RNG-k-epsilon with realisable contraint
 

Hello Foamers!

Does anybody know if the RNG-k-e model has a realisable constraint? If not, then I am wondering if it even need one, i.e. are there a other properties which makes the need for readability redundant?

Cheers in advance

In the sense of realizability of the Reynolds stresses, the RNG k-e model has no explicitly imposed constraint, so it could not be considered as "realizable". However, the latest version of this model (which is from Yakhot et al., 1992) has some other interesting features, such as improved performance for homogeneous shear flows and geometrically separated flows (like the backward facing step test case). If you want to learn something more about the RNG model and/or about realizability matters, I can direct you to the following references:

- P. A. Durbin, "On the k-e stagnation point anomaly", Int. J. Heat and Fluid Flow 17: 89-90, 1996.

- Tsan-Hsing Shih et al., "A new k-e eddy viscosity model for high Reynolds number turbulent flows", Computers Fluids Vol 24, No 3, pp 227-238, 1995.

- C. G. Speziale and S. Thangam, "Analysis of an RNG turbulence model for separated flows", Int. J. Engng Sci. Vol 30, No 10, pp 1379-1388, 1992.

- V. Yakhot et al., "Development of turbulence models for shear flows by a double expansion technique", Phys. Fluids A 4 (7), July 1992.

Regards

V.

The thing is, that I need to know if it is better to have a realizable constraint in the RNG model or not. Because, I computed a simulation which aborted due to a stagnation point. I changed to realisableKE to pass the critical point. Afterward, I changed back to RNG and the simulation went through. This is not a nice solution, but i worked. Therefore, I am interesting to implement the realizable constraint in RNG model.

In practice, the realizable k-e model of Shih et al. (which is the realizableKE implemented in OpenFOAM) possesses both the realizability costraint on the Reynolds stresses and some characteristics which closely resemble that of the improved RNG k-e model: so if you succeed to impose the realizability constraint, probably you will have something very similar (in terms of computed results) to the realizable k-e itself...So my advice would be to use directly the realizableKE model for your problem, but apart from my opinion a good idea to impose the realizability maybe could be a limiter on the turbulent time-scale k/epsilon (see Durbin's technical note about this matter)

Regards

V.