|
[Sponsors] |
September 24, 2010, 14:54 |
kOmegaSST: bounding of arg1 and arg2
|
#1 |
Senior Member
Felix L.
Join Date: Feb 2010
Location: Hamburg
Posts: 165
Rep Power: 18 |
Hello, fellow Foamers,
I'm currently investigating the (incompressible) kOmegaSST turbulence model with respect to it's behaviour on different grid resolutions using a simple 2D zero pressure gradient flat plate testcase. While studying the implementation of said turbulence model in OpenFOAM 1.7.1 I stumbled across something I can't explain: The arguments arg1 and arg2 are implemented as follows: Code:
volScalarField arg1 = min ( min ( max ( (scalar(1)/betaStar_)*sqrt(k_)/(omega_*y_), scalar(500)*nu()/(sqr(y_)*omega_) ), (4*alphaOmega2_)*k_/(CDkOmegaPlus*sqr(y_)) ), scalar(10) ); Code:
volScalarField arg2 = min ( max ( (scalar(2)/betaStar_)*sqrt(k_)/(omega_*y_), scalar(500)*nu()/(sqr(y_)*omega_) ), scalar(100) ); What puzzles me is that both arg1 and arg2 are being bounded to the minimum values 10 and 100 respectively, resulting in F1 and F2 always having the value 1 - regardless of the properties of the flow. This leads to the fact that the kOmegaSST-model implemented in OpenFOAM only uses the k-omega approach and forfeits the blending between k-omega and k-epsilon which was the original motivation of the SST turbulence model. What's the reason for omitting the blending mechanism in OpenFOAM? Is this intended or am I missing something? I'll be grateful for any kind of clarification. Greetings, Felix. |
|
|
|