Disks tilt

Regarding Keplarian disks, with gamma=1.001, which move about a point gravity without any disturbance from the ambient medium. I've found a correlation between the point gravity softening radii, rs, and the undesired disk gas motion. This is for the binary simulation setup (i.e. grid size, BC, and scales).

Basically, for 4 amr levels, disks are symmetric and stable (i.e. neither tilt nor explosion) for rs>0.5, but unstable for smaller rs. More exploration is needed, but thing are looking pretty good.

r=0.5 is about 3% of the grid size. The wind-capture disks that we've been seen so far had a radius close to 0.5, but his may change now with a different r_s. for 4 amr, r_s this was resolved with 8 cells

The stable disks I've been seeing have a radius of .5 too. This means that all the initial disk gas is in the flat part of the grav potential. I was using much smaller r_s because I wanted to model the grav potential as accurate as possible.

Comments

1. martinhe -- 13 years ago

I see table disk for rs=0.25 and for a disk with the same temperature than the ambient medium. Thus I'm trying this setup in the binary problem. I can see that the very first concentration of gas about the secondary is different than it was with the "tilting" setup.

2. martinhe -- 13 years ago

And by table disk, I meant stable disks. :P

3. martinhe -- 13 years ago

I've been running the binary simulation, but using a gravity soft radius of .5. Part 1 of the simulation has ran; when the primary fills the grid with its spherical wind. I've just restarted the simulation but now the secondary will pull the wind gas towards it (with rs=0.5). I should find whether or not the formed disk tilts or not by tonight.

4. martinhe -- 13 years ago

To understand the effect of a wind on disks, I've been running a simulation where the initial conditions are a Keplerian disk with a static point gravity particle, gamma=1.001, a density contrast of 100 relative to the ambient, a radius to height ratio of 4, and rgrav-soft=0.5.

http://www.pas.rochester.edu/~martinhe/2011/binary/6dec11b.png

I let the disk to relax until it reached a quasi-steady state:

http://www.pas.rochester.edu/~martinhe/2011/binary/6dec11c.png

Then I inject a wind from one of the domain boundaries (left in the figure below) which has a constant velocity of about 10mks-1 and is perpendicular to the angular momentum vector of the disk:

http://www.pas.rochester.edu/~martinhe/2011/binary/6dec11d.png

There are some asymmetries in the flow, likely produced by the particle grid refinement, but not a pronounced tilt. I'll then try with a wind with a spherical distribution, injected by an outflow object located at, again, the left boundary.