Fixing/removing BoundaryMesh is in my sights for when MeshView is
implemented (https://bitbucket.org/fenics-project/dolfin/issue/329).
There are problems with BoundaryMesh (particularly in parallel) that have
been causing headaches.

Garth

On 21 January 2015 at 07:55, Martin Sandve Alnæs <[email protected]> wrote:

> Hi Matthew,
> a BoundaryMesh is 'just a mesh' and FEniCS does not
> preserve cell orientation, instead the vertex list in each
> cell is sorted by global vertex number. This gives implicit
> agreement of the edge direction between cells.
>
> Martin
>
> On 20 January 2015 at 18:17, Matthew Scroggs <[email protected]
> > wrote:
>
>> Dear all,
>>
>> I have been attempting to implement FEM BEM coupling with Fenics and
>> BEM++ and have encountered the following issue:
>>
>> When importing a mesh into Fenics then exporting the Boundary mesh, the
>> normal vectors for each element are a mixture of inward pointing and
>> outward pointing. (Before importing, they were all outward pointing).
>>
>> The following code is where the issue occurs:
>>
>> import dolfin
>> from bempp.grid import grid_from_element_data
>> from bempp.file_interfaces import export
>>
>> mesh = dolfin.Mesh('sphere.xml')
>> bm = dolfin.BoundaryMesh(mesh, "exterior", False)
>> bm_coords = bm.coordinates()
>> bm_cells  = bm.cells()
>> bm_nodes  = bm.entity_map(0).array().astype(np.int64)
>> bd_grid = grid_from_element_data(bm_coords.transpose(),bm_cells.
>> transpose())
>> export(bd_grid,'grid.msh')
>>
>>
>> I have attached the imported and exported meshes.
>>
>> Matthew
>> --
>> Matthew Scroggs
>> PhD Student
>> Department of Mathematics
>> University College London
>>
>> _______________________________________________
>> fenics mailing list
>> [email protected]
>> http://fenicsproject.org/mailman/listinfo/fenics
>>
>>
>
> _______________________________________________
> fenics mailing list
> [email protected]
> http://fenicsproject.org/mailman/listinfo/fenics
>
>
_______________________________________________
fenics mailing list
[email protected]
http://fenicsproject.org/mailman/listinfo/fenics

Reply via email to