About devReff() and divDevReff() in kEpsilon.C
 

In incompressible case , in kEpsilon.C , devReff() = -nuEff()*dev(twoSymm(fvc::grad(U_))) , we know divU=0 in incompressible case , so why not use devReff() = -nuEff()*twoSymm(fvc::grad(U_)) , is there any CFD programming technique to use dev() function ?

And in incompressible case , divDevReff() = -fvm::laplacian(nuEff(),U) - fvc::div(nuEff()*dev(fvc::grad(U)().T())) , is there any special reason use dev() function ? or why not use dev2(fvc::grad(U)().T()) as in compressible case ?

thanks a lot .