Toulouse 2D numerical radiative transfer codes : MALI and Gauss-Seidel
Authors : F. Paletou and L. Léger (Contact: frederic.paletou/at/univ-tlse3.fr)
- Purpose:
We propose two numerical codes for 2D multilevel non-LTE radiative transfer, assuming complete frequency redistribution, for static, homogeneous and isolated structures. They were initially developed for solar prominence radiative modeling of various spectral lines of H or He. Two versions are proposed, respectively using the following numerical schemes:
- Multilevel Accelerated Lambda Iteration (MALI),
- A Gauss-Seidel iterative scheme in a 2D cartesian grid.
- More informations:
- Léger, Chevallier & Paletou, 2007, A&A, 470, 1-9
- Paletou & Léger, 2007, Journal of Quantitative Spectroscopy & Radiative Transfer, 103, 57-66
- Lambert, Paletou, Josselin & Glorian, 2016, Eur. J. Phys., 37
- Auer & Paletou, 1994, A&A, 284, 675-686
- PhD, L. Léger, 2008, "Transfert de rayonnement hors-ETL multidimensionnel. Application au spectre de l'hélium dans les protubérances solaires" (text in French, but many developments worth it for anyone)
More references :
Léger & Paletou, 2009, A&A, 498, 869-875 / Paletou, 1997, Astron. Astrophys., 317, 244-247
Paletou & Auer, 1995, Astron. Astrophys., 297, 771-778 / Heinzel, 1995, A&A, 299, 563-573
Trujillo & Fabiani, 1995, the Astrophysical Journal, 455, 646-657 / Rybicki & Hummer, 1991, Astron. Astrophys., 245, 171-181
Olson, Auer & Buchler, 1986, J. Quant. Spectros. Radiat. Transfer, vol 35, N°6, 431-442
Mihalas, Auer & Mihalas, 1978, the Astrophysical Journal, 220, 1001-1023 / Avrett, 1968, Resonance lines in Astrophysics, NCAR
- Fortran files and documentations:
MALI code and Gauss-Seidel code were successfully tested on PCs running linux with gfortran compiler. The source codes are available here : MALI, Gauss-Seidel. For running the code, see the README file. For more details on the codes, see the documentation.
Tutorial about iterative methods including Gauss-Seidel iterations in radiative transfer : http://rttools.irap.omp.eu/
- Initiation to MALI : 1D MALI code. The source codes are available here.
- Input files (test case):
MALI code :
- input : file including the number of reduced frequencies, the number of directions, the grid in Z (max. optical depth, first optical depth step, number of points for a logarithmic-type mesh for isothermal slabs), the grid in Y (max. optical depth, first optical depth step; same as Z), the choice to read the grid from extra file, the maximum number of iterations, the uniform slab temperature (K), the model-atom (H, Ca or Na), the (fixed) background continuum opacity, the number of 'natural' iterations before applying acceleration of convergence, the expected relative error (from an iteration to another)
- several reference data files of populations (H, Ca, Na) from 2D slab models, resulting from 'fully converged' (i.e. using a large number of iterations) populations: populationsH.save, populationsCa.save, populationsNa.save
- several reference data files of populations (H, Ca, Na) from 1D slab models of similar vertical extension as 2D slabs: populationsH1d.res, populationsCa1d.res, populationsNa1d.res
Gauss-Seidel code :
- input : file including the number of reduced frequencies, the number of directions, the grid in Z (max. optical depth, first optical depth step, number of points for a logarithmic-type mesh for isothermal slabs), the grid in Y (max. optical depth, first optical depth step, same as Z), the maximum number of iterations, the uniform slab temperature (K), the model-atom (H, Ca or Na), the (fixed) background continuum opacity, the option SOR (Successive Over-Relaxation)
- Output files (test case):
MALI code :
- population2d.res : density of populations in 2D obtained after convergence
- population.res : density of populations along mid symmetry axis for 1D comparisons, obtained from 2D computations
- mali2d.res : file including three sets of maximum relative errors on source function, populations and vs. 1D populations of reference
Gauss-Seidel code :
- sor2d.res : file including the SOR parameter value, the density of populations in 2D, the maximum relative error on source function, and on populations from an iteration to another
Back to the other pages :
PRESENTATION DATA TOOLS PUBLICATIONS NEWS