Processing line sampling data

[joejoeyjoseph]

Hello, I am trying to generate and analyze the data vertically at a point, and I turned on the LineSampler option. After the simulation is finished, in /post_processing/sampling0xxxx, I notice different types of binary files exist (DATA_0xxxx and Particle_H). I am unsure which files contain the line sampling data, I cannot differentiate the two points/labels I specified (ls1 and ls2 below), and I do not know how to use the binary files efficiently. Please kindly give me some suggestions.

The following is the setting in my abl file:

incflo.post_processing = sampling
sampling.output_frequency = 3000
sampling.output_format = "native"
sampling.labels = ls1 ls2
sampling.fields = velocity temperature

sampling.ls1.type = LineSampler
sampling.ls1.num_points = 200
sampling.ls1.start = 500.0 500.0 0.0
sampling.ls1.end = 500.0 500.0 1000.0

sampling.ls2.type = LineSampler
sampling.ls2.num_points = 200
sampling.ls2.start = 2800.0 2800.0 0.0
sampling.ls2.end = 2800.0 2800.0 1000.0

Thank you in advance.

Joseph

[sayerhs]

@joejoeyjoseph The data is saved in native AMReX particle format. You can use the python script in the tools directory to extract the data into a Pandas Dataframe.

from amrex_particle import AmrexParticleFile

# Load data from 1000th timestep in  the run directory (containing input file)
pfile = AmrexParticleFile.load(1000)

# get dataframe
df = pfile()

[joejoeyjoseph]

Thank you!

[joejoeyjoseph]

Hello, I tried to use the AmrexParticleFile function, and it gave me errors when I was loading the data. I figured out that the error is related to pidx and the offset between different sample towers in amrex_particle. I tried to debug it but I could not completely eliminate the error.

I uploaded a directory of my sampling data at
https://github.com/joejoeyjoseph/awaken-code/tree/main/eagle/sampling60000

Please let me know. Thank you in advance.

[sofYarikova]

Hi @sayerhs!
I am trying to turn on the LineSampler option and Averaging.
The settings in my input file:

incflo.post_processing = averaging line_sampling
averaging.type = TimeAveraging
averaging.labels = means
averaging.averaging_window = 10.0
averaging.averaging_start_time = 0
averaging.means.fields = velocity
averaging.means.averaging_type = ReAveraging
line_sampling.output_frequency = 10
line_sampling.labels = line0
line_sampling.fields = velocity_mean
line_sampling.line0.type = LineSampler
line_sampling.line0.num_points = 100
line_sampling.line0.start = 0.1 0.0 -0.1
line_sampling.line0.end = 0.1 0.0 0.1
line_sampling.output_format = ascii

But after the simulation in finished files there are -nan velocity, i.e. I think that field was created, but nothing was determined there. Why might this be so?

[sayerhs]

@sofYarikova I am tagging @tonyinme as he wrote the time averaging code and has been using it for his work.

I don't see anything obvious that could be wrong in the snippet you posted. I have a few things that I would suggest that we check to understand the root cause. If you have already done these steps, can you share your full input file so that we can recreate the error at our end.

• The log file doesn't show any solution blow up in the equation residuals?
• The velocity field itself doesn't have any NaNs?

[tonyinme]

@sofYarikova , I agree with @sayerhs , please let us know if the solution itself if giving NaNs. If not, can you please share the input file?