[Mesa-users] New Public Release 10398
ebauer at physics.ucsb.edu
Wed Mar 21 17:47:02 EDT 2018
Dear MESA Users,
We’re happy to announce a new public release of MESA: version 10398. Release notes below.
Evan (on behalf of the MESA developer team)
MESA 10398 Release Notes
Equation of State: PTEH, DT2, and ELM (Bill)
Several new options for the mesa/eos have been added, all aiming for more accurate partials for the Newton solver. All of these new eos options use bicubic spline interpolation in tables of lnPgas, lnS, and lnE as a way to get numerically accurate 1st and 2nd partial derivatives with respect to lnRho and lnT. The partials are directly calculated from the interpolating bicubic polynomials to give numerical accuracy, but this comes at a cost in thermodynamic consistency since the actual thermodynamic relations can only be approximated by bicubic splines.
The new eos options are called “PTEH”, “DT2”, and “ELM”. The PTEH tables are created using the approach of Pols, Tout, Eggleton, and Han (1995) as implemented by Paxton (2004) in a program derived from Eggleton’s stellar evolution code (1973). PTEH extends the mesa/eos coverage to lower densities than allowed by OPAL (down to 10^-18 g cm^-3) and higher metallicity than covered (OPAL stops at Z = 0.04 while PTEH covers all Z). When PTEH is enabled, it is used for low densities and for high Z in cases that for lower Z would be handled using data from OPAL/SCVH tables. In the old MESA EOS we fell back to HELM to provide approximate results for the cases now covered by PTEH.
The mesa/star default controls enable PTEH for both low densities and high Z.
use_eosPTEH_for_low_density = .true.
use_eosPTEH_for_high_Z = .true.
Z_for_all_PTEH = 0.040d0
Z_for_any_PTEH = 0.039d0
The two remaining new eos options, DT2 and ELM, provide high resolution tables in logRho and logT for values from mesa/eos for OPAL/SCVH values and for HELM respectively. These cover a subset of the standard eos domain with standard eos results for logPgas, logS, and logE in a form suitable for bicubic spline interpolation in order to give 1st and 2nd partials with high numerical accuracy. However, since use of DT2 and ELM will give decreased thermodynamic consistency that might not be compensated for by better residuals, these are both disabled by default in mesa/star.
use_eosDT2 = .false.
use_eosELM = .false.
Opacities (Josiah, Aaron)
The opacity module (kap) underwent some internal restructuring. The kap module now exposes only a single kap_get interface instead of separate kap_get_Type1 and kap_get_Type2 subroutines. This has two user-visible consequences.
* The control kappa_type2_logT_lower_bdy was removed. That control was no longer needed, as the existing control kappa_blend_logT_lower_bdy now also applies to Type2 opacities. All other related opacity controls (e.g., use_Type2_opacities) remain unchanged in name and behavior.
* Previously, there were separate "other" hooks for Type1 and Type2 opacities. Now, there is only one hook, other_kap_get. It has the call signature of the previous Type2 hook, which is a super-set of the arguments to the Type1 hook (see star/other/other_kap.f90).
In previous versions opacities where clipped to the edge values of the tables when logR=logRho-3logT+18<-8. This has been replaced for a blend to Compton opacities between logR=-7.5 and logR=-8.
Element Diffusion (Evan)
Fixed a bug in the ionization treatment for diffusion in the pressure ionization routine. This was due to a typo in the original paper that presented the ionization scheme. Restored the missing factor of rho^1/3 thanks to a later presentation of this same scheme (Dupuis et al. 1992) and a note here: http://www1.astrophysik.uni-kiel.de/~koester/astrophysics/astrophysics.html <http://www1.astrophysik.uni-kiel.de/~koester/astrophysics/astrophysics.html>.
Added a user control (D_mix_ignore_diffusion) for when to ignore element diffusion in surface or core mixing regions. Previously, diffusion would be turned off for surface mixing regions of ANY strength, even very weak mixing where diffusion might still be relevant. Now this control is set to a D_mix of 10^5 (cm^2/s), so that mixing that will obviously overwhelm diffusion (like convection) will turn it off, but weaker mixing won’t.
Gravity Darkening (Aaron)
Added options to include gravity darkening, in the form of projected (surface-averaged) luminosities and effective temperatures of the star viewed along the equator and pole, to the history file. Assumes the star is an oblate spheroid; see https://github.com/aarondotter/GDit <https://github.com/aarondotter/GDit> for more info.
Isomers (Frank, Josiah, Bill)
The isomers of 26Al can now be added to a reaction network. To use them, include the isomers in your network specification file. Two examples include
h 1 1 ! hydrogen
he 4 4 ! helium
mg 25 25 ! magnesium
al26-1 ! ground state
al26-2 ! meta-stable excited state
One may use either al26 or al26-1 and al26-2. Reaction rates for the 26Al isomers with other isotopes are picked up from the JINA reaclib file. Reaction rates for al26-1 <-> al26-2 are from Gupta & Meyer (2001) [http://adsabs.harvard.edu/abs/2001PhRvC..64b5805G <http://adsabs.harvard.edu/abs/2001PhRvC..64b5805G>] and located in data/rates_data/rate_tables along with the new default rate_list.txt file.
if you want a local rate_tables directory, http://mesa.sourceforge.net/star_job_defaults.html#rate_tables_dir <http://mesa.sourceforge.net/star_job_defaults.html#rate_tables_dir>, and you want the 26Al isomers, then the two al26-1 <-> al26-2 rate files must be copied from their default location to your local rate_tables directory and your local rate_list.txt modified to include these two rates.
Installation Debugging (Rob, Josiah)
There is a new command, $MESA_DIR/help which outputs system information we need when debugging installation issues and/or MESA crashes. ./install will now also log its output to a file $MESA_DIR/build.log, if you have an installation issue please include this file when reporting an issue to mesa-users.
Miscellaneous improvements (Rob, Josiah):
You can now use the MESA_INLIST environment variable to set the name of the main inlist file when using MESA binary.
The output cadence of MESA binary has been tweaked to that its behavior is the same as MESA star. (If you use the same options, you should get output at the same steps.)
There is now a flag b% need_to_update_binary_history_now, which if set forces binary history output to occur at the current step.
Put calls to extra_header_items back into standard_run_star_extras and provided working examples of how to call all of them. These are useful for adding extra information to the history and profile headers beyond what is provided by default, such as including mixing_length_alpha in the history file header.
Building with Other Compilers
MESA currently does not compile with ifort.
Other non-SDK compilers that are known to work (at the bit-for-bit level):
Gfortran 7.3.1 (fedora 27)
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the Mesa-users