[dreamwork]

Hi
I am trying to model vortex shedding caused by 2D airflow over a circular cylinder. The Reynolds number is about 100000. The field of study is rectangular with velocity inlet for upstream and pressure outlet for downstream and symmetry for top and bottom of cylinder. I choose K-Epsilon turbulant model and steady solver and very fine mesh. Oscillation is detected in momentums residual and does not converge to zero. What should I do for this problem? I look forward to hearing back from you. Thanks a lot.

[flotus1]

You cant capture transient phenomena (vortex shedding) with a steady-state approach.
This is also the reason why the case shows poor convergence.

[virendra_p]

ke with std wall functions over very fine mesh will cause problems....use scalable wall functions instead...also you can try kwSST model which is popular for separating flows....good luck!

[dreamwork]

Thank you both. Good points. Although it is not yet been thoroughly resolved and the residuals become stationary at a high value but the vortex is visible past the cylinder. But what about BCs and fine mesh?Are they true? I tried a coarse mesh and surprisingly the vortex vanished and the velocity contour became like the velocity of a low Reynolds number flow!!! any help would be appreciated.

[virendra_p]

The bc's you used are fine...this problem is easy to mesh...you can use icem to create high quality..well controlled and resolved mesh to capture tiny details...have a look..upstream part is shorter with coarse mesh...cylinder adjacent area is finely meshed to resolve bl...and the downstream part is longer to place outlet far away and is finely meshed to capture the vortices...

[dreamwork]

Now I got. Thank you Virendra.