Top") = {1,2,3} )
- run the python script provided by the developers, that digests the aforementioned .msh file and creates files, needed for xmeshfem2d to proceed.


That works fine with me, and I managed to run the simulations using this approach. I was not quite happy with the numerical reflections though, so decided to fiddle with CPML a bit.


b)  As I 've gone through the examples provided by developers, the CPML feature implementation becomes more and more vague to me ;

Therefore, I am wondering whether :

- I need to create additional physical surfaces and name them somehow

- xmeshfem2d will understand the CPML layers in my mesh file

-there is any script that will , analogous to the one I mentioned above for stacey's conditions, read the msh and output the necessary files, so I do not need to worry about them.

- CPML files need to be created manually, and if yes, then what the easiest way would that be.


I sincerely appreciate all the suggestions you might have for me,



Best regards,


Anton.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.geodynamics.org/pipermail/cig-seismo/attachments/20150529/a06e98d1/attachment.html>