Hi all,

could someone share his experience with the use of the V2F turbulence model with regard to the performance in separated flow regime and numerical robustness ?

Thanks,

Mark

(1). What is V2F turbulence model? (2). Is it a two equation model? high Reynolds number model? or Low Reynolds number model? (3). Another black box code?

V2f is a 4 equation relaxation model. Besides k and epsilon equations, it also solves velocity fluctuations perpendicular to the streamline and a global relaxation function f. The prediction for separation flows is much better in genreal. But the model requires resolution all the way to the wall without wall function approximation, which is lot more expensive.

(1). So, it is a four-equation, low Reynolds number k-epsilon model. (2). And it is likely to have the same Y+=<1 constraint at the wall?

Hi,

Well the v2f model is more costly than the more or less standard low-Re k-e model. But only about 20 - 25 %.

So compared to Full Reynolds Stress model is rather cheap. And it does performe better than RSM (my point of view).

Regards

Jens

To begin with take a look at :

P.A. Durbin - "Near wall turbulence closure without 'damping functions'" - Theoret. Comput. Fluid Dynamics, 3 , 1-13, 1991

If you want to know more about test cases involving this model, take a look at :

R. Manceau & S. Parneix - "Computations of turbulent flows using the V2F model in a finite element code" - Engineering Turbulence Modelling and Experiment 4, 319-328 - W. Rodi and D. Laurence (Editors) - Elsevier - 1999

An extension of the elliptic relaxation to RSM models is introduce in :

D. Laurence and P. Durbin - "Modeling near wall effects in second moment closure by elliptic relaxation" - C.T.R. Proceedings of the summer programm 1994, 323-338 - 1994

I do believe that there is also an article in the J.F.M. from P. Durbin.

Several CFD codes editors are going (or have already) implement this model.

(1). Thank you for the information.

Hi Jens,

what type of flows did you calculate with the V2F model ? What were you experiences concerning physical correctness (e.g. prediction of separation) and numerical robustness ?

Thanks,

Mark

Hi,

Well, most basic flow since I am currently implementing this model into oure reserach code.

Stuff like Backfacing step, simple room geometry, cube in a channel.

But Star-CD has just release an add-on modul with the turbulence model and they have reported 20 % increase in cpu time.

So I think my 20 - 25 % is within limits.

About the physical correctness I have good experience, especially for separations point in transient simulation. And also numerical it is good, I have not observed any problem once it has been correct implemented.

Regards

Jens DIAS