Breakup, splashing, and wall film models #17
Conversation
… difference directory
| @@ -0,0 +1,5063 @@ | |||
| #ifndef MECHANISM_H | |||
There was a problem hiding this comment.
Feels like the chem files should belong in PelePhysics?
| @@ -0,0 +1,38 @@ | |||
| # AMReX | |||
There was a problem hiding this comment.
Part of me is a bit worried about all these Exec tests in PeleMP. I can easily see us changing things in LMeX or C and forgetting that these exist. Maybe we should think about moving these problems into the actual code bases?
There was a problem hiding this comment.
This is a large data file. Does it belong here?
|
Another thing that would be super helpful here because we're not as familiar with the models being implemented is some brief documentation on the theory/models, how to run them, and the expected results from a test case. |
| @@ -204,9 +226,17 @@ SprayParticleContainer::uniformSprayInit( | |||
| } | |||
| part_vals[SprayComps::pstateT] = T_part; | |||
| part_vals[SprayComps::pstateDia] = dia_part; | |||
| for (int spf = 0; spf < SPRAY_FUEL_NUM; ++spf) { | |||
There was a problem hiding this comment.
Was this removed intentionally? Test cases in LMeX and C are failing with NaNs due to these components not being initialized.
There was a problem hiding this comment.
I believe this was mistakenly removed and should be added back in. Good catch
There was a problem hiding this comment.
Thanks for the quick replies. This has been done in AMReX-Combustion/PelePhysics#452
| // | ||
| // Set the number of particles per parcel | ||
| // | ||
| pp.query("parcel_size", parcel_size); |
There was a problem hiding this comment.
The test cases in LMeX and C are set up to have parcel_size of 10 and 100, so this change, which effectively makes parcel_size always 1 without splash/breakup models, leads to different results for those tests. Running the old MP code with a parcel_size of 1 results in no diffs vs. this code. Just want to make sure that this is the correct anticipated behavior after this PR, and we can simply remove the parcel_size input file entries going forward.
There was a problem hiding this comment.
With these changes, there should not be a parcel_size input file entry. Instead, the parcel size is set based on how the particles are initialized. For jets, it is an input for the particular jet, and for uniform initialization the parcel size is set as an input here.
PeleMP/Source/PP_Spray/SprayInjection.H
Line 207 in 81ae98e
The correct fix would be to have a problem specific input from the input file for the uniform spray cases that gets read in and passed through here.
There was a problem hiding this comment.
Ok. Will leave this alone for now because the cases can still run, but will look into adding this later
This PR contains