[rolando]

Hi all,
comparing the RANS-solver turbFoam and the LES-solver oodles there is one eye catching difference.
In the LES solver the sgs model is calculated at the beginning of a time step and in the RANS solver the turbulence model is calculated at the end of it.
Does it make a difference at all where I place the turbulence/sgs model?
What happens with moving meshes, when a phi correction is done at the beginning of a time step?

Rolando

[rolando]

Hi,
I want to write a hybrid solver, which can be used both for RANS and LES calculations.
Comparing the two basic solvers of that type, "oodles" and "turbFoam", I noticed, that in "oodles" the turbulence equations are solved at the begin of the timestep, whereas in "turbFoam" the equations are solved at the end of the timestep.
Does it make any difference where the turbulence equations are solved?

Rolando

[eugene]

No, not really. But why would you want to write a new solver? The only real difference between oodles and turbFoam is the turbulence models. In fact the SpalartAllmaras and oneEqEddy LES models are DES models which blend between LES and RANS modeling based on wall distance.

[rolando]

Hello Eugene,
that´s what I want to do.
I want to do DES calculations.
Further I want to do some calculations with the models in RANS mode.
There are also some small additional features I integrated in my solver. Having a single hybrid solver would ease the maintenance of the it.
Can you tell me, why it is a difference to calculate the equations at the beginning or the end of the loop?
Is the placement at the beginning all right for DES calculations?

Rolando

[eugene]

The only difference between placing it at the beginning or end of the loop is during the first timestep.

The question is simply, what do you think is better: calculating the flow first with rubbish turbulence, or calculating the turbulence first with rubbish flow field. Since everything is pretty rubbish at first I doubt it will make a significant difference.

I would use the oodles solver as my base simply because it is more advanced (uses the SIMPISO algorithm). This algorithm will probably be applied to turbFoam in due course.

[rolando]

Then it should be possible to do RANS calculations with the oodles solver, right?

What do you mean with SIMPISO algorithm? In the oodles solver there is the same PISO algorithm as in turbFoam.

Rolando

[eugene]

Yes, I dont see why not, as long as you use a "RANS" LES model.

oodles has a SIMPISO/PIMPLE solver, dont let the PISO entry in the fvSolution fool you. It only does full convergence on the final corrector step and is about twice as fast as PISO as a result.

[rolando]

Thanks a lot Eugene,
I think with "RANS" LES model you mean, that the RANS model is derived from "LESmodel" and not from "turbulenceModel", right?

Rolando

[eugene]

That would be one way yes.

[rolando]

Thanks a lot Eugene.