It is well known what should be the value of Y+ in a plate moving on the air. We know it depends on the length. What if the wall where we are calculating Y+ is a closed surface (of a rotating cylinder, for instance). What should be the L used ? Should it be the perimeter?

I am trying to run CFD codes to predict the induced air speed around a rotating cylinder (as in Magnus effect with no forward velecity to begin with), but no one seems to give me the result I expected. Migth be two things: 1- At the high speeds I am using, there is not too much induced velocity, for some effect due to turbulence. 2- Less probably, these CFD commercial codes does not work well with righ speed rotations...

Does anyone has had any experience with similar studies?

What should be the best turbulence model (and best comercial CFD) to face this problem?

Does anyone have any analytical solution for the turbulent boundary layer and flow field?

I am sorry if I am asking stupid question, but I am beginning in the field...

Thank you all, Luiz

(1). If you are using a commercial code, then the best help you can get is from the vendor's support engineer. (2). Turbulence modeling is in itself a big and difficult research field. My suggestion is: first study the book by Schlichting, the boundary layer theory first to get some basic understanding of the turbulent boundary layer flows. (3). After that, if you are interested in the modeling, you can take a look at the forum/Books section. Resources section also has some turbulence modeling information.

Thank you for yor answer. I knew I would have to do that eventually, but I need answers in a much shorter terms. Meanwhile, if anyone could help me find some papers / experimental results onthis subject, I would be very happy. Best Regards, Luiz

(1)."It is well known what should be the value of Y+ in a plate moving on the air. We know it depends on the length. " (2). Y+ is a dimensionless coordinate of y, which is the cross-stream coordinate normal to the local wall. (3).So, the statement of item-1 does not make any sense at all. In other words, your question needs to be re-phrased so that our readers can understand it.

Check out:

Launder, B.E., Pridden, C.H., Sharma, B.I., (1977), 'The Calculation of Turbulent Boundary Layers on Spinning Curved Surfaces' Journal of Fluids Engineering. pg 231 - 239

This suggests a correction to the k-epsilon turbulence model when dealing with swirling flows. There have been several previous posts on this list about the best turbulence model to use in swirling flows and generally people like to recommend Reynolds Stress Models. This does capture the most physics, but is not always practical in terms of coding, stability and running times. The correction dealt with in this paper has been shown to work well in many geometries with only one tunable constant. This tunable constant always seems to take the same optimal value in these many different geometries!

Enrico