Re: [mesa-users] Yet about Brunt calculation ...

[Bill P. <pa...@ki...>]

Hi Ehsan,

Thanks for your email and your thoughts on how we might make progress getting mesa to be able to support asteroseismology.  As you know, that's a high priority and lots of people are helping, yourself included.  

At this point it is too early to predict where the problems lie or how they will be solved, and we welcome input and participation in figuring things out.  Happily mesa is modular!  So you might try writing a routine to call the eos to calculate the brunt the way you suggest.  You can get all the things you need in the profile output from star to be able to calculate a brunt profile using a different scheme and compare it what we're doing now.  That would be very interesting to see.

I'm confident we'll get there -- but there's still some hard work to be done.

Best wishes,
Bill




On Jul 9, 2011, at 11:35 AM, Ehsan Moravveji wrote:

> Dear Bill and mesa friends,
> Hi.
> I'm totally aware that implimentation of Brunt-Vaisala is challenging for you, and you made a lot of efforts to implement Brassard et al. receipe into MESA. Great ...
> 
> But, 
> I read the first 4 pages of the paper. They clearly state in a paragraph above Eq. 14 that "We thus may use Y ... as a unique indicator of composition change" and indeed they are true for a model of white dwarf, which is quite simple star. One can even model a white dwarf with a composite polytrope of varying n and still succeed to explain many phenomena. But, in my opinion this cannot be the case for evolved stars off the main sequence.
> 
> Here, I have attached a plot of the abundance profile variations within a model of Rigel. We are all familiar with similar plots.
> 
> 1- The H profile variation could be quite profound outside the core.
> 2- For other heavier species, their abundances are small, but the differential of their logarithm may still contribute appreciably to the last term in Eq. 13 of Brassard et al. at least for the first few abundant species in normal stars.
> 3- caling multiple times to EOS could be demanding and CPU expensive, but in my opinion, a better evaluation of N2 is by considering Eq. 13 and letting the sum over just few species, say those that are followed by the rates.
> 4- I admit that I might be totally wrong, but need to be justified.
> 5- I have no idea how difficult it might be to implement numerically the more "standard" (in the sense of literature) N^2 using \nabla_\mu rather than B term in Brassard et al. 
> Physically, \nabla_\mu is very meaningful; it carries the weighted presence of all species in each layer, plus electrons! Not only He.
> 6- Unfortunately, I'm not familiar with the Lars formula called "simple". But, what I intend to do is to scape from free parameters once a quantity is free from this, and N^2 is. I'm really annoyed when I have to decide upon brunt_H_frac and test the effects of this keeping all parameters fixed, and repeating and repeating. And even after that, being yet worried of it's potential effects with the changes in the model. Of course, this is my problem and comes from my lack of knowledge.
> 
> N^2 has a very critical role in asteroseismology, and that's indeed why it is worth to discuss. Specially, it is an input to pulsation codes, and when N^2 profile has some spikes, it makes the solution of pulsation equations tough.
> 
> With kind regards.
> 
> Moravveji, Ehsan.