Dear Bärbel,
> I extended the tutorial step 31 to a Navier-Stokes solver. There is going > to be a problem with the strong boundary conditions. If I use > MatrixTools::apply_boundary_conditions() then the GMRES solver takes more > than 6000 interations for 7000 dofs in 2d. Is it possible to avoid the use > of MatrixTools::apply_boundary_conditions? Do the problems appear already when you use the Stokes equations and Dirichlet boundary conditions? If not, then the problem is more likely caused by the different equation structure of the Navier--Stokes equations compared to the Stokes equations (non-symmetry in case you discretize fully implicitly, and you need another Schur complement approximation). If the problems do appear already for Stokes, then I guess that you are trying to solve a rank-deficient system. That arises for example when you set Dirichlet conditions on velocity, but no conditions on the pressure. In such a case, the pressure is only determined up to a constant, so you would need to fix one pressure node value. Anyway, it can be useful to have a look at a solution field (if you get that far). It usually tells you what went wrong. Best, Martin _______________________________________________ dealii mailing list [email protected] http://poisson.dealii.org/mailman/listinfo/dealii
