We've tried a few techniques for integrating the partial elements. For our
problems, we're really interested in using the original integration points
to avoid the need to map any stateful data to new integration points.
We've experimented with some techniques to use the original integration
points on the partial elements, calculating new weights for those
integration points to give an optimal integral of the functions over the
physical part of the element. Those techniques give improved results, and
they compare well with a "gold standard" solution that we obtained by using
an extremely high order integration rule. Really, the best thing to do for
that "gold standard" solution is to use your code that triangulates the
split elements, and we should probably do that. The main conclusion,
though, is that the results aren't that much better than what we get just
multiplying the weights by the fraction of the volume that is physical for
the partial elements. We're working on writing up our findings.
-Ben
On Fri, Nov 21, 2014 at 12:43 PM, Kirk, Benjamin (JSC-EG311) <
benjamin.k...@nasa.gov> wrote:
> I'm curious if you've had any issues relating to integrating over
> discontinuities, as that's the first & only thing I've taken a stab at.
>
> As for the old dof indices bit, I'd be happy to participate in that
> discussion, especially if you can create the problem with a minimal example.
>
> -Ben
------------------------------------------------------------------------------
Download BIRT iHub F-Type - The Free Enterprise-Grade BIRT Server
from Actuate! Instantly Supercharge Your Business Reports and Dashboards
with Interactivity, Sharing, Native Excel Exports, App Integration & more
Get technology previously reserved for billion-dollar corporations, FREE
http://pubads.g.doubleclick.net/gampad/clk?id=157005751&iu=/4140/ostg.clktrk
_______________________________________________
Libmesh-devel mailing list
Libmesh-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/libmesh-devel
