[mesa-users] accretion_powered_irradiation
Bill Wolf
wmwolf at physics.ucsb.edu
Tue Jun 14 12:15:58 EDT 2016
Hi Wen-Cong,
It looks like you're off to a great start! Something about the
irradiation is causing the problem (well obviously, since it's the only
thing you changed), where lower irradiated fluxes don't give the solver
too many problems (as you would expect, since the test case should work
fine as Flux -> 0).
I'm not familiar with the irradiation controls for binary, but it seems
like fiddling with the maximum allowed irradiated flux or the depth at
which it is deposited (col_depth_for_eps_extra) would be good places to
start. Note also that the default radius for the accretor is set to 1d6
cm, or 10 km (i.e., the accretor is assumed to be a neutron star). I
don't know what scenario you are trying to model, but if it isn't a NS
X-ray binary, this matters a lot.
Did you do what I suggested and look at real-time plots during the
convergence troubles? I can't stress how important and useful this is in
both diagnosing problems and building intuition for your stellar models.
Regards,
Bill
On 06/14/2016 09:01 AM, Chen WC wrote:
> Dear Bill,
> Just now, I use your binary_donor_only_implicit_mdot (in
> /binary/test_suite) to test irradiation process.
>
> In inlist_project, I add two lines as follows (other files have not
> been revised):
> ________________
> accretion_powered_irradiation = .true.
> col_depth_for_eps_extra = 100.0
> ________________
>
> The calculation shows a same problem. It seems that a strong
> irradiation flux may be responsible for the converge problem.
> If I take a relatively maximum irradiation flux such as max_F_irr =
> 1.0d11, the calculation can be continuous.
>
> Some test results in the terminal are as follows:
> _______________________________
>
> hydro_newton_step returned ierr
> retry cnt, step, log10(dt/yr), retry_factor 1 1672
> -0.679740 0.500000
> 1 1672 4.655311795E+07
> 2.575229314E+00 -5.660728564E+00 4.462894360E-03 photos1/x650
> lg F_irr 1672
> 1.2698970004336019D+01
> 1 1673 4.655311845E+07
> 2.575228218E+00 -5.660728564E+00 4.463616201E-03 photos1/x650
> lg F_irr 1673
> 1.2698970004336019D+01
> 1 1674 4.655311906E+07
> 2.575226903E+00 -5.660728564E+00 4.452531008E-03 photos1/x650
> lg F_irr 1674
> 1.2698970004336019D+01
> 1 1675 4.655311978E+07
> 2.575225325E+00 -5.660728564E+00 4.435920888E-03 photos1/x650
> lg F_irr 1675
> 1.2698970004336019D+01
> 1 1675 4.655311978E+07
> 2.575225325E+00 -5.660728564E+00 4.452531008E-03 photos1/x650
> hydro_newton_step returned ierr
> retry cnt, step, log10(dt/yr), retry_factor 1 1675
> -0.743226 0.500000
> 1 1675 4.655311942E+07
> 2.575226114E+00 -5.660728564E+00 4.464310264E-03 photos1/x650
> lg F_irr 1675
> 1.2698970004336019D+01
> 1 1676 4.655311985E+07
> 2.575225167E+00 -5.660728564E+00 4.465242717E-03 photos1/x650
> lg F_irr 1676
> 1.2698970004336019D+01
> 1 1677 4.655312037E+07
> 2.575224031E+00 -5.660728564E+00 4.453366447E-03 photos1/x650
> lg F_irr 1677
> 1.2698970004336019D+01
> 1 1678 4.655312100E+07
> 2.575222668E+00 -5.660728564E+00 4.434346597E-03 photos1/x650
> lg F_irr 1678
> 1.2698970004336019D+01
> 1 1678 4.655312100E+07
> 2.575222668E+00 -5.660728564E+00 4.453366447E-03 photos1/x650
> hydro_newton_step failed to converge
> retry cnt, step, log10(dt/yr), retry_factor 1 1678
> -0.806713 0.500000
> 1 1678 4.655312068E+07
> 2.575223350E+00 -5.660728564E+00 4.465193254E-03 photos1/x650
> lg F_irr 1678
> 1.2698970004336019D+01
> 1 1679 4.655312106E+07
> 2.575222532E+00 -5.660728564E+00 4.468752930E-03 photos1/x650
> lg F_irr 1679
> 1.2698970004336019D+01
> star model age mass
> lg_mdot (r-rl)/rl last photo
> 1 1680 4.655312151E+07
> 2.575221550E+00 -5.660728564E+00 4.456992571E-03 photos1/x650
> lg F_irr 1680
> 1.2698970004336019D+01
> 1680 7.331477 1.209E+04 1.150339 1.150339 2.575222
> 2.575222 0.636153 0.007871 0.280005 0.684625 -3.262558 915
> 363
> -0.347320 1.570276 0.194863 -37.917546 -0.070347 -5.660729
> 0.000000 0.343696 0.006040 0.020000 0.295185 0.040638 3
> 36
> 4.6553E+07 17.009051 1.672812 -14.449440 3.060987 -9.084476
> 0.000000 0.000049 0.002099 2.015E-02 2.019E-02 0.120E-03 max
> increase
>
> 1 1681 4.655312205E+07
> 2.575220372E+00 -5.660728564E+00 4.437264242E-03 photos1/x650
> lg F_irr 1681
> 1.2698970004336019D+01
> 1 1681 4.655312205E+07
> 2.575220372E+00 -5.660728564E+00 4.456992571E-03 photos1/x650
> hydro_newton_step failed to converge
> retry cnt, step, log10(dt/yr), retry_factor 1 1681
> -0.870199 0.500000
> 1 1681 4.655312178E+07
> 2.575220961E+00 -5.660728564E+00 4.467380124E-03 photos1/x650
> lg F_irr 1681
> 1.2698970004336019D+01
> 1 1682 4.655312210E+07
> 2.575220254E+00 -5.660728564E+00 4.468891483E-03 photos1/x650
> lg F_irr 1682
> 1.2698970004336019D+01
> 1 1683 4.655312249E+07
> 2.575219406E+00 -5.660728564E+00 4.460107099E-03 photos1/x650
> lg F_irr 1683
> 1.2698970004336019D+01
> 1 1684 4.655312296E+07
> 2.575218388E+00 -5.660728564E+00 4.441551583E-03 photos1/x650
> lg F_irr 1684
> 1.2698970004336019D+01
> 1 1685 4.655312351E+07
> 2.575217167E+00 -5.660728564E+00 4.420451056E-03 photos1/x650
> lg F_irr 1685
> 1.2698970004336019D+01
> 1 1685 4.655312351E+07
> 2.575217167E+00 -5.660728564E+00 4.441551583E-03 photos1/x650
> hydro_newton_step failed to converge
> retry cnt, step, log10(dt/yr), retry_factor 11685 -0.8545040.500000
> __________________________________________
>
>
> Cheers,
>
> Wen-Cong
>
> 在 2016年6月14日，上午1:39，Bill Wolf <wmwolf at physics.ucsb.edu
> <mailto:wmwolf at physics.ucsb.edu>> 写道：
>
>> Hi Wen-Cong,
>>
>> We can’t reproduce this without also having inlist1, and any model
>> loaded by it (as well as your run_star_extras, if you edited that).
>>
>> But from the output you’ve provided, it looks likes the solver is
>> indeed struggling periodically amidst a sea of “max_increase”. I’m
>> guessing this is the case based on the “hydro_newton_step failed to
>> converge” and the other related retry about halfway down.
>>
>> This can be a hard type of problem to diagnose. Make sure you are
>> tracking the structure of your star with pgstar, because it may
>> become clear to you what’s difficult for the star to deal with. If
>> nothing jumps out at you there, you’ll have to go for full-blown
>> hydro debugging tools I linked to earlier.
>>
>> Neither of these give you a simple fix in the form of an inlist
>> control or some hack in run_star_extras, but an understanding of the
>> state of the star can inform new strategies for getting out of this
>> retry limit cycle. For instance, problems very near the surface can
>> be “ignored” by setting tau_factor to a sufficiently high number so
>> that the problem is “outside the model”. Obviously this comes at some
>> cost to the fidelity of your model, so experiment with caution.
>>
>> Good luck!
>>
>> Regards,
>>
>> Bill
>>
>> ________________________
>> William Wolf
>> wmwolf at physics.ucsb.edu <mailto:wmwolf at physics.ucsb.edu>
>>
>> UCSB Department of Physics
>>
>> On June 11, 2016 at 10:22:52 AM, Chen WC (chenwc at pku.edu.cn
>> <mailto:chenwc at pku.edu.cn>) wrote:
>>
>>> Hi, Bill,
>>> Thank you for your response.
>>>
>>> My inlist_project file is as follows:
>>> ******
>>> &binary_job
>>>
>>> inlist_names(1) = 'inlist1'
>>> inlist_names(2) = 'inlist2'
>>>
>>> evolve_both_stars = .false.
>>>
>>> / ! end of binary_job namelist
>>>
>>> &binary_controls
>>>
>>> m1 = 3.0d0 ! donor mass in Msun
>>> m2 = 1.4d0 ! companion mass in Msun
>>> initial_period_in_days =1.0d0
>>>
>>> fr = 0.05
>>> fr_limit = 1.0d-2
>>>
>>> limit_retention_by_mdot_edd = .true.
>>> accretion_powered_irradiation = .true.
>>> col_depth_for_eps_extra = 100.0
>>> accretor_radius_for_irrad =1.0d6
>>> ! max_F_irr = 5.0d11
>>>
>>> use_other_jdot_mb = .true.
>>> ! use_other_jdot_ml = .true.
>>> use_other_mdot_edd = .true.
>>> ! keep_mb_on = .true.
>>> max_tries_to_achieve = 50
>>> mdot_scheme = 'Ritter'
>>> / ! end of binary_controls name list
>>> *****
>>>
>>> Some output results in the terminal are as follows:
>>>
>>> ******
>>> hydro_newton_step failed to converge
>>> retry cnt, step, log10(dt/yr), retry_factor 1 2737
>>> -0.788149 0.500000
>>> 1 2737 2.216895761E+08
>>> 2.117224534E+00 -5.713074589E+00 6.755361544E-03 photos1/x700
>>> lg F_irr 2737 1.2698970004336019D+01
>>> 1 2738 2.216895765E+08
>>> 2.117223777E+00 -5.713074589E+00 6.749303810E-03 photos1/x700
>>> lg F_irr 2738 1.2698970004336019D+01
>>> 1 2739 2.216895770E+08
>>> 2.117222869E+00 -5.713074589E+00 6.752663357E-03 photos1/x700
>>> lg F_irr 2739 1.2698970004336019D+01
>>> 1 2740 2.216895775E+08
>>> 2.117221779E+00 -5.713074589E+00 6.763619824E-03 photos1/x700
>>> lg F_irr 2740 1.2698970004336019D+01
>>> 1 2741 2.216895782E+08
>>> 2.117220471E+00 -5.713074589E+00 6.776681402E-03 photos1/x700
>>> lg F_irr 2741 1.2698970004336019D+01
>>> 1 2742 2.216895790E+08
>>> 2.117218902E+00 -5.713074589E+00 6.785790596E-03 photos1/x700
>>> lg F_irr 2742 1.2698970004336019D+01
>>> 1 2742 2.216895790E+08
>>> 2.117218823E+00 -5.691885290E+00 6.706529261E-03 photos1/x700
>>> lg F_irr 2742 1.2698970004336019D+01
>>> 1 2742 2.216895790E+08
>>> 2.117218823E+00 -5.691885290E+00 6.776681402E-03 photos1/x700
>>> hydro_newton_step returned ierr
>>> retry cnt, step, log10(dt/yr), retry_factor 1 2742
>>> -0.693273 0.500000
>>> 1 2742 2.216895786E+08
>>> 2.117219686E+00 -5.713074589E+00 6.770370486E-03 photos1/x700
>>> lg F_irr 2742 1.2698970004336019D+01
>>> 1 2743 2.216895791E+08
>>> 2.117218745E+00 -5.713074589E+00 6.770294906E-03 photos1/x700
>>> lg F_irr 2743 1.2698970004336019D+01
>>> 1 2744 2.216895797E+08
>>> 2.117217615E+00 -5.713074589E+00 6.774228347E-03 photos1/x700
>>> lg F_irr 2744 1.2698970004336019D+01
>>> 1 2745 2.216895804E+08
>>> 2.117216259E+00 -5.713074589E+00 6.779083382E-03 photos1/x700
>>> lg F_irr 2745 1.2698970004336019D+01
>>> 1 2746 2.216895812E+08
>>> 2.117214632E+00 -5.713074589E+00 6.796953754E-03 photos1/x700
>>> lg F_irr 2746 1.2698970004336019D+01
>>> 1 2746 2.216895812E+08
>>> 2.117214551E+00 -5.691885290E+00 6.718226481E-03 photos1/x700
>>> lg F_irr 2746 1.2698970004336019D+01
>>> 1 2746 2.216895812E+08
>>> 2.117214593E+00 -5.702907945E+00 6.758997650E-03 photos1/x700
>>> lg F_irr 2746 1.2698970004336019D+01
>>> 1 2747 2.216895822E+08
>>> 2.117212595E+00 -5.702907945E+00 6.746944676E-03 photos1/x700
>>> lg F_irr 2747 1.2698970004336019D+01
>>> 1 2748 2.216895835E+08
>>> 2.117210196E+00 -5.702907945E+00 6.725341516E-03 photos1/x700
>>> lg F_irr 2748 1.2698970004336019D+01
>>> 1 2749 2.216895849E+08
>>> 2.117207318E+00 -5.702907945E+00 6.695434808E-03 photos1/x700
>>> lg F_irr 2749 1.2698970004336019D+01
>>> 1 2749 2.216895849E+08
>>> 2.117207455E+00 -5.724097244E+00 6.778111874E-03 photos1/x700
>>> lg F_irr 2749 1.2698970004336019D+01
>>> 1 2749 2.216895849E+08
>>> 2.117207390E+00 -5.713849448E+00 6.739682584E-03 photos1/x700
>>> lg F_irr 2749 1.2698970004336019D+01
>>> star model age
>>> mass lg_mdot (r-rl)/rl last photo
>>> 1 2750 2.216895867E+08
>>> 2.117204022E+00 -5.713849448E+00 6.714233214E-03 photos1/x700
>>> lg F_irr 2750 1.2698970004336019D+01
>>> 2750 7.343212 1.214E+04 0.969037 0.969037 2.117204
>>> 2.117204 0.289812 0.011725 0.280162 0.601416 -3.186960
>>> 904 567
>>> 0.241207 1.727500 0.311526 -36.435469 -0.214368 -5.713849
>>> 0.000000 0.690597 0.001600 0.020000 0.378468 0.058189 4
>>> 49
>>> 2.2169E+08 17.035797 1.912756 -14.750470 2.908839 -9.265046
>>> 0.000000 0.000076 0.002099 1.959E-02 2.012E-02 0.630E-04 max
>>> increase
>>>
>>> save photos1/x750 for model 2750
>>>
>>> ***********
>>>
>>>
>>>
>>>
>>>
>>> Cheers,
>>>
>>> Wen-Cong
>>>
>>>
>>>
>>>
>>> 在 2016年6月10日，下午6:20，Bill Wolf <wmwolf at physics.ucsb.edu
>>> <mailto:wmwolf at physics.ucsb.edu>> 写道：
>>>
>>>> Hi Chen,
>>>>
>>>> Full inlists and model files would be helpful for us to reproduce
>>>> this error.
>>>>
>>>> My guess, though, is that you’re having periodic retries and/or
>>>> backups. Often I’ll have a series of steps where the tilmestep is
>>>> increasing as much as it can (max_increase) and then once it has
>>>> gotten to some critical timestep size, the hydro solver struggles
>>>> to converge and forces a retry, bringing the timesteps back down.
>>>>
>>>> If this is a case, I have no simple solution for you, but you might
>>>> try to use the hydro debugging tools to see what the physical
>>>> situation is that is causing problems. See my guide over here:
>>>> http://wmwolf.github.io/projects/mesa_debugging/ for an example of
>>>> how to do this.
>>>>
>>>> Bill Wolf
>>>>
>>>> ________________________
>>>> William Wolf
>>>> wmwolf at physics.ucsb.edu <mailto:wmwolf at physics.ucsb.edu>
>>>>
>>>> UCSB Department of Physics
>>>>
>>>> On June 10, 2016 at 9:59:52 AM, Chen WC (chenwc at pku.edu.cn
>>>> <mailto:chenwc at pku.edu.cn>) wrote:
>>>>
>>>>> Dear All,
>>>>> I use the “binary_donor_only_implicit_mdot” to calculate the
>>>>> evolution of an intermediate-mass X-ray binary including a 1.4
>>>>> Msun NS and a 3.0 Msun donor star (with a solar composition),
>>>>> and the initial orbital period is 1.0 day.
>>>>>
>>>>> In calculation, I consider the irradiation process by X-ray
>>>>> luminosity of the NS. In the list_project:
>>>>> ****
>>>>> accretion_powered_irradiation = .true.
>>>>> col_depth_for_eps_extra = 100.0 ! g/cm^2
>>>>> ****
>>>>> Other input parameters such as accretor_radius_for_irrd, max_F_irr
>>>>> were adopted to be default.
>>>>>
>>>>>
>>>>> However, after the mass of the donor star decrease to be ~ 2.3-2.4
>>>>> Msun, the calculation is very difficult to continue.
>>>>> The timestep decrease to be <0.1 yr, and the terminal output for
>>>>> timestep control displays “max increase”.
>>>>>
>>>>> I would appreciate if someone can help me to solve this problem.
>>>>>
>>>>>
>>>>> Cheers,
>>>>>
>>>>>
>>>>> Chen Wen-Cong
>>>>>
>>>>>
>>>>> ------------------------------------------------------------------------------
>>>>>
>>>>> 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
>>>>> <mailto:mesa-users at lists.sourceforge.net>
>>>>> https://lists.sourceforge.net/lists/listinfo/mesa-users
>
>
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.mesastar.org/pipermail/mesa-users/attachments/20160614/24ee8dcf/attachment.html>
More information about the Mesa-users
mailing list