Timo, That's exactly what I was looking for. It was dumb of me to not see it from the example title. Thanks for the pointer !
Just a thought: why is this not the default behavior when doing AMR ? Perhaps a user flag to control whether the transfer should be done would be a nicer API. There is nothing to be gained in solving the same refined system but with a zero guess unless in say modified systems when computing only the change in solution. It could be that the examples just weren't updated but it'll be useful to do that for relevant cases. My 2 cents. Vijay On Thu, May 10, 2012 at 2:23 PM, Timo Heister <[email protected]> wrote: > Hi Vijay, > > you are looking for the SolutionTransfer class. Take a look at step-15 > for example. > > Best, > Timo > > On Thu, May 10, 2012 at 1:59 PM, Vijay S. Mahadevan <[email protected]> wrote: >> I am trying to preserve the computed solution on a coarser grid and >> project it to a refined mesh after performing AMR based on a specified >> criteria, after performing the solve in a cycle. I see that all of the >> examples (say step-6), the solution is reset to zero when performing >> solution.reinit() but I do not want this to happen. Is there a way to >> use VectorTools::interpolate or VectorTools::project to consistently >> perform an L2 projection on the locally refined mesh ? My initial >> tries with these functions did not yield much because I do not store a >> copy of the old dof_handler. Would this be the way to go ? > > > -- > Timo Heister > http://www.math.tamu.edu/~heister/ _______________________________________________ dealii mailing list http://poisson.dealii.org/mailman/listinfo/dealii
