[mesa-users] New eps_grav Test

[Broxton Miles]

Hi Bill,

Ah thanks, that makes sense.

Broxton


On Mon, Oct 21, 2013 at 12:35 PM, Bill Paxton <paxton at kitp.ucsb.edu> wrote:

>
> On Oct 21, 2013, at 9:58 AM, Broxton Miles wrote:
>
> The question I have is why does the time step size seem to bounce around?
> I may be missing something pretty simple, but it's not really clear to me
> why that should be happening.
>
> Anyhow, I hope this has been informative, and any insights, comments, or
> questions is more than welcome.
>
>
> Hi Broxton,
>
> Your new results look very encouraging!  Thanks for sending them.
>
> The bouncing in the timestep is because the code doesn't know how to
> adjust the next timestep in order to hit the exact value you have requested
> for CpT_absMdot_div_L_limit.  It just uses a simple linear adjustment --
> dt_next = dt * desired_ratio / actual_ratio.  This will of course lead to
> bouncing around the desired result.  If it becomes a priority, I could
> reduce the bouncing by introducing a 2nd order low-pass filter similar to
> what I already have for the varcontrol_target control.  (see hydro_timestep
> in star/private/timestep.f -- H211b "low pass" controller from Soderlind &
> Wang).
>
> Cheers,
> Bill
>
>
>
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.mesastar.org/pipermail/mesa-users/attachments/20131021/0d9748cf/attachment.html>