WITH six crew members, and some 1200 cubic metres of space to keep habitable
24 hours a day, the International Space Station will be one of the most
power-hungry objects ever to orbit the Earth. When the Sun shines on it, the
space station鈥檚 huge array of solar panels will generate more than enough
electrical energy for this purpose. But every 60 minutes or so, it will plunge
into darkness as it passes into the Earth鈥檚 cold shadow. As the power from the
solar panels fades, batteries will take over.
During the next 30 minutes, the cells must provide 110 kilowatts of
electricity at 110 volts. After this sharp burst of activity the cells will
spend the next 60 minutes recharging with power from the solar arrays. This
charging and discharging takes its toll: after five years and some 30 000
cycles, the batteries, which are hugely expensive, will be exhausted and will
have to be replaced.
But there is an alternative. Two small American companies are developing
flywheels that might one day store and provide enough energy to power satellites
and even the space station for twenty years or more in orbit. Engineers at
NASA鈥檚 Lewis Research Center near Cleveland, Ohio, are eager to test the
competing versions of the flywheel. They say that flywheels have the potential
to be lighter and less complex than today鈥檚 battery stacks, that they can store
more energy and that they do not need any of the toxic chemicals that batteries
rely on.
Advertisement
What鈥檚 more, by careful regulation of the way they spin, a set of flywheels
could at the same time be used to control the attitude of a spacecraft.
In March, NASA will test these ideas using a flywheel developed by US
Flywheel Systems based in Newbury Park, California. About a year later, SatCon
Technologies in Cambridge, Massachusetts will deliver its own version so the two
can be compared. If the tests are successful, the plan is to use the flywheels
to power not only the International Space Station when its first set of
batteries run out early next century, but communications, remote sensing and spy
satellites. Flywheels might even help power NASA鈥檚 X-33 rocket when it flies,
crewless, in 1999.
The two companies are locked in battle to produce the flywheels of the
future鈥攂ut their approach is based on the same principles. When a flywheel
rotates, it stores kinetic energy. This energy can be retrieved by using the
flywheel to drive a dynamo that produces electricity. And the energy can be
replaced by passing a current through the dynamo so that it works like a motor,
increasing the flywheel鈥檚 speed.
Explosive flaws
The design of the flywheel is crucial. Double its mass and you double the
energy it stores at a given speed of rotation. But doubling the speed of
rotation quadruples the amount of stored energy. So the big challenge is to
design a flywheel that can rotate tens of thousands of times a minute, and it is
no mean task.
The breakthrough that is making such flywheels possible is the development of
extremely strong lightweight materials such as composites of carbon fibre and
epoxy resin. 鈥淩ight off the bat, carbon composite is a quarter of the weight of
steel and four times as strong,鈥� says Jack Bitterly, the veteran engineer who is
the brains behind the work at US Flywheel Systems.
Using this material, his team has developed a flywheel the size of a small
bicycle wheel that is capable of spinning at 55 000 revolutions per minute. The
outer rim of the flywheel travels at more than 5000 kilometres per hour,
creating forces that would tear apart any other material.
So Bitterly builds his wheels carefully. Any tiny flaws in a composite wheel
could cause the device to explode. Bitterly makes each wheel from a single
strand of carbon filament that is drawn through a resin bath and then wound onto
the flywheel, like thread on a cotton reel.
The forces generated in the flywheel are so great that they stretch the
fibres. 鈥淭he composite material is very plastic and so the wheel expands as it
spins,鈥� says Ray Beach, the NASA engineer who is in charge of flywheel research
at Lewis. 鈥淚f you watched a flywheel in slow motion you would see the material
诲别蹿辞谤尘颈苍驳.鈥�
Since this stretching and deforming cannot be prevented, it must be
controlled. For Bitterly and his 20-strong team, this means that any deformation
must occur symmetrically around the flywheel鈥檚 axis of rotation. If there are
any holes in the material to start with, the process of deforming forces the
fibres to realign to fill them. This in turn changes the centre of mass, causing
vibrations that send the wheel spinning out of control.
Just how Bitterly prevents these holes is a carefully guarded secret,
although he acknowledges that maintaining a constant tension in the filament
during the winding process is crucial.
SatCon has a different approach, which it hopes will allow flywheels only 15
centimetres across to spin at up to 90 000 rpm. The team begins with a woven
carbon composite blanket and moulds it into a disc. Then they inject epoxy resin
into the mould, which hardens to form a solid disc that can be used as a
flywheel.
However each team does it, there is no room for error. 鈥淣ASA headquarters has
told us that we can鈥檛 allow one of these to fail in orbit,鈥� says Beach. A
failure is a spectacular event in which the wheel explodes into a cloud of
talcum powder-like dust. In orbit, this cloud of debris would become a
significant hazard to other spacecraft.
To determine the limits of the performance, NASA is testing computer models
that simulate the way a wheel deforms. 鈥淲e have a lot of confidence that we
understand what鈥檚 happening,鈥� says Beach. Once the limits of failure have been
established, the wheels will operate at only 75 per cent of this speed to
guarantee a safe margin.
Designing the wheel is only one part of the problem: engineers must also
perfect magnetic bearings that support the wheel while it is spinning, without
causing any friction. 鈥淲e鈥檙e lucky because we can rely on others for this
technology,鈥� says Beach, pointing to the next generation of gas turbine aircraft
engines which will use magnetic bearings instead of mechanical ones. Today鈥檚
flywheels use a combination of active and passive components鈥攁 permanent
magnet supports the wheel while sensors and electromagnets monitor the position
of the axle and nudge it into line when it strays.
Changing attitude
A magnetic bearing has other advantages too. For example, it allows the
spinning wheel to find its own axis of rotation, unlike a mechanical bearing
where the axis is fixed. This allows for small movements in the axis that occur
when the material deforms. And since the entire wheel spins in a vacuum, such a
bearing does not suffer from friction with the air. 鈥淭here is very little to
slow the flywheel down,鈥� says Bitterly.
However, the relative motion of magnets on the axis and in the bearing
induces stray currents that generate heat and this loss of energy does cause the
wheel to slow down. Beach says that this will be less of a problem in orbit. In
microgravity conditions, the forces required to control the flywheel are small,
so engineers are working on entirely active bearings in which the magnetic field
can be varied. Once in orbit, the field strength would be reduced, minimising
the losses caused by heating.
Magnetic bearings must also be able to cope with conditions on Earth. In
future, for example, Beach expects the flywheels to be used and tested by
satellite manufacturing companies in California. 鈥淪o they鈥檒l have to be
earthquake rated.鈥� And during a launch, when forces can reach several times that
of gravity, the flywheels will sit in a mechanical bearing.
Perhaps the most difficult challenge is to work out how to retrieve energy
from a flywheel while at the same time using it for attitude control. Spinning
masses called reaction wheels are already used routinely to control the attitude
on spacecraft. These are small metal wheels spinning at a few thousand rpm that
can transfer their angular momentum to and from the spacecraft to change its
attitude.
But this task will be more difficult with flywheels because they spin much
faster. 鈥淥ur flywheels will have several orders of magnitude more momentum,鈥�
says Beach. So small changes in their speed can result in a relatively large
transfer of momentum. Consequently, the attitude control system will have to
monitor the wheels鈥� rotational speed extremely accurately.
In March, Beach will begin testing an attitude control platform consisting of
two flywheels spinning in opposite directions. By changing the speeds of the
wheels, it is possible to transfer momentum to and from the platform, thereby
changing its attitude. The tests will involve floating the platform on an
air-bearing table at the Lewis Research Center and monitoring the vibrations
that these changes create. 鈥淲e have to demonstrate that vibrations can be
controlled and minimised,鈥� says Beach.
He is confident that these problems can be solved. 鈥淭here are engineering
challenges but there are no showstoppers,鈥� he explains. With a working system,
engineers will be able to take advantage of the biggest benefit of
flywheels鈥攖he amount of energy that they can store in a given mass. This
energy density is measured in watt-hours per kilogram, and for conventional
rechargeable batteries that can be bought in shops it is 40 Wh/kg.
The batteries used in spacecraft don鈥檛 come close to this figure. To reach
the relatively high voltages needed, these cells have to be mounted in series
and so require complex electronics to determine their charge and to spread the
load equally among them. This dramatically increases their weight, so that
although spacecraft batteries rely on the same chemistry as terrestrial cells,
their energy density is much less than 10 Wh/kg.
Bitterly鈥檚 flywheels, on the other hand, even when all their accompanying
electronics have been taken into account, have an energy density of 44 Wh/kg.
What is more, this energy can be converted into useful electricity with an
efficiency of up to 96 per cent compared with only 80 per cent for
batteries.
And flywheels have the potential to do even better. SatCon is working on
smaller flywheels with a diameter of only 15 centimetres that have an energy
density of 88 Wh/kg. The improvement is due mainly to the use of lightweight
electronic components. This design will be ready in a year or so. Taking a
longer view, James Kirk, a mechanical engineer at the University of Maryland,
College Park, is working on a design that he hopes to test on a satellite within
ten years.
He claims that his flywheel can spin at up to 600 000 rpm. The weak point in
the other designs, he says, is the link between the fibre and the axle. So his
flywheel is a cylinder with a hole through the middle instead of an axle,
allowing it to spin much faster. The device can store as much as 250 Wh/kg, he
says.
Beach believes that flywheels will play an important part in powering
satellites in the future. As satellites become bigger, their power demands will
increase and building bigger and bigger stacks of cells is less viable .
鈥淔lywheels have great promise,鈥� he says. But he admits that there will always be
a place for the humble battery in small satellites that require low power.
鈥淏atteries will always be up there,鈥� he says.
