So a subgrid RSM model?

Or do you mean to calculate UPrime2Mean by any chance?

Because when you put it the way you put it, it sounds like a custom turbulence model which should feed back into UEqn

not a subgridmodel. the current idea is to postprocess the transport of the reynolds-stresses - the result could be the time averaged solution of the reynolds stress transport equation in the domain

Wait why are you calling it transport, instead of calculation. You already have LES for transport

Does the fieldAverage object with UPrime2Mean with a custom window not give Reynolds stresses?

The resolved fraction of it atleast...

Seems like a lot of extra equations to solve just to get to the same result as what UPrime2Mean gives, no?

Normally you calculate TERMS of the Reynolds Stress eQuation in order to get some structure from your turbulence field. Solving the whole REQ is expensive.

Please show us what you want to solve

Notice also that you dont solve the material derivative normally, REQ is derived directly from NSE

True.

I need to look into my project again check the main goals. As I am currently traveling that will take until Wednesday 😦

Hello. Does anyone know why in the kOmega SST implementation the production term is written like that ? : alpha()*rho()*gamma*GbyNu0 - fvm::SuSp((2.0/3.0)*alpha()*rho()*gamma*divU, omega_) How do we get that ?

it's fine I get it

Hello turbulence guys ! What do you think of using inlet mapped BC for U, k and omega to get fully developped inlet condition ?

for example in a case of a pipe flow

Mr. Guererro from WolfDynamics does that and it seems to work pretty well

On meshes where there's no place for flow to steady

Does it use also for k, omega, epsilon or R ?

From my test it seems to work

Yes, epsilon/k/omega/U

ok !

I'm in the process of doing a Sparlat-Allamaras DDES simulation and I have a question regarding wall functions. I was wondering if the turbulence model use RAS automatically if y+>1? How does this work?

I'm wondering if there is some sort of guidelines in order to choose the right DES turbulence model for the type of geometry, nature of the flow, etc., that I want to study?

For a DFSEM inlet condition for a LES simulation, is there any reason to change the "nCellPerEddy" option to anything else than 1 ? I assume it has something to do with the density/number of the synthesized eddies?

It puts a lower limit on the size of the generated eddy

Afaik it should be more like 2-4

If an eddy is the size of one cell it can't be convected

Alright Thanks !