This set of simulations consists of 4 simulations that use the same geometry and mesh, but different values of the kinematic visosity
Flow around a cylinder and the phenomenae that occur at various values of Re is a well known canonical case in fluid mechanics. I did theese simulations to solidify what i learned about CFD and get some practice in OpenFOAM. I ran simulations at Re values of: 1, 20, 55 and 100. The following image found in this paper ilustrates the resulting phenomenae.
Because of the simplicity of the geometry, I constructed it and its mesh using just BlockMesh. I divided the mesh into regions in the way shown in the picture below.
The mesh density is greater in the zones close to the cylinder and in the zones where the trailing wake is expected to be. I also used simple grading to create finer meshing in the most important regions.I used the icoFoam solver for all of the cases. The discretization schemes i used are fairly standard and can be found in the fvsSchemes dictionary. The same can be said for fvSolution. I used Scotch decomposition to split the domain into 8 parts and ran the simulations on my AMD FX-8320 CPU which were completed in reasonable time (less than 20 minutes for each).
I visualized the results with the following animation which presents one seceond of the simulations with the first 0.1 seconds playing in a slower speed.
resU.mp4
The resulst are as expected. For the lowest Reynolds number there is no backflow. In the case in which the Reynolds number was 20 a steady backflow appears. At the Reynolds number of 55, we can see unsteady shedding of the flow, and at 100 we can see a fully devoloped Karman street.
I also made an animation of the case with a reynold's number of 100 with particle tracing added on.