Difference between revisions of "ET Workshop Fall 2011 RAD"
(→Implementations) |
(→Implementations in "GR") |
||
| (One intermediate revision by the same user not shown) | |||
| Line 32: | Line 32: | ||
** in 3D competitive with other methods in terms of computational needs | ** in 3D competitive with other methods in terms of computational needs | ||
| − | ==Implementations== | + | ==Implementations in "GR"== |
* Shibata et al. | * Shibata et al. | ||
** Leakage | ** Leakage | ||
| Line 45: | Line 45: | ||
** Monte-Carlo | ** Monte-Carlo | ||
*** 1D only | *** 1D only | ||
| + | * AEI | ||
| + | ** Leakage | ||
Latest revision as of 16:31, 2 November 2011
Possibilities
- Leakage scheme
- very ad-hoc
- only mimicks loss due to radiation
- Diffusion equation
- accurate only when mean free path is really short
- almost always implicit to avoid strong CFL limit
- cheap, usually inaccurate
- Flux limited diffusion
- Diffusion + "trick" at lower optical depths
- also simple, and has been used for supernovae simulations
- unclear if working in GR
- limits radiation speed: ad-hoc
- also not very accurate
- Two-Moment-Scheme / Truncated moment scheme
- start with Boltzmann equation
- decompose into moments
- truncate series
- close using some prescription (approximate the rest of the series)
- Full Transport equation
- Full discretization of Boltzmann equation
- extremely complex
- used, e.g., in reactor simulations
- SN/PN discretize equation in different way
- SN not scaling well
- PN might work better
- Monte-Carlo
- no equations, direct simulations
- very simple to implement complicated processes
- more expensive
- given enough computational power, very accurate
- in 3D competitive with other methods in terms of computational needs
Implementations in "GR"
- Shibata et al.
- Leakage
- Truncated Moment Formalism
- Shapiro et al.
- Diffusion approximation, grey, 3D
- Garching
- Full Boltzmann transport
- 1D, ray-by-ray
- Caltech / Cactus
- Leakage scheme
- Monte-Carlo
- 1D only
- AEI
- Leakage
