Hello everyone,

I am trying to compute the skin friction of a flat plate boundary layer (validation), using a wall function boundary condition. Unfortunately I get strange results (skin friction too low). The computation has been made with a commercial as well as a "home-grown" code and both give wrong results. Both codes produce reasonable results if the viscous sublayer is resolved, though.

Any advice on that?

cheers, Oliver

Dear Sir,

My experiance using wall function (which we reluctanly do need to use sometimes) is that the only cases where these give fairly consitent result with simulation for which the near-wall region is resolved, are channel flows, and flat plate boundary layer flows. How low is too low? 10% could possibly be acceptable even for these simple flows. I've seen more than 5% variation between different low-Reynolds number turbulence model for channel flow. As you use two different codes, my advice would be to check the mesh. If the first node is not located in the wall function valid range (y+>30 to the end of the logarithmic region) you may get erronous results. Check by doing simulations on different meshes.

Regards Jonas

Hi there, I looked in detail into this problem a few years back using a commercial code. Try and get hold of a copy of my paper from you local library or online.

Date, JC & Turnock, SR, "Computational Fluid Dynamics Estimation of Skin Friction Experienced by a Plane Moving Through Water", Transactions of the Royal Institution of Naval Architects, Vol. 142, 2000, ISSN: 00358967.

Regards James