Hi Markus,
Yes, it works to use the original local stiffness matrix. But I want to
later separate the assemblage of the matrix and the right-hand side,
similar to what tutorial 32 does. So I was testing this easier case.
Somehow it doesn't work with this auxiliary matrix. I am just wondering
if there is anything wrong or missing in the code.
Thanks,
Huan
On Mon, 2010-05-17 at 08:29 +0200, Markus Bürg wrote:
> Hello Huan Sun,
>
> why do you create another matrix for the boundary conditions? You have
> already created the problem's matrix for this cell right above. Thus
> you can hand over just this matrix:
> constraints.distribute_local_to_global (local_rhs, local_dof_indices,
> rhs, local_matrix);
>
> Best Regards,
> Markus
>
>
>
> Am 17.05.10 08:12, schrieb Huan Sun:
> > Hi all,
> >
> > I was following tutorial 22 and 34 to implement inhomogeneous Dirichlet
> > boundary conditions using the ConsraintMatrix class but couldn't get it
> > right. I think the problem mainly lies in the assemblage part (i was
> > solving the basic Poisson's eq.)
> >
> > if(constraints.is_inhomogeneously_constrained
> > (local_dof_indices[i])){
> > for(unsigned int j=0; j <dofs_per_cell; ++j){
> > matrix_for_bc(j,i) +=
> > grad_basis_phi[i] * grad_basis_phi[j] *
> > fe_values.JxW(q);
> > }//j-loop
> > }//if
> > ...
> > constraints.distribute_local_to_global(local_rhs,
> > local_dof_indices,
> > rhs,
> > matrix_for_bc);
> >
> > However, when I comment out the if statement, the program seems to work
> > correctly. The code is attached in this email.
> > Your help will be greatly appreciated.
> >
> > Thanks!
> >
> > Huan
> >
> >
> > _______________________________________________
> > dealii mailing list http://poisson.dealii.org/mailman/listinfo/dealii
> >
> _______________________________________________
> dealii mailing list http://poisson.dealii.org/mailman/listinfo/dealii
_______________________________________________
dealii mailing list http://poisson.dealii.org/mailman/listinfo/dealii
