(4) What is the fundamental difference between the TKE tarnsport eqns in kEquation (LES) & ~KOmega~ kEpsilon (RAS)? Can I get RANS from kEquation?

"get RANS" is a bit... Wrongly said, but anyways

One simple way to test this "theory", is to use PIMPLE and Co >> 1.

oh dear, that's a lot to digest, it'll take a while

This is something I've also been wondering, as kEqn looks a lot like epsilon expression plugged back into tke.. in the numerical implementation sense I could see there's no linear fluctuation term in RANS implementation of Reff, while those terms wont disappear in LES, so we will need to account for one uu, and one Uu interaction.

my guess is that a grid that's too coarse for DNS and no turbulence model won't find a solution at all, like running a 'laminar' case where turbulence is expected. both LES and RAS will model turbulence each in its own way. can't tell about the spectrum though...

Define "not find a solution at all"

explode 🙂

You can perfectly run coarse DNS simulations and get results, sometimes even quite acceptable

(please bear in mind my theoretical knowledge is about an Ångstrom deep)

Suggest the following reading then : https://www.sciencedirect.com/science/article/pii/S0021999117304059?via%3Dihub

I think they even use FOAM for this paper

These are expected to be accounted for by the two coefficients of the model. The dynamic version alleviates this a bit by filtering over a "test" grid

I always get better results with a LES without any model than with Smagorinsky (not the dynamic)

and I think thats funny because a lot of commercial CFD tools recommend to use Smagorinsky

Smagorinsky is a family of models, not just one.

lagrangian Smagorinsky is the de-facto model in atmospheric flows with LES

i can only find Spalart-Allmaras RAS and DES, where the difference is in the d. Am I on the right track? Can't find what's the meaning of y here: https://www.openfoam.com/documentation/guides/latest/doc/guide-turbulence-des-spalart-allmaras-des.html

this might be due to the relatively strong dissipation in OF. On our chair such results have also been shown for channel flow

without subgrid scale model was better because there was already enough numerical dissipation so no subgrid scale dissipation was necessary

Yeah you're on the right track Original SA uses distance from the wall to calculate nut The DES formulation works by just modifying that wall distance term (the length scale) You can just read Spalarts 1997 paper on it if you're interested, he sets things out in a pretty easy to follow way

cool, i'll do that

I have a question regarding the usage of K omega SST model with transition (4 equations)

I know it would be very useful in predicting laminar-to-turbulent BL transition when Reynolds number is in the order of 1e5

I have a rotating airfoil think of wind turbine, where the Reynolds number based on the chord of the airfoil is only 40 000

Do you think the 4 equation SST model could be beneficial in such cases?

does it make sense to use it in the first place?

This is not a particularly helpful answer, but aerofoils have been modelled once or twice over the years - there's probably a lot of literature out there comparing different models. Maybe look up the original paper for the model, they might have even used an aerofoil case to prove it works

Thank you

Yes, I am looking my way through several papers