[Ya_Squall2010]

(1) for incompressible solvers, while discretizing momentum equations, why is the following term considered only when turbulence modelling is turned on:
- fvc::div(nuEff()*dev(fvc::grad(U)().T()))

see :

GenEddyVisc::divDevBeff(volVectorField& U) const
00091 {
00092 return
00093 (
00094 - fvm::laplacian(nuEff(), U) - fvc::div(nuEff()*dev(fvc::grad(U)().T()))
00095 );
00096 }

(2) in pEqn.H of interFoam, the 'guess' fluxes phiU is first evaluated and then corrected, the question is, why the correction (and adjustment) is performed without considering the source terms (gravity, surface tension, etc.)

(3) in UEqn.H :

00019 if (momentumPredictor)
00020 {
00021 solve
00022 (
00023 UEqn
00024 ==
00025 fvc::reconstruct
00026 (
00027 fvc::interpolate(rho)*(g & mesh.Sf())
00028 + (
00029 fvc::interpolate(interface.sigmaK())*fvc::snGrad(a lpha1)
00030 - fvc::snGrad(p)
00031 ) * mesh.magSf()
00032 )
00033 );
00034 }

The source terms are first computed at cell faces and then reconstructed to cell centres, why bothering? Why not just compute them at cell centre in the first place? Is there any special issue I missed out?

Many Thanks!