> On Mar 5, 2016, at 1:12 AM, Giorgio Bornia <giorgio.bor...@gmail.com> wrote: > > My starting point is a fine mesh of a real geometry that was generated by > someone else. > To be more precise, what I have is an STL geometry of a surface which is made > of rather fine triangles. > I have an algorithm to generate the mesh in the volume enclosed by the > surface, but this volume mesh will be very fine, as it starts from a fine STL > file. > > So, I was thinking of coarsening the STL first (at least where I need less > resolution, say in flatter regions) > and rerun my algorithm, which will then create a coarser mesh. > > I need a mesh as coarse as possible to begin with, because I am building > geometric multigrid operators on it, > so I start from a coarse level and I build projections (I will then get the > restrictors as transposes). > > Any suggestions?
I just heard a talk that mentioned GRUMMP from UBC. http://tetra.mech.ubc.ca/GRUMMP/ It is supposed to do volume preserving coarsening, and the speaker was using it for MG type methods. I haven't used it myself though. ------------------------------------------------------------------------------ _______________________________________________ Libmesh-users mailing list Libmesh-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/libmesh-users
