Levitated flywheels with conventional magnetic bearing have fairly high energy loss rate cause by eddy currents. In an early posting I pointed out that Pentadyne's composite flywheels for UPS systems will consume their entire stored power in a little over two hours. More than an order of magnitude improvement in energy consumption would be required for 12 hours load shifting. I found an article concerning superconducting magnetic bearings for flywheels posted a the website of the International Technology Research Institute at Loyola College in Maryland. They claim that orders of magnitude improvement in flywheel energy consumption can be achieved with superconducting bearings. Here are some excerpts from Chapter 4 (Other Applications) of the web page Power Applications of Superconductivity in Japan and Germany:
Flywheel energy storage is not a new technology. In fact, it is widely used in industry in applications as diverse as punch presses and high field pulsed magnets. Its stored energy density (usually expressed in Watt-hour/kg) is, however, relatively low. In addition, its practical use has been limited to areas where the energy holding time is short because of the excessive rotational loss caused by bearings and windage. The NEDO program addresses the energy density issue through work on high-strength, high-speed flywheels and addresses holding time through development of a very low friction HTS magnetic bearing.
Most commercial flywheels are made of metal and rotate at relatively low speeds to maintain the tensile stresses in the flywheel within reasonable limits. Since the stored energy is directly proportional to the mass of the flywheel but proportional to the square of the rotational speed, increases in rotational speed yield a large benefit in energy density. Today's state-of-the-art flywheels therefore employ fiber-reinforced plastics (FRP) rather than metal. These materials can be engineered to have very high strength in the radial direction, thus permitting higher operational speeds. In addition, they can be designed to fracture into many small pieces in the event of a structural failure. This reduces the size of the required containment vessel and allows the system to be mounted above ground. (Safety considerations have led to designs of metal flywheel systems where the flywheel itself is underground). While the first demonstration flywheel in the NEDO program was made of steel, all subsequent devices will employ FRP. An additional benefit of the FRP flywheel, as seen in the following section, is that the lighter flywheel simplifies the design of the magnetic bearing.
Superconducting magnetic bearings
The primary factor preventing the application of flywheels to long-term energy storage is loss in the bearings. Any mechanical bearing with contact between the stationary and rotating parts will have enough loss to render the system uneconomical (Higasa 1994). One solution to the problem is to use a non-contact active magnetic bearing that employs conventional electromagnets. The rotational loss of such a bearing is 1-10% that of a mechanical bearing under the same operating conditions. The problem, however, is that the bearing itself consumes power, which is dissipated as heat in the copper electromagnets, and the bearing and cooling system power consumption must be included in the calculation of the overall system efficiency. A reasonable magnetic bearing consumes a few watts for each kilogram of flywheel weight, depending on the structure of the bearing and the control system, and this loss is sufficient to make a system using copper electromagnets uneconomical. Superconducting magnetic bearings, on the other hand, have demonstrated losses of 10-2 to 10-3 watts per kg for a 2,000 rpm rotor. This translates to an overall one-day, "round-trip" system efficiency of 84%, which is acceptable.
Figures 4.3 and 4.4 show the basic operation of the flywheel and bearing. The flywheel is at room temperature and carries on its lower surface a permanent ring magnet that rides above the superconducting portion of the bearing, which is simply an array of pellets of YBCO. The YBCO is kept at 77 K by an external supply of liquid nitrogen. The YBCO traps the magnetic flux produced by the permanent magnet, and as long as the pinning force of the YBCO is not exceeded, the bearing generates restoring forces to counter any relative motion of the permanent magnet and the superconductor. Thus, the bearing is completely passive and does not require the complex feedback and control circuits needed for conventional magnetic bearings.
In Germany, the Forschungszentrum Karlsruhe (FZK) directs a flywheel energy storage program that uses melt-processed YBCO for the bearings. FZK's Institut für Technische Physik (INFP) has advertised and offered for commercial sale semifinished cubes, cylinders, and rods of melt-processed YBCO. The institute has built and tested a 300 Wh flywheel model using superconducting bearings combined with permanent magnets. The flywheel system uses disks constructed from advanced carbon fibers. It was operated from 30,000 to 50,000 rpm and produced 10 kVA/300 Wh at 50,000 rpm. The electric drive system for the flywheel was developed in cooperation with the Institut für Elektrische Maschninen und Antriebe of Stuttgart University. The next step in the program is to develop a larger flywheel with a capacity of about 7 Wh/250 kVA, together with Siemens and accompanied by application studies with utilities. This flywheel will probably be a "stacked" design similar to that used in the NEDO program. A composite, rather than a steel, flywheel is planned with a diameter of approximately 80 cm. Large, above-ground installation with appropriate protection, is favored over below-ground excavation, due to the projected lower installation expense. It is expected that large energy delivery at > 20 MW would most likely experience some problems with respect to heating of the motor/generator. Flywheels appear attractive also for railway application combined with regenerative braking for charging. Siemens is also conducting studies using flywheels for spinning reserve.
The above extracts are largely self explanatory. The article does not explicit mention the energy input requirements for cooling the high Tc superconductor, but presumably they are relatively small or the bearing could not achieve the advertised energy consumption rate of a few watts per kilogram.
This article was written in 1997, and I have not been able to detect any signs that flywheels with superconducting bearings are heading toward real world applications. Either this technology has unresolved problems or the economics does not look attractive. My feeling is that the economics of flywheels are best in applications with many short cycles such a frequency regulation or regenerative braking. For such applications the extra expense and complication of superconducting bearing may not be justified.