Re: [Getfem-users] Is there any good way to contact between different dimension model?

[Konstantinos Poulios]

Dear Tetsuo,

If you edit this page you can also remove the reference to "boost intrusive_ptr". We use standard shared_ptr now. You could also add some mention to projected_fem which is the same as interpolated_fem but performing a projection on a surface instead of an interpolation within a volume.

BR

Kostas

On Tue, Feb 5, 2019 at 11:28 AM Tetsuo Koyama <address@hidden> wrote:

Dear Yves

I think it is good to add this information to

mixed methods with different meshes in
http://getfem.org/userdoc/interNMM.html

If I could have agreement, I'll add it myself.
Could you give a comment?

Best Regards,

Tetsuo

2019年2月5日(火) 19:09 Tetsuo Koyama <address@hidden>:
>
> Dear Yves
>
> Thank very much you for your answer.
>
> Best Regards,
>
> Tetsuo
>
> 2019年2月5日(火) 19:05 Yves Renard <address@hidden>:
> >
> >
> > Dear Tetsuo,
> >
> > The standard way in GetFEM is to define a fem in the whole domain and use
> >
> > md.add_filtered_fem_variable(name, mf, region)
> >
> > where region is the boundary. This will select the degrees of freedom whose shape function is non-zero on the given region.
> >
> > You can either select the degrees of freedom yourself and use a "partial_mesh_fem".
> >
> > Best Regards,
> >
> > Yves
> >
> >
> > Le 05/02/2019 à 10:52, Tetsuo Koyama a écrit :
> >
> > Dear Yves
> >
> > I'm searching to how to make MeshFem on left and right and top mesh
> > border region in user document, but I couldn't find it.
> > Where can I get information about that?
> >
> > Thank you.
> >
> > Best Regards,
> >
> > Tetsuo
> >
> > 2019年2月5日(火) 18:02 Tetsuo Koyama <address@hidden>:
> >
> > Dear Yves
> >
> > I could understand that model object have no restriction for mesh dimension.
> > I'll tried to use it. I will ask you a question if there is an uncertain point.
> >
> > Thank you for your support and development. GetFEM++ is very useful.
> >
> > Best Regards,
> >
> > Tetsuo
> >
> > 2019年2月5日(火) 17:20 Yves Renard <address@hidden>:
> >
> > Dear Tetsuo,
> >
> > The ability to couple problems defined on meshes of different dimensions
> > is one of the strenght of GetFEM.
> > You can indeed handle meshes of different dimensions in a single model.
> > If you have some variables defined only on a boundary, you can add some
> > weak form terms on that boundary. If you have a different mesh there,
> > you can add the coupling terms using the interpolation transformations.
> > And if you want to solve some sub-parts of your model, in order for
> > instance to solve you global model in an fixed point way, this is also
> > possible to make some partial solve (you can disable some variables to
> > freeze them).
> >
> > Best Regards,
> >
> > Yves
> >
> >
> >
> > Le 05/02/2019 à 08:22, Tetsuo Koyama a écrit :
> >
> > Dear GetFEM++ project
> >
> > Hello. Now, I'm making 2D mesh model which have 1D mesh model in LEFT,
> > RIGHT and TOP side.
> > I have to solve 1D mesh model and solve 2D mesh model by using 1D mesh
> > model result.
> > I know that GetFEM++ couldn't solve different dimension model in one
> > object. So I made 4 model object and tried to move 1D mesh model
> > object result to 2D mesh model object result. But, it is a little
> > bothering to take correspondence between 2 model object.
> >
> > Is there any good way to solve the probelm?
> >
> > Best Regard Tetsuo,
> >
> > --
> >
> >    Yves Renard (address@hidden)       tel : (33) 04.72.43.87.08
> >    INSA-Lyon
> >    20, rue Albert Einstein
> >    69621 Villeurbanne Cedex, FRANCE
> >    http://math.univ-lyon1.fr/~renard
> >
> > ---------
> >
> >
> > --
> >
> >   Yves Renard (address@hidden)       tel : (33) 04.72.43.87.08
> >   INSA-Lyon
> >   20, rue Albert Einstein
> >   69621 Villeurbanne Cedex, FRANCE
> >   http://math.univ-lyon1.fr/~renard
> >
> > ---------