using non-orthogonal correction for the laplacian term
 

Dear all,

Hopefully one of you can help me with the following question:

If one uses (in fvSchemes):

laplacian(nu,U) Gauss linear corrected;

which correction is referred to? minimum correction, orthogonal correction or over-relaxed? (as referred to in the thesis of Jasak)

Thanks in advance.

Hello,

I believe this is simply an orthogonal correction, used mainly for highly orthogonal grids (such as high aspect ratio anisotropic tets).

Hi travis,

Ok, I understand this, I try to use it on polyhedral grids. But do you also know how this orthogonal correction is implemented? Jasak refers to three different possibilities (minimum correction, orthogonal correction or over-relaxed) which are all orthogonal corrections bu with different numerical behaviour.

Until we've studied over-relaxed approach is what is implemented. Take a look of gaussLaplacianScheme.C lines 169-173, correctedSnGrad.C lines 56-108 and surfaceInterpolation.C lines 270-384.

Regards.

Great references to the code lines Santiago!

It is very clear now to me.