Help with postprocessing multi-zone case.
 

Attachment 88418

Attachment 88419

Attachment 88420


I'm attempting to post-process a case to prepare it for adjoint calculations and wish for some "hand-holding." I'm looking for a place to start understanding and analyzing what is going on, using a known-case that works.



This base case involves a hub, nacelle, propeller, and wing. I have attached the main configuration and the sub-zone configuration files here. I believe I have the results of running this preliminary case, but I don't really know where to start in terms of post processing this.


I have a .vtu file for each 500 iterations, for each surface zone, up to files numbered 69000. Zone_1.cfg appears to be the zone with motion and zone_2 is static.


I'm looking at the trend lines of the residual data and it looks like things have mostly settled but there is a some up-tick of the line at the end.


Thanks in advance.

I don't know what you're asking for. What is the next step that you would like to achieve?

I guess I'm looking for help with postprocessing. First, I like to spot check to make sure and confirm what I set as output is what I have actually stored on the cluster computers. Second, I'd like to develop some skills in loading the result files correctly and understand what I'm seeing on ParaView for a case like this. Then I'd like to understand the adjoint solvers a bit more.

Thanks in advance. I'm sorry of it sounds a little like to broad of an ask, but whatever the case, let me know your thoughts.

OK, so if I understand correctly, you have done the CFD simulation, you have the results and you want to make sure that they are OK? Is this also about how to effectively use paraview for postprocessing?

The first thing to do is check the residuals to make sure that the solution has converged. Can you share an image of the convergence?

Then, In paraview just open the most recent .vtu files in paraview and select the velocity or pressure as the coloring and first check that your boundary conditions are correct. You can select 'hover points on' (a small icon with a question mark on it directly above the render view window) to quickly get the solution values at a point.

You can also load all .vtu files over all iterations at once by selecting the 'main' .vtu file at the top of the file list in the file selection menu. Then paraview will load all files at once and you can 'play' through the results with the animation buttons.

Is this the kind of information you are looking for?

you have the results and you want to make sure that they are OK?
>> Correct.
... how to effectively use paraview for postprocessing?
>> Yes!
Can you share an image of the convergence?
>> See attached output1.jpg for the end of the .out file.

history_1.csv and history_0.csv (both screenshotted) show something similar.


The residuals don't seem low enough but maybe it's different for this type of simulation? What do you think so far?

How do you load all VTUs at the same time without downloading them off of the server? These are fairly large files. I don't see a "main" vtu file.

I'm going to download the most recent VTU for each surface and post it here. So far I see 4 of the most recent VTU files, two for each surface or zone:
flow_0_69000.vtu (~571 MB)
surface_flow_0_69000.vtu (~9 MB)
flow_1_68999.vtu (~183 MB)
surface_flow_1_68999.vtu (~2 MB)



Thanks in advance!!

you can also directly load your history.csv in paraview. (1) create a new layout tab and (2) select 'Line Chart' and (3) make the data visible for this viewing window in the pipeline browser. You can then check how many orders your residuals have dropped, assuming you saved the residual data.

When all data is stored remotely, you have to work with a paraview server on the remote machine. If the paraview server was set up by the administrator then you can simply use 'connect' : https://docs.paraview.org/en/latest/...alization.html

When your vtu files are numbered sequentially, paraview will recognize it as a sequence and group them. When you want to load a file, they appear grouped and you have to expand the file list to see all the files. files flow001.vtu, flow002.vtu and flow003.vtu will appear as

|> flow...vtu
If you do not expand and select the list, all files in the list will be selected and loaded as a sequence.

But I suggest to first check if the server/client setup works.