You should detail the parameters of your code setting. The vortex shedding appears since laminar flow conditions, therefore I think something is wrong in your simulation.

Then, have you performed the run until the flow has fully developed?

@filippo

My mesh us much fine. Time step is also according to CFL number. I have run simulations untill there is no further change in vorticies.

These are not details… and what do you mean for :

I have run simulations untill there is no further change in vorticies.

By using URANS you are actually assuming a fully turbulent flow. when the boundary layer arr turbulent, flow separation is delayed. It may be suppressed. This is what normaly occurs for a smooth cylinder at Reybolds numbers of order 2E5 and can occur at much lower Reynolds when tge wall is rough (this is usef in golf and tenis balls). Try a laninar simulation. Starting around Re=50, periodic vortex sgedding will occur. You may find is difficult to reach Reynolds numbers higher than 1000 or 2000 with a 2D laminar flow simulation. You could use a RANS model in the wave if the cylinder. Regards. Mico

It is known that a regular vortex shedding is observed in the range of Reynolds numbers VD/v from ca. 60 to 5000, where V the upstream velocity, D the cylinder diameter, and v the kinematic viscosity of the fluid. Refer p.31 in the book entitled "Boundary-layer theory" by H. Schlichting(1968).

If your computational results show no vortex shedding in this Reynolds numbers, you should consider that your computation has some faults,

Periodic vortex shedding does indeed always occur in the range 60

I agree with Avraham' s comments.

Try to give a small perturbation to the flow at the beginning (several time steps), like a ramp profile at the inlet. Then move back to the original inlet boundary condition (and wait some mass flow through to start sampling)

I do suppose that your DES model is looking for unsteady (time dependent) solutions?

regards

mico