Difference between revisions of "ET Workshop Fall 2011 MHD"
| Line 11: | Line 11: | ||
** all NR-based, only choice of variables | ** all NR-based, only choice of variables | ||
** Multiple guesses sometime necessary: should be "supported" | ** Multiple guesses sometime necessary: should be "supported" | ||
| − | |||
* Bruno: divergence cleaning worked smoothly until merger | * Bruno: divergence cleaning worked smoothly until merger | ||
** Putting everything in source-terms doesn't work well | ** Putting everything in source-terms doesn't work well | ||
| + | * Whisky implements | ||
| + | ** contraint transport (problems with AMR) | ||
| + | ** divergence cleaning (not very well tested) | ||
| + | ** vector potential | ||
| + | *** at the center of the cell works, using KO-Dissipation | ||
| + | *** at the edge works as well (e.g. Zach Etienne), but requires special operators in Carpet -> not public | ||
Revision as of 15:30, 31 October 2011
Current Status
- MHD implemented in "parallel" to non-MHD code in GRHydro
- Variables:
- HydroBase: primitives
- GRHydro: conservatives
- GRHydro: Calculate magnetic pressure in calculation of Tmunu, and store it
- Riemann: only HLLE and local Lax-Friedrich, no Roe, no Marquina (Bruno: HLLE)
- HLLC or HLLD should be better than HLLE
- Con2Prim
- works for polytropic cases
- all NR-based, only choice of variables
- Multiple guesses sometime necessary: should be "supported"
- Bruno: divergence cleaning worked smoothly until merger
- Putting everything in source-terms doesn't work well
- Whisky implements
- contraint transport (problems with AMR)
- divergence cleaning (not very well tested)
- vector potential
- at the center of the cell works, using KO-Dissipation
- at the edge works as well (e.g. Zach Etienne), but requires special operators in Carpet -> not public
