Hello,
You can put your boundary condition after the submerged hydraulic jump so that all the domain has a fluvial regime (in this case prescribed velocity and free water depth), or you can roughly model the tunnel as a square with prescribed velocity and depth, but in this latter case you may have to tune the velocity to get the right discharge (the prescribed depth will replace the continuity equation at the boundary, so the discharge effectively entered into the domain may be different from the expected value velocity * section).
With best regards,
Jean-Michel Hervouet