adds option to output velocity gradient
All threads resolved!
All threads resolved!
This expands get_rfields with the option to output velocity gradient as well.
Merge request reports
Activity
requested review from @g-lbent
assigned to @chichi
@g-lbent : please test this by (1) calling the
get_rfields
tool with all output options turned on and (2) writing a python code from scratch that reads velocity gradient, vorticity and TrS2, and then compares vorticity and TrS2 from file with (python-side) newly computed vorticity and TrS2.- Resolved by Cristian Lalescu
It's not working, yet, I believe.
- Iterations > 0 of my test simulation match very well in TrS2. Vorticity matches with some error bars (atol=1e-5, rtol=1e-5) but only up to a sign. I'm computing vorticity as
vort[..., 0] = grad[..., 2, 1] - grad[..., 1, 2] vort[..., 1] = grad[..., 0, 2] - grad[..., 2, 0] vort[..., 2] = grad[..., 1, 0] - grad[..., 0, 1]
- Iteration 0 of the test simulation (Gaussian initial condition) matches in TrS2 but not in vorticity. Could that be related to some incompressibility issue?
- I am getting this warning on running get_rfields: "file /localdisk/bt307719/simulations/test_rfields/test_checkpoint_0.h5 does not have velocity group" It seems a bit strange to me that for this postprocessing, a test_fields.h5 is created linking to the checkpoint file, and then the output of the postprocessing is written back into the checkpoint file. In this case, the extra fields file does not seem necessary.
added 7 commits
-
9ccab40d...415a97ea - 6 commits from branch
master
- 681fdd89 - adds option to output velocity gradient
-
9ccab40d...415a97ea - 6 commits from branch
mentioned in commit e76cf3b1
Please register or sign in to reply