[gniticxe]

Hi,
It sounds like you may have a cool project behind this, I'll help where I can:

If you have a disc, and not mass on a ring, I = 1/2*M*R*R

I'm sure you know already, but Ek = 1/2*I*(rad/s)*(rad/s), instead of rpm

As far as a physical way to transfer the energy, how about using a commutated shaft and using an electro magnet to extend a piece of steel past the radius of the wheel (you'll need a spring to keep it from deploying early and also a balancing weight on the opposite side, depending on how large/fast this is). In less than a revolution, you'll have contact with the waiting rod.

Using an efficiency of 0.6 for energy transfer calculations is probably plenty conservative, especially since the disc may spin backwards after hitting the rod.