Hi Wolfgang, Thank you for your reply. Where the discontinuities are depends on the initial_refinement integer. When this was 6 it produced the figure I sent originally - sorry I should have said this. This is what I found particularly odd, that for different levels of refinement, the behaviour is different, as the VectorTools::projection occurs after the refinement of the grid.
Many thanks, Katie Leonard DPhil student in Computational Biology, The University of Oxford. ________________________________________ From: Wolfgang Bangerth [[email protected]] Sent: 11 November 2011 13:43 To: Katie Leonard Cc: Guido Kanschat; [email protected] Subject: Re: [deal.II] Projecting Initial Values without interpolation Katie, > I do apologise for that mistake. I hope that what I attach now is OK. The picture I get from running this program looks different than the one you had posted before (mine is attached, I used visit for visualization but it looks similar with gnuplot). In particular, the overshoots are now along the entire line of discontinuity and I do in fact get undershoots. Does the solution not look like this for you? Also: This is how the projection of a discontinuous function looks like. You get over and undershoots, that's just how it is -- it's called Gibbs phenomenon. If you don't want those, use an interpolation instead. Best W. ------------------------------------------------------------------------- Wolfgang Bangerth email: [email protected] www: http://www.math.tamu.edu/~bangerth/ _______________________________________________ dealii mailing list http://poisson.dealii.org/mailman/listinfo/dealii
