[make]

In OpenFOAM-1.4.1 the reynolds stress tensor in the kEpsilon model has been calculated using
(2.0/3.0)*I)*k_ -(nut_)*2*dev(symm(fvc::grad(U_)))
which seems plausible to me.

Why is it in the RNGkEpsilon
((2.0/3.0)*I)*k_ - nut_*2*symm(fvc::grad(U_))
Where is here the deviatoric part? Isn't that wrong? Shouldn't it be
((2.0/3.0)*I)*k_ - nut_*2*dev(symm(fvc::grad(U_)))

Also for the divR term it is calculated using
- fvm::laplacian(nuEff(), U)
- fvc::div(nuEff()*dev2(fvc::grad(U)().T())
Shouldn't it be dev2?

Greetings,
Matthias

[pbo]

Matthias,

it seems you're looking at the incompressible models here, for which

dev(symm(fvc::grad(U_))) = symm(fvc::grad(U_))

since

tr(symm(fvc::grad(U_))) = div(U) = 0

Patrick

[make]

yep I was too small minded and didn't catch that point.
Nevertheless, no I can put my mind at rest and perhaps others too.
Patrick, thank's a lot for your quick reply.

Matthias