[Mesa-users] Finding ZAMS

Sat Mar 16 09:10:46 EDT 2019

Hello,

I have thought a little bit more about stellar properties on the main  
sequence. Then I remembered that MESA uses varcontrol_target as a  
measure for the variation in the structure of the model and  
correspondingly chooses the next timestep. Therefore, MESA timesteps  
are a measure for how much the structure is changing. As we expect a  
somewhat 'constant structure' on the main sequence, these timesteps  
may work as indicator if the star is on the main sequence. I therefore  
calculated the change in timestep between models  
(timestep_differences) and plottet them against the star age. I used  
the maximum timestep_difference to normalize this differences to 1.  
Then I checked at which timestep these normalized timestep_differences  
reach a value bigger than 0.25 and attributed the step before with  
being ZAMS. I have attached corresponding plots in a zipped file.  
These are plots for 28 different masses that show log_L and log_Lnuc  
in the upper panel, the ratio Lnuc/L in the middle panel and these  
normalized timestep_differences in lower panel. The green line shows  
the determination of ZAMS according to LH/L ratio going back to 0.9995  
and the magenta line shows the ZAMS point according to  
timestep_differences. It seems to me that in the mass range in which  
the LH determination works, the timestep_difference gives the same  
result (maximum one or two models appart). Additionally, the  
timestep_difference method seems to work for any mass.

Now the way MESA calculates the timesteps is very complicated  
(star/private/timestep.f90). Do you think there is a way to translate  
this timestep_difference method to stellar properties?

Thank you again for your help!
Thomas

Zitat von Thomas Steindl <Thomas.Steindl at student.uibk.ac.at>:

> Hello,
>
> thanks to all of your input.
>
> First, it seems that the criterions of minimus radius and maximum   
> log_g work a little bit better, but again fail for specific masses.  
> I  have attached corresponding plots in a zip file.
>
> Well the properties of a star on the main sequence include as far as  
> I  know thermal and hydrostatical, constant burning of hydrogen in  
> the  star, constant total gravitational energy, etc. I have tried to  
>  incorporate some of this conditions, yet I failed. I also tried to   
> look at the abundances of the components of the cno- and pp-cycle   
> (C12, O16, ..). So far I am out of ideas but looking forward for any  
>  suggestions! :)
>
> Again, thank you all for your time!
> Thomas
>
>
> Zitat von Aaron Dotter <aaron.dotter at gmail.com>:
>
>> Hi Thomas,
>>
>> OK, now I understand!  Thanks.
>>
>> I still maintain that the ZAMS is hard to define precisely, especially in
>> terms of L_nuc/L or L_H/L, and so I would suggest something that is both
>> easy and straightforward to define and measure in your models.
>>
>> I have found that the minimum radius--or maximum log(g), signalling the end
>> of pre-main sequence contraction, is a useful proxy that is also easy to
>> locate.  You can use the same test on center H1 that I suggested yesterday
>> to define the domain over which you search for min(r) or max(log(g)).
>>
>> Aaron
>>
>> On Fri, Mar 15, 2019 at 1:03 PM Thomas Steindl <
>> Thomas.Steindl at student.uibk.ac.at> wrote:
>>
>>> Hi Aaron,
>>>
>>> I am sorry, I forgot that. I am trying to investigate the probing
>>> power of g-modes on the pre-MS and I would like to compare models that
>>> have the same percental age in respect to the model's age at ZAMS (
>>> age(model)/(age(ZAMS_model) ). Without a proper definition of ZAMS I
>>> am not able to be certain I am comparing the 'matching models'.
>>>
>>> Thomas
>>>
>>> Zitat von Aaron Dotter <aaron.dotter at gmail.com>:
>>>
>>>> Hi Thomas,
>>>>
>>>> In order for us to advise you, it would be helpful for you to answer my
>>>> first question:
>>>>
>>>> "The ZAMS is a hard thing to define precisely.  What does it mean for the
>>>> project you're working on?"
>>>>
>>>> Why do you need a precise ZAMS definition?  I think you're finding, as
>>>> others have before, that it is not something that is easily and
>>> succinctly
>>>> defined for the full range of stellar masses.
>>>>
>>>> Cheers,
>>>> Aaron
>>>>
>>>> On Fri, Mar 15, 2019 at 12:24 PM Thomas Steindl <
>>>> Thomas.Steindl at student.uibk.ac.at> wrote:
>>>>
>>>>> Hello,
>>>>>
>>>>> I have now tried your stopping condition. Therefore, I calculated 21
>>>>> models in a mass range of 0.5 to 6 solar masses (the 0.775 M_sun
>>>>> modelled would not converge). The stopping condition works well for
>>>>> the masses 1.05 - 5ish solar masses depending on where you define the
>>>>> cut off criterion. For the models with higher masses the ratio L_H/L
>>>>> never drops below 1 after it first rises above 1. I could avoid this
>>>>> problem by fitting a linear function in the points on the main
>>>>> sequence and determine the ZAMS by the first point which is below this
>>>>> linear fit. Unfortunately, this would not work for the 0.5 -1ish solar
>>>>> mass as the ratio converges from values above the main sequence value.
>>>>> What I could do is determing the ZAMS by finding the point which
>>>>> differs from the linear fit using the main sequence point by some
>>>>> value. Yet, I wonder if this stopping criterion would still be
>>>>> physical and not just some way to deal numerical issues?
>>>>>
>>>>> I have again attached a zipped folder of the plots that show what I
>>>>> refer to. Here i have plotted log_L and log_LH in the upper pannel and
>>>>> LH/L in ther lower pannel. Also the green line shows the determined
>>>>> ZAMS timestep which is the point on which the ratio drops below 0.9995
>>>>> when looking at the inverse time.
>>>>>
>>>>> Thank you in advance for any help!
>>>>>
>>>>> Thomas Steindl
>>>>>
>>>>>
>>>>> Zitat von Aaron Dotter <aaron.dotter at gmail.com>:
>>>>>
>>>>> > Hi Thomas,
>>>>> >
>>>>> > [Josiah beat me to it!]
>>>>> >
>>>>> > The ZAMS is a hard thing to define precisely.  What does it mean for
>>> the
>>>>> > project you're working on?
>>>>> >
>>>>> > You're seeing the different behavior because of the transition from
>>>>> > H-burning dominated by the pp chain vs. the CNO cycle.  The basic net
>>>>> > ignores several lesser species that equilibrate once the core gets hot
>>>>> > enough for the reactions to proceed.
>>>>> >
>>>>> > If it's acceptable for you to evolve slightly past the ZAMS and then
>>>>> > identify a previous timestep as the ZAMS, the following is the
>>> procedure
>>>>> I
>>>>> > use (in post-processing a MESA history file).
>>>>> >
>>>>> > Stop once the central H mass fraction has been reduced by 0.0015 from
>>> its
>>>>> > initial value.  Move backward from this point until you find the
>>> timestep
>>>>> > for which the model first L_H/L_tot crosses 0.99, or whatever you want
>>>>> your
>>>>> > ZAMS definition to be.
>>>>> >
>>>>> >
>>>>> > Cheers,
>>>>> > Aaron
>>>>> >
>>>>> >
>>>>> >
>>>>> >
>>>>> > On Thu, Mar 14, 2019 at 5:30 PM Thomas Steindl <
>>>>> > Thomas.Steindl at student.uibk.ac.at> wrote:
>>>>> >
>>>>> >> Hello,
>>>>> >>
>>>>> >> I am working on pre-MS Asteroseismology. Therefore, I am calculating
>>>>> >> MESA models which I use as input for GYRE. In my analysis, I need to
>>>>> >> know the timestep at which the star reaches ZAMS. Unfortunately, I
>>>>> >> fear the stopping conditions implemented in MESA do not work. To show
>>>>> >> this I have used three different inlists. A first, very simple inlist
>>>>> >> (Basic_inlist), an inlist with additional varcontrol
>>>>> >> (additional_varcontrol) and one with additional varcontrol and using
>>>>> >> the ppandcnoextras net instead of the basic net
>>>>> >> (additional_varcontrol_and_net). I have also attached the
>>>>> >> corresponding inlists. In the corresponding png files, I have plotted
>>>>> >> the values for log_L and log_Lnuc in the upper panel and the ratio
>>>>> >> L/Lnuc in the lower panel. As you can see, the ratio seems to
>>> converge
>>>>> >> to the value 1 for the first two inlists (for which this stopping
>>>>> >> criterion works). But for the latter of the three inlists, Lnuc
>>>>> >> exceeds L at some point in the evolution and the stopping criterion
>>>>> >> fails.
>>>>> >>
>>>>> >> The zip folder 'ZAMS_second_stop_option' contains plots for 30
>>> stellar
>>>>> >> models computed with the 'additional_varcontrol_and_net' inlist and
>>>>> >> different masses. In the 30 pngs in this folder I have plotted log_L
>>>>> >> and log_Lnuc in the upper pannel and
>>>>> >> abs(log(power_hburn)-log_surface_luminosity) in the lower panel. The
>>>>> >> title shows the mass of a specific model and the value of
>>>>> >> L_nuc_counter which states how many timesteps Lnuc/L has been between
>>>>> >> 0.99 and 1.01. As you can see, the value for the second stop option
>>>>> >> exhibits drops (mostly while log_Lnuc > log_L) which in terms makes
>>>>> >> this stop option unreliable and I don't see how it can function as
>>>>> >> global stop option for all masses.
>>>>> >>
>>>>> >> I have tried other options (gravitational energy, contraction speed,
>>>>> >> etc.), yet I cannot find a criterion. Do you have any idea on which
>>>>> >> stellar properties I could use? Any help would be greatly
>>> appreciated.
>>>>> >>
>>>>> >> Thank you in advance,
>>>>> >> Thomas Steindl
>>>>> >>
>>>>> >> _______________________________________________
>>>>> >> mesa-users at lists.mesastar.org
>>>>> >> https://lists.mesastar.org/mailman/listinfo/mesa-users
>>>>> >>
>>>>> >>
>>>>> >
>>>>>
>>>>>
>>>>>
>>>>
>>>
>>>
>>>
>>
>
>
>

-------------- next part --------------
A non-text attachment was scrubbed...
Name: ZAMS_timestep.zip
Type: application/x-zip-compressed
Size: 706055 bytes
Desc: not available
URL: <https://lists.mesastar.org/pipermail/mesa-users/attachments/20190316/6d2eaf6e/attachment.bin>