Refactor SST model implementation and addition of variants/corrections

[WallyMaier]

Proposed Changes

This PR refactors the SST model implementation to include (hopefully) the KL, RC, V, m, SUST models.
Mainly for the models SST-V(m) and SST-Vsust(-m). These tend to be used in higher speed flows. https://turbmodels.larc.nasa.gov/sst.html

A quick summary of some of the changes are:

Adding the"V" term simply changes the P-term from
Eddy_Viscosity_i * pow(StrainMag,2) to
Eddy_Viscosity_i * pow(VorticityMag,2)

Adding the "KL" source term changes the P-term to
Eddy_Viscosity_i *StrainMag*VorticityMag

Adding the "RC" term:
Multiplying the P-term by a function, f.

Many of the Options can be used together -> SST-V-sust-m. The NASA TMR doesn't explicitly mention something like SST-2003-V-RC-sust-m,so I don't know if that is a valid permutation. That being said, the amount of options is fairly so I'm not sure the best way to approach this formulation, without unnecessarily bloating the code. Hence this isn't operational yet. Work will need to be done to validate the addition.

Related Work

This is part of the larger turbulence fixes/additions/modernization efforts seen in #1557. It also addresses some comments in #1551 and #1413

Also as @pcarruscag pointed out below, this falls out in the larger discussion mentioned in #1364.

PR Checklist

• I am submitting my contribution to the develop branch.
• My contribution generates no new compiler warnings (try with the '-Wall -Wextra -Wno-unused-parameter -Wno-empty-body' compiler flags, or simply --warnlevel=2 when using meson).
• My contribution is commented and consistent with SU2 style.
• I have added a test case that demonstrates my contribution, if necessary.
• I have updated appropriate documentation (Tutorials, Docs Page, config_template.cpp) , if necessary.

[pcarruscag]

The direction we are going is to have turbulence model variants/corrections specified according to the strategy in #1364

TURB_MODEL= SST
SST_OPTIONS= 1994/2003, m, V, KL, RC, sust, UQ

The code change for V is the 3 lines below, if you want to start the implementation of "SST_OPTIONS" that would be awesome.

[WallyMaier]

Ill go ahead and start the SST_OPTIONS option. I imagine with the amount of variants/corrections, we will need to follow the SA implementation style from @suargi ?

Its a bit tricky since it seems like "m" can be added to all the mentioned models: base/standard model, SST-V, SST-SUST, KL, RC, etc. Also it seems like SUST and V are can be modifiers as well.

[bigfooted]

I moved some things around to make it run. I also added sst1994, sst1994m,sst2003,sst2003m.

[bigfooted]

typical result of comparison of original SU2 (1994m) vs SU2-V2003m. The difference (< 2%) is due entirely to the change from vorticity to strain in the computation of muT (CturbSSTSolver.cpp)

[bigfooted]

NOTE: for this result, the finest grid was not computed yet
For bump in channel: same trend. For this case, V2003m results are available of FUN3D and CFL3D. Here, the FUN3D and CFL3D results of V1994m are still shown, but the difference between V2003m and V1994m is almost negligible. This does mean that SU2-V2003m does not correspond as well to the FUN3D/CFL3D results as the SU2-V1994m results

[pcarruscag]

It would be nice to match their results, but I think ours look better, i.e. seem to converge with grid size.

[WallyMaier]

The presented 1994m were results from the "bad" 1994 version in SU2?
Also, I am trying to find some cluster space to run the SWBLI case to check vorticity model.

[pcarruscag]

That case has some problems, probably because of #1631 (supersonic inflow is also affected)

[pcarruscag]

There is a shock near the start of the plate, and it's vertical position changes 1:1 with mesh size

[WallyMaier]

I can just run the flat plate with the V and KL options, if that works.

[pcarruscag]

Up to you but given that it's a one line change... I would give more priority to resurrect the V&V pages for the plate and bump cases, and adding something to the vandv.py regressions.

[pcarruscag]

Nice 👍
Please cleanup the configs from comments and for V&V regressions all cases should be restarted from a converged solution, we want to track the values at convergence.

I'm working on having the SA options in the same format as SST btw.

[bigfooted]

ok, great, I've modified the python scripts for the postprocessing, I'll update them as well (upgraded to use pandas framework)

[aeroamit]

There is a shock near the start of the plate, and it's vertical position changes 1:1 with mesh size

I had also tried this case sometime back with SU2. Faced the same issues as reported by Pedro. Also, SST was not tallying with the results presented in the reference (https://www.grc.nasa.gov/WWW/wind/valid/m5swbli/study02/m5swbli2.html). There seems to be a shift of 5mm. It will be useful to sort out this case...
See some details in the attached pdf file...
SWBLI_case_issue.pdf

[pcarruscag]

Do you want to do the honors @bigfooted ?

[WallyMaier]

This looks good to me! Thanks for taking the lead @bigfooted and @pcarruscag!