It seems that two-equation RANS models dampen (or kill) the unsteadiness of problems that is known to be "very" transient, like the sub-critical flow past a circular cylinder. They kill the unsteadiness so that a steady-state simulation can be used which in turn results in a converged steady-stae like solution. The total, local pressure and drag coefficients obtained using this steady stae simulations (i.e. by dropping the transient term in the RANS equation) were comparable to the time-averaged coefficients I obtained using the time-accurate simulation.

This made me believe that the dissipative effect of these models overwights any numerical dissipation a discretization scheme can cause.

I wonder if some one will comment on such observation and then help me understand the reason behind that. I am particulary, interested in the effect of dropping the transient term on the pressure term in RANS equation.

Thanks I will appreciate any comment

I am also interested in the explanation for this.

I know that some researchers claimed that they can use the two-equation model to predict unsteady flows such like vortex shedding, but others claimed that only RSM models work.

I guess this is related to the Boussinesq¡¯s hypothesis in the two-equation model.

May be I don't understand the problem, but I do have an experience of modelling unsteady flow with two-equation model (k-epsilon RNG, to be exact). I was vortex shedding at the circular cylinder.