AIAA SciTech Forum 10.2514/6.

2019-0786
7-11 January 2019, San Diego, California
AIAA Scitech 2019 Forum

Magnetic Brake testing for Hyperloop Pod Design

Abhishek Soni1, T. S. Indraneel1, Vignesh Jayakumar1, Dhaval Shiyani2, Pushkaraj Bhagwat1, and Dr. Shabaan
Abdallah3
University of Cincinnati, Cincinnati, Ohio, 45219, USA

Hyperloop is the fifth mode of transportation with the aim of drastically reducing the
time over medium distance travel. A Hyperloop pod was designed for the SpaceX
competition to travel at 110 m/s and in an environment of 1000 Pa absolute pressure. Due to
the high-speed nature of the travel, the system requires a sophisticated braking mechanism.
A non-contact magnetic braking mechanism was designed for this application with analysis
Downloaded by IOWA STATE UNIVERSITY on January 12, 2019 | http://arc.aiaa.org | DOI: 10.2514/6.2019-0786

performed on various types of magnetic configurations. The primary focus of this

investigation was to identify a magnetic configuration to maximize the braking force. The
braking forces generated by this optimal configuration were measured for further analysis.
The task to select the optimal orientation was performed in two parts, simulations were
carried out using ANSYS Maxwell and the forces generated by the braking arrays were
measured experimentally.

I. Introduction
Proposed in a white paper by Elon Musk and SpaceX in 2013 [1], the Hyperloop is a high-speed transportation
system that involves passenger pods travelling at subsonic speeds inside an evacuated tube. SpaceX, following the
release of the white paper, announced a student competition to spur the growth of the technology. Hyperloop UC
(HUC) was one of the 30 finalists for the competition out of the 1000 participants. As a part of the competition a
design and fabrication activity for a Hyperloop pod prototype was undertaken. This work focuses on the design and
testing of HUC pod’s eddy current based braking system.
Eddy current based braking systems have been widely used in the maglev industry. Magnetic field lines when
cut by a conducting non-ferromagnetic material produces eddy currents in the non-ferromagnetic material. The eddy
currents generated in turn produce magnetic fields that oppose the change in magnetic flux of the material, which
generates an opposing force to the relative motion. This opposing force can be used as a braking force [2]. Previous
research indicates that the braking force is a function of three main parameters - the relative velocity, the distance
between the magnet and the conductive non-ferromagnetic surface, and the configuration of the magnetic array.

Fig. 1 Braking force vs. velocity characteristics [3] (FL - Lift Force or Repulsion Force; FD - Drag Force)

__________________________________
1
Graduate Student, Department of Mechanical and Materials Engineering.
2
Graduate Student, Department of Aerospace Engineering.
3
Professor, Department of Aerospace Engineering.

Copyright © 2019 by the American Institute of Aeronautics and Astronautics, Inc. All rights reserved.
 The nature of the relationship between the braking forces and the relative velocity is shown in Fig. 1. The
opposing forces are also known to be inversely proportional to the cube of the distance between the two surfaces
[3][4]. Consequently, designing a braking system with good braking characteristic across the velocity range and
determining optimum location for engaged positions of the magnetic brakes was important and challenging.
Fig. 2 provides a schematic representation of the braking system. The reaction surface for the magnets was the
centering I-beam provided by SpaceX on the test track. In order to avoid large unbalanced forces on the I-beam,
permanent magnets are placed on either side. The actuation of the brake system between the engaged and
disengaged positions was implemented through use of linear actuators. Each brake pad on either side of the I-beam
comprises of a 5-magnet array with each magnet of size 1" × 1" × 1". The selection of the array size was based on
ease of handling, space and packaging constraints.
This work details the selection and testing of appropriate magnetic arrays for the brake system, beginning with
the simulations performed using ANSYS Maxwell to narrow down the types of magnetic configurations considered
for testing. The test setup and measurement system are described, and the results obtained for various magnetic
configurations is analyzed and discussed. The types of magnetic configurations considered are presented in Table 1
Downloaded by IOWA STATE UNIVERSITY on January 12, 2019 | http://arc.aiaa.org | DOI: 10.2514/6.2019-0786

below.

a) b)
Fig. 2 Pod braking mechanism in engaged position against the central I-beam on the track

II. Magnetic Arrays

The configuration of the magnetic array used affects the magnetic flux passing through the aluminum I-beam,
and the drag forces generated are directly proportional to the rate of change of flux [5]. Hence the selection of an
appropriate array is important to obtain good braking characteristics across the velocity range of interest. Magnetic
arrays can be arranged in multiple configurations, three of which are the focus of this investigation; unidirectional,
alternating, and Halbach. The unidirectional array (Fig. 3(a)) has all magnets placed with the north and the south
poles oriented in the same direction. Whereas, the alternating array (Fig. 3(b)) has each magnet placed to the next
with a 1800 rotation of the field. The Halbach array (Fig. 3(c)) has each magnet placed to the next one with a 90 0
rotation in the field direction. Since the flux patterns of the arrays are different, the force generating capability also
vary.
Fig. 4 highlights a plot of the braking force estimates from a simulation using ANSYS Maxwell. The plot
compares the expected braking forces at two different extremes of the expected velocity profile of the vehicle, 7 m/s
and 110 m/s. Comparison between the models indicate that the Halbach array can be expected to provide a greater
braking force than the unidirectional and alternating configurations. However, to further validate the choice of
orientation, the unidirectional and the Halbach array were chosen for the experimental quantification of the drag
forces.

III. Test Setup

A rotating aluminum circular disk, with approximately the same thickness as the I-beam was used to mimic
linear motion for testing magnetic brake arrays [6][7]. The test-setup and the magnetic arrays are shown in Fig. 5.
The disk is directly mounted onto the shaft of a 3-phase motor. The motor was operated using a variable frequency
drive (VFD) to allow for varying speeds, and the disk speed was captured using a laser tachometer.
A mounting structure was built using 80-20 beams for the load cells and magnetic arrays, with the arrays placed
against the rotating circular disk. A single 2000 lbf load cell was attached on an array in the drag direction and

2
 another on the second array in the repulsion direction to avoid cross coupling of the drag and repulsion forces. This
structure was fixed to the ground using a T-slot table. For more details on the setup, refer to Table 1.
Downloaded by IOWA STATE UNIVERSITY on January 12, 2019 | http://arc.aiaa.org | DOI: 10.2514/6.2019-0786

Fig. 3 a) (Top Left) Unidirectional array b) (Top Right) Alternating Array c) Halbach Array [8]

a) b)
Fig. 4 Comparison of Braking Forces as a function of magnetic orientation a) At a velocity of 7 m/s b) At a
velocity of 110 m/s for 0.25” between magnets and the disk

IV. Test Analysis Procedure

Before the start of each measurement, the drag load cell and the repulsion load cell were calibrated in the
assembled state, and the load cell amplifier balanced. The angular velocity of the disk was ramped up to 2000 RPM
and then the motor turned off. Data was captured from the time the disk reaches its maximum velocity until the disk
came to a halt. Angular speed of the disk was obtained from the tachometer data using an adaptive nth pulse
algorithm [9]. Linear velocity was approximated as 𝑅Ω, where 𝑅 is the distance to the center of the magnetic array
and Ω is the angular velocity of the rotating disk. Force values were plotted against this linear velocity.

3
 Magnet Array
34-inch diameter
Aluminum disk

Strain Gauge
Load Cell
80-20 beam setup
Downloaded by IOWA STATE UNIVERSITY on January 12, 2019 | http://arc.aiaa.org | DOI: 10.2514/6.2019-0786

Fig. 5 Test Setup

Table 1 Specifications of the setup

Component Specification Value

Aluminum Radius 17”

Thickness 0.5”

Motor Power 20 HP

Strain based load cell Capacity 2000 lbf

Nominal sensitivity 20 mV/lbf

Neodymium magnets Size 1” x 1” 1”

N52 Grade strength (residual induction) 14,800 Gauss

80-20 beams Cross section 1.5” x 3”

Amplifier Model number Vishay 220A

Data Acquisition System Model number NI 4431

V. Results
A. Single Sided Unidirectional and Halbach
The braking forces generated by a single array of unidirectional and Halbach configurations on one side of the
disk were compared to confirm the trends observed with the FE models (Fig. 4). As expected, the Halbach array is
observed to provide better braking forces as compared to the unidirectional array. The single sided setup produces
very low forces which are hard to capture using a 2000 lbf load cell. It can be noted that the force values for the uni-
directional array at higher speeds, lower end of the force spectrum, indicates negative values which is likely due to
the zero error of the load cells.

4
Downloaded by IOWA STATE UNIVERSITY on January 12, 2019 | http://arc.aiaa.org | DOI: 10.2514/6.2019-0786

Fig. 6 Drag estimates for single sided Halbach and Unidirectional configuration

B. Attracting vs Repelling
The double-sided magnetic arrays can be arranged in two different configurations - attracting and repelling (Fig.
7). The drag forces generated by attracting and repelling configurations were both measured using the test setup.
Both configurations have similar braking estimates in the higher velocity range. However, the attracting Halbach
array is seen to provide significant braking advantage in the lower velocity ranges and the overall braking
characteristics generated by the attracting configuration yielded shorter braking times from top speed to pod stop.

a) Repelling configuration b) Attracting configuration

Fig. 7 - Array configurations

Fig. 8 Drag estimates for double sided Halbach - Attracting Vs. Repelling

5
 C. Double Sided Attracting Halbach Array
To optimize the engaged and disengaged positions for the brake, the forces were measured for 8 distances
(between the disk and the magnets) from ½” and 5/4” with a ⅛” step size. The data obtained is shown in Figure 8.
Downloaded by IOWA STATE UNIVERSITY on January 12, 2019 | http://arc.aiaa.org | DOI: 10.2514/6.2019-0786

Fig. 9 Drag estimates for double sided Halbach 0.5” to 1.25’’ from the disk

The repulsion forces (FL from Fig. 1) were not the focus of this work. However, it was necessary to measure
these forces for linear actuator selection and the space frame design. As the amplifier was being balanced to zero
before each test, the static repulsion or attraction loads for the arrays had to be measured separately to offset the
dynamic measurements appropriately. Fig. 9. and Fig. 11. show the estimated drag force from each magnetic array
(𝐹𝐷 /2) and the corresponding repulsive loads (𝐹𝐿 ) that the individual actuators would experience for a case of
magnets placed at 0.5’’ from the rotating disk. Additionally, the drag forces highlighted in the previous plots
correspond to forces experienced by a single array of the pair. The braking force generated by a pair of magnetic
arrays would be double the numbers shown.
For configurations closer than 0.5 inches, the braking forces were too high for the 20 HP motor rotating the disk
to overcome. Hence the trials were restricted to 0.5-inch distance and the data from the testing was used to
extrapolate and predict the braking forces and distances at 0.25 inches from the I-beam.

Fig. 10 Forces generated by a double-sided magnetic array on a moving disk

6
Downloaded by IOWA STATE UNIVERSITY on January 12, 2019 | http://arc.aiaa.org | DOI: 10.2514/6.2019-0786

Fig. 11 Drag estimates for double sided Halbach - 0.5’’ from the disk

VI. Conclusion
Attracting Halbach configuration provides the best braking forces amongst the compared linear magnet array
configurations due to the increased magnetic flux lines between the two opposing poles in this configuration. The
test setup was found restrictive in the power of the motor used, and the ability to measure braking forces for magnet
arrays placed at less than 0.5’’ from the disk was limited as a result. It is possible to extrapolate the data at 0.5’’ to
predict the braking forces at 0.25’’ distance from the rotating disk. Alternately, a higher-powered motor can be used
to measure braking forces for closer distances. The overall dimensions of the magnetic array lead to some
approximations in linear velocity calculations from the rotating velocity. Use of a larger disk size would help reduce
these approximations in the linear velocity estimations. However, the use of a larger disk would need more
sophisticated manufacturing techniques.
Based on the measurement results and force estimates for smaller distances arrived based on theoretical
extrapolation, four pairs of linear braking arrays were designed. Each pair containing five neodymium magnets of
size 1" × 1" × 1", arranged in a Halbach configuration, with each pair of arrays in attraction. An electro-mechanical
actuator was used to translate each array of magnets from its resting non-braking position at 1.5” distance from the
I-beam to the fully actuated braking position of 0.25” away from the I-beam. Moreover, two pairs of friction brake
pads were installed to achieve dead stop due to the loss in braking force from the permanent magnets at very low
velocities.

Acknowledgments
The Hyperloop pod design venture was a massive team effort of over 50 student volunteers across multiple
colleges at the University of Cincinnati. The authors would like to acknowledge the entire Hyperloop UC team for
their efforts as a collective group in the design of the pod and accompanying test setups. Hyperloop UC was
fortunate to have sponsors who provided technical advice and manufacturing help particularly Justin Atkins
(Cincinnati Incorporated), Tri-State Fabricators, and Ansys Inc., their support is highly appreciated. Additionally,
we would like to thank SpaceX for organizing this competition. The authors would also like to acknowledge the
contributions of Ronald Hudepohl for his manufacturing guidance and Dr. Randall Allemang for his valuable
technical guidance and for allowing to use the Structural Dynamic Research Lab (SDRL) for testing and assembly
process.

7
 References

[1] E. Musk, “Hyperloop alpha,” PDF). SpaceX. Retrieved August, vol. 13, 2013.
[2] E. Lenz, “Ueber die Bestimmung der Richtung der durch elektrodynamische Vertheilung erregten galvanischen Ströme,”
Ann. Phys., Vol. 107, no. 31, pp. 483–494, 1834.
[3] T. D. Rossing and J. R. Hull, “Magnetic levitation,” Phys. Teach., vol. 29, no. 9, pp. 552–562, 1991.
[4] J. Hribar, “Magnetic braking,” Univ. Ljubljana, vol. 172, 2008.
[5] S. Sainjargal and J.-K. Byun, “Analysis and case study of permanent magnet arrays for eddy current brake systems with a
new performance index,” J. Magn., vol. 18, no. 3, pp. 276–282, 2013.
[6] J. WANG, Y. LI, and L. YAN, “Study on Applying the Linear Halbach Array to Eddy Current Brake System for Maglev
[J],” Electr. Drive, Vol. 5, p. 3, 2010
[7] M. Z. Baharom, M. Z. Nuawi, G. Priyandoko, and S. M. Harris, “Eddy current braking experiment using brake disc from
aluminium series of A16061 and A17075,” in IOP Conference Series: Materials Science and Engineering, 2012, vol. 36, no.
1, p. 12005.
[8] A. San-Millan, “Design of a teleoperated wall climbing robot for oil tank inspection,” in Control and Automation (MED),
Downloaded by IOWA STATE UNIVERSITY on January 12, 2019 | http://arc.aiaa.org | DOI: 10.2514/6.2019-0786

2015 23th Mediterranean Conference on, 2015, pp. 255–261.

[9] A. Vemuri, R. J. Allemang, and A. W. Phillips, “Estimation of Instantaneous Speed for Rotating Systems: New Processing
Techniques,” in Rotating Machinery, Hybrid Test Methods, Vibro-Acoustics & Laser Vibrometry, Volume 8, Springer, 2016,
pp. 521–535.