Meeting Update 06/26/2012 - Eddie

I ran some very low res sims for the 3D cooling jets to make sure the problem worked and I had the dimensions right. The left image is what it looked like at frame 70 (100 frames total). I made the domain a bit bigger in all directions so that frame 100 now looks like the right image.

movie​

Since these test sims looked okay, I submitted 4 jobs to bluegene this morning (Monday, 6/25). The wait in bluegene looks like it's about 3 days, so I guess I'll have to run the analysis when I get back from vacation.

With the ambient density held constant at 10 cm-3, I change the jet density to change the cooling strength chi. When chi is really small, this implies very strong cooling, and it becomes difficult to resolve the cooling region. All runs have a jet velocity of 250 km/s and a jet radius of 2e16 cm. Here are the varying parameters for 5 different runs:

n jet eta chi required x res
0.1 0.01 0.0015 78,335
1 0.1 0.034 3495
10 1 0.358 335
100 10 1.367 88
1000 100 2.428 50

The required resolution in the x-direction is based on a desired resolution of 10 cells per cooling length. You can see that the low chi runs require a lot of resolution, but there are ways around this. I can probably come up with a clever way to only track the jet head and region immediately behind it. For now, I did not submit a job for the first run in the table (njet = 0.1), and I threw some AMR levels at the other runs.

So here is my plan for when I get back:

  • analyze these runs that I have submitted
  • code something to force refinement only where I need it (this will help with low chi sims)
  • run some sims with Francisco's parameters (Adam, can you forward this info to me please?)

Attachments (3)

Comments

No comments.