Re: [mesa-users] diffusion causing model to get stuck

[Evan B. <eb...@ph...>]

Hi Jean,

Another control you might find useful is

         !### diffusion_v_max

         ! Max velocity (cm/sec).
         ! We can get extremely large velocities in the extreme outer envelope
         ! that cause problems numerically without really effecting the results,
         ! so we allow a max for the velocities that should help the numerics
         ! without changing the results.
         ! Note: change `diffusion_v_max` to at least 1d-2 when using radiative levitation.

      diffusion_v_max = 1d-5

Setting this to a smaller number can help models that are having trouble with fast moving material in the outer atmosphere. In your model, I suspect the stuff that’s causing problems at the surface will eventually sediment away from where it’s moving so fast, and there will be nothing left that requires enforcing an artificial velocity cap, so you shouldn’t hurt the overall physics of diffusion much.

Another option is just to force the model to take smaller time steps for a while until things sediment away from the very surface layers. When all else fails, I find it is often useful to set max_years_for_timestep to something rather small to keep diffusion well behaved for a while. Then I turn that control off and restart after the fast sedimentation at the surface has taken place.

Cheers,
Evan


> On Jun 5, 2017, at 12:52 PM, Aaron Dotter <aar...@gm...> wrote:
> 
> Hi Jean,
> 
> I recommend trying the following control;
> 
>          ! treat at least this much at surface as a single cell for purposes of 
> 
>       diffusion_min_dq_at_surface = 1d-9
> 
> 1d-9 is the default value.  Try setting it to a larger number, say 10^-4 or 10^-3.  This tells the diffusion module to use treat the outer fraction of the star by mass as one zone in the diffuison calculation.  That calculation can get messy near the stellar surface, especially in stars in around ~1.5 Msun (as yours is) because they have a very thin -- or no -- surface convection.
> 
> Good luck!
> 
> Aaron
> 
> 
> 
> 
> 
> 
> On Mon, Jun 5, 2017 at 3:42 PM, Jean McKeever <je...@nm... <mailto:je...@nm...>> wrote:
> Hello,
> 
> I was trying to create a grid of models using the astero module and the grid got stuck on a particular point. After some testing of turning things on and off I have discovered that do_element_diffusion appears to be the culprit. For a 1.4 solar mass star with a metallicity of -.4, the model seems to get caught in a series of retries where dt keeps decreasing to be prohibitively small. This happens very early on in the model. If I turn off diffusion, the model proceeds normally. 
> 
> Can anyone provide insight as to why diffusion is causing the model to get stuck? I have attached some modified inlists and starting model to help reproduce this. I checked the newest version of mesa and it is still a problem in 9793.
> 
> Thank you,
> Jean
> 
> ------------------------------------------------------------------------------
> Check out the vibrant tech community on one of the world's most
> engaging tech sites, Slashdot.org! http://sdm.link/slashdot <http://sdm.link/slashdot>
> _______________________________________________
> mesa-users mailing list
> mes...@li... <mailto:mes...@li...>
> https://lists.sourceforge.net/lists/listinfo/mesa-users <https://lists.sourceforge.net/lists/listinfo/mesa-users>
> 
> 
> ------------------------------------------------------------------------------
> Check out the vibrant tech community on one of the world's most
> engaging tech sites, Slashdot.org! http://sdm.link/slashdot_______________________________________________
> mesa-users mailing list
> mes...@li...
> https://lists.sourceforge.net/lists/listinfo/mesa-users