Add automated test for ion impact ionization #5761
Conversation
|
Just a side comment, as far as creating the checksum file is concerned, I think this could be a good opportunity to verify that the steps in our how-to documentation (https://warpx.readthedocs.io/en/latest/developers/how_to_test.html#how-to-add-automated-tests) work properly, in particular step 7. below (I'm not sure if developers actually do this, instead of copying existing checksums and then overwriting them):
|
Signed-off-by: roelof-groenewald <[email protected]>
|
@EZoni I was able to reset the benchmarks using the instructions you linked 🎉 . @RemiLehe I modified the scattering logic to avoid the Btw, I checked that the electron impact ionization test still match theory (as close as it did before), and it looks good: |
@RemiLehe are you planning on updating the analysis script to get the theoretical ion impact ionization rate and the effect on the ion temperature? |
roelof-groenewald
added
bug
| @@ -21,7 +22,7 @@ warpx.const_dt = 1e-09 | |||
| geometry.dims = 3 | |||
| geometry.prob_hi = 0.1 0.1 0.1 | |||
| geometry.prob_lo = 0 0 0 | |||
| amr.max_grid_size = 8 | |||
| amr.max_grid_size = 4 | |||
There was a problem hiding this comment.
I noticed the test runs on 2 procs but with max_grid_size=8 there was only 1 tile, so parallel execution was not really tested.
Source/Particles/Collision/BinaryCollision/DSMC/SplitAndScatterFunc.H
Outdated
Show resolved
Hide resolved
Signed-off-by: roelof-groenewald <[email protected]>
|
The DSMC ionization tests require BLAST-WarpX/warpx-data#6 to be merged to add the necessary cross-section files. Now using the hydrogen cross-sections and having turned off the field solver in the test, we get pretty close match to the theoretical model: |
|
Relaunching the CI tests, now that BLAST-WarpX/warpx-data#6 is merged. |
roelof-groenewald
changed the title
| 3 # dims | ||
| 2 # nprocs | ||
| inputs_test_3d_ionization_ion_dsmc # inputs | ||
| OFF # "analysis_ionization_dsmc_3d.py" # analysis |
There was a problem hiding this comment.
Could we turn the automated test on here?
There was a problem hiding this comment.
The analysis would have to be modified substantially. The current test implicitly assumes that the neutral is much heavier than the impacting particle - a good assumption for electron impact but not for ion impact. So I'd say we should merge this as is and then possibly add a similar but modified analysis for the ion-impact case in a future PR.
TODO: update analysis script and recreate checksums