[Mesa-users] mesh_dlogX controls for gravity modes

[Bill Paxton]

In response to Thomas’ email, I’ve implemented a new hook in mesa so that you can try new ideas for adjusting the mesh resolution without having to wait for the mesa team to add new features for you.   This hook will be in the next public release, but I expect that you’d like to start trying it sooner.   I’ll be happy to get you going with a private version whenever you want.

Cheers,
Bill




> On Dec 2, 2019, at 6:57 AM, Conny Aerts <conny.aerts at kuleuven.be> wrote:
> 
> 
>> On 2 Dec 2019, at 07:51, Thomas Steindl <thomas.steindl at uibk.ac.at> wrote:
>> 
>> Hello Conny,
>> 
>> At this point I am just trying to get MESA through some tough accretion. But you are right, at some point we want to calculate gravity modes and I was not aware that the mesh could have such large effect. Thank you very much for this information!
>> Could you explain how I can make sure that the mesh does not have an effect that is too large? Are you testing this as a resolution study?
>> 
> 
> Dear Thomas,
> 
> getting proper eigenvectors and eigenfrequencies of high-order gravity modes in stars with a convection zones requires specific attention to the meshing, with higher-than-usual resolution in the convective/radiative boundary layers. I suggest we follow-up this conversation off-list as it is probably too technical and of little interest to the majority of users on this email list. I’ll email you separately; very welcome in Leuven for live tutorials as well.
> 
> Cheers,
> Conny
> 
>> Thomas
>> 
>> 
>>> Am 30.11.2019 um 19:21 schrieb Conny Aerts <conny.aerts at kuleuven.be>:
>>> 
>>> Dear Bill, Thomas,
>>> 
>>>> Message: 1
>>>> Date: Sat, 30 Nov 2019 11:48:42 +0100
>>>> From: Thomas Steindl <thomas.steindl at uibk.ac.at>
>>>> To: Bill Paxton <paxton at kitp.ucsb.edu>
>>>> Cc: "mesa-users at lists.mesastar.org" <mesa-users at lists.mesastar.org>
>>>> Subject: Re: [Mesa-users] mesh_dlogX_species problem
>>>> Message-ID: <B8933AEF-56A3-468A-B32A-76D7BECE70BF at uibk.ac.at>
>>>> Content-Type: text/plain; charset="utf-8"
>>>> 
>>>> Hello Bill,
>>>> 
>>>> Thank you for quick response!
>>>> 
>>>> 
>>>>> Would you be willing to try these code changes in star/private/adjust_mesh_support and let me know if they work.   I think this now matches the >documentation, but I?d appreciate it if you?d run it and check and let me know.
>>>> 
>>>> I tried the code changes. I actually had to change the line
>>>> 
>>>> 	 X_min_for_extra = exp10(max(-50d0, logX_min_for_extra))
>>>> 
>>>> to 
>>>> 
>>>> 	 X_min_for_extra = exp10_cr(max(-50d0, logX_min_for_extra))
>>>> 
>>>> for it to compile. Now it matches the documentation and does exactly what it's supposed to do. Thank you!
>>>> 
>>>>> I can fix the code, but frankly I?d rather delete the feature and add a hook so that user?s like yourself can add new things.   That seems to me better >than having lots of rather obscure options which are never or rarely used (you are probably the 1st to try this option since it came into existence >years ago).  
>>>> 
>>> 
>>> no he is not...we gravity-mode asteroseismologists use that frequently...
>>> 
>>>> I think that?s a fantastic idea. There are actually a lot of options to change the meshing and frankly I was a little overwhelmed when I dove into it. I would love to have a hook for this and to be able to check whatever I want to increase the mesh. That would be great. 
>>> 
>>> @Thomas, given we are “scientific family”, I suspect you may be computing gravity modes and hit problems with spiky BV profiles. You can download Cole Johnston’s inlist on the mesa marketplace to see how we have set the mesh options for proper high-order g-mode frequency computations. I hasten to add that this was for v10108 and that we face the challenge of failures in our in-house g-mode tests whenever functionalities or numerical interpolation/integration schemes change in new MESA/GYRE versions; such changes tend to have an effect on the BV frequency that is larger than Kepler data precision for observed g modes (and then it is a no-go for my standards…). So don’t use the mesh settings that worked for v10108 without carefully testing their effect on the latest version of MESA.
>>> 
>>> Best, Conny
>>> _______________________________________________
>>> mesa-users at lists.mesastar.org
>>> https://lists.mesastar.org/mailman/listinfo/mesa-users
>>> 
>> 
> 
> _______________________________________________
> mesa-users at lists.mesastar.org
> https://lists.mesastar.org/mailman/listinfo/mesa-users
>