[mrblond]

Hi everyone,
I'm new in this forum and I'm currently trying to learn using star ccm+ vertion 9. I'm analyzing the infinite lentgh cylinder case, see the next figure for the geometry.



the mesh settinggs are:
base size on volume:1m
target size on cylinder: 0.02m
prysm layer thickness: 0.02m
number of layers: 20
layers stretching: 1.5

I also used a refined volume with a base size of 0.1m behind the cylinder to better capture the wake.

the boundary condition are:

inlet: velocity inlet (10,0,0)m/s
outlet: pressure outlet 101325 Pa
top and bottom: wall (slip condition)
sx and dx boundaries: periodic

I tried different physics but more or less I obtained the same results. Mainly these are the settings:

steady
coupled flow
constant density
turbulent
k-omega

all with "auto-select recomnended models" activated.

The problem is that th Tke converges until the 10th or 50th iteration depending on other settings I choose but then it diverges strongly. also the y-momentum does not have a strong convergence. Does anybody have some suggestions?

Thanks

[solsidan]

A few suggestions:
1) Have you tried the segregated solver?
2) Activate "Temporary Storage Retained" in the k-omega solver and step the solution forward one iteration to visualize the TKE residuals in a scalar scene. This may give you some hints if there are localized issues with the mesh or the geometry.
3) Check the "Simulation divergence check-list" post in the knowledge base on the Steve Portal.