power densities 10 percent lower than those quoted for a cell. The basic cells have also exhibited
long life (over 100,000 charge/discharge cycles) and have very low open circuit current loss, with
self discharge to half the original voltage occurring in about four days. 1l5SRI International is developing a thin-film lithium polymer ultracapacitor, and it has projected a
specific energy of 70 Wh/kg and a specific power rating of 50 kW/kg,ll6 which corresponds to an
order of magnitude increase over other ultracapacitor types. It is not clear whether such goals
actually will be achieved.Although the progress in ultracapacitor technology has been remarkable, it should be noted that
the technology is still in the early development stage. It is difficult to forecast the performance and
cost parameters for a “fill-scale” ultracapacitor that can contain 5 kWh of energy, for example.
Many in the ultracapacitor industry believe that the DOE midterm goals of 10 Wh/kg energy
density and a cost of $1/Wh could be attained in the next five to eight years, suggesting that a
commercial product could be introduced in about 10 to 12 years. Peak power densities of over 2
kW/kg appears to be feasible for such devices, with storage efficiencies in the 93 percent to 95
percent range.^117
Flywheel energy storage has been researched for decades, but recent progress has been
attributed to improvements in materials and bearing technology. The energy stored by a flywheel
is directly proportional to its mass but proportional to the square of its rotational speed, so the
key to storing large quantities of energy is to increase speed--speeds of 100,000 rpm and higher
have been contemplated. The flywheel can absorb and release energy very quickly, with the major
limitation being the capability of the power electronics and stator to handle high peak power.
Energy storage capability is limited by flywheel material properties, as well as safety
considerations in the event of rotor failure (the cost and weight of the containment system
increases with energy stored).
The only flywheel actually installed and tested in an automotive environment for which data are
publicly available is a relatively low-performance system built by Magnet Motor MIX. The
system uses a rotor operating at a maximum speed of 12,000 rpm, to provide performance levels
of 750 W/kg power and about 5 Wh/kg energy, levels similar to those of an ultracapacitor. The
system uses conventional bearings and has worked satisfactorily in an urban bus.ll8
Oak Ridge National Lab has constructed an experimental system using samarium-cobalt
permanent magnets and a water cooled stator with a carbon-fiber flywheel rim. The estimated
performance characteristics of such a system are 50 Wh/kg energy density and 1.5 kW/kg density
power, indicating an energy density roughly comparable to an advanced Ni-Cd battery.^119 These
figures, however, seem very high relative to other flywheels that have been built. American
Flywheel Systems (AFS), in conjunction with Honeywell, claims even higher figures for its
1l5E. Blank, “Ultracapacitors for Automotive Applications," paper presented at the Automotive Technology Development Contractors
Coordination Meeting, U.S. 116 Department of Energy, October 1994.
CALSTART, letter to OTA, May 1995.(^117) Murphy andKramer, see footnote 114.
(^118) G. Reiner,"Experience with the Flywheel Storage System in Diesel Electric and Trolley Public Transport System,” Flywheel Energy Systems
Technology Workshop, 119 November 1993.
M. Belanger, “Workshop Summaryand Observation” Flywheel Energy Systems Technology Workshop, November 1993.