Simon,

Yes, you should be able to skip the constrained hanging nodes. Like you 
said the value will be overwritten when you call distribute(). Let us know 
if you get something unexpected but that should work.

Best,

Bruno

On Wednesday, June 23, 2021 at 5:00:06 AM UTC-4 Simon wrote:

> Dear all,
>
> I am dealing with adaptivity, so hanging nodes are present in my mesh 
> which are handled by a constraints object like in the following:
> AffineConstraints...constraints;
> make_hanging_node_constraints...();
> constraints.close();
>
> My solution vector is filled via a local approach. Basically I loop over 
> all active vertices, use cell->vertex_dof_index() to access the dofs living 
> on them and write appropriate values in the global solution vector. So I 
> don't use the before mentioned constraints object to make a local-global 
> distribution or something similar as it is done in most tuturial programs. 
> After my loop is finished I call constraints.distribute().
>
> So by looping over all active vertices I also visit the constrained 
> hanging nodes and write some values for these dofs in the solution vector.
>
> My question is if I can just skip the hanging nodes in my loop because the 
> call "constraints.distribute()" will overwrite these values anyway?
> This is at least my understanding of what constraints.distribute() does, 
> i.e. computing constraind dof values from unconstrained ones. But I am not 
> quite sure if problems can come up by modifyiing the constrained dofs 
> beforehand. 
>
> Best
> Simon
>

-- 
The deal.II project is located at http://www.dealii.org/
For mailing list/forum options, see 
https://groups.google.com/d/forum/dealii?hl=en
--- 
You received this message because you are subscribed to the Google Groups 
"deal.II User Group" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to [email protected].
To view this discussion on the web visit 
https://groups.google.com/d/msgid/dealii/6288fd1c-fa80-4798-ac7b-49a904f1ab40n%40googlegroups.com.

Reply via email to