[aspect-devel] Van Keken et al. 2008 subduction benchmark

[Max Rudolph]

Scott and Wolfgang,
I did not try the modification to allow a linear ramp. The problem that I
encountered was not associated with the pressure field in the wedge. It was
actually along the subduction interface where the downgoing plate meets the
over-riding plate. There is a discontinuity in the velocity field there.
The velocity field is locally divergence-free but it may not be
divergence-free in the integral sense when interpolated using the finite
element basis functions. One possible work-around would be to prescribe
zero dynamic pressure throughout the rigid downgoing and over-riding plate
(the correct solution). Doing so would mean prescribing all of the degrees
of freedom for every node in the prescribed velocity region. I think that
this would be a very straightforward modification of the prescribed
velocity plugin. Maybe we can chat for a few minutes about this at the
ASPECT dinner?

Max

On Fri, Dec 9, 2016 at 1:22 AM, Scott King <sdk at vt.edu> wrote:

> Max;
>
> Do you use a linear ramp something from the point where the slab meets the
> over-riding plate down into the wedge?   This *smooths* the discontinuity
> (at least as far as the wedge sees it).  With this the pressure fields in
> the wedge seems to become more smooth for many codes.   I always had to you
> a much larger "ramp" (25,-50 km) because my elements were not refined as
> much as Herr. Dr. Prof. van Keken.   This puts the discontinuity into the
> part of the domain where the velocities are entirely prescribed.  It seems
> to be ok for some codes because they don't "solve" any equation when all
> the velocities are prescribed.  Years of arguing and twiddling.
>
> Scott
>
> On Dec 8, 2016, at 10:25 PM, Wolfgang Bangerth <bangerth at tamu.edu> wrote:
>
> > On 10/27/2016 10:13 AM, Max Rudolph wrote:
> >> I was wondering if anyone has been successful in setting up the Van
> Keken et
> >> al. 2008 subduction benchmark using ASPECT? I have been using this as a
> >> starting point to set up a more complicated corner flow model and have
> not yet
> >> been successful.
> >>
> >> This benchmark solution requires a rigid overriding plate and kinematic
> >> subducting plate. There is a discontinuity along the subduction
> interface. I
> >> tried imposing kinematic boundary conditions on the left and bottom of
> the box
> >> and using Jonathan Perry-Houts' prescribed velocity plugin
> >> (cookbooks/prescribed_velocity), but the Stokes solver fails to
> converge.
> >> Looking at the pressure field, it's not surprising to see very large
> >> oscillations in this region where the velocity field is discontinuous.
> It
> >> occurred to me that using the locally conservative discretization might
> help.
> >> However, when the locally conservative discretization is enabled, the
> internal
> >> prescribed velocities are no longer enforced anywhere. Is it obvious
> why this
> >> might be the case?
> >
> > Max -- I haven't tried implementing the benchmark in ASPECT, but recall
> trying to do so in an early version of step-32 when I visited Scott in
> Blacksburg for the very first time many many years ago. I remember it to be
> very difficult because the benchmark definition really doesn't fit the
> finite element method.
> >
> > One of the issues with the use of the prescribed velocity is that the
> prescribed velocity needs to be divergence free on each cell. Is it, in
> your case? Even on cells where you only prescribe values at some nodes?
> >
> > Best
> > W.
> >
> > --
> > ------------------------------------------------------------------------
> > Wolfgang Bangerth          email:                 bangerth at colostate.edu
> >                           www: http://www.math.colostate.edu/~bangerth/
> >
> > _______________________________________________
> > Aspect-devel mailing list
> > Aspect-devel at geodynamics.org
> > http://lists.geodynamics.org/cgi-bin/mailman/listinfo/aspect-devel
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.geodynamics.org/pipermail/aspect-devel/attachments/20161209/294bcf0f/attachment.html>