[kkpal]

Dear all,
I am simulating a 2D kepsilon case of flow around cylinder, Re=100.
A rough look at the drag coefficient plot looks good. But when I zoom in, it seems there are two parallel lines forming the plot, and the value of Cd is wiggling in a small range.
any idea why?

[wyldckat]

Hi Kai,

Since you mentioned this thread on your other thread: http://www.cfd-online.com/Forums/ope...les-cd-cl.html - here's what I can figure out:
The mesh is very likely to have the exact size, plus the "deltaT" ahve the right value, for this phenomenon to occur. This has to do with the effect of discretization... for which I can never find a good example on the web

But imagine the following scenario: you have a plot of a sine with a period of 1 second and amplitude of 1.

• If your sampling rate is exactly 0.5 s. What do you see for each period? 3 points with 0 value.
• You change the sampling rate to 0.25s. What do you see? 5 values, where 3 are zeros, a "+1" and a "-1". This is basically what you're seeing in this plot.

This is why I mentioned in the other thread about the "deltaT". If it were small enough, you would see the complete wobbling plot. And maybe even the flow would change considerably, because it had more iterations for resolving the cumulative effect of the kinetic energy blobs moving around the mesh

Best regards,
Bruno