On Thu, 11 Mar 2010, Karen Lee wrote:

> To recap, I guess I wasn't very clear (regarding what Roy mentioned) how to
> store the data on a degree 1 LAGRANGE basis in the code...

Add a second system, an ExplicitSystem, to your EquationSystems.  Add
a single variable to it.  Load data as a solution into that, and query
it when you're integrating your real system.

> I've used the original Poisson example with a constant term for fxy
> to test the program, but I ought to be feeding that from my data
> (which I have successfully read into MeshData, just not sure how to
> access the correct points...) I know that the quadrature points are
> not the node values... How do I assign the correct values and not
> have MeshData think I'm giving it a node it doens't know about?
> Also, should I be using mesh_data.operator()?
> 
> Would I be better off storing data to be used in the RHS for elements
> instead of for the nodes instead?

These are primarily questions about your formulation; figuring out how
to implement it is impossible until you've got that pinned down.  What
is your forcing function, conceptually?  A bunch of delta functions at
each node?  Then you want to integrate your forcing function by
looping over all nodes instead of all elements.  A smooth or a
discontinuous function on your domain?  Then you want to integrate
over all elements, but interpolate a continuous or a discontinuous
finite element solution at each quadrature point.
---
Roy

------------------------------------------------------------------------------
Download Intel® Parallel Studio Eval
Try the new software tools for yourself. Speed compiling, find bugs
proactively, and fine-tune applications for parallel performance.
See why Intel Parallel Studio got high marks during beta.
http://p.sf.net/sfu/intel-sw-dev
_______________________________________________
Libmesh-users mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/libmesh-users

Reply via email to