[mart.hein]

Hello everybody,

I'm using the compressible k-Epsilon turbulence model and I've got a question about the formulation of the total stress term devRhoReff and the source term for the momentum equation divDevRhoReff. The former is defined as

-muEff()*dev(twoSymm(fvc::grad(U_)))

whereas the latter is defined as

- fvm::laplacian(muEff(), U) - fvc::div(muEff()*dev2(fvc::grad(U)().T()))

Mathematically the second equation should just be the divergence of the first one. However, when I calculate the divergence of the first one and replace the second one with

fvc::div( -muEff()*dev(twoSymm(fvc::grad(U_))) )

the solver crashes after some iterations with a time step continuity error .

Can someone explain this behaviour to me? Are my assumptions not correct?

Thanks a lot

martin