ELECTROMAGNETIC GUNS, LAUNCHERS and REACTION ENGINES*
Massachusetts Institute of Technology
Francis Bitter National Magnet Laboratory**
Cambridge, Massachusetts, 02139
Henry Kolm, Kevin Fine, Fred Williams and Peter Mongeau
Recent advances in energy storage, switching and magnet technolgy make electromagnetic acceleration a viable alternative to chemical propulsion for certain tasks, and a means to perform other tasks not previously feasible. Launchers of interest include the dc railgun driven by energy stored inertially in a homopolar generator and transferred through a switching inductor, and the opposite extreme, the synchronous mass driver energized by a high voltage alternator through an oscillating coil-capacitor circuit. A number of hybrid variants between these two extremes are also promising. A novel system described here is the momentum transformer which transfers momentum from a massive chemically driven armature to a much lighter, higher velocity projectile by magnetic flux compression. Potential applications include the acceleration of gram-size particles for hypervelocity research and for use as reaction engines in space transport; high velocity artillery; stretcher-size tactical supply and medical evacuation vehicles; the launching of space cargo or nuclear waste in one-ton packets using off-peak electric power.
Magnetic guns and launchers have received periodic attention for many years, and several large systems have actually been built. The fact that none of these evolved into a practical device reflects largely the immaturity of required support technology and lack of coordinated follow-up programs. The most recent survey of the field was made by the Naval Weapons Laboratory in 1972, and the report contains all significant prior references.
Since 1972 considerable attention has been devoted to linear electric motors in the context of air cushion and magnetically levitated high speed trains; an extensive review published in 1975 contains over 140 references. Most early efforts utilized linear induction motors (LIMs) which do not lend themselves to high acceleration. There evolved one concept, however, the linear synchronous motor (LSM) first proposed by Powell and Danby and ultimately implemented by Kolm and Thornton at MIT; it is synthetically synchronized and is capable of very high acceleration, efficiency and speed. G.K.O'Neill of Princeton University proposed using the LSM for launching lunar raw materials into very precise orbits to permit interception at a space manufacturing site, thus re-inventing a concept first proposed by Arthur C. Clarke in 1950. O'Neill and Kolm developed the "mass driver" as part of two NASA-AMES summer studies in 1976 and 1977, and a group of students constructed the first demonstration model at MIT. It was exhibited at the 1977 Princeton Symposium on Space Manufacturing and also on the occasion of the first flight of the orbiter Enterpise in August 77. A second, more sophisticated mass driver is presently under construction at Princeton and MIT, with support from NASA-Lewis.
Another significant effort was made recently by Marshall and Barer who used the world's largest homopolar generator at the Australian National University in Canberra to power a series of exerimental dc railguns. Their spectacular success might not have been of much practical interest, had it not been accompanied by equally spectacular progress in the design of practical pulse-rated homopolar generators by Woodson, Weidon and others at the University of Texas in Austin. The group also invented a new inertial energy storage device, the "compensated alternator", or "compulsator". There has also been a great deal of other work in the area of energy storage in relation to requirements for ohmic heating of plasmas in toroidal fusion experiments, laser-induced fusion, particle beam weapons research and laser weapons research. Much of this work is directly applicable to accelerators. Equally applicable is work done in the development of large, high-intensity magnet coils, superconducting as well as normal, for MHD power generation and for solid state research. The MIT National Magnet Laboratory is a center of expertise in this area. Related work which is doubly applicable is the development of large superconducting magnet systems for inductive energy storage at Los Alomos and Sandia.
In March 1977 Dr. Harry Fair, head of the Propulsion Technology Branch of the Army Research and Development Command in Dover, N.J., inquired whether any of the MIT Magneplane or Mass Driver work might have ordnance applications. It was immediately obvious that the potential applications and related comcepts and technologies spanned such a vast range as to require a nationally coordinated effort. Peter Kemmey and Ted Gora of ARRADCOM were assigned to the task of coordinating the effort within DOD, and the present authors were funded to conduct a preliminary study. In addition, we have assembled an inter-agency steering committee and a technical advisory panel to ensure liaison with other centers of expertise.
We are concerned here with linear motors which are capable of very high acceleration. This excludes at the outset the sizeable literature of linear motors developed over the years for a variety of purposes, including traverse curtain rods, conveyor belts, solid waste separation, liquid metal pumps, high speed ground transportation, and even certain attempted launch devices. We shall characterize the features and limitations of our basic arsenal of accelerator concepts.
The classic railgun is the simplest and also the most high perfected accelerator. It consists of two parallel rails connected to a source of dc current, the projectile consisting of a short-circuit slide propelled between the rails by the Lorentz force F = BLI/2 newton, where B is the magnetic field instensity between the rails in tesla, L is the length of the current path through the slide, or the gap between rails in meters, and I is the current in amperes. The factor of 1/2 accounts for the fact that the field is B behind the slide and zero in front of it, the average being B/2.
The classic railgun has been studied extensively by Brast and Sawle of the MB Associates in the mid-sixties under NASA contract, and more recently by Marshall and Barber using the world's largest homopolar generator at the Australian National University in Canberra; it is capable of storing 500 MJ.
Railguns can operate in two distinct modes. In the metallic conduction mode, current flows through the sliding projectile itself, and this mode has been demonstrated to a performance level of about 1 kg mass and 2,000 g (20,000 m/s2) acceleration by the switching gun used in the Canberra installation to feed the main gun. Marshall and Barber tound that if the railgun is driven very hard, a plasma arc tends to bypass the projectile, leaving it behind. By using a non-conducting lexan projectile and confining the arc behind it they were able to achieve a performance level of 16 gram accelerated at 250,000 g along a 5 m barrel to a final velocity of 5.9 km/s. As railguns are extrapolated to large projectile sizes, the distinction brush conduction mode and plasma mode is likely to vanish: brush conduction will be supplemented by arc conduction as the limit of brush current is exceeded.
The practical limit of railgun performance in regard to projectile size, acceleration, length and velocity will have to be explored by progressive refinement of material and engineering details, as in the case of any new technology. The Canberra work has provided sufficient information to justify the first attempt in this direction. Westinghouse, with support from DARPA, will construct a practical railgun system including the first pulse-rated homopolar generator designed with attention to overall weight. The objective is to demonstrate feasibility of accelerating a 0.33 kg (.73 pound) projectile to a velocity of 3 km/s (9.8 ft/s), corresponding to a muzzle energy of 1.5 MJ.
To a great extent, the practical limit of rail guns will depend on acceptable cost and service life. The problems relate to mechanical containment of the percussive expansion force which tends to blow the rails apart, the electromagnetic analog of barrel pressure in a chemical gun, with the important difference that the railgun maintains more or less constant pressure throughout the acceleration. Instead of chemical corrosion, there is the destructive effect of high brush current density and the related metal vapor arc. The body of knowledge available from the study of brushes and circuit breakers does not extend to the current densities and velocities in question.
In addition to these limits, the classic railgun also faces certain fundamental lmimits which are not related to acceleration, but to maximum possible length or maximum muzzle velocity. As a railgun is lengthened, the resistance and inductance of the rails eventually absorb a dominant fraction of the energy. The effect is seen to begin at about five meters in the Canberra tests. Increasing velocity also causes an increasing back-emf. Current will continue to flow, even if this emf exceeds the output voltage of the homopolar generator, because the intermediate storage inductor acts as a current source. However, there is a practical limit to the voltage which can be stood off by the gap between rails, and this scales about linearly with size. Thus there are two fundamental effects which limit the amount of energy that can be transferred to the projectile, regardless of how much is available.
Another shortcoming of the railgun is its inherent inefficiency. An appreciable amount of energy is contained in the rail inductance at the instant the projectile leaves, and this energy must be absorbed by a muzzle blast suppressor. A fraction migh conceivable be returned to the homopolar generator. There are several means for circumventing the limitations of the classic railugn.
The magnetic field between the rails can be augmented by supplementary current which does not flow through the sliding brushes. This current can be carried by separate conductors flanking the rails (which must be farther from the projectile), or it can be added to the rail current itself by simply terminating the rails with a load resistor or inductor at the muzzle to carry a fraction of the current. The rails themselves will obviously contribute more field than auxiliary rails located farther away, but the use of superconducting auxiliary rails might be expedient in some applications. It should be noted that railgun fields are much higher than the critical fields of superconductors. Augmentation has the obvious effect of reducing the amount of current flowing through the brushes and the projectile, and thereby the necessary conductor mass which must be accelerated.
It should also be noted that the augmenting field is twice as effective as the rail field itself. The augmenting field prevails in front of the projectile as well as behind it, thereby elimintating the factor of 1/2 in the Lorentz force expression. This fact is important inasmuch as it reduces to one half the rail bursting force which must be contained for a given acceleration.
Augmentation therefore ameliorates both the brush current density limitation and the bursting force containment limitation of classic railguns.
The length limitation imposed by rail resistance and rail inductance can be circumvented by simply subdividing a long railgun into short segments, each fed by an independent local energy source. This will of course involve certain commutation problems as the projectile transitions between segments, but will permit using part of the energy stored in each segment to energize the subsequent segment. The segmented railgun seems promising for launching large masses such as aircraft at low acceleration. In very long launchers, the use of multiple independent energy supplies will have other advantages as well.
As mentioned in the introduction, the mass driver is a direct adaptation of the linear synchronous motor first conceived and developed as the MIT Magneplane system in 1970-75, a high-speed magnetically levitated train. The mass driver can be planar or axial depending on requirements. The axial configuration permits higher efficiency and is therefore preferred for high acceleration, while the planar configuration will accommodate payloads which need not be cylindrical and may have any arbitrary shape.
In both cases, the payload is carried by a reuseable vehicle, called the bucket, which is provided with two superconducting coils carrying a persistent current and guided without contact by repulsive eddy currents induced by the bucket motion in an aluminum guideway. The bucket is propelled by a series of drive coils which are pulsed in synchronism as the bucket passes by. The bucket operates like a surfboard riding the forward crest of a magnetic travelling wave, the wave being generated by the drive coils and synchronized by position sensors. Buckets can be launched at repetition rates of 10 per second. Each bucket releases its payload at a precise speed, is decelerated, and then returns to the starting point on a return track to be reloaded and relaunched.
Mass drivers can operate in the "push-only" mode as in the case of Mass Driver One, or in the pull-push mode of Mass Driver Two, now under construction, in which each drive coil undergoes a complete sinusoidal oscillation by being connected synchronously to a supply capacitor line. By tuning this cycle to the effective wavelength of the bucket it is possibe to achieve energy transfer efficiencies, electric-to-mechanical, of better than 90 percent. We should add that the bucket-to-payload ratio is about unity, and that about half the bucket energy is recoverable by regenerative braking.
For all practical purposes, mass drivers have no velocity limit and no length limit. Acceleration has been limited thus far by the current and voltage capacity of the SCRs used for switching. Using shelf components, Mass Driver Two should achieve 500 to 1,000 g. If the SCR limitation is removed, by using ignitrons, spark gaps, or direct contact switching, performance will be limited by mechanical and thermal failure of the drive coils. Some preliminary calculations based on a four inch caliber mass driver using aluminum bucket coils and copper drive coils suggest an acceleration limit between 100,000 and 250,000 g. This is comparable to railgun performance. however, the failure mode of drive coils under fast pulse conditions is a very complex subject requiring experimental study.
All previous mass driver designs are based on a bucket coil current density of 25 kA/cm2 of cable, achieved in an operational model of the MIT magneplane. Superconductors should withstand up to four times this current density at the low field intensity and stored energy involved. It should also be pointed out that mass drivers do not necessarily require superconducting bucket coils. For periods of the order of 0.1 second it is actually possible to maintain higher current densities in normal conductors. Maximum performance mass drivers are therefore likely to utilize aluminum bucket coils, possibly precooled to liquid nitrogen temperature, fed by sliding brushes, and drive coils triggered by physical contact. Of course this would eliminate the non-contact advantages.
A unique feature of mass drivers bears emphasis: although they are energized by capacitors, the costliest, heaviest and bulkiest energy store known, each capacitor is used hundreds or thousands of times during each launch cycle by being connected to many drive coils through feeder lines. This permits the use of an efficient but slower intermediate energy store, such as a compulsator or MHD generator.
The railgun is in essence a single-turn motor. A multi-turn railgun would reduce the rail current and the brush current by a factor equal to the number of turns. It therefore seems worth-while to study a "helical railgun". In this hybrid device, the two rails are surrounded by a simple helical barrel, and the projectile or re-useable carrier is also helical. The projectile is energized continuously by two brushes sliding along the rails, and two or more additional brushes on the projectile serve to energize and commute several windings of the helical barrel direction in front of and/or behind the projectile. The helical railgun is in fact a cross between the railgun and the mass driver.
Accelerators based on mechanical energy storage have not been used since the day of the bow and medieval catapult, with the exception of naval aircraft launching. Mechanical energy storage devices are bulky, heavy, and slow to release their energy. The advent of practical superconducting magnets provides a good mechanical storage mechanism, the magnetic slingshot.
Consider a short superconducting solenoid which is free to slide inside a long one. The travelling solenoid will be either attracted to or repelled from the center of the long solenoid, depending on the direction of relative magnetization. Either configuration can serve as an electromagnetic slingshot.
In the attractive configuration, the travelling solenoid can serve as a payload-carrying shuttle bucket. Released at the breach end of the barrel coil, it will accelerate to the center, where it will release its payload at maximum velocity, come to rest at the muzzle, and then return empty to a position short of its release point, from where it can be returned to the release point by mechanical force, possiby by a thermal cycle. This oscillation is inherently loss-less, except for possible eddy currents induced in nearby metal.
In the repulsive configuration, the travelling solenoid will be moved by mechanical force from the breach to a point just beyond the center of the barrel. When released, it will be expelled from the muzzle as part of the projectile. Veclocities up to several hundred m/s are attainable by slingshots.
By successively quenching a line of adjacent coaxial superconducting coils forming a gun barrel, it is possible to generate a wave of magnetic field gradient travelling at any desired speed. A travelling superconducting coil can be made to ride this wave like a surfboard. The device in fact represents a mass driver or linear synchronous motor in which the propulsion energy is stored directly in the drive coils.
A brass washer placed on top of a vertically oriented pulsed field coil is driven upward, accelerated by eddy currents which tend to be 180 degrees out of phase with the inducing field pulse. The resulting impulse has been used commercially since 1962 for metal forming operations, for instance by swaging terminal fittings around aircraft control cables. The process has certain applications for acceleration. It can be made into a synchronous induction motor whose performance is limited by the thermal inertia of the sliding member.
A novel concept described here for the first time is what we shall call the "momentum transformer". It makes use of a so-called "flux concentrator", first studied by Howland at MIT Lincoln Laboratory in 1960. A flux concentrator is simply a conducting cylinder with a funnelled bore, and at least one radial slot extending from the inside to the outside surface. The cylinder is surrounded by a pulsed field winding, preferably imbedded in a helical groove to minimize hoop stresses. A fast pulsed current in the winding induces an opposite image current in the outer surface of the cylinder. Due to the radial slot, this induced current is forced to return along the inner perimeter of the cylinder, thereby generating a magnetic field in the funnelled bore. All of the magnetic flux which would have filled the pulsed field winding in the absence of the conectrator is thus compressed into the central bore, resulting in a field intensity which is higher than it would have been by about the outside-to-inside cross section ratio.
The device was used at MIT for high field research and also for industrial metal forming. In 1965, Chapman used a flux concentrator with a tapered bore for accelerating milligram metal spheres to hypervelocities. Using a first stage explosive flux compressor, Chapman managed to reach peak fields in excess of 7 megagauss, starting with an initial field of only 40 kilogauss.
The momentum transformer proposed here uses a flux concentrator as the armature or sabot in a chemically driven conventional gun. The bore of this sabot is occupied by a much smaller projectile, for instance a rod-shaped armor penetrator. The muzzle end of the gun is a pulsed field winding imbedded in a helical groove, which is excited with a current pulse sufficiently slow to penetrate the barrel and fill the bore with magnetic flux. When the sabot enters this flux region so rapidly that the effective penetration depth of the field is small, it compresses the flux into its inner bore, decelerates drastically, and expels the projectile contained in its bore at a much higher velocity. The device should have very little recoil because the muzzle coil acts like a muzzle brake, transferring much of the sabot momentum to the barrel. The process can be multi-staged with a series of nesting sabots.
The acceleration of milligram to gram size pellets to hypervelocities, i.e., 10 to 100 km/s, already has a literature of three decades. Research areas include micrometeorite impact studies, equation-of-state research, terminal ballistics, etc. A new application of current interest involves the achievement of fusion by pellet impact at several hundred km/s.
Projectiles in the range of ten grams to a kilogram accelerated to 3 to 10 or 20 km/s have foreseeable applications. The destruction of missiles in space, where mass is at a premium is one obvious use. Another is the possible interception of incomgin rounds by ships and armored vehicles. This requires small projectiles travelling at speeds much greater than the incoming round, capable of detonating, deforming, or just deflecting them. Plasma-driven railguns already have the required capability on a laboratory basis. If incoming round interception can be accomplished with good reliability, it will make armored vehicles as obsolete as knights on horseback.
An armor penetrator fired at 3 km/s, twice present speed, needs only to be about one fifth the size to inflict equal damage. If in addition it can be propelled with available diesel fuel, tanks can be given five times present capability with drastically reduced vulnerability. We are dealing here with energy pulses in the 1 to 3 MJ range, supplied by the primary propulsion engine of the tank.
It is an irony of modern tactical warfare that an armored advance can be supported with many tons per minute of artillery, but not by a single gallon of fuel or pound of food. Helicopters and parachutes are too vulnerable for battlefield use, and the chemical gun does not lend itself to logistic supply applications. Electromagnetic launchers can fill this need.
A 300 pound stretcher or supply module can be launched from a 100-foot, truck-mounted ramp to 100 mph at 3.3 g acceleration, using only 0.14 MJ of energy. It could easily be guided to a soft landing by microwave or conventional ILS type guidance system located at the destination point. The vehicle would operate at high speed, low trajectory, be relatively invulnerable and weather-independent, and significantly less expensive and fuel-consumptive than a helicopter. It could be built using available technology.
It is interesting to study the generation of STOL aircraft which could be designed by eliminating the requirement of inordinate take-off thrust from on-board engines.
The application of mass drivers for lunar launching and for use as reaction engines in orbital transfer has already been studied extensively. However, the possibility of electromagnetic earth-based launching, proposed by science fiction writers since the forties, has never before been considered seriously. On the basis of computer software developed by NASA in connection with the Venus lander, it appears quite practical.
A telephone-pole shaped vehicle 8 inches in diameter and 20 feet in length, weighing 1.5 tonnes, accelerated to 20 km/s at sea level would traverse the 8 km atmosphere in half a second, emerging at 16 km/s, which is enough velocity to escape the solar system. It would lose 3 to 6 percent of its mass by ablation of a carbon shield. Initial projectile energy would be 300 x 10^9 joule, one third of which would be lost in traversing the atmosphere.
The launch energy may seem formidable, but it amounts to only 83 MW-hrs, which represents several minutes of output by a large metropolitan utility plant. The required launcher would be 20 km long at 1,000 g acceleration; it would be only 2 km long, less than a small airport runway, at 10,000 g, which should be easily attainable. Such a launcher could be installed on a hillside, or in a vertical hole made by an oversize rotary well drilling rig.
One potential application is the disposal of nuclear waste. 2,000 tons of waste will be generated between 1980 and 2000. This waste could be launched out of the solar system by using off-peak power from a utility plant at a cost corresponding to only 2 cents per kw-hr of generated power which produced the waste. Considering that the average cost of power during the period will be 22 cents per kw-hr, this waste disposal cost is very low.
Rotary motors have not yet approached the conceptual or practical limits of their potential, even after a century of intensive evolution. Fundamental innovation still occurs under the stimulation of new technology and new needs.
Linear motors have not been pursued to anywhere near a comparable degree, although an appreciable literature exists. Linear motors might be on the threshhold of an evolution comparable to the evolution of rotary motors. The above survey indicates that there is no shortage of new concepts or uses. What makes this field exciting is the advent of new pulsed energy sources, and the challenging fact that a motor of zero curvature is virtually free of all fundamental limitations on size, acceleration and velocity.
* Study supported by U.S.Army Armament Research and Development Command, Dover NJ, under ARO Grant No. JAAG 19-78-G-1047
** Laboratory supported by the National Science Foundation.
Last update May 7, 2007 by Barry Hansen ©1998-2007