krome_equilibrium doesn't reach equilibrium

[Valentin Mauerhofer]

Hello,

I just found out that there was a "krome_equilibrium" routine,  it's great for what I want to do.

I have simulations with a few millions of cells and I want to compute things with Krome in post-processing, so at equilibrium.

But for around ~30% of the cells of my simulation,  Krome says :

krome_equilibrium: Did not converge in    50000000000.000000       years.

As you can see, I increased max_time from 500 Myr to 50 Gyr,  but the problem is still here.

Do you know why it happens ?

It can be because my network is a bit weird.  Hydrogen and helium recombination/collision rates are set to 0,  because I don't want to change their ionization fraction, I want to keep them as in the simulation. Is that something I shouldn't do ?  

Here is my network :

 

@format:idx,R,R,P,rate 

1,H+,E,H, 0d0 

2,He+,E,He, 0d0 

3,He++,E,He+, 0d0 

4,Si+,E,Si, auto

5,Si++,E,Si+, auto

6,Si+++,E,Si++, auto

7,Si++++,E,Si+++, auto

 

@format:idx,R,R,P,P,P,rate 

8,H,E,H+,E,E, 0d0 

9,He,E,He+,E,E, 0d0 

10,He+,E,He++,E,E, 0d0 

11,Si,E,Si+,E,E, auto

12,Si+,E,Si++,E,E, auto

13,Si++,E,Si+++,E,E, auto

14,Si+++,E,Si++++,E,E, auto

#Photoionization 

 

@photo_start 

@format:idx,R,P,P,rate 

15,H,H+,E,  0d0

16,He,He+,E, 0d0

17,He+,He++,E, 0d0 

18,Si,Si+,E, auto 

19,Si+,Si++,E, auto 

20,Si++,Si+++,E, auto 

21,Si+++,Si++++,E, auto 

@photo_stop 

 

@format:idx,R,R,P,P,Tmin,Tmax,rate 

22,Si++,H,Si+,H+,1d0,1d5, 6.77d-9*(T/1d4)**7.36d-2*(1 - 4.3d-1*exp(-1.1d-1*(T/1d4))) 

23,Si,H+,Si+,H,1d0,2d5, 9.2d-12*((T/1d4)**1.15d0)*(1 + 8d-1*exp(-2.4d-1*(T/1d4))) 

24,Si+,H+,Si++,H,1d3,1d5, 2.26d-9*((T/1d4)**7.36d-2)*(1 - 4.3d-1*exp(-1.1d-1*(T/1d4)))*exp(-3.1882d0/(T/1d4)) 

25,Si,He+,Si+,He,1d1,1d4, 1.3d-9 

26,Si+,He+,Si++,He,2.511d3,3d5, 1.5d-10*((T/1d4)**2.4d-1)*exp(-8.0187d0/(T/1d4)) 

27,Si++,He+,Si+++,He,3.162d3,3d5, 1.15d-9*((T/1d4)**4.4d-1)*exp(-1.0335d1/(T/1d4)) 

Thanks,

All the best,

Valentin

[Troels Haugbølle]

Dear Valentin,

I wrote large part of that routine. Sometimes the integration can get unstable and the solution start to oscillate if you take very long time-steps. As far as I can see below, you are looking at ionisation equilibrium, which I presume is driven by photo-chemistry.

That should reach equilibrium much faster than on a Gyr time-scale.

I think rather you may have to reduce the initial time-step in the routine. Maybe you can also print out abundances in the individual steps, as it moves towards equilibrium, for problem cells, to diagnose if indeed the solution oscillates in key species.

Best,

Troels

--
You received this message because you are subscribed to the Google Groups "KROMEusers" group.
To unsubscribe from this group and stop receiving emails from it, send an email to kromeusers+...@googlegroups.com.
To post to this group, send email to krome...@googlegroups.com.
Visit this group at https://groups.google.com/group/kromeusers.
To view this discussion on the web visit https://groups.google.com/d/msgid/kromeusers/9382475a-8110-49af-b4e8-755800f7d857%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.