CU Hyperloop

2018 Technical Report

June 16, 2018

 Contents
I Introduction 3
I.A About the Hyperloop . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
I.B About the Competition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
I.C About CU Hyperloop . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3

II Pod Design 3
II.A Mechanical . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
II.A.1 Frame and Structure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
II.A.2 Attitude Control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
II.A.3 Brakes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
II.B Aerodynamics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
II.B.1 Shell Design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
II.C Propulsion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
II.C.1 High Pressure Tanks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
II.C.2 Low Pressure Tank . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
II.C.3 Piping, Manifold, and Electronics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
II.D Electrical . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
II.D.1 Power Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
II.D.2 Data Acquisition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
II.D.3 Processing and Data Handling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
II.E Software . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
II.E.1 Communication . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
II.E.2 Software Flowcharts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16

III Sling Design 19

III.A Mechanical . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
III.A.1 Frame and Structure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
III.A.2 Stability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
III.B Electrical . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
III.B.1 Power Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
III.B.2 Propulsion System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
III.B.3 Data Acquisition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
III.B.4 Processing and Data Handling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
III.C Software . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
III.C.1 Communication . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
III.C.2 Software Flowcharts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23

IV Safety 25
IV.A Hazardous Materials on the Pod . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
IV.B Pod Safety Mechanisms and Protections . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
IV.C Hazardous Materials on the Sling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
IV.D Sling Safety Mechanisms and Protections . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26

V Analysis 26
V.A Mechanical . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
V.A.1 Structure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
V.A.2 Braking . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
V.A.3 Sling Launch . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
V.B Aerodynamics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
V.B.1 Shell . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
V.C Propulsion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
V.C.1 Rocket Combustion Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
V.C.2 High Pressure Air Thrust Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31

1
 V.D Electrical . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
V.D.1 Power Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
V.D.2 Interference . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35

VI Conclusion 35

2
 This paper contains a comprehensive overview of CU Hyperloop’s 2018 hyperloop pod
design, which utilizes a carbon fiber body around an aluminum frame supporting three high
pressure gas tanks serving as the primary propulsion method. Electricity is stored in a single
Lithium Metal Hydride battery and distributed using a custom power distribution board. All
on-board processing is done on a TIVA-C series LaunchPad, which is constantly polling sensors
that determine the pod’s position, velocity, acceleration, critical temperatures, vibration, and
pressure, among other data points. In addition to the pod, the team designed a sling capable
of accelerating the pod up to an initial speed prior to the main propulsion system activation.

Introduction

About the Hyperloop

The hyperloop concept was first described by Elon Musk in 2013. In short, by utilizing a low pressure tube and
specially designed capsules, pods carrying people or cargo could travel down the tube between cities at speeds of over
700 mph [1]. Musk outlined his vision for the hyperloop design, an encouraged others to contribute as well. To this end,
SpaceX started the Hyperloop Pod Competition in 2015.

About the Competition

The SpaceX Hyperloop Pod Competition is an annual completion held at SpaceX headquarters in Hawthorn, CA
since 2015. Supported by SpaceX and The Boring Company, the Hyperloop Pod Competition aims to challenge
primarily university students to design and build a hyperloop pod in less than a year. Past competitions have been judged
in multiple categories, but the 2017 and 2108 competitions were judged solely on maximum speed with successful
deceleration [2].

About CU Hyperloop
Formed in the Fall of 2017, CU Hyperloop is a competition team based out of the University of Colorado Boulder.
The 2018 team was made up of approximately 20 undergraduate students from a variety of engineering disciplines and 2
faculty mentors. We strongly believe in open and accessible information, which is why all of our design is detailed in this
report. While our team was not selected to advance to the testing phase of the 2018 Hyperloop Pod Competition, we are
spending the off-season building and testing critical subsystems in preparation for the 2019 competition. Additionally,
our team will be attending the 2018 competition as attendees and guests.

Pod Design

Mechanical
The CU Hyperloop mechanical team was responsible for the overall frame and support structure of the pod. To that
end, the team is also responsible for all CAD work and structural analysis in order to ensure that the pod will not fail
under design conditions.

Frame and Structure

The frame is designed to provide a platform for all components of the hyperloop pod. In order to remain modular,
80/20 Aluminum extrusion was used for the majority of the structure. In order to maintain stiffness, joints would be
welded rather than fastened with traditional fasteners, as seen in Fig. (1). This provides for a strong, stiff frame, and
mounting interfaces would be slotted into the 80/20 rails so they could be repositioned or removed at a later date.

3
 Fig. 1 Locations of welds on the frame

In addition to supporting the critical hyperloop systems, the pod is able to support a single, average sized human
passenger. No life support systems are included, and due to the nature of our design and the scale of the test track, the
ride for the passenger would likely be very uncomfortable (see Sections V.3.A and V.C)

Attitude Control
In order to remain upright and in line with the hyperloop test track, a tricycle wheel configuration was used for the
ground interface and clamping stability wheels are used on the rail, as shown in Fig. (2). The main, weight bearing
wheels were to be made of liquid-cast polyurethane around an aluminum cylinder. The tricycle configuration limited
rotation along the roll axis and additionally limited the maximum translation if any of the stability wheels failed (the
main tricycle wheels would be unable to jump the central rail in the test track).

Fig. 2 Wheel configuration on the pod frame

Suspension aligned in the forward, downward facing direction (see Fig. (3)) allowed the pod to skip over any track
variances with ease. Additionally, the use of springs meant that there wouldn’t be any compressed fluids or gases to
worry about.

4
 Fig. 3 Close up of the rear pod suspension

Figure (4) should the configuration of the alignment wheels on the central track rail. These wheels are wider than the
suspension stroke of the pod, meaning that even in full compression, the stability wheel will remain clamped on the rail.

Fig. 4 Detail view of the stability wheels around the track rail

Brakes
Brakes which clamped onto the center rail were used to bring the pod to a stop after the test, as seen in Fig. (5)
and (6). The brake pad was designed to be as large as possible in order to create the most friction and made of a
yet-to-be-determined polymer so that it was softer than the central rail. The electronic linear actuators are able to output
4450N of dynamic force and 6000N of static force, with variable amounts being applied over time to allow for gradual,
controlled decelerations.

5
 Fig. 5 Overview of the braking mechanism

Fig. 6 Diagram showing the different states of the brakes

6
 For redundancy, the pod has two electric brakes and an additional 2 pneumatic brakes, as shown in Fig. (7). When
pressurized, the pneumatic actuators are able to clamp with 4670N of force, and automatically close if pod power is lost.

Fig. 7 Overview of the braking mechanism compartment showing both types of braking actuators

Aerodynamics

Shell Design
In order to make cooling electronics easier, the chosen operating pressure within the tube was determined to be 10%
atmosphere, or approximately 1.5 psi. As the pod will be operating in some atmosphere, an aerodynamic cover for the
pod was required to reduce drag and protect internals as well as future passengers.
Utilizing research done by Opgenoord and Caplan from MIT [3] and analysis conducted by AzLoop [4], as well as
internal simulation and analysis, the team decided to choose a shell the resembled an airfoil. Aside from producing less
profile drag than a standard pill shaped pod seen in Fig. (8), the airfoil geometry generates a small amount of lift in the
thin atmosphere, resulting in a reduction in the effective weight of the pod. This reduces wheel friction. A number of
design iterations like the one’s shown in Fig. (9) were tested using Star-CCM+. Each model was simulated, areas of
high pressure and turbulence were identified, and optimizations were performed to produce future iterations until the
final design was chosen.

7
 Fig. 8 Example of a pill-shaped pod, courtsey of TU Delft [5]

Fig. 9 Design iterations to the final CU Hyperlop shell design

The final design was simulated with a laminar model for aerodynamic constants and a K-Epsilon model for thermal
profiles with a turbulence intensity of 0.01. The mesh, consisting of 578,000 cells, was tested at a velocity of 90 m/s at
1.47 psi. This resulted in a CL of 0.013 and a CD of 0.026. Flow velocity can be seen in Fig. (10). Anticipated lift and
drag were 23.86 N and 47 N, respectively.

8
 Fig. 10 Flow velocity contour plot of the final shell design at nominal performance conditions

Thermal analysis was also conducted on the shell. Because of the low tube pressure, skin friction is low which
resulted in shell temperature increases of only 4 K. Additional shell testing needs to be conducted to validate the
aerodynamic models. Scale test models will be manufactured over the summer of 2018 and tested in a wind tunnel.

Propulsion

High Pressure Tanks

The pod’s primary method of propulsion comes high pressure Nitrogen tanks produced by Steelhead Composites
[6]. Originally designed for compressed natural gas, these tanks can support pressures up to 3600 psi. Three of these
tanks would be pressurized to 3600 psi, providing a peak thrust of 2000 N per tank (6000 N total). A rendering can be
seen in Fig. (11). These tanks are relatively lightweight and have ports on both ends, allowing for the nozzle hardware
on one side and a fill manifold on other.

9
 Fig. 11 CAD rendering of the 3 propulsion tanks and the nozzle hardware

Nitrogen was chosen due to it’s low cost and ease of accessibility. It’s non-volatile and can be purchased relatively
pure, reducing the need for additional filters. Additionally, since Earth’s atmosphere is primarily Nitrogen, it has
scalability benefits for full scale systems.
Figure (12) shows the different parts of the thrust nozzle. The nozzle piping is straight with exception of the high
pressure electronic valve, minimizing obstacles to the flow. The nozzle itself would be optimized for near-vacuum
expansion with yet-undetermined chamber pressure. Since the chamber pressure is continuously decreasing as gas is
expelled, selecting a single pressure for optimization is difficult and having an adjustable nozzle is overly-complex.
Additional analysis is being conducted to determine the best pressure to optimize at. For safety, a manual in-line
ball valve separates the electronic valve and the tank. This allows for safe transport of the pod after the propellant is
loaded, yet prior to tube closeout. The final step in the pod loading procedure would be to open the manual safety valve.
Additionally, all nozzle parts are certified to 6000 psi, yielding a factor of safety of 1.67.

10
Fig. 12 1. Brass fitting adapting from the tank outlet to the in-line safety valve; 2. In-line ball valve acting as a
manual safety; 3. Brass fitting adapting the in-line safety valve to the electronic valve; 4. ASCO electronic valve;
5. Vacuum optimized nozzle

Low Pressure Tank

A 4th, low pressure propellant tank is also stored on-board the pod. The purpose of the low pressure tank is to
actuate the propulsion valves (item 4 in Fig. (12)) as well as actuate the pneumatic brakes in Fig. (7).

Piping, Manifold, and Electronics

A single fill manifold allows tank pressures during propellant loading to remain equal. A separate plumbing system
would be used to fill the low pressure tank for pneumatic control. Wherever possible, pipes would be solid tubing rather
than flexible, reducing variations in tube diameter around corners and making management of internal systems less
cluttered.
Synchronization of the three thrust valves is key to a successful test. If the valves open asynchronously, imbalanced
thrust could torque the vehicle off of the center rail. With the chosen electronic valves, the maximum time to open is
0.08 seconds, and the maximum software delay to each valve is capped at 20 nanoseconds. Additionally, all of the
piping coming from each propulsion tank is of equal length, lending a further physical mechanism to syncronyze the
thrust. Given this, we anticipate no problem with imbalanced thrust at the start of the thrust phase. Additional research
and testing needs to be conducted to confirm this.

Electrical

Power Systems
All power requirements are met with a single 12.8 V, 40 Ah lithium iron magnesium phosphate battery from Valance
Technologies (now Lithium Werks). The U1-12 XP battery can provide a continuous 80 A and up to 120 A for up to 30

11
sec. The pod power budget can be seen in Fig. (13), and it was determined that the pod needs 197,309 Ws of power.
The battery provides 1,843,000 Ws.

Fig. 13 Power budget for the pod

Battery management is handled by the recommended Battery Management System (BMS) from Lithium Werks.
Using this system, the pod’s battery cell voltages, stack voltages, current draw, charge state, and errors can all be
monitored and data can be relayed to the pod’s control board (discussed further in Section III.D.3).
Power regulation and distribution is handled by a custom manufactured Power Distribution Board (PDB). The PDB
will provide 3V3, 5V, 12V, and 24V lines out to the required pod systems. Current sense and fault monitoring is done
using MSP430FR5989. TI Webench was used for part selection, schematics, recommended board layout, and thermal
analysis (discussed further in Section V.D.1). Schematics for each of the voltage lines can be seen in Fig. (14), (15),
(16), and (17).

Fig. 14 3V3 power rail schematic on the PDB

12
 Fig. 15 5V power rail schematic on the PDB

Fig. 16 12V power rail schematic on the PDB

Fig. 17 24V power rail schematic on the PDB

Data Acquisition
Multiple sensors are used to obtain real-time telemetry on the pod.
There are 5 vibration sensors on the pod. One on the leading edge of the shell, one on the front wheel, one on the
left wheel, one on the left electric brake actuator, and one on the central propulsion tank.
The pod has 7 temperature sensors (thermistors). One on the leading edge of the shell, one on the left brake actuator,
one on each propulsion tank near the exhaust manifold (total of three), and two on the battery (one for redundancy).
An optical camera is located on the shell and pointing upward to count stripes in the track.
Finally, one Inertial Measurement Unit (IMU) is located near the center of mass (CoM) of the pod.

13
Processing and Data Handling
Only some of the data that the sensors collect is actually sent back to the ground station, the rest is used in real time
and simply logged for later analysis by the team.
Since the pod has not been tested in the real world, all any operational limits are determined solely from manufacturer
datasheets and computational models. All of the sensors will feed data to the processing and control board, and
depending on the sensor, a braking command will be sent.
Since we can’t validate an appropriate magnitude of acceptable vibration without an actual pod test, the vibration
data will be compared against a maximum acceptable value. This magnitude will be set high for the first few tests. Data
will be used to validate models and set a more reasonable limit.
The same approach will be used for the thermistors, with exception of the battery temperature. The battery
temperature limit will be set based off the manufacturer recommendations and will not change without extensive in-house
testing and acceptable reasoning.
The camera, a LUX 330 EVS, is the primary method of determining the pod’s position withing the test track. It has
a built-in FPGA for rapid image processing, and can sample the environment at up to 8,000 fps, allowing for up to 4
images of a single stripe as the pod moves through the tube. With SpaceX’s approval, our team will opt to shut off the
built-in track lighting and instead use our light ring to illuminate the stripes. This allows for more consistent lighting
and therefore more consistent stripe counts. The camera+FPGA would determine stripes by measuring the intensity of
the red, green, and blue color channels.
The IMU serves as a backup for the camera+FPGA. It is susceptible to temperature fluctuations and error in
acceleration is compounded during integration, but it still serves as a reasonable secondary system. Additionally, the
IMU is required by SpaceX.
The final method of pod positioning is a simple timer. If all else fails, the pod will still trigger it’s brakes after a set
amount of time, serving as a final safety mechanism preventing the pod from crashing into the end of the tube.
The actual pod processing and control comes from a TIVA C. This was selected because of the built-in Ethernet
port, fast processing times, and heritage at CU. The biggest downside of the board is that there aren’t enough analog
inputs for all of our sensors. This problem is solved using the AD5592R analog to digital converter (ADC) from Analog
Devices. This would interface with the TIVA C via SPI, and has a sampling rate of 400 kHz with 12-bit resolution.

Software

Communication
Intra-pod communication depends on the devices that need to talk to one another. The FPGA from the camera uses
UART. The thermistors, vibration sensors, and pressure transducers use analog inputs. The IMU communicates over
I2C, and the TIVA C communicates to the SpaceX network access panel (NAP) over Ethernet. The NAP then sends
data to the ground station laptop over a dedicated network set up by SpaceX. Access to this network within the test track
is met with a radiating Ethernet cable, or a leaky feeder.
The data that is sent over this network changes slightly depending on the state of the pod. Figure (18) details the
data that is sent to the ground station when in the ready to launch (R2L) state, and Fig. (19) shows what is sent post
launch (PL).

14
 Fig. 18 Communication table when in the R2L state

Fig. 19 Communication table when in the PL phase

This difference in sent telemetry is due to the fact that the sling does not have a dedicated NAP and therefore cannot
communicate over the network without a physical connection to the pod. Therefore, sling status is only sent when in
R2L mode, as the pod will detach itself from the sling once the launch command is given.
The launch command can only be sent from the ground station laptop. The laptop also controls the manual emergency
stop, but the likelihood of an operator being able to recognize erroneous streaming data before the pod is highly unlikely.
The pod would trigger a stop command before the ground station operator. If at any point communication between the
ground station and the pod is interrupted, the pod will issue a stop command.

15
Software Flowcharts

Fig. 20 Software overview

16
 Fig. 21 General program

Fig. 22 Failsafe sub-process

17
 Fig. 23 Safe state sub-process

Fig. 24 Sling activation sub-process

18
 Fig. 25 Pre-Launch sub-process

Sling Design
CU Hyperloop’s propulsion system works best when it’s at maximum pressure. As the air is used up, pressure
decreases and overall thrust is reduced. Therefore, any initial speed that can be given to the pod before the main
propulsion systems fire will greatly increase the maximum attainable speed, as the inertia of the entire system will be
reduced. In past years, SpaceX has provided a pusher kart to all teams to serve this function. This kart is not provided in
the 2018 Competition.
Because of this, we set out to design our own. Our solution involved the use of an electric winch system to ’slingshot’
the pod to an initial velocity. Furter detail will be discussed in the sections to follow.

Mechanical

Frame and Structure

In order to impart the lowest g-force on the pod, the sling needs to be long. Per SpaceX competition requirements
however, the maximum pod length is 24 ft. This was the largest constraint on the mechanical design of the pod, and in a
full scale hyperloop, we invision a 300 ft long or larger sling at every hyperloop station.
The structure of the sling is built to support the electric motors and the winch system. A CAD model can be seen in
Fig. (26), and the mass breakdown by system is in Fig. (27). Unlike the pod, the sling never has to move, so mass is not
a concern other than to keep the entire design (pod + sling) under the weight limit set by SpaceX. Where the pod is
made of aluminum and carbon fiber to reduce weight, the sling is constructed from primarily steel in order to increase
strength and keep manufacturing simple.

19
 Fig. 26 CAD model of the sling

Fig. 27 Mass breakdown of the sling system by major components and assemblies

The sling’s main support comes from the two L-beams that run along the length of the device. Additional 80/20
supports are used for mounting the majority of the hardware. The overall design is fairly simple, as most of the team’s
focus went into the pod. In the future, we want to further develop the sling concept.

Stability
As the sling remains stationary, "stability" refers to the sling’s ability to stay clamped to the center rail and ramain
pointing down the tube. To accomplish this, the sling has four of the same electronic actuators that the pod uses
for braking. These provide 6000 N of force each, keeping the sling in one place during launch. The rail clamping
mechanism is exactly the same as on the pod, as seen in Fig. (28).

20
 Fig. 28 The brakes used on the sling in order to maintain stability

Electrical

Power Systems
The electric motors for the sling propulsion (discussed further in Section IV.B.2) are fairly power hungry, but they
only operate for a short time. The faster these motors can turn, the more of a boost the pod will receive. Because of this,
we opted to use two batteries on the sling, one for each motor. These batteries are from the same manufacturer as the
pod batteries, but they’re one size larger at 110 Ah. They can take a peak current draw of 300 A for up to 30 seconds,
which is more than enough for our needs. Just like the pod, the sling has it’s own seperate battery management system
from the manufacturer.

Propulsion System
The propulsion system consists of two 10.75 hp electric motors which serve as a winch as seen in Fig. (29). The team
considered other methods of propulsion, such as magnets, compressed air, and hydrolics, but chose electric motors due
to their simplicity and ease of manufacturing and integration. In the future, these other methods will be further explored,
as they offer significantly greated performance (see Section VI.A.3 for further analysis on the sling propulsion).

21
 Fig. 29 Concept image of the sling propulsion mechanism with winch lines drawn

Data Acquisition
The sling will have up to seven thermistors monitoring the health of various critical subsystems. Firstly, there will
be one thermistor for each battery. Another two thermistors will monitor the temperature of the electric motors, and two
more will be positioned to take temperature readings of the pulleys supporting the winch. THe final thermistor will
monitor the temperature of the TIVA C LaunchPad processor. No further sensors are included on the sling, as no other
data is important to it’s operation.

Processing and Data Handling

Processing on the sling will be handled by a dedicated TIVA C LaunchPad, though a smaller, less powerful control
board might be chosen in the future. Since the sling only takes data from thermistors, there is not much data handling to
actually be done. Temperature limits will be set for each sub-system, and any systems outside of this limit will trigger a
warning on the ground station (more information in Section IV.C.2).

Software

Communication
In order to communicate with the ground station, the sling has to be connected to the SpaceX wireless network.
Without it’s own NAP, it can not do this, and the team’s NAP is being utalized on the pod. However, communication
with the sling is not necessary after the launch of the pod. Therefore, sling communication will occur through a wired
connection to the pod prior to launch through a break-away communication line. After launch, this connector breaks
away and the sling goes into a safe state and is ready to be recovered. Any information pertinant to the recovery will be
displayed through an on-board LCD screen, and critical safety information will be broadcast through the use of flashing
lights.

22
Software Flowcharts

Fig. 30 Pod-Sling launch states and their outcome

Fig. 31 Pod-Sling launch states and their outcome

23
Fig. 32 Pod-Sling launch states and their outcome

Fig. 33 Pod-Sling launch states and their outcome

24
 Fig. 34 Pod-Sling launch states and their outcome

Safety
Safety of all systems is critical for success and a number of safety features are in place in order to ensure safe
operation.

Hazardous Materials on the Pod

The two sources of hazard from a materials perspective on the pod are the batteries and the three propellant tanks.
The battery is a Lithium-Iron Magnesium-Phosphate battery in a sealed container. The team still needs to verify
vacuum compatibility, but the battery type is know to be quite safe.
The propellant tanks are aluminum lined, composite overwrapped pressure vessels rated to 3600 psi and designed
for natural gas. Transportation of the tanks will do done to DOT regulations and the tanks will be chained up whenever
they aren’t in use.

Pod Safety Mechanisms and Protections

The pod is most susceptible to damage in a fault during the acceleration phase. Because of this, there are a number
of software and hardware safety features in place to prevent failures and subsequent pod or track damage.
First, the electronic brake actuators cannot change position without a signal from the processor. This helps prevent
accidential braking,but it also means that in the event of a power loss, the pod cannot stop using the electronic brakes.
To combat this, the pod is also equipped with pneumatic brakes which are electronically held open. This means that in
an electrical failure, the pneumatic brakes would automatically close and stop the pod before the end of the track.
There are 2 batteries on board, and each can handle the full load of the pod. This adds considerable weight, so in
thefuture, the team may decide to drop the redundant battery and replace it with a smaller emergency battery that can
trigger emergency systems in the case of a critical failure.
Since the propulsion system contains a large volume of high pressure gas, numerous safety measuers are in place in
order to ensure safe operation. First, all the tanks are certified from the manufacturer. Aside from the software safety on
the nozzle valves, there is a physical safety valve between the tank and the nozzle. Activating this valve would be the
final step in the pod loading prcedure, as this valve is completely independent of any electronics and acts as the final
safety to the propellant system. On the software side, the propulsion valves are isolated depending on the launch state

25
(i.e. the valves cannot be toggled when in the "braking" state). Additionally, like the pneumatic brakes, the propulsion
valves are normally closed, meaning that if the pod experiences a loss of power, the propulsion valves will close.
The pod has minimal single points of failure. Of note are the wheels and suspension and the breakaway connection
between the sling and the pod. If the wheels or suspension fail, pod and/or track damage is likely, though depending on
the speed, the pod would likely skid to a stop. The sling -> pod connection is more uncertain, because the severity of the
failure depends on the conditions under which it failed. It could be as beneign as having to reset the device, or it could
be as severe as the pod propulsion firing while still connected to the sling (of which, the final outcome is unknown).

Hazardous Materials on the Sling

The only source of hazardous materials on the sling are the two high power batteries. These batteries are of the
same type as the ones in the pod and are of no more concern than the pod batteries.

Sling Safety Mechanisms and Protections

The sling also has a number of safety features in place in order to prevent accidents. Firstly, the voltage, current, and
temperature of each battery is monitored by the LaunchPad and checked against known limits. Additionally, the sling is
always in standby mode unless the launch signal is received, and in order to launch, a physical safety switch needs to be
activated on the sling.
Two seconds after receiving a launch command, the sling automatically re-enters standby mode. This ensures that
the sling remains safe to approach regardless of if the pod actually launched. Since the electronic brakes don’t move
without an explicit command and power, once they are secured, they cannot unclamp from the rail without instruction.
This ensures safe operation even in the case of a power failure.
During the launch, a detachable connector disconnects from the pod, signalling a successful seperation. Once this
connection is broken, the motors on the sling disengage, preventing overdraw from the motors.
The biggest single point of failure on the sling is the winch wire. A failure here would result in imbalanced pull
from the sling without any indication to the sling controller (as would be the case in a motor failure). A failure of the
winch wire would likely cause considerable damage to the pod and track. Presumably, minimal damage would occur to
the actual sling, as the number of moving parts and mechanisms beyond the end of the pod are minimal.

Analysis

Mechanical
Structural analysis was done using ANSYS. A factor of safety of 2 was used for all structural analysis.

Structure
The two mainpoints offocus for structure analysis were the tank mounts and the frame assembly.
A force of 4000 N was applied to each bell housing as shown in Fig. (35). The maximum stress was determined to
be 4.18 ∗ 107 mN2 , which is 75% of the yeild stress of the material.

Fig. 35 FEA analysis of the propellant tank mounting bells

26
 For the frame, a global deceleration of 2 g was applied to the entire structure. Using "Aluminum Alloy" as the
material (actual alloy would be Al 6105-T5), the maximum stress was determined to be 2.4 MPa, which is only 8%
below the tensile stress limit of Aluminum. While this leaves a narrow margin, we postulate that the carbon fiber shell
would inclrease the stiffness and strength of the frame considerably. Further analysis needs to be done to connfirm this.
An image of the alaysis shown using a logarithmic scale is shown in Fig. (36).

Fig. 36 FEA analysis of a 2 g deceleration of the entire frame

Braking
The extruded aluminum brake mounts were also analyzed for stress and displacement. Two 9000 N loads were
applied to the face of where each linear braking actuator attaches. Stress was low in these beams, and the overall
maximum displacement was 0.22 mm, as seen in Fig. (37).

27
 Fig. 37 FEA analysis of the brake mounts

Sling Launch
The sling was analyzed to ensure the structure could withstand the forces of a launch. Force was applied to the
vertical supports at the end of the sling, as that is where the pullies would wrap around and pull the winch cables and
thepod from.
The maximum displacement was nearly 4 mm at the ends of the aluminum supports, and the maximum stress was
also on these supports at 62 MPa, as seen in Fig. (38). This is well below the limit of Al 6061-T6, though the high
displacement is concerning if we’re planning on doing multiple launches. Further analysis and trade studies will go into
selecting a more suitible material for these vertical supports if we decide to persure this launching mechanism in the
future.
Note that the overall displacement of the L-beams in the sling is low, due mostly to two reasons: First, there is
inherant strength in the geometry of the L-beam in the direction that the moment is being imparted; Second, the L-beams
are made of Structural Steel rather than Aluminum.

Fig. 38 Displacementof the sling during intial acceleration

28
Aerodynamics

Shell
Aside from the analysis done in Section III.B.1, the shell was also analyzed for thermal variations at speed. Across
the entire shell at 90 m/s, the temperature only varried by 4 K as seen in Fig. (39). Note that none of the internal
electronics or propulsion tanks were examined in this analysis, so actual temperatures would likely be different. However,
this does prove that in the low pressure environment in the tube, temperature fluctuations from skin friction on the shell
are negligible.

Fig. 39 Displacementof the sling during intial acceleration

Propulsion
Two different models were developed to measure the performance of the pod’s propulsion mechanism. An important
note is that all the propulsion analysis was done independently of the sling. Therefore, for simplicity, any velocity from
the sling would add to the overall velocity of the pod in a linear fashion.

Rocket Combustion Model

The first model was based off of rocket engine combustion performance and from Wheeler [7], Evans [8], and
Anderson [9]. It is a simple model using simple integration and considering only wheel friction as an external force
working against the pod. This simplicity makes it great for trade studies and initial down-selection, but doesn’t
necessarily make it a reliable final result.
Some of the main downsides are that the model assumes perfect expansion about an ideal nozzle. It also assumes
that there are no pressure or ṁ losses due to the piping geometry. Finally, this model doesn’t take into account the
changing mass of the pod over time.
As seen in Fig. (40), this model predicts that the pod reaches it’s maximum velocity of approximately 179 m/s after
about 4.5 seconds. Without truncating the thrust period, the pod would coast past the end of the track at these speeds
given our current braking setup, which can be seen in Fig. (41).

29
 Fig. 40 Predicted velocity profile of the pod

Fig. 41 Predicted position profile of the pod with the track end shown

30
High Pressure Air Thrust Model
Model two is a bit higher fedility, being developed for a water bottle rocket pressurized with high pressure air
[10]. This model uses differential equations to model changing mass due to ṁ, accounts for realistic nozzle dynamics
like choked and un-choked flow, and has the capability to further model converging-diverging nozzle geometries.
Additionally, this model has been validated against a real-world system (a water bottle rocket, where it achieved 95%
model accuracy) and has been further refined for hyperloop applications [11] [12].
The biggest downsides of this model are the linear thrust scaling and that there are no losses due to piping geometry.
The second model predicteda lower top speed of around 116 m/s, which was reached within the same 4.5 sec as seen
in Fig. (40). In this case, the thrust phase would not need to be truncated in order to avoid colliding with the end of the
wall, as the top speed is low enough to warrant a small coast phase prior to braking, which can be seen in Fig. (41).

Electrical

Power Systems
Analysis on the efficiancy of each of the power lines given arrying battery supply voltages was performed and can be
seen in Fig. (42), (43), (44), and (45). The effieicency of each line is reletively similar for varying voltages in all but the
12 volt line where the efficieny drops at a supply voltage of 11.5 V as seen in Fig. (44). However, since the nominal
supply is 12.8 V from the battery, this isn’t a concern.

Fig. 42 Transmission efficiency of the 3V3 power rail given varrying input voltages

31
Fig. 43 Transmission efficiency of the 5V power rail given varrying input voltages

Fig. 44 Transmission efficiency of the 12V power rail given varrying input voltages

32
 Fig. 45 Transmission efficiency of the 24V power rail given varrying input voltages

Additionally, thermal analysis was performed on each layer of the distribution boards, taking into account the lower
pressure within the tube and the wait time during pump-down. The results of this analysis can be seen in Fig. (46), (47),
(48), (49), (50), and (51). The only concern with regards to thermal was on the 5V line, where the TPS565201 chip
reached tamperatures of around 100 o C. The operating range of this chip goes up to 125 o C.

Fig. 46 Thermal response of the 3V3 power rail

33
Fig. 47 Thermal response of the 5V power rail

Fig. 48 Thermal response of the 12V power rail

Fig. 49 Thermal response of the 12V power rail

34
 Fig. 50 Thermal response of the 24V power rail

Fig. 51 Thermal response of the 24V power rail

Interference
CU Hyperloop didn’t prioritize electrical and magnetic (EM) interference analysis, but we do recognize that the
radiating ethernet cable could pose a problem for our electronics. Further analysis needs to be performed to ensure that
all electrical systems are sufficiently shielded from any EM radiation.

Conclusion
The CU Hyperloop pod is still in active development. This is the first year we’ve been a team, and the first time any
of us have undertaken a project of this scope. There is lots to learn, and we will be moving forward and continuing
development of our pod and sling design in the years to come.
The entire team worked very hard all year to make the contents of this paper possible. Nonetheless, there are still
numerous things was can improve on. From a managerial standpoint, finding a balance between school and hyperloop
proved challenging for everyone, and this is something I hope we can work out better in the 2019 season.
On the technical side, I think we are on the right track, but more analysis should be conducted earlier. Of priority for
the summer and next season is vibrational analysis and vacuum compatibility analysis.
We also want to start working more with both the campus community and industry. If you’d like to get involved,
don’t hesitate to reach out to me at [email protected].

35
 Acknowledgements
I’d like to specifically acknowledge the entirety of the CU Hyperloop team:
Jack Maydan
Gavin Decamillis
Meg Vader
Bum Soo Kim
Christian Carmack
Nolan Ferguson
Ashley Montalvo
Conrad Trybus
Jordan Hammer
Kevin Guinn
Markia Schubert
Boskin Erkocevic
Alanna Siler
Alex Bertman
Samuel Leon
Eric Daugherty
Christopher Mann
Yuma Yagi
Dishank Kathuria
Jett Moore
Stephen Moore
Stephen Peng
Lauren Lee

I’d also like to thank our faculty advisors:

Arielle Blum
Phil Larson

Additional support for this report comes from Trevor Weschler.

Team Sponsors
Without the support of our industry sponsors, CU Hyperloop would be unable to continue active development of our
hyperloop system.
The Engineering Excellence Fund
ANSYS
CU Representitive Council
The Vice Chancellor’s Office for Student Affairs
Steelhead Composites

References
[1] Musk, E., “Hyperloop Alpha,” Tech. rep., SpaceX, 2013.

[2] SpaceX, “2018 SpaceX Hyperloop Pod Competition Rules and Requirements,” Tech. rep., SpaceX, 2017.

[3] Opgenoord, M. M. J., and Caplan, P. C., “On the Aerodynamic Design of the Hyperloop Concept,” Tech. Rep. AIAA 2017-3740,
Massachusetts Institute of Technology, 2017.

[4] AzLoop, “TU Delft Hyperloop Team Website,” , 2017.

[5] Hyperloop, D., “Arizona’s SpaceX Hyperloop Competition Team Website,” , 2018.

[6] Composites, S., “Steelhead Composites Website,” , 2017.

36
 [7] Wheeler, D., “Modeling the Thrust Phase,” Tech. rep., 2002.

[8] Evans, J., “University of Colorado Boulder ASEN 3113,” , 2017.

[9] Anderson, J. D., Fundamentals of Aerodynamics, 5th ed., Boston: McGraw-Hill, 2010.

[10] Ann, and Department, H. S. A. E. S., “University of Colorado Boulder ASEN 2012/ASEN 2004,” , 2017.

[11] Chiasson, T. M., and Lozano, P. C., “Modeling the Characteristics of Propulsion Systems,” Tech. rep., Massachusetts Institute
of Technology, 2012.

[12] Anis, A., “Cold Gas Propulsion System - An Ideal Choice for Remote Sensing Small Satellites,” Tech. rep., NED University of
Engineering and Technology, Pakistan, 2012.

Appendix

Technical Risk Assesment as of Jan. 2018

Fig. 52 Risk assesment plot

37
Fig. 53 Description of the risks presented in Fig. (52)

38