[CIG-LONG] Under resolved cell problem

[Louis Moresi]

Charmaine,

I think the dtFactor should take a value 0.0 < dtFactor <= 1.0 since is
governed by numerical stability criteria.
This parameter is the safety margin in the timestepping scheme which ensures
you do not take timesteps which
are large enough to cause problems. It is a ratio and does not change with
scaling the problem.

1.0 should be stable, but, in 2d and 3d, it is not guaranteed, so we usually
recommend using 0.5 or so.

L.


On 17 June 2010 11:40, Charmaine Thomas <charmthomas at gmail.com> wrote:

> Hello everyone!
>
> I am currently using Gale to run a 2D long-term extension model using very
> low velocities (~2 mm/year) over a ~40 million year time period, and with a
> resolution of 135 x 90.
> I try to choose a dtFactor such that the amount of time in each timestep is
> roughly 100k.
> Everything runs smoothly until after only the first or second timestep,
> when Gale gives the following disheartening error:
>
> "Something went horribly wrong in _PCDVC_Calculate2D: Problem has an under
> resolved cell (Cell Id = 738), check or tune your population control
> parameters"
>
> No such error comes about when I leave the dtFactor as 1, so it would seem
> the problem is with increasing the dtFactor. Also, if I just use dt the same
> error comes about.
> Increasing the resolution and number of particles seems to help, but I can
> only increase so much before I run out of memory (I am not using compiled
> version of Gale), so is this the only solution?
>
> Has anyone come across this error?
> If anyone can shed some light that'd be really great!
>
> Cheers,
>
> Charmaine Thomas
> M.Sc. Candidate
> School of Geosciences | University of Sydney
>
>
>
>
> _______________________________________________
> CIG-LONG mailing list
> CIG-LONG at geodynamics.org
> http://geodynamics.org/cgi-bin/mailman/listinfo/cig-long
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://geodynamics.org/pipermail/cig-long/attachments/20100617/2b488b75/attachment-0001.htm