I am trying to simulate vortex shedding off of the rear of hexahedral shaped vehicle. Clearly defined vortices are formed in the steady state solution, however during the unsteady time simulation there is very little change and no movement of vortex centers. The timestep is set at 1/10th of the theoretical shedding period. Any ideas on why the vortex shedding is not being properly simulated? I was thinking perhaps convergence criteria or perhaps a different scheme for pressure-velocity coupling.

Ben

Use higer order discretization schemes. For pressure-velocity coupling use PISO.

What turbulence model have you used? A standard k-eps will never show vortex shedding, no matter what discretisation scheme, or which order, ...

Laika, still orbiting

I have been using the realizable k-eps turbulence model. Also I have been having major difficulties gettting the solution to converge with the PISO scheme

Hi

Make sure you're grid is fine enough to resolve the vortex shedding and also make sure that the time step is small enough so as to satisfy the CFL criteria. Start with a time step that is much smaller than the one needed for CFL criteria and then increase it whilst checking the accuracy. i.e. check that the shedding frequency does not alter with increasing time step.

James