This is a very similar concept to a flywheel except that instead of levitating a rotor on magnetic bearings, we use robotic maglev cars speeding around a track instead. The track would be several kilometres in diameter and would be heavily banked to resist centrifugal forces.

The cars have regenerative
brakes that release energy whenever required. Likewise whenever excess energy is available to the grid, they are sped up.

To further reduce idling losses it would be beneficial to enclose the whole track in a vacuum. This would also allow a higher maximum speed.

The energy stored would obviously be our friend œmv².

The best way to store a lot of energy would be to maximize V since total energy stored is proportion to V squared.

The limiting factor to speed is probably the forces exerted on the track by the hurtling cars..

So lets say we have 100 cars running round the track each weighing 100,000 kg (about 220, 500 lbs). Lets be optimistic and estimate the speed to be 3,600 kph/2,237 mph/1000 m/s (I know!). So our energy stored is 100 x 0.5 x 10^5 x (10^3)^2 = 0.5 x 10^2 x 10^5 x 10^6 = 5 x 10^12 Joules / 3.7 x 10^12 foot pounds.

(Or 1.2 Kilotons of TNT, but why would we measure this in Kilotons...)

This is equivalent to 1.4 million Kwh. So quite a lot of leccy. If the turnaround efficiency was 100%. Which it wouldn't be..

Someone is going to tell me they can't go that fast. I just know it..

The centripetal force would be

F = mv²/r

So given a radius of 2 km, the force would be:

F = 10^5 x (10^3)^2 / (2x10^3)
F = 5 x 10^10/(10^3)
F = 5 x 10^7 Newtons

F = 50 MN / 11.24 million lbs

Uh.. thats actually quite a lot.. and the cars would be dealing with a force of ~50g.