Hi:

I believe there are a lot of turbulence model expert in this forum. May I ask two fundamental questions:

1) How can we determine if we should use turbulence model rather than only solving laminar problem? I know in heat transfer textbook we usually consider the Reynold number. But when we deal with complex geometry, it is hard to define characteristic dimension and no experimental data is available. In some special cases the flow has both external and internal characteristics, how can we choose the critical Re number?

2) As I understand, we should only implement the Algebraic mixing-length turbulence model in the boundary layer. But if the free stream is inside an large enclosure and still have some transverse velocity gradient, should we also add the turbulence viscosity and diffusivity to the whole domain or we just leave it as laminar?

Thank you very much for you time and help.