CFD Online Discussion Forums

CFD Online Discussion Forums (
-   OpenFOAM (
-   -   Use of face centered differences in gradient computation (

henningh August 25, 2011 05:56

Use of face centered differences in gradient computation

According to the recent work of Richrds and Norris (J. Wind Eng Ind.Aerodyn 99 (2011) 257-266), the near ground spike in turbulent kinetic enery in the k-epsilon
model originates from an inconsistency in the discretization. This error can be removed by computing all gradients in terms of face centered differences.

Can this shift be achieved by any of the options in the fvSchemes file?

Best regards,
Henning Heiberg-Andersen

stevenvanharen August 25, 2011 09:34

Are you sure this is not already being done?

I think discritization (53) from the paper you refer to is not possible in an unstructered code.

Furthermore, in kEpsilon.C:


volScalarField G("RASModel::G", nut_*2*magSqr(symm(fvc::grad(U_))));
where fvc::grad(U_) are the gradients at the face.

Is this what you were referring to?

praveen August 25, 2011 11:41

I have asked this question before

and I think there is no such scheme available. Please correct me if I am wrong.

All times are GMT -4. The time now is 13:33.