Dear Yves,
thank you very much for your comprehensive answer. That solved my
problem. I forgot, that the gradient of the level-set function provides
the normal-vector, but it totally makes sense to me now.
Thank you,
Patrick
On 08/29/2017 10:51 AM, Yves Renard wrote:
Dear Patrick,
The "Normal" vector is indeed only defined on element faces. In order
to obtain the normal vector of a level-set, the standard way is to use
"Normalized(Grad_lambda)" where "lambda" is the scalar field of the
level-set function (that you have to add as a data to the model). Note
also, that you have access to discontinuities across the level-set of
Xfem fields (mesh_fem_level_set objects) thanks to the Xfem_plus and
Xfem_minus command of the language (see
http://getfem.org/userdoc/gasm_high.html#xfem-discontinuity-evaluation-with-mesh-fem-level-set)
Best regards,
Yves.
Le 28/08/2017 à 20:44, Patrick Schmidt a écrit :
Dear Getfem users,
I am new to getFem and made my way through the demo_crack.py file in
order to use the implemented XFEM formulation.
Now I would like to set up a boundary value problem where a pressure
acts orthogonal to the surface of the discontinuity. I tried to find
some way in order to apply source terms orthogonal on the surface of
the discontinuity. In terms of the boundary condition that's as far
as I got:
applied_force = 1.e7
md.add_initialized_data('Fx', [applied_force, 0]);
mim_bound1 = gf.MeshIm('levelset', mls, 'boundary(a)',
gf.Integ('IM_STRUCTURED_COMPOSITE(IM_TRIANGLE(6),3)'))
md.add_source_term_generic_assembly_brick\
(mim_bound1, 'Fx.Test_u')
I tried to use the 'Normal' keyword within the string, but since the
integration takes part within the element and the normal seems to be
defined on surfaces only, my python script failed.
Is there a way to apply Neumann Boundary conditions orthogonal to the
surface of the discontinuity?
Kind regards,
Patrick
--
Yves Renard ([email protected]) tel : (33) 04.72.43.87.08
Pole de Mathematiques, INSA-Lyon fax : (33) 04.72.43.85.29
20, rue Albert Einstein
69621 Villeurbanne Cedex, FRANCE
http://math.univ-lyon1.fr/~renard
---------
