2012 - Cryocoolers For Aircraft Superconducting Generators and Motors
2012 - Cryocoolers For Aircraft Superconducting Generators and Motors
2012 - Cryocoolers For Aircraft Superconducting Generators and Motors
Ray Radebaugh
ABSTRACT
INTRODUCTION
NASA has set the following performance goals for future transport aircraft: (1) reduced
airport noise, (2) reduced emissions (both pollutants and greenhouse gasses), and (3)
reduced fuel burn [1]. With reduced noise, aircraft could fly into and out of small centrally
located metropolitan airports. To do so also requires a capability for short take-off and
*
Contribution of NIST, not subject to copyright in the US. The use of company names does not imply
endorsement by NIST, but they are mentioned only to identify unique equipment relevant to this study.
171
TABLE 1. Dates and technology development for future aircraft [1].
Generation Approximate date Technology
N+1 2015 Conventional tube & wing
N+2 2020 Unconventional hybrid wing-body
N+3 2030 Advanced aircraft concepts
landing (STOL). Future generations of aircraft are designated as N+1, N+2, etc., with N
being today’s technology. TABLE 1 gives the approximate time frame for new aircraft
technology development to meet the NASA performance goals. The performance goals for
N+3 aircraft set by NASA are very stringent, such as a 55 dB noise reduction, 75 %
emission reduction, and a 70 % fuel burn reduction. One concept being investigated by
NASA to meet the N+3 goals is a distributed turboelectric propulsion concept as illustrated
in FIGURE 1 [2]. With this concept superconducting motors are used to drive propulsor
fans distributed over much of the upper surface of the trailing edge of the wings. With
current high-bypass jet engine technology, the fans connected to the engine provide about
85 % of the engine thrust with the trailing exhaust providing only 15 % of the thrust. The
fan speed is limited to a few thousand RPM in order to prevent supersonic speeds at the
blade tips. Electric power to the motors can be provided by two superconducting
generators driven by high-speed gas turbines. The generators and turbines can be made
very efficient and compact by running them at much higher speed than that of the motors.
The generators and motors would be connected through an “electrical gearbox.” This
revolutionary concept, which decouples speed and torque, provides many control
advantages in addition to the potential of meeting the N+3 performance goals.
The electrically driven propulsion system is feasible only if the electrical motors can be
about the same size as, or smaller than, existing gas turbines. FIGURE 2 shows a
comparison of power densities for electric motors and aircraft engines taken from data by
Masson, et al. [3]. The values shown in this figure represent general trends, but values for
various motors and engines can vary significantly. The specific power of gas turbine cores
(without propulsor) for use on the Boeing 787 is about 16 kW/kg, whereas the specific
power of most conventional electrical motors is limited to about 0.5 kW/kg, although very
large industrial motors can have power densities up to about 3 kW/kg. Some electric
motors developed for use on automobiles have power densities as high as 4.8 kW/kg [4].
Therefore, because of their low power density, conventional electric motors cannot be used
for aircraft propulsion. High power density electric motors are a part of NASA and DoD
research and development plans. Higher power densities require higher current densities in
the conductors and/or higher magnetic energy densities in permanent magnets. The output
Superconducting motor-driven-
fans in a continuous nacelle
Wing-tip mounted
superconducting
turbogenerators
FIGURE 1. Artist’s concept of future turboelectric aircraft utilizing superconducting generators and
motors [2].
172
FIGURE 2. Power density comparison for gas turbines and electric machines.
REFRIGERATION REQUIREMENTS
173
use. First, we must establish the requirements for the cryocooler, such as the operating
temperature, the net refrigeration power, the maximum input power, and the maximum
weight.
A superconductor normally operates within a three-dimensional space limited by
magnetic field, temperature and transport current. The superconductor goes normal
whenever any one of these limits is exceeded. To remain superconducting while carrying
high currents, the superconducting wire must operate at temperatures well below the
critical field value to allow for an adequate margin. For Bi-2223, or first generation HTS
wire, the operating temperature must be around 30 K to 40 K. Second generation HTS
wires, made with YBCO, can operate at about 50 K to 60 K. For N+3 aircraft operating
around 2035, YBCO and possibly even better HTS wire would be available. For this study,
operating temperatures of 50 K to 70 K are used as the goal to be provided by cryocoolers
in future aircraft [7].
The heat load on the cryocooler for any HTS motor or generator consists mainly of (1)
the active heat dissipation from the HTS winding due to AC losses, (2) heat conduction in
the current leads, and (3) background heat leaks due to radiation and due to conduction
through shafts and supports. One quality measure of any superconducting machine is the
ratio of required refrigeration power to the electrical (generator) or mechanical (motor)
power output of the superconducting machine. This quality measure can be defined as the
cryogenic inefficiency of the superconducting machine, given by
Q
λsc = 1 − ε sc = net , (1)
Ps
where λsc is the cryogenic inefficiency, εsc is the cryogenic efficiency, Q is the net heat
net
load on the cryocooler, and Ps is the power rating of the superconducting machine.
The cryogenic efficiency is different from the overall efficiency of the motor or
generator, which is given by the ratio of the output power to the input power. The
difference between the input and output powers of an electrical machine is the total loss,
which is in the form of heat. However, some of the losses in a superconducting machine
will occur in the ambient-temperature components, such as the copper windings of the
stator or hysteresis losses in iron. Also, the background thermal loss in a cryogenic system
is not part of the overall loss associated directly with the superconducting machine.
Published numbers for λsc in discussions of HTS machines are seldom reported, but such
numbers would be very helpful in planning for future refrigeration requirements and sizing
of cryocoolers. In 2009 Sivasubramaniam reported measurements with a 1.3 MW
generator designed for airborne applications [8] that had a design power density of 8.8
kW/kg at full power. It was a HTS homopolar inductor alternator that utilized a stationary
HTS field excitation coil, a solid rotor forging, and an advanced stator. The HTS coil for
this configuration of generator operates with DC current, so AC losses are not important.
The cooling requirement for this machine was 40 W at about 30 K. The cryogenic
inefficiency for this machine is then λsc = 3 x 10-5. A 4 MVA generator developed by
Siemens that used HTS for the rotating field coil required a refrigeration power of about 50
W at 30 K [5]. For this machine λsc = 1.3 x 10-5.
To achieve the very high power densities discussed earlier (as high as 40 kW/kg), all-
superconducting machines (rotor and stator) would be necessary. In such machines, the
cryogenic losses are dominated by the AC losses in the superconductor. AC losses in
currently available YBCO tapes are much too high to consider their use in the armature,
where large AC currents are required and high frequencies are needed for high rotational
speeds. However, much research is underway to reduce AC losses in YBCO. When low
AC loss YBCO becomes commercially available, such conductors could be used in future
174
high-power density HTS machines for aircraft use. Barnes et al. [9] estimated total losses
for such advanced machines. For a 5 MW machine, they estimated that the total loss would
be about 500 W at 65 K, which results in a cryogenic inefficiency λsc of 1 x 10-4.
For a 30 MW machine with a cryogenic inefficiency of 1 x 10-4 the heat load to 65 K
would be 3 kW. A somewhat more conservative projection for AC losses in YBCO would
have λsc = 5 x 10-4, in which case the refrigeration power is 15 kW at 65 K. For
demonstrations within the next ten years we may need up to about 10 kW of refrigeration at
a temperature of 50 K. However, as advances are made in HTS, the required refrigeration
power may decrease somewhat and the operating temperature may increase to about 65 K.
For an aircraft with the size and power of the Boeing 737-200, the expected electrical
power for thrusters would be about 10 MW, which requires 1 kW of refrigeration when the
cryogenic inefficiency is 1 x 10-4. For purposes of this report we focus on refrigeration
powers in the range of 1 kW to 10 kW for temperatures in the range of 50 K to 65 K.
Superconducting generators and motors offer many advantages for use on aircraft, but
the power input and mass of cryocoolers required to maintain the necessary cold
temperature results in a disadvantage for such systems. The overall system becomes
successful only if the cryocooler input power and mass are small compared with the power
output and mass of the superconducting machinery. We define these ratios as
P
XP = c (3)
Ps
and
M
XM = c , (4)
Ms
where Pc is the input power to the cryocooler, Ps is the power rating of the superconducting
generator or motor, Mc is the mass of the cryocooler, and Ms is the mass of the
superconducting machine. The power fraction of the cryocooler can be expressed in terms
of the machine cryogenic inefficiency λsc and the second-law efficiency ηc of the
cryocooler as
λsc (Th − Tc )
Xp = , (5)
ηcTc
where Tc is the cold temperature and Th is the hot or ambient temperature of the cryocooler.
FIGURE 3 shows how the power fraction varies with λsc and ηc for the case of Tc = 50 K
and Th = 300 K. (Th could be reduced in flight.) This figure shows that cryocooler
efficiencies of only 10 % to 20 % of Carnot yield power fractions less than 0.1 for λsc <
0.001. However, we shall see in the following paragraph that to achieve low mass
fractions, the cryocooler efficiency must be closer to 30 % of Carnot. For a net
refrigeration power of 10 kW at 50 K, the input power becomes 167 kW when the
efficiency is 30 % of Carnot.
The mass of the cryocooler can be expressed as
M c = mc Pc = mc X P Ps , (6)
where mc is the specific mass of the cryocooler (mass per unit of input power). When both
sides of equation (6) are divided by the mass of the superconducting machine, we have
X M = mc X P ps , (7)
175
FIGURE 4. Ratio of cryocooler mass to HTS
FIGURE 3. Ratio of cryocooler power to HTS machine machine mass for various cryocooler specific
power for various cryocooler efficiencies vs. cryogenic masses vs. power fraction. HTS machine power
inefficiency of the HTS machine. density is 20 kW/kg for this figure.
where ps is the power density of the superconducting machine. FIGURE 4 shows how the
mass fraction varies with power fraction and the cryocooler specific mass for the case of a
superconducting machine power density of 20 kW/kg. We note that with an expected
power fraction of only 0.01, the specific mass mc of the cryocooler must be less than about
3 kg/kW if the mass fraction is to be kept below about 0.6. Reducing the power fraction to
0.005 would allow the specific mass to be increased to 6 kg/kW for the same mass fraction,
but the efficiency would have to be doubled to 60 % of Carnot for the same λsc and ps. For
a cryocooler specific mass of 3 kg/kW, the power density is 0.33 kW/kg. This power
density is much less than that of conventional high-power electric motors, as shown in
FIGURE 2. An alternate expression for equation (7) is XM = mc*λscps, where mc* is the
cryocooler mass per unit of net refrigeration power. However, most of the cryocooler mass
is that of the compressor, so mc* becomes a strong function of the cryocooler efficiency.
In order to compare the efficiency and specific mass of existing cryocoolers to the
requirements discussed in the previous section, we turn to the results of cryocooler surveys.
A survey by Strobridge [10] in 1974 is often used as a comparison for any newly developed
cryocooler. The extensive survey of 235 cryocoolers by ter Brake and Wiegerinck [11] in
2002 provides a valuable update of the Strobridge survey. However, many of the surveys
lump cryocooler cold-end temperatures together, which result in 80 K cryocoolers
dominating the reported performance. The recent survey by Ladner [12] shows how
efficiency and mass vary with temperature. Ladner’s survey included most new Stirling
and pulse tube cryocoolers reported between 1999 and 2009.
Efficiency
FIGURE 5 shows how the efficiency of cryocoolers varies with refrigeration power.
The lower curve in this figure is from the 1974 survey of Strobridge, and the upper curve is
the upper limit of efficiency from the ter Brake survey [11], as given by Kittel [13]. This
figure shows that in the range of 1 to 10 kW of refrigeration power, the maximum
efficiency is about 20 to 25 % of Carnot and has not increased much in 35 years. FIGURE
6 shows data from Ladner [12] for the highest efficiencies of Stirling and Stirling-type
pulse tube cryocoolers and how they vary with temperature. We note that maximum
176
FIGURE 5. Cryocooler efficiencies from Strobridge [10] and ter Brake [11] with the upper solid line
from Kittel [13] showing the maximum values. Capacity is the same as net refrigeration power.
efficiencies of about 25 % of Carnot are reported for Stirling cryocoolers in the temperature
range of 50 K to 65 K, whereas pulse tube cryocoolers show maximum efficiencies in the
range of 10 % to 15 % of Carnot in that temperature range. Most of these cryocoolers have
refrigeration powers of only a few watts, although one of the pulse tube cryocoolers
considered in the survey had a refrigeration power of about 1 kW, and two large
commercial Stirling crocoolers of 1 kW and 4 kW at 77 K have efficiencies of about 25 %
of Carnot in this temperature range. The largest Stirling cryocooler ever made provided 20
kW of refrigeration at 70 K with an efficiency of 41 % of Carnot [14]. Although this
industrial cryocooler was much too heavy (49 kg/kW) for aircraft application, it shows that
an efficiency goal of about 30 % of Carnot in large Stirling cryocoolers for temperatures of
50 K to 65 K should be realistic. Further research on large pulse tube cryocoolers is
necessary to overcome flow nonuniformities if such high efficiencies are to be achieved.
Specific Mass
FIGURE 7 shows how cryocooler mass varies with input power from the CILTEC
survey of ter Brake [11] and the Ladner survey [12] compared with the goal for aircraft use.
The graph also shows mass data for five relevant cryocoolers with power inputs greater
177
FIGURE 7. Comparison of cryocooler mass from surveys with mass goal for aircraft applications.
Because the specific mass goal is a greater challenge than the efficiency goal, we
discuss here some approaches to reducing the mass of cryocoolers. Obviously, several
engineering tasks include the use of high-strength and light-weight materials in place of the
more customary materials wherever possible. Compact packaging arrangements, such as
those used in the MCC Stirling cryocooler mentioned in the previous section, are also
needed. Here we discuss three science-based approaches that could be useful in reducing
mass.
178
Higher Frequency and Pressure in Regenerative Cryocoolers
The compressor accounts for most of the mass in Stirling and pulse tube cryocoolers.
Operating them at higher frequency increases the power density proportional to the
frequency. However, frequencies higher than about 60 Hz have usually resulted in
significantly reduced efficiencies because of poor regenerator performance. Radebaugh
and O’Gallagher [19] showed that high efficiency can be maintained at higher frequencies
by simultaneously making the following changes to the operating conditions and the
regenerator geometry: (a) decrease regenerator volume, (b) decrease hydraulic diameter,
and (c) increase average pressure. Experiments with a 120 Hz pulse tube cryocooler
verified that high efficiency can be maintained when the modifications discussed above are
made [20]. FIGURE 8 shows calculated results for the cold-head efficiency of a 50 K
regenerative cryocooler operating at 60 Hz and at 120 Hz. The peak values are at optimum
conditions for each frequency. The calculations assume an expansion-space efficiency of
85 % due to nonisothermal processes in the expansion space (either a displacer or a pulse
tube). For the same refrigeration power (same cold-end mass flow m c ), the higher
frequency leads to a smaller gas cross-sectional area Ag and a shorter length L. The
regenerator volume and mass are reduced by a factor of three when increasing the
frequency from 60 Hz to 120 Hz and the pressure from 2.5 MPa to 5.0 MPa. The higher
pressure used for 120 Hz requires twice the tube wall thickness (taken into account in the
calculations for efficiency), but the additional mass for the regenerator is negligible. The
cold head efficiency decreases only slightly, from 42 % at 60 Hz to 38 % at 120 Hz. The
higher frequency and pressure should lead to a compressor volume that is reduced by about
a factor of four and a mass that is reduced slightly less than a factor of four due to the
thicker walls needed for the higher pressure.
With the cold head efficiency of about 40 % of Carnot, the ambient temperature part
(aftercooler and compressor) must be about 75 % efficient in converting electrical power
into acoustic power at 300 K to achieve a total system efficiency of 30 % of Carnot. That
requires careful matching of optimum compressor and optimum cold head acoustic
impedances. It also requires a displacer to recover the cold-end acoustic power, which is
about 17 % of the acoustic power at the regenerator warm end. A displacer is an integral
part of the Stirling cryocooler, but it can also be used at the warm end of the pulse tube in a
pulse tube cryocooler.
FIGURE 8a. Calculated second-law efficiency of FIGURE 8b. Calculated second-law efficiency of
a regenerative cryocooler cold head at 50 K optimized a regenerative cryocooler cold head at 50 K optimized
for a frequency of 120 Hz and 60 Hz vs. specific area. for a frequency of 120 Hz and 60 Hz vs. length.
179
Higher Pressure and Pressure Ratio in Brayton Cryocoolers
The Brayton cryocooler is a recuperative type with steady flow and pressure. For
long-life space applications, centrifugal compressors and turbo-expanders supported on gas
bearings are used. Such devices work best at low pressures and pressure ratios, which lead
to low power densities. However, the very high rotation speed of the compressor and
expander (1000 to 4000 rev/s) compensates for the low pressure and results in very high
power densities for the compressor and expander. In the recuperative heat exchanger, the
relevant power density is the Gibbs free energy flow per unit volume flow rate, which for
an ideal gas is given by
0T ln( PH / PL )
nR
gv = = P ln Pr , (8)
V
where n is the molar flow rate, V is the volume flow rate in the heat exchanger, R0 is the
universal gas constant, T is temperature at any location, PH is the high pressure, PL is the
low pressure, Pr is the pressure ratio, and P is the pressure in the flow channel of interest.
The low pressure side (typically about 0.1 MPa) of the heat exchanger has the lower power
density and will dominate the size of the heat exchanger. With the low pressures and
pressure ratios (about 1.6) used in turbo-Brayton cryocoolers, the power density in the heat
exchanger is very low, so the heat exchanger is the largest and heaviest component in the
system. The low power density also requires very high effectiveness in the heat exchanger
to yield high cryocooler efficiency. Research on ways to use higher pressures and pressure
ratios in Brayton cryocoolers is needed to reduce their specific mass to 3 kg/kW.
Multistage turbines could provide higher pressure ratios, but would add to the cost.
About 80 % of the mass of large Stirling or Stirling-type pulse tube cryocoolers is due
to the pressure oscillator (compressor) [6, 12]. Presently about 30 % of the compressor
mass is due to the linear motor. As efforts are made to reduce the mass of the compressor
through the use of lightweight materials, the motor mass could make up 50 % of the
compressor mass because of the requirement for using high density magnets, iron, and
copper. About 50 % of the linear motor mass is due to the iron in most conventional linear
motors [6]. Thus, research into ironless linear motors could lead to low-mass compressors.
The use of a Halbach array [21] for the permanent magnets is one approach that can be
used to eliminate the back iron in a linear motor [22]. In a Halbach array a set of
permanent magnets are arranged with the magnetization direction rotated by 90° with each
adjacent magnet, as shown in FIGURE 9. The magnetic field lines nearly cancel on one
side of the array and are combined and amplified by a factor of about 1.4 on the other side
FIGURE 9. Halbach array of permanent magnets to enhance magnetic field on one side and cancel field
on other side.
180
compared with a conventional array [22]. The higher field leads to an increase by a factor
of two for the power density, and no back iron is required in the magnetic field path. A
second Halbach array parallel and near the first array results in a very high field between
the two arrays with very little field on the outside. For a cylindrical device the arrays can
be curved into a circle with the use of magnetized arc segments. A comprehensive study of
power density of a linear motor with Halbach arrays would be very useful. A
superconducting linear motor [23] is also a possible option to increase the power density of
a linear compressor, but it complicates the startup procedure before the cryocooler reaches
the low temperature necessary for the superconducting linear motor to become effective.
CONCLUSIONS
Transport aircraft with significantly reduced airport noise, emissions, and fuel burn are
desired and planned for the year 2035. One proposed approach to meet these goals relies
on distributed turboelectric propulsion in which two high-speed gas turbines drive close-
coupled superconducting generators that power a set of distributed lower-speed propulsor
fans driven by superconducting motors. The superconducting machines will require about
1 to 10 kW of refrigeration at about 50 to 65 K. This paper shows that cryocoolers for this
application need to have an efficiency of about 30 % of Carnot and a specific mass less
than about 3 kg/kW of input power. Surveys of recent cryocoolers show that most high-
efficiency cryocoolers have efficiencies of about 20 % of Carnot and specific masses of
about 10 to 30 kg/kW. Considerable research is needed to meet the low specific mass goal.
The use of lightweight materials, higher frequencies, higher pressures and higher pressure
ratios are suggested to meet such goals. Innovative approaches to the elimination of iron in
cryocooler motors can also lead to reduced mass.
ACKNOWLEDGEMENT
Funding from the NASA/Glenn Research Center for this study and valuable
discussions with Gerald V. Brown are gratefully acknowledged.
REFERENCES
1. Luongo, C.A., Masson, P.J., Nam, T., et al., “Next Generation More-Electric Aircraft: A Potential
Application for HTS Superconductors,” IEEE Trans. Appl. Superconductivity 19 (2009) pp. 1055-1068.
2. Felder, J.L., Kim, H.D., Brown, G.V., “Turboelectric Distributed Propulsion Engine Cycle Analysis for
Hybrid-Wing-Body Aircraft,” in 49th AIAA Aerospace Sciences Meeting, Orlando (2011), AIAA-2011-
0300.
3. Masson, P.J., Brown, G.V., Soban, D.S., and Luongo, C.A., “HTS machines as enabling technology for
all-electric airborne vehicles,” Supercond. Sci. Technol. 20 (2007) pp. 748-756.
4. FutureDrive editorial, “Electric Motor Comparison,” Oct. 3, 2007, Word Press;
http://futuredrive.wordpress.com/2007/10/03/electric-motor-comparison/.
5. Masson, P., “Superconducting Machines,” Presentation at Workshop on Turboelectric Propulsion for
Transport Aircraft, NASA/Glenn, Feb. 10, 2009.
6. Radebaugh, R., and Ladner, D., “Cryocoolers for Airborne Superconducting Generators and Motors,”
NIST Tech Note, to be published.
7. Barnes, P.N., Levin, G.A., and Durkin, E.B., “Superconducting Generators for Airborne Applications
and YBCO Coated Conductors,” in Proceedings of the Power and Energy Society General Meeting –
Conversion and Delivery of Electrical Energy in the 21st Century, (2008) IEEE, pp. 1-4.
181
8. Sivasubramaniam, K., Zhang, T., Lokhandwalla, M., Laskaris, E.T., Bray, J.W., Gerstler, B., Shah, M.R.,
and Alexander, J.P., “Development of a High Speed HTS Generator for Airborne Applications,” IEEE
Trans. Appl. Superconductivity 19 (2009) pp. 1656-1661.
9. Barnes, P.N., Sumption, M.D., and Rhoads, G.L., “Review of high power density superconducting
generators: Present state and prospects for incorporating YBCO windings,” Cryogenics 45 (2005) pp.
670-686.
10. Strobridge, T.R., “Cryogenic Refrigerators – an Updated Survey”, NBS Tech Note 655 (1974).
11. ter Brake, H.J.M., and Wiegerinck, G.F.M., “Low Power Cryocooler Survey”, Cryogenics 42 (2002) pp.
705-718.
12. Ladner, D.R., “Performance and Mass vs. Operating Temperature for Pulse Tube and Stirling
Cryocoolers,” Cryocoolers 16 (2011), pp.633-644.
13. Kittel, P., “Cryocooler Performance Estimator,” Cryocoolers 14 (2007) pp. 563-572.
14. Dros, A.A., “An industrial gas refrigerating machine with hydraulic drive,” Philips Technical Review 26
(1965) pp. 297-308.
15. Breedlove, J.J., Magari, P.J., and Miller, G.W., "Cryocooler for Air Liquefaction Onboard Large
Aircraft," Adv. Cryogenic Engineering 53 (2008) American Institute of Physics, pp. 838-845.
16. Potratz, S.A., Abbott, T.D., Johnson, M.C., and Albaugh, K.B., "Stirling-Type Pulse Tube Cryocooler
with 1 kW of Refrigeration at 77 K," Advances in Cryogenic Engineering, 53, (2008) pp. 42-47.
17. http://www.stirlingcryogenics.com/ (Efficiencies are reduced from manufacturer’s input power vs. Tc
and capacity vs. Tc data curves in “Process Coolers” section).
18. White, M., Infinia, private communication (2010).
19. Radebaugh, R. and O’Gallagher, A., “Regenerator operation at very high frequencies for
microcrycoolers” Adv. Cryogenic Engineering, 51 (2006) pp. 1919-1928.
20. Vanapalli, S., Lewis, M., Gan, Z., and Radebaugh, R., "120 Hz pulse tube cryocooler for fast cooldown
to 50 K,” Appl. Phys. Letters, 90, issue 7 (2007) 072504.
21. Halbach, K., “Design of Permanent Multipole Magnets with Oriented Rare Earth Cobalt Material,”
Nuclear Instruments and Methods 169 (1980) pp. 1-10.
22. Jang, S.M., Choi, J.Y., Cho, H.W., and Lee, S.H., “Thrust Analysis and Measurements of Tubular Linear
Actuator With Cylindrical Halbach Array,” IEEE Transactions on Magnetics 41 (2005) pp. 2028-2031.
23. Oswald, B., Best, K-J., Maier, T., Soell, M., Freyhardt, H.C., “Conceptual design of a SC HTS linear
motor,” Supercond. Sci. Technol.17 (2004) S445-S449.
182
Copyright of AIP Conference Proceedings is the property of American Institute of Physics and its content may
not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written
permission. However, users may print, download, or email articles for individual use.