Sign inconsistency in the compressible Launder-Sharma k-epsilon model
 

Hi all,
checking the epsilon transport equation in the Launder-Sharma low-Re model I've found a clear inconsistency in the C3 constant sign (at least it is there in the OF-2.1.0 release, I don't know if it has been already fixed in 2.1.x).

The following is the (correct) compressibility production term in the standard k-epsilon model implementation:

- fvm::SuSp(((2.0/3.0)*C1_ + C3_)*rho_*divU, epsilon_)

As you can see, the term multiplied by C1 enters with a negative sign, while the term multiplied by C3 actually becomes positive, as in OF C3 is equal to -0.33 while the correct constant value in the RHS of the epsilon equation should be +0.33.

This is the same (incorrect) term in the Launder-Sharma model:

+ fvm::SuSp((C3_ - 2.0/3.0*C1_)*rho_*divU, epsilon_)

The C1 term is again negative, but so it is the C3 one as in OF +C3=-0.33.

So, I don't know what is the real influence of such an error, but to me this is clearly a bug (I will report it to the bug tracker in a while).

Greetings to all

V.

(Reported)
http://www.openfoam.org/mantisbt/view.php?id=462