Prof. Bangerth, Thank you for the advice. I've spent some time with step-61 (and step-51 as well) and I think I understand the distinction between the two classes. I am using FEFaceValues to assemble the edge term from the interior faces, and attempting to use FEInterfaceValues to access the values from the cell face above.
One thing I am still unclear on is if FEInterfaceValues is the best (or only) way to access the face values of a field on a neighboring cell, or if this can be done manually. Right now, I'm using an approach like std::vector<double> field_at_qpts(sd.fe_interface_values.n_quadrature_points); sd.fe_interface_values.get_fe_face_values(1).get_function_values(field, field_at_qpts); where 'sd' is a typical scratchdata object with an FEInterfaceValues instance as a class attribute. Is this typically how neighboring data is accessed for DG-like methods? Thank you, Corbin On Friday, April 9, 2021 at 7:17:54 PM UTC-4 Wolfgang Bangerth wrote: > > Corbin, > > > I'm attempting to solve a local discontinuous Galerkin problem on each > cell in > > my mesh. I've attached a screenshot with the cell-local equation. > > > > Screen Shot 2021-04-09 at 13.56.15.png > > > > I am solving these cell problems in a very specific order to explicitly > > propagate information downwards in depth, so assembling an entire system > like > > in step-12 is not what I am looking to do. > > > > I'm able to assemble the interior terms without a problem, but I'm a bit > > unclear on how to assemble the edge terms. > > > > Should I be using FEFaceValues or FEInterfaceValues? On the top of the > cell, I > > need the face data from the cell above (for u_hat), whereas on the > bottom > > cell, I need to integrate over the bottom face. In either case, it's > unclear > > to me how to map the integrals involving face basis functions to the > volume > > basis functions in the cell. > > The distinction is that FEFaceValues evaluates the values/derivatives of > solutions and shape functions on faces *as seen from one cell*, whereas > FEInterfaceValues does so for averages/differences between adjacent cells. > In > your case, it seems to me that you are only interested in evaluating \hat > u as > seen from one side of the face, so FEFaceValues is the right tool. > > You might be interested in spending an hour or two on step-61, which also > solves local problems (in a different context, but maybe there is a thing > of > two for you to learn from). > > Best > W. > > -- > ------------------------------------------------------------------------ > Wolfgang Bangerth email: bang...@colostate.edu > www: http://www.math.colostate.edu/~bangerth/ > > -- 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. To view this discussion on the web visit https://groups.google.com/d/msgid/dealii/a25dffdb-74da-438a-b576-521858337b60n%40googlegroups.com.
