I'm a beginner with CFD (CFX4.3): in dealing with the k-epsilon turbulence model I've to set the first node at y+>11.Since y+ is a function of the unknown friction velocity how can I estimate the first grid position? Are there any empirical relationships which allow a crude estimate of the friction velocity? Thank you in advance for anybody helping me.

Yes, you can estimate the wall-distance you need to get a certain y+. There is even a nice little free online calculator which will do it for you - you can find a link to it in the Resources/Calcs/Misc section of CFD Online.

In the manual you will see that you can set a USRSCAL (user scalar) to return the y+ values and plot it.

Regarding u*, for open channel for example you can use the crude assumption that a mean u* = sqrt(g*R*So) where g is the gravitational acceleration, R the hydraulic radius and So the energy gradient.