Posts for the month of September 2016

A good read

I updated the developer guide with the various graphs showing the stencil components and their dependencies… See SweepScheme

meeting update

Rotating, supercritical Bonnor Ebert sphere forms a disk, protostar, and outflow.

Here is the density right before the outflow leaves the sphere:

right after:

Looks like a RT instability along jet axis.

Radial velocity (ignore the legend label — it's wrong) shows that the kernel is not directed outward, but beyond that, it is… Not sure why yet. Infall is happening in the disk though, which is as expected:

The temperature shows the outflow is decades larger than the core. This may need to be adjusted:

A movie of density can be found here:

https://www.youtube.com/watch?v=rpUXAqMHjCs

Update 9/21

  • Debugging MHD simulation runs. Starting with Bp = 0.54 G, Bs = 13.5 G, cutoff radii of 0.125 and 1*1010 respectively. First run, time to completion approached infinity - fixed this, but simulations are still extremely slow. Currently running short time, high framerate. Also calculated Alfven speed () and plotted contours - high (3.5*107 cm/s) near poles of planet. This seems to be higher than I would expect for the maximum Alfven speed, which should be near 5*106 cm/s.

  • Read three papers on WASP-12b observations (Absorbing Gas, Metals in Exosphere, NUV absorption. Observations of WASP-12 have found lines missing from the magnesium spectrum, particularly, which are extremely unlikely to not be radiating from the star. In order to get absorption to the observed levels, it requires a density of a least an order of magnitude more than the interstellar medium. It is therefore hypothesized that WASP-12b is creating a cloud of gas that is capable of creating the observed magnesium absorption lines.

Cooling Test Results for ThermalPulse module

2 or 3 levels of AMR

DMcooling with floorTemp=100K DMcooling with floorTemp=500K
density http://www.pas.rochester.edu/~bliu/Matthias/rhoScaled_DM100K_fr200.png http://www.pas.rochester.edu/~bliu/Matthias/rhoScaled_DM500K_f200.png
Temp http://www.pas.rochester.edu/~bliu/Matthias/Temp_DM100K_fr200.png http://www.pas.rochester.edu/~bliu/Matthias/Temp_DM500K_f200.png
Velocity http://www.pas.rochester.edu/~bliu/Matthias/vel_DM100K_fr200.png http://www.pas.rochester.edu/~bliu/Matthias/vel_DM500K_f200.png
Movies density; temperature; velocity density; temperature; velocity

A summary

Update 9/13

  • Determining parameters for magnetic fields of planet and star for HD209458b simulations:

Matsakos et al. use the plasma at the base of the outflows (surfaces) to define the magnetic field strength. With , the magnetic field at the surface is 11.5 G, or ~10 Bs.

The magnetic field of HD209458b was estimated to be ~10% that of Jupiter (Ref), or ~0.5 G. This gives , which is right in the middle of the range used in Matsakos (0.002 - 400).

Currently, calculations of the bow shock radius use . Will need to convert this to using by relating magnetic field:

Or, in terms of knowns,

With the values above, this gives a bow shock distance of 0.155 (in computational units).

(What does from PlanetaryAtmospheres represent? to plasma )

The other parameter that needs to be set is the cutoff radius for the fields. Currently planning on large (essentially infinity) for the star and near the edge of the box for the local simulations (0.125 orbital radii).

Once these are set, I'll be ready to start running simulations. Plan to begin with isotropic temp profile and no rotation, then jumping to anisotropic with rotation (& the global model) if it works - are we interested in the intermediate models still?

  • Also beginning to look at simulating WASP-12b, to attempt to reproduce the observed absorption lines of Mg II and Ca II. Physically, this requires that a fairly dense (109/cm3 of H) and stable torus be created by the capture of the planetary wind. Will possibly require inclusion of radiative transfer, depending on level of ionization of torus.
  • Passed prelim.
  • Began application for NSF GRFP.

Meeting update...

  • Galactic code not constrained by memory… Running 105 particles

  • Data from new Hot Jupiter Run with solar ionizing flux values..

Meeting Update --09/13/16

  • XSEDE proposal
  • ThermalPulse module with high temperature inside the envelope.
    1. Overshoot expansion velocity problem with temperature inside the
frame 1 temp with DMcooling (minTemp 1000K) http://www.pas.rochester.edu/~bliu/Matthias/temp.png
overshoot expansion velocity http://www.pas.rochester.edu/~bliu/Matthias/overshoot_1000K.png
low temp velocity http://www.pas.rochester.edu/~bliu/Matthias/vel_DMcooling.png
  1. fixed two bugs related to the conical wind nozzle in August.
  2. waiting for confirmation from Bruce's new runs.

Some results from outflow testing.

The following are plots of the outflow kernel, that is, the profiles of injected density and velocity of the two-component sub-grid model:

In fixed grid we see a very nice outflow. Here is the density (rho_outflow=rho-rho_amb):

But in AMR, the outflow looks like:

This particle has a buffer around it of finest cells that is r=16 zones=kernel of outflow. We still see edge effects however. I tried making the buffer larger, to r=32 zones, but this doesn't impact the mesh strangely:

Looking at the outflow velocity (which is along z, the particle's spin axis), we have in fixed grid:

And in AMR with 'r_sink=32',

I also played with a collapsing BE sphere. Here is a movie of a marginally stable BE sphere that is perturbed to collapse by a density enhancement. The outflow of the particle is not fixed, and so I think it is aligning to random noise in the accreted spin ,which is why it looks off-center by the end of the movie. We should probably initialized the BE sphere with some rotation.

4AU binary simulation with great pulsation in the co-rotating frame.

This post analyze the 4 AU binary simulation with great pulsation in the co-rotating frame. The pulsation is from t=15yrs to t=25yrs which last for 10 years. The wind speed during the pulsation is 20km/s and the escape speed is 40km/s at that radii.

This image shows the density looking from top. There is a spiral structure that extends outward indefinitely. This structure may be a consequence of the great pulse or just the normal AGB pulsation. But a direct reason for this structure is that the secondary can not accrete all the gas that has been pulled to it. Therefore the secondary will periodically release some gas and absorb the rest.

I do not think the spiral structure is starting from the L2 point. Because the L2 point is not located on the high density arm that extends outward. As I can see, the gas that leaking from the L2 point is gravitationally bound.

This image shows another information. First I will define a quantity, angular momentum per unit mass:

where subscript b stands for binary. is the distance to the center of mass and is the velocity with respect to the center of mass in the lab frame. This value for such system is .

Then we calculate the actual angular momentum per unit mass in the simulation and name it . This image shows the value of . The greater the the higher the angular momentum per unit mass. We can see that this value is above 4 in most area in this image.

Next I calculated the mass of the gas that leave this system, it is . The accretion rate of the secondary is . Therefore the massloss rate of the AGB star is . In the single AGB star case, the massloss rate is around . So the existence of the secondary double the massloss rate of the AGB star.

I also calculated the angular momentum loss rate, that is, the angular momentum in the z direction that is carrying away by the runaway gas per unit time. It is . Hence the angular momentum per unit mass in the z direction in the runaway gas is:

compared to , it is:

Now, we can see that the runaway gas is carrying more angular momentum than the average angular momentum in the binary system. More specifically, it is 4.613 times of the average angular momentum per unit mass.

If we write in terms of , and , it is:

(1)

On the other hand, by the conservation of angular momentum:

where 0 stands for the initial values.

If we let this angular momentum and accretion run for under such condition, then at that time. Use equation (1) and substitute the correct numbers we can get that the separation will be 2 AU.

Researchers have put much attention to the angular momentum carried away by the gas leaking from the L2 point. The angular momentum loss in this simulation is different from that and it is more effective. However, I can not tell if this situation is going to last because it has only been 9 years after the pulsation. However, considering that the massloss rate for the single AGB star is low. I would argue that if the massloss rate of the AGB star is very high, such expulsion may last because the secondary can not accrete that much of gas. I will run the 4 AU simulation without the great pulsation and check whether if will also be there.

I think when there is no formation of circumbinary disk, there is strong orbital decay.

topdensity.gif

topdensityzoomin.gif

sidedensity.gif

6au simulation

I have done the 6 AU binary simulation. The circumbinary disk forms after I enlarge the AMR region. We can not see the formation of circumbinary disk and the accretion disk in low resolution simulation.

topdensity.gif

topdensityzoomin.gif

sidedensity.gif