Anna,

> For now, to refine the mesh, I use KellyErrorEstimator, and in general it
> produces different error_estimates for the two solutions.
> So I can have two different meshes after mesh refinement step. However since
> the system matrix is the same I would like to have a single mesh for both
> polarizations. So my question is what is the best way to do this?

The correct way would be to design your own estimator that treats the
system you are looking at (Kelly is designed for Laplace problems and
might not even be robust and reliable for your PDE). But you can
probably argue that the sum of the two estimators is a reasonable
estimator for the problem. You can achieve this by adding the two
Vector<double> that contain the estimates per cell given by the Kelly
estimator before doing the refinement with it.
This of course assumes that the errors are scaled in a similar way
(same units, etc.). Otherwise you might need to rescale and/or do
something more clever. We discuss some strategies in the paper
High Accuracy Mantle Convection Simulation through Modern Numerical
Methods. II: Realistic Models and Problems
Geophys J Int, 2017, 210, 833-851.

-- 
Timo Heister
http://www.math.clemson.edu/~heister/

-- 
The deal.II project is located at http://www.dealii.org/
For mailing list/forum options, see 
https://groups.google.com/d/forum/dealii?hl=en
--- 
You received this message because you are subscribed to the Google Groups 
"deal.II User Group" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to [email protected].
For more options, visit https://groups.google.com/d/optout.

Reply via email to