[mesa-users] Test suite CCSN problem

[Francis Timmes]

hi manos,

regarding point 1, recall that the E in mesa stands for experiment.
i suggest experimenting with other values for the energy injection, 
calculating the binding energy of the pre-sn model yourself, and so on. 

fxt




> On Jun 18, 2015, at 5:39 PM, Manos Chatzopoulos <manolis at flash.uchicago.edu> wrote:
> 
> Hi Robert,
> 
> Yeah that seems to work so far for point 2, thanks!
> 
> Manos
> 
> On 6/18/15 3:59 PM, Robert FARMER wrote:
>> On point 2)
>> you can remove the 
>>          max_number_backups
>>          max_number_retries
>> 
>> lines from your inlists so that mesa can try unlimited retries. Though there is no guarantee that mesa can find a converged solution with unlimited retries.
>> 
>> Rob
>> 
>> On Thu, Jun 18, 2015 at 1:41 PM, Manos Chatzopoulos <manolis at flash.uchicago.edu> wrote:
>> Hi all,
>> 
>>   The original, test_suite ccsn problem in the latest public release of
>> MESA runs fine but I have a couple of question
>> with regards to optimizing it for another application:
>> 
>> 1) Is there a special reason this very particular and accurate value for
>> SN energy injection in the inlist_edep
>> was chosen?:
>> 
>>       inject_until_reach_delta_total_energy = 1.51935238181291335e51
>>        inject_extra_ergs_sec = 1.51935238181291335e51 ! erg/s
>> 
>> Is that a value determined based on the gravitational bounding energy of
>> the core, an integrated
>> thermal energy required to unbind the star or something else? Because if
>> I choose this value
>> even by a bit (1.6e51 erg) the ccsn run does not work (it fails during
>> edep and before explosion).
>> 
>> 2) Even without changing the injected energy in inlist_edep I wished to
>> follow the SN shock for
>> longer; obviously not until shock break-out (that cannot be done
>> properly at this time) but at least
>> up to mass of ~10-11 Msun (out of ~12 Msun that is left - therefore
>> still within the optically thick
>> region of the star, deep beneath the surface and in the hydrogen envelope).
>> I changed   x_ctrl(1) to 10. thinking this will do the trick but the run
>> failed again with convergence problems:
>> 
>>                        dt_div_dt_Courant        3.8739805487249512
>>                              total_energy    1.2004485474821978D+51
>>           log_rel_cumulative_energy_error       -2.8599403101283558
>>               rel_cumulative_energy_error       -0.0013805739990469
>>          rel_error_in_energy_conservation   -7.4721951660307834D-10
>>                            avg_E_residual   -2.8042773685002612D-10
>>                            max_E_residual   -1.4603660217645308D-06
>> 
>> hydro_newton_step returned ierr
>>         retry cnt, step, log10(dt/yr), retry_factor     1  1672
>> -6.324885        0.500000
>>                                              num_retries         401
>>                                       max_number_retries         400
>> 
>>   stopping because of convergence problems -- too many retries
>> 
>> 
>> terminated evolution: nonzero_ierr
>> 
>> termination code: max_number_retries
>> 
>> Anyone has a trick that will allow me to follow the shock for longer and
>> further out in the star?
>> 
>> Thanks,
>> 
>> Manos
>> 
>> --
>> Dr. Emmanouil Chatzopoulos
>> Enrico Fermi Postdoctoral Fellow
>> Department of Astronomy & Astrophysics
>> FLASH Center for Computational Science
>> University of Chicago
>> https://www.linkedin.com/pub/manos-chatzopoulos/a1/688/a03
>> https://astro.uchicago.edu/people/manos-chatzopoulos.php
>> 
>> 
>> ------------------------------------------------------------------------------
>> _______________________________________________
>> mesa-users mailing list
>> mesa-users at lists.sourceforge.net
>> https://lists.sourceforge.net/lists/listinfo/mesa-users
>> 
> 
> -- 
> Dr. Emmanouil Chatzopoulos
> Enrico Fermi Postdoctoral Fellow
> Department of Astronomy & Astrophysics
> FLASH Center for Computational Science
> University of Chicago
> 
> https://www.linkedin.com/pub/manos-chatzopoulos/a1/688/a03
> https://astro.uchicago.edu/people/manos-chatzopoulos.php 
> ------------------------------------------------------------------------------
> _______________________________________________
> mesa-users mailing list
> mesa-users at lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/mesa-users