IEEE TRANSACTIONS ON MAGNETICS, VOL. 25, NO. 1, JANUARY 1989

~

243

APPLICATION OF SUPERCONDUCTING TECHNOLOGY TO EARTH-TO-ORBIT ELECTROMAGNETIC LAUNCH SYSTEMS

John R. Hull Argonne National Laboratory, Argonne, IL 60439

Lynnette M. Carney National Aeronautics and Space Administration, Lewis Research Center, Cleveland, OH 44135

ABSTRACT

This paper explores the potential benefits that may occur as a result of incorporating superconductors, both existing and those currently under development, in one or more parts of a large-scale electromagnetic launch (EML) system that is capable of delivering payloads from the Earth's surface to space. Both rail accelerator and coaxial magnetic accelerator devices are considered, using reference concepts developed under previous NASA studies as the basis for compari- son. The use of superconductors for many of the EML components results in lower system losses; conse- quently, reductions in the size and number of energy storage devices are possible. Applied high-temperature superconductivity may eventually enable novel design concepts for energy distribution and switching. All of these technical improvements have the potential to reduce system complexity and lower payload launch costs.

INTRODUCTION

Compared to a rocket, electromagnetic launch (EML) of payloads from deep gravity wells is attractive because of its intrinsically higher energy effi- ciency. Several mission studies have examined Earth- to-space EML of nonfragile cargo [l-41 and moon-to- space launch of raw materials and nonfragile cargo [5- 81. The use of either rail or coaxial accelerator

PROJECTILE VELOCITY,

KM/SEC

LOO

10

1

0.1 0.0

systems appears technically feasible for these missions; however, large capital costs for contemporary designs indicate that such systems will not be economi- cally attractive until some time in the future when high launch rates are needed - of the order of 6000 kg of payload per day.

EML technology has been seriously studied in recent years [9], with a large amount of current effort focused on impact-fusion and defense-oriented applications with relatively small-scale devices. EM accelerators that have been tested are typically only a few meters long with bore sizes less than 100 cm2. As indicated in Fig. 1, the projectile masses, accelerator lengths, and energy requirements of existing EML facilities are many orders of magnitude smaller than those required for large-payload space launch systems. Given these large differences, simple scale up of con- ventional designs is not likely to result in optimized designs. Rather, novel designs may be necessary to make EML technology superior to competing technologies at this scale.

This paper discusses the incorporation of super- conducting technology in the design of Earth-to-orbit (ET01 EML systems in ways that have the potential to significantly reduce the capital costs. We consider both rail accelerator and coaxial EML technologies and compare designs that use nonsuperconductors, conven- tional superconductors, and high-temperature supercon-

- LINES OF CONSTANT PROJECTILE ENERGY

\r- 1 0.1

\ \ \ EML-MIT \

\ I 1 \ I I 10 100 1KG 10 100 1 TON 10 10

0 Demonstrated Technology 0 Mission Studies

PROJECTILE MASS LEGEND : ANU - Australian National University LANL - Los Alamos National Laboratory LeRC - NASA Lewis Research Center LLNL - Lawrence Livermore National Laboratory MIT - Massachusetts Institute of Technology

Fig. 1. (in terms of projectile mass and final velocity); only open literature references included.

Comparison of representative existing EML devices with advanced space propulsion mission requirements

0018-9464/89/0100-0243$01 .WO 1989 IEEE

244

ductors (HTS), currently under development. HTS refers to the class of ceramic superconductors (e.g., yttrium- barium-copper-oxide and bismuth-strontium-calcium- copper-oxide) that have been shown to definitely have superconducting properties at temperatures above the boiling point of nitrogen ( 7 7 K).

While the use of superconductivity may increase the benefits of EMLs at any location, this paper focuses on an Earth-based launch system. An Earth- based EML could be developed in the shortest time period, and it would have the largest potential near- term impact, because present high costs for taking payloads to orbit are generally considered the largest obstacle in the commercialization and development of the space habitat. The advantages of an Earth-based system include relatively low-cost construction, maintenance and resupply. Disadvantages include the necessity of punching through a relatively thick atmosphere and a relatively high launch velocity, requiring a longer accelerator.

The mission reference concept [2] chosen for this analysis is the launch of 6000 kg (rail) or 3000 kg (coaxial) projectiles to a velocity of 7 km/s at an acceleration of 1225 g's through a 2 km-long launcher with a bore of about 1 m, elevated 20 degrees from the Earth's surface. Each projectile carries propellant and rocket for final insertion of the 600 kg payload into Earth orbit.

MISSION STUDIES

Large-scale EMLs have been suggested for delivering bulk cargo [i.e., nonfragile supplies, orbital transfkr vehicle (OTV) propellants, and materials for space processing] to space. Locations for EML devices are usually suggested for the bottom of large gravity wells, such as the surface of the Earth, Moon, or Mars, where rocket transport of payloads to orbit is inefficient and expensive. Some proposals have considered orbiting EML devices, either for transport from low Earth orbit (LEO) to geosynchronous Earth orbit or to take surface launched objects up to orbital velocity. One of the earliest suggested applications was ejection of radioactive wastes from the Earth to interstellar space [l]. More recent thinking has concentrated on application of EML systems to commercial space-based development [l-81.

For Earth-based launchers the ablation of projec- tile mass and kinetic energy loss is an important scaling parameter. The losses essentially scale with surface area, whereas the kinetic energy of the projec- tile scales with volume for a fixed density. Thus, for a constant aspect ratio projectile, the fraction of mass and energy lost during a launch through the Earth's atmosphere diminishes as the size increases.

Rice et al. [l] assessed the potential feasibility and benefits of a conceptual Earth-to-space rail launch (ESRL) system capable of a variety of missions. The primary application, that of deep-space disposal of nuclear waste, required a launch velocity of 20 km/s at a maximum of 10,000 g ' s acceleration. The secondary application, Earth-to-orbit launching of nonfragile cargo, required velocities of 5 to 10 km/s at 2500 g's. Cost estimates for the ESRL system performing both applications range from $5 to 8x10' (1981 dollars). The costs included total research, develop- ment, and investments with a 30 yr amortization. Annual operating expenses (personnel and supply costs only) were estimated as $60M. At two launches per day, it would cost $560/kg to dispose nuclear waste. At a rate of eight launches per day, the cost of launching bulk cargo to Earth orbit was estimated as $590/kg.

Miller et al. [2] investigated all types of EML concepts, in addition to the rail accelerator, for various space propulsion missions. The primary mission application chosen for development was the ET0 launch of bulk cargo to support an orbiting Space Station with

the delivery of supply items, OTV propellants, and materials for space processing facilities. Of all EML types reviewed by Miller et al. [2], only the coaxial magnetic accelerator showed promise equal to or superior to that of a rail accelerator. Total research, development and investment costs were estimated as $2.2~10' in 1981 dollars. Annual operating costs were estimated as $40M. At a launch rate of one payload per day, the cost per kg would be $757 when amortized over a 30 yr period. At a rate of 10 launches per day the cost would be $234/kg.

The above studies [1,2] concluded that large-scale applications of both the rail accelerator and the coaxial EHL are not only technically feasible but also economically beneficial as a means of delivering nofragile bulk cargo to space. The technology assess- ments required substantial extrapolation of the state- of-the-art in EML technology; however, the two studies found no insurmountable technical barriers to exist and no areas that required major technological break- throughs. Advanced superconducting technology was not incorporated in the above analyses. Although the ET0 system has a relatively low operating cost, there is an extremely large capital expenditure, and therefore the need for such a system is predicated on a large material delivery requirement.

DESIGN OPTIONS

The major obstacle for ET0 EML is the large capital expenditure, and designs that significantly reduce this cost are of most value. Although the launch energy itself is very inexpensive (<$l/kg to LEO), the cost of storing it with rapid access is expensive. A major challenge to EML development is therefore power conditioning, or more specifically, power compression. With this in mind, this section examines the impact of applied superconductivity on the various EML subsystems of both rail accelerator and coaxial magnetic accelerator systems, evaluated with regard to the promise it has in reducing either system complexity or cost.

Energy Storage

A limitation of the railgun is that it is basically a single-turn motor, and it therefore requires a very high current at relatively low voltage. Investigations of multi-turn railgun concepts are still very preliminary [lo]. Typical rail ac elerato operation involves currents of the order of

Homopolar generators (HPGs) are an attractive power source because they are very efficient at the high currents; however, they do not provide sufficient voltage output (back emf) to drive a rail accelerator. Consequently, HPGs are generally matched to a low impedance inductor to provide the required current and voltage.

The previous Battelle studies [1,2] envisioned cryogenically cooled [using liquid nitrogen (LN2)] inductors matched to the homopolar generators for both the rail accelerator and coaxial EML concepts. The rail accelerator system, operating at high current at relatively low voltage, requires many BPGs with individual LN2 inductors; the coaxial magnetic launcher, high voltage at relatively low current, would require a single LN2 coil.

Use of superconducting inductors between the HPGs and the rails offers the opportunity to increase the energy transfer efficiency in the switching circuits and thereby significantly reduce the number of HPG/inductors required. Preliminary estimates of benefits derived from using superconducting materials in the inductors are shown in Table I.

The primary function of the HPG is to bring the inductor up to voltage. For nonsuperconducting inductors this "charging" needs to be relatively fast

10 s to 10' (or higher) Amperes.

245

Table I. Distributed Energy Storage Benefits for an Earth-to-Orbit Rail Launcher as a Function of Energy Transfer Efficiency

Overall Stored Energy No. of HPG/Inductor Distance Transfer Inductor Delivered Units Required Apart Efficiency Energy to Rails (percent) (MJ) (MJ) (m)

72 48.0 40.8 3585" 75 48.5 42.0 3448 80 50.1 44.8 3232 85 51.6 47.6 3044 90 53.1 50.4 2875 95 54.6 53.2 2722 99 55.7 55.4 2613

0.57 0.59 0.63 0.67 0.71 0.75 0.78

* Battelle study [2] results, based on the following assumptions: - Overall transfer efficiency 72% (0.85 HPG to inductor, 0.85

- 56 MJ HPG coupled with an inductor that stores 48 MJ at 4 MA. inductor to rails)

to minimize resistive heating losses in the coils. However, for superconducting coils the charging can be as long as the period between lauches. The inductor charging could then be readily accomplished by means of a Graetz bridge circuit [Ill, similar to that envisioned for superconducting magnetic energy storage coils for utility load leveling or pulsed fusion magnets. The input to the Graetz bridge would lead directly to the power transmission lines, the output to the inductors. Each HPG would be replaced by three thyristor (or possibly superconducting - see below) switching units. F o r a charging time of one hour, each Graetz bridge unit would need a power rating of 16 kW to charge a 56 MJ coil.

With inductor energy storage, advantage is gained if the energy can be stored in one large inductor. For example, a thin coil of radius R and fixed ampere turns has stored magnetic energy that is approximately proportional to Rln(R). The cost of the coil is roughly proportional to R, and therefore the cost of stored energy varies as l/ln(R).

Rails and Components of Rail Accelerator Systems

One of the biggest drawbacks to a simple rail accelerator system is that it is an inherently inefficient device in that only a fraction of the energy supplied to the accelerator is converted into the kinetic energy of the projectile. This inefficiency is due to the high resistive heating losses in the rails and armature and also to the energy which remains stored in the magnetic field between the rails. Excessively long accelerators can have prohibitive resistive losses.

In the reference design [ 2 ] the required current in the 2-km long rail is 16.8 MA. Because of high- frequency skin effects, the current flows on the inside portion of the rail with a skin depth 6 , given by [12]

where p is the electrical resistivity of the rails, t is the time elapsed since the projectile passed, and is the permeability of the rails. The time of flight in the launcher is about 4/7 sec, s o when the projectile is leaving the muzzle, the usuable depth of copper at the breech is only 16 cm, even if the rail is much thicker than this. Over most of the rail length, the skin depth will be much l e s s . The resistance through both rails R as a function of time T since onset of acceleration is

R(T) = (pu/~)'/2[8a/(3n)1 T3'2 , ( 2 )

where a is acceleration (12250 m2/s), and W is the rail height (1.0 m). For t = 4/7 sec, when the projectile

is at the end of the muzzle, the resistance is 1.3~10-~ Q. The joule heating power dissipated in both rails is P(~=4/7 s ) = 367 GW, which compares with 490 GW being delivered to the projectile at that time. The energy E dissipated in joule heating is found by integrating eq. ( 2 )

E = (2/5)rP(~) , ( 3 )

E(~=4/7 S ) = 84 GJ , which compares with 145 GJ of kinetic energy delivered to the projectile. The above calculations typically lead to the suggestion of a distributed energy system (DES) to increase the efficiency of an ESRL, such as that proposed by Refs. 1 and 2 .

An alternative option would use actively cooled superconducting rails to carry the large currents, with a copper ( o r other highly normal conducting, such as amzirc) veneer to transfer the current to the arma- ture. The advantage of this option, which would r quire HTS to be viable, is that the absence of large

single large inductor coil connected to the rails at the breach of the accelerator. Such a design would benefit from the scaling laws discussed in the previous sect ion.

Resistive l o s s e s would still occur in the arma- ture, as in DES designs. There would be some Joule heating in the copper as the projectile passes, until the current could diffuse into the superconductor. Using the ESRL concept developed by Miller, et al. [ 2 ] as the basis for comparison, this scheme greatly reduces the number of homopolar/generator inductor units required in a conventional segmented rail accelerator. Table I lists the benefits for the multi- stage rail launcher as the energy transfer efficiency is increased through the use of superconducting materials.

If the resistance heating is decreased by a factor of ten and all of the rail heat is transf rred to

approximately 50 m3 of LN2 is required (about $10K per launch. used instead, about 3000 m3 is required (about $6M) per launch. Based on cooling costs alone, it appears that this design option will require HTS to be feasible.

Of equal concern as resistive heating is, that in a breech-powered circuit, an amount of energy equivalent to the projectile kinetic energy is stored in the magnetic field of the rail inductance. This energy must be dissipated or recovered after the projectile leaves the muzzle. With a DES system only a small amount of magnetic energy is in the rail system, and these l o s s e s are then not significant.

I s R losses enables the launcher to be powered by a

boiling liquid nitrogen (latent heat 161 MJ/m 9 ), then

If liquid helium (latent heat 2.62 MJ/m 3 ) is

246

Another reason for a DES design is that the railgun is usually operated at the maximum acceleration that either the launcher or projectile can tolerate. The accelerating force is limited by the need to hold the rails together against magnetic bursting forces, which are the equivalent of the gas pressure in a gun barrel. It is then desirable to have a constant current in the rails to keep the acceleration at its maximum. If the current falls with increasing projectile position, then the launcher must be made longer to make up for the reduced acceleration. To minimize this effect in a single coil system, the energy stored in the large coil must be significantly larger than that delivered to the projectile.

Superconducting Augmentation to a Railgun

A third major option for railguns involves the use of a superconducting dipole magnet to augment the force on the projectile of an otherwise conventional railgun system, allowing the use of smaller currents in the rails and possibly alleviating some of the switching constraints and component stresses. A dc dipole magnet would provide a strong magnetic field along the length of the rails. Except for the ends, the dipole would consist of a set of long superconducting cables of a relatively simple fabrication.

Related studies [13,14] have shown that a super- conducting augmented railgun offers both increased muzzle velocities of the projectiles and/or higher launch efficiencies over an unaugmented accelerator.

One major technical issue with this type of system is that the size and weight of the superconducting magnet makes the augmentation impractical. HTS could potentially reduce the requirements, also increase the current capacity of the magnet, make the massive cryogenic portion unnecessary, and provide significantly higher magnetic fields.

Projectile Coils for Coaxial Magnetic Accelerator System

Coaxial EMLs can be classified by the way synchronization is achieved and by how the projectile coil current is obtained. For low velocity devices the projectile current can be obtained by commutator brushes, and this technique can be used to accelerate very large masses. The barrel coils can take the form of a single helically-wound coil. However, for velocities higher than about 1 km/s, commutation is no longer feasible [15].

There are two alternatives to brush-fed excitation. One is to short-circuit the projectile coils and induce a current into them [16]. The second is to use superconducting projectile coils energized with a persistent current before the bucket is launched [SI. In either case, the barrel coils usually take the form of discrete loops, each separately pulsed. An advantage of the noncontacting nature of the projectile coils in this approach is that there is no wear on the system, an advantage over the rail accelerator as well

Based on an analysis of pinning forces [17], superconductors can easily withstand the generated forces in EML devices. Of greater concern is the heat generated in the superconductors. In a coaxial EML the bucket coils experience a constant magnetic field to first order, i.e., they see essentially a dc magnetic field. An ac component to this field occurs because of the discreteness of the drive coils and because the bucket does not travel on its equilibrium trajectory, but rather experiences small oscillations about it. An ac field component may also appear due to nonperfect timing in switching circuits. The ac field induces eddy currents in the projectile coil. While these currents are much smaller than those that must be induced in a nonsuperconducting coil, they are

~ as the brush-fed method.

nevertheless large enough to drive a conventional superconducting coil nonsuperconducting. If this occurs, the projectile coil current soon disappears and acceleration stops.

Wipf [17] estimates that the ac loss of a su erconductor for a 0.1 T peak field is 1 to 10 pJ per cm' surface per cycle. In the reference design [2] each projectile coil carries a current of 250 kA along a circumference of about 70 cm. Drive coils are spaced every 16 cm. If wg assume a superconducting current density of lo5 A/cm , then each projectile coil has a volume of 245 cm3 and a surface area of 450 cm2. Assuming 10 uJ/cm2 is deposited for every drive coil, there is a total energ3 deposition in the projectile coils of about 200 mJ/cm .

Conventional superconductors can only tolerate 10 (for NbTi) to 100 (for Nb3Sn) mJ/cm3 of heat addition before losing the superconducting state, and in this example the nonuniform field experienced by an accelerating projectile would add more than this amount of energy to the projectile coils. By compari on, the new HTS materials may absorb about 100 J/cmp before losing superconductivity. This 1000 to 10,000-fold increase definitely enhances the prospects of super- conducting projectile coils.

Drive Coils for Coaxial Magnetic Accelerator System

Superconductivity can be used in the drive coils to reduce the coil size and increase the mutual inductance between drive and projectile coils. HTS materials will tolerate, in theory, much higher magnetic fields than conventional superconductors, thus allowing a smaller clearance gap between drive and projectile coils and therefore higher accelerations, at least from a magnetics criterion.

Energy Switching

Energy switching can be a large cost component of either a rail accelerator or coaxial EML system. For both rail accelerators and coaxial launchers, switches must be able to handle currents on the order of 1 MA, several tens of kV, and switching times of microseconds to milliseconds.

For coaxial magnetic launchers, switching and control is even more critical than for a rail accelerator [16]. The drive coils in a coaxial EML can be made with multiple turns, to make them operate at lower currents and higher voltages than rail accelerators. At the same time, the coils are only closely coupled when they are close together, say within one coil diameter. This requires that the power pulse in the drive coil be closely synchronized with the passing of the projectile. The projectile's short transit time through the active region of a drive coil reduces the time for signal propagation and switch turn-on. Because of the short time available for pulsing the drive coil, a large dI/dt is necessary, and higher speeds result in higher voltage requirements. At 7 km/s, the drive-current pulse voltages are about 100 kV [2,18]. The switching issue is not one of technical feasibility (large accelerator magnets are pulsed with comparable power and shorter times), but rather of economics.

While considerable technological improvement has been made in recent years [19,20] in large-power, pulsed power components and supplies, and continuing improvement is expected, large amounts of power in short pulses still require a premium price. Most solid-state switches have voltage limitations or are difficult to open. Recently, however, Gate-Turnoff Thyristors have been operated in a series configuration that promises high voltage capability [21,22].

The availability of HTS may open the way for low- cost, large-current, high-voltage, opening switches with fast switching times. Several designs using

247

to zero. At the time of this writing, the highest verified Tc is about 110 K. For operation at liquid nitrogen temperature (77 K), it is important to have as high a T as possible, because critical current and critical field fall off rapidly as Tc is approached.

The difficulty for future HTS applications arises from the present fabrication techniques, which make wires and tapes in a polycrystalline form. Critical transport currents in such structures have so far been low, because of randomly oriented crystal alignment and poor conduction at crystal grain boundaries. In addition to a poor geometric match inherent in polycrystalline structures, there appears to be an intrinsic insulator at the boundary of each individual crystal.

While up to 200 T has been estimated for the upper critical field in HTS materials, the ability to achieve fields over 10-20 T will undoubtably be limited in practice by the strain induced from stresses connected with containment of the high magnetic fields. For pulsed fields, the constraint is even more severe. It is customary to have maximum conductor tensile strains of several tenths of a percent for NbTi, a very ductile material, and it is unlikely that such strains can be supported by the brittle ceramic HTS, even with the coils designed to always remain in compression. Thus, unless a breakthrough occurs in the mechanical proper- ties of the HTS, significant quantities of high-modulus support structure, closely coupled to the HTS, will be needed to limit the strain on any magnet coils.

conventional superconductors have already been investi- gated [23]. HTS has several advantages over low- temperature superconductors (LTS) in switching appli- cations. The volume V of superconductor needed for a switch is given by [23]

v = , ( 4 )

where Pma is the peak power to be transferred, pN is the normaf state electrical resistivity, and Jc is the critical current density. The normal state resistivity of the ceramic HTS is much higher than its LTS counter- parts. Assuming equal Jc, the volume required for a HTS switch should be small. A second advantage is that the auxiliary cooling requirements for HTS is much smaller than for LTS.

In the switch the HTS would be combined with a structural material of low conductivity. It would be connected to a nonsuperconducting shunt of high conduc- tivity, such as a pair of rails or a drive coil. When the switch is in the closed position, the HTS is in the superconducting state. When the switch is to be opened, an external magnetic field, laser-driven heat pulse, or current pulse is imposed to drive the HTS normal and rapidly transfer the current to the shunt circuit. A preferred geometry for such a switch is a thin film [23]. Thin films may be the first fabricated forms for the new HTS material. The availability of such switches may allow direct coupling of EML devices to superconducting inductors.

Transmission Line Storage

If superconductivity enables a low-loss system, then it may be practical to construct the coaxial coils (or rail accelerator inductors) in such a way that the system forms a transmission line, with all the energy stored for launch stored in it. The energy would travel down the transmission line at a speed synchronized with a projectile launch and be reflected at each end. The synchrotron oscillation principle could be used to keep the projectile in phase with the traveling wave. The disadvantage of such a method is that the drive coils would be regularly pulsed, adding to the fatigue wear on them. In addition, the scaling of the launcher may make such a system unfeasible. Nevertheless, the potential for minimal-switching, traveling wave accelerator concepts need to be explored further [24].

CHALLENGES

Clearly, any significant technology development that minimized the complexity in the design requirements of the ET0 EML system or components, maximized the launcher system efficiency, or in some other way, reduced the large up-front capital investment, is an "enabling" technology. As we have seen in the previous section, superconductivity, and especially HTS, has the potential to significantly reduce the capital costs.

Malozemoff et al. [25] have described some of the expected properties of the developing HTS, and the general impact of these properties on some potential applications. While the field is developing very fast, and the final properties that may become available cannot be known at this time with any degree of certainty, it may be valuable to report some of their observations [25]. Current densities on preferentially oriented thin films at 77 K have already demonstrated more than lo5 A/cy2, and single crystal currents as large as 3x106 A/cm

As is well known, the state of superconductivity exists inside an envelope with temperature, current density, and magnetic field strength as parameters. As one approaches the critical value for any of these parameters, the available values for the other two go

have been seen at 4 K.

CONCLUDING REMARKS

This paper has examined the incorporation of superconducting technology in the design of Earth-to- orbit electromagnetic launch systems. Both rail accelerator and coaxial magnetic accelerator devices were considered, using the reference concepts developed under previous NASA studies as the basis for compari- son. The use of both conventional and advanced, high- temperature supeconductivity can potentially improve EML technology at the subsystem and component levels. Some of the major technical benefits in the use of superconductivity may be realized in the following areas: EML components (rails, coils, power lines, etc. 1; energy storage; energy switching; supercon- ductive magnetic augmentation (rail accelerator); projectile and drive coils (coaxial magnetic accelerator); and transmission line storage (coaxial design).

One primary application of superconductivity would be to improve the energy storage and distribution system required for a large-scale EML. Reductions in the size and number of energy storage devices are possible with applied superconductivity. The use of superconductors for many of the other EML components would result in lower system resistive losses and may even enable new design concepts for energy distribution and switching. All of these technical improvements have the potential to reduce system complexity and lower payload launch costs.

While the field of high-temperature superconduc- tivity is developing very fast, the final properties that may become available cannot be known at this time. One critical issue for future high-temperature superconductivity applications is that the materials strength properties of the new ceramics is very low. Significant improvement in the strength properties of bulk HTS materials is required before HTS can be considered a viable option for EML technology areas.

Acknowledgements

Part of this work was sponsored by the U.S. Dept. of Energy under Contract W-31-109-Eng-38.

248

REFERENCES

1. E. E. Rice, L. A. Miller, and R. W. Earhart, Preliminary Feasibility Assessment for Earth-to- Space Electromagnetic (Rail Gun) Launchers. NASA CR-167886 (1982).

2. L. A. Miller, E. E. Rice, R. W. Earhart, and R. J. Conlon, Preliminary Analysis of Space Mission Applications for Electromagnetic Launchers NASA-CR- 174748 (Aug. 30, 1984).

3. L. M. Zana, W. R. Kerslake, and J. L. Sturman, Rail Accelerators for Space Transportation, NASA TP-2571 (May 1986).

4. L. M. Zana and W. R. Kerslake, Rail Accelerator Technology and Applications, 1985 JANNAF Propulsion Meeting, NASA TM-86947 (1985).,l

5. H. Kolm, An Electromagnetic Slingshot" for Space Propulsion, Technology Review, pp. 61-66 (June 1977).

6. G. K. O'Neil and B. O'Leary (eds.), Space-Based Manufacturing from Nonterrestrial Matedials, Prog. Astron. Aeron., Vol. 57 (1977).

7. J. Billingham, W. Gilbreath, and B. O'Leary (eds.), Space Resources and Space Settlements, , NASA-SP-428 (1979).

8. T. A. Heppenheimer, Achromatic Trajectories and Lunar Material Transport for Space Colodization, Jr Spacecraft, Vol. 15, 176-183 (1978).

9. S. C. Rashleigh and R. A. Marshall, Electromagnetic Acceleration of Macroparticles to High Velocities, J. Appl. Phys., Vol. 49, 2540-2542 (1978).

10. J. G. Moldenhauer and G. E. Hauze, Experimental Demonstration of an N-Turn EML, IEEE Trans. Magn., Vol. MAG-20, 283-286 (1984).

11. H. A. Peterson, N. Mohan, W. C. Young, and R. W. Boom, Superconductive Inductor-Converter Units for Pulsed Power Loads, in Energy Storage, Compression,

. 309-317, Plenum Press, New York (eds. W. H. Bostick, V. Nardi, and

(1976).

12. D. R. Sadedin, Efficiency Equations of the Railgun, IEEE Trans. Magn., Vol. MAG-20, 268-271 (1984).

13. C. G. Homan, C. E. Cummings, and C. M. Fowler, Superconducting Augmented Rail Gun (SARG), Trans. Mag., Vol. MAG-22, 1527-1531 (1986).

14. C. G. Homan and . W. Schottz, Evaluation of Superconducting Augmentation on a Rail Gun System, IEEE Trans. Mag., Vol. MAG-20, 366-369 (1984).

15. P. Mongeau, Analysis of Helical Brush Commutation, IEEE Trans. Magn., Vol. MAG-20, 231-234 (1984).

16. K. McKinney and P. Mongeau, Multiple Stage Pulsed Induction Acceleration, IEEE Trans. Magn., Vol. MAG-20, 239-242 (1984).

17. S. L. Wipf, Concepts and Limitations of Macropar- ticle Accelerators Using Travelling Magnetic Waves, IEEE Trans. Mag., Vol. MAG-18, 121-126 (1982).

18. H. Kolm and P. Mongeau, An Alternative Launching Medium, IEEE Spectrum, pp. 30-36 (April 1982).

19. W. F. Weldon, Pulsed Power Packs a Punch, Spectrum, pp. 59-66 (March 1985).

20. D. Y. Chen, Power Semiconductors: Fast, Tough, and Compact, IEEE Spectrum, pp. 30-35 (Sept. 1987).

21. W. McMurray, Feasibility of Gate-Turnoff Thyristors in a High-Voltage Direct-Current Transmission System, EPRI EL-5332 (1987).

22. W. McMurray, Series Connection of Gate-Turnoff Thyristors, EPRI EL-5331 (1987).

23. 8. L. Laquer, Superconductivity, Energy Storage and Switching, in Energy Storage, Compression, and Switchin (eds. W. H. Bostick, V. Nardi, and 0. e ' p p . 279-305, Plenum Press, New York (1976).

24. M. D. Driga, W. F. Weldon, and H. H. Woodson, Electromagnetic Induction Launchers, IEEE Trans. s, Vol. MAG-22, 1453-1458 (1986).

25. A. P. Malozemoff, W. J. Gallagher, R. E. Schwall, Applications of High Temperature Superconductivity, ACS Symp. Ser. (Sept. 1987).