New spin for flywheel technology
Leland Teschler
Executive Editor
Aside from a few modest successes, flywheels have not lived up to this sort of hype. The basic science hasn't been in place to conquer fundamental problems that crop up when wheels spin at the 30,000 rpm-and-higher speeds necessary to deliver performance promised by the most wild-eyed flywheel proponents.
The conundrum is explained by NASA flywheel project Chief Engineer Ray Beach. NASA's Glenn Research Center in Cleveland has had flywheels under development for several years with the idea of applying them in scenarios ranging from orbiting space stations to satellites. But engineers there found a lot of holes in the technology when they first started serious work on the subject.
"People have been talking for 20 years as though this technology was right around the corner. It hasn't been around the corner," says Beach. "When NASA first got into flywheels about seven years ago, we assessed the different technologies available and felt the physics was there to build a unit that would handle the requirements for a space application. Initially we worked with several flywheel firms. We had some success but we found we were significantly off from what you needed to put something into space."
THE BASICS OF HIGH-TECH FLYWHEEL DESIGN
The allure is easy to understand. Flywheels potentially have a high-energy density and are pollution-free. That's why they have garnered interest through the years as a source of clean power for trains, trucks, and automobiles.
Many of those drawn to flywheel research have focused on getting high energy density by means of a light wheel that spins fast, usually 30,000 to 100,000 rpm. Managing the high inertia that results from such a design has proven to be problematic on a variety of fronts. Flywheel energy storage potential is proportional to mass moment of inertia and the square of rotational speed. Speed is limited by the strength-to-density ratio of the rotor material, so strong-but-light composite materials have gotten most of the attention.
Problem is that composites can delaminate or be subject to failures in the fiber/matrix material under the right conditions.
Flywheel researchers say they now have a handle on how to design wheels that won't burst during their projected lifespan. Even so, containment schemes able to hold debris flying off a high-energy wheel have gotten attention as well.
Flywheels operate in at least a partial vacuum to avoid the aerodynamic drag losses that would otherwise result. This means they must incorporate a vacuum pump (if they are not sitting in the vacuum of outer space), which itself dissipates power. The vacuum can complicate the dissipation of heat generated by ohmic losses in the bearing electromagnets and rotor. In addition, active magnetic bearings require sophisticated computer control to maintain levitation.
These systems usually have a rotating-field generator with the magnetic field supplied by rare-earth permanent magnets. The specific strength of these magnets is typically much less than that of the composite flywheel. Thus they must spin at much lower tip speeds and sit near the inner portion of the flywheel. This may compromise the power density of the generator.
The experience eventually led NASA to seek help from academia. "In the end, we found all the flywheel companies we worked with had similar capabilities. Most of the ones we've had experience with are small and have really sharp people. And most of them are underfunded so they are generally selling more than what they really have had in hand," says Beach.
One of the institutions tapped by NASA was Texas A & M's Center for Space Power, which has been researching low-loss and lowmass magnetic bearings. The term " bearing" is actually something of a misnomer here. Magnetic bearings on flywheels actually refer to a levitation system that raises the rotor shaft with magnetic fields so there is little or no physical contact while it turns. Electronic controls monitor the shaft position in real time and adjust fields to keep the rotor shaft centered.
Similarly, NASA went to the Center for Electromechanics (CEM) at the University of Texas in Austin for help with rotor technology. The problem NASA encountered initially was that rotor technologies wouldn't provide the kind of tip speeds (linear velocity at the outside radius of the flywheel) needed to give energy densities space applications require. Researchers at U of T came up with an all-composite rotor-plus-arbor design. This approach replaces composite material close to the shaft with a light and flexible structure dubbed an arbor. The arbor expands outward with higher rpm's at the same rate as composite material at the outer-most wheel diameter. This keeps the arbor from loading the outer-most ring elements and boosts flywheel life.
CEM has worked on several advanced flywheelprojects over the years, including efforts sponsored by Darpa (Defense Advanced Research Projects Agency) as well as feasibility studies for transit buses and a 3-MW, 15,000-rpm unit for trains. One conclusion coming out of this work: "Space is probably the best application for advanced composite flywheels right now," opines CEM program manager Joe Beno. "Economically it makes the most sense because there is such a premium put on weight and efficiency. But once you get flywheels made in sufficiently large quantities and their price comes down, they can be reasonable for a lot of areas."
One of these areas may be mass transit. In studies funded partly by the Houston Metropolitan Transit Authority, CEM used a 150-kW flywheel hitting 40,000 rpm as a replacement for chemical batteries on a hybrid-electric city bus. The wheel could be made to last the life of the bus, unlike chemical batteries which needed periodic replacement. CEM figures the flywheel was responsible for boosting fuel efficiency on a demonstration bus by 30% and acceleration by almost a factor of two.
CEM's transit bus project sheds light on some of the practicalities of fielding flywheels in vehicles. For example, researchers studied how bumps and potholes might affect a spinning wheel. They came up with a magnetic bearing for the device that kept the wheel suspended to within a 20-mil tolerance for impact loads of 3 g or less. More severe jarring (which topped out at about 8 g in tests there) would put the wheel momentarily down on a backup mechanical bearing.
Also of concern was the containment of wheel debris in the event of a failure. The need to reliably contain high-energy-density flywheels was highlighted nine years ago by the death of an engineer in Germany during a spin test of a composite flywheel. The same year, Darpa sponsored the Flywheel Safety Project to find means of safely keeping a disintegrating wheel from causing damage.
CEM participated in the Safety project and says that effective containment schemes are now well established. The flywheel on the transit-bus study, for instance, employed an aluminum housing that contained composite rings encircling the flywheel which serve as a soft landing zone for debris.
But the usual design approach is to sense when the flywheel is wearing out and retire it before catastrophic failure, says CEM's Beno. The typical sign that it's time for a new wheel is an out-of-balance condition. This happens when the outer portion of the wheel, which is the part under the most stress, begins to separate from the inner rings. Transducers in the magnetic bearings can note this problem and provide enough warning to spin down the wheel well before there's a chance of disintegration.
EFFICIENCY CONCERNS
The efficiency of driving the flywheel up to speed was another area of concern. NASA relied on fieldoriented control techniques borrowed from ac motor drives to ensure high torque and efficiency in delivering power to the wheel.
An extra complexity in spacecraft is the use of momentum wheels to provide attitude control and stability. The addition or removal of energy from one such flywheel applies torque to the corresponding axis of the spacecraft, causing it to react by rotating. Keeping the flywheel rotating at a constant velocity stabilizes that axis of the spacecraft. Several momentum wheels can be used to provide full three-axis attitude control and stability.
The problem when combining the momentum wheels and a power-generating flywheel is that both couple to the same motor generator. Disturbances caused by changes in the power load must not affect the momentum wheels. Lockheed Martin Corp. and Northrop Grumman Corp. helped NASA develop a control algorithm that effectively isolates the two systems.
NASA has put what it has learned into demonstrator wheels and is now working what's called the G3 flywheel. This is a 15-in.-diameter unit designed to reflect requirements for space flight. These requirements include redundancy in the key components. In the magnetic bearing, for example, loss of power to one magnet pole won't affect suspension of the rotor.
Wheels much like the G3 will eventually be used in space missions for efforts aimed at going to the Moon and Mars, says Beach. "The infrastructure for Moon and eventually Mars missions would have to be in place for many decades. You don't want the assets wasting away up there," explains Beach. "That's a strong selling point of flywheels. A flywheel sitting in space isn't degrading. And it is not like electrochemical power sources which have a specific number of charge and discharge cycles."
NASA thinks the basic flywheel technologies rotors, bearings, containment housings, and power conversion are in place. Efforts going forward in the field will mainly focus on optimizing systems for specific missions. One of these missions is in powering small satellites for low-earth orbits. The main problem such applications present is that of optimizing wheel performance in a small package. Flywheels would serve as the primary power source when the satellites lose sunlight for the solar panels 30 min out of every 90-min orbit. They would put out on the order of 4 kW.
Beach estimates that it would take the Agency about four years to develop a flywheel optimized for a specific deep-space mission; less if the need is for something closer to earth. "The further away you go, the less mass you want, and the more effort it takes," he says.
BACK-UP POWER
Despite all the interest in spacecraft and vehicles, the first practical applications for flywheel power has been where the wheel can sit in one place. Indeed, flywheel power far less exotic than what NASA is working on has been available for years. In 1996, a company originally known as Magnetic Bearing Technologies Inc. commercialized flywheel technology for use in back-up power supplies. Now known as Active Power, the Austin, Tex., firm incorporates flywheels in products that combine power line conditioning with uninterruptible power sources.
Active Power's flywheels and the shaft on which they spin are forged as one piece from 4340 steel. They rotate at the relatively leisurely rate of 7,700 rpm, well below the plastic deformation level of the steel. Even so, the company sacrifices every fiftieth wheel it makes to check for anomalies that might make the wheel disintegrate.
Because rpms are low, the wheel housing primarily serves as a means to pull a rough vacuum rather than as protection. The flywheel spins on bearings that are not particularly high tech, though Active Power has patented their design. The levitation system unloads only about 80% of the flywheel weight. The rest, about 100 lb, sits on ceramic ball bearings. The benefit of this scheme, says Active Power, is more stability than where the wheel is to be completely suspended by magnetic fields.
The typical application for these flywheels is to provide power between the time the mains go dead and when an emergency supply (usually a diesel generator) kicks in. These transition times generally span tens of seconds at most. The flywheels work well here because statistics show that 96% of all power interruptions span 10 sec or less. Customers for the units include hospitals, process plants, and broadcasting facilities.
Power needs spanning tens of seconds are the sweet spot for flywheel backup units, says Active Power. The company foresees making incremental improvements in its flywheels. But it sees no need for more sophisticated flywheel technology, such as wheels made of composites, to be in its offerings anytime soon.
VOLTAGE SUPPORT
Nevertheless, composite flywheels have recently made a debut in dc power devices. Pentadyne Power Corp. in Chatsworth, Calif., produces flywheel systems for power quality that are based on technology originally devised to power vehicles. The company received its first purchase orders two years ago. Pentadyne focuses just on flywheels; it partners with other firms that incorporate the wheels into end products for the UPS market.
Pentadyne's carbon-fiber rotors spin at about 55,000 rpm and are designed to handle 60,000 rpm. They are levitated by magnetic bearings that consume about 150 W.
Its vacuum system differs from that of many other flywheel systems for terrestrial use in that it is integrated into the enclosure that houses the wheel. It uses a turbomolecular-style pump whose rotors mount on the flywheel shaft. (As a quick review, turbomolecular pumps create a vacuum with rapidly spinning rotors that push gas from their inlet towards the exhaust. Most employ multiple stages consisting of rotor/stator pairs mounted in series. Gas captured by the upper stages gets pushed into the lower stages and successively compressed.) Because compression varies linearly with circumferential rotor speed, the shafts on Pentadyne's high-speed flywheels serve as highly efficient basis for these pumps.
The housing system surrounding the flywheels is designed to contain flywheel shrapnel in the event of a catastrophic failure. It uses inner and outer housings made of mild steel with coolant fluid filling the area between them. The two housings connect to each other by light attachments that are designed to shear in the event of a wheel disintegration.
If the worst happens, the inner housing is designed to contain shrapnel from the wheel, helped by a carbon fiber overwrap banding its center section. Meanwhile, torque imparted to the inner housing from the high-energy debris shears the attachments to the outer housing. The inner housing begins rotating, opposed by the friction of the coolant between the inner and outer housing. The idea is to spread torque transmitted by flywheel components out over time. This heats the fluid somewhat, but keeps the burst energy from ever reaching the outer housing.
Like the flywheel systems from Active Power, Pentadyne dc power sources target applications needing about 20 sec of energy or less. (However, the company says one customer uses the device to deliver three minutes of power, though at a low rate of energy use.)
Are such systems practical? It would seem so for at least 23 different installations. That is the number of units Pentadyne says it has either installed or is in the process of bringing on line.
MAKE CONTACT:
Active Power, www.activepower.com
NASA Glenn Research Center, www.grc.nasa.gov
Pentadyne Power Corp., www.pentadyne.com
Texas A & M Center for Space Power, www.engineer.tamu.edu/tees/csp.
University of Texas Center for Electromechanics, www.utexas.edu/research/cem