Hi there,
I am studying how to code up the Neumann boundary condition with step-7. I
am confused by following segment of code in step-7
**********************************************************************************************************
for (unsigned int face=0; face<GeometryInfo<dim>::faces_per_cell; ++face)
if (cell->face(face)->at_boundary()
&&
(cell->face(face)->boundary_indicator() == 1))
{
fe_face_values.reinit (cell, face);
for (unsigned int q_point=0; q_point<n_face_q_points;
++q_point)
{
const double neumann_value
= (exact_solution.gradient
(fe_face_values.quadrature_point(q_point)) *
fe_face_values.normal_vector(q_point));
for (unsigned int i=0; i<dofs_per_cell; ++i)
cell_rhs(i) += (neumann_value *
fe_face_values.shape_value(i,q_point) *
fe_face_values.JxW(q_point));
}
}
*******************************************************************************************************
If I understand it correctly, for the 2D problem, a particular cell, you
detect one of four faces of this cell with indicator 1. Then you do the
integration on this face. You need two values on two n_face_q_points
(bilinear shape function) using fe_face_values. But what
is fe_face_values.shape_value(i,q_point) where i=1->4? Because you are
using shape function on a face, the i in
the fe_face_values.shape_value(i,q_point) is the dofs_per_face and can only
be 1 and 2. However, you use dofs_per_cell that is 1 to 4. I am really
confused. Is the shape function for the Neumann boundary, in the weak form
f(K,i), a shape function on face or on cell?
--
Dr. Nan Zhang
Postdoctoral Research Associate
Department of Geological Science
Brown University
324 Brook Street
Box 1846
Providence, RI 02912
HomePage: http://www.geo.brown.edu/People/Postdocs/Zhang/NanZhangHome.html
_______________________________________________
dealii mailing list http://poisson.dealii.org/mailman/listinfo/dealii
