[mesa-users] Radiative-Convective boundary discrepancy

[Bill Paxton]

Hi,

The counter-intuitive degradation of the Brunt with increasing number of grid points may be a result of the dependence of the brunt "B" term on a numerical approximation to a derivative (see eqn 8 in 2nd mesa instrument paper).  As the cells get smaller the numbers being subtracted get smaller and numerical problems get larger.  All very ugly.

Since you are looking at a solar model, you might consider turning on overshooting at the base of the convection zone if you haven't already done that.  It might smooth out the composition gradients enough to help the Brunt.

-B


On Jun 9, 2014, at 3:10 PM, Momchil Molnar wrote:

> Dear All, 
> 
> I am writing with the following concern - after encountering an unexpected behaviour of the Brunt frequency around the radiative-convective boundary for a Solar model and sending an email to the list, I continued looking through my results and found that the Brunt becomes less and less well defined in this region as spatial resolution increases (i.e. increase in number of grid points). 
> 
> I am attaching plots of the Brunt frequency squared for different number of grid points for a solar model, evolved from the integrated one in mesa/star/test_suite/solar_calibration/solar_calibration_input.mod, using the same inlist differing only in the mesh_delta_coeff (accordingly 1 for 1500 model; 0.15 for ~10000 pts model and 0.08 for ~20000 pts model). 
> 
> The previous reply, which I received on the topic was leading in a direction, which suggests that the boundary layer is very thin and using higher resolution models will constrain the inconsistency of the Brunt frequency only in a near-by area around the transition region. However, as we see, there is an unexpected bump inside of the radiation zone for the models with higher spatial resolution, further away from the radiative-convective boundary.
> 
> Do you think this can be avoided by using different parameters for describing the convection mechanism or this might be an implementation problem in MESA? 
> 
> Cheers, 
> Momchil Molnar 
> MIT 
> <diff_num_grid_pts_1.pdf><diff_num_grid_pts_2.pdf>------------------------------------------------------------------------------
> HPCC Systems Open Source Big Data Platform from LexisNexis Risk Solutions
> Find What Matters Most in Your Big Data with HPCC Systems
> Open Source. Fast. Scalable. Simple. Ideal for Dirty Data.
> Leverages Graph Analysis for Fast Processing & Easy Data Exploration
> http://p.sf.net/sfu/hpccsystems_______________________________________________
> mesa-users mailing list
> mesa-users at lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/mesa-users

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.mesastar.org/pipermail/mesa-users/attachments/20140609/6a71a526/attachment.html>