heels
(Post 1973) to provide electricity storage for the US
power grid. Unfortunately, achieving dynamic sta-
bility and structural integrity in composite flywheels
proved far more difficult and costly than expected.
The largest commercial units constructed to date are
400 times smaller than those Dr. Post envisioned
in spite of a critical need and a huge potential mar-
ket. Now, discoveries made during the course of a
magnetic levitation transportation project have fi-
nally opened the door to construction of utility-scale
flywheel electricity storage systems. We call these
devices Power Rings. Figure 1 shows what a 400
kilowatt-hour Power Ring might look like.
LaunchPoint began development of Power Ring
technology in 2002, under contract to the US Navy,
with a study of the potential for short duration, very
high power units. This was followed by a contract in
2003 from the New York State Energy R&D Au-
thority (NYSERDA) to design a long duration elec-
tricity storage unit for utility applications. These
contracts enabled detailed analysis of the Power
Ring concept, and led to additional development
contracts from the Department of Energy and the
National Science Foundation, which are in progress
now (2006).
Figure 1. A 400 kWh Power Ring
2

ELECTRICITY STORAGE TECHNOLOGIES
Technologies such as pumped hydro, compressed air
energy storage (CAES), batteries, fuel cells, super-
conducting magnetic energy storage (SMES), ultra-
capacitors (or supercapacitors), and flywheels can all
be used to provide electricity storage. Flywheels,
SMES, and ultracapacitors all cost too much for use
in large installations. Fuel cells and flow batteries
show promise but are also projected to be expensive.
Lead-acid batteries are used in small systems, such
Third Generation Flywheels For High Power Electricity Storage
O.J. Fiske, M.R. Ricci
LaunchPoint Technologies, Inc., Goleta, California, USA


ABSTRACT: First generation flywheels of bulk material such as steel can mass tens of tons, but have low en-
ergy storage density. Second generation flywheels of composite materials have higher energy storage density
but limited mass due to structural and stability limitations. LaunchPoint is developing high energy third gen-
eration flywheels "Power Rings" using radial gap magnetic bearings to levitate thin-walled composite
hoops rotated at high speed to store kinetic energy. Power levels exceeding 50 megawatts and electricity stor-
age capacities exceeding 5 megawatt-hours appear technically feasible and economically attractive. Power
Rings can be used to decrease the peak power requirements of electric transportation systems by supplying in-
termittent high power for vehicles such as maglev trains. They can also store braking energy, isolate the
power grid from surges and spikes, reduce the incidence of transportation system power outages, and provide
back-up power in case of blackouts. as uninterruptible power supplies, and a few moder-
ately large installations. Costs, concerns regarding
toxic materials, sheer mass, and space requirements
prevent their widespread use in large installations.
CAES systems can be scaled up to large capaci-
ties, but need a fuel supply and underground com-
pressed air storage caverns or land area for com-
pressed air storage pipelines. They are similar to
turbine power plants, i.e. large and noisy, making
them unsuitable for many areas. They also have a
cold spin-up time of 15 minutes, making them im-
practical for some applications.
Pumped hydro was the premier storage system for
decades, with over 22 gigawatts of capacity installed
in the US and over 30 gigawatts in Japan (Bradshaw
2000). Capital costs for existing plants were low
when they were constructed $250/kW for a Ten-
nessee Valley Authority installation in the early
1980s and $800/kW for the Rocky Mountain
Pumped Storage facility constructed near Rome,
Georgia in the early 1990s but are now estimated
to be $1,100-$2,000/kW, making them economically
much less attractive. A 1600 MW system completed
in Japan in 2001 cost $3.2 billion. Prospects for ad-
ditional pumped hydro facilities in many countries
are limited.

The best sites have already been used
and the remaining sites are remote, requiring new
transmission lines. Reservoir construction also pre-
sents major environmental impact issues.
Flywheels of various forms have been used in in-
dustry for hundreds of years or more, and both first
generation (iron or steel) and second generation
(composite) flywheels are now used for electricity
storage. Power costs for commercial flywheels are
higher than some other energy storage technologies,
but data from the Electricity Storage Association,
which factor in efficiency and expected longevity,
show flywheels to be highly competitive for applica-
tions involving frequent charge-discharge cycles
(ESA 2006). Unfortunately, for fundamental techni-
cal and economic reasons, they have been restricted
to 6 kWh or less in most commercial applications. If
they could be scaled to larger capacities at reason-
able cost, they would clearly provide great benefit in
many applications.
3

POWER RING TECHNOLOGY
3.1

The Flywheel Dilemma
Figure 2 illustrates the basic design of a second
generation flywheel. The rim is attached by spokes
or a hub to a central shaft, which is supported by
bearings. A motor-generator operates as a motor to
spin the flywheel to store energy, and as a generator
to extract stored energy. The kinetic energy stored in
Figure 2. Flywheel structure
the rotor (rim) is proportional to the mass of the ro-
tor and the square of its velocity. The equation for
stored kinetic energy is:

K.E. = ½ J 2
= ½ kmr
2 2
(1)


where
is the rate of rotation in radians per second,
J is the moment of inertia about the axis of rotation
in kilogram-meters squared, m is rotor mass, r is ro-
tor radius (also known as the radius of gyration), and
k is an inertial constant dependent on rotor shape.
Stress produced in the rim is proportional to the
square of linear velocity at the tip. When rotor speed
is dictated by the rotor fabrication material, the
maximum linear tip velocity is constant, regardless
of rotor radius. The maximum rotation rate is then
inversely proportional to rotor diameter.
The best materials for flywheels are not the dens-
est, or even the strongest they are those with the
highest specific strength, i.e. the ratio of ultimate
tensile strength to density. For a thin rim, the rela-
tionship of maximum rim stress to specific energy
(energy stored per unit mass) is:

K.E./m = h
/2
(2)

where h
is the maximum hoop stress the ring can
withstand in N/m
2
and
is the density of the ring
material in kg/m
3
. So, specific energy corresponds
directly to specific strength, h
/
, and filament-
wound rotors made of high strength, low density fi-
bers will store more energy per unit weight than
metal rotors.
Since energy is proportional to the square of
speed, high performance is attained at high tip
speed. Carbon fiber rims have attained tip speeds in excess of 1000 meters per second and are housed in
evacuated chambers to minimize energy losses and
heating due to friction.
As rotational velocity increases, the rotor experi-
ences increasing radial force causing it to expand
faster than the shaft. The spoke or hub assembly
must compensate for this differential growth while
maintaining a secure bond with the rim. High-speed
carbon composite rims can expand by more than 1%
in normal operation, E-glass even more. Hoop stress
is highest at the inner boundary of the rim and
causes a common failure mode in which the rim
separates from the spokes.
Hoop stress decreases rapidly from the inner
boundary of the rim to the outer boundary. The fi-
bers used in the construction of the rim are ex-
tremely strong along their length, but are held to-
gether in the radial direction only by relatively weak
epoxy binder. This results in another common fail-
ure mode in which the rim delaminates or fractures
due to radial stress, which peaks at a point part way
between the inner and outer edges. The longer the
rim radius the higher the forces become.
Many methods have been proposed to alleviate
these problems, but a fundamental limitation re-
mains in all present designs the rotating mass is far
from the axle while the stabilization system (bear-
ings and actuators) operates directly on the axle. If
the arbor or spokes are flexible enough to expand as
rpm increases, then the stabilization system must
transmit control forces to the rim through a floppy
structure an impossible task but if the structure is
rigid it will delaminate under high radial stress. The
only way to resolve this conflict, so far, has been to
restrict composite flywheels to small diameters.
3.2

A new class of magnetic bearing
In a permanent magnet Halbach array (Halbach
1985), the field produced by each magnet reinforces
the fields of all the other magnets on the active
side of the array, and cancels them on the other side.
The result is, in essence, a one-sided permanent
magnet with an intense field. When two identical
Halbach arrays are placed with their active sides fac-
ing each other, they produce powerful repulsive, at-
tractive, or shear forces, depending on alignment.
As compared to simple opposed pole faces, a 5-
element Halbach array provides more than three
times as much force per unit volume of magnet.
This provides the basis for the shear-force levita-
tor, shown in Figure 3.
Here two Halbach levitation arrays are arranged
vertically with the static array attached to a sta-
tionary support. If the moving array is now offset,
from the initial position shown, upward to th