I think all that's been done is to provide an active magnetic levitation system
that incorporates permanent magnets to provide nearly all of the lifting force.
This is a good way to reduce flywheel energy consumption (equivalent to self
discharge in a chemical battery).
This doesn't solve the main market acceptance problem for flywheels, which is
catastrophic failure.
The technical issue, somewhat surprisingly, isn't fragment containment. That's
pretty easy, especially because composite fiber rotors produce more benign
fragments than metal rotors.
It's the angular momentum. Angular momentum is conserved, which means that
however the failing flywheel contacts the non rotating external environment, it
will transfer angular momentum to that environment. That generates torque. Lots
of it. Enough to easily twist the flywheel housing loose from its mount and
send the whole thing spinning through the building, through a wall, and out
into the street.
You can get a feeling for the torque by imagining that you have a 1 hp (0.75
kW) motor to charge the flywheel for 5 hours. That will store 3.75 kWh, a
modest amount of energy. Most of us have a feel for the torque on a 1 hp motor.
Not something you would want to resist by hand.
Now imagine the flywheel fails, coming to rest with respect to its case in 10
seconds - a fairly slow failure, if fragmentation is involved. All of that
stored angular momentum gets transferred to the case in 10 seconds. You can
roughly estimate the torque pulse by multiplying the charging torque times the
ratio of chart time/discharge time. In this case, it's (5 hours)/(10 sec) =
1,800 as the torque multiplier.
This is a really tough failure mode to deal with economically. Either you
design mounts that can take the maximum possible torque pulse ($$) or use
counter-rotating wheels that crash into each other (net external angular
momentum = 0, $$$) or mount the case in bearings with a brake so the case can
spin up during failure and then be braked slowly thereafter ($$$).
None of the solutions looks cheap enough to make it into residential mass
markets. I doubt we'll see mass deployment of residential units. The only
economic use case I know of is phase matching at grid intersection points. Big
flywheels in an application that can pay for all the failure mitigation
measures required.
