I am solving the 2D Stokes equations with Dirichlet BCs, so I have 3 unknowns (u1, u2, p) with Dirichlet BCs on u1 and u2. Rather than interpolate, I'd like to project the Dirichlet boundary values.
Looking at the documentation for VectorTools::project_boundary_values(), the description of the "std::vector< unsigned int > component_map" input describes exactly this scenario, seemingly indicating that using a component_map will allow me to project values to 2 of the 3 components of my finite element. However, reading the remainder of the description, and attempting to use the component_map, this seems to be not the case. What I'd need is a component_map of {0,1}....that is, the first boundary_function should apply to u1, and the second boundary_function should apply to u2. However, the map goes in the opposite direction....that is, it is required to have 3 components {0,1,?}: u1 gets the first boundary_function applied to it, u2 gets the second boundary_function applied to it, and p gets the ? boundary function applied to it. Is this part of the documentation incorrect when it mentions the Stokes problem as a use for the component_map? In any case, what is a recommended method to projecting the boundary values for the Stokes problem? Thanks, Stephen -- The deal.II project is located at http://www.dealii.org/ For mailing list/forum options, see https://groups.google.com/d/forum/dealii?hl=en --- You received this message because you are subscribed to the Google Groups "deal.II User Group" group. To unsubscribe from this group and stop receiving emails from it, send an email to dealii+unsubscr...@googlegroups.com. For more options, visit https://groups.google.com/d/optout.