[mesa-users] accretion

[Bill Paxton]

Hi Dave,

On Oct 2, 2013, at 12:31 PM, David Arnett wrote:

> Historically we have used accretion as a way to change mass slowly, often without worrying about the entropy of the accreted matter or more rapid time dependence. That is actually a complex multidimensional radiation-hydrodynamics problem itself, as Dean knows perhaps too well ;-)

wow!  "complex multidimensional radiation-hydrodynamics" -- it makes my head spin just to read the words.  I'll be happy to leave that to you and Dean and friends!
> 
> 
> I have a question: ...  what MESA is trying to do here?

The new scheme I just introduced in the last release is using Lagrangian 1st order difference to approximate the time derivatives that show up in the calculation of eps_grav.  So it is nothing new --- that's how we treat all of the time derivatives.  The only thing new is that we are using the same scheme now in material accreted during the timestep as in material that was already part of the star at the start of the timestep, namely lagrangian Df/Dt = (f_final - f_initial)/(time_final - time_final) with the f's evaluated at the same mass coordinate.

So it is consistent and I believe it makes sense as a 1st order solution --- but of course the need for accuracy adequate for a particular application may require a 1st order scheme to take painfully small timesteps.    I don't need to tell you about the dangers of letting an implicit code take such large timesteps that it glosses over important aspects of the evolution!

My last email was to suggest it might be possible to use other more accurate side-computations to replace the standard eps_grav calculations in material near the surface when doing accretion.  Perhaps we can get the accuracy where we need it without forcing the entire calculation to go to small timesteps.

As you know, I'm trying to ensure that mesa/star is flexible enough that it will allow this sort of experimentation and innovation by anyone who would like to dig in and try.  I hope Dean will take up the challenge on this one --- he'd be a great one for it. ;D

Cheers,
Bill