Limiting convective velocities with max_conv_vel_div_csound

[JRGarza]

Hi,
I noted the control max_conv_vel_div_csound is deffined in the documentation (for 23.05.1 and later versions), as well in ctrls_io, controls.defaults and star_controls but it is not used in the code. I notice this after trying to use the control to limit convection velocities below sound speed but did not see any difference, I am using Mihalas MLT option. Is there a reason for this, or it is just a remanent code from previous MESA versions?

[Debraheem]

Thanks for opening this issue. Hmm, It looks like a remanent deprecated control, but it should be functional in theory, although I don't see it being used internally anywhere. It appears to be a topic of an open issue that was never addressed, so as of now there is no conv_vel limiter.

see #373

perhaps, we can revive it... but is it necessary?

For now, this control should either be removed, or put to use.

[Debraheem]

We are having git-lfs bandwidth issues at the moment, but when they are resolved, I can push a branch with a conv_vel limiter.

[Debraheem]

Perhaps you could git checkout this branch https://github.com/MESAHub/mesa/tree/EbF/add_Cs_limit_to_TDC? Let me know if you have any thoughts. I'm not immediately sure how to limit the conv_vel in mlt self-consistently like we do in TDC. The hack might have to cut it.

[JRGarza]

Thanks for the answer. For now I required the option for the 'Mihalas' option (as in my science case I have issues using TDC, but thanks for referring to the add_Cs_limit_to_TDC branch). I made a PR #781 for this control in case it is useful. As for the justification, the MLT prescriptions (at least Cox and Mihalas) are not valid for supersonic convection velocities. The argument in Cox and Giuli's book (chapter 14) states that turbulent pressure by non-supersonic convection is negligible and not taken into account for the derivation, contrary to the case of supersonic convection.

[Debraheem]

Check out the the add_Cs_limit_to_TDC again. I added an option for mlt too. It seems similar to what you added?

On the opacity floor, that's a good idea, I've crashed many expanding models becauase the surface hits the opacity table boundary...

[Debraheem]

I think the only difference is that you are using the actual sound speed at the cell face, where as I just used an estimate. I can correct that.

[JRGarza]

add_Cs_limit_to_TDC

I just saw the similarities between both. Sorry I thought that such branch was only usable for TDC. And it seems that the main difference between both is the csound considered for the limit as you mention.

[Debraheem]

Yes, no worries! Happy to see we thought along the same lines. When I merge your branch into the add_Cs_limit_to_TDC branch, I'll squash the differences and pick one of our approaches.