[Mesa-users] Query regarding implementing hydrodynamics

Wed Mar 20 15:35:30 UTC 2024

Hi Ebraheem,

Thanks a lot! This was really helpful indeed. I shall enter
dragon territory now, hopefully I'll survive :)

-A

On Tue, Mar 19, 2024 at 9:24 PM Ebraheem Farag <ekfarag at asu.edu> wrote:

> Hi Abha,
>
> Very Cool Science you're exploring!
>
> to answer your questions:
> 1: yes, those two controls are all you need. u_flag is off by default, so
> it shouldn't show up in your inlist unless you are going to turn v_flag
> off, and u_flag on to enter the HLLC Riemann hydro mode.
> 2: split_merge_amr is for u_flag only, so don't worry about it, MESA uses
> separate meshing functions for the finite difference solver, see
> mesh-adjustment
> <https://docs.mesastar.org/en/release-r24.03.1/reference/controls.html#mesh-adjustment>
> .
> 3: 10 years is fine, but if you do capture a shock, it might crash. artificial
> viscosity
> <https://docs.mesastar.org/en/release-r24.03.1/reference/controls.html#pvsc-zsh> can
> help prevent a shock from forming and maybe velocity drag
> <https://docs.mesastar.org/en/release-r24.03.1/reference/controls.html#drag-coefficient> can
> help damp dubious velocities, but beyond this, Riemann hydro is the path
> forward. There is no shame in trying to use it.
>
> -EbF
>
> On Tue, Mar 19, 2024 at 8:34 AM Abha Vishwakarma <
> abha.vishwakarma at niser.ac.in> wrote:
>
>> Dear Jared and Ebraheem,
>>
>> Thanks a lot for your reply. I had some misunderstanding about the v_flag
>> and u_flag parameters and this does seem to clarify it.
>>
>> To answer Jared's question, I am trying to model a common envelope
>> system, in which the companion is a neutron star with a 12 solar mass red
>> giant primary, and it is modelled by injecting extra_heat(as accretion
>> energy) into the star at a radius corresponding to the orbital evolution of
>> the system. To model the envelope expansion due to the accretion energy,
>> and capturing potential shocks, I want to switch on hydrodynamics. It seems
>> like I don't need to use u_flag for this purpose, as it is not an
>> explosive system.
>>
>> I had a few more questions regarding this:
>>
>>    1. So, just to confirm, for my purposes all I need to do to keep
>>    hydrodynamics on throughout the evolution is to set change_v_flag =
>>    .true. and new_v_flag = .true., and not do anything to u_flag
>>    parameters ?
>>    2. Do I need to set the use_split_merge_amr = .true. because I will
>>    be injecting extra_heat at specific zones which might need more refinement?
>>    3. My evolution will not be on the scales of stellar evolution, it
>>    should be around the order of 10 years or so, and I just wanted to know if
>>    this timescale is consistent with the typical timescales in which MESA
>>    simulations run with v_flag = .true. for red giant systems.
>>
>> I shall give more thought to the surface phenomena and boundary
>> conditions as Ebraheem has pointed out and come back to this if I have any
>> questions. Any help on the above listed questions will be very helpful!
>>
>> Best,
>> Abha
>>
>> On Tue, Mar 19, 2024 at 3:18 AM Jared Goldberg <
>> jgoldberg at flatironinstitute.org> wrote:
>>
>>> Dear Abha,
>>>
>>> What science case are you trying to do? MESA has 2 hydro solvers
>>> implemented, including the HLLC (u_flag) and the non-HLLC (v_flag) solver.
>>> The latter may be more stable for your purposes, depending on your desired
>>> use case.
>>>
>>> You mention that you change both u_flag and v_flag to true.
>>>
>>> change_initial_u_flag = .true.
>>>> change_initial_v_flag = .true.
>>>>
>>>
>>> U_flag is for the newer (as of instrument paper IV) cell-centered,
>>> better-shock-capturing hydro scheme useful for core-collapse supernova
>>> explosions. V_flag turns on the older, more traditional face-centered hydro
>>> solver with artificial viscosity. It’s not a good idea to have both on at
>>> the same time (though only one will take priority anyway). See Paxton+18
>>> <https://urldefense.com/v3/__https://ui.adsabs.harvard.edu/abs/2018ApJS..234...34P/abstract__;!!IKRxdwAv5BmarQ!d6n78vYvXa4k38_9Sj_ByNW2v1XKQ4AR4SsIxZS_srBBmuQn1rVm578bfMjRaTMGTDDAHCMnXM8HeMR7UwhfJiC8a3zqjDJfJg$> for
>>> a detailed explanation of the different hydro solvers and various tests.
>>>
>>> In the various test_suite cases bringing 12 and 20Msun stars to
>>> core-collapse (e.g. 20M_pre_ms_to_core_collapse), the v_flag hydro is set
>>> on and u_flag is off. Maybe try using one of those test cases as a point of
>>> reference instead.
>>>
>>> Cheers,
>>> ~Jared
>>>
>>> On Mar 18, 2024, at 2:35 PM, Ebraheem Farag via Mesa-users <
>>> mesa-users at lists.mesastar.org> wrote:
>>>
>>> 
>>> Hello Abha,
>>>
>>> I don't have a whole lot to say here beyond "Here be dragons".
>>>
>>> In the current form we only use the HLLC rieman solver for explosion
>>> modeling, not normal evolution calculations.
>>> Perhaps investigate the $MESA_DIR/star/test_suite/PPISN test_suite to
>>> get an idea for what might be needed to get rieman hydro working during a
>>> normal evolution run with the core intact.
>>> Also note that this was only done for very short timescales and a large
>>> mesh with tight time steps might be necessary to get anywhere. Sometimes
>>> the numerical initialization of riemann hydro can even drive the surface
>>> crazy as well if done during an evolution calculation. I have some working
>>> fixes for this but they are quite involved and do not exist in the current
>>> release version. Just know that we're working on it and we'll likely
>>> implement a future test suite to showcase how it's done :).
>>>
>>> -EbF
>>>
>>> On Fri, Mar 8, 2024 at 5:15 AM Abha Vishwakarma via Mesa-users <
>>> mesa-users at lists.mesastar.org> wrote:
>>>
>>>> Dear MESA Users,
>>>>
>>>> I have been trying to switch on hydrodynamics in a 12 solar mass red
>>>> giant star. I first ran the hse_reimann test suite which implements
>>>> hydrodynamics in a 12 solar mass star with the core excised. But, I wish to
>>>> implement hydrodynamics with the core intact.
>>>>
>>>> I switched on hydrodynamics by adding the following lines to my inlist:
>>>>
>>>>         change_initial_u_flag = .true.
>>>>     change_initial_v_flag = .true.
>>>>     new_v_flag = .false.
>>>>     new_u_flag = .true.
>>>>
>>>> My program takes as input a 12 solar mass red giant. But, the program
>>>> crashes because it hits the timestep limit or because of crossing some
>>>> tolerances. But, if I instead take the 12 solar mass star with the core
>>>> removed, as is done in the hse_reimann module, the code works fine.
>>>>
>>>> I am not sure what I am missing here, but I feel it has something to do
>>>> with the core of the red giant, or some boundary conditions or physics that
>>>> I might have missed. I have attached my inlist to this email, and any help
>>>> in this regard would be greatly appreciated!
>>>>
>>>> Sincerely,
>>>> Abha Vishwakarma
>>>> _______________________________________________
>>>> mesa-users at lists.mesastar.org
>>>>
>>>> https://urldefense.com/v3/__https://lists.mesastar.org/mailman/listinfo/mesa-users__;!!IKRxdwAv5BmarQ!bzUw4XcFA1iASqOxUxHhwcARQ6OtC83-QQ27p6R8v-bzLG-pTromCBdxIPBnsTAnsCwM93sbobPbhoMSOIB8KgNHttU90w$
>>>>
>>>> _______________________________________________
>>> mesa-users at lists.mesastar.org
>>> https://lists.mesastar.org/mailman/listinfo/mesa-users
>>> <https://urldefense.com/v3/__https://lists.mesastar.org/mailman/listinfo/mesa-users__;!!IKRxdwAv5BmarQ!d6n78vYvXa4k38_9Sj_ByNW2v1XKQ4AR4SsIxZS_srBBmuQn1rVm578bfMjRaTMGTDDAHCMnXM8HeMR7UwhfJiC8a3zGYiKI1g$>
>>>
>>>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.mesastar.org/pipermail/mesa-users/attachments/20240320/878ee178/attachment.htm>