Hi all, Does anyone here have any experience applying mean value constraints (specifically with periodic boundary conditions)? I'm having some trouble. As far as I can tell, there are two approaches (both of which I have been able to get working). The first approach is to just add the mean value constraint straight into the constraint matrix. I got this working just fine and was able to get an output but the problem is that this condition seems to make my matrices dense(?) or, at the very least, require a huge amount of memory hindering this approach for the types of fine spatial meshes I require for this application. The other approach is to just apply periodic boundary conditions to the matrices, try to solve the system anyway and, because UMFPACK is awesome, somehow manage to get a solution out of all of this (which will of course only be defined up to a constant). The problem here is that the solution I am getting is of order 10^12! When I postprocess and subtract the mean value, what I get looks correct but I am losing a lot of accuracy in the process but it does work for the fine meshes that I require. Anyone have any ideas or do I just have to live with the reduced accuracy here? Thanks!
-- 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/a8f1c121-14a6-48c0-9127-288fdf63840dn%40googlegroups.com.
