[linkamp]

Hi everyone,

I'm trying to simulate a t-type shaped vortex shedder in a 2D "pipe" as used in a vortex flowmeter using rhoPimpleFoam. My basic problem is that I cannot get near the vortex shedding frequencies I obtained in experiment (factor 1.3 to 1.4). I tried everything I came up with.
I'm simulating RANS since I don't have the capacity (nor experience) for LES, not to mention DNS. I tried k-eps, k-omega, k-omega-SST. The vortex frequency is a bit different for each model, but for none of them it's near what I expected. I also tried using wall functions as well as resolving the boundary layer. Not that big difference. I carried out mesh refinement and higher temporal resolution, but even with the highest resolution (time and space) it doesn't really get "better".
Any idea what I do basically wrong? What irritates me is that there is some literature available, where vortex shedders are simulated in 2D using eddy viscosity models (mostly k-epsion) getting excellent agreement with experiments.