search

mwiesenberger / feltor

Numerical methods for edge and SOL blob and turbulence simulations
http://feltor-dev.github.io
MIT License
4 stars 2 forks source link

Added chi weighted jump factor to the elliptic operator. #5

Closed aslakp closed 4 years ago

aslakp commented 4 years ago

Possibility to weight the jump term by \chi in the elliptic operator.

mwiesenberger commented 4 years ago

Also, does the addition of the weights preserve the symmetry of the operator?

aslakp commented 4 years ago

Hi Matthias, I changed the documentation to include the class documentation. I tried to be a concise as possible, but let me know if it is unclear. In regards to the convergence, both cases seem to converge just fine. For the 3d case however, the weighted case need a few more iteration. For the 2D case, it is more or less the same. Note, this was not a rigorous testing. In regards to the symmetry, I can't say for sure. Mathematically i believe it should be conserved, but I'm not sure how to check it numerically.

mwiesenberger commented 4 years ago

Regarding symmetry I think chi breaks it because (Chi J)^T = J^T Chi^T = J Chi, so maybe it should not be used in connection with CG?