[Getdp] boundry elements

[Christophe Geuzaine]

"ghather at uclink.berkeley.edu" wrote:
> 
> I'm trying to find the field inside of a dialetric object in a uniform field.  I decided to test a sphere of radius 1 inside of a box of width 14 (the boundry is infinite, so I had to trucate it with the box).
> 
> So I'm going to try to couple the finite element method with the boundry element method because the boundry element method can handle infinite boundrys, and only the boundry needs to be meshed.
> 
> Can Getdp do boundry elements?
> 
> If so, can you give me a hint as to what a boundry element formulation would look like?
> 

Greg,

If you increase the interpolation order and the error does not decrease,
you may be right: the error may be due to to the fact that you truncate
the domain. In this case, using boundary elements is indeed one
alternative. Nevertheless, boundary elements in getdp are not completely
ready for the end user (no documentation, and the public version of
getdp does not contain all the analytic integration routines needed to
use boundary elements efficiently. And I unfortunately don't have time
to work on any of these two problems at the moment...). However, there
is an alternative to boundary elements: you can simply use the special
transformation jacobians. See getdp/demos/test.pro.

Christophe

-- 
Christophe Geuzaine

Tel: (626) 395-4552    http://www.geuz.org
Fax: (626) 578-0124    mailto:geuzaine at acm.caltech.edu