On Wed, 11 Feb 2015, Vikram Garg wrote: > More than visualization, I need to be able to associate sensitivities with > spatial locations for uncertainty propogation. I can do this via the > GMV/Exodus route but was looking for a more direct route via libMesh.
Once you've got a NumericVector with the sensitivities set (whether for the original system or for a separate ExplicitSystem with fewer variables as Derek suggested), then you can use that in a MeshFunction to evaluate values at points, including node points. You'll want to be really careful here, though. If you refine some boundary elements, for example, then the sensitivities to boundary condition values on the refined grid won't be a higher-accuracy representation of some infinite dimensional function, the way refined solutions are. The point-evaluated sensitivities will *decrease* on the refined parts of the mesh. You really need to pair this sort of dual-solution with a standard solution (e.g. a BC perturbation function) in the original space to get a sensible output from it. --- Roy ------------------------------------------------------------------------------ Dive into the World of Parallel Programming. The Go Parallel Website, sponsored by Intel and developed in partnership with Slashdot Media, is your hub for all things parallel software development, from weekly thought leadership blogs to news, videos, case studies, tutorials and more. Take a look and join the conversation now. http://goparallel.sourceforge.net/ _______________________________________________ Libmesh-users mailing list Libmesh-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/libmesh-users
