I posted a ring flywheel idea a while ago and, having thought a bit more about it, have made a few improvements.
The key elements of both designs are that the flywheel is a ring shape (thus maximizing the rotational inertia) and the ring levitates above the stationary housing (thus no contact, thus
minimizing friction).
I'll briefly discuss three aspects to the invention: shape, levitation and propulsion.
Shape
My idea is to use a half toroidal shell as a housing around the ring flywheel (see illustration). Using a half toroidal shell will restrain the flywheel's horizontal and vertical movements while allowing rotation. I think the ring will be stable and levitate at any speed - even when not moving.
The ring is lumpy and has coils spaced around its inner periphery. The coils are for propulsion, the lumpiness is so the ring has uniform resistivity all the way round.
Levitation
The toroidal shell (purple) has current traveling in one direction (say clockwise), while the ring (silver) has current traveling in the other direction (counter-clockwise); the opposing currents causes repulsive magnetic fields hence levitation. The current in the shell is sustained by a current generator. The current in the ring is induced into the ring by a changing magnetic field (generated by an alternating (sawtooth) current in concentric non-touching wire loop (orange)).
Propulsion (i.e. spinning the flywheel)
Some of the current that is traveling around the ring will flow through the coils. These coils create a magnetic field that can be pushed against by carefully timed magnetic field pulses produced by the stationary coils (green) thus propelling the ring to store energy, or to induce current in the green coils to draw energy from the flywheel.
There will be energy losses from the electrical resistance in both the toroidal shell and the ring. However, my rough calculations from my previous idea showed that the resistive losses would be acceptable for a suitably dimensioned flywheel.
As before, the whole apparatus could be mounted on a tilt according to latitude to prevent precession forces due to the Earth's rotation causing destabilization or unnecessary energy drain on the flywheel.
I realize that there are other magnetic bearing flywheel systems, however they suffer from two disadvantages. Firstly, they usually rely on permanent magnets which tend to produce counterproductive eddy currents in the flywheel; secondly, they usually require active electromagnetic control to keep the flywheel levitating. My design is relatively simple and avoids both of these issues.