[mesa-users] Time-step dependent behavior on stars with high mass loss

[Pablo Marchant]

Hi!

While running some comparisons between MESA and the BEC code used here at
Bonn, I found that stars seem to react differently between the codes to
high mass loss rates that drives them out of thermal equilibrium.

To see this, I setted up a simple experiment, where I would let a 20 solar
mass star with solar metallicity evolve on the main sequence and at some
point initiate mass loss at a rate of 10^{-3} solar masses per year. To do
this, I use the attached inlists, "inlist_start" to save a particular main
sequence model, and "inlist_load" to load it specifying the timestep I want
to use and set the mass loss rate.

I also repeated the experiment using BEC, ensuring that at the point where
mass loss starts the two stars are at the same point of their evolution in
the main sequence.

What I see with MESA (image "mass_loss.eps" attached), is that there are
important differences on the evolution of the radius of the star depending
on the timestep used during the mass loss phase, with the star shrinking
more as one uses smaller timesteps. There is no difference though with
different choices of mesh_delta_coeff. Using BEC, the radius of the star is
consistently smaller than any of my MESA runs, and I only see a small
difference when changing the timestep.

As the resolution does not seem to play an important role I chose to run
many models using mesh_delta_coeff=2.0, and using smaller timesteps to see
if things ever converge. This is what I show on the file "mass_loss2.eps",
and the models don't appear to be converging. In fact, the one with the
smaller timestep at one point starts to expand slowly and its radius
digresses significantly from the others.

I have not been able to find what causes this discrepancy, so if anyone
could provide me with an additional clue on what is going on or where  can
I look in mesa, I'd be grateful.

Cheers!

P.D. I also attempted the same thing with a significantly smaller mass loss
rate, 10^(-5) solar masses for year, for which the star should be able to
adjust and remain in thermal equilibrium, and I see no strange
timestep dependence.

-- 
Pablo Marchant Campos
M.Sc on Astrophysics, Universidad Católica de Chile
PhD student, Argelander-Institut für Astronomie
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.mesastar.org/pipermail/mesa-users/attachments/20130319/5a14dac5/attachment.html>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: inlist_start
Type: application/octet-stream
Size: 10705 bytes
Desc: not available
URL: <https://lists.mesastar.org/pipermail/mesa-users/attachments/20130319/5a14dac5/attachment.obj>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: inlist_load
Type: application/octet-stream
Size: 10684 bytes
Desc: not available
URL: <https://lists.mesastar.org/pipermail/mesa-users/attachments/20130319/5a14dac5/attachment-0001.obj>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: mass_loss.eps
Type: application/postscript
Size: 43860 bytes
Desc: not available
URL: <https://lists.mesastar.org/pipermail/mesa-users/attachments/20130319/5a14dac5/attachment.eps>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: mass_loss2.eps
Type: application/postscript
Size: 100305 bytes
Desc: not available
URL: <https://lists.mesastar.org/pipermail/mesa-users/attachments/20130319/5a14dac5/attachment-0001.eps>