[waynezw0618]

Hi everyone:

i have find in kEpsilon and KomegaSST,both k , epsilone,Omega equation has the advection term like this:
+ fvm::div(phi_, k_)
- fvm::Sp(fvc::div(phi_), k_)
but i think the div(phi_)should be "0" for incompressible flow. and why not write as :
- fvm::Sp(fvc::div(k_), phi_)
or
- fvm::Sp(fvm::div(k_), phi_)

thanks

yours wayne

[villet]

Hi Wayne, there has already been discussion about this matter. Try to search for it, because I cannot remember the details. You shouldn't worry too much about it, since the divergence term goes to zero as the solution converges.

[waynezw0618]

hi Ville

thanks，i will search it now.if you find please tell me .there are some similar term in the diffusion term too :

tmp<fvvectormatrix> kEpsilon::divDevReff(volVectorField& U) const
{
return
(
- fvm::laplacian(nuEff(), U)
- fvc::div(nuEff()*dev(fvc::grad(U)().T()))
);
}

the red part will be 1/3*nuEff*grad(div(U)),and for incompressible flow it will be zero too.

the problem is that these terms should be zero in theory,but not in numerical,so i don`t know why add these terms here with the danger of introduce the numerical error. i think maybe there is something i do not know.so any one could help me ?

thanks

yours wayne