On Thu, 9 Feb 2012, Derek Gaston wrote:

On Feb 8, 2012, at 6:14 PM, Roy Stogner wrote:

Only with some difficulty; it'd be the first time we put variables on
elements that didn't technically exist.

It'd be easier (though still quite difficult) to do something we've
been wanting to for a while: extend libMesh to handle Mesh objects
with Elems of more than one dimension inside.  Then you just create
the boundary mesh, add those elements into the Mesh with their own
subdomain, set a per-subdomain variable on for them, and off you go.

Are you sure this does't work right now?  We are using lower
dimension meshes in higher dimension spaces right now…

Lower dimension meshes in higher dimension spaces should definitely
work; we've also been using them for a while and I only recall finding
(and fixing) one or two bugs before we started getting good results.

and I can't see why having some elements that are 2D and some that
are 1D in the same mesh wouldn't work.

There are lots of places where we build an FEBase object of
mesh_dimension() and then assume that we can use it on every element,
for starters.

You might even be able to tie them together by using the same nodes
(so that the solutions are continuous and the sparsity pattern is
correct).

This is a great idea but would almost certainly break on
find_neighbors().
---
Roy
------------------------------------------------------------------------------
Virtualization & Cloud Management Using Capacity Planning
Cloud computing makes use of virtualization - but cloud computing 
also focuses on allowing computing to be delivered as a service.
http://www.accelacomm.com/jaw/sfnl/114/51521223/
_______________________________________________
Libmesh-users mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/libmesh-users

Reply via email to