(Note: Figures which are in the thesis document are missing from this webpage)

Magnetic levitation for high speed ground transportation (Maglev) has been an active area of research and development for over 30 years. This new transportation mode has been proposed as an alternative to air, automobile, and high-speed train travel. Various experimental and a large number of theoretical studies have been done. Work has been continuous in Japan and Germany, while U.S. research support has been sporadic. The development in the past 10 years of high-temperature superconductors may alter the economics of high-speed ground transportation for the better. The goal of this thesis has been to design and test a novel prototype electrodynamic magnetic suspension using high-temperature superconductors and an iron core.

High-speed Maglev systems have been designed and tested with cruising speeds of up to 500 km/hour (~300 miles/hour). By comparison, commercial high speed rails such as the German I.C.E., the French T.G.V. and the Japanese Shinkansen which use standard wheel-on-rail technology reach approximately 200 miles/hour. Furthermore, Maglev systems offer higher acceleration rates and a smaller curve radius, which provides the potential for significantly higher point-to-point speeds than high-speed rail.

Maglev systems may be classified into two types, depending on the details of the magnetic suspension. The attractive type Maglev (called the electromagnetic suspension, or EMS), uses electromagnets on the train which are attracted to an iron rail. The vehicle magnets wrap around the iron guideway, and the attractive force in the +z direction lifts the train. The repulsive system (electrodynamic suspension, or EDS) is levitated by reaction forces caused by induced circulating currents in a passive conductive guideway. In either case, the levitating magnets are mounted to a number of “bogies” which in most Maglev systems are connected to the main train body by a secondary suspension consisting of springs and dampers. The secondary suspension maintains acceptable ride quality in response to external disturbances such as guideway imperfections and aerodynamic forces.

The EMS system was first commercially implemented for the Birmingham International Airport shuttle with two vehicles on line on a 900 meter run at 48 km/hour. The Birmingham system was operated from 1985-1995 with high reliability.

Research on high-speed long distance EMS systems has been done primarily in Germany by the Konsortium Magnetbahn Transrapid. The Transrapid TR07 is a train that has on-board iron-core copper coils with a pole pitch of approximately 0.25 meters. The vehicle wraps around a T-shaped guideway of steel and concrete beams constructed to a very tight tolerance. Propulsion is provided by a long-stator linear synchronous motor, and primary braking is accomplished by regeneration through the linear synchronous motor. Emergency braking is provided by eddy current braking and high-friction skids. Power limitations in the copper levitation coils limit the height of the air gap as doubling the air gap requires doubling the magnet current, quadrupling the magnet power. Therefore, this system operates with a small air gap (~ 1 centimeter), and is unstable if operated with constant current in the levitation coils. Therefore the system requires an active control system with redundancy to meet safety standards.

A 40 kilometer test loop is in operation in Emsland, Germany [219, pp. 2]. In 1988, the Transrapid TR06 test vehicle was operated at 257 miles per hour with 8-10 minute transits of the Emsland loop. Intensive testing on the TR07 prototype revenue service vehicle has resulted in the accumulation of more than 100,000 kilometers of tests by 1990. Revenue-producing operation is scheduled to begin in 2004 from Hamburg to Berlin.

The electrodynamic superconducting magnetic suspension (EDS) is a repulsive suspension invented in the United States in the 1960’s and has been actively researched in the United States, Japan, and to a lesser extent Germany. Superconducting coils on bogies mounted to a moving train induce circulating currents in a conducting guideway, with resultant levitation and guidance forces. The currents in the guideway also result in a drag force which in addition to aerodynamic drag must be overcome by the train propulsion system. At low speeds, the drag force increases with increasing train speed until the “drag peak” velocity is reached. A unique feature of the EDS system is that at speeds higher than the drag peak, the drag force decreases. Aerodynamic forces, on the other hand, increase as velocity squared.

The EDS suspension is stable without an active control system. At normal cruising speed, the levitation and guidance forces are weakly dependent on train velocity. The dynamics of motion of the train in multiple degrees-of-freedom is underdamped, and must be controlled to achieve acceptable ride quality. However, there is no levitation force at zero velocity, so rubber wheels or some other low-speed suspension is required.

A discrete-coil EDS Maglev vehicle was described as part of the National Maglev Initiative by a team at M.I.T. [178]. The train shown has two superconducting magnet bogies attached to each side of the train. A propulsive force Fp propels the train at velocity v in the +y direction. This force is provided by a linear synchronous motor. The high field (~ 1 Tesla at the guideway conductors) generated by the superconducting coils in the bogie induces currents in the guideway conductors. The high field also allows operation of the full-scale EDS Maglev train at a significantly higher airgap than the EMS system (~ 10-25 cm). The interaction of the field and the induced current produces a shear levitating force in the z direction. At the “lift-off” speed, the levitation force is sufficient to lift the weight of the train and the passengers. There is a sideways guidance force in the x direction acting on the bogies which tends to center the vehicle on the track. The circulating currents in the guideway coils also produce a drag force Fd in the -y direction which is overcome by the propulsion force.

The forces acting on the guideway coils are transmitted to the guideway superstructure. The guideway must be sufficiently stiff to handle the dynamic mechanical load without significant flexure. Mechanical design of guideway structures is discussed in the thesis by Phelan [118], and mitigating strategies for reducing eddy current stresses in guideways is discussed in the paper by Zahn [120].

Research effort in the U.S. focused initially on the EDS system using conductive aluminum sheets, suggested by Guderjahn, Wipf, Coffey and others [154, 158, 159]. Other geometries have been developed since the mid-1970’s, with significant work being done on discrete coil and ladder guideway designs. The null-flux suspension, first proposed in 1966 by Powell and Danby [168 - 171] at the Brookhaven National Laboratory, has coils arranged in a differential geometry with a high magnetic field gradient. This suspension has high stiffness and low magnetic drag, and the advantage that there are hold-down forces.

The first scale-model demonstration of EDS Maglev was the Magneplane of Kolm, Thornton, Iwasa, Brown and others at M.I.T [183, 188, 189]. A 1/25 scale model was demonstrated with superconducting magnets and a linear synchronous motor. The motor was used for active heave damping.

The first full-scale system using superconducting magnets was the Japanese National Railways’ ML-100, a 3.5 ton vehicle propelled tested on a 500 meter track by a linear induction motor [51]. This full-scale null-flux EDS system has evolved from the Powell and Danby concept and has been further developed and tested at the Miyazaki test site. The Japanese Maglev design uses niobium-titanium low temperature superconductors operating at 700 kA-turns and cooled by liquid helium at 4.2 K. The superconducting coils have a pole pitch of 2.1 meters. The ML500 test vehicle, based on an inverted-T guideway, was tested at 517 km/hour in December 1979. Since then, construction has begun on a new test line at Yamanashi. The new test track will include some gradient sections and other sections with tunnels [51, pp. 16]. The MLU-series test vehicles are based on a U-shaped guideway with separate levitation and guidance loops, operating with an air gap of approximately 20 centimeters. The levitation force is provided by sidewall levitation coils in a shear force arrangement. The target maximum commercial speed of the Maglev vehicle will be approximately 500 km/hour, almost twice the maximum speed of the Shinkansen high speed train.

The German and Japanese Maglev development programs, and a variety of results from other test fixtures and theoretical studies have proven the technical viability of Maglev as reliable high speed transportation mode. The prospects for German EMS Maglev look good, with significant government and private research support and a timetable in place for revenue service. The future of Japanese EDS Maglev is less clear, but significant resources are being focused on the development of an air core, low temperature superconducting Maglev system.

The focus of this thesis project has been the design and test of a new magnetic suspension suitable for EDS Maglev using high temperature superconductors and an iron core. The suspension is based on a “flux-canceling” geometry and a new low-cost guideway design. We take advantage of the electro-thermal stability of high-temperature superconductors by implementing an active secondary suspension control system with dynamic control of the current in the superconducting coils. Suspension models have been developed which capture the details of the electrodynamic interactions without being overly complicated.

1.1. Issues Related to EDS Maglev Using High Temperature Superconductors

The discovery in 1987 of the first high-temperature ceramic superconductor (HTSC) with a transition temperature above 77K [29] was promising for Maglev applications, for several reasons. The economics of using low-temperature superconductors is affected by the cost and weight of the liquid helium cryostat. Size scaling also come into play; in general, the power rating of a magnet is proportional to its volume and the heat leak is proportional to surface area, so a larger refrigerator is more efficient. For electric motors, it was shown that low-temperature superconductors were cost-effective only for larger machines [3].

The critical field is the maximum externally applied field for which the superconductor will conduct without loss. This parameter varies with operating temperature and current density inside the superconductor. High-temperature superconductors offer very attractive electrical properties when operated at low temperatures; several materials have a very high critical field at 4.2K (Table 1-1). Several materials have been found which have critical fields higher than 100 T, which is much higher than conventional low-Tc superconductors. For this reason, high-Tc materials have been used as inserts in high field magnets, operating at 4.2K, where their properties are very advantageous for magnet designs. One such NMR magnet design, using high-Tc superconductors, operated in a background field higher than 17 Tesla at 4.2K [7].

Table 1-1. Properties of some commonly-used superconductors at 4.2K

showing transition temperature (Tc), critical field at zero current and self-field (Bc),

and critical current density at zero field (Jc).

Values for Jc are approximate, since this value depends strongly on sample preparation

Material Tc (K) Bc (T) Jc (A/cm2 )

Nb-Ti 9.5 13 106 @ zero field

Nb3Sn 18 38 107 @ zero field

YBa2Cu3O7 (YBCO) 95 > 50

Bi2Sr2Ca2Cu3O10 (BSSCO) 110 > 50 2´105 @ zero field

High temperature superconductors may extend the range of cost-effective solutions to smaller machines and magnet designs with a relatively simple cooler. A further benefit is the cost of refrigeration. Separate analyses by American Superconductor Corporation [19] and Iwasa [8] (Table 1-2) shows an energy cost comparison, in terms of refrigeration energy needed per Watt of power dissipation, of cooling a refrigerator to various temperatures. These numbers are for a closed system, with boil-off not allowed, and show that much less energy is needed for refrigeration at 77K than at 4.2K. The values are approximate, and change with refrigerator size.

Table 1-2. Approximate refrigeration power required per Watt of dissipation in a cryostat

at various refrigerator operating temperatures

4.2K 20K 50K 77K

Refrigeration energy needed, Watts/Watt 1400 140 25 10

A prototype high-temperature superconducting magnet may be operated inside a simple insulated vessel. The low cost of the liquid nitrogen allows boil-off; re-liquefaction is not required. Furthermore, the latent heat of evaporation for liquid nitrogen (hL » 161 J/cm3) is much higher than that of liquid helium (hL » 2.6 J/cm3) so the volume rate of evaporation of liquid nitrogen per Watt of dissipation will be much less. Liquid nitrogen is relatively cheap (approximately $0.10/liter in quantity) compared to liquid helium (approximately $10/liter). Giese [270] approximates Maglev magnet coolant boil-off costs of $50,000 annually for a liquid helium cooled magnet, and significantly less for a liquid nitrogen cooled magnet, based on two cryostat designs with similar geometries.

Material properties also favor operation of a superconducting magnet at a higher temperature in terms of stability against magnet quenching. The specific heat of metals approaches a constant at temperatures above the Debye temperature (approximately 315K for copper, and 395K for aluminum) [203]. Below the Debye temperature and at cryogenic temperatures, the variation of heat capacity with temperature is as:

[Eq. 1-1]

where g and A are material constants. At very low temperatures and for most materials the T3 term dominates. so an increase in operating temperature can result in a drastic increase in heat capacity. At 4.2K, the heat capacity of Nb-Ti is Cv » 5.76´103 J/m3 K, while the heat capacity of a silver-clad high-temperature superconductor at 77K is approximately 106 J/m3 K [8]. Therefore, it takes much more energy to raise the temperature of the high-Tc superconductor than a low-Tc superconductor, for a given temperature change, and an HTSC coil is less prone to quenching due to magnetic and mechanical disturbances.

The difficulties associated with using high-Tc material for magnet designs are also numerous. The high-Tc materials discovered so far are oxide ceramics; the materials are brittle. Nb-Ti, on the other hand, is ductile and may be handled similarly to copper magnet wire using standard coil winding processes. Due to the chemical properties of the high-Tc ceramic material, a costly process to sheath the ceramic in a silver matrix is necessary. The resultant high-Tc superconductor is not a wire (as in the case of Nb-Ti) but a flat superconducting tape. The tape geometry requires subtleties in the details of the magnet winding, such as the use of the pancake winding.

The resulting high-temperature superconducting tape exhibits anisotropic electrical characteristics. At a given temperature and operating current, the tape can withstand a much higher field parallel to the tape flat surface than perpendicular to the same surface, before it becomes normal [23]. Therefore, conventional winding geometries designed for copper wire and low-Tc magnets may not work in a high temperature superconducting magnet, since their performance is not dependent on the direction of the field impinging on the conductor.

Also, the critical current density and critical field in the high-Tc material are degraded significantly when operating at 77K as compared to lower temperatures. Data for Intermagnetics General Corporation’s silver-sheathed Bi-2223 superconductor is given in Table 1-3 [6]. Performance at 77K with a 1 Tesla field perpendicular to the flat surface of the tape results in very poor performance, even compared to copper wire.

Table 1-3. Properties of Bi-2223 high-Tc superconductor, operating at 77K, 1 Tesla applied field

Jc (A/cm2 ) Test condition

30,000 Zero field

5,000 1T field parallel to tape

300 1T field perp. to tape

The high-Tc materials are also fragile; a tensile strain of 0.2% or greater is sufficient to degrade the superconductor’s properties [21, 26]. As compared to Nb-Ti, (where bending strains of 1% are easily handled), much more care must be taken in the winding of the magnet coils with high-Tc material. High-Tc magnet designs have been cited with wire bending radii of a few centimeters [1, 2, 5, 6, 7, 9, 10, 11, 18, 122, 123, 124, 125].

The science of large-scale high-Tc superconducting magnet design is still in its infancy. However, it is possible to make some intelligent approximations, if not to define the exact properties, of a high-Tc magnet for Maglev. For a practical full-scale EDS Maglev magnet design, the following specifications for superconducting wire performance must be met [3, 126]:

· Critical current density greater than 105 A/cm2. Special care must be taken in this specification as the tape is anisotropic with regard to fields parallel and perpendicular to the wide tape surface.

· Critical magnetic field of greater than 1 T at 77K.

· Long wire lengths (> 100 m) so that windings need not be formed in multiple sections.

· High strength to withstand Lorentz forces and forces due to thermal expansion.

· Robustness with regard to AC losses, wire uniformity, and quenching.

· Wire must be ductile, and able to withstand significant strain due to bending during the coil winding process.

While critical temperature and critical field in a superconductor are dependent on the chemical structure of the superconductor, the critical current density achievable is primarily dependent on details of the material processing. Steady progress has been made to increase the critical current density of high-Tc materials. Experimental magnets using high-Tc superconductors have been built, although most have been run at temperatures significantly lower than 77K (Table 1-4) .

Table 1-4. Some prototype high-temperature superconducting magnet designs

[1, 2, 7, 10, 11, 122-126]

Group Operating temperature Test Condition

University of Oxford 4.2K to 64K > 0.4 T@ 4.2K

American Superconductor 27K > 2 T

Oak Ridge 4.2K 1.14 T

Mitsubishi 20K High-Tc coil for Maglev, air core

Grumman 20K iron core magnet for Maglev

Intermagnetics General 4.2K >0.2 T in 17T background field

University of Wollongong, Australia 77K Bi-2223, wind and react coils, resulting in 15,000-20000 A/cm2, self field

1.1.1. AC Losses in HTSC

One significant advantage is that high-temperature superconductors have been proven to be much more robust with regard to losses due to AC coil currents and external AC fields than their low-temperature counterparts. HTSC coils have a higher heat capacity than LTSC coils operated at 4.2K, reducing the likelihood of thermal runaway which results in a magnet quench. Various workers have published data which shows that HTSC coils require a higher dissipated energy density by factors of 104~106 (Table 1-5). A minute amount of energy is required to quench a LTSC coil, on the order of microJoules [19]. Therefore, LTSC coils subjected to AC background fields and/or mechanical strains must have magnetic shielding and significant mechanical structural integrity to limit dissipation in the windings. Also, magnet driving currents must be limited to DC currents, as AC currents cause loss due to eddy currents, hysteresis effects, and filamentary coupling.

The Japan National Railways has found that a fundamental limitation to maximum Maglev train speed is vibration-induced mechanical losses inside the LTSC magnet winding [200]. The losses were especially pronounced when the coils were excited at structural resonant frequencies in the 100-300 Hz range. Mechanical losses required an expensive re-design of the superconducting levitating coils and support structure.

Wilson [295, pp. 70] offers a further example of mechanically-induced losses in superconducting wire operating in a field of 5T and with a current density of 3x104 A/cm2. A movement of 10 mm will result in 1.5 J/cm3 work being done on the superconductor. If only a small fraction of this energy is dissipated as heat due to friction, quenching will result.

Assuming that sufficient cooling is provided, it is expected that a HTSC coil will not quench in a Maglev environment. During operation, “hot spots” may sporadically develop due to external disturbances and disappear. These normal regions inside a high-temperature superconducting coil will grow and shrink slowly and the magnet is unlikely to quench. As a result of this slow normal region growth it is difficult to detect a normal region in a HTSC coil. Further research is needed to determine if HTSC coils operated at intermediate temperatures (20-40K) will result in significant performance improvements. However, the fundamental physics shows that HTSC coils will be more useful in AC applications that LTSC coils.

Table 1-5. Approximate energy density required for magnet quenching

Comparison of low-Tc and high-Tc materials [8, 19]

Material Quench Energy Test Condition

Nb-Ti 0.01 J/cm3 4.2K

BSSCO 15 J/cm3 60K, 1T

1.1.2. Applications of HTSC to Transportation Research

Maglev trains are subjected to mechanical disturbances under normal operating conditions in the 0 - 20 Hz range [148], due to imperfections in the guideway. EDS magnetic suspensions, and especially the flux-canceling topology, have little intrinsic magnetic damping for vertical motion. Therefore, some form of secondary suspension is required to meet necessary ride quality standards. The secondary suspension provides damping which minimizes the effects of external mechanical disturbances.

Economic considerations demand that there be a reasonable tolerance on the alignment of adjacent Maglev guideway sections. For example, misalignments of 25 meter guideway spans with a train operating at 200 meters/second will result in mechanical disturbances with a fundamental frequency of 8 Hz. The disturbance spectrum will change as the vehicle speed changes, and the EDS primary resonant frequency may be excited at select train speeds. Adequate ride quality is achieved with the use of feedback control which maintains stable suspension while keeping the vertical acceleration in the passenger cabin to acceptable levels.

For low temperature superconducting Maglev designs, most proposed designs have a mechanical secondary suspension provides the necessary damping. A mechanical linkage connects the main levitating magnets to the train passenger compartment. The suspension may have shock-absorbers or air-springs. This mechanical suspension adds weight, complexity, and cost to the overall system design.

Our suspension is based on a secondary magnetic suspension concept using high-temperature superconductors, to take advantage of the robustness of HTSC material with regard to AC currents and AC magnetic fields. The necessary damping is provided by a control system which controls the superconducting magnet currents; in order for the control system to maintain the magnet position, transient currents are put through the superconducting coils. This sort of active control is impossible with low temperature superconductors, due to their AC loss characteristics.

With an active secondary magnetic suspension, the necessary ride quality may be achieved without the mechanical secondary suspension, with associated decrease in train weight and complexity. The use of HTSC coils in an active secondary suspension is a significant contribution to the state-of-the-art in Maglev suspensions.

1.2. Test Fixture System Overview

The design and test of a fixture for the evaluation of a new EDS Maglev system is the focus of this thesis. The fixture has been built, and the resultant Maglev test facility has been used to predict behavior of a full-scale flux-canceling Maglev system, based on a high-temperature superconducting coil. One side of a Maglev octapole has been built and operated with copper coils.

The guideway is composed of multiple conductive copper coils. When the train is in the vertical null position at z = zo and traveling in the +y direction, there is no net flux through the levitating coils, and no net current induced around the loop. However, if the train’s vertical position deviates from equilibrium, the net changing flux through the loop induces currents in the loop. The Lorentz force is a restoring force in this structure, with the magnetic suspension acting as a linear spring with spring constant kz. The suspension behaves like a mass and a linear spring, with a resulting resonant frequency .

Eddy currents and circulating currents create a drag force acting on the train in the -y direction. There is also a sideways force (+x) which pushes the train away from the guideway. For an actual Maglev train, there will be another octapole on the other side of the train, and the sideways force will tend to center the train on the track.

The rotating test wheel is composed of an aluminum hub, with a non-conductive fiberglass/epoxy composite rim. The wheel has a diameter of 1.7 meters, and is designed to operate at a maximum speed of 1000 R.P.M., which corresponds to a linear peripheral speed of approximately 100 meters/sec. The speed of the wheel is regulated by a 10 horsepower motor and speed controller. The magnets are mounted such that the two rows of magnet polefaces line up with the wheel outer radius.

Guideway coils are embedded in the epoxy rim and every effort has been made to reduce the amount of conductive and magnetic material in the vicinity of the magnet pole faces. Furthermore, there is no ferromagnetic material in the vicinity of the guideway conductors. This will allow extrapolation of forces for the full-scale design. The relative motion between the stationary superconducting magnet and the moving guideway coils create the magnetic forces.

The rotating test wheel has been used to test the viability of using high temperature superconductors for Maglev and linear motors. We compare lift, drag, and guidance forces with calculations derived from simple models. The fixture has been operated in two different modes:

1. For measurement of magnetic forces and torques, and aerodynamic forces, the superconducting magnet assembly is mounted to a 6-axis force sensor. This sensor measures forces and moments in 6 degrees of freedom, with forces 0-200 Newtons and torques 0-100 Newton-meters. The force sensor restricts the translational and rotational motion of the magnet assembly. Forces may be measured for various x, y, and z magnet positions relative to the guideway conductors.

2. For study of vertical dynamics, the magnet is mounted to a one degree-of-freedom air bearing, which allows low friction vertical movement of the magnet. Using this bearing, the vertical motion with and without an active secondary suspension has been studied.

For future tests, it is possible to modify the test wheel and magnets for alternative tests on guideway and/or linear motor topologies.

1.3. Goals of this Thesis

Specific goals of this thesis have been:

· Design, build and test a working model of an iron-core Maglev electrodynamic suspension suitable for use with high temperature superconductors. In the design, low cost construction techniques and materials have been used.

· Design a new low-cost, high-performance loop guideway. Emphasis has been placed on a guideway design which will give an acceptable lift-to-drag ratio for the full-scale design, and which is easy to manufacture.

· Develop hybrid electrodynamic/circuit models for prediction of lift and drag forces. There has been limited use of detailed electrodynamic analysis and finite element analyses; rather, approximate techniques are used which lend themselves as practical design tools.

· Prepare a methodology to collect data for future Maglev electrodynamic stability analyses.

1.4. Organization of Thesis

This thesis covers the design and analysis of a fixture for the testing of candidate Maglev magnet and coil geometries. The Maglev test fixture is of sufficient interest and design complexity that the detail of the design has been presented. Chapter 2 is a literature review, which covers much of the history of Maglev research, including significant works on the development of Maglev circuit and electrodynamic models. Other peripheral Maglev topics, such as guideway design and superconducting coil design are also referenced.

Chapter 3 covers the mechanical design of the test fixture, with details given on the design of the test wheel, cryostat, air bearing, and nitrogen delivery system. Chapter 4 covers the design and construction of the electrical system, including detail of the iron core, copper coil, and superconducting coil designs.

Chapter 5 develops a model for calculation of forces for the magnet and guideway conductor geometries. The approach taken is to produce simple circuit models which may be used as a design aid for a full-scale system. Therefore, the focus has been to utilize available linear circuit-analysis tools without becoming inundated in the detail of the mathematics of the electrodynamic interactions..

Chapter 6 presents the test results and analysis of design models. In this section, measurements of lift force, drag force, and drag peak are compared to predictions based on the circuit models, with good results. Further tests were performed to generate lift force at zero velocity by AC excitation of coil currents.

Chapter 7 concludes the thesis, and summarizes test results and formulates recommendations for future study. Chapter 8 contains appendices covering peripheral topics such as circuit schematics, data on the strength and properties of materials used in the construction of the thesis project, and Matlab programs. Chapter 9 concludes with an extensive indexed bibliography, covering many aspects of the history of Maglev research and details of the thesis design.

Contact Information:

Marc T. Thompson, Ph.D.
Thompson Consulting, Inc.

9 Jacob Gates Road 

Harvard, MA  01451
Phone: (978) 456-7722
Business website: 

Home Page  |  Curriculum Vitae  |  Consulting Projects  |  Maglev Research at MIT   |   Teaching at WPI   |  Links