[Bazinga]

I am trying to compute the unsteady flow around a circular cylinder using an URANS approach and foam extend 4.0. The problem is that I get some weird pressure regions in the wake very close to the surface of the cylinder as shown in the attached figure. I already performed multiple simulations with the below-mentioned setup with different rectangular shapes without any similar issues. For the rectangular shapes, I also used this setup for free and forced oscillations of the cylinder.

It is a low-Re approach using the k-omega SST model only using a wall function for omega. I keep the max Courant Number < 1 and the average y+ value is ~0.5. In time I discretize using Crank Nicolson and 2nd order upwind for the divergence schemes. The Reynolds number is somewhere around 50.000.

I already played around with 2nd order discretization schemes and the solution procedure. Tested with OpenFOAM 5 and got similar results. Also, the problem occurs on different meshes, however, the area of the high suction behind the cylinder gets a bit smaller with refined mesh.

The 0, constant and system directories of the simulation and some results can be found here:

https://drive.google.com/open?id=1t3...8xOhxXI4Z_PNtm

Wondering if anyone got similar problems or can point me in the right direction. Again, what bothers me a bit is that for various rectangular shapes I never got similar problems.

[piu58]

> it is a low-Re approach ... Reynolds number is somewhere around 50.000.

50000 is not a low Reynolds number. It is quite high instead. I recommend to read the results of CFDing with Re=3900 and the problems whick ocured at this much lower Re:
Detached–eddy simulations of the flow over a cylinder at Re = 3900 using OpenFOAM, Valerio D’Alessandro, Sergio Montelpare, Renato Ricci

https://www.researchgate.net/publica...using_OpenFOAM

[Bazinga]

Originally Posted by piu58

> it is a low-Re approach ... Reynolds number is somewhere around 50.000.

50000 is not a low Reynolds number. It is quite high instead. I recommend to read the results of CFDing with Re=3900 and the problems whick ocured at this much lower Re:
Detached–eddy simulations of the flow over a cylinder at Re = 3900 using OpenFOAM, Valerio D’Alessandro, Sergio Montelpare, Renato Ricci

https://www.researchgate.net/publica...using_OpenFOAM

Dear Piu58,

Yes, indeed 50000 is not a low Reynolds number! With low-Re approach I was referring to the modeling at the wall (see for example: https://www.cfd-online.com/Wiki/Two_...bulence_models).

Thanks for the reference, I will check it out.