Hi,
No, you can't use a face quadrature rule in one-dimension as the edge
is simply a point and thus there is nothing to do. Thus, deal will
throw an exception.
This should work:
for (unsigned int vertex = 0; vertex <
GeometryInfo<1>::vertices_per_cell; ++vertex)
if (cell->at_boundary(1) == true) {
const double neumann_value = -0.5;
cell_rhs(fe.dofs_per_cell) += (neumann_value);
}
Andrew
On 29 Sep 2009, at 9:26 AM, Pietro Maximoff wrote:
From: Andrew McBride <[email protected]>
To: Pietro Maximoff <[email protected]>
Cc: [email protected]
Sent: Tuesday, September 29, 2009 7:39:28 AM
Subject: Re: [deal.II] 1D Boundary-Value Problem
Hi Pietro
Why are you note using a volume integral to evaluate the Neumann
contribution at a face (i.e. a point)? Surely you should be using
the equivalent of the face quadrature in 1D? I know this is covered
in the docs somewhere.
regards
Andrew
============
Do you mean using a face quadrature formula like so:
QGauss<dim-1> face_quadrature_formula(3); ?
Well, QGauss<0> generates an exception. So, doing it this way has
been impossible so far.
Incidentally, the exact solution is: 0.5ln(x) + 2/x.
Please help.
Thanks.
Pietro
_______________________________________________
dealii mailing list http://poisson.dealii.org/mailman/listinfo/dealii
