dolfin team mailing list archive

["jhameed2@xxxxxxxxxxxx" <question108100@xxxxxxxxxxxxxxxxxxxxx>]

Question #108100 on DOLFIN changed:
https://answers.launchpad.net/dolfin/+question/108100

Description changed to:
I have two equations to solve. One of them is solved on a subdomain, and
the other on the full domain. Once each problem is solved, I need to use
its results in the second problem. So, if the problem on subdomain (say
problem Small) is solved, I need to use this as a variable in the
equation on the bigger domain (say problem Big). Once Big is solved,  I
need to restrict it to the smaller domain, and so on.

I can see two ways of going about it:

1) Store two different meshes for the two domains. Once I have the
values for one of the problems, use interpolation to get the values for
the other problem. This doesnt seem very elegant, as one of the meshes
is simply a sub-mesh of the other.

2) Use one mesh to start off with, but then use sub-mesh to solve
problem Small. Here I had a couple of questions. I want to store the
boundary markers for the subdomain in the mesh file. So, these boundary
markers will not be boundary of the mesh, but should be markers for the
boundary of the submesh. Can I do this? Once I extract a submesh, will
the boundary indicators for the submesh also be extracted? Boundary
markers are needed by me to mark the boundary as Neumann, or Dirichlet
etc.

Which of the two methods is better, or is there a better third option?
Is there any demo which sort of does this in the examples?

Thanks,
-Jehanzeb

-- 
You received this question notification because you are a member of
DOLFIN Team, which is an answer contact for DOLFIN.