[aspect-devel] Questions about surface process

[Yimin Jin]

Dear developers of ASPECT:

I am a student of University of Chinese Academy of Sciences and I have been
using ASPECT to do simulations on crack propagation of small-scaled
geological bodies in the passing year. I have found that the existing
velocity pojection strategies in the free-surface module introduces too
much error to the result (the 'normal' strategy leads to severe distortion
of cells, and the 'vertical' strategy ignores horizontal displacements),
while the sticky-air method will double the computational complexity of the
problem (my model can 'grow' tall so the air layer must be thick enough).
Therefore I intended to apply a diffusive surface process on the
free-surface, but when I looked into this problem I found it quite
difficult because the partition changes each time we execute refinement and
coarsening. The finest way seems to be transporting faces between processes
using functions provided by p4est (I don't know if there are any), then it
will be unavoidable to read the codes of p4est. Could you please tell me
whether you have started to do such things or not?  If it has been
implemented and will soon be released, then I will be happily looking
forward to the next version of ASPECT; otherwise, I would like to make my
effort in this part of work.

Cincerely,
Yimin Jin

Key Laboratory of Computational Geodynamics, UCAS
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.geodynamics.org/pipermail/aspect-devel/attachments/20161209/d2459070/attachment.html>