Hi,

Since the NS-eqs. are time-averaged in RANS, is it then possible to make a steady state calculation of a flow with turbulence modelled with a RANS-model? (if the bc's also are steady)

/Rasmus

I consider the NS-eqs. are ensemble-averaged in RANS in which case a RANS simulation is not neccessarily steady. However, both time-averaging and ensemble averaging give rise to the same kind of closure problem and the same kind of models so the answer is not clear. A good question is what do you expect to get out of a RANS simulation of the shedding flow behind a cylinder? Would you expect a steady solution corresponding to the time-averged flow? I don't think that is what most RANS models will provide. Some will give a near-steady but inaccurate solution for the average while most will do a pretty decent job of simulating the shedding and even get the frequency about right.

One way out of this dilemma is to use LES for simulating unsteady problems which unambiguously and correctly produces an unsteady solution as it should.