On Wed, 6 Feb 2013, Kyunghoon Lee wrote: > I'm trying to solve a simple Laplace equation with FEMsystem based on > introduction_ex3.C and fem_system_ex1.C (It seems easy to expand a steady > model to a transient model with FEMsystem). One thing I cannot figure out > is how to add the element matrix and RHS with FEMsystem. Do I need to > override element_time_derivative or assembly? I'd appreciate if someone > could briefly describe necessary steps.
If you're solving only the steady case, then you could choose to override element_time_derivative or element_constraint; which one you choose would determine what happens if you later switch to a transient solver. My apologies for the inadequate FEMSystem documentation; I'm not sure when one of us is going to find time to fix that. If you want some code to start from, we've actually just added a FEMSystem based du/dt=div(grad(u)) solver in adjoints_ex5 in the git master. It's got a bunch of code you don't need for doing transient adjoint-based sensitivity evaluations, but if you ignore those parts then heatsystem.h and heatsystem.C are pretty much what you're looking for. --- Roy ------------------------------------------------------------------------------ Free Next-Gen Firewall Hardware Offer Buy your Sophos next-gen firewall before the end March 2013 and get the hardware for free! Learn more. http://p.sf.net/sfu/sophos-d2d-feb _______________________________________________ Libmesh-users mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/libmesh-users
