[mesa-users] Calculating oscillation modes of higher degree

[Wiebke Herzberg]

Hi everyone,

I am having trouble calculating oscillation modes of higher degree (l >= 
8) for subgiant models created with MESA.

I first tried this some months ago using GYRE for the mode calculation, 
but for higher l values I noticed inconsistencies in the radial mode 
order n and the corresponding eigenfunctions appeared strangely shifted. 
I went ahead and reported the problem on the GYRE forums 
(http://www.astro.wisc.edu/~townsend/gyre-forums/viewtopic.php?f=3&t=88). Rich 
suggested, that this problem was likely caused by the model not being in 
complete hydrostatic equilibrium as assumed by the pulsation codes and 
his advice was to try and increase the grid points of the model.

When I recently picked up this work again, I tried all kinds of (more or 
less basic) things, such as:

- increasing the mesh up to mesh_delta_coeff = 0.1
- smaller time steps, down to max_years_for_timestep = 1d7
- turning diffusion on/off
- starting with a pre-MS model instead of loading a ZAMS model
- different convection/overshooting settings

but nothing seemed to help. Of course, I am not an experienced modeler, 
so maybe I just did not try the "right" thing, since I could not really 
localize the problem?

I also used the minilab provided by Pieter Degroote 
(http://www.ster.kuleuven.be/~pieterd/mesa/run_star_extras.html) to have 
a look at the hydrostatic equilibrium in my models. And while there are 
some deviations from the equilibrium at the surface, the bottom of the 
convection zone and the core of the model, they are generally small, of 
the order of 1e-9 in the core region up to 1e-6 at the surface.

Additionally, I tried adipls for the mode calculation and encountered 
the same (I assume?) problem there, although it showed itself 
differently. For modes with l>=8 inconsistencies in the radial mode 
order n appeared and when looking at the corresponding eigenfunctions 
they showed a small region close to the core where the function was no 
longer smooth, but looked noisy/jumpy. A plot showing this behavior for 
three example eigenfunctions is attached. (Note: This problem only 
occurred for the p-dominated mixed modes whose eigenfunctions have small 
amplitudes in the core region.)

Has anyone experienced the same difficulties? Is it indeed a problem 
with the hydrostatic equilibrium that hampers the mode calculation? And 
if so, is there a remedy or maybe a workaround for it? I would be 
interested in calculating modes up to about l=20 for my MESA models. Any 
comments or suggestions are greatly appreciated.

Wiebke


PS: I also attached the inlist I used recently, just in case.

-- 
Wiebke Herzberg               Kiepenheuer-Institut für Sonnenphysik
Phone: +49 761 3198-232       Schöneckstraße 6
Fax:   +49 761 3198-211       79104 Freiburg, Germany
wiebke at kis.uni-freiburg.de    http://www.kis.uni-freiburg.de/
-------------- next part --------------
A non-text attachment was scrubbed...
Name: eigfun.pdf
Type: application/pdf
Size: 24653 bytes
Desc: not available
URL: <https://lists.mesastar.org/pipermail/mesa-users/attachments/20141125/590a53d2/attachment.pdf>
-------------- next part --------------

! inlist_subgiant

! contains the parameters controlling this job

! for the sake of future readers of this (yourself included),
! ONLY include the controls you are actually using.
! DO NOT include all of the other controls that
! simply have their default values.



&star_job

! Leave the plot windows open when run is finished
      pause_before_terminate = .true.

! control output:
      profile_columns_file = 'profile_columns.list'
      
! pre main sequence:
      !create_pre_main_sequence_model = .true.

! nuclear reactions:
      change_net = .true.
      new_net_name = 'o18_to_mg26.net'

! eos:
      eos_file_prefix = 'mesa'

! opacities:
      kappa_file_prefix = 'gs98'

! choice of variables:
      change_lnPgas_flag = .true.
      new_lnPgas_flag = .true.

! Plots:
      pgstar_flag = .true.

/ !end of star_job

&controls

  ! starting specifications
      initial_mass = 1.25 ! in Msun units
      initial_Z = 0.02
      initial_Y = 0.28

  ! output to files and terminal
      calculate_Brunt_N2 = .true.
      max_num_profile_models = 500
      photostep = 50
      profile_interval = 10
      history_interval = 1
      terminal_interval = 10
      write_header_frequency = 10

  ! time step / accurracy of calculation
      !varcontrol_target = 1d-3       !default is 1d-4, do only change if you want quick tests
      !max_years_for_timestep = 1d7

  ! improve grid sampling
      max_allowed_nz = 15000     !maximum number of grid points allowed
      mesh_delta_coeff = 0.3     !default is 1, lower values give more grid points

  ! when to stop
  
      max_age = 4.6d09                !in years, typical subgiant ages are 3.5-7.0 Gyr

      !log_center_temp_limit = 7.33131
         ! stop when log10 of the center temperature exceeds this limit.
      
      Teff_lower_limit = 5.768d3
         ! stop when Teff is less than this limit.
         
  
  ! atmosphere boundary conditions
      which_atm_option = 'grey_and_kap'  ! default is 'simple_photosphere'

  ! convection
      mixing_length_alpha = 1.8
      !use_Ledoux_criterion = .true.
      !alpha_semiconvection = 0.01
      !min_overshoot_q = 0.05          
 
  ! overshooting, (use the exponential setting)
  !    overshoot_f_above_nonburn = 0.014
  !    overshoot_f_below_nonburn = 0.014
  !    overshoot_f_above_burn_h = 0.014
  !    overshoot_f_below_burn_h = 0.014
  !    overshoot_f_above_burn_he = 0.014
  !    overshoot_f_below_burn_he = 0.014
  !    overshoot_f_above_burn_z = 0.014
  !    overshoot_f_below_burn_z = 0.014
  

  ! asteroseismology
      write_pulse_info_with_profile = .true.
      !pulse_info_format = 'GYRE'
      pulse_info_format = 'FGONG'
      add_center_point_to_pulse_info = .true.

/ !end of controls