Hey Soon Hoe, one thing you can do for your initial time step (and also if you suddenly change your external heat source), is to redo those time steps a couple of times until the mesh had time to adapt. For example: 1. do first time step 2. adapt mesh 3. go back to 1) for x number of times 4. your normal code with adaption every y steps.
We are doing something like this in step-32. Does that make sense? On Thu, May 3, 2012 at 4:09 PM, Soon Hoe Lim <[email protected]> wrote: > > Hi. I am solving a 2-d time independent heat conduction equation using > deal.II using theta scheme (implicit Euler) as my time discretization > scheme and local adaptive refinement for mesh refinement. The domain is > [-1,1], with initial condition T=1 for a square of length 0.2 at the > center, and T=0 elsewhere. Boundary condition is 0 at all the boundaries. I > realize that if I start with a rather coarse mesh (say put an initial > global refinement of 3), the resolution of the mesh at initial time steps > is very poor and the solution won't be accurate. But if I start with a fine > mesh (say start with global refinement of 5), the resolution improves. > > But if I wish to add in external heat source (say T=1) at some later time > > 0 at the location where the meshes are coarse, I will get poor resolution > at that location. Any idea on how to deal with this problem? Or is it fine > if I get poor resolution? Thank you! > > Sincerely, > Soon Hoe > _______________________________________________ > dealii mailing list http://poisson.dealii.org/mailman/listinfo/dealii -- Timo Heister http://www.math.tamu.edu/~heister/ _______________________________________________ dealii mailing list http://poisson.dealii.org/mailman/listinfo/dealii
