[mesa-users] density and sound speed profiles with "ripples" in solar calibration example

Francis Timmes fxt44 at mac.com
Thu Aug 28 13:24:07 EDT 2014


hi ana,

another resource for learning how to use run_star_extras is 
kevin moore’s excellent summer school material:

http://mesastar.org/teaching-materials/2014-mesa-summer-school-working-dir/moore

fxt



On Aug 28, 2014, at 9:57 AM, Bill Paxton <paxton at kitp.ucsb.edu> wrote:

> Hi Ana,
> 
> Good progress.   Your new plots seem to show the "ripples" gone from the mesa results, but there is still an offset from your cesam results.  Perhaps this is not a big deal since we also see a fairly large difference in mesa vs your cesam results compared to the solar sound speed data.
> 
> The sound speed data used for comparison gives values at points, so radius and sound speed come from the same location in the model.  So the comparison data from mesa needs to come from same locations as well -- but that means an interpolation is needed since radius is defined at cell  boundaries and sound speed at cell centers.  When you correctly do the interpolation, the results are very good; if you don't do the interpolation you are taking values from slightly different locations in the model.  The magnitude of the offset depends on the size of the cell and there is variation in the cell sizes, so it doesn't act as a simple uniform offset. instead it acts as a noise source and produces poor results, including "ripples".
> 
> It seems your next step is to learn more about using run_star_extras.  I suggest you start here: 
> http://mesa.sourceforge.net/run_star_extras.html
> 
> There are lots of folks on mesa-users who will be able to assist you.
> 
> Good luck,
> Bill
> 
> 
> 
> On Aug 28, 2014, at 6:11 AM, Ana Olímpia wrote:
> 
>> Hi Bill,
>> 
>> 
>> Quick comment: since r is defined at cell boundaries, while csound and density are defined at cell centers, you can generate visual ripples in a plot by showing values at cell center offset to cell boundary.  To get around this, we need to either use the radius at the cell center (called rmid) for the x axis, or switch to plotting center values that have been interpolated to the cell face (e.g., csound_at_face instead of csound).
>> 
>> Have you done that in the plots you sent?  If not, that would be 1st thing to try. 
>> 
>> The first thing I did was recompute the acoustic potential with all the values at cell boundaries: rmid, csound and logRho. It worked fine. The "ripples" in density and sound speed have disappeared completely. I attached a file with the all the new figures.
>>  
>> For your calculations, you might want to do everything at cell boundaries -- so for values defined at cell centers, you'll need to interpolate.  There is a routine to make that easy in star/public/star_lib: it is called star_interp_val_to_pt.  It does a monotonic cubic interpolation.
>> 
>> Yes, I would like to do everything at cell boundaries also. But here, I could not finish, I have a problem. Since the radius (or logR) and csound_at_face I can take directly from profile columns I just need to interpolate the density. I'm not calling the routine star_interp_val_to_pt properly. I read this: http://mesa.sourceforge.net/run_star_extras.html and tried to adapt to this case. Among other things I copied the routine directly to the run_star_extras.f file and/or added a line star_interp_val_to_pt = .true. to the inlist_solar, etc...Nothing works. I was also wondering if to use the routine star_interp_val_to_pt I have to add new columns to the profile output or it it will just return the interpolated values of density (temperture and others) at boundaries. This is for sure very easy to do but I am a bit lost...
>> 
>> In the very first part of the link above, in the section dedicated to activate the run_star_extras.f file, I noticed that I don't have a line like this: include 'standard_run_star_extras.inc' inside the file run_star_extras.f. It looks like the file (standard_run_star_extras.inc) has been already copied to the run_star_extras.f
>> 
>> When we do the comparison of the star model sound speed profile to the data, we use csound_at_face rather than csound --- it makes a huge difference.  
>> 
>> Why is that difference so huge? And why the values at cell boundaries produce a "better" comparison than the values at cell centers?
>> 
>> 
>> Regards and thank you very much in advance!
>> 
>> Ana
>> 
>> 
>> 
>> 
>> 
>>  
>> 
>> 
>> On Aug 20, 2014, at 5:13 PM, Ana Olímpia wrote:
>> 
>> > Hello all,
>> >
>> > I am trying to use the solar_calibration example in the test suite directory to compute the solar acoustic potential (formula 1 in the attached file) with mesa.
>> >
>> > To do this I just need the radius, the sound speed, and the density profiles which I took from the profile.data file of the final model in the LOGS directory.
>> >
>> > I computed this potential, several times before, with the CESAM code but I can't reproduce the results with mesa.
>> > I decided to compare, step by step, the process of computing the potential with MESA and with CESAM. A summary of this comparison is in the file AP_with_MESA.
>> > In the final plot of the acoustic potential for the Sun, we should be able to identify the superadiabatic region, the second Helium ionization and the base of the convective zone (figure 8 in the attached file AP_with_MESA)
>> >
>> > I suspect that those "ripples" in figure 2 and 3, respectively in the sound speed and density, are responsible for the final result of the potential with MESA. I tried to play with the mesh refinement parameters but I got always the same final result for the acoustic potential that can be seen in figure 8.
>> >
>> > Why am I getting those "ripples" in the density profile and in the sound speed profile?
>> >
>> > Thank you very much in advance!
>> >
>> > Ana
>> >
>> > PS1 - I use exactly the same process to compute the potential with mesa and with cesam. The difference is only in the input data: radius, sound speed and density.
>> >
>> > PS2 - Ubuntu 14.04, Mesa version 6794, Mesa SDK version 20140715
>> >
>> > <AP_with_MESA.pdf>------------------------------------------------------------------------------
>> > Slashdot TV.
>> > Video for Nerds.  Stuff that matters.
>> > http://tv.slashdot.org/_______________________________________________
>> > mesa-users mailing list
>> > mesa-users at lists.sourceforge.net
>> > https://lists.sourceforge.net/lists/listinfo/mesa-users
>> 
>> 
>> 
>> <AP_cell_centers.pdf>
> 
> ------------------------------------------------------------------------------
> Slashdot TV.  
> Video for Nerds.  Stuff that matters.
> http://tv.slashdot.org/_______________________________________________
> mesa-users mailing list
> mesa-users at lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/mesa-users





More information about the Mesa-users mailing list