A binary simulation
There is a 1 solar mass AGB and a 0.5 solar mass secondary with 3 AU separation in the simulation. In this simulation, we include physics such as cooling, central luminosity, dust formation zone (LTE condition) and pulsating AGB star. The luminosity of the AGB star is 3500 solar luminosity.
The contour lines still have the meaning as before.
In this simulation, a gravitationally bound circumbinary disk appear and disappear from time to time. The pulsating AGB star extends its atmosphere periodically.
I think the most important parameters include:
- luminosity/mass
- binary mass ratio
- separation
Progress since 2015 summer
- Wrote "The creation of AGB fallback shells" and published on MNRAS.
- Went to WorkPLANS workshop in Leiden, it was a very enjoyable trip. I met some astronomers.
- Wrote "Three dimensional hydrodynamic simulations of L2 Puppis" and submitted it to MNRAS. The paper is in its second version.
- Pass the Ph.D qualify exam.
- Developed the pulsating AGB model, added cooling (400K - 2000K) to AstroBEAR.
Meeting update
Writing the MHD shock roerientation paper. Expect to have a first cut of the paper through the results section by the end of the week. Will like to discuss this draft next week and next steps for the paper.
Would like to look through the plots of the 2D runs in the meeting and decide on a set to focus on.
Oh, and reorienation did not occur as strongly in the 60 runs as we saw last week when I made the box big enough to contain all the field lines. At this point, am seeming the strongest reorientation occur in the beta=10 cases (not the beta=1 case like we saw last week).
Update 5/11/16
Started catching up a bit from the end of the semester. Read Schneiter et al. 2016 and Christie et al. 2016:
- Schneiter paper makes synthetic observations of Lyman-alpha absorption in tails created by interacting solar and planetary winds, with photoionization included. They have nineteen models of varying stellar UV flux (photoionization rate), stellar wind conditions, and mass-loss rate of the planet. Both the stellar and planetary winds are isotropic, and radiation pressure from the star is approximated by reducing stellar gravity. They find that by including photoionization, a smaller neutral tail is formed; they also find a lower time to a stationary state than in previous models without photoionization. The most absorption is found in the blue-shifted side, between -130 and -40 km/s, the extent of which is dependent on the mass loss rate of the planet and on the ionizing flux. By comparing their models to observation, the heat efficiency of HD 209458b can be predicted to be less than 50%. In addition, it can be seen that the observed Lyman-alpha absorption does not necessarily require charge exchange to accelerate the neutral hydrogen sufficiently.
- 2.5D spherical simulations of planetary and stellar wind interactions, including charge exchange, were performed. Density was fixed at the base of the planetary wind and an inflow boundary condition on one half of the simulation served to emulate the stellar wind. In addition to charge exchange, advection, photoionization and recombination, and collisional ionization were included. The escape parameter lambda was used to categorize the models; it was found that there were two distinct regimes, with a transition region between. With lambda ⇐ 4 (high planetary temp), the planetary wind becomes transonic before colliding with the stellar wind, creating a large tail that takes a significant amount of time to mix. With lambda ≥ 6 (low planetary temp), the planetary wind has no chance to become transonic before it encounters the stellar wind, and the winds interact turbulently rather than collide, resulting in a well-mixed, barely evident tail. The transition region between these is also shown clearly in the calculated mass-loss rates of the simulations.
Meeting Update 05/11/2016 - Eddie
- Finished thesis last week! woohoo!
- Clumpy paper and response to referee is written. Ready to resubmit?
- Cooling paper is written. Need more details on other codes. Pat will add some writing in 1-D radiative shocks section?
- Resumed 3-D pulsed jet sims. Hydro and beta = 1 runs are done. beta = 5 is 95% done, beta = 0.4 is 66% done and may take a few more weeks to finish on bluehive. Maybe only 2 weeks if I can get a reservation for more nodes?
- Finished poster for next week's HEDLA conference:
Meeting Update 05/11/2016
- Wire Turbulence
- fixed a bug in the script extracting spectra data and so the Box 1 results. The updated results for frame 195 can be found in blog:bliu05032016 . x-y axis are both in log scale except for Box 1 which has a different x-range for showing the delta function…
- Results for other frames
Box 1 | Box 2 | Box 3 | |
frame 1 | ![]() | ![]() | ![]() |
frame 3 | ![]() | ![]() | ![]() |
frame 7 | ![]() | ![]() | ![]() |
short-range Box 5; short-range Box 9; short-range Box 10
- Debugging code for Bruce's module (#445)
Some thoughts on oblique shocks (both hydro and MHD)
1D Hydro Oblique Shocks.
Here is the pseudo color plot of rho with velocity vector field:
Here are lineouts of fluid vars:
The initial conditions have no gradients in y (the Riemann discontinuity is in x). This means the d/dy term in the y-momentum equation go to zero, and py is essentially advected through the grid passively (this would explicitly happen if nDim=1 in the code). There are periodic boundary conditions in y, as well, so this is effectively an infinite 1D problem.
Now, because there are no pressure gradients in y, we would not expect vy to change across the outer shock front. Instead, the only thing that could happen is a change in the x quantities. This is consistent with the lineouts above.
Also interesting is how the component of the velocity perpendicular to the shock front (vx) goes exactly to zero across the shock. At first I thought this must again be because there are no gradients in the setup that would drive changes in vy across the shock. So, if vx did not go to zero across the boundary, vy would change. Thus, vx necessarily must go to zero across the shock, regardless of mach, angle, etc. However, upon further thought, I don't think this is valid. Vy is a parallel component to the shock, and thus should be constant across the shock no matter what. So why does vx go to zero?
1D MHD Oblique Shocks.
The case is a little different in MHD. In MHD, each of the fluid velocities are kept in the Euler equations for our solvers, even when nDim=1. Is this because now, there can be gradients in y generated from the field, even when none might exist initially? We certainly see different behavior in the lineouts for the MHD case. Where there was no change in vy across shocks in 1D hydro, there are changes in vy across 1D MHD shocks:
Wire Turbulence Spectra-MHD
Implemented the Spectra object with 10 boxes/windows (each with size of Ly or Lz or 1/10Lx; the wire is around the center of box 2) in the wireTurbulence module. Worked on scripts to extract the data out and make plots. Here's the testing results of frame 195 for the MHD runs for each box:
box | spectra | zoomed-in |
1 | ![]() | ![]() |
2 | ![]() | ![]() |
3 | ![]() | ![]() |
4 | ![]() | ![]() |
5 | ![]() | ![]() |
6 | ![]() | ![]() |
7 | ![]() | ![]() |
8 | ![]() | ![]() |
9 | ![]() | ![]() |
10 | ![]() | ![]() |
Some filament/magnetic field observations
The integral-shaped cloud, and field
The following images depict the integral-shaped filament in Orion-A, and some have corresponding magnetic field vectors overlaid. One hypothesis is that this filament is wrapped by a helical magnetic field.
This filament/field morphology also bears striking resemblance to an oblique, reoriented colliding flow (60-degrees):
Herschel view of Taurus
Here is another example of a well-studied filament and corresponding field topology. I am pretty sure I've read that the uncertainty on the field vectors increases in the densest regions of the gas.
Reorienting MHD Colliding Flows
MHD, w/cooling, mach=1.5, beta=1, beta_ram=3.8
30-degree angle
(Movie here)
Here we see that the outer shock stalls on the top right and bottom left of the collision interface. This has the effect of 'reorienting' the shock front over time, so that it becomes normal to the magnetic field. Internally, it takes much longer for the internal collision surface to reorient, and then, it only becomes normal to the flows near the flow/ambient boundary. That happens to coincide with the field becoming parallel to this inner shock layer. Late in the sim, we also see what appears to be magnetic field waves, leaving the collision region and heading toward the flow boundaries. The entire interface then becomes unstable, and begins to break down.
Note also that while the 1D solution (without cooling), as well as the cooling case of this run, show velocity vectors going to zero across the slow shock, the present case with cooling shows strong velocity vectors 'falling onto the main filament'. This must be do to the cooling/compression of the gas. This motion in the cooling case leads to a more kinked field.
60-degree angle
Wow, we get much stronger re-orientation in this case, and in fact, it seems from a slightly different mechanism — one in which the deformity of the field plays a big role. Notice in the below movie of rho, that instead of 'blow-out' regions from the central shock zone, we start to see the flow turn around and re-enter the collision region! The flow is following the strongly deformed field. The flow vectors seem to be pushing on the central interface, causing the entire thing to reorient.
The field structure is also fascinating to watch. Over time, we see a central 'line' of black appear… this is the strongest field strength in the plot.
It is interesting that despite the cooling, the shock fronts are so nicely supported. Clearly, they are magnetically supported.
MHD, w/cooling, mach=10, beta=1, beta_ram=38
The strong shock creates strong post-shock pressures that cause a fast ejection of material from the collision region. Because of the strong shock, we also get strong compressions and thus, strong cooling. This reduces the thermal pressure in the collision zone, which triggers thermal instability, and subsequent NTSI, KH, RT modes. The fast shock is not longer supported by either thermal or magnetic pressure, so it collapses down on the thin boundary. If the field were stronger, we would likely see more magnetic support and thus a stronger fast shock front.
Despite the ripples of the collision interface, we do see reorientation occur. While the instabilities are growing in the center, you can see the field 'ballooning'. Oppositely directed field comes together, and looks to dissipate in strength (go from dark to light in color).
MHD, w/cooling, mach=1.5, beta=10, beta_ram=38
These are the same parameters as the 3D MHD colliding flows paper. In that case, the 30 degree shock didn't seem to reorient, but the 60-degree shock did.
30-degree angle
Weakening the field leads to less reorientation. Both the outer shocks and the inner, do not reorient as much as the beta=1 case.
We also see that the ejected material pushes the field out to a further radius, which is expected. This is the same as saying the flow can escape the collision region easier in this case, and this is why we start to see a straightening of the upper and lower 'knobs' of ejected material. They are the only parts of the flow that reorient, and become normal to the flows. The internal collision layer doesn't reorient
movie of rho movie of field
60-degree angle
MHD, w/cooling, mach=1.5, beta=.1, beta_ram=.38
Keeping the mach constant, but increasing the beta, we see that the field is strong enough to resist significant ejection from the collision region. Without this ejection, the outer shocks do not stall, and therefore, we do not see reorientation occur.
MHD, w/cooling, mach=10, beta=.1, beta_ram=3.8