No-slip condition for non-resolved boundary layer in open channel banks
 

Dear all,
I am using a commercial code and applying it to open channel flows. The question is: it is really wrong to use a non-slip condition in channel lateral banks if I use very coarse meshes, like 10-20 elements for channel width? In the user manual of the model I am using it is written:

"When the grid size near the boundaries is larger than the thickness of the boundary layers
occurring in the flow, the shear-stresses along the lateral boundaries can be neglected. A so-called free slip boundary condition is applied at all lateral boundaries. "

"The choice of the slip condition depends on the grid size versus the boundary layer thickness.
It is of no use to apply a no-slip condition when the grid size cannot resolve the boundary layer."

So the questions are:

1)do I really make a big error to use no-slip condition if the cell next to the bank is larger then the BL thickness (defined with the 99% of the free stream velocity, and also the notion of free stream in a channel is relative, since there is also the boundary layer generated at the bottom and that reaches the water surface).

2)In literature I have found that the boundary layer thickness near the banks scale with the depth. But I was able to find this information only in one conference paper (that does not references any source). Anybody has any good references for bank boundary layer thickness? Could it actually be that it extends for all the channel width how it does in the vertical?

3)It is true that if I do not resolve the boundary layer I should prescribe slip condition?

thanks
A.

No, is not "wrong" to prescribe the no-slip condition, but you have to take into account for your choice... I mean that you can not hope to compute the wall stress or other wall parameters... on the other side, if you fix the slip condition you are prescribing somehow the wall stress. This is an assumption and can be poor for complex geometry.