Hey there,
I have some familiarity with the transfinite interpolation implementation
and am looking into this now. It looks like, in the transformation, one of
the 'reference cell points' is calculated as {20, 20}: I don't yet see why.
I suspect that we now create a twisted cell (or something else that is
poorly conditioned) or that there is an issue with how we compute chart
points: either way, its either something we need to catch with a better
error message or a bug we need to fix.
Over the last few months Martin has added a lot of optimizations to the TFI
code, in addition to using it in a more general setting during mesh
refinement: put another way, our implementation has changed substantially,
so this problem is at least in an area of the library where we have
recently made huge changes.
Thanks,
David Wells
On Wednesday, January 17, 2018 at 3:05:48 PM UTC-5, Juan Carlos Araujo
Cabarcas wrote:
>
> Dear all,
>
> Since the inclusion of Transfinite interpolation, I have been successful
> on working with this powerful technique in my research. I had coded a mesh
> implementing concentric circles, where the inner most is shifted a small
> distance s. All concentric circles are labeled 100+i, where i is in a loop
> on all circles.
>
> The mesh was working after 4/Apr/17 when I added the following entry in
> the forum:
> https://groups.google.com/forum/#!topic/dealii/hCZqv9g6mKk
>
> I installed the development version of dealii on the 17/nov/17. The
> details of the installation are also in the forum:
> https://groups.google.com/forum/#!topic/dealii/ee2w2X987ig
>
> and recently I went back and tried to run thhis code, obtaining the error
> at the end of the email.
>
> I prepared a minimal example that includes the definition of my grid. It
> includes how I colorize each face as explained before, and the rest are
> colorized "1" as the documentation suggests.
>
> The symptoms are the following:
> - If I use refinement=0, everything works and the mesh looks quite nice in
> zeros.vtk and surface with lines mode in Paraview.
> - For any higher refinement, it gives the error below.
>
> Maybe there has been a major change in TransfiniteInterpolation since
> Apr/2017, or maybe the way I colorize is not good anymore for some weird
> reason. Any ideas or comments would be greatly appreciated.
>
> Juan Carlos Araújo,
> Umeå Universitet.
>
>
>
> terminate called after throwing an instance of 'dealii::Mapping<2,
> 2>::ExcTransformationFailed'
> what():
> --------------------------------------------------------
> An error occurred in line <1554> of file
> </path/dealii/source/grid/manifold_lib.cc> in function
> typename dealii::Triangulation<dim, spacedim>::cell_iterator
> dealii::TransfiniteInterpolationManifold<dim,
> spacedim>::compute_chart_points(const dealii::ArrayView<const
> dealii::Point<spacedim> >&, dealii::ArrayView<dealii::Point<dim> >) const
> [with int dim = 2; int spacedim = 2; typename dealii::Triangulation<dim,
> spacedim>::cell_iterator = dealii::TriaIterator<dealii::CellAccessor<2, 2>
> >]
> The violated condition was:
> false
> Additional information:
>
> --------------------------------------------------------
>
> Aborted (core dumped)
>
>
>
--
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].
For more options, visit https://groups.google.com/d/optout.
