Continuing with my quest: - Turbulence quantities change did not address the oscillations at the LE - the simulation converged to below 1e-06, so convergence is not the issue - wall boundary conditions were adjusted and resulted in no change - Discretization schemes did also not address the oscillation issues - I dont think fvSolution changes to the Matrix solvers can change anything, so I wont inspect that - I am guessing its a bad mesh now, first I am going to try a coarser leading edge seeing if that may fix the underlying issue, or atleast improve it somewhat - If this doesnt help I will try to adjust nNonOrthogonalCorrectors from 1 to 2 or even 3 and seeing if that fixes the simulation. My max non orthogonality is 57. If that doesnt help then I am all out of ideas and I dont have a deep enough understanding to really pinpoint what is causing this. Anyone else that could give advice?

Also your Quest deserves a separate thread

Probably not a bad idea, I agree

I haven't been keeping up with what you've been doing, but have you tried different wall functions?

So originally I was setting the wall for the turbulence quantities as: k - uniform 0 fixed value nut - nutUSpaldingWallFunction omega - omegaWallFunction U - noSlip obviously Now I switched to: k - kLowReWallFunction, which seems odd because according to literature the lowRe k BC is just uniform 0, which makes sense. Now I have a k-distribution across my surface mesh. nut - nutLowReWallFunction, which just sets it to 0 omega - omegaWallFunction

Almost looks like TS waves lol, but it's not that

This didnt really change anything

Are the waves steady or transient? Like, does the pattern stay the same at different iterations

For k I am still unsure if... having k values across the surface of magnitudes if around 1e-04 is realistic. I mean its similar to my inlet conditions but yeah its just odd considering k=0 makes sense. Omega is also an interesting one, I get magnitudes going to 1e+08, with the magnitude really depending on the refinement of the surface. Although it is specific dissipation rate, so I guess it would make sense.

That's what I meant by TS waves

oh you mean like transition

I see, but yeah I have no model implemented to predict this

But if the waves are constant then it's not some sort of transition, it's most likely a mesh thing

My best go to now is trying to increase the BL size again and seeing if 20 layers then behave better, even if my y+ goes up a little bit

I mean for fun you could try a transition model

I probably will eventually to compare it to a single state and seeing how things change. But it also means I need less ambigious turbulence BC's

What does a slice of tke and nut look like through the full domain? I find inflow values can drop off surprisingly quickly, you might find that the turbulence hitting the wing is much less than specified at the inlet

I know, and thats definitely happening as I dont have decay control set. But those shouldnt influence the simulation too greatly imo. I was also doing a test with ridicoulusly high turbulence values and they had similar effects

Yeah what I'm wondering is whether it decays so much that the inflow makes no difference, it decays to zero either way?

hm, not necessarily? Spalding did a study on this a while ago.

If you're following the literature it's probably fine

For my inlet turbulence im sitting at omega = 80, k = 0.0006, which gives me a viscosity ratio of 0.5

I'm surprised how common it seems to be that people do nothing to stop the turbulence decaying

seems to not be a hugeissue I guess

I'm getting my gut feel mixed up a bit between epsilon and omega, but that seems like a very high omega

typical for atmosphere apparently, Spalding suggets for general use of typical external aerodynamic applications with two equation turbulence models: k/U^2 = 1e-06, so k = 1e-06 * U^2 and omega * chord / U = 5, so omega = 5 * U / chord

Yeah okay, all seems fair enough then

Im referencing "Effective Inflow Conditions for Turbulence Models in Aerodynamic Calculations - Philippe R. Spalart, Christopher L. Rumsey 2007"

Think I've read it before but can't remember it

Its pretty much my only reference point for the turbulence quantities

You've probably checked already, but any correlation between the wave pattern and the triangles in your STL?

my STL is like 800 Mb big and has a very high LE refinement

Yeah if it were me I'd be adding source terms to maintain the free stream values, but that's probably more time consuming that you're willing to get into

Its not, OpenFOAM has it built in

Didn't work for my case when I tried it

oh

But maybe my case was different somehow 🤷

Okay I did recheck STL and surface mesh again, and in some areas behind the LE the tri's of my STL do become bigger than the cells of my surfaceMesh, I guess I can try to refine the STL even more

50gb STL here we come

I mean I feel like this shouldnt be that big of a problem but yeah, its snappyHexMesh so who knows

Do the triangle edges line up with the waves at all?

hehehehe

Yup that fixed it, still some issues with generating the STL as my pc just crashes with salome with even higher tessellation, but w/e. There is still a dip of shear stress on the wall on the LE but this is very likely due to the jump in layer height caused by the underlying meshing method. So not much I can do about it.

I'm not sure what you're referring to - what does the Y+ look like?

The y+ looks totally fine, but there is a dip at the leading edge between the spike at the nose and the transition into the flat region which I cant explain.

Yeah so what is the y+ either side of the transition

lemme draw a quick paint showing what I meant lol