[mesa-users] Problems with Rochelobe overflow

Tue Jun 14 04:33:09 EDT 2016

You can try to relax that option to see if timesteps become larger. Its
usually the case whn this happens that increasing the timestep will only
result in convergence issues, you need to experiment to see what happens.
El 14 jun. 2016 10:19 a. m., "Hannah Brinkman" <brinkmanhe at gmail.com>
escribió:

> My apologies for not sending a message to all. I just clicked on reply and
> didn't pay attention to the recipients.
> So, if I understand correctly, I just have to experiment a bit with the
> amount of Newton-iterations I allow the solver to take until I reach a
> point that the evolution continues properly?
> With kind regards,
> Hannah
>
> 2016-06-13 18:23 GMT+02:00 Pablo Marchant <pamarca at gmail.com>:
>
>> Hannah, please keep mesa-users in the loop.
>>
>> That's what's controlling your timestep. If you check carefully, as long
>> as the reason for the timestep limit is max_increase, timesteps will be
>> increasing, until once you get newton_iters, it will decrease. This option
>> is to reduce the timestep when the newton solver has do to many steps to
>> get a solution. You can relax this by increasing the
>> newton_iterations_limit option. Beware though, you likely have many
>> iterations because the solver is having trouble finding a solution, likely
>> increasing the timestep too much will result in convergence issues. But you
>> can experiment with that and see how it works.
>>
>> On Mon, Jun 13, 2016 at 5:04 PM, Hannah Brinkman <brinkmanhe at gmail.com>
>> wrote:
>>
>>> Dear Pablo,
>>> the terminal output is already set to every step. It just says max
>>> increase all the time at the end of the simulation, sometimes interupted by
>>> newton_iter, but mostly max increase.
>>> For the binary system I do not get neon-ignition, just shell-burning.
>>> And as soon as I see neon-ignition in the Kippenhahn-diagram, the star is
>>> no longer relevant for me since these stars won't evolve into
>>> electroncapture supernovae. The problem is that from about model 7000,
>>> nothing seems to happen anymore to the system and I need to know if there
>>> is a next phase of Rochelobe overflow to do a proper prediction whether the
>>> star will become an ECSN. This depends highly on the final mass of the core
>>> of the star, and if I have rochelobe overflow in the final stages of the
>>> evolution, the core  might not be able to become heavy enough to go
>>> supernova.
>>> The helium-coremass at this point is for the 8.5Msun primary is 1.33Msun
>>> which is also the total mass of this star. This rules out this star to
>>> become an ECSN, unless it accretes mass from its companion at a later
>>> stage. There is shell-burning, but no energy generation in the center of
>>> the star.
>>> Is there a way to get through this post-carbon burning phase in a
>>> smoother way? Because I don't think simply running the code longer will
>>> solve this problem.
>>> With kind regards,
>>> Hannah
>>>
>>> 2016-06-13 12:59 GMT+02:00 Pablo Marchant <pamarca at gmail.com>:
>>>
>>>> Hannah, you are likely miswing terminal output for every step. Set the
>>>> control terminal_cnt to 1 and you will get the reason for the timestep
>>>> control at every step.
>>>>
>>>> Max_increase means the timestep should be steadily increasing, so it
>>>> cant be thr single thing controlling your timesteps.
>>>>
>>>> What is the size of your helium core when you run into issues? If the
>>>> mass is not high enough you can get off-center neon ignition, which is
>>>> simply computationally expensive to model. If this is what is happrning,
>>>> further evolution should proceed rapidly and you might be able to simply
>>>> ignore it. Do some Kippenhahn diagrams to see what the core is doing at
>>>> those late phases.
>>>> Am 13.06.2016 12:10 nachm. schrieb "Hannah Brinkman" <
>>>> brinkmanhe at gmail.com>:
>>>>
>>>>> I checked the limitation on the timestep for the binary system, but it
>>>>> gives me simply max increase, which has no counterpart in the control
>>>>> files. This leaves me clueless on what to increase to get a bigger timestep
>>>>> in the final phases of the evolution. Any ideas how to proceed?
>>>>> The timesteps for the single star go down slowly from about a year to
>>>>> 10^-5 years at the very end of the evolution. This seems indeed to be the
>>>>> post-carbon burning state. It reaches this point in only 2000 timesteps and
>>>>> uses the other 8000 to get barely any further in the evolution. The
>>>>> limiting factor for the single star is dX_nuc_drop. But since these stars
>>>>> don't have Roche-lobe overflow that changes the mass drastically, I think I
>>>>> will be able to predict the final state of the star based on the core just
>>>>> after carbon-burning.
>>>>> With kind regards,
>>>>> Hannah
>>>>>
>>>>> 2016-06-10 17:22 GMT+02:00 Hannah Brinkman <brinkmanhe at gmail.com>:
>>>>>
>>>>>> Thanks for giving me some advice! At the moment I have no acces to my
>>>>>> data, so I will look at these suggestions on Monday.
>>>>>> With kind regards,
>>>>>> Hannah
>>>>>>
>>>>>> 2016-06-10 16:33 GMT+02:00 Robert Farmer <rjfarmer at asu.edu>:
>>>>>>
>>>>>>> What happens to the timestep if you run your primary star inlist as
>>>>>>> a single star? Post carbon burning can be difficult without the added
>>>>>>> complications of mass transfer.
>>>>>>>
>>>>>>> Also in your terminal output what's limiting the timestep? This will
>>>>>>> give you a clue as to what to change to get bigger timestep. Your looking
>>>>>>> for dt_limit in the output. Then search the binary and single star
>>>>>>> control.defaults file to match that limit to a inlist parameter.
>>>>>>>
>>>>>>> Rob
>>>>>>> Dear all,
>>>>>>> for my masterthesis I am using MESA to evolve binary systems in
>>>>>>> search for electron capture supernovae in binary systems. However, I am
>>>>>>> experiencing some troubles with the Rochelobe overflow, especially later in
>>>>>>> the evolution. This limits my ability to predict if a star will become a
>>>>>>> supernova or not since this depends a lot on the final mass of the star. If
>>>>>>> there is a phase of Rochelobe overflow at the end of the run and this
>>>>>>> removes enough mass, the supernova can be prohibited.
>>>>>>> The problem is that at the end of the simulation run, after
>>>>>>> carbon-burning, the timesteps become very small, as in seconds, and it
>>>>>>> seems that the code is a bit stuck. When checking the radius and radius of
>>>>>>> the Rochelobe, this seems to be the same moment as a second phase of
>>>>>>> Rochelobe overflow begins.
>>>>>>> Is there a way to get the binary system through this second
>>>>>>> Rochelobe overflow more smoothly?
>>>>>>> I have attached the inlist-files I am using. I use MESA-version 8118.
>>>>>>> With kind regards,
>>>>>>> Hannah Brinkman
>>>>>>>
>>>>>>>
>>>>>>> ------------------------------------------------------------------------------
>>>>>>> What NetFlow Analyzer can do for you? Monitors network bandwidth and
>>>>>>> traffic
>>>>>>> patterns at an interface-level. Reveals which users, apps, and
>>>>>>> protocols are
>>>>>>> consuming the most bandwidth. Provides multi-vendor support for
>>>>>>> NetFlow,
>>>>>>> J-Flow, sFlow and other flows. Make informed decisions using capacity
>>>>>>> planning reports.
>>>>>>> https://ad.doubleclick.net/ddm/clk/305295220;132659582;e
>>>>>>> _______________________________________________
>>>>>>> mesa-users mailing list
>>>>>>> mesa-users at lists.sourceforge.net
>>>>>>> https://lists.sourceforge.net/lists/listinfo/mesa-users
>>>>>>>
>>>>>>>
>>>>>>
>>>>>
>>>>>
>>>>> ------------------------------------------------------------------------------
>>>>> What NetFlow Analyzer can do for you? Monitors network bandwidth and
>>>>> traffic
>>>>> patterns at an interface-level. Reveals which users, apps, and
>>>>> protocols are
>>>>> consuming the most bandwidth. Provides multi-vendor support for
>>>>> NetFlow,
>>>>> J-Flow, sFlow and other flows. Make informed decisions using capacity
>>>>> planning reports.
>>>>> https://ad.doubleclick.net/ddm/clk/305295220;132659582;e
>>>>> _______________________________________________
>>>>> mesa-users mailing list
>>>>> mesa-users at lists.sourceforge.net
>>>>> https://lists.sourceforge.net/lists/listinfo/mesa-users
>>>>>
>>>>>
>>>
>>
>>
>> --
>> Pablo Marchant Campos
>> M.Sc on Astrophysics, Universidad Católica de Chile
>> PhD student, Argelander-Institut für Astronomie
>>
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.mesastar.org/pipermail/mesa-users/attachments/20160614/1e7ae009/attachment.html>