Linear Electric Motors for Aerospace Launch Assist

Matthew T. Caprio, Sid Pratap, William A. Walls, Raymond C. Zowarka

Center for Electromechanics, The University of Texas at Austin
Mail Code R7000, Austin, Texas 78712 USA
(512) 471-4496 / Fax (512) 471-0781 [email protected]
www.utexas.edu/research/cem

Keywords
High speed motor, launch assist, linear motor design, linear propulsion, power system simulation

Abstract
This paper summarizes the results of a design study investigating the use of linear electric motors
in an aerospace electromagnetic launcher application requiring 7g’s of horizontal acceleration to
a velocity of 300m/s. The study initially reviews the current state of high-speed electric
machines in applications similar to the one proposed. Induction and synchronous linear motors
are then evaluated for suitability to the application by analyzing their characteristics in this high-
speed operating regime. A detailed design approach is used to synthesize machines capable of
meeting the launcher’s performance requirements, since the conditions fall outside of the
conventional design envelope. Realistic physical features of all electrical power system
components are included to ensure the evaluated systems are physically realizable. The two
motor types are compared for suitability to the application based on performance, cost, and
system integration issues. Finally, a prototype demonstrator design is proposed to verify the
results of the study.

1 Background and Motivation

This study into the use of induction and synchronous linear motors for an aerospace launch assist
application was performed in support of NASA’s Reusable Launch Vehicle initiatives. NASA has a
goal of reducing the cost of launching cargo into space by one or two orders of magnitude within
25 years. One means of achieving this goal is to develop a more cost efficient method of
accelerating the craft than the current two-stage rocket engine system.

NASA is pursuing an electric powered launch assist scheme that accelerates the craft horizontally
along the ground (using a power source external to the vehicle), thereby eliminating one stage of
rockets. This approach of replacing the first rocket stage with an electric launcher track could
reduce the per-launch cost and improve safety and reliability.

This study for NASA focused on investigating the high-speed use of conventional, continuous
linear motors (induction and synchronous) to accelerate launch craft in numerous scales. NASA
intends to begin with a 45 kg demonstration, and increment the scale of prototypes to ultimately
apply this technology to a full size launch craft. Masses of 45, 455, 4550, 45500, and 455,000
kg were considered in this work.

2 High Speed Linear Motor Review

As a first step in evaluating the use of linear motors in a launch assist scheme, a literature review
was undertaken to investigate whether electric machines are currently in operation in related
applications with similar requirements.

The major uses of continuous linear motors found in the literature review were ground
transportation (high speed rail and amusement rides), machining operations, materials transport /
handling, and other launcher systems under development [1]-[11]. The highest speeds
encountered in these results were MagLev trains tested at 153 m/s [12]. This is significantly lower
than the NASA requirement of 300 m/s.
Very high speeds have been achieved with pulsed linear motors used as accelerators and rail guns.
Velocities well above 5 km/s have been demonstrated [13]. However, these machines are not well
suited to launch assist due to implementation difficulties in long lengths. Nevertheless this fact
serves as a data point for velocities achievable with electric machines.

Figure 2.1 summarizes the demonstrated speed ranges of electric machines in numerous high-
speed applications. From this chart it is apparent that there are no known high-power
applications demanding 300 m/s operation from a linear electric motor.

Figure 2.1 Speed range of linear electric motors

Given the apparent “gap” in applications in the area of 300 m/s, no currently existing machinery
can be identified as suitable and proven for the launch assist application under study. With this is
mind, it is very likely that one of the existing classes of machines will need to have its operating
range extended to accommodate the launch assist conditions. At this point it is unclear whether it
will be better to advance continuous motors (induction and synchronous) beyond their existing
ranges to higher speeds, or to refine pulsed machines for lower speed, lower acceleration,
continuous operation over long distances.

3 Launcher Mechanical Requirements

As an initial estimate of the scale and capabilities of linear motors that would be necessary to
meet the requirements of the study, preliminary calculations were performed. The NASA
specifications require constant acceleration of the craft from rest to 300 m/s at 7g’s. These
specifications were expanded into size and parameter estimates by applying them to a mechanical
system and computing the approximate levels of force, power, consumed energy, and other
factors.

Table 3.1 Mechanical requirements of launchers

Launch Parameter units 45kg 455kg 4550kg 45500kg 455000kg
Frontal area (input ) m2 0.25 1.16 5.38 25 116
2
Constant acceleration (input ) m/s 7 7 7 7 7
Launch velocity (input ) m/s 300 300 300 300 300
Time to launch s 4.62 4.62 4.62 4.62 4.62
Launch distance m 656 656 656 656 656
Constant inertial force kN 3.09 31.2 312 3,125 31,245
Drag force at launch kN 7.40 34.3 159 740 3,433
Total force at launch kN 10.49 65.6 472 3,864 34,678
Delivered inertial energy MJ 2.03 20.5 205 2,048 20,478
Drag loss energy MJ 1.71 7.9 37 171 792
Total expended energy MJ 3.73 28.4 242 2,219 21,270
Peak inertial power at launch MW 0.93 9.4 93.7 937.4 9,374
Peak drag power at launch MW 2.22 10.3 47.8 222 1,030
Total peak power at launch MW 3.15 19.7 142 1,159 10,404

A kinematic model was created to compute the dynamic mechanical loads including inertial forces
and aerodynamic drag (substantial in this application) encountered at all speeds throughout the
launch. With this model, a speed profile is imposed and the non-linear drag forces are computed
over the entire speed range, and the resulting energy loss and power requirements quantified. The
results of these mechanical load estimations for all craft sizes can be seen in Table 3.1. In
reviewing the results for the 45 kg craft, it is apparent that the high drag force has a substantial
impact on the peak required power as the drag loads exceed the 7g’s acceleration load.

4 Detailed Analysis of Linear Motors

Induction and synchronous linear motor technologies are at mature levels, however in this case,
each type would need to be pushed beyond the existing design envelope to meet the application
requirements [14]-[16]. Only through the process of synthesizing and simulating each motor for
these conditions can it become clear if the necessary materials and operating characteristics can
be realized.

To this end, a design study is performed for both induction and synchronous linear motors in this
operating regime. Both types are analyzed in detail with their parameters iterated and optimized,
and a preliminary design of each is simulated to evaluate the operation under these high force,
high frequency, high current conditions.

4.1 Linear Induction Motor Design

The general configuration assumed for the launcher induction motor concept is a long, double-
sided stator that sandwiches a moving secondary conductor sheet. The stator is fixed to the
ground and takes on the form of a track. The moving secondary is a continuous sheet of
conductor material that attaches to a carrier mechanism to drive the craft. A schematic of this
configuration can be seen in Figure 4.1.

moving sheet secondary,

short length V

z
y x

fixed double stator,

very long length

Figure 4.1 Linear induction motor configuration

Because of the long length of motor required for this application, only a small fraction of the
stator is being used at one time. For this reason, the stator of this induction machine is assumed as
divided into separate segments. This “block switching” allows short portions of the stator to be
energized two-at-a-time in succession, to avoid resistance losses in the long unused portions. The
length of an individual portion, and total number of portions is a parameter that is determined
through trade studies.
To solve for baseline values defining the general size of the induction motor, an electromagnetic
code was written. This “continuum model” is a closed-form field solution of the electromechanics
of the linear induction motor that evaluates the critical electrical and physical parameters of a
design.
The inputs to the code specify the induction machine geometry and material properties. The
routine then applies inverse-dynamics to solve for the operating conditions necessary to
complete the prescribed acceleration of the mechanical load from 0 to 300 m/s including the
aerodynamic drag forces. The outputs are the critical electrical and thermal parameters of the
system that determine whether the design can be physically realized. The geometry parameters of
the linear induction motor are shown in Figure 4.2.

STATOR SINGLE PHASE WINDING POLE LOCATIONS

8
y
µr =

s = winding pole pitch

a = winding spacing
a s b = conductor height
µo l l = air gap length
µo = permeability of air
µr = infinite permeability
b (of iron)
x

ARMATURE SHEET INDUCED POLE LOCATIONS

8

y
µr =

s = induced pole pitch

a s a = induced winding
µo spacing
l t t = sheet conductor
thickness
l = air gap length
µo = permeability of air
µr = infinite permeability
x (of iron)

THREE PHASE STATOR AND ARMATURE

y

a = phase "A" winding

a' = phase "A" opposite
direction winding
b = phase "B" winding
b' = phase "B" opposite
direction winding
b' a c' b a' c b' a c' b a' c
c = phase "C" winding
c' = phase "C" opposite
x direction winding

Figure 4.2 Induction motor geometry parameters

The operation of the closed form field solution model proceeds as follows. The algorithm steps
through increments of velocity to represent the imposed constant acceleration of the payload
and secondary. At each speed increment, the vector control law determines the necessary applied
stator frequency to establish the required force, without exceeding the limit of maximum flux
density. In each of these computations, the effects of the harmonic current interactions are
considered, including power loss and drag forces resulting from negative sequence harmonics. The
computed values of fundamental current density, slip, power loss, secondary temperature rise and
total force are recorded for post analysis. The values include mechanical, electrical, and thermal
parameters as follows:

v the velocity profile of the secondary and payload

Ft the total force applied by the secondary to overcome inertia and losses
Fd the internal drag forces of the machine throughout acceleration
Fnet the net force exerted on the mass (controlled to be constant for constant
acceleration)
Akz harmonic components of stator current density
s slip: the relative speed between the traveling flux wave and the secondary sheet
P loss the power loss due to internal drag forces
dT the temperature rise of the secondary due to power loss heating
ni the stator ampere-turns required to generate the magnetic field
By the stator flux density (controlled to be constant)
The continuum model was employed to choose optimized parameters of an induction motor
suitable for driving the 455 kg mass craft. Table 4.1 summarizes the final values selected through
the design study trades. Table 4.2 indicates the performance parameters of this design.
 Table 4.1 LIM baseline geometry Table 4.2 LIM field solution baseline
configuration performance
Stator Quantities
Pole pitch 0.3 m Velocity 300 m/s
Active height 0.3 m Constant acceleration level 7 g's
B max 1.25 T Maximum generated force 73222 N
Electrical frequency (max) 500 Hz Total energy consumed 34.0 MJ
Secondary Maximum power level 21.8 MW
Length 3.0 m Electrical power efficiency 96.6%
Active height 0.3 m Max line current density 221 kA/m
End turn overhang 0.15 m (per side) Slip (relative speed) 0.35 m/s
Material Aluminum Max ampere-turns 28.2 kA-t
Mass 78 kg Secondary temp. rise 56.3°C
Total air gap 36 mm

Numerous stator winding schemes and patterns could be used to generate the stator flux
distribution imposed in the field solution—it is winding independent. However, to quantify how
the LIM interacts with its power supply, stator winding parameters were selected and the resulting
impedances calculated. The winding configuration was chosen through optimization; results are
shown in Table 4.3. A full electrical power system and mechanical load dynamic simulation was
performed using these parameters.
Table 4.3 LIM electrical parameters
Stator Winding Configuration
Number of turns per coil 3
Coils per segment 10
Segment uncovered length 10 poles
Stator conductor cs area 0.0012m2
EQ Circuit Parameters
Rs 0.0034 Ohm
Ls 41 uH
Lm 78 uH
Rr 5.4e-4 Ohm
Lr 1.38 uH

4.2 Linear Synchronous Motor Design

The general configuration assumed for the launcher synchronous motor concept is essentially
identical to the induction machine, with the exception of the armature. The moving secondary
armature of the synchronous machine must contain persistent magnetic field poles. These poles
could be established by many different devices, but for the purposes of this study, a particular type
is not chosen. The poles are simply imposed mathematically. In actual implementation,
conventional DC windings, superconductors, or permanent magnets could be employed.

The stator of the synchronous motor concept is a long, double-sided stator that sandwiches the
moving secondary armature. The stator is fixed to the ground and takes on the form of a track.
The moving secondary establishes a permanent magnetic field and attaches to a carrier
mechanism to drive the craft. A schematic of this configuration can be seen in Figure 4.3.

Wherever possible, the approach used to configure and design the synchronous motor is
essentially the same as for the induction motor, except where differences must naturally exist.
This allows the baseline machines to be as similar as possible, so they are easily compared.
Additionally, the simulation architecture and codes were recycled to a large degree.

As the secondary current is controlled rather than induced in the synchronous machine, the
selection of its geometry parameters is less complex than for the induction motor. A field
solution code is not required, instead an equivalent persistent current can be explicitly solved to
generate the required force for a given stator flux distribution.
 moving permanent pole
armature,short length V

S
N
z
y x

fixed double stator,

very long length

Figure 4.3 Linear synchronous motor configuration

The synchronous motor baseline design was developed to meet the force and high-speed
requirements of this application for the 455 kg craft case. The parameters of the baseline
synchronous machine geometry are listed in Table 4.4. The baseline stator winding configuration
for the synchronous motor and its inductance and resistance parameters are shown in the Table
4.5.

Table 4.4 LSM baseline geometry Table 4.5 LSM computed electrical
configuration parameters
Stator Stator Winding Configuration
Pole pitch 0.3 m Number of turns per coil 3
Active height 0.3 m Number of coils per segment 10
B max 1.25 T Segment uncovered length 10 poles
Electrical frequency (max) 500 Hz (max) Stator conductor cs area 0.0012 m2
Armature Stator Electrical Parameters
Length 3m,10 pole pitch Phase self inductance 0.4788 mH
Active height 0.3 m Phase to phase mutual ind. -0.18819 mH
End turn overhang 0.15 m (per side) Phase resistance 7.8 m Ohm
Mass 78 kg
Total air gap 36 mm
Equivalent persistent current 57 kA

5 Launch Power System Dynamic Simulation

To evaluate linear motors for the launcher application, the combined electro-mechanical system
was dynamically simulated through a launch to verify its performance. For this, all components
of the electric power supply system had to be represented with sufficient realism and fidelity.
This need stems from the fact that the power system interacts dynamically with the motor and
the load, and that the physical limits of the power supply impose operational limits on the
motor, and therefore the overall system performance. The design of a motor is therefore
substantially dependent on the characteristics of its power supply, as well as its load. For this
reason, the major components of the launcher power system must be appropriately characterized.
For this study, no components were completely idealized.
 Rs Ls Lr

Lm Rr/slip

THREE PHASE D.C. RECTIFIER D.C. FILTER PWM INVERTER INDUCTION MOTOR PER PHASE CIRCUIT
ALTERNATOR (Energy Storage)
(Prime Power)

Figure 5.1 Simulated power system schematic

Figure 5.1 schematically represents the assumed major power system components of the 455 kg
craft launcher system for both motor types, shown supplying a single phase of a multiphase
induction motor. With the exception of the motor component, the power system schematic is
identical for supply of a synchronous motor.

As seen in the figure, the major components are the alternator, the DC rectifier, the DC filter /
link, and the PWM inverter. Not shown is the energy storage device, presumed in this schematic
to be a mechanical component (not an electrical one) supplying torque to the alternator.
Additionally, the impedance of the block switching and distribution system was included. These
devices are modeled with key physical characteristics for use in performance simulations.

The mechanical loads applied in the simulations include the inertial load of the craft, non-linear
aerodynamic drag, and propulsion loads required to provide magnetic levitation based on an
assumed passive electro-dynamic levitation (EDL) scheme.

The launch operation of both the induction and synchronous motor systems for the 455 kg craft
was dynamically simulated. Both systems demonstrated the ability to successfully complete the
7g’s acceleration launch to 300 m/s. Additionally, it was found that system-wide operating
conditions were within reasonable physical ranges. A summary of the results of the dynamic
simulations can be seen in Table 5.1.
Table 5.1 Dynamic launch simulation summary
Quantity Induction Synchron.
Electric energy input 37.2 MJ 34.35 MJ
Converted mech. energy 34.4 MJ 33.65 MJ
Delivered mech. energy 25.2 MJ 24.86 MJ
Converted energy efficiency 92.47% 97.96%
Delivered energy efficiency 67.74% 72.37%
Max. phase voltage 6.6 kV 6.6 kV
Max. phase current rms 11.3 kA rms 10.6 kA rms
Max. supply frequency 505 Hz 500 Hz
Max. electrical power 29.4 MW 29.25 MW
Max. force 73.2 kN 73.2 kN
Max. velocity 300 m/s 300 m/s

Another notable result of the simulations involved the electrical distribution system. The
impedance of a distribution bus to the block switched stator segments was included in the
simulations, and as a result, it was concluded that a single power source could be located near the
launch end of the track, rather than multiple distributed power sources as initially assumed. This is
due to the fact that the highest current draw occurs when the aerodynamic drag becomes
substantial, at the high-speed end of the track. Transmission losses are therefore manageable at
the beginning of the launch where the currents are lower, and the frequency is low. Figure 5.2
shows this possible distribution schematic.
 approx. 700m

MOTOR

DISTRIBUTION BUS
INVERTER

ENERGY STORAGE TO
& ALTERNATOR GRID

Figure 5.2 Single power supply distribution system

6 Performance Evaluation Comparison

From reviewing the simulation results, it is apparent that there is very little performance
difference between the induction and synchronous designs in this case. Both systems met the
launch performance requirements. Further, as their electrical operating characteristics are very
similar, the power supply system required for each is also virtually identical. This result seems to
indicate that in the linear launcher application, both types of motor can be satisfactorily designed
for the task, and neither exhibits obvious performance advantages.

In general, the evaluation indicates that the comparison and trade-offs of the two motors in this
application are similar to what would be expected from ordinary applications, under normal
operating conditions. However, one notable difference exists in controls. Due to the high-speed
operation of this machine, position error may become a dominant controls issue in properly
orienting the stator field in the synchronous motor to maintain synchronicity. This effect even
caused some difficulty during mere simulation of the machine. In contrast, a relatively large error
in position and velocity can be robustly tolerated with the induction motor.

Estimated cost of the induction and synchronous motor based launch systems was very similar.
The cost analysis figures for the two complete prototype launcher systems for the 455 kg craft
were approximately $27 million and $29 million USD respectively, a difference of less than 10%.

7 Rotary Demonstrator
For testing the first prototype launcher of the 45 kg craft scale, a rotary embodiment of the
linear motor is proposed. The test could demonstrate only the highest speed operation of the
linear motor, which is presumed to be the most challenging aspect of this machine. This test unit
can be designed to behave in a manner physically equivalent to a linear system, although being in
a convenient rotary configuration.

The rotary prototype is a convenient test configuration for this system because it is far simpler
to implement. It requires only a small portion of the linear motor to be constructed in order to
demonstrate the machine. Additionally, a conventional motor could be used to bring the system
to a high initial speed (at a low acceleration rate) so that only a short duty would be required of
the prototype. For a demonstration of 7g’s acceleration from 292 to 300 m/s, the motor and
power system equipment could be built for only transient duty to substantially lower the cost and
complexity.

8 Conclusions
A detailed design study of linear motors for aerospace launch assist led to baseline designs for
induction and synchronous motors that were dynamically simulated in a launch, with realistic
mechanical loads applied and realistic power supply characteristics. The performance of these
machines was analyzed and both designs met the performance specifications, all with reasonable
system operating parameters. As a next step, an experimental test of a prototype of each design
is recommended to verify the simulation results.
It is significant that both motors exhibited normally expected behavior, even outside of their
typical operating ranges. Neither design proved to have major performance advantages in this
application beyond the conventionally held trade-offs between induction and synchronous
motors. The selection of the preferable machine type will therefore be influenced by factors
outside the scope of this study such as reliability, maintenance concerns, control system design,
and integration with the MagLev scheme.

9 Acknowledgements
The authors thank NASA Kennedy Space Center for sponsoring and funding this research.

10 References
1. Cowan, M., “Ultimate velocities for induction launchers,” Sixth IEEE Pulsed Power Conference – Digest of
Technical Papers, Jun 29 – Jul 1 1987.
2. Eastham, J.F., et at, “Comparison of some propulsion methods for magnetically-levitated vehicles,”
International Conference on Maglev Transport Now and for the Future, London ,1984, p. 111.
3. Gilliland, R.G., et al, “A linear synchronous unipolar motor for integrated magetic propulsion and
suspension,” International Conference on Maglev & Linear Drives, 1986, Vancouver, B.C., p. 149.
4. Kozoriz, V.V., “New principles of maglev and traction underlying transportation,” SAE Technical Paper
Series, Aug 5-7 1991, p. 1-6.
5. Laithwaite, E.R., “Adapting a linear induction motor for the acceleration of large masses to high velocities,”
IEE Proceedings Electric Power Applications, v. 142 n 4, July 1995.
6. Murray, C.J., “Coil gun drives high-speed-rail technology,” Design News, v. 50, Sept 1995, Cahners
Publication Company.
7. Panaitescu, A., “Linear induction motor for high speed trains,” International Journal of Applied
Electromagnetics in Materials, v. 2 n 4, Apr 1992, p. 345-352.
8. Poloujadoff, M., The Theory of Linear Induction Machinery, Oxford University Press, New York, 1980.
9. Schaaf, J.C., Davey, K, Zowarka, R.C., Weldon, J.M., “AMT Maglev sled – EML weapons technology
transition to transportation,” IEEE Transactions on Magnetics, v. 33 n 1 pt 1, Jan 1997, p. 379-383.
10. Shokair, I.R., et al, “Performance of an induction coil launcher,” IEEE Transactions on Magnetics, v. 31
(Jan 1995), p 510-515.
11. Valenti, Michael, “Designing the ultimate thrill machine,” Mechanical Engineering, v. 117 (August 1995),
p. 70-8.
12. Okumura, F., “Superconducting linear motor car reaches a speed of 552 km/h in manned operation,”
Japanese Railway Engineering, v. 143, 1999, p. 28-29.
13. Weeks, D.A., Weldon, W.F., Zowarka, R.C., “Hypervelocity macroparticle accelerator experiments at
CEM-UT,” IEEE Transactions on Magnetics, v. 27, n 1 Jan 1991, p.85-90.
14. Nasar, Syed A., Boldea, I., Linear Electric Motors, Prentice Hall, New Jersey, 1987.
15. Gieras, J.F., Linear Induction Drives, Oxford University Press, New York, 1994.
16. Gieras, J.F., Linear Synchronous Motors, CRC Press, Florida, 1999.