On Wed, 14 Oct 2009, Joa Ljungvall wrote: > So returning to the "fix" using the constraint equations... I don't quite > understand the idea behind this. If my hanging node no longer is at the right > position, my mesh is broken, no?
Not quite. If your hanging node is no longer at the right position, then the mesh is geometrically broken, but that can be fixed by simply moving the node to a consistent position immediately, before the mesh is topologically changed. The refinement cycle is what turns a geometrically broken mesh into a topologically broken mesh, and that's what can't be fixed. > So fixing this with the constraint equations is a way of making the > calculations go on, assuming that the error we introduced this way > is small, and not a complete solution to the problem? I believe this is a complete solution to the problem for adaptive refinement - it would "un-break" the mesh before the breaks could have any real effect. For adaptive coarsening, however, it now occurs to me that we wouldn't be able to use the DofConstraints to restore hanging node locations, because the constraint coefficients are built based on geometry-dependent evaluations and would give the wrong results for newly-hanging nodes. I'm not sure what the best solution is in the coarsening case. I suggested using the DofConstraints because that code already works out the hairy problems (as in the example I mentioned) of recursively expanding constraints with nested dependencies... > This is way I came up with the idea of moving the nodes in groups... This wouldn't work for adaptive coarsening either, I'm afraid. --- Roy ------------------------------------------------------------------------------ Come build with us! The BlackBerry(R) Developer Conference in SF, CA is the only developer event you need to attend this year. Jumpstart your developing skills, take BlackBerry mobile applications to market and stay ahead of the curve. Join us from November 9 - 12, 2009. Register now! http://p.sf.net/sfu/devconference _______________________________________________ Libmesh-users mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/libmesh-users
