[Mesa-users] Model terminating after main sequence?

[Francis Timmes]

hi trevor,

varcontrol_target as a timestep controller is a bad idea because its not really a timestep controller. 
it is wiser to use limits on physical parameters to control the timestep.
a solution that has been tested over a wide range of masses in 12115 and 12778 is

! temporal resolution
     varcontrol_target= 5.0d-3


! time step resolution on temperature and density changes
      delta_lgT_cntr_limit   = 0.001   ! default 0.01
      delta_lgT_limit = 0.05d0         ! default 0.5

      delta_lgRho_cntr_limit = 0.005   ! default 0.05
      delta_lgRho_limit = 0.1          ! default 1.0


as the snippet below has a recognizable piece of one of my inlists, 
i get to flog myself repeatedly and often. i invite bill to pile on
about misuse of varcontrol_target.

fxt








> On May 18, 2020, at 3:27 PM, Francis Timmes <fxt44 at mac.com> wrote:
> 
> hi trevor,
> 
> i recognize the inlist from a mesa summer school ...
> 
> ebraheem farag and i have been seeing exactly what you describe.
> we've made some inlist adjustments and are running a grid under release r12778 as i write.
> so far so good, but give us another day or two before replying more substantially.
> 
> fxt
> 
> 
> 
> 
> 
> 
> 
> 
>> On May 15, 2020, at 2:51 PM, Trevor Dorn-Wallenstein via Mesa-users <mesa-users at lists.mesastar.org> wrote:
>> 
>> Hi There!
>> 
>> I've been running a grid of massive (15-30 M_sun in 1 M_sun steps) Z=0.006 models. I'm running into an odd problem where above 17 or 18 M_sun, the star leaves the main sequence, does a Henyey hook, then the timestep abruptly drops to zero and the model terminates. Tweaking the time/spatial resolution might change the mass where this starts by +/-1 M_sun, and sometimes fixes it for one of the higher mass models, but nothing works for the whole grid. It also seems to be insensitive to the input physics. I'm running the models with macOS 10.15.2, MESA version 12778, MESA SDK version 20.3.1.
>> 
>> I admit that in my quest to find an answer, I've thrown what is probably too many controls at this problem. I've attached an HR diagram, ZAMS model that I'm starting the 18 M_sun run from, and the inlist. Weirdly enough, for my most recent runs, the "successful" models have very low time resolution. The time/space resolution controls I'm using are:
>> 
>>  ! Finer mesh
>>    mesh_delta_coeff = 0.5 !0.75
>>    mesh_delta_coeff_for_highT = 1.0
>>    max_allowed_nz = 50000
>> 
>>  ! some mesh settings to increase around sensitive areas
>>    mesh_dlog_pp_dlogP_extra = 0.015
>>    mesh_dlog_cno_dlogP_extra = 0.15 
>>    mesh_dlog_3alf_dlogP_extra = 0.15 
>>    mesh_dlog_burn_c_dlogP_extra = 0.15 
>>    mesh_dlog_burn_n_dlogP_extra = 0.15 
>>    mesh_dlog_burn_o_dlogP_extra = 0.15 
>>    mesh_dlog_burn_ne_dlogP_extra = 0.15 
>>    mesh_dlog_burn_na_dlogP_extra = 0.15 
>>    mesh_dlog_burn_mg_dlogP_extra = 0.15 
>>    mesh_dlog_cc_dlogP_extra = 0.15 
>>    mesh_dlog_co_dlogP_extra = 0.15 
>>    mesh_dlog_oo_dlogP_extra = 0.15 
>> 
>>  ! temporal resolution
>> 
>>    varcontrol_target = 5d-5
>>    min_timestep_limit = 1d-12 ! (seconds) 
>>    max_timestep_factor = 1.10d0 !1.15d0
>>    min_timestep_factor = 0.05
>> 
>>  ! resolve the depletion of each fuel
>>    delta_lg_XH_cntr_limit = 0.05d0
>>    delta_lg_XH_cntr_max   = 0.0d0
>>    delta_lg_XH_cntr_min   = -6.0d0
>>    !delta_lg_XH_cntr_hard_limit = 0.02d0 !turned off for most recent run
>> 
>>    delta_lg_XHe_cntr_limit = 0.05d0
>>    delta_lg_XHe_cntr_max   = -1.0d0
>>    delta_lg_XHe_cntr_min   = -6.0d0
>>    !delta_lg_XHe_cntr_hard_limit = 0.02d0
>> 
>>    delta_lg_XC_cntr_limit = 0.05d0
>>    delta_lg_XC_cntr_max   = -1.0d0
>>    delta_lg_XC_cntr_min   = -6.0d0
>>    !delta_lg_XC_cntr_hard_limit = 0.02d0
>> 
>>  ! this is mainly to resolve properly when the star goes off the main sequence
>>    delta_HR_limit = 0.0025d0
>>    delta_lgL_He_limit = 0.1
>> 
>> Thanks in advance for any help or guidance.
>> 
>> Best,
>> 
>> Trevor Dorn-Wallenstein
>> University of Washington
>> Astronomy Department
>> he/him
>> <18M_at_ZAMS.mod><inlist><example_hr.pdf>_______________________________________________
>> mesa-users at lists.mesastar.org
>> https://lists.mesastar.org/mailman/listinfo/mesa-users
>> 
>