[Mesa-users] numerical noise in Brunt-Vaisala -- eos issue

[Alina Istrate]

Dear all,

I am trying to understand what looks like a numerical noise in the Brunt
-Vaisala frequency (attached figure[ brunt_N2.png]). The peak is due to one
cell (unless I use *num_cells_for_smooth_brunt_B, *which will then move it
across a few cells).
The peak comes from the brunt_B term (attached figure[brunt_B.png]) and
there is no meaningful composition gradient at that cell to reproduce such
a jump (attached figure[composition.png]).

I am using version 12778 and the system I am looking at is a WD (0.37 Msun,
CO-core, on the cooling track).  I just want to note here that this type of
noisy behaviour occurred also in a more massive CO WD.

The brunt_B term is computed as *s% brunt_B(k) = (lnP1 -
lnP2)/delta_lnP/chiT_face*. Investigating all these terms separately, I
narrow it down to the difference* (lnP1 - lnP2). **lnP1* and *lnP2* are
computed by calling the EOS using the temperature and density at the face
of cell k, but evaluated with the composition of cells k and k-1
respectively.
Having the suspicion on eos discontinuity, I made a small experiment. I
interpolate the composition vector between these two cells and get the
*res(i_lnPgas) *from a get_eos call (with T and rho set to the values at
face k) .
 The result is attached (figure[check_eos.png]) and indeed shows a jump. I
also attach the folder [eos_sample.zip]  with the files used for the eos
sampling, but if needed I can also provide the files to reproduce the
behaviour on the WD cooling track.

 Does anyone have a suggestion on how to get rid of this type of noise ?

Thanks a lot in advance,
Alina

-- 
*--------------------------------------------------------------------*
Dr. Alina Istrate
Institute for Mathematics, Astrophysics & Particle Physics
Department of Astrophysics
Radboud University
P.O. Box 9010, MC 62
NL-6500 GL Nijmegen, The Netherlands
r: HG02.738
a.istrate at astro.ru.nl
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.mesastar.org/pipermail/mesa-users/attachments/20200715/f9d28a7f/attachment-0001.htm>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: brunt_n2.png
Type: image/png
Size: 90171 bytes
Desc: not available
URL: <https://lists.mesastar.org/pipermail/mesa-users/attachments/20200715/f9d28a7f/attachment-0004.png>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: brunt_B.png
Type: image/png
Size: 50713 bytes
Desc: not available
URL: <https://lists.mesastar.org/pipermail/mesa-users/attachments/20200715/f9d28a7f/attachment-0005.png>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: composition.png
Type: image/png
Size: 118021 bytes
Desc: not available
URL: <https://lists.mesastar.org/pipermail/mesa-users/attachments/20200715/f9d28a7f/attachment-0006.png>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: check_eos.png
Type: image/png
Size: 62960 bytes
Desc: not available
URL: <https://lists.mesastar.org/pipermail/mesa-users/attachments/20200715/f9d28a7f/attachment-0007.png>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: eos_sample.zip
Type: application/zip
Size: 486082 bytes
Desc: not available
URL: <https://lists.mesastar.org/pipermail/mesa-users/attachments/20200715/f9d28a7f/attachment-0001.zip>