I was rather surprised that the energy transfer loss is only 10% in batteries, but 15% in flywheels. It's pretty amazing that the chemical process in a li-ion battery (developed in the last 50 years) is more efficient than our best engineering efforts in the simple wheel and axle design (over 5500 years old).<p>On a side note, an engineer at our local power plant once told me that the inertial energy available on "the grid" was sufficient to cause one of the steam turbines at our local power plant to "jump out of the floor" should someone fully engage it out of phase.
AES did a pilot project of the same type with Altairnano batteries back in 08, you can see some of the numbers and details here (warning pdf): <a href="http://www.b2i.cc/Document/546/KEMA_Carina_validation_report_public_final.pdf" rel="nofollow">http://www.b2i.cc/Document/546/KEMA_Carina_validation_report...</a>
What about some sort of hydraulic "battery"? I have seen water pumps that can be used as generators and vice/versa, especially during offpeak periods, water is pumped uphill, and during peak demand, water is sent downhill to generate electricity.<p>Would a hydraulic turbine/pump/generator be feasible?