Dear CFD people,

I modified 2 dimensional density based finite volume laminar flow solver, which is based on cell centered scheme such that it can simulate unsteady flows using Dual Time Stepping. However, while trying to simulate vortex shedding past 2D cylinder, it didn't converge and negative density values are resulted at computational boundaries (treated using farfield BC). Do we need modifying farfield BCs for simulating unsteady flows?

Best regards.

Dear CFD student,

When treating the unsteady problem as a pseudo-steady stae problem using DTS, the steady state tools can be used as it is, so no modification needs to be done to the farfield BCs, when running the code. If you are not using a freestream condition at the boundary, you could try with this condition, provided the boundary is far off as this is a simple BC, and could help you to understand if the problem is in the BC or anywhere else in the code. Also, it would be nice if you stop the simulation before the shed vortices( although getting dissipated) do not cut the farfield boundary, precisely for which reason the farfield should be taken really far.

Hope this helps

Regards,

Ganesh