Context Navigation

← Previous Post
Next Post →

Disk formation in binaries using a co-rotating frame

The simulations in this blog are based on these ideas:

⁺https://clover.pas.rochester.edu/trac/astrobear/blog/johannjc04032012

⁺⁺https://clover.pas.rochester.edu/trac/astrobear/blog/johannjc03312012

8May '12

30AU, r_Bondi=6, vel_w=10km/s, t_w=1000K, \gamma=1 with 96x96x144 + 2amr res, 4hrs, 40 afrank procs, bhive

30AU, t_w=1000K, \gamma=1 with 100x100x160 + 4amr res. No disk formed after .07 orbits (3days, 1024 ranger procs).
- Movies of logarithmic density (grayscale)and linear velocity [Mach]:
  - orbital plane http://www.pas.rochester.edu/~martinhe/2011/binary/30au-1000k.gif
  - orbital plane, zoom in http://www.pas.rochester.edu/~martinhe/2011/binary/30au-1000k-z.gif
  - longitudinal plane, zoom in http://www.pas.rochester.edu/~martinhe/2011/binary/30au-1000k-z-l.gif

27 Apr '12

40AU, t_w=3000K, \gamma=1 with 168³ + 4amr res is running at ranger's long (1024p, 2days) queue.

10AU, t_w=3000K, \gamma=1 with 112x144x160 + 3amr res. No disk formed after .21 orbits.

Orbital plane Longitudinal plane

Understanding/solving the dynamics of the 40au, \gamma=1.1, t_w=300K sim (below).
- Here's a temperature map of the whole orbital plane http://www.pas.rochester.edu/~martinhe/2011/binary/corot-40au-sol2-300K-gamma1.1e.gif.
  - Hot gas (10⁵ K) seem to discontinually emanate from the particle, specially early in the sim.
- Now running a short test, same parameters, with 10 times more frames.
  - Temperature movie 2. 20 times faster frame rate than the temp movie above and zoomed in to the particle, http://www.pas.rochester.edu/~martinhe/2011/binary/corot-40au-sol2-300K-gamma1.1f.gif.

26 Apr '12

10AU, t_w=3000K, \gamma=1 with ~64³ + 2amr. No disk formed after 2 orbits (16hrs, 256 ranger procs). Trying 2 times more res.

40AU, t_w=3000K, \gamma=1 with ~128³ + 2amr. No disk formed after .44 orbits (1day 256 ranger procs). Images show a zoom in to the particle.

Orbital plane Longitudinal plane

40AU, t_w=300K, ideal gas solver, \gamma=1.1. Very interesting dynamics.
- ~128³+2amr, 0.2 orbits / 1day, 256 ranger procs.
- Movies of logarithmic density (grayscale)and linear velocity [Mach]:
  - orbital plane http://www.pas.rochester.edu/~martinhe/2011/binary/corot-40au-sol2-300K-gamma1.1a.gif
  - orbital plane, zoom in http://www.pas.rochester.edu/~martinhe/2011/binary/corot-40au-sol2-300K-gamma1.1b.gif
  - longitudinal plane http://www.pas.rochester.edu/~martinhe/2011/binary/corot-40au-sol2-300K-gamma1.1c.gif
  - longitudinal plane, zoom in http://www.pas.rochester.edu/~martinhe/2011/binary/corot-40au-sol2-300K-gamma1.1d.gif
  - Jet? This corresponds to the last frame of this short test sim.

Though I was expecting differences between the \gamma=1 and \gamma=1.1 cases, I'm surprised that they are SO different. This is inconsistent with http://adsabs.harvard.edu/abs/2000MNRAS.316..906M. One has to be careful though for their parameters are in general different than ours.

24 Apr '12. Research meeting. To try:

10AU and 40AU low res and t_wind=3000K, instead of 300K. Tests ran
40AU, t_wind=300K, ideal gas solver, \gamma=1.1. short test ran
Jonathan's latest thread implementation on the global (2 particle) sims.

23 Apr '12

separation = 40AU
m_prim= 1.5 M_o
m_sec = 1 M_o
temp_w = 300 K
dens_w = 1.5x10¹¹ gr/cc
v_w = 15 km/s = 15 Mach
v_orbit-sec = 4.5 km/s
r_Bondi = Gm_sec/ ( v_w² + v_orbit-sec² )= 3.6 AU
r_soft = 4 cells
r_Bondi/r_soft = 19.3
r_orbit/r_Bondi = 6.62

A) low res=56x72x80 + 3 particle Ref. .43 orbits / 26 hrs with ~ 32 afrank nodes in bhive. Running.

Movies:

dens+vels, http://www.pas.rochester.edu/~martinhe/2011/binary/corot-40au-sol2.gif

dens+vels, secondary zoom in, http://www.pas.rochester.edu/~martinhe/2011/binary/corot-40au-sol2z.gif
dens+vels, secondary zoom in, longitudinal plane http://www.pas.rochester.edu/~martinhe/2011/binary/corot-40au-sol2l.gif

Tilt?. At some point during this simulation I see a rather brief (lasting for ~ 0.005 orbits) tilted disk-like structure similar to the ones I saw in the global (2 particles) sims (https://clover.pas.rochester.edu/trac/astrobear/blog/binaryBondi, top table).

http://www.pas.rochester.edu/~martinhe/2011/binary/question0000.png

B) high res= 56x72x80 + 5 particle Ref. .05 orbits / 20 hrs with 256 ranger procs. Running. Note the images below show gas at a point at which the mass of the secondary is increasing to reach its final 1M_o value. This was dome for numerical stabilization and, given that we want to run the sim for several orbits, it should not affect the formation/evolution of the disk.

Gas momenta. I'm using this plot trying to understand why gas is not going about the seconday as expected.

http://www.pas.rochester.edu/~martinhe/2011/binary/idea1.png

20 Apr '12

We've determined that Jonathan's module template does not have a frame/secondary which follow a circular orbit and has extreme parameters. Presumably, a disk is formed in such setup, with a separation as large as 350AU, because the unstable orbit is subject to a lot shear from the flow which might make rotation about the secondary easier.

The 1st, intermediate-res test with the full wind solution (wind+secondary's gravity) has completed for

separation = 10AU
m_prim= 1.5 M_o
m_sec = 1 M_o
temp_w = 300 K
dens_w = 1.5x10¹¹ gr/cc
v_w = 15 km/s = 15 Mach
v_orbit-sec = 8.9 km/s
r_Bondi = Gm_sec/v_w² = 4 AU
r_soft = 4 cells
r_Bondi/r_soft = 21
r_orbit/r_Bondi = 1.52
running time ~ 1 orbit/11hrs (16afrank, bhive nodes)

Movies:

dens+vels, http://www.pas.rochester.edu/~martinhe/2011/binary/corot-10au1.gif
dens+vels, secondary zoom in, http://www.pas.rochester.edu/~martinhe/2011/binary/corot-10au2.gif
ram pressure, zoom in, http://www.pas.rochester.edu/~martinhe/2011/binary/corot-10au3.gif

Higher res test running now. Note that we've not used this (new) wind solution for the 40AU case.

19 Apr '12 Trying to understand why Jonathan's setup⁺ forms a disk while mine (below) doesn't. Here's a parameter comparison.

	Jonathan	Martin
m_prim [M_o]	1.1	1.5
m_sec [M_o]	.19	1
separation [AU]	350	40
temp_w [K]	0.01	300
dens_w [part cm^-3]	1^a	1.5x10¹¹
v_w^b [km/s]	.9	15
v_orbit-sec [km/s]	.54^c	4.5
r_Bondi [AU]	200	4
r_soft [cells]	4	4

^a \dot{M} = \rho 4 \pi r_{sec}^2 v_w. \text{~Thus using~} \rho=1 ~~\text{part/cc and~} r_{sec}= 50 \text{~AU (the position of the grid boundary), yield}

\dot{M}=1.7 \times 10^{-14} M_o / yr << \dot{M}_{AGB} \sim 1 \times 10^{-5} M_o/yr.

^b \text{~From~} m_{pri}/(m_{pri}+m_{sec}) \sqrt{G(m_{pri}+m_{sec})/a}, \text{~where~} a \text{~is the stellar separation.~}

^c \Omega r = 2*3=6 \text{~cu~} = .54 \text{~km/s~}.

The combined mass and the separation in the run of the left column seem extreme.

16 Apr '12

Same parameters as the run below, but a slightly larger grid and 3 spatially fixed refinement (particle) levels, http://www.pas.rochester.edu/~martinhe/2011/binary/corot1.gif.

Discussion and further analysis needed.
I do not see a disk forming. Will try with 2 more refinements in order to have ~ 40cells/r_Bondi.
The flow shows some strange pulsations not related to protections.
The accretion tail is not a collimated one, like that in the low res run (below), and it broadens up.

6 Apr '12

I've got Jonathan's new co-rotating frame setup^+,++ running for a binary system with the parameters below which are representative of my previous lab frame binary runs (see the table below):

Temp_wind=300K
AGB mass-loss = 10^-5 M_o yr^-1
vel_wind=9 km/s (as in Mastrodemos & Morris)
vel_wind/vel_orbit_sec = 2.01
a=40AU
circular orbit
q=m_pri/m_sec=1.5
r_soft=4 cells
64x64x128 cells, fixed⁺⁺⁺
grid: 40x40x80 AU.
t_run~.5 orbit/hr in 16 debug bluehive procs
dx_co-rotating ~ 13 dx_{lab frame}

Log(dens) and vel field movie. Pole-on (left) edge-on (right). The green diagonal line in the left panel show the direction of the plane corresponding to the right panel. A disk forms. The tail formed by accretion eventually gets inflow from -Y. The tail is bent then. The edge-on panel shows asymmetries. We are unresolved with respect to the global (lab frame) simulations to fairly compare/conclude anything about disk tilting at this point.

http://www.pas.rochester.edu/~martinhe/2011/binary/coRot1.gif

Here's a second try at this sim but with a larger grid. This last sim will run, any time now, at bhive with 512procs and 128³+3amr.

astrobear

Context Navigation

Disk formation in binaries using a co-rotating frame

Comments

Download in other formats: