← Back to team overview

dolfin team mailing list archive

Re: Petrov-Galerkin FEM (Upwinding)

 

On Thu, Mar 05, 2009 at 10:54:02AM -0500, Hatef Monajemi wrote:
> That works fine when you have one Testfunction like "v", How about when
> you have the following set of Testfunction:
> -->
> P2 = VectorElement("Lagrange", "triangle", 1)
> P1 = FiniteElement("Lagrange", "triangle", 1)
> TH = P2 + P1
> 
> (v,w) = TestFunctions(TH)
> -->
> analogous with what anders said, one should be able to do something
> like :
> 
> (v,w)=( v + delta*A(v), w)
> 
> However, this creates problem in ffc. An alternative is to multiply the
> whole equation sets by the added stabilization term "delta*A(v)" and
> add the new terms to the unstabilized variational form. But it is much
> easier if one can just change the weighing functions v to v+delta*A(v).
> is there anyway to do that?

This should be perfectly doable in FFC. What problem are you running
into?

It might be some small bug related to applying the operator to v
(which happens to be a Python list to since v is a subset of a mixed
test function).

-- 
Anders

Attachment: signature.asc
Description: Digital signature


References