open TELEMAC-MASCARET The mathematically superior suite of solvers

Hello,

I'm running a simulation in a laboratory flume. In calibrating the roughness, I found that the Nikuradse roughness ks=0.01m produces good results. However, based on how telemac computes friction velocity (attached), a mesh with vertical mesh size smaller than the roughness will result in a negative u*, assuming u_plane1 is a positive value (In my case, the vertical grid size, i.e. dz, is about 0.003m). I thus wonder how and in which routine u* is computed & is there any recommendation for vertical resolution with respect to roughness height?

A related question is: how is the velocity at the bed (a solid boundary) computed? I noticed that it's non-zero and it's assumed to be at the height of about 1/10 of the vertical grid size. I'm wondering if it's computed by fitting a log curve with u* obtained by knowing u_plane1 (attached screenshot).

Any input would be welcome and very much appreciated!

Thank you,
Jason