hi all! I'm studying the turbulent boundary of the flat-plate now. First,the turbulent model I used is the k-epsilon, but i find the velocity on the wall isn't zero.(is about 7~8m/s,if the inlet velocity is 20m/s) Using the same grid ,if the turbulent model is the SST or k-omega, the velocity on the wall is right ,zero. I don't know why！

Hello Wei,

I would suggest you to have a look at HYBRID vs CONSERVATIVE option. You can select one or the other option when you color a plane. The difference is well explained in the documentation and selecting hybrid should allow you to see what you want !

Also, pay attention when using the same grid for SST and k-epsilon. This is usually poor practice.

Regards,

Felix

Felix, thank you very much! the problem is solved, and i find the velocity profile near the wall is too different from the Coles, SO ,is it the wall funciton is not suitble to solve the problem near the wall.

Hello Wei,

There are some good books describing the wall function approach and its validity. I suggest you have a look at Ferziger & Peric (2002) or Wilcox (1993).

If the behavior of the flow in the vincinity of the wall is important to you, try solving right through the viscous sublayer. Some k-epsilon models can do that, the k-omega too and the SST when you keep y+ < 2.

Felix

I wouldn't jump into that conclusion too fast. Many factors will affect the quality of your results. Try improving your mesh etc... Refer to the user manual. Lots of tips there to help you improving your results.

Felix,thank you very much ,i will do that as you suggested,

That's right TB, it is good to be kept in mind that the wall function is only one parameter among many others ! Solving through the viscous sublayer is not a miracle solution. It only allows to get computed results near the wall.

Felix

Hey up yours TB!

Felix