Re: [Libmesh-users] empirical convergence study

[Derek G. <fri...@gm...>]

Just in case it isn't completely clear... what you need to do is construct an ExactSolution object using your coarse grid solution and use the "attach_reference_solution()" method to attach your fine grid solution.  Then when you call functions on it like l2_error() it will compare the two solutions in that norm.

Derek

On Jul 30, 2010, at 1:00 PM, Roy Stogner wrote:

> 
> On Fri, 30 Jul 2010, Boyce Griffith wrote:
> 
>> I'm trying to setup a simple empirical grid convergence study in
>> libMesh, and was hoping you all could point me in the right direction.
>> Specifically, if I have a solution on a coarse Cartesian grid, and a
>> solution on a fine Cartesian grid which is a factor of two finer than
>> the coarse grid, is there a mechanism in libMesh that would allow me to
>> restrict the fine grid solution onto the coarse mesh so that I may
>> compute norms of differences between the coarse and fine solutions?
> 
> You ought to integrate your norms on the fine grid (otherwise you may
> get a fraction of your approximation error but only after having
> thrown away all your interpolation error), and you can do so with the
> ExactSolution class.
> 
> That class currently only has functions for returning norms in one
> variable at a time; if you've got a multiphysics system and want to be
> able to calculate a SystemNorm based result then we'd be happy to have
> a patch.  ;-)
> ---
> Roy
> 
> ------------------------------------------------------------------------------
> The Palm PDK Hot Apps Program offers developers who use the
> Plug-In Development Kit to bring their C/C++ apps to Palm for a share
> of $1 Million in cash or HP Products. Visit us here for more details:
> http://p.sf.net/sfu/dev2dev-palm
> _______________________________________________
> Libmesh-users mailing list
> Lib...@li...
> https://lists.sourceforge.net/lists/listinfo/libmesh-users