[Delcraft]

Hello,

I'am trying to train the application of k-epsilon turbulence model in OpenFOAM on a benchmark problem of turbulent incompressible flow over a flat plate. I used the following boundary conditions (they were taken from this link: http://turbmodels.larc.nasa.gov/flatplate.html):
Inlet Velocity: 0.2 M (Re = 8.5E6)
Flat plate length: 2 m
Inlet total temperature: 302.4 K
Outlet and freestream pressure: 101325 Pa

I solved this problem on two computational meshes, with Y+ = 112 and 32. After the solution I plotted the dimensionless velocity profile in a boundary layer expecting to obtain the law of the wall resolution. But instead I obtained a strange distribution of the velocity for both of comp. meshes (you could find the plot attached). For both of solutions there was a linear velocity interpolation below of given Y+ values and something like logarithmic distribution above. But even for the latter if we project the "logarithmic line" down to Y+=1, we won't obtain the value of U+ = 5.1 (what we should obtain from the log-law).

Finally, I have a question - am I don't understand the work of wall functions in k-epsilon model (and wall functions in general) or these results are strange indeed? Because I think that the wall function must interpolate the velocity profile correspondingly to the law of the wall (and viscous and log. sublayers must present) if I place the first grid point correctly (in this case, into the logarithmic sublayer, Y+>= 30).
Thanks in advance,

Delcraft

[Delcraft]

I suppose that I caught the sense of these approximations... The law-of-the-wall formula is used just for a calculation of dependent variables (like wall shear stress, K and epsilon) on the wall from known parameters like Y coordinate and Velocity, and it doesn't relates to interpolation between grid nodes.
So, this theme is closed now.