Boundary Layer in Duct Flow
 

I am trying to verify the accuracy of CFX by modelling some simple flows. One such is flow through a square duct. I am noticing an odd velocity profile within the boundary layer. It starts at zero at the no-slip wall, climbs and peaks at a velocity greater than the free stream velocity, then lowers to a steady value approximately equal to the free stream velocity within the core of the duct.

As far as I understand flow theory, the velocity within the boundary layer should not exceed the inlet velocity. Does anyone know why this happens and if it can be fixed?

Thank you,

Brady

Yes, the result is correct.

If your inlet is a plug flow (ie constant velocity over the whole boundary), as the boundary layer grows it is a region of slower velocity. To maintain continuity then the flow in the middle must be faster than the inlet.

Attachment 6220

That makes sense, but should the boundary layer show a profile such as this? Sorry for the crude drawing, it should be symmetric.

Yes, that is normal. As the boundary layer develops the "bump" smooths itself out and ends up being a small increase in the duct centre velocity. But in the early stages of BL grow it starts as a bump in the velocity profile.

Don't take my word for it, do a literature search and find out for your self. Also, do a sensitivity study on your simulation (mesh, convergence, time step) and if it still gives the bump you can be confident it is real.