Hi Ulrich,

since I was the one intoducing anisotropic refinement in deal.II I thought I 
might as well look at the problem. Wolfgang forwarded your email to me for that 
purpose.

I checked something into the svn right now and hope it solves your problem, 
please let me know. If it does: It is one of the basic problems with the 
anisotropic implementation in deal.II: The level information has no real 
meaning anymore except to the internals of the Triangulation class and its 
accessors. It cannot be used to decide upon finer or coarser neighboring cells 
at a face. Obviously, there always is a way around this problem using some 
other function, otherwise I would not have been able to use the anisotropic 
stuff successfully, but except from me nobody is really constantly aware of 
this issue.

By the way, it is very interesting to know that someone other than myself is 
using these extensions. What is your application?

Best,
Tobias



-----Ursprüngliche Nachricht-----
Von: [email protected] im Auftrag von Wolfgang Bangerth
Gesendet: Di 12.01.2010 21:18
An: Ulrich Matthes
Cc: deal.II
Betreff: Re: [deal.II] deal.II, problems with refinement,deduging mode and 
Raviart-Thomas elements##SPAM ASA_schwacher Verdacht
 

Ulrich,

> The error occurs in the debuging mode in the case of non-global (but
> without hanging nodes) refinement with cut_y using
> cell->set_refine_flag( RefinementCase<2>::cut_y );
> and
> m_triangulation.execute_coarsening_and_refinement ();
> in combination with the use of Raviart-Thomas elements.

I'm afraid the Raviart-Thomas element has never been converted to support 
anisotropic refinement :-(


> The error does not occur without the debuging mode. In this case the
> program runs and produces at least plausible results.

The check isn't run in non-debug mode, so the program would not abort. I would 
at least doubt the correctness of results in that case, though it is entirely 
possible that the check is actually in error -- it was not written with 
anisotropic refinement in mind...

I guess we'd need a volunteer to try to understand what would have to be done 
to make this work.


> The error does not occur in the case of the deal.II tutorial Step-3 even
> for non-global refinement with cut_y. A probable reason is that this
> step does not use Raviart-Thomas elements.

Yes.

Best
 W.

-------------------------------------------------------------------------
Wolfgang Bangerth                email:            [email protected]
                                 www: http://www.math.tamu.edu/~bangerth/

_______________________________________________
dealii mailing list http://poisson.dealii.org/mailman/listinfo/dealii

_______________________________________________
dealii mailing list http://poisson.dealii.org/mailman/listinfo/dealii

Reply via email to