On Tue, 11 May 2010, David Knezevic wrote: >> Oh, my mistake, I just needed to add system.mesh_position_set() after >> system.time_solver->advance_timestep(). It works nicely now, cool! > > One last post about this: I guess you didn't expect the numerical jacobian > verification to work with mesh motion? When I turned the numerical jacobian > verification on it reports an error, even when the x and y variables are > fixed at their initial values.
The numerical jacobian verification should give false positive failure reports, yeah. You've still got a CFDLab login, right? Check out ~roystgnr/pecos/ale/doc/paper.pdf for some discussion of the problem. Basically there's no way to fix it for existing codes, but the most straightforward fix wouldn't take too much in the way of application changes and would also make it easy to add well-modularized Petrov-Galerkin method support at the library level, so that's at least on the long-term TODO list. --- Roy ------------------------------------------------------------------------------ _______________________________________________ Libmesh-devel mailing list Libmesh-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/libmesh-devel
