shift WindNinja momentum solver runs from using the kEpsilon model to using the RNG kEpsilon model

[latwood]

As we found in our soon to be published Perdigao paper, the RNG kEpsilon model performed better for turbulence outputs than the kEpsilon model, so now we are making RNG kEpsilon the primary run turbulence model for WindNinja momentum solver simulations in OpenFOAM.

It's straightforward to make the change in OpenFOAM 8, but may need a bit more tweaks for OpenFOAM 2.2.0, specifically with the inlet boundary condition application files.

[latwood]

Just reviewed this. This branch is fully ready to go for foam 8 and 2.2.0, though need to recompile the 2.2.0 BCs for the windows build with MinGW to truly go forward with it in the master branch.

In the meantime, we need to appropriately merge the master branch into this branch, then extend the same changes done to foam 8 to the foam 9 BCs. Then look over the foam 9 code to make sure that it is really using the values properly. The foam 9 code SHOULD behave the same as the foam 9 code, but just want to double check to be sure.

With these tasks completed, the code would be good to go, using hard coded values, where to swap back and forth between RNG-kEpsilon and standard kEpsilon you would have to recompile the BCs for the given hard code values, not just set the appropriate values in the NINJAFOAM directory.

However, some future work past this working point would be helpful. Future work would include tossing out some of the extra leftover not used constants from foam 2.2.0, setting up the BCs to be better about more consistently printing values, and setting up the BCs to just lookup the appropriate constants from the NINJAFOAM directory rather than having hard coded values. But this future work would need checked at each step of the way to make sure that it is all still working properly.

[ECheynet]

If the goal is to simulate natural wind flow in complex terrain, one could also consider the realizable k-epsilon model as another viable option. It is available in openFoam and has been used in a few studies [1,2].

[1] Blocken, B., van der Hout, A., Dekker, J., & Weiler, O. (2015). CFD simulation of wind flow over natural complex terrain: Case study with validation by field measurements for Ria de Ferrol, Galicia, Spain. Journal of Wind Engineering and Industrial Aerodynamics, 147, 43-57.
[2] Cheynet, E., Liu, S., Ong, M. C., Jakobsen, J. B., Snæbjörnsson, J., & Gatin, I. (2020). The influence of terrain on the mean wind flow characteristics in a fjord. Journal of wind engineering and industrial aerodynamics, 205, 104331.

[nwagenbrenner]

@ECheynet Thanks for pointing this out. Yes, we could also look into the realizable k-epsilon model performance. Sorry for my oversight in the original comment on #627 - I originally read "realizable" but was thinking "RNG". Thanks for the references. This is something we'll look into.

[latwood]

With this last pull request merged to the master, this leaves WindNinja running the various OpenFOAM versions doing the following:

• OpenFOAM 2.2.0 -- standard kEpsilon (for windows build)
• OpenFOAM 8 -- RNG-kEpsilon
• OpenFOAM 9 -- RNG-kEpsilon

We would update the windows build to use RNG-kEpsilon as well, but unfortunately we cannot easily recompile the old OpenFOAM 2.2.x binaries. So we will leave the windows build as is until we can find a workaround.

There is a little bit of cleanup work that could be done, just in comments (see commit 421a78e), and possibly updating the inlet BCs to autoselect the turbulence model coefficients rather than having to be recompiled with hard coded values. Also, I saw some old foam versions in header files which is small leftover cleanup related to the work of issue #458. But that work can wait until just before we are ready to try upgrading to an even later version of OpenFOAM for Ubuntu 24.04 builds.

I find the pull request #613 to be helpful to quickly see what work was done in the past for shifting from kEpsilon to RNG-kEpsilon.