Degenerate segment

[clau]

Hi,

I'm having problems running GYRE for some red giants.
MESA version-r7503.
One of the profiles is attached (+inlist). GYRE error is:

Code: Select all

gyre [6.0.1]
------------

OpenMP Threads   : 2
Input filename   : gyre.in

Model Init
----------

Reading from FGONG file
   File name profile43.data.FGONG
   File version 300
   Read 1423 points
 ASSERT 'COUNT(gr%pt%s == s) >= 2' failed at line 114 <gyre_grid:grid_t_x_>:
 Degenerate segment

I'd appreciate your suggestions.
Claudia.

[rhtownsend]

Dear Claudia --

Thanks for your question, and welcome to the forums. It appears that your FGONG model file has three points at the outer boundary with the same radial coordinate. This is what is causing the error.

In reality, it's unlikely that the numerical model (MESA, right?) has this issue. However, the FGONG file is in a limited-precision format that doesn't include enough significant figures in the radial coordinate data to adequately resolve the three points.

This issue was addressed in 2015 by extending the FGONG format to include more significant figures -- see

https://gyre.readthedocs.io/en/stable/_ ... format.pdf

...and support for the extended format was incorporated in MESA. Are you perhaps using a very old version of MESA, which doesn't have these changes?

Best wishes,

Rich

[warrick]

Claudia mentioned that she's running MESA r7503, which has the old control and default

Code: Select all

      format_for_FGONG_data = '(1P5E16.9,x)'

I think changing that to

Code: Select all

      format_for_FGONG_data = '(1P,5(X,E26.18E3))'

will adhere to the new standard but GYRE might get confused by ivers (the last integer on line 5 of an FGONG file) still being 300. If GYRE doesn't run, try changing ivers manually to 1300 in the FGONG file.

Newer versions of MESA respect the FGONG standard better and only allow the user to specify ivers as 300 or 1300, rather than allowing the user to choose an arbitrary float format.

[clau]

Thank you both for your help. Changing the FGONG format and ivers solved the problem.
In a few cases though I still saw two radial coordinates with the exact same value, so I manually added a really small number to the first column of line 9 to avoid the error.
Regards,
Claudia.

[rhtownsend]

Not that it is allowable for two points to have the same radial coordinate -- just not three or more!

When two points have the same radial coordinate, they are treated as a "double point" where there can be a discontinuity in variables such as density. GYRE handles such discontinuities by using appropriate jump conditions in the pulsation equations.

So, I recommend not editing pairs of points to make their radial coordinates different!

[clau]

I looked at my FGONGs more closely and there was a second pair of points with same radial coordinate.
Taking Rich's advice to avoid editing the points manually, I tried saving the profiles in MESA/GYRE format instead of FGONG and this has worked well so far.
Claudia.