[mesa-users] crazy C shell burning

[Francis Timmes]

hi tuguldur,

> ... due to lack of NSE treatment you're likely to get into trouble during Si burning. 

i take issue with this as a general statement. 
there are certainly other stellar evolution codes that do not 
invoke qse/nse and manage just fine. qse/nse is a convenience 
used to decrease the overall computational effort, but it is not 
a fundamental requirement for "proper" behavior.

one interesting challenge with qse/nse is to assure that a network 
calculation and a qse/nse calculation give the same abundances 
when they should. in general, these two calculations will not give the 
same abundances at the ~10% level largely because of the differences in
coulomb/screening corrections applied (see the appendix in 
calder et al apj 2007 656 313 for a treatment). of course, abrupt ~10% 
jumps in the composition can cause convergence issues during the 
newton iterations of a timestep, assuming an implicit integration.

this all said, qse/nse is a planned capability addition in mesa.

fxt








On Apr 16, 2013, at 10:24 PM, Tuguldur Sukhbold <sukhbold at ucolick.org> wrote:

> Dear MESA users,
> 
> many of us are now well aware that due to lack of NSE treatment you're likely to get into trouble during Si burning. But in my 2+ year experience with MESA I have seen many times this weird behavior well before Si ignition, during the C burning. The attached plot shows the central T and rho evolution during C burning for two identical 15Msun models with different alpha_mlt. You see the alpha_mlt=2 case goes smoothly, but the alpha_mlt=1 case is doing some crazy stuff. I could wait for thousands of time steps and it would be just doing its weird loops back and forth, and never manage to ignite Ne.
> 
> With older MESA versions I noticed sometimes even very slight changes in the composition could result into this problem during C burning. If I remember correctly, during the last summer school Josh Shiode also reported similar results when he played with the time step and zoning (?) parameters.
> 
> I was wondering if anybody had looked into details of why this happens? how to avoid? previously I used to almost 'randomly' play with certain parameters until I find that sweet spot to make it through C burning. But now I am trying to make a detailed comparison with another code, so i need to be a bit careful.
> 
> thanks,
> Tuguldur
> 
> ***just to clarify, in this plot the alpha_mlt=1 case only shows data since the ignition of C in the core. While the other case shows data from previous evolution.
> 
> <Screen Shot 2013-04-16 at 9.51.50 PM.png>
> ------------------------------------------------------------------------------
> Precog is a next-generation analytics platform capable of advanced
> analytics on semi-structured data. The platform includes APIs for building
> apps and a phenomenal toolset for data science. Developers can use
> our toolset for easy data analysis & visualization. Get a free account!
> http://www2.precog.com/precogplatform/slashdotnewsletter_______________________________________________
> mesa-users mailing list
> mesa-users at lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/mesa-users