# [Mesa-users] Relaxing a variables value between Newton-Raphson solver iterations

Francis Timmes fxt44 at mac.com
Wed Aug 31 13:47:52 UTC 2022

```hi arman,

> ... may help the convergence of my star when the outer envelope is rapidly expanding.

this is not specific enough, but mesa has many options for mass loss and gain

https://docs.mesastar.org/en/release-r22.05.1/reference/controls.html#mass-gain-or-loss

and atmosphere boundary conditions

https://docs.mesastar.org/en/release-r22.05.1/reference/controls.html#atmosphere-boundary-conditions

maybe explore some plots on how your specific model is evolving and why?

fxt

> On Aug 30, 2022, at 1:22 AM, Arman Aryaeipour via Mesa-users <mesa-users at lists.mesastar.org> wrote:
>
> Dear MESA-users,
>
> Is there a way to relax the value of a variable for a cell, like velocity for cell 1, from an initial value to a final value between a set number of the Newton-Raphson solver iterations for every timestep?
>
> For example, is there a way to set v(1)=10^8 at solver_iter=1 and allow it to relax to v(1)=5x10^7 at solver_iter=5? in this scenario, I would set the minimum number of solver iterations to 10 so that the value of v(1) can converge on the correct final solution between solver iteration 6 to 10, and I would also do this for every timestep. I ask as I think this may help the convergence of my star when the outer envelope is rapidly expanding. I used the velocity variable in this example, but I am interested to know if there is a way to do this for any variable.
>
> I'd be grateful for any suggestions, and thank you in advance!
>
> Best regards,
> Arman Aryaeipour
> _______________________________________________
> mesa-users at lists.mesastar.org
> https://lists.mesastar.org/mailman/listinfo/mesa-users

```