So I've finished writing the ​Disk module tutorial that I mentioned yesterday at the meeting.
You are invited to give a glance to the page and let me know if you think anything should be added, removed or changed to make it better. You can also directly edit the page if you need to make minor changes.
Any type of feedback wanted !

So…. I was looking at Erica's BE setup and decided to start advecting the stable sphere and ambient - to test the stability/accuracy of self-gravity… First I noticed that occasionally the value of PhiDot would be very large… This was because a root step would finish just before a frame dump - so the code would take a very small time step - this would cause noise or small fluctuations in the potential to generate very large time derivatives… If these large derivatives are then used for a normal size time step - they will produce drastic errors… This is because of the drastic extrapolation in time of the potential. One way to mitigate this is to avoid having to take very small time steps by anticipating such scenarios and taking two half-size steps instead of a full step and a tiny step.

(Errors end up being amplified by the ratio of time steps)

Fortunately we no longer use old time derivatives for subsequent steps (except to set the initial guess for the solution vector which could make converging a bit difficult for visit) - but it does make the value for PhiDot plotted in visit look way off…

The other problem is regridding… As the grid adapts - regions that previously were refined will continue to have a correct fine-grid solution while regions that were not previously refined will have prolongated solutions… If a cell goes from having prolongated to a calculated solution - the time derivative will have a component in it that is just because of changes in the discretization of the grid - which could then be used to set boundary conditions for finer cells etc… Here the solution is to calculate the potential after regridding - before advancing - and then calculate it again after the advance to get a time derivative for the potential that is solely due to gas motions and not regridding effects… It is a little more expensive (though hypre should converge quickly since the initial guess will be fairly accurate) but well worth it in terms of stability etc…

• Worked on static AMR improvements. #242
• Found a bug with self gravity and sink particles.

movie​

• Updated my tickets report {7} to show due date.
• Noticed that data files explained section of wiki should be updated with golden release… Created an enhancement ticket #246 - but it is currently unassigned…

• Got PhD

• Golden Version Status:
  1. ​https://clover.pas.rochester.edu/trac/astrobear/wiki/GoldenVersion

• Tickets:
  1. New: #243 (heat conduction and resistivity solver), #244 (Segmentation fault with UniformCollapse), #245 (Installing libs and astrobear on Blue Streak—the Q)

• Yat-Tien from UCLA
  1. Will arrive at 630pm on Sep 12th
  2. Talk/dinner before training—asked by Yat-Tien (Volunteers needed)
  3. Prepare training (temporary account on local machines, modules)
  4. Yat-Tien's Project: ​http://www.pas.rochester.edu/~bliu/Yat_Tien_UCLA/Yat-Tien_UCLA.pdf

• Worked on: #237 (test modules), #240 (Molecular clouds), #244 (Segmentation fault with UniformCollapse), #245 (Installing libs and astrobear on Blue Streak—the Q)

What I've been doing:

Revised the ApJ paper: one more appendix on initial field setup, one more paragraph on instabilities with one additional figure to support it, misc modifications according to Eric and Adam's previous suggestions, and also the referee response from the HEDLA proceeding. (Both Eric and I thought it is good enough to submit, so waiting on Adam's feedback.)

Making test modules for the gold version: 50% done.

Adapting the multiphysics code to the gold version: currently working on the communication stuff of the subcycle control routines.

Resistive paper: low res runs are shown in a previous blog post, high res runs are currently running on bluegene (We could have some movies to look at but unfortunately I found something wrong in the job that was running last Friday. So I started everything over new, and they just started running this Monday). It looks like we can finish them by the end of September. We will do R=10,100,1000,infinite for both Bx, By configurations, and also one hydro case and one Mach 10 case (the jobs use the experiment setup data the same as the NLUF proposal, Mach is 5) for comparison's sake, according to Gianluca's advice.

Qual: I'm planning on an October date. Eric Blackman and Chuang Ren both have time and have agreed to help. Currently waiting for Eric Mamajek's response. (Dan will not be available during that time frame) I will probably start writing up the brief in early September.

After fixing bugs (previously it does not diffuse the field lines correctly due to a sign mistake) and adding the magnetic energy flux, the resistive solver work better now. About the energy flux:

Previously we diffuse the magnetic field according to the equation:

\frac{\partial \bf{B}}{\partial t}=-\nabla \times \bf{J}

But we do not explicitly treat the change of energy due to this movement of magnetic field lines. Therefore the total energy of one cell stays the same before and after resistive diffusion. This is not right since the total energy per cell will change (thermal energy stays the same). In the new code, we calculate the energy flux resulted from the resistive diffusion explicitly at the cell faces:

\bf{F_m} = \bf{J} \times \bf{B}

From this flux we update the total energy at each cell using the divergence law. With this calculation, the cell where field is diffusing away from does not magically get instantly heated up like before. There can be an inaccuracy resulting from the finite grid size during this calculation. This inaccuracy can be observed by looking at the thermal energy contained in one cell and test whether it's kept the same. At low resolution (16 zones per clump radii), this error gets to 0.3%. The solver is completely divergence free: during resistive diffusion, the magnetic field divergence is kept zero.

The next figure is the "Bx" (magnetic field parallel to the shock normal) case at around 3 clump crushing time scales. The Rm labeled in the figure is the magnetic Reynolds number. The plot at the bottom is the case where the fluid is perfectly conductive (Rm is infinite).

Here is a movie for it:
​http://www.pas.rochester.edu/~shuleli/Resistive_Clump/denpara_lr.gif

The following figure is the "By" (magnetic field perpendicular to the shock normal) case at around 3 clump crushing time scales.

Here is a movie for it:
​http://www.pas.rochester.edu/~shuleli/Resistive_Clump/denperp_lr.gif

We can see that although at Rm=100 still show quite a difference in the two field configurations, the difference is beginning to diminish. We will be doing production runs with high resolution (64 zones per clump radii) on the bluegene cluster in the next few weeks.

• I've been working on getting ready for my defense - Tomorrow at 2:30…
• I've finshed the talk - just need to run through it and time it a couple times

I've been working on the resistive MHD simulations for the next paper. Here's a image for the R_M = 100 case in low res:



We can see differences between this and the ideal MHD case (have to verify high res).

​http://www.pas.rochester.edu/~shuleli/RMag1/rm1.gif

We are heading towards an hydrostatic equilibrium solution.
There are 2 possible approaches, we are currently evaluating the best path to follow.

* Scenario 1 (Manual density profile):

User defined variables: density profile ⇒ rho₀*(r/r₀)-2
Pressure would be calculated accordingly by the hydrostatic equation.
Temperature can be changed by adjusting the density profile.

Pros:

• Is the classic implementation of the hydrostatic equilibrium
  
• Density profile is consistent with AGB stars models

Cons:

• P(r) involves integral calculation
  
• Scarce usage of astrobear functions, introduces a window for errors
  
• Temperature profile could not be what we want for our type of simulations
  
  

* Scenario 2 (Isothermal profile, suggested by Jonathan):

User defined variables: constant ratio (pressure/rho)
Pressure and density would be calculated accordingly

Pros:

• Usage of astrobear common functions, less possbility of errors
  
• Does not involve manual calculation of integrals
  
• Gives us the temperature profile that we need
  

Cons:

• Density profile could not be disk friendly

• Golden Version Status
  1. ​https://clover.pas.rochester.edu/trac/astrobear/wiki/GoldenVersion

• Tickets
  1. New: #241, #242
  2. Closed: #216, #241

• Worked on #231, #237, #241

Finished the second revision of the clump paper. Eric has commented on it after his second read, and his suggestions are put into the latest version. Currently waiting for Adam's second read and responses. Hopefully it can be submitted very soon.
​http://www.pas.rochester.edu/~shuleli/paperpics_clump1/Paper_shule_revised.pdf[[BR]][[BR]]

Currently working on the testing modules for the Gold version. I have generated the low res reference frames for the Magnetized Clump (using 3D toroidal aligned geometry) problem, and am working on the other problem modules since I'm not familiar with some of them.

Talked with Laura about the Qual. We need:
(1) setup the committee with four people: Adam, an experimentalist, a faculty member from the same area, a faculty member from a different area (can be from another department).
(2) setup date, time and room.
(3) submit a 10 page handout outlining the research and goals, each faculty member should get one 2 weeks prior to the actual Qual.
(4) prepare a presentation.

• Golden Version Status:
  1. ​https://clover.pas.rochester.edu/trac/astrobear/wiki/GoldenVersion

• Tickets:
  1. New tickets: #239, #240
  2. Closed tickets: #147, #171, #198, #202, #232, #233, #239
  3. ​https://clover.pas.rochester.edu/trac/astrobear/query?status=accepted&status=assigned&status=new&status=reopened&group=priority&col=id&col=summary&col=status&col=owner&col=type&col=priority&col=due_date&col=component&col=ttd&col=reporter&order=priority&report=28

• Worked on #231, #240, Gantt Chart plugins on wiki system

• Will be on vacation on Thursday (Aug 9th) and Friday (Aug 10th)

The quarter grid was pretty simple to set up…I changed the domain and boundary conditions in global.data. Also, a couple things for the shape refinement. Run A took just under an hour on 8 procs on alfalfa (those images are below). Runs B and C will need bluegene due to their very high resolution.

The resolution had to be altered a little bit, so just as a reminder here are the parameters for my runs:

Technically, the cells per jet radius increased for B and C in order to keep the proportions the same from sim to sim, but the overall domain is smaller so these should run faster now.

I need to spend most of my time studying for the prelim now, so I don't expect to be getting much research done for the rest of August.

• Magnetic tower paper accepted in ApJ!'''

• Spherical wind against constant vertical mag fields (from HELDA). 1st test with revision 951 ran well in Bgene: completed in 28hrs with 256 procs. ​http://www.pas.rochester.edu/~martinhe/2012/andrea6aug12.gif

Noticed quite a bit of protections in jets runs (AGN and the above ones) using a grid boundary in r951, which I didn't see before.

• Binary disks. 15au run added to the paper, but its disk outer radius -formed by light gas- is larger than the 10au case (paper Figure 4). Running 10au case with 4amr and latest module to make sure the disk sizes we're getting are physical. Added more text into the paper, will hand it to Jason next week while I'm in Mex. draft:

​http://www.pas.rochester.edu/~martinhe/2012/binary/paper2.pdf

• AGN jet traction by AGB winds. 3 HD sims are running (strong jet with AGB trajectories that cross both the jet's edge and axis; weak jet with AGB trajectory that crosses the jet's edge) with the latest revision and in Bgene/Bhive.

• Off to Mexico City from the 13th of Aug to the Sept. the 1st for US visa renewal.

• Adam: I need a recommendation letter for the Marie Curie Fellowship by Aug the 15th please.

New movies:

Here are larger resolved runs ~ 2x clump radius:

Matched case: ​http://www.pas.rochester.edu/~erica/MatchedJuly26.gif

Light case (1/1000th ambient rho): ​http://www.pas.rochester.edu/~erica/NewLight1000thJuly28.gif

Going into a hole for this upcoming month, studying for prelim. See you on the other side !

So far we have done different runs:

Batch 1 (4 runs):
Constant parameters:
Central mass = .4 M_Sun
Disk mass = 10 M__Jupiter
Disk radius = 2e10 cm
Disk height = 2e9 cm
Disk temperature = 4000 K
Ambient density = 0.001 g/cc
Variable parameters:
Ambient temperature = 1e6K | 5e6K | 1e7K | 5e7K
Preview:​Batch1_preview

Batch 2 (4 runs):
Constant parameters:
Central mass = .4 M_Sun
Disk radius = 2e10 cm
Disk height = 2e9 cm
Disk temperature = 4000 K
Ambient density = 0.001 g/cc
Ambient temperature = 5e6 K
Variable parameters :
Disk Mass = .1 | .5 | 1 | 10 M_Jupiter
Preview: ​Batch2_preview

In all the cases the ambient is attracted towards the central object, this generates velocity and temperature, and a heat bubble embeds the disk. A new hydrostatic equilibrium module (the current one implements constant pressure, which is not appropriate in our case) is needed to prevent the ambient from generating heat.