The Future of Hyperloop

By Delft Hyperloop

A report presented to the Dutch

Ministry of Infrastructure and Water Management

Full Scale Hyperloop Department

Delft Hyperloop
Delft
June, 2019
 Full Scale Hyperloop
Stichting Delft Hyperloop
Delft University of Technology
www.delfthyperloop.nl
[email protected]

An overview of the current state of hyperloop

development and future recommendations as
envisioned by Delft Hyperloop

All information given in this report is gathered by extensively studying the hyperloop environment,
together with the help of industry experts. All findings in this report are described as veracious and
objective as possible, to our best intent.

Contributors:
J.K. van Leeuwen
J.M.P. Lohle
T.R. Speelman
Y. van der Tang
M.H. Teeuwen
T. Vleeshouwer

This work is licensed under a Creative Commons “Attribution-NonCommercial-ShareAlike 4.0 International”

license.
This page was intentionally left blank.
Executive Summary
All common modes of passenger transportation show an increase in demand, leading to a growth in problems: capacity
shortage, environmental pollution and disruptive situations. A promising solution for these problems is the hyperloop:
a high-speed transportation system using near vacuum tubes in which pressurised vehicles travel. This report, commis-
sioned by the Dutch Ministry of Infrastructure and Water Management, provides an update on hyperloop development.
The main goal of this report is to give an objective overview of hyperloop knowledge, in order to accelerate the realisation
of a hyperloop. Furthermore, recommendations are giving to the Ministry of Infrastructure and Water Management. At the
time of publishing, multiple student teams, companies and non-profit organisations are working to realise a hyperloop.
However, they have different visions of a hyperloop system and convergence towards certain technologies or parameters
is yet to be reached.

Technology
Multiple technologies have potential for a hyperloop, and additional research and development has to be conducted to
determine the most feasible technologies. Delft Hyperloop has researched several subsystems, and based on this research,
recommendations are given.

Levitation: Electrodynamic Suspension (EDS) and Electromagnetic Suspension (EMS) are the two most promising tech-
nologies for levitation, as they are currently used in Maglev trains. For the operational speeds of a hyperloop, both tech-
nologies still need to be proven. EMS is very promising because of its low energy consumption. However, as the system
needs an active control system to levitate, it is less reliable from a safety point of view. EDS, which is a fail-safe passive
levitation method with relatively large air gaps, is therefore opted for as the most promising levitation option as safety is
an important design criteria.

Propulsion: A Linear Synchronous Motor (LSM) and a Linear Induction Motor (LIM) are the two most promising tech-
nologies for propulsion. A LIM performs best in terms of costs and reliability, however current technology is not able to
reach the same speed as an LSM. Furthermore, LSM is the best option regarding energy consumption. During further
development, the reliability and safety of an LSM should be increased.

Passenger Pod: Although the near vacuum environment, passenger pod design should take aerodynamically optimised
shapes into account to minimise aerodynamic drag. The capacity of a pod is recommended to be around 50, to cope
with the expected demand in a European network. Based on aircraft design guidelines, a pod diameter of 2.7 m is recom-
mended to fit three seats abreast, including an aisle. This includes 0.2 m for structural components including isolation.
Furthermore, to diminish station size, it is recommended to have a bidirectional pod.

Tube: With a blockage ratio of 0.7, the tube diameter becomes 3.5 m. Steel is recommended as material, however it is
useful to closely pay attention to new technological developments of other materials. Developments to watch are steel
weight optimisations and fibre reinforced polymers. If additional aesthetic value is needed, acrylic transparent tubes
might be suitable, although expensive. A steel tube thickness of 25 mm is advised to withstand vacuum buckling, including
a safety factor. More extensive calculations are needed for the final design of the tube.

Vacuum: The optimal tube pressure depends on the pod frequency in the tube: the more pods in the tubes, the more
efficiency is gained when the pressure is lower, because all pods will experience a reduced aerodynamic drag and there-
fore reduced power consumption. For a frequency of 2 pods per minute, the optimal tube pressure is 3 P a, under the
assumption that the pumping speed and the power required per vacuum pump do not vary with pressure. It is advised to
have a variable tube pressure, that varies dependent on the changing pod frequency.

Communication: Current communication technologies used in high-speed railways are not suited for hyperloop speeds.
The main challenge lies within the communication from the pod to the outside world, which is necessary to exchange data.
Optical fibre is a technology that has potential to solve this challenge. Furthermore, there is a possibility that new commu-
nication protocols get developed in the coming years, such as 5G. This would erase the need for current technologies in a
hyperloop system. However, as the development of new technologies is uncertain, it is useful to further develop current
technologies in order to make them suitable for high-speed transportation.

i
Artificial Intelligence: Application of Artificial Intelligence is promising for the hyperloop. This can be used during
designing, building, operating and maintaining the system. For example incident detection, ensuring on-board safety
and prediction of (sub)system failure, can be areas where Artificial Intelligence can be applied. Next to that, it can be used
for optimising security checks, timetables and scheduling.

Impact of a Hyperloop
A European hyperloop network designed by Delft Hyperloop, is able to transport over 300 million passengers yearly by
replacing a share of the short-haul air passenger transportation. The network connects 48 of the largest cities in Europe
and can take over two-thirds of all passengers flying between these cities. By doing so, the environmental impact of
transportation will be diminished, as a hyperloop is fully electric. It is estimated that the hyperloop infrastructure costs
approximately €38 million per kilometre above-ground, and €61 million per kilometre underground. This means that the
complete designed network, which requires 19.700 km of bidirectional tube, would cost close to a trillion euro. Next to
providing high-speed transportation, a hyperloop network will positively impact society. By increasing the connectivity
between cities, welfare will be enlarged.

Safety
Safety is the most important aspect of a hyperloop system. To increase passenger safety, hazard mitigation methods need
to be incorporated during design. It is important to maintain and monitor all subsystems thoroughly to ensure a reduced
likelihood for failures. Furthermore, a redundant power supply system is desired to ensure safe operation in case of a
power outage. As communication is one of the most critical subsystems, it is advised to incorporate a secondary commu-
nication system. Next to that, the tube is the most sensitive subsystem for black-swan risks: risks with a low likelihood
and severe consequences.

To guarantee safety, it is important to test all subsystems thoroughly, which shows the need for a test facility where pods
can reach speeds over 1000 km/h. A high-speed test facility is also necessary for certification purposes. For this certifi-
cation, it is advised to found an agency that is responsible for hyperloop certification. Multiple stakeholders should be
involved in this agency.

To minimise the cost for emergency exits, while still guaranteeing safety, it is recommended to further investigate the
potential of Safe Havens. Safe Havens are intermediate emergency exit stations, where pods can stop to provide safe exit
to the passengers. Whilst doing so, impact on the operation of the rest of the system is minimised.

Challenges
Hyperloop is a promising innovation in order to decrease the rising problems in transportation, however there are still
various challenges. To increase interoperability, it is important that there will be a single European standard. However, it
is necessary that this does not happen too early in the process, as multiple techniques have to be developed first in order
to research their potentials. To accelerate development of these techniques, policy support for infrastructure is essential.
Standardisation is hard to achieve, as it is difficult to converge to an optimal system. Furthermore, similar to existing
infrastructure, hyperloop infrastructure is expensive. It is nearly impossible to find a single party that is able to finance the
construction of a complete European network, which is why a public-private partnership is most likely needed. However,
as governments will be involved, political aspects will make this more complex. Finally, high-speed switches are essential
for realising efficient point-to-point connections. However, as these switches do not yet exist, these need to be developed.

Conclusions
The hyperloop provides a sustainable solution to the growing demand for high speed travel within Europe. In order to
successfully realise a hyperloop system, two main focus points can be identified: working towards standardisation and
setting up the foundation for implementation. It is recommended to start setting up a framework for standardisation
in the future. The three main steps will be continuing research and development, sharing knowledge and investing in
a long test track. Planning for actual implementation should already start in the early phases of the project in order to
create maximum benefit for future passengers and European citizens in general. The three main steps will be to found a
European agency for certification, secure financing and determine the location of the first links.

ii
Table of Contents
Executive Summary i

1 Introduction 1

2 Current State Hyperloop Parties 3

2.1 Hyperloop University Teams . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
2.2 Hyperloop Companies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
2.3 Non-profit Hyperloop Organisations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
2.4 Websites . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
2.5 Test Tracks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5

3 Levitation 6
3.1 Subsystem Requirements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
3.2 Analysis of Potential Levitation Mechanisms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
3.3 Trade-off . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
3.4 Recommendation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13

4 Propulsion 14
4.1 Subsystem Requirements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
4.2 Analysis of Potential Propulsion Mechanisms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
4.3 Trade-off . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
4.4 Recommendation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18

5 Pod Characteristics 19
5.1 Situation Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
5.2 Recommendation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20

6 Tube Characteristics 21
6.1 Subsystem Requirements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
6.2 Tube Material . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
6.3 Tube Thickness . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
6.4 Pillar Distance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
6.5 Thermal Expansion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
6.6 Recommendation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25

7 Vacuum Analysis 26
7.1 Vacuum Pumps . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
7.2 Situation Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
7.3 Recommendation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27

8 Pod Communication 29
8.1 Theoretical Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
8.2 Situation Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
8.3 Communication for Sensor Data and Commands . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
8.4 Location and Pod ID . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
8.5 Required Hardware . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
8.6 Recommendation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34

9 Artificial Intelligence Applications 35

9.1 AI in Current Public Transportation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
9.2 Requirements for the Applications of AI . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
9.3 Areas to Apply AI in Hyperloop . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
9.4 Recommendation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38

10 Cost Estimation 39
10.1 Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
10.2 Infrastructure Costs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
10.3 Station Cost . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
10.4 Pod Costs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
10.5 Recommendation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44

11 European Hyperloop Network 45

11.1 Methodology for Network Design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
11.2 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46
11.3 Societal Impact of a Hyperloop Network . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47

iii
 11.4 Recommendation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48

12 Safety Analysis 49
12.1 Importance of Safety . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
12.2 Research Approach . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50
12.3 Hazard Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
12.4 Safe Haven Design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
12.5 Recommendation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58

13 Regulatory Implications 59
13.1 Standardisation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59
13.2 Legislation and Policies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59
13.3 Certification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59
13.4 Recommendation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60

14 Future Points of Notice 61

14.1 Main Barriers for Hyperloop Development . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61
14.2 Challenges . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61

15 Conclusions 63

Bibliography 65

Glossary 67

Appendices 70

A Environmental Analysis 72

B Levitation Design Criteria Description 73

C Levitation Post-Chapter Remarks 74

D Technology Readiness Level 75

E Pod Characteristics Methods 76

F Tube Characteristics Tables and Calculations 78

G Vacuum Analysis Background 80

H Network Analysis 81

I Safety Hazard Analysis 82

iv
1.Introduction
Each year, people tend to travel further and further. Except for sea transport, all modes of transportation show an in-
crease in passenger kilometres travelled [European Environment Agency, 2018]. This growth unfortunately also leads to
an increase in problems. The capacity of existing infrastructure does not suffice for the growing demand, leading to large
waiting queues and traffic jams. Additionally, current modes of transportation are energy inefficient and thereby large
contributors to rising carbon dioxide levels and climate change. Furthermore, current modes of transportation are sensi-
tive to external influences, leading to large disruptive situations.

A promising solution for these problems is the hyperloop: a high-speed transportation system using near vacuum tubes
in which pressurised vehicles travel. The concept of hyperloop was initiated back in 1799 as an atmospheric railway by
George Medhurst and got reintroduced by the publication of the white paper of Elon Musk in 2013. By reducing the air
resistance and rolling resistance, a hyperloop could travel efficiently with speeds over 1000 kilometres per hour. The en-
closed environment of the system ensures reduced external influences.

Goal of this Report

This report is commissioned by the Ministry of Infrastructure and Water Management to stay updated on hyperloop de-
velopment. Therefore, this report is written for the Ministry of Infrastructure and Water Management. However, anyone
interested in hyperloop, with different levels of expertise, is encouraged to read this report. The report is written to be
comprehensible for everyone with general hyperloop knowledge. For people with less specific knowledge, a glossary is
provided at page 67 of this report. Readers with a more technical interest that want to go more into detail on specific
subsystems can find according information in the appendices.

The main goal of this report is to give an overview of hyperloop knowledge, in order to provide recommendations to ac-
celerate the realisation of a hyperloop. An overview of companies involved in development is given, as well as appropriate
technologies for each subsystem. Trade-offs are made to indicate the advantages and disadvantages of every technology.
Another important subject is safety: the main safety risks per subsystem are determined to show the critical points in
hyperloop design. Furthermore, the main barriers and challenges for a hyperloop are emphasised, in order to determine
the key steps in hyperloop development. Ultimately, all information and recommendations provided by Delft Hyperloop,
offers the Ministry a framework to help decide on how to proceed as the Netherlands in hyperloop development.

Delft Hyperloop
This report is written by Delft Hyperloop III, a student team from Delft University of Technology. Delft Hyperloop com-
petes in the SpaceX Hyperloop Pod Competition initiated by Elon Musk, where they received the first and second place
in the previous years, out of student teams all over the world. Besides the competition, a part of Delft Hyperloop focuses
on open-source researching the implementation of a future hyperloop system, the results of which are described in this
report. The goal of this department of Delft Hyperloop is to share knowledge in order to stimulate hyperloop development.

Overview of Report
In order to give a comprehensive overview of hyperloop knowledge, a broad overview of related subjects is discussed.
First, an overview of companies, non-profit organisations and student teams working on hyperloop is given in Chapter 2
to list the most relevant hyperloop parties. Top-level system parameters are listed for each of the companies to determine
their overlap and differences in vision.

Afterwards, the main subsystems of a hyperloop are described in Chapters 3 to 8, i.e. levitation, propulsion, pod char-
acteristics, tube characteristics, vacuum analysis and pod communication respectively. Multiple potential techniques to
be used are compared, and trade-offs are made to determine the most feasible techniques. This is followed by Chapter 9
that highlights the possible applications of Artificial Intelligence. Every chapter is concluded with a recommendation by
Delft Hyperloop.

With the parameters of all main subsystems determined, the expected costs are estimated in Chapter 10, based on the
techniques recommended by Delft Hyperloop. The infrastructure, station and pod costs are presented. These costs are
subsequently used as input to develop a European hyperloop network. The network, presented in Chapter 11 is based on

1
Chapter 1. Introduction

replacing short-haul air passenger transport.

Afterwards, Chapter 12 describes the safety of a hyperloop system on a top-level. A Hazard Analysis is conducted to
determine the largest safety risks. Hazard mitigation methods are defined to illustrate important design implications.
Furthermore, a new safety concept (Safe Haven) is described, which optimises passenger safety while minimising emer-
gency exit costs. This is followed by Chapter 13, discussing the legislation and certification implications of a hyperloop.

Finally, future points of notice are discussed in Chapter 14, covering the main barriers and challenges for hyperloop de-
velopment. Chapter 15 concludes this report by highlighting the focus points in order to successfully realise a hyperloop.

As the hyperloop system consists of many different subsystems, it is too elaborate to discuss everything in detail. There-
fore, some interesting subjects are left out of this report. On the one hand, because the topics are still uncertain, such as
exact energy consumption or the choice for airlocks or jetways at stations. On the other hand, because the subjects relate
less to the goals of this report compared to other subjects discussed. Examples of this are secondary suspension in pods,
energy supply to the system and solar panels.

2
2.Current State Hyperloop Parties
The purpose of this chapter is to provide an overview of everyone actively contributing to hyperloop development as
of 2019. This can be divided into university teams competing in the SpaceX Hyperloop Pod Competition, commercial
companies and non-commercial organisations. Besides, two websites tracking hyperloop development are mentioned.
Finally, an overview of the operational and planned hyperloop test tracks is provided.

2.1 Hyperloop University Teams

The student teams invited to join the final round of the SpaceX Hyperloop Pod Competition 2019 are listed in Table 2.1.
TUM Hyperloop and Delft Hyperloop are the only two teams that are actively researching the implementation of a full-
scale hyperloop concept besides working on the competition. The only teams besides Delft Hyperloop and TUM Hyper-
loop that are expected to make an impact with their competition vehicle are MIT Hyperloop II, Swissloop and EPFLoop as
they have shown good results in past competitions.

Table 2.1: SpaceX Hyperloop Pod Competition 2019 finalists

Team Name University Location

Avishkar Hyperloop Indian Institute of Technology Madras India
Badgerloop University of Wisconsin – Madison USA
Delft Hyperloop Delft University of Technology Netherlands
EPFLoop EPFL – École Polytechnique Fédérale de Lausanne Switzerland
HYPED The University of Edinburgh Scotland
Hyperloop at Virginia Tech Virginia Polytechnic Institute and State University USA
Hyperloop UPV Universitat Politècnica de València Spain
Hyperlynx University of Colorado – Denver USA
HyperXite University of California – Irvine USA
Purdue University;
Midwest Hyperloop University of Cincinnati; USA
University of Illinois Urbana-Champaign
Massachusetts Institute of Technology;
MIT Hyperloop II USA
University of Texas at Austin
OneLoop University of California – Davis USA
Northeastern University;
Paradigm Hyperloop Memorial University of Newfoundland; USA
College of the North Atlantic
Queen’s Hyperloop Design Team Queen’s University Canada
SLOLoop California Polytechnic State University – San Luis Obispo USA
Swissloop ETH Zurich Switzerland
TUM Hyperloop Technical University of Munich Germany
UMD Loop University of Maryland USA
UNSW Hyperloop The University of New South Wales Australia
uWinLoop & SCCLoop University of Windsor; St. Clair College Canada
Washington Hyperloop University of Washington USA

2.2 Hyperloop Companies

Several companies are working on commercialising the hyperloop concept. An overview of the companies as of 2019 is
provided in Table 2.2. This table indicates the propulsion and levitation system that these companies are (expected) to use,
along with the number of passengers per pod. Unfortunately, this information is lacking or might be outdated for various
companies mainly due to two reasons. Either they do not want to share this information because of intellectual property
reasons, or they have not yet decided upon a definitive method. It can be concluded that various companies have different
ideas of a hyperloop system and convergence is not yet reached. The fact that several technologies are investigated might
actually be beneficial even though it will become difficult for standardisation. However, knowledge on these technologies
is only available within specific companies, due to intellectual property reasons. To accelerate hyperloop development, it
is important to share knowledge and results in order to work towards the best concepts.

3
Chapter 2. Current State Hyperloop Parties

Table 2.2: Commercial hyperloop companies in 2019

Organisation Name Location Propulsion Levitation Passengers per pod

DGWHyperloop India - - 44
Hardt Global Mobility Netherlands LSM EMS 58
HyperloopTT USA LSM EDS 28-40
Hyper Poland Poland - EMS -
TransPod Canada LIM - 40
Virgin Hyperloop One USA LIM/LSM EDS 12-80
Zeleros Spain Compressed air EMS 50

2.3 Non-profit Hyperloop Organisations

Besides commercial companies, two non-profit organisations are performing research towards the implementation of
hyperloop as shown in Table 2.3. It is noteworthy that Eurotube is planning to build a full-scale test facility to eventually
host a European hyperloop competition in the early 2020s.

Table 2.3: Non-profit hyperloop organisations in 2019

Organisation Name Location

Eurotube Switzerland
RWTH Aachen Germany

2.4 Websites
Several websites and platforms online exist that are contributing to hyperloop development.

Hyperloop Connected is an initiative of Delft Hyperloop, with the goal to create a platform with an overview of all Hyper-
loop related knowledge. At the time when the platform was launched, no comprehensive overview of who was working
on Hyperloop and what they were working on existed. Delft Hyperloop noticed this problem and decided to create Hy-
perloop Connected. Currently, research performed by Delft Hyperloop is shared via this site. Other parties such as several
companies and teams mentioned here before have shown interest in the platform and mentioned the importance of an
independent knowledge hub. Besides articles, the website currently contains a frequently asked questions page and an
interactive map showing everyone in the world who is contributing to hyperloop. SpaceX engineers mentioned that the
website is fascinating and are inspired by the enthusiasm. Everyone is able to contribute to Hyperloop Connected by shar-
ing their knowledge, research or vision for hyperloop. Several articles from contributors around the world have already
been posted as well. See www.HyperloopConnected.org.

IntheLOOP brings all the new information concerning the hyperloop industry into one place. It provides regular updates
from all companies and teams that are working on developing hyperloop and is therefore a good source of hyperloop
news. The site also started a ‘tech series’ in which more technology related articles are posted. See www.intheLOOP.news.

4
Chapter 2. Current State Hyperloop Parties

2.5 Test Tracks

Full-scale test tracks are going to be key to develop the technology required for a hyperloop system. As of 2019, several
companies and organisations already have a test track operational. Moreover, several organisations have presented plans
for future test tracks. An overview is given in Table 2.4.

Table 2.4: Operational and planned hyperloop test tracks

Organisation Name Location Test Track Length Tube Diameter Status

Hardt Hyperloop Delft, the Netherlands 30 m 3m Operational
Virgin Hyperloop One Nevada, USA 500 m 3.3 m Operational
SpaceX Los Angeles, USA 1.25 km 1.8 m Operational
HyperloopTT Toulouse, France 320 m 4m Under construction
Transpod Droux, France 3 km 2m Planned (2020)
HyperloopTT Toulouse, France 1 km 4m Planned
Hyper Poland Poland 500 m full-scale Planned
Eurotube Valais, Switzerland 3 km unknown Planned
Hardt Hyperloop The Netherlands unknown unknown Planned

5
3.Levitation
This chapter focuses on levitation, which is the phenomenon of a floating object due to a lift force counteracting the grav-
ity on that object. Maglev has proved the advantages of levitating passenger transport. No physical contact between the
vehicle and the guideway results in less noise production, less maintenance and less energy consumption compared to
high-speed rail [Wang & Zong, 2010]. Therefore, Delft Hyperloop believes that levitation is the best option for a hyperloop
pod.

In this chapter, the four potential levitation mechanisms known by Delft Hyperloop will be explained. The advantages
and disadvantages are listed and the mechanism will be assessed on design criteria. After assessing all four mechanisms,
a trade-off is be made to determine the most promising levitation system. Eventually, a recommendation is given on
which levitation method seems optimal for a hyperloop. The flowchart (see Figure 3.1) represents the structure of this
chapter. The top-level system requirements and the environmental boundaries are input values for this chapter. Within
these limits, the levitation systems will be evaluated in Sections 3.1, 3.2 and 3.3. The final section of this chapter gives a
recommendation.

Figure 3.1: Flowchart representing the structure of this chapter.

3.1 Subsystem Requirements

The top-level system analysis and the environmental analysis (see Appendix A) determine the boundaries for the levita-
tion subsystem analysis. The top-level system requirement is that the pod shall be able to levitate during operation. This
results in the primary subsystem requirement: the subsystem shall be able to lift the pod for all attainable operating ve-
locities. Secondary system requirements are listed below.

The subsystem shall:

• provide a sufficient air gap to prevent physical contact with the guideway.
• be able to lift the expected loaded pod mass with an additional safety margin.
• be mass producible, it should therefore not consist of rare or extremely expensive materials.
• provide air gap data.
• be fully autonomous.
• be able to perform switches on speeds above 600 km/h.
• provide a guidance function such that the vehicle is stable and controllable at all times.
• be able to function within the limits given by the environmental analysis (see Appendix A).

The design criteria on which the potential subsystems will be assessed are listed below, ordered by priority. More infor-
mation about these design criteria is given in Appendix B.

• Reliability
• Power consumption
• Stability

6
Chapter 3. Levitation

• Cost
• Safety
• Complexity

Another important feature to take into account is the TRL1 of technology. It is a measure of how far a technology is into
development, a high TRL means that the technology is further in the development and closer to implementation. For
more information, see Appendix D.

3.2 Analysis of Potential Levitation Mechanisms

In this section, several levitation methods are elaborated further upon. For each levitation mechanism, the current situ-
ation including TRL is stated and the advantages and disadvantages are given. Succeeding, the levitation mechanism is
briefly summarised and will be assessed according to the design criteria. Eventually, after all levitation mechanisms are
discussed, a trade-off will be made in Section 3.3. The mechanisms that will be discussed are respectively electromagnetic
suspension, electrodynamic suspension, air bearings and electrodynamic wheels. These are all known potential levitation
mechanisms by Delft Hyperloop. The structure of this section is depicted in Figure 3.2.

Figure 3.2: Flowchart of the levitation mechanism analysis

3.2.1 Electromagnetic Suspension (EMS)

The main characteristic of EMS is that it uses the attraction of electromagnets and ferromagnetic material in order to
levitate [Han & Kim, 2016]. Electromagnets are used to provide a variable attractive magnetic force. There are two options
for a guideway configuration: top- and bottom-EMS. Figure 3.3 shows the two potential guideway configurations for EMS,
known by Delft Hyperloop. The left (top-EMS) is a situation where the levitation of the pod is generated by electromagnets
on top of the pod. The pod hangs below the track. The right (bottom-EMS) requires arms around the track in order to make
use of the EMS levitation, this configuration is used in the Shanghai Transrapid.
1 Technology Readiness Level: a level indicating the development of a certain technology, originally invented at NASA [Mankins, 1995]

7
Chapter 3. Levitation

Figure 3.3: Schematic of two EMS guideway configurations. Left: top-EMS. Right: bottom-EMS, the traditional EMS Maglev guideway
(Transrapid).

Current Situation
EMS is currently operational in the Transrapid, Shanghai. It functions as a line connecting Shanghai Airport and Longyang
Road Station with a top speed of 431 km/h [Han & Kim, 2016]. Another example of an operating EMS system is the Linimo.
This is a low speed train that has an operating speed of 100 km/h. The operations have not always gone smoothly. Two
accidents with the Transrapid Maglev occurred in 2006, one of which resulted in 23 fatalities. The TRL depends on the
guideway, for the top-EMS the TRL is 5 or 6. For the bottom-EMS the TRL is 9, because it is already implemented into the
Shanghai Transrapid. Several companies and student teams are developing EMS (see Chapter 2).

Advantages and Disadvantages

The advantages and disadvantages are listed in Table 3.1.

8
Chapter 3. Levitation

Table 3.1: EMS - Advantages and Disadvantages

Advantages Disadvantages
• Low magnetic fields inside and outside the vehicle. • Inherently unstable, which means it requires con-
The fields have to be limited for medical electronic de- stant monitoring to maintain the distance between the
vice wearers and persons with pacemakers and other guideway and the pod (both for levitation and guid-
implanted electronic devices. For the Transrapid in ance) Han & Kim [2016].
Shanghai the stray magnetic fields were measured and
satisfy the requirements for urban transit [Han & Kim, • Electromagnets are positioned on-board, which re-
2016]. The induction outside the vehicle is 10 µT and quires batteries that increase the mass of the vehicle.
inside 25 µT [Kircher et al., 2018]. This is below the Moreover, the batteries have to be recharged during
magnetic field of the Earth (30-50 µT ) and within the operation.
limits imposed by the German regulation BImSchV 26 • Small gap distance (approximately 10 mm [Han & Kim,
[Kircher et al., 2018]. 2016]) resulting in the following: at speeds around 300
• Less guideway material required because it already m/s, a very fast control loop is necessary because the
works with two ferromagnetic beams, no complete pod should be able to provide stability. At high veloc-
guideway required. ities the irregularities in the guideway should be ob-
served very quickly. Bumps in the track could result in
• Some forms of EMS have already been certified, proven the pod colliding and potentially crashing. The follow-
by the operations of the Transrapid Maglevs. ing potential hazards should not be neglected when re-
searching EMS: load changes, unsteady aerodynamic
• Good reliability of system is proven by 10 years of forces, guideway induced vibrations and guideway
problem-free operation [Han & Kim, 2016]. misalignment [Han & Kim, 2016].
• Achieved speed of 430 km/h (Shanghai Transrapid)
[Han & Kim, 2016][Cassat & Bourquin, 2011].

• Energy efficient, a measure for this is the Lift-over-

Drag. The Transrapid reaches magnetic Lift-over-Drag
values above 500 resulting in a low energy consump-
tion for generating lift [Wang & Zong, 2010]. A Boeing
has a Lift-over-Drag of approximately 15. This means
that the hyperloop generates 500/15 ≈ 33 times lower
(magnetic) drag when lifting the same mass. Note that
the aerodynamic drag is not taken into account.

Design Criteria Assessment

• Reliability: EMS involves several active systems that increase the complexity of the overall system. Proof has to be
delivered that the system can be operated reliably in the hyperloop environment (vacuum and velocities twice to
triple the velocity of the Transrapid). Transrapid has proven safe operation but only within its own conditions.
• Power consumption: there are losses due to the electrical current through the electromagnet (I 2 · R-loss) but there
still is a Lift-over-Drag of over 500, therefore it has a low power consumption compared to other methods of levita-
tion.
• Stability: this can be achieved to a high order since all electromagnets are controllable separately.
• Cost: the cost of the guideway (only the levitation parts considered) will be comparable to that of Maglev. The
operational cost will be lower since the power consumption is lower than that of Maglev. Maglev is the most realistic
comparison that can be made at this point.
• Safety: the magnetic field generated by the EMS levitation is low because the field goes into the ferromagnetic guide-
way. However, in case of a crash the pod gets a large impact, significantly endangering the passenger on-board.
• Complexity: the pods are scalable in terms of length and weight. However, if there is a mass increase, the electro-
magnets require higher power or the gap to the guideway will be smaller. The system is not very complex and can
be made scalable and mass producible.
These characteristics will be used in a trade-off, summarised by Table 3.5.

3.2.2 Electrodynamic Suspension (EDS)

Electrodynamic suspension uses repulsive forces in order to levitate. These repulsive forces are often the result of a mag-
netic field moving relative to a conductor. The moving magnetic field is generated by the pod and the guideway functions

9
Chapter 3. Levitation

as conductor. According to Faraday’s and Lenz’s Laws a lift force is created and the pod is able to levitate. The options for
passive levitation that will be discussed in the paragraph are:

1. Passive levitation with superconducting magnets

2. Passive levitation with permanent magnets, Inductrack is one example of this type

Current Situation
Chuo Shinkansen L0 is one example of Maglev that uses superconductivity, it holds the speed record for Maglev trains2 .
Moreover, Virgin Hyperloop One has shared a version of a hyperloop vision with electrodynamic suspension, but they
are also interested in electromagnetic suspension. Hyperloop Transportation Technologies makes use of electrodynamic
suspension (Inductrack) as well. In the U.S., four major industrial teams (Foster Miller, Grumman, MIT and Bechtel)
were initiated to research the potential of, and eventually develop, Maglev trains. Concepts were worked out but the U.S.
government stopped funding for these programs after one and a half years [Lever, 1998][Rose et al., 2019]. The TRL on
EDS with superconductors is 8 or 9. The TRL of Inductrack is unknown.

Advantages and Disadvantages

Advantages and disadvantages of EDS are listed in Table 3.2.

Table 3.2: EDS - Advantages and Disadvantages

Advantages Disadvantages
• Highest maglev speed record of 603 km/h set by the • Necessary use of magnetic shielding due to strong
Chuo Shinkansen L0. fields (pacemakers, data storage media) outside the ve-
hicle. Because the magnetic field can reach 300 µT .
• Large gap between guideway and pod (at least 25 mm However, this is still under the ICNIRP guideline that
[Han & Kim, 2016]). The Chuo Shinkansen has an air put the limit on 2500 µT [Kircher et al., 2018]. Inside
gap of around 100 mm [Kircher et al., 2018]. the vehicle the magnetic field still passes the guidelines
• Guideway can be on the bottom of the tube, this en- with 100 µT without shielding [Kircher et al., 2018].
ables wheels as auxiliary suspension system at low • Vehicle must be wheeled to travel at low speed. In case
speeds. of superconductivity, the cryogenic system has to keep
• No additional on-board batteries have to be installed the pod levitating.
to facilitate the EDS levitation. • Eddy current losses resulting in additional drag. Addi-
• Energy required for the levitation is taken from the ki- tional drag involves an increased energy consumption.
netic energy of the pod. Therefore, no additional bat-
teries have to be on-board to have an operational pod.

• Passive and inherently stable, this means that no con-

stant monitoring and no feedback loop are required for
this system. It is already stable when the lift-off veloc-
ity is achieved.

Design Criteria Assessment

• Reliability: in general the reliability of an EDS system is high. Mainly because it does not rely on any power source (it
is completely passive, disregarding the cooling of the superconductors). There is a large gap height (25 mm) which
means that irregularities in the track will not be catastrophic.
• Power consumption: depends on the guideway, Inductrack has a Lift-over-Drag resulting in an energy consump-
tion similar to that of a train. A conducting guideway with copper coils results in too much magnetic drag and is
therefore not feasible. The power consumption for superconducting EDS is lower than permanent magnet EDS (e.g.
Inductrack) but it does require energy to cool the magnets.
• Stability: an active secondary suspension mechanism can guarantee stability at all times. However, the air gap
height is not and depends on the velocity.
• Cost: if Inductrack is going to be realised, guideway cost will be high, an indication cannot be given because the
information about Inductrack is not published regularly. However, the fact that the guideway costs are high does
2 Land Speed Records - https://en.wikipedia.org/wiki/Land_speed_record, consulted on 31 May, 2019

10
Chapter 3. Levitation

not imply that the total system is more expensive. It could mean that there have to be less safety measures with
respect to other levitation systems.
• Safety: magnetic shielding is required outside the vehicle. However, passenger are safe within the vehicle because
the magnetic field is not oriented in that direction. Moreover, the large air gap significantly decreases the probability
of a pod colliding with the guideway.
• Complexity: EDS works simple since there are no active systems required. Moreover, the pods are scalable in terms
of mass and length.

3.2.3 Air Bearings

The lift force is created by an air bearing, formed by an air compressor. Air bearings were proposed as the hyperloop
levitation system by Elon Musk in the Alpha Paper [Musk, 2013]. First, the current situation will be given. Then, the
advantages and disadvantages of this mechanism are listed in Table 3.3. Finally, air bearings as levitation mechanism will
be assessed on the design criteria.

Current Situation
Research has been performed on air bearings as a levitation method, for example with the TLRV (Track Levitated Research
Vehicle) [Fischer et al., 1974]. It had air-cushioned suspension and was designed to reach speeds up to 480 km/h. Unfortu-
nately, the Urban Mass Transportation Administration stopped funding for this project. Delft Hyperloop has no knowledge
of other hyperloop corporations or teams researching air bearings for a future hyperloop. However, the hyperloop team
of MIT has developed a vehicle with air bearings that will probably participate in the 2019 SpaceX Pod Competition. The
TRL of a complete system for air bearings in high-speed transport has not reached 9, the best guess is 6. The TLRV is not
considered in the TRL estimation.

Advantages and Disadvantages

Advantages and disadvantages of air bearings are listed in Table 3.3.

Table 3.3: Air Bearings - Advantages and Disadvantages

Advantages Disadvantages
• Extremely low drag [Musk, 2013]. • Fans, motors and hover-pads are heavy components
increasing the pod weight that has to be lifted.
• The generated magnetic fields are negligible, since no
electromagnetism is required for creating the levita- • Compressors can be noise producing, the noise-level is
tion force. type dependent.

• The compressor sucks additional flow from the front • Compressors require high-maintenance [Opgenoord
of the pod through the pod (decreasing the blockage et al., 2017].
ratio) [Opgenoord et al., 2017].
• On-board power is required to power the compressors
[Musk, 2013].

Design Criteria Assessment

• Reliability: the method has been implemented in several prototypes. For example, it was implemented in the TLRV
[Fischer et al., 1974]. However, funding for this project was stopped. The TLRV is too outdated (1974) to assume as
reliable reference project. A full-scale system is never tested in a near vacuum environment.
• Power consumption: expected to be higher than EMS or EDS due to the energy consumption of the compressors
and the losses in the required interfaces. No sources were found to support this claim.
• Stability: the stability can be controlled but the gap height is small [Musk, 2013].
• Cost: if made in mass-production, the air compressors can be made relatively cheap. However, operational cost are
high due to the energy consumption and the performance of high-maintenance.
• Safety: currently using air bearings is not very safe; transonic compressors at such low Reynolds numbers would
require more research and development. Moreover, it is expected to be difficult to control the motion and inflow of
air. This could result in small gap heights and a higher risk to crash. This claim was not based on solid sources.
• Complexity: high complexity since there are many moving components and interfaces. Moreover, the system is
subjective to high-maintenance and is not easy to mass-produce.

11
Chapter 3. Levitation

3.2.4 Electrodynamic Wheels (EDW)

Mechanical rotation of radial magnets (electrodynamic wheels) can provide guidance, lift and thrust forces. The rota-
tion of the wheels generates a varying magnetic field in a conducting guideway. The varying magnetic field creates Eddy
Currents that generate an opposing magnetic field, with a repulsive lift force as a result.

Current Situation
No companies are developing Electrodynamic Wheels. Some hyperloop student teams researched this levitation mecha-
nism for competition purposes, see Chapter 2. As far as Delft Hyperloop knows, EDWs have not yet been used for trans-
portation purposes. The TRL is presumed to be 4.

Advantages and Disadvantages

Advantages and disadvantages of Electrodynamic Wheels are listed in Table 3.4.

Table 3.4: EDW - Advantages and Disadvantages

Advantages Disadvantages
• Suspension and propulsion can be realised at the same • Wheels rotate at high rotations per minutes (RPM’s).
time [Bird & Lipo, 2003]. Magnetic fields induced in the direction of the passen-
ger compartment.
• Can ride on the electrodynamic wheels on low speeds.
• The wheels will have a relatively high mass [Bird &
• Peak thrust efficiency almost independent of operating Lipo, 2003]. Moreover, a higher mass results in a higher
speed [Bird & Lipo, 2019]. energy consumption.

• Requires active propulsion on pod, resulting in me-

chanical losses for rotating the wheels.

• Does not outperform other levitation systems in terms

of efficiency [Bird, 2007].

• The power supply is positioned on-board, this in-

volves a higher mass and necessity to charge induc-
tively while moving.

Design Criteria Assessment

• Reliability: many moving components; motors, wheels axles and electrodynamic wheels rotating at very high RPM’s
(over 50000 RPM [Bird, 2007]). Moreover, fast rotating magnetic fields which are also directed to the passenger
department. But also wheels rotating at very high RPM’s and transmission from the motor to the wheels. Therefore
the reliability of the system is low.
• Power consumption: compared to EDS and EMS [Bird, 2007], the power consumption will be high due to mechanical
losses, moving components and eddy current losses.
• Stability: expected to be sufficient because every wheel can be regulated independently. This enables the operating
system to control the movement of the wheels and ensure stability.
• Cost: cannot be estimated with the current knowledge. Expected to be higher than EDS and EMS.
• Safety: one has to deal with high rotating magnetic field. Therefore shielding is required for the passenger module.
• Complexity: high complexity since there are many moving components and interfaces. Also because of this, there is
high-maintenance.

3.3 Trade-off
The trade-off for the levitation mechanism is performed in Table 3.5. EMS is used as benchmark in this trade-off.
From the information that is available, it can be determined that EDW provides the least potential as levitation mech-
anism. Therefore Delft Hyperloop recommends not to use EDW as levitation mechanism for the future hyperloop. Air
bearings are complex, more research and development has to be performed for the use of air compressor systems at ve-
locities similar to those of the hyperloop. Moreover, the air gap is small, making it difficult to respond to variations in track
height in time. Therefore Delft Hyperloop recommends not to use air bearings for a hyperloop.

12
Chapter 3. Levitation

Table 3.5: Trade-off levitation mechanisms

Concept 0 Concept 1 Concept 2 Concept 3 Concept 4 Concept 5

EMS - bottom EMS - top EDS - Permanent EDS - Supercon- Air Bearings EDW
Magnets ducting
Magnets
Reliability 0 0 + + - --
Power consumption 0 0 - - - -
Stability 0 0 + + - -
Cost 0 + - - - -
Safety 0 - + + 0 -
Complexity 0 0 0 0 - -

EMS and EDS are very promising options for levitation. EDS costs more energy since there are eddy current losses present.
However, they could be reduced significantly by developing a suitable guideway (e.g. Inductrack) and therefore it is still
assumed to be a feasible levitation method. EMS also has large potential since it is energy efficient and the operator is able
to control the gap height. The focus of EDS research and development should be on the guideway configuration, working
towards a guideway that has a Lift-over-Drag ratio of over 200. Inductrack could be an option but there has not been any
published progress around Inductrack lately. The EMS development should be on the reliability of the system, especially
on high speeds (Delft Hyperloop envisions velocities over 1000 km/h for the hyperloop). Being able to control the system
at all times is a necessity with such a small air gap. Redundant systems should be installed to guarantee the levitation
of the pod in case of a power outage. The track has to be capable to charge on-board batteries to provide power for the
levitation.

3.4 Recommendation
An EMS system is very promising because of its low power consumption, this is definitely an advantage with respect to
EDS. However, if an EMS system is able to reach the operating speeds presumed by Delft Hyperloop, one has to verify that
the system’s small air gap is sufficient, as a crash should be prevented at all times. To prove the reliability of the EMS system
at, the system must be tested at high-speed of up to 1000 km/h. Moreover, an EMS system does depend on active control
systems for it to levitate. Therefore it is dependent on the power supply of the electromagnets which makes it inherently
more prone to failure than a passive levitation system such as EDS.

The recommendation of Delft Hyperloop for the levitation system is EDS. Primarily because Delft Hyperloop prioritises
safety, especially when developing a high-speed passenger transport mode. The relatively large air gap of EDS significantly
lowers the probability of crashing due to track irregularities or other effects on the movement of the pod. A remark on EDS
is that more research and development into the energy efficiency in combination with guideway configuration has to be
performed. This is necessary to increase the energy efficiency, as this is still a disadvantage of EDS that must be improved.

Post-chapter remarks are provided in Appendix C.

13
4.Propulsion
Propulsion is the subsystem that generates the movement of the pod. The two most promising propulsion mechanisms
for the hyperloop, will be introduced and explained. The advantages and disadvantages are listed and the mechanism will
be assessed on design criteria. After assessing all mechanisms, a trade-off will be made to determine the best propulsion
system. Eventually, a recommendation is given by Delft Hyperloop on which propulsion method to use for a hyperloop.
The flowchart (see Figure 4.1) represents the structure of this chapter. The top-level system requirements and the envi-
ronmental boundaries are input values for this chapter. Within these limits, the propulsion systems will be evaluated in
Sections 4.1, 4.2 and 4.3. The final section of this chapter gives a recommendation on which propulsion mechanism to
use.

Figure 4.1: Flowchart representing the structure of this chapter.

4.1 Subsystem Requirements

The propulsion mechanism has to be capable of accelerating the pod to the cruising speed and to maintain this speed.
Because the pod levitates, the propulsion has to function without any contact. A complete subsystem analysis of the
propulsion subsystem is given in this section. Two mechanisms will be introduced; Linear Induction Motor (LIM) and
Linear Synchronous Motor (LSM). In the next section they are discussed in more detail and a trade-off is made to deter-
mine the most feasible options. The subsystem requirements are determined based on the top-level system requirements
and an environmental analysis (Appendix A). The latter determines the limits of the environment in which the system
should be able to operate.
The vehicle will levitate in order to be as energy efficient as possible, as was discussed in Chapter 3. Therefore, no propul-
sion mechanisms can be used that operates on wheels. The Linear Induction Motor (LIM) and Linear Synchronous Motor
(LSM) produce a propulsion force electromagnetically and work in combination with levitation. The subsystem require-
ments for the propulsion are given by the list below.

The subsystem shall:

• be able to accelerate the vehicle to cruising speed (assumed to be 1080 km/h).

• be able to maintain cruising speed (assumed to be 1080 km/h).
• be able to reach high accelerations (compared to accelerations limited by passenger comfort) for goods transporta-
tion.
• have the ability to brake.
• transform kinetic energy to electrical energy while braking (regenerative braking).
• The magnetic fields in the passenger compartment should not exceed the values indicated by human safety in mag-
netic field guidelines [Kircher et al., 2018].
• gather data on temperatures, velocities, acceleration and power consumption [Kaye & Masada, 2004].
• be able to function within the limits given by the environmental analysis (see Appendix A).

The design criteria on which the potential subsystems will be assessed are listed below. More information about these
design criteria is given in Appendix B.

14
Chapter 4. Propulsion

• Reliability
• Power consumption
• Cost
• Safety
• Complexity

The TRL1 of the technology is important to take into account as well. For more information, see Appendix D.

4.2 Analysis of Potential Propulsion Mechanisms

In this section, two propulsion methods are elaborated further upon. For each mechanism, the current situation and TRL
is stated, after which the advantages and disadvantages are given. Succeeding, the propulsion mechanism is briefly sum-
marised and will be assessed according to the design criteria. With this information, a trade-off will be made in the next
section. The mechanisms that will be discussed are respectively Linear Induction Motors (LIM) and Linear Synchronous
Motors (LSM). The structure of this section is depicted in Figure 4.2.

Figure 4.2: Flowchart of the propulsion mechanism analysis.

4.2.1 Linear Induction Motor (LIM)

A Linear Induction Motor originates from a rotary motor and works in a similar way. A rotary motor consists of a stator and
a rotor. However, the rotor is cut open and laid out flat. The stator and rotor are respectively called primary and secondary
in the linear motor.

The function of the primary part is to generate a varying magnetic field across the air gap, this is called the stator. The
magnetic field induces an electromotive force in the secondary part of the LIM, a conductor. This electromotive force
creates eddy currents which interacts with the magnetic field of the primary, thereby creating a propulsion force.

Two variants of LIM exist: short primary (SP) and long primary (LP). The stator is always the primary and the conduc-
tor is always the secondary.

• Short primary: stator is on-board of the pod and the conductor is in the guideway.

• Long primary: stator is in the guideway and the conductor in on-board of the pod.

Current Situation
A few examples that use LIM for propulsion are listed below. Note that they all of them operate on relatively low speeds.
All examples use Short Primary LIM as propulsion mechanism.

1. HSST (Japan): High Speed Surface Transport, maximum speed of 100 km/h [Cassat & Bourquin, 2011] and it uses
electromagnetic suspension to levitate.
1 Technology Readiness Level: a level indicating the development of a certain technology, originally invented at NASA [Mankins, 1995]

15
Chapter 4. Propulsion

2. Korean Hyundai Rotem (South Korea), maximum speed of 110 km/h. The used technology is similar to the HSST
technology. [Cassat & Bourquin, 2011]
3. Changsha (China): maximum speed of 100 km/h [Yang et al., 2018].

The TRL of LIM is 9 because it has been researched, developed and implemented in multiple projects already.

Advantages and Disadvantages

The advantages and disadvantages of a Short Primary LIM are represented in Table 4.1.

Table 4.1: LIM Short Primary - Advantages and Disadvantages

Advantages Disadvantages
• Proven its operation for Maglev [Cassat & Bourquin, • A reduced energy efficiency because of end effects.
2011][Yang et al., 2018][Thornton, 1993]. Moreover, the energy efficiency is reduced because the
air-gap flux is inductively created. The leakage induc-
• Implemented in applications besides transportation tance for the LIM is inherently large and gets worse
(e.g. cranes, pumps, actuators, conveyor systems and with an increased air-gap.
accelerators). Because of its many applications, the
system has been researched many times which results • Heat generation in the short primary (pod).
in a higher reliability.
• Because of the large air gap of EDS, LIM is not efficient
• Relatively cheap guideway because no active parts and in combination with EDS.
power supply have to be installed (a conducting sheet
is sufficient).

• No gears required (no mechanical rotary to linear con-

verter).

The advantages and disadvantages of a Long Primary LIM are represented in Table 4.2.

Table 4.2: LIM Long Primary - Advantages and Disadvantages

Advantages Disadvantages
• Implemented in high speed Maglev. • Air-gap flux is inductively created, the leakage induc-
tance for the LIM is inherently large and gets worse
• No gears required (no mechanical rotary to linear con- with an increased air-gap, therefore it has a low energy
verter). efficiency.
• Relatively high acceleration and deceleration is achiev- • Increased guideway cost because all the active parts of
able since the active parts are in the guideway and the LIM have to be positioned in the guideway.
therefore no heavy batteries are required on-board, re-
ducing the mass to be accelerated. • Heat generation in the guideway.

• Able to reach high speeds and do this more efficient • Because of the large air gap of EDS, LIM is not efficient
than a SP LIM. in combination with EDS.

• Flexibility to respond to variable or uncertain demand,

number and size of pods can be adjusted without
problems [Kaye & Masada, 2004].

Design Criteria Assessment

• Reliability: Linear Induction Motors are being used in several applications with a high rate of success. However, the
LP LIM has not been in operation for high speed transportation.
• Power consumption: energy efficiency of the LIM is low because of end effects [Yang et al., 2018][Liu et al., 2015]. The
larger the database with information about LIM performance (e.g. data of LIM drives in steel-wheel transit systems),
the higher the reliability of the system will be [Kaye & Masada, 2004].

16
Chapter 4. Propulsion

• Cost: in case of an LP LIM, guideway cost will rise significantly because the guideway has to be active and installed
with coils. The guideway cost of an SP LIM are lower.
• Safety: the system has not been implemented in high speed passenger transport, therefore it should be tested thor-
oughly to guarantee the safety. The magnetic fields created by the LIM will not be dangerous for passengers located
in the passenger compartment. At low speeds, the HSST has proven to be safe.
• Complexity: electrical complexity of the LIM is high, maintenance has to be performed regularly.

These characteristics will be used in a trade-off, summarised in Table 4.4.

4.2.2 Linear Synchronous Motor (LSM)

The Linear Synchronous Motor is a propulsion mechanism in which the mechanical motion is synchronous with the
magnetic field. The thrust force is not created by an induced magnetic field, but the magnetic field is created by windings.
For this it is necessary to know exactly at what position the vehicle is and at what speed it is travelling.

Current Situation
The LSM propulsion has been implemented in multiple Maglev projects already. Examples are listed below:

• The German Transrapid (TR07) uses LSM propulsion in combination with EMS levitation. The top design speed is
482 km/h [Lever, 1998].
• Japanese MLX01 currently holds the world Maglev speed record (581 km/h) while using LSM for propulsion and
EDS with null-flux coils to provide lift and guidance.
• Bechtel, this project was started but eventually the US government stopped its funding [Lever, 1998], the design
cruising speed was 482 km/h.
• Grumman, this project was started but eventually the US government stopped its funding, the design cruising speed
was 482 km/h [Lever, 1998].

The TRL of LSM is 9 because it has been researched, developed and implemented in multiple projects already.

Advantages and Disadvantages

The advantages and disadvantages of a LSM are represented in Table 4.3.

Table 4.3: LSM - Advantages and Disadvantages

Advantages Disadvantages
• Already in operation in the Transrapid Maglev Cassat & • Higher cost compared to LIM because the LIM only re-
Bourquin [2011]. quires one active part (in the pod or in the guideway).

• The energy efficiency is high (0.8-0.9) compared to that • LSM lacks the flexibility to easily change with system
of a LIM. capacity and operational modes. Moreover, the system
needs to be designed for the highest expected demand.
• Ability to reach high speeds (e.g. the Japanese MLX: If the demand is lower than expected, the system turns
581 km/h). Almost all high-speed Maglev designs use out to be over-designed [Kaye & Masada, 2004].
a LSM for the propulsion [Han & Kim, 2016]).
• Requires data for the exact position of the on-board
• Energy does not have to be transferred to the pod since magnets to ensure that the vehicle is synchronous with
the primary can be in the track [Kaye & Masada, 2004]. the traveling wave generated by the stator winding in
• High accelerations can be reached, this could be useful the guideway [Kaye & Masada, 2004].
for transporting cargo. However, for passenger trans-
port, the acceleration is limited by safety requirements
and passenger comfort [Kaye & Masada, 2004].

Design Criteria Assessment

• Reliability: by placing various sensors the reliability enlarged. It is important to know the position and velocity of
the pod at all times to ensure the reliability [Kaye & Masada, 2004]. The reliability of active guideways and switches
must be established with tests.
• Power consumption: the energy efficiency on high speeds around 300 km/h is above 80% and can be even higher at
higher speeds.[Kaye & Masada, 2004].

17
Chapter 4. Propulsion

• Cost: high investment cost because the active guideway and power supply installation needs to be designed for the
highest expected demand. If the demand is lower than expected, the system turns out to be over-designed [Kaye &
Masada, 2004].
• Safety: the safety has not been implemented often in high speed passenger transport, therefore it should be tested
thoroughly to guarantee the safety. The magnetic fields created by the LSM will not be dangerous for passengers
located in the passenger compartment.
• Complexity: the system is not complex and has been built before. However, maintenance may be time-intensive. If
a segment of stator windings malfunctions, repair and re-qualification testing results in the whole track being out of
service [Kaye & Masada, 2004].

These characteristics will be used in a trade-off, summarised in Table 4.4.

4.3 Trade-off
In Table 4.4, the discussed propulsion methods are assessed according to the design criteria. The trade-off in Table 4.4

Table 4.4: Trade-off propulsion mechanisms

Concept 1 Concept 3 Concept 4

LSM LIM SP LIM LP
Power consumption 0 -- -
Cost 0 + -
Reliability 0 + +
Safety 0 0 0
Complexity 0 + -

compares the possible methods for the propulsion of the hyperloop with LSM as the baseline measurement. It can be
derived from this trade-off that a LIM with a long primary (active guideway) is the least optimal propulsion mechanism
for a hyperloop.

An LSM is the best option if the system is assessed from an energy efficiency point of view. Especially for high speed oper-
ations the LSM is preferred because of the low energy efficiency of LIM at high speeds. Moreover, the ability to transfer the
high electrical power to the pod is impractical at speeds of the hyperloop [Kaye & Masada, 2004]. However, the reliability
and safety of the LSM system is less than the LIM because the LIM has been put into practice many times before.

4.4 Recommendation
Two propulsion mechanisms were introduced as potential propulsion subsystems for the hyperloop. Namely, Linear Syn-
chronous Motors and Liniear Induction Motors. For the Linear Induction Motors two types were assessed: LIM with a
short primary and LIM with a long primary. The mechanisms were assessed according to the following design criteria:
power consumption, cost, reliability, safety and complexity. A trade-off was made to determine the best option.

If it is desired to develop a system that is energy efficient and is able to achieve high velocities (over 1000 km/h), one
should opt for LSM. Moreover, LIM is not compatible with EDS levitation and requires non-contact charging of the batter-
ies during the journey. However, the guideway costs and complexity of an LSM propulsion system are higher than those
of a LIM.

Since the envisioned hyperloop system of Delft Hyperloop must be energy efficient and requires speeds of approximately
1000 km/h, the LSM is the best option for the propulsion system of a hyperloop. Future research must prove the safety
and reliability of a LSM system at high-speeds of 1000 km/h. This should be achieved through extensive testing of the
system at a high-speed test facility. Only when safety and reliability are guaranteed, could LSM be a real option for the
hyperloop.

18
5.Pod Characteristics
To size the hyperloop pod, two different approaches can be used; the inside-out approach and the outside-in approach.
Using the outside in approach, the diameter of the tube is set fixed. From there, the pod can be sized based on the
constraints of the tube and the aerodynamic performance. However, this could lead to unrealistically small hyperloop
pods if the tube diameter is based on other requirements than pod size and is therefore not a recommended approach. In
this chapter, the inside-out approach is explained. The flowchart (see Figure 5.1) represents the structure of this report.
By starting with assuming the amount of passengers per pod, the pod can be sized and the tube follows afterwards.

Figure 5.1: Flowchart representing the structure of this chapter to size the pod and the tube

5.1 Situation Analysis

A hyperloop pod is best fitted with 40 to 60 passenger seats in order to both cope with the expected demand and to be
able to reach a frequency of 2 pods per minute during rush. To continue further analysis and designing, the number of
seats per pod has been assumed to 50 in the design by Delft Hyperloop. From this number, the pod can be sized based on
aircraft sizing guidelines. By analysing the aerodynamics, the tube can be sized afterwards.

5.1.1 Fuselage Sizing

A hyperloop pod is effectively a pressure vessel with similar characteristics to the fuselage of aircraft. Both travel at high
speeds, have to cope with pressure differences and most importantly, transport people. For the conceptual design of a
hyperloop pod, design guidelines for aircraft can be used due to these similarities. Similar to aircraft, a hyperloop pod
should be of cylindrical shape since this is the most efficient way to distribute the forces of the pressure difference.

The parameters required for the fuselage sizing process are taken from reference aircraft. This method is explained in
detail in Appendix E. The seat width is set to 22 i n or 56 cm, which is a compromise between 20 i n used for short range
economy seats and 24 i n used for mid range business class seats. In this way, seats in a hyperloop pod will offer a great
level of comfort while still being space efficient. Using the values as given in Table 5.1, the total outer diameter of the pod
becomes 2.70 m. The required length of the pod to accommodate 17 rows of seats, one toilet for passenger convenience
and four exits is 22 meters. To optimise the aerodynamic efficiency, an aerodynamic cone will be added to the front and
back adding another 10 meters.

5.1.2 Aerodynamic Analysis

A pressure of 0.1% atmosphere will normally result in 0.1% aerodynamic drag compared to the drag one would experience
at sea level. However, this only holds if the volume around a vehicle can be assumed to be infinitely large, which is a
perfectly valid assumption for aircraft. A hyperloop pod violates this assumption by having the walls of the tube closely
around the vehicle. In this way, a hyperloop pod somewhat resembles a supersonic wind tunnel with choked flow and
shock waves as a result.

19
Chapter 5. Pod Characteristics

Table 5.1: Values used for calculating the pod diameter

Parameter Value
Seats per row (n sa ) 3
Seat width (w seat ) 0.56 m
Number of aisles (n ai sl e ) 1
Armrest width (w ar mr est ) 0.05 m
Aisle width (w ai sl e ) 0.51 m
Side wall clearance (s cl ear ance ) 0.02 m
Structural depth 0.2 m

The aerodynamic drag is highly dependent on the blockage ratio (which equals to the ratio of the cross section of the pod
to the cross section of the tube). In Appendix E, an extensive aerodynamic analysis is presented to cover this topic. For the
input of the aerodynamic analysis, the operating pressure was set to 100 P a and the operating velocity was set to 300 m/s.

5.2 Recommendation
It can be concluded from the aerodynamic analysis that the aerodynamic drag ranges between 2 kN and 10 kN for reason-
able operating speeds (>500 km/h) and blockage ratios (0.6 - 0.8) at an internal tube pressure of 100 P a. The aerodynamic
drag for the configurations where the flow is not chocked is several orders of magnitude smaller than the aerodynamic
drag of the chocked flow. A blockage ratio between 0.6 and 0.8 is therefore recommended. To continue further analysis
and designing, the blockage ratio has been set to 0.7 in the design by Delft Hyperloop. More elaborate research is required
to determine the optimal blockage ratio.

Based on aircraft design guidelines, a pod diameter of 2.7 m would be recommended to fit three seats abreast. The cross-
section of a hyperloop pod is presented in Figure 5.2, based on the parameters of Table 5.1. A blockage ratio in the order of
0.7 would mean an inner tube diameter of 3.5 m. The pod length including aerodynamic noise and tail cones (to reduce
pressure drag) would become 30 meters.

This is a top-level design based on the system parameters analysed by Delft Hyperloop. The next step is to move to-
wards subsystem design using a systems engineering approach. The choices made in this chapter need to be evaluated
and possible other pod concepts need to be investigated. A more elaborate market and cost analysis can be conducted to
optimise the number of passengers per pod. Once converged to one concept, subsystem requirements need to be set-up
and the interface between all the subsystems of the pod need to be analysed. Safety and costs are two main criteria that
will be important for every subsystem. From here, each subsystem of the pod can be designed in detail. This includes
the pod structure, levitation and propulsion system, life support systems including environmental controls, power sub-
system and navigation and communication systems. Full-scale testing of these subsystems in a simulated environment
or test-track will be important in order to make progress in pod design.

Figure 5.2: Hyperloop pod cross section with three seats abreast.

20
6.Tube Characteristics
A hyperloop tube consists of many components. In this chapter only the main cylindrical tube and its pillars are con-
sidered. Other components are too detailed for this stage of development. Since physical infrastructure is needed for all
links, the hyperloop tubes are prominent in the total system.

This chapter explains the current situation for the hyperloop tube. First, the requirements and boundary conditions for
the tube are explored. After that, different materials are explained and a qualitative trade-off is made. Following this,
basic calculations for the tube thickness and heat expansion are done and the optimal pillar distance is shortly addressed.
Finally, a recommendation for the tube configuration is given, based on material choice, thickness, pillars and thermal
expansion.

6.1 Subsystem Requirements

A hyperloop tube has two main functions which it should fulfil, namely to confine the low pressure environment and to
protect the pods from all external conditions. These functions provide the low aerodynamic drag and the ability to operate
continuously in a safe manner. To maintain the low pressure environment, the tubes should be airtight and strong enough
to prevent puncture.

Furthermore, there are requirements from other subsystems which the tube should satisfy. Such as, the size of the pod and
the aerodynamics, which determine the size of the tube. The tube should be 3.5 metres in inner diameter, as was found in
Chapter 5. Also, the tube should be sufficiently stiff and strong to prevent excessive deformations.

6.2 Tube Material

The material choice for a hyperloop tube has a wide variety of consequences. On the one hand, the lowest price per kilo-
metre is desired, but on the other hand, producibility and transportability are important. Furthermore, the tube should
have all the desired structural properties. A qualitative trade-off is made based on the following criteria:

• Span suitability
• Cost
• Thermal expansion
• Workability
• Air tightness

A further explanation of the criteria is given in Section 6.2.2. A table containing the most important characteristics of the
materials can be found in Appendix F.

6.2.1 Material descriptions

In this subsection, different possible materials for a hyperloop tube are explored. These materials are either conventional
construction materials or promising materials from other industries.

Concrete
Concrete is a good all-round material. Using reinforcement bars, pre-stressing and other advanced techniques, robust
structures and desirable properties can be achieved. Concrete might not always be airtight, thus special coating is re-
quired. Concrete is ideal for compression, but concrete is significantly weaker under tension. For a hyperloop tube, this
property is not ideal.

Currently new techniques are under development by combining concrete with glass or carbon fibres. These developments
look promising as they greatly reduce the required thickness of the concrete while maintaining the same strength. In this
paper these technologies are not considered, since developments are in their early stages.

21
Chapter 6. Tube Characteristics

Steel
Steel is a material with a high strength to weight ratio for both tensile and compressive stresses. Strong structures and
large spans can be created with relatively little material (compared to concrete for example). Due to the characteristics of
steel, it is easily moulded into any shape while maintaining its structural strength and stiffness. A problem for steel is that
it is prone to corrosion.

Current new techniques are under development to make a steel tube with the same strength significantly lighter. This is
done by using innovative wall geometries (TRL 6). Structurally, these innovations are promising, however, the effects on
the aerodynamics, the producibility or the track have not been investigated yet.

Aluminium
Aluminium is lighter compared to steel, but has a lower stiffness. This means more aluminium is needed than steel for the
same deflections. The strength of the aluminium highly depends on the type of alloy.

Aluminium is often used in the aerospace industry, due to its low weight and high strength. However, weight optimisation
is of great importance for airplanes, but not for a hyperloop tube. Aluminium is more expensive than steel, what often
makes it less preferred. This is especially true for the hyperloop, where the tubes make up a large portion of the costs.

Acrylic
Acrylic is a transparent and strong material, which finds its use, amongst other things, for large (aquarium) tanks and
can be made in any shape. It is corrosion free and can withstand various weather effects. Acrylic is significantly more
expensive than steel or concrete, rising the same cost problem here as for aluminium. Acrylic is aesthetically appealing
and could be used for small portions of tube.

Fibre Reinforced Polymers

A wide variety of composite materials is eligible for a hyperloop tube. Composites can be formed into any shape and have
a high strength and a low weight. Different techniques and sandwich structures can be used to increase the stiffness of
the material while barely adding weight. Fibre reinforced polymers consist of a fibre net and resin, which together form a
strong material.

Fibre reinforced polymers are expensive and are mostly used in high-tech environments, for instance, in race cars and
airplanes. The same costs problem arises as for aluminium and acrylic.

New developments are made to reduce the costs and to speed up the production process. Currently, it is investigated to
what extent glass fibre composites are suitable for a hyperloop tube. Glass fibre composites are currently being developed
for offshore windmills (TRL 6). Experts on the subject have stated that glass fibre composites would be similar in costs
compared to steel tubes after it is fully developed. A problem that currently remains is the ability to recycle fibre reinforces
polymers. If the tubes cannot be recycled, the sustainable goal of the hyperloop is neglected.

6.2.2 Trade-off
In order to choose the best tube material, a trade off is made based on five qualitative characteristics of the materials.
These five qualitative characteristics are explained below in descending order of importance.

I Cost: Hyperloop tubes contribute a large portion to the total cost and economical feasibility of the system. Reducing
the costs of the tube therefore is beneficial for the success of the hyperloop. Costs is the strongest weighing factor,
as it determines the economic feasibility of the hyperloop.

II Span suitability: A span is the unsupported distance a between two points in a construction. Not all materials are
as suitable as others for the application in spans. For spans, the material should have a high strength (yield stress),
to prevent material failure in the middle of the span. Furthermore, the material should have a high stiffness (Young’s
Modulus), to minimise occurring deflections.

III Thermal expansion: All materials expand and contract due to thermal fluctuations. For large construction works,
this expansion and contraction can lead to significant size differences of the tube throughout the year.

IV Workability: Workability is a combination of producibility and transportability. It is important that big numbers of
tube can be produced in a short time and that transportation to the construction site is efficient. This reduces cost
and construction time.

22
Chapter 6. Tube Characteristics

V Air tightness: To maintain the low pressure environment inside the tube, air tightness is desired. Tubes that leak
will result in more pumping effort, higher drag and higher energy usage.

Table 6.1: Trade-off table for materials, concrete used as benchmark

Concrete Steel Aluminium Acrylic

Cost 0 0 - --
Span suitability 0 ++ + +
Thermal expansion 0 0 - --
Workability 0 0 0 0
Air tightness 0 + + +

For the material, steel was chosen as the best option due to Table 8.1. Since a European hyperloop network requires large
amounts of tube, material and cost reduction will significantly decrease construction time and costs. Steel has a high
strength to weight ratio and stiffness compared to concrete, so less material is needed for the same goal. Aluminium and
acrylic have higher prices than steel and would make the hyperloop system less to not economically feasible.

A problem with all materials is thermal expansion due to temperature fluctuations. This could create problems in areas
where large temperature fluctuations are typical. Of all trade-off materials, steel and concrete have comparable heat
expansion, whilst aluminium and acrylic have significantly higher heat expansion. In Section 6.5 this problem is addressed
shortly.

New developments are not opted for, as there are no clear conclusion and results. It is advised to keep an eye on the
current developments as promising ideas might emerge. Criteria are that new developments are both low in cost and easy
to produce and construct.

From this qualitative trade-off, steel is chosen as the best material.

6.3 Tube Thickness

A first order of magnitude value for the steel tube thickness is calculated in this section. This value is sufficient at this
stage of hyperloop development to be able to calculate other parameters of the hyperloop system. The tube thickness
determines how much steel is needed for the infrastructure and allows for a more accurate cost estimation.

To determine a value for the tube thickness it is assumed that the vacuum buckling criteria is governing. This assumption
is made, since designing for maximum deflections is not possible at this stage in hyperloop development. More research
is needed to determine the allowable deflections and the optimal pylon distance. Furthermore it is assumed that the tube
is an infinite long vacuum chamber, that the tube is thin-walled (thickness/diameter < 1/20) and that the tube holds a
pressure of 3 P a. For the complete calculation, see Appendix F.

It is found that the tube thickness should be 25 mm to withstand vacuum buckling including a safety factor of 1.5.

6.4 Pillar Distance

The pillar distance is the intermediate distance between two pillars on which the hyperloop tube is constructed. A large
pillar spacing requires the tube to have more strength and more stiffness to carry its own weight and to minimise deflec-
tions. On the other hand, a small pillar spacing allows for a more slender tube. However, for decreasing pillar spacing, the
pillars converge to a wall. A wall is not desirable, since this can lead to segregation on different levels, as experienced in
American cities. So there is a lower limit for the pillar spacing coming from social effects and an upper limit for the pillar
spacing coming from structural properties of the tube.

Calculating the optimal pillar spacing is out of the scope of this research, as it has no large influence on the complete
system. It is expected that the pillar distance will be around 30 metres.

23
Chapter 6. Tube Characteristics

6.5 Thermal Expansion

Steel hyperloop tubes constructed above-ground will experience thermal expansion. Due to thermal expansion and the
tube being ‘clamped’ in between two stations, great compressive forces in the tube may occur. This could result in lateral
buckling of the tube, which has occasionally already happened with train tracks. Buckling occurs due to compressive
stresses in the material, which should therefore be minimised.

A solution utilised in railway engineering is to heat the track before connecting the different rails. this results in the desired
stress in the track at the design temperature. In the winter, tensile stresses occur in the track and in the summer, little to
no compressive stress occur in the track. In this way, Buckling of the track is prevented.

For above-ground hyperloop tubes a solution to thermal expansion has to be thought of as well. The tubes are significantly
bigger than rails, so using the exact same techniques and machines is not possible. In the following list possible and
existing ways of countering thermal expansion are explored:

I Mechanical prestressing: Mechanical prestressing means that the material is mechanically elongated which in-
duces a stress. Under this stress the sections are connected. This means that there is a stress in the tube at installa-
tion temperature. For higher temperatures, the stress reduces and compressive stresses can occur. However, these
stresses are smaller due to the mechanical prestressing.

II Thermal prestressing: Thermal prestressing uses the same principle as mechanical prestressing, except that the
initial elongation is caused by heating up of the material.

III Gaps in between sections: This option allows for all sections to elongate since there is room in between the sections.
This space should be filled with elastic material (e.g. rubber) that ensures an airtight gap.

IV Expansion joints: Expansion joints can be put in between any connection to allow for elongation of the two adjacent
sections into this joint.

V Shaft at station: This is essentially the same as the expansion joints. However, in this case there is one expansion
joint located at a station. The entire elongation of the tube should roll into this shaft.

VI Stress in tube: This option makes use of the stiffness of the tube. The stress in the tube builds up and a sufficient
mass moment of inertia of the tube should prevent buckling.

In order to achieve an order of magnitude of the expansion of the tube, the maximum thermal elongation for a segment
of tube and a link can be seen in Table 6.2. These values (order of magnitude) are chosen to show that thermal expansion
for a link looks more crucial than it is. For an explanation on how these values are calculated, see Appendix F.

Table 6.2: Maximum elongation of the tube for different tube lengths

Tube length [m] Elongation [m]

400,000 m 316.8 m
30 m 0.024 m (2.4 cm)

Because of the large lengths of tube needed for the hyperloop, thermal expansion on the large scale is significant. Dealing
with thermal expansion at one point (option V) is therefore unfeasible. However, a small part of 30 metres elongates by
just 2.4 centimetres. It is feasible to accommodate for this, by using rubber spacers (III) and prestressing (I and II). Not all
elongation needs to be obviated, since some elongation could be converted to stresses in the tube.

24
Chapter 6. Tube Characteristics

6.6 Recommendation
The goal of this chapter is to give a recommendation for the right configuration of the tube. In this chapter, the tube
material, tube thickness, pillar distance and thermal expansion are discussed. To summarise this chapter, the following is
recommended:

• Tube material: Steel is recommended as the tube material. Steel has a high strength and stiffness and is suitable for
both compressive and tensile stresses, which makes it ideal for the creation of spans. However, it is useful to keep
an eye out for new technological developments happening on a large scale, as they might lower the cost or ease
construction. New materials or innovative techniques might develop in the coming years. Developments to watch
are steel weight optimisations and fibre reinforced polymers. If extra aesthetic value is needed, acrylic transparent
tubes might be suitable.

• Tube thickness: A tube thickness of 25 mm is advised as a first value. This tube should be able to hold up against
vacuum buckling with a safety factor of 1.5. More extensive calculations are needed for the final design for the tube.
These calculations are too detailed for the current state of hyperloop development and are out of the scope of this
research.

• Pillar distance: The pillar distances is a compromise between the structural properties of the tube on one side, and
social impact, like segregation, on the other side. Calculating the exact pillar distance is out of the scope of this
research and should be analysed further. However, an intermediate pillar distance of approximately 30 metres is
expected.

• Thermal expansion: Thermal expansion is not as big of a problem on a local level as thought of on a global level.
Instead it is a classic engineering problem all projects are facing, with a set of solutions. Rubber gaps in between
section, prestressing and minor stresses in the tube can be used to counter thermal expansion.

25
7.Vacuum Analysis
The low pressure environment reduces aerodynamic drag allowing for high speed transportation with relatively low energy
consumption compared to other modes of transportation. Since the vacuum environment is a key element in a hyperloop
system, this chapter provides an analysis on the major design choices with regard to the vacuum system. This chapter
covers the analysis of the vacuum installation required for pump-down, the vacuum installation required for maintaining
the vacuum level and the operational pressure level. Appendix G covers the physics and equations used for this analysis.

7.1 Vacuum Pumps

Several types of vacuum pumps for different applications exist. For the medium vacuum level required for hyperloop
tubes, the best option is a combination between roots pumps with backing pumps. Backing pumps must be used to
produce a rough vacuum (down to 3, 000 P a or 0.03 at m) from atmospheric pressure as the initial part of the pump down.
From here, the roots pump take over to bring the pressure to the required level. Roots pumps operate by pumping air
using a pair of rotary vanes and can pump small pressure differences between its inlet and outlet. The roots pumps will
be used to maintain the pressure level as well [Pfeiffer Vacuum GmbH, 2013].

7.2 Situation Analysis

The operating vacuum level is analysed to find an optimum between power required to maintain pressure and power
required to overcome the aerodynamic drag. Furthermore, the operations of the vacuum pumps can be split into pump
down from atmospheric pressure and maintaining operational pressure.

7.2.1 Vacuum Level

The aerodynamic drag scales with tube pressure, so when decreasing the pressure, the aerodynamic drag also decreases.
Lower aerodynamic drag means less propulsion required, thus saving energy. However, to maintain a lower pressure, the
roots pumps require more energy. Figure 7.1 represents an example of the energy consumption at a range of operating

1e12

1.4

1.2
Total energy use per day [J]

1.0

0.8

0.6

0.4

0.2

0.0
0 20 40 60 80 100 120 140

Pressure [Pa]

Figure 7.1: Energy consuption for vacuum pumps and overcoming aerodynamic drag for different operating pressures.

pressures for a frequency of 12 pods per hour. For this scenario, the optimum tube pressure would be 10 P a. This energy
consumption does not yet include the propulsion required to overcome magnetic drag. However, since magnetic drag is
only dependent on the velocity and not on the tube pressure, it does not influence the optimum pressure.

The pod frequency hugely influences the optimum pressure where the energy consumption is minimised. Table 7.1 pro-
vides an overview of the estimated optimum pressure for different pod frequencies. These pressures were established for a

26
Chapter 7. Vacuum Analysis

tube length of 500 km and an assumed trip length of 35 mi n. For this analysis, it was assumed that the power consumption
and pump speed of the vacuum pumps do not change with operating pressure for simplification purposes.

Pods/hr Optimum pressure [Pa]

1 35
2 25
4 18
6 14
12 10
60 5
120 3
Table 7.1: Optimum tube pressure for different pod frequencies

7.2.2 Pump down

For this analysis, it is assumed that one kilometre tube has to be pumped down to 1 mbar . With a tube diameter of 3.5
m the volume is 9,621 m 3 . If it has to be pumped down to 1 mbar within six hours, the total capacity of the vacuum
pump installation should be: S = 33, 230 m 3 /h (see Appendix G). The roots pumps can approximately pump 10,000 m 3 /h,
thus three pumps are required. However, the roots pumps can only pump small pressure differences. Therefore, backing
pumps are needed to reduce the pressure difference (backing pumps are capable of reducing the pressure from atmo-
spheric pressure down to 7 mbar ). The pressure is 7 times larger for the backing pump so the pump capacity can be 7
times smaller to distribute the same amount of air particles. The capacity of backing pumps are often lower, only a capac-
ity of 33230/7 = 4747 m 3 /h is required. This can be achieved with two backing pumps. For this case three roots pumps
and two backing pumps are required.
The power consumption is calculated using Equation G.3,for a pressure difference of 6 mbar , to be 5.5 kW . The backing
pumps require around 2.5 times as much as the power consumption of the roots pumps.

The pressure is not discrete, which means there is always a pressure gradient in the tube. For example, when an air-
lock opens, one could intuitively think that the column of air from the airlock moves as column through the tube. This is
probably not the case, the pressure gets distributed along the tube (note that the tube acts as huge vacuum pump as well).
When vacuum pumps are not evenly distributed, the pressure will also not be the same everywhere. The vacuum pumps
should be distributed as evenly as possible.

7.2.3 Maintaining Pressure

During the initial pump down, the pressure in the tubes is reduced to the required level for efficient operations. However,
the tube pressure will start to rise eventually when all the pumps are shut down after pump down. This will be due to
leakage of air through the steel of the tube and leakage at connections, welds, O-rings, airlocks. Leakage of air through a
material is considered to be out-gassing and is dependent on the material type and area. Thus several pumps are required
to operate continuously to counteract the leakage in order to maintain the operating pressure.

7.3 Recommendation
The optimal tube pressure is a balance between the energy required to maintain the vacuum level and the energy required
to overcome the aerodynamic drag. Therefore, the optimal tube pressure strongly depends on the pod frequency in the
tube. The more pods travel in the tube, the more efficiency is gained when the pressure is lower because all pods will expe-
rience a reduced aerodynamic drag and therefore a reduced power consumption. Therefore, the optimum tube pressure
will be variable based on the pod frequency and will be vary between 50 P a and 3 P a according to this analysis. The range
for the optimal operational pressure could however be different based on the variance in leakage, power-consumption of
the pumps and the frequency of pods.

The exact values of the tube pressure should be determined with a more in-depth analysis of the vacuum pumps which
should include an increase in power and decrease in pump speed at lower pressures. Extensive CFD research and a long
test track would be required to test the exact behaviour of the integrated components for the vacuum system. Out-gassing
and leakage can be determined and different pump configurations can be tested in order to optimise the vacuum system.
This research should also give more insights into shockwaves in the tube.

27
Chapter 7. Vacuum Analysis

To facilitate a vacuum environment it is recommended to install three roots pumps and two backing pumps together
in one unit. The amount of units required to pump down the entire tube from atmospheric pressure to the desired oper-
ating pressure heavily depends on the time this process is allowed to take. From the assumption that this must be able to
happen overnight, within a span of 6 hours, this would result in a vacuum system with 2 units every 5 km.

28
8.Pod Communication
In each transportation system, a secure line of communication with the moving vehicle is essential. A hyperloop system is
no exception. Only by a proper functioning communication system, the moving vehicle can be monitored and actuated.
This is essential for the functioning of the vehicle.

In this chapter, the communication for the hyperloop is explored. First, the theoretical background of communication
will be examined. This is followed by an overview of the communication systems in current methods of transportation.
Then, the situation with respect to communication in a hyperloop system will be analysed. In this section, all methods
for communication will be explained, followed by an overview of the system requirements. This will be concluded by a
recommendation.

8.1 Theoretical Background

The hyperloop communication system can learn from the communication used in high-speed transportation. To give an
idea of communication systems used for high-speed transportation, high-speed railways (HSRs) are looked upon. Rele-
vant parameters and challenges in HSRs are discussed.
Currently, most of the HSRs use GSM-R (Global System for Mobile Communications - Railway) as primary communica-
tion system. Limitations of GSM-R relate to interferences with other public networks, capacity and capability in terms of
transmission rate [He et al., 2016]. Therefore, HSRs are evolving into the next generation communication system: LTE-R
(Long Term Evolution - Railway). To improve security, quality of service and efficiency, LTE-R features [Ai et al., 2014]:

• Information transmission of control system via wireless communication with <50 ms delay.
• Real-time monitoring with <300 ms delay.
• Train multimedia dispatching.
• Railway emergency communications.
• Railway Internet of Things.

The use of LTE-R will face some issues. Implementation of LTE-R for HSRs already has some issues, which will be of greater
influence for a hyperloop due to the high speed. Some challenges regarding LTE-R are listed below:

• Working frequency bandwidth;

For coverage without interference, no other services can use the frequency range of hyperloop communication.
This can be a problem, since Hyperloop communication will not be the only user of the LTE bandwidth. For ex-
ample, mobile and fixed communication, radio positioning and satellite communication services all use different
frequencies. For a network involving multiple countries, this can be a problem.

• Doppler effects at high speeds;

High relative velocities lead to a shift of the received frequency, which is called the Doppler Effect. An increase in
velocity leads to an increase of the shifted frequency, based on Equation 8.1.

Ft
Fd = ·v (8.1)
c
Here F d is the Doppler frequency shift, F t the frequency of the signal. The doppler-effect implies a loss or misinter-
pretation of the data signal, due to the shifted frequency.

• Handovers between cells;

The cell range for LTE-R is 4-12 km, which means that a pod travelling with 1080 km/h would be in a cell for at most
40 seconds. When the inter-cells overlap is 1 km, the handover needs to occur in slightly more than 3 seconds. This
can be challenging. However, the handover success rate of LTE-R at 500 km/h is >99.9% [He et al., 2016].
Furthermore, the handover procedure is soft, which means that there is no data loss [He et al., 2016].

8.2 Situation Analysis

In this section, the situation related to communication in a hyperloop system is explored. This is followed by an overview
of various manners to collect data within the pod and tube. This is an essential factor for proper functioning communica-
tion. Finally, the required components for data communication are explored.

29
Chapter 8. Pod Communication

8.2.1 Introduction to Pod Communication

To guarantee safety and comfort while travelling with hyperloop, communication is an absolute necessity. Therefore, a
secure and reliable communication system has to be present at all times. There are multiple ways for communication in
a hyperloop system with each their own function. The pods need to communicate with each other, and a pod needs to
communicate with an external computing unit that processes data, makes decisions for the pod and actuates the pod.
Two types of communication will be discussed.

• The communication of the pods sensor data and commands to and from a centralised data processor
• The communication with information about the pods location between the pod and the tube

First, the communication of the pods sensor data and commands to and from a centralised data processor gets explained.
Then, the communication of the pods location between the pod and the tube will be discussed.

8.2.2 Data Collection

In the communication system, passenger safety is the top priority. The data required to ensure this safety gets collected
by a variety of sensors. These sensors are placed in two separate ways: at the pod and at the infrastructure. The data
collected at the pod contains information about the status and parameters of the pod. These are all observed by sensors
placed on the pod. This information needs to be communicated between the moving pods, the tube and a centralised data
processing station. The data collected at the infrastructure relates to location and velocity data of the pod, the propulsion
mechanisms of the infrastructures and ambient data (e.g. the air pressure and temperature in the tube).

8.2.3 Centralised Data processor

All computations and decisions are made at a centralised station, to minimise the number of decisions taken in the pod
or the tube. This station will be connected to the pod and the tube. This station is responsible for the operating of the
hyperloop pods. The station can actuate an emergency brake, determine the location and monitors the status of the pod.

8.3 Communication for Sensor Data and Commands

This segment of the pod communication system is divided in two sorts of communication: pod to pod (P2P) and pod to
infrastructure (P2I). P2I means the communication from the pod towards the centralised data processor. This is not a
direct connection; this traffic passes the tube. Therefore, P2I then can be divided further into pod to tube (P2T), and tube
to centralised processing unit (T2C). Since all of these options will be used for separate means and come with their own
challenges and requirements, P2P, P2T and T2C all require the use of different communication technologies.

P2I
To establish a secure line of communication between a moving pod and the outside world, some challenges occur. Since
the tube is expected to be made of steel, it is impossible to send and receive data directly with any existing method of
wireless data transfer. The combination of the material and thickness of the tube blocks the radio waves. Therefore, a new
method of communication can solve this problem. This challenge can also be overcome by making adjustments to the
tube. One possible way to ensure fast and reliable data transport between the tube and the outside world would be the
use of fibres that connects an antenna inside and outside the tube. Such a communication system is illustrated in Figure
8.1 and works as follows:

P2C
1. The pod sends the collected data with an antenna in the pod to the closest antenna on the surface at the tube.
2. The antenna receiving data in the tube sends the data via fibre to an antenna outside of the tube.
3. The antenna outside of the tube sends this data to the central data processor.

C2P
1. The central data processor sends this data to the antenna outside of the tube.
2. The antenna receiving data in the tube sends the data via a cable to an antenna within the tube.
3. The antenna within the tube sends this data back to the pod

T2C and C2T are the same protocol in reversed direction. The dataflow for both P2C and C2P is shown in Figure 8.2

30
Chapter 8. Pod Communication

Figure 8.1: Representation of the communication system

P2P
Since the propulsion mechanism and the tracking of speed and location (this will be elaborated further upon later in this
section) are located at the tube rather than on the pod, communication regarding safety between pods is completely out-
sourced to hardware integrated in the infrastructure. The functioning for the communication between pods relies only on
passenger comfort and wellbeing, and for operational purposes. Since there is no wall of steel between the pods them-
selves, communication between moving pods in the tube seems less complex than communication with the outside world.
The main counterarguments for completely wireless communication between pods, using radio waves, is the Doppler ef-
fect, as described in Equation 8.1. The Doppler effect occurs when the multiple pods move with a relative velocity not
equal to zero.

Doppler Effect One potential way to overcome the Doppler problem, is to eliminate the completely wireless communi-
cation and transfer data via a combination of wired and wireless communication. This can be achieved by sending data
from the antenna of the pod to the antenna within the tube. This antenna then transfers the data to a next antenna located
near the target pod via a cable within the tubes. The receiving antenna then forwards the data to the target pod.

31
Chapter 8. Pod Communication

Figure 8.2: Schematic view of the communication between the pod and the outside world

8.4 Location and Pod ID

To obtain the position of the moving pods, a system based on signalling can be implemented in the hyperloop system.
This is used in the European train network. The main principle is, that a vehicle moves past a marker that observes its
presence. This sensor then knows the time at which the vehicle passes, alongside its location. By placing multiple of these
sensors along the way, the velocity of the moving vehicle can be approximated. The accuracy of this approximation de-
pends on multiple factors, such as the number of sensors placed on the track, and their relative location. Also, the delay of
the sensors are crucial for accurately approximating the velocity. The determination of the location (and thus its speed) in
trains is a combination of the method explained above, and the use of odometers. An odometer measures the amount of
‘spins’ a wheel has made, and then calculates the travelled distance by multiplying the odometers value with the perimeter
of the wheel (under the assumption that the wheel does not slip). However, this is not possible for a hyperloop, since it
is propelled by magnetic propulsion, and uses no wheels on high speeds. Therefore, a new way needs to be invented, to
make sure the signalling-based method can be accurate and reliable enough to be used for location and velocity determi-
nation.

To use a signalling system for a hyperloop, photoelectric sensors will be placed onto the tube, and light strips will be
attached to the pods. By this way, the tube will detect when a pod moves by. This data will be send to the central pro-
cessing unit, that computes the velocity based on this received data. To perform this process successfully, it is essential
that a pod has a unique ID. Light strips located on the pod function as a bar code. This way, the photoelectric sensors can
separate every pod by their unique ID. More details on this photoelectric sensors are provided in Subsection 8.5.3. By this
process, the measured locations, and therefore the calculated velocity, get matched to the right pod. This communication
system is illustrated in 8.3.

Figure 8.3: Schematic view of the communication between the photoelectric sensor on the tube and the pod. Here, 1 represents the
photoelectric sensor, 3 represents the pods unique bar code and 2 the communication between this components. This line of
communication passes the ID located on 3 to sensor 1.

32
Chapter 8. Pod Communication

8.5 Required Hardware

In the previous section, several approaches for communication are discussed, both internal- and external communica-
tion. The conclusion remains that the majority of the communication will be externally. This means that the infrastructure
needs to support the pod, by placing communication hardware both inside and outside of the tube.

8.5.1 Wires in the Tube

The wires in the tube create a network where all antennas on the surface of the tube (both within the tube as outside
the tube) are connected. This way, all antennas can communicate, and by this, the pods can also communicate with
each other. The requirements for these wires relate to safety, durability, velocity and bandwidth. With respect to these
requirements, optical fibre seems a viable option.

Speed
Properties of optical fibre depends on the diameter and the number of fibres. In 2012, researchers from Necam and
Corning reported an ultra-large capacity transmission, with 1 petabit per second over 52.4 km of 12-core (light paths)
optical fibre [Peach, 2012].

Bandwidth
Current technology supports wavelengths up to 1675 nm. This equals 0.18 MHz. This is sufficient for the required data
communication. The current practical limits in bandwidth do not approach the theoretical limit of optical fibre.

Based on the speed and bandwidth, optical fibre is a viable option for the wired communication between the antennas in
the tube [M, 2009].

8.5.2 Internal Communication

To be connected with the outside world, the antenna on the pod needs to communicate with the antenna in the tube. This
antenna needs to be connected to the sensors on the pod. The measured data in the pod will be communicated to the an-
tenna via a wired connection. A possible way to establish this connection between the antenna in the pod and the sensors
at the pod, is to set up a Controller Area Network (CAN) bus. This is a vehicle bus standard to allow micro controllers and
devices such as sensors to communicate with each other without a host computer.

The bar codes placed on the pod need to be readable at all times. To ensure this readability, a trade-off should be made
between a static or hard coded bar code, and an adjustable bar code projected on a screen. The advantages for the vari-
able bar code are its property to add, delete or edit data. The disadvantages are related to its decreased reliability. For now,
a static bar code has been chosen, due to the fact that the pod only needs to show its presence, and its unique ID that
functions as a name tag.

8.5.3 External Communication

Inside the tube, antennas of the same kind as in the pod will be placed. This is essential, as the pod needs to communicate
with these antennas to reach the outside world. The number of antennas depends on the time it takes to send data from
the pod to the central data processor and back, and the distance the pod travels in this time. To read the bar code on a
pod, photoelectric sensors need to be installed in the tube. A photoelectric sensor (also known as diffuse sensors) is a
device that measures the distance to, or the presence or absence of an object. The photoelectric sensor uses infrared light
transmission and a photoelectric receiver. There are three types of photoelectric sensors: opposed (Through-Beam), retro-
reflective, and proximity-sensing (diffused). A trade-off with the advantages and disadvantages for each photoelectric
sensor is shown in Table 8.1.

Table 8.1: Trade-off for photoelectric sensors 1

Through-Beam Retro-reflective Diffuse

Accuracy + 0 -
Range + 0 -
Reliability + 0 -
Costs - 0 +
Set Up - 0 +

33
Chapter 8. Pod Communication

Through-Beam photoelectric sensors are the best functioning sensors. These consist of the most advanced hardware.
However, these are unpractical in terms of set-up time and costs, since these sensors require more labour and mate-
rial. Diffuse sensors perform less than Through Beam photoelectric sensors. Diffuse sensors are easier to set up. Retro-
reflective sensors lay in the middle of these two.

8.6 Recommendation
For the current status of pod communication, challenges have to be solved to implement a safe and reliable system. These
challenges lie within the communication from the pod to the outside world. Such a connection is necessary for the ex-
change of data with a central processing unit or help desk, or to establish an on-board internet connection. Another
potential problem lies within the collection of data from the pod at the infrastructure (external communication).

To overcome these challenges and to decrease errors and increase effectiveness, the technologies described in the sec-
tion should be developed further. Optical fibre is a technology with great potential. However, the type of fibre depends on
the size and speed at which data needs to be communicated. This should be researched in a later phase in the pod com-
munication. Wireless communication between antennas with radio waves still remains one of the most effective ways for
wireless data transportation.

This whole section is based on the technologies that are currently available. There is a high probability, that future devel-
opments in technology will erase the need for technologies described in this chapter. An example is the communication
from pod to the outside world. There is a possibility that within the coming years, a new communication protocol that
allows data travel through steel gets invented, such as 5g 2 . However, as these new technologies are still in development,
and since there is uncertainty about their performance, it is useful to further develop current technologies.

1 Photoelectric Sensors Selection Guide - https://cdn.automationdirect.com/static/specs/peselection.pdf, consulted on 31 May, 2019

2 Qualcomm - https://www.qualcomm.com/invention/5g/what-is-5g, consulted on 31 May, 2019

34
9.Artificial Intelligence Applications
Designing, building and operating a transportation system is a complex matter, and the complexity keeps increasing as
the demand and the size of the transportation system will continue to grow. This phenomenon has been happening since
the arrival of the first computers, when people got access to methods to process more data than ever before. Computers
surpassed humans ability to make complex calculations a long time ago. This has lead to an exponential growth in tech-
nologies. Traditional means of transportation are making use of computational power for years now. Now, in the recent
years, a new development within the field of computer science is evolving; Artificial Intelligence (AI).
Artificial Intelligence can be defined as follows:

Machines that mimic "cognitive" functions that humans associate with other human minds, such as "learning" and "prob-
lem solving"[Russell & Norvig, 1995].

AI is applied in a great variety of business cases and is promising for the hyperloop as well. AI can be of great contribution
to the designing, building and operating of a hyperloop transportation system. In this chapter, the possible application of
Artificial Intelligence in a hyperloop system are explored. First, the current applications of AI in transportation are exam-
ined. Then, this view of AI is projected on the hyperloop development, and a recommendation to apply AI in designing
and operating a hyperloop system is given.

9.1 AI in Current Public Transportation

AI already gets applied in transportation. AI gets applied in a great variety of transportation industries, both in the de-
velopment and the operating processes. These industries include the automobile industry, the aviation industry and the
railway industry.

Aviation
The AI systems used in the aviation ranges from failure detection to flight route optimisation. Since a numerous amount
of ways in which AI gets applied in the aviation industry, only a few examples are explored:
• Anomaly Detection in airline safety.
• Aviation turbulence detection.
• Route optimisation.
These subjects are elaborated further upon Section 9.3.

Railway
Similar to the aviation industry, the decision making in the railway industry gets dominated by AI algorithms. It can be
applied in the same fields as in aviation. One difference is the route optimisation. In aviation, the direction that an airplane
travels in depends on many variables. The direction of the train depends on other factors; the scheduling, the capacity of
stations and the passenger demand. This course optimisation results in the determination of the timing of a train, and it
does not alter the direction since it is a fixed route.

9.2 Requirements for the Applications of AI

To successfully apply Artificial Intelligence, there are two requirements:
1. Ensuring sufficient computing power
2. Collecting sufficient data
The first requirement is relatively easy. A party in charge of the deployment of the AI should own a computing station
operating with sufficient computing power, or rent cloud computing power [M. Dikaiakos et al., 2009]. The second re-
quirement is the biggest challenge in most AI applications. To successfully learn a specific task, AI algorithms need to
learn from many given examples.

Therefore, to make fully use of the benefits of AI, a hyperloop system needs to start collecting data from the first oper-
ational vehicle. Data collection is a process where data gets measured and saved. Both the pods and the infrastructure
should be supplied with sensors, measuring as much data as possible. Depending on the specific task the AI has to fulfil,
different data is necessary. For example, for predicting mean time to failure (MTTF) (9.3.7) for the pressure vessel of the
pod, the pressure and temperature inside and outside the pod have to be measured, since this is related to the vessels
characteristics.

35
Chapter 9. Artificial Intelligence Applications

9.3 Areas to Apply AI in Hyperloop

As described in section 9.1, AI can be applied in a great variety of areas. In this section, the application of AI is divided in
two main areas:

• Designing and building a hyperloop

• Operating and maintaining a hyperloop

Designing and Building

When designing and building a hyperloop system, AI can be used as a tool for:

• Define the network.

• Define station capacity.

Operating and Maintaining

When operating and maintaining a hyperloop system, AI can be used as a tool to:

• Incident detection.
• Ensuring on-board safety.
• Security and staff.
• Timetables and scheduling.
• Mean time to failure.

9.3.1 Define the Network

Determining the best locations for stations, and the links between these nodes is a complex process with multiple vari-
ables. An AI model should iterate on potential configurations, and be based on the demand per area, the demand for
flight- and train tickets, and the costs per connection. Such an AI model can provide an approximation of a network,
potentially much more suitable than a network provided by humans. The process of finding an optimal path between a
variety of multiple separately located nodes is called Path Finding. For Path Finding, usually Dijkstra’s algorithm is used,
as in this example where an optimal train route through a city is found[M. Agarana et al., 2016].

9.3.2 Define Station Capacity

For the size and capacity of the station, a similar model can be used. This model requires different data, and should be
trained in a different way. By feeding the network data about the number of people and the times they visit, alongside with
the paths they will most likely travel, an approximation for the structure of the station can be made.

9.3.3 Incident Detection

Incidents happen unexpectedly. The reason for this phenomenon is the lack of indicators for this incident. However, these
indicators probably did exist, they were just not observed since they are subtle signs. When every aspect of a hyperloop
pods and tubes is monitored, and everything gets measured, such indicators could be detected and reveal information
about the potential incidents. This is where AI can be applied. The subfield of AI responsible for these type of tasks is
called Anomaly Detection. By creating an abstraction of all available data, the model compares every new datapoint with
a predicted value. If this new datapoint differs from the predicted value (with a standard deviation included), the point of
interest seems like an outlier. This can than be checked, either by human or by AI, to see whether this outlier indicates a
potential incident. This same strategy can be used for safety management, mean time till failure (9.3.7) and more.

9.3.4 Ensuring On-board Safety

Emergencies effecting passenger safety is a possible option. It could be due to sudden sickness, or violence between pas-
sengers. To ensure safety during the transport, measurements should be implemented to avoid passenger emergency.
One convenient way to achieve this, would be the placement of an agent in the pod. This, however, has some downsides.
A human agent needs to be paid and connected to the outside world. In cases of communication loss, this is a problem.

An alternative for the human agent would be the guidance of an artificial intelligent assistant. Employing an AI agent
rather than a human agent solves the problems relating the costs and the constant connection to the internet, as men-
tioned above. The costs for an AI agent will be lower than the amount that the human agent would be payed. Also, an AI
agent does not require such a connection, since all the required knowledge can run locally on its software system.

This AI assistant could manifest in a robot, a computer-graphic generated ‘person’ on a screen, or a mere voice on a
telephone. The AI needs to meet the following standards:

36
Chapter 9. Artificial Intelligence Applications

• The AI interacts humanly.

• The AI runs locally.
• The AI reads input given by the user.

These requirements are explained further upon this section.

The AI Needs to Interact Humanly: To give the passengers the feeling they are helped and understood, it is important
to create an AI that feels and interacts as close to a human as possible. People will feel more comfortable when they are
helped by a human voice rather than a robotic voice.

The AI Needs to run Locally: To ensure the assistant will work at any time, it is important that it will always be connected
to a source of power, and a source of information, since the assistant will also be used in situations where power or access
to the internet is limited. The power can come from an emergency battery located in the pod. Therefore, the entire
operating system for the AI needs to run locally, so that it will still function even without an active internet connection.

The AI Needs to be Able to Read Input Given by the User: It would be difficult and inefficient to sense every possible
situation, based on sensors connected to the AI assistant. To overcome this challenge, the user should be able to give input
(e.g. with a tablet or with voice commands).

9.3.5 Security and Staff

At hyperloop stations, there should be a high level of security, since hyperloop transportation will be continental. However,
the security for hyperloop stations should be faster than security at airports. One way to achieve this, is the use of AI
systems. This can be applied to multiple security-related areas:

• Checking identification.
• Luggage security check.
• Using AI agents as staff (such as Pepper 1 ).

These security-related areas are elaborated further upon this section.

Checking Identification: The scanning of identification can be automated. Some airports (such as Schiphol Airport 2 )
already use this technology. Scanning identification with a computer saves time and money. It is unclear whether these
systems are more accurate than human agents. However, it is more time and cost effective. The recognising of instances
on images, also known as Pattern Recognition, is a subfield within the Artificial Intelligence. A recommendation for AI
techniques that can be used for this purpose are Convolutional Neural Network [Saha, 2018].

Luggage Security Check: The x-ray scans of the luggage should recognise forbidden goods that can not enter the ve-
hicle. Today, the x-ray images are still analysed by human employees. This is already changing; the technique to use
Artificial Intelligence for this purpose is gaining popularity, and it will be a matter of years before the first airports will
deploy these techniques. At hyperloop stations, these techniques should be applied, since it is faster, more accurate and
more cost-efficient than human employees. Similar to checking identification, the AI models will most likely be based on
convolutional neural networks.

Using AI Agents as Staff: The functioning staff at the station could consists of AI agents (robots) designed to interact and
help the human passengers. The current status of these robots is already employable. However, they are still less advanced
than their human operators. This gap will decrease in the coming years, as the functioning hardware and software of
these agents are currently evolving. Development in computer science and Artificial Intelligence is an exponential growth
process. This growth is described as Moores Law, that states that the number of transistors at dense integrated circuits
increase at 100% approximately every two years [Moore, 1965]. Since computing power is one of the factors that determine
the performance of an AI system, an increase in computing power enhances the performance of an AI system.

9.3.6 Timetables and Scheduling

Determining a schedule with the timetable for the departure and arrival of vehicles in a station is a complex problem,
since many different variables are involved. In train systems, this scheduling makes use of AI applications. For example,
the Dutch Railways, has the scheduling of train departures and arrivals fully outsourced to AI systems.
1 Pepper - https://www.humanizing.com, consulted on May 31, 2019
2 Schiphol security - https://www.schiphol.nl/en/security-check/, consulted on May 31, 2019

37
Chapter 9. Artificial Intelligence Applications

9.3.7 Mean Time To Failure

Another possible application of AI in hyperloop systems is the determining of mean time to failure (MTTF). This technique
is widely used in a great variety of industries. For example, in the aviaton industry, car industry and train industry, but also
for maintenance of machineries and other production tools. MTTF is the prediction of the time until a (sub)system fails
to function. This can be used to replace or maintain parts, to enhance their lifespan and to avoid accidents.

9.4 Recommendation
Within the field of Artificial Intelligence, there is a great variety of techniques with each their own application, advantages
and disadvantages. The type of AI techniques depends fully on the specific problem characteristics and the available data.
These applications of AI are not specific for hyperloop; it can be applied in any other transportation system.

In a hyperloop system, multiple areas to apply Artificial Intelligence are present, which all require data. To gain full ad-
vantage of the AI technologies, ways to capture and collect data should be implemented in hyperloop design.

In this chapter, various ways to implement AI in the designing and operating of hyperloop are examined. Especially for the
operating and optimising of hyperloop, AI can play a crucial role. For this purpose, the most important recommendation
is related to the collecting of data. Sensors should be placed on the pod and at the tube. The type of sensors depends on
the specific areas where AI can be applied. These areas include:

• Incident detection.
• Ensuring on-board safety.
• Security and staff.
• Timetables and scheduling.
• Mean time to failure.

If the hyperloop system meets the requirements for the application of AI, based on computing power and data, then these
processes could contribute greatly to the feasibility of a hyperloop system. It is therefore recommended to closely follow
the development of Artificial Intelligence.

38
10.Cost Estimation
The total cost of the hyperloop system is an important factor that determines if the hyperloop can be realised in the fu-
ture. Similar to other infrastructure projects, the investment costs for a hyperloop system are very high. For example, the
Betuwe route, a 160 kilometre long freight railway between Rotterdam and Germany, cost over €29 million per kilometre.
Moreover, the metro connection between the North and South of Amsterdam cost almost €320 million per kilometre.

The costs for a hyperloop system can be divided into two categories: investment costs and operational costs. Infrastruc-
ture, stations and pods are taken into account for the investment costs. The operational costs include costs for mainte-
nance, personnel and energy use. This chapter will only focus on the investment costs. First, an overview of the total
investment cost will be presented, followed by three sections covering the infrastructure, station and pod costs respec-
tively. The final section of this chapter provides a recommendation.

10.1 Overview
The investment costs cover the costs required for the realisation of the system. These cost have to be made upfront, after
which the basis of the system is realised. The investments cost are characterised as sunk investment cost (incurred and
cannot be recover post investment). For the hyperloop system, the investment cost cover the cost for the infrastructure,
stations and pods. Table 10.1 presents an overview of the estimated investment costs which are elaborated upon in the
following sections.

Table 10.1: Investment cost of a hyperloop system

Part Cost
Infrastructure costs per km
• above-ground (two-way) € 37,923,655
• underground (two-way) € 60,577,955
• high-speed switch € 27,934,298
Station € 700,000,000
Pod € 8,300,000

10.2 Infrastructure Costs

The total infrastructure costs are determined by the tube costs, land costs and certification costs. For underground infras-
tructure, tunnel costs are included. In Table 10.2 a breakdown of the estimated infrastructure cost is provided. Each item
is explained in the following sections.

Table 10.2: Infrastructure costs overview for a two-way link per kilometre (high-speed switches not included)

Part Above-ground costs Underground costs

Tube € 32,018,174 € 29,018,174
Tunnel €- € 25,430,000
Land € 5,530,000 € 5,530,000
Certification 1% of infrastructure costs 1% of infrastructure costs
Total € 37,923,655 € 60,577,955

10.2.1 Tube Cost

The breakdown of the tube cost is presented in Table 10.3. Each of the cost aspects is elaborated on in the sections below.
The total cost given here are for above-ground tubes. The estimated costs for the above-ground tubes will be €32,018,174.
For underground tubes, the concrete pillars, solar panels and an aesthetic cover are not required. This reduces the tube
costs to €29,018,174. Construction costs for all the components are covered by the costs for man hours in Table 10.3.

39
Chapter 10. Cost Estimation

.
Table 10.3: Tube costs overview for a two-way link per kilometre (high-speed switches not included)

Part Costs
Concrete pillars € 200,000
Steel tubes € 8,631,524
Inductrack € 918,000
LSM propulsion € 12,900,900
Guidance € 200,000
Solar panels € 800,000
Aesthetic cover € 2,000,000
Vacuum pumps € 160,000
Additional and communication € 1,207,750
Man hours € 5,000,000
Total € 32,018,174

Concrete Pillar Cost

The concrete pillars are the supports of the infrastructure. The pillars should be strong enough to support the weight of
the tube. It is necessary to begin with a proper foundation for the pillars and place them deep in the ground. The following
assumptions were made for the concrete pillars:

• Foundation piles are negligible (20 m x 0.2 m cylinder, around €750 per pile1 ).
• Concrete price of €250 per m 3 [Ocean Concrete, 2019].
• Height of 5 m, this value is taken to place the tube at a height for safety based on engineering judgement.
• Cross-sectional area of 4 m 2 , based on reference infrastructure projects.
• Support pillar interval of 30 m, based on the Alpha Paper [Musk, 2013] and several reference projects.

With an interval of 30 m, 33.3 pillars are required per kilometre. Each pillar has a volume of 24 m 3 , resulting in a total
concrete cost of €200,000. This price is for a pillar supporting a two-way tube.

Steel Cost
The tube is completely made of steel. The following assumptions were made to determine the total steel cost per kilometre
tube:

• No stiffeners required, since the tube design has redundancies.

• Tube diameter of 3.5 m.
• Tube thickness of 25 mm.
• Steel density of 7,860 kg /m 3 [Cutnell & Johnson, 2005].
• Steel price of €2/kg .

The tube diameter is 3.5 m and has a thickness of 25 mm. Therefore, 2,157,881 kg of steel is used for a one-way tube.
Multiplying this by the price per kilogram for steel, a single tube will cost €4,315,762 per kilometre. For two-way tubes, the
price will be €8,631,524 per kilometre.

Inductrack Cost
There is currently no information available regarding the costs of Inductrack. This is currently estimated by Delft Hyper-
loop to cost €918,000 per kilometre based on the price for an aluminium sheet, but should be investigated more thor-
oughly.

LSM Propulsion and Guidance Cost

The propulsion is necessary to generate the movement of the pod. A Linear Synchronous Motor (LSM) will be used for
both propulsion and braking. The following assumption has been made:

• LSM will be installed along the entire length of the tube, in order to reach a safety exit at all times.
1 Cost Heiwerken - https://www.offerteadviseur.nl/categorie/bouw/verbouwing/kosten-heiwerken/, consulted on 31 May, 2019

40
Chapter 10. Cost Estimation

In Lever [1998], the costs of guideway propulsion (LSM) are calculated for four Maglev concepts. The average cost is as-
sumed to be a good indication for the hyperloop guideway propulsion. This contains guideway magnetics (€3,603,961),
wayside control and communication (€870,106), power distribution (€1,044,126) and converter/inverter (€932,256). Re-
sulting in a total cost of €6,450,450 per kilometre for the propulsion. For two-way tubes this is €12,900,900, in which the
guidance is partially included. An additional guidance cost of €200,000 is taken into account for the required guidance
coils.

Solar Panels and Aesthetic Cover Cost

When tubes go above-ground, solar panels will be installed to generate energy and aesthetic covers to reduce the land-
scape pollution. Solar panels are not crucial for the functioning of the hyperloop system. However, solar panels make the
system more self-sufficient and reliable. The following assumptions have been made:

• Mono crystalline solar panels will be used.

• Solar panel cost is €200 per square metre based on commercially available solar panels.
• 4 m 2 of solar panels per metre of bidirectional tube.
• An aesthetic cover of €2,000,000 per kilometre.

The above-ground tubes could be installed with 4,000 m 2 of solar panels, resulting in a total cost of €800,000 per kilometre
for a two-way tube. The aesthetic cover is estimated to be €2,000,000 per kilometre for a two-way tube. The cost for solar
panels and aesthetic cover add up to €2,800,000 per kilometre of tube. This includes additional electrical components and
wiring.

Vacuum Pump Cost

The cost estimate for vacuum pumps and required installation is €80,000 per kilometre. The following assumptions have
been made for this estimation:

• Desired tube pressure ranges between 3 and 50 P a.

• Pump-down should take 6 hours maximum.
• Pump efficiency is assumed to be 100%.
• Water vapor is negligible.
• Outgassing independent of tube pressure.
• Power required independent of tube pressure.
• Pump 500 km down to the desired tube pressure.
• Cost of 1 pump is €10,000, based on commercially available industrial vacuum pumps.
• Installation, ventilation, housing, pump tubing, maintenance are assumed to have a cost similar to 3 times the total
price of the vacuum pumps.

According to these assumptions, about 1.2 roots pumps and 0.8 backing pumps are required per kilometre. This results
in a total cost of €20,000 per kilometre. This amount increases by a factor four because of the installation, ventilation,
housing. The total cost for the vacuum system is estimated to be 80,000 per kilometre per tube.

Additional Cost
Additional cost for cables and lightning amount to €10,000 per kilometre [Lever, 1998]. Emergency exits, excluding air-
locks, are estimated to be €100,000. This includes ventilation, first aid kits, emergency staircase and ways to exits [Eden-
baum et al., 2015]. De-pressurisation units are expected to be about €1,000,000. The sealing is expected to be about 1% of
the total tube cost. The total additional cost for two-way tubes are therefore €1,207,740.

41
Chapter 10. Cost Estimation

High-speed Switches Cost

In the envisioned network switches are required. This analysis is not part of the cost breakdown of the infrastructure cost
but provides a overview of the expected costs of a high-speed switch. The high-speed switch only includes the section
where the pod will switch from one track to another. The following assumptions are made for the high-speed switches:

• A switch requires a tube with a two-way tube width.

• Total amount of steel and aluminium doubles on the places where the switch is situated, because almost two parallel
tubes are present.
• Time it takes to switch is 10 seconds, such that the traffic flow remains uninterrupted.
• Thickness of steel tube doubles because the tube width has increased.

The assumption that the pod needs 10 seconds to make the switch result in a total switch length of 3 km, since the cruising
speed is 300 m/s. The total amount of steel needed for a switch of 3 kilometre with a thickness of 50 mm is 2,850 m 3 . This
results in a steel cost of €44,733,813 for a 3 kilometre switch. Moreover, the man hour cost is expected to increase for the
high-speed switches due to the higher complexity. Track costs will also increase. The total high-speed switch cost come
down to €27,934,298 per kilometre (including the tube).

10.2.2 Tunnel Costs

For underground infrastructure, it is assumed that both tubes will be placed in a single tunnel. The following assumptions
were made to determine the tunnel cost:

• Tunnel material is concrete.

• Tunnel diameter of 8.5 m.

According to the British Tunnel Society, based on 21 tunnels in European Union countries, the tunnel cost ranges from
£18 to £39 million (2010 pounds) per kilometre for tunnels with a similar diameter [HM Treasury and Infrastructure UK,
2010]. Tunnel costs are predominantly determined by diameter, however an increased tunnel length decreases the unit
costs. On average, a 12-km long tunnel costs per kilometre half of the costs of a 6-km long tunnel. As hyperloop tunnels
are expected to be very large, the minimum cost range for 8.5 m diameter is taken. In 2019 euros, this comes down to
€25.43 million per kilometre. This includes portals and emergency shafts.

For verification purposes, a new high-speed rail in the United Kingdom is examined by studying the tunnel cost esti-
mate and breakdown. For a pair of two 7 km 8.8 m internal diameter spaced tunnels, £33,07 million of civil engineering
costs are calculated per kilometre [UK Government, 2015]. For a single tunnel in 2019 euros, this is €22.21 million. This
is in the same order of magnitude as the 21 EU tunnel projects that is used. Therefore, tunnel costs for an underground
hyperloop is estimated to be €25.43 million per kilometre.

10.2.3 Land Costs

According to Campos & De Rus [2009], planning and land costs can take up a significant amount of infrastructure projects
budgets. For high-speed railways, this often represents between 5% and 10% of the total investment amount. These costs
include feasibility studies (both technical and economic), technical design, land acquisition and others such as legal and
administrative fees, licenses and permits. The following assumptions were made to determine the land costs:

• Tube outer diameter of 3.55 m based on an inner diameter of 3.5 m and a tube thickness of 25 mm.

For 45 high-speed rail projects researched by Campos & De Rus [2009], the total investment cost was €17.5 million per
kilometre on average (2005 euros), ranging from €6 million to €45 million per kilometre. Planning and land costs were on
average €2.15 million per kilometre, with a maximum of €5.53 million per kilometre (2019 euros). As the hyperloop will
have high infrastructure costs and will cross multiple dense areas, this maximum amount for ground costs is used.

For verification purposes, a new high-speed rail in the United Kingdom is used for comparison. For the planned high-
speed rail from London to West-Midlands (Phase One), land and property cost estimates are made by the UK Government
[2016]. This trajectory includes tunnels and both urban and rural areas, which is similar to hyperloop infrastructure. In to-
tal, 70 square kilometre of land and property will be bought for an estimate of £3.3 billion. This includes space for shunting
operations and maintenance, which is also needed for a hyperloop system. The length of the trajectory is 225 km, which
comes down to £14.64 million or €16.25 million per kilometre. However, as the trajectory mainly has four tracks, approx-
imately 50-80m in width is required to place the tracks [UK Government, 2016]. For two hyperloop tubes with an outer
diameter of 3.55 m, it is expected that approximately 15 m in width is needed as land area. With a linear cost decrease,
€4.875 million per kilometre is needed. This is in range with the previous estimate. Therefore, land costs for a hyperloop
will be estimated to be €5.53 million per kilometre, both above-ground and underground.

42
Chapter 10. Cost Estimation

10.2.4 Certification Costs

To implement a hyperloop, the system must meet certain safety standards. For existing modes of transportation, this is
regulated via certification of the system, which is done by specialised companies. This certification has to be done for
both the infrastructure and the pods, based on regulation. However, no regulation for hyperloop exists yet, and therefore
expectations based on railway systems are made. For railways, certification consists of four subjects:

• European regulation regarding interoperability (Notified Body)

• European regulation regarding safety (Assessment Body)
• European regulation regarding maintenance
• National regulation (Designated Body)

The regulations for a hyperloop system will be different, but these subjects are expected to be the same. Various railway
standards can be used, such as the life cycle model for the RAMS (Reliability, Availability, Maintainability and Safety).
However, these standards need to be adapted to future hyperloop regulations.

The costs for certification of a new train is in the range of several million euros. According to industry experts, for in-
frastructure projects, 1% of total investment cost is used as a guideline for certification costs in general. Therefore, this
number is also used to estimate the certification costs of a hyperloop.

10.3 Station Cost

To estimate of the cost of a hyperloop station, rail stations are used as reference. By comparing existing stations with the
predicted hyperloop station design, a cost estimate is made. To estimate the cost of a hyperloop station, it is important
to have a general idea of the layout of a hyperloop station. The layout will be mostly determined by the expected pod
frequency and the size of the pods. The pod size is estimated at a length of 30 metres, as explained in Chapter 5, and the
pod frequency is assumed to be 2 per minute.

To determine the station size, knowledge about the amount of platforms is required. With the assumption that the tran-
sition time of a hyperloop pod within a station is 6 minutes, 12 platforms are required. The assumption for the transition
time is based on the time it takes to move the pod from the platform to the airlock and the time it takes passengers to
embark and disembark the vehicle. In the hyperloop system design the embarking will be made as fluent as possible.

Table 10.4: Station space needed for main categories

Segment Area
Escalators and platforms 6,600 m 2
Platform-to-airlock area 7,920 m 2
Security 1,000 m 2
Shops -

In Table 10.4 the expected areas for the main components of a hyperloop system are listed. All values are indications and
are based on back of the envelope calculations and common sense. The escalators and platforms area are based on and
escalator length of 15 m and a platform length of 40 m. The width of this area is taken to be 120 m.

For the platform-to-airlock area, the width of 120 m is taken again and the expected length is 132 m. This area is ex-
pected to take the shape of a triangle. Security is based on the maximum pod frequency of 2 pods per minute. With a pod
size of 50 passengers, this results in an average interarrival time of 0.6 seconds. With new security systems, it is expected
that passengers will spend around 15 seconds at security gate. To prevent congestion, 25 security gates are needed. Se-
curity gate dimensions are estimated to be 10x4 metre, resulting in a total security area of 1,000 m 2 . The remaining space
can be filled with shops and the area needs to be determined on demand.

Since the hyperloop pods will be shorter than trains, it is possible to use the space of a regular rail platform more ef-
ficiently. The squared size needed to house the 12 hyperloop platforms is therefore expected to be similar to the size
needed for the 5 large platforms used in in rail stations. With this estimate the hyperloop station size is comparable to
the Berlin rail station. This station cost €700 million to build. The construction of the station had a lot of problems and
therefore this estimate for station cost is high. However, to make a conservative estimation and to cover the extra cost for

43
Chapter 10. Cost Estimation

a hyperloop station (e.g. airlocks and tracks), this value is kept for calculation of the station cost.

As additional reference for the station cost, the Delft rail region project is considered [NRC, 2016]. This project com-
prised the whole Delft rail region in the centre of Delft. In this project not only a station was build, but also tubes and
roads were constructed, overpasses were demolished, channels were dug and bike garages were build. This entire project
cost around €790 million. In reference to the estimate for the hyperloop station cost, it can be concluded that €700 million
is a decent estimate.

10.4 Pod Costs

As a hyperloop pod can be compared to a Maglev train, the costs of a Maglev train have been investigated. A Maglev train
with a capacity of 90 seats costs €12.5-15 million [Goeverden et al., 2018]. As the structure of a hyperloop pod needs to be
stronger due to the pressure level in the tubes (which is the main difference between a hyperloop pod and a Maglev train),
the higher estimate of 15 million is taken. When calculating the costs per seat, the Delft Hyperloop design with 50 seats
will cost €8.3 million. This is a very rough estimate but can be used in further analysis.

10.5 Recommendation
With the costs for the hyperloop infrastructure ranging from approximately €38 to €61 million per kilometre, the infras-
tructure costs are significant. Even though this is not incomparable to other infrastructure projects, research in ways to
reduce the costs for infrastructure could increase the economic feasibility of a hyperloop system. Furthermore, the costs
presented in this chapter are approached for a conceptual link. Therefore factors such as link trajectory and geographic
factors are not considered. For further research into the cost of the hyperloop infrastructure these factors should be in-
cluded.

44
11.European Hyperloop Network
The key features of a hyperloop transportation system are low energy use, high speed travel and the ease of a turn-up-
and-go system. It is because of this unique combination that the hyperloop system is generally considered to be the fifth
mode of transportation. An efficient (high-speed) rail network within Europe is already well established and therefor there
is no need for a hyperloop to compete with trains on smaller distances. It is assumed that a hyperloop system is a direct
competitor of short-haul (up to three hour) airline flights. Both airlines and airports have indicated in conversations with
Delft Hyperloop that they cannot handle the expected growth in aviation the upcoming decades. A properly designed
hyperloop network is therefore able to take over a significant amount of demand for short-haul flights within Europe, and
provide a valuable addition to the European transportation system.

This chapter presents a European hyperloop network designed by Delft Hyperloop. This network is optimised to make
efficient use of the unique characteristics of a hyperloop system. First, the database and methodology used to deter-
mine this network are covered. The method used to determine the design of the European hyperloop network is visually
presented in a flowchart in Figure 11.1. This is followed by a section covering the results of the method presenting the
actual network. Afterwards, the societal impact of a European hyperloop network is described. Finally, a section with a
recommendation for the possible European network is presented.

Figure 11.1: Flowchart representing the method used to determine the network design

11.1 Methodology for Network Design

The data that is used for this network design problem is gathered from the air travel database [Eurostat, 2017b] of Eurocon-
trol provided by the European Commission. Detailed air passenger transport data is available for the 34 largest countries
in Europe (excluding Russia) including all the 28 current members of the European Union. This database contains the
passenger numbers travelling between every commercial airport in the respective country and all its destinations. Differ-
ent metrics to address passenger numbers exist in the database. The total passengers boarded (‘PAS_BRD’) is used in this
analysis. This data was filtered to only include intra-European flights. Since the network will consists of connected cities
rather than airports, the airports in the database were converted to their respective cities. Several cities within Europe
contain multiple airports, for which all the passenger data of these airports were summed together. Finally, the coordi-
nates of every city centre were added to the database. The minimum distance between two points on a globe is found by
using the great circle distance between these points. Based on the coordinates of every city, the minimum distance for
every origin-destination pair (referred to as OD-pair) was determined. In this way, the actual passenger numbers between
over 3,000 OD-pairs were retrieved, combined with the minimum distance between each OD-pair.

The data used origins from the year 2017. It is assumed that it will take at least 15 years before a hyperloop system is
fully developed and ready to be implemented on a large scale. Therefore, a hyperloop network needs to be designed for
the future based on future demand, instead of the demand in 2017. Eurocontrol has presented several growth scenarios
for air traffic between 2017 and 2040 [Eurostat, 2017a]. Scenarios range from most optimistic to least optimistic, with the
scenario ‘regulation’ presented as most realistic. This scenario provides expected growth factors per country with a Euro-
pean annual average of 1.6%, resulting in an increase of 40% in 2040. To determine the growth for the demand of every

45
Chapter 11. European Hyperloop Network

single OD-pair, the growth factors of the country of origin and destination are averaged.

The goal of an infrastructure network is to connect cities. The most optimal way is to connect two cities directly. However,
this is both highly inefficient and completely infeasible, since this requires far too much infrastructure. The power of an
efficient network lies in connecting different cities with only a few links. It is assumed that about 50% of a hyperloop net-
work needs to be installed underground for reasons including space restrictions and horizon pollution. This brings the
average cost of one-kilometre tube to about 50 million euro for bidirectional infrastructure as explained in Chapter 10. The
shortest distance between any two points on the globe is defined by the great circle distance. However, it is never possible
to travel completely straight because of the restrictions provided by existing infrastructure and urban areas. The distance
cars travel between two cities is roughly 1.3 times the great circle distance. This turns out to be a good estimation for all
cross-border car travel distances in Europe. A hyperloop system will be largely installed next to existing infrastructure
where possible. If straight connections above-ground are not feasible, underground tubes will be used. The total distance
of a hyperloop link between two cities is thus assumed to be 1.2 times the great circle distance.

To determine the shortest route over any network design, a shortest-path algorithm was implemented. In this way, the
shortest path between two nodes over any generic network can be found. A penalty function for longer distances between
OD-pairs was implemented using this algorithm. The percentage of total demand for each OD-pair was set equal to the
ratio of shortest path distance between two cities and the actual distance travelled over the network. It should be noted
that even direct links are 1.2 times longer than their respective great circle distance resulting in a maximum demand for a
link of 1/1.2 or about 83% of current air passengers. A different travel time results in a different demand which is already
taken into account by the change in travelled distance.

11.2 Results
Using the model described above, the network was heuristically optimised for maximum passenger throughput with min-
imum cost per passenger kilometre to break-even. This was achieved iteratively by manually changing, adding and re-
moving links. During this process, the performance of the network and of every individual link was determined at every
iteration by calculating the following key performance indicators:

• Total number of passengers using the network.

• Total amount of required investment for the infrastructure.
• Price per passenger kilometre.
• Number of passengers travelling over each individual link.

While changing the network structure, it was assured that the OD-pairs with the most demand were included in an ef-
ficient manner. Finally, socio-economical, demo-graphical and political issues were taken into account by respecting
political differences between different countries. This resulted in the network design visualised in Figure 11.2 containing
48 stations with 51 links.

An overview of the links within the network is provided in Appendix H which contains all the 51 (bidirectional) links of the
network alongside with the link distance and the expected amount of passengers travelling over each link. This passenger
number is not the demand for direct travel between the destination connected by the link, but rather the total passengers
that use that link in the network to get from their origin to their final destination. With these links, a total of 672 OD-pairs
are created carrying a total of 308.5 million passengers per year. This equals two-thirds of all the passengers travelling
between these 48 cities and one-third of the total intra-European air passengers in 2040.

This network requires 19.700 km of bidirectional tube to be installed. This results in a total investment cost close to one
trillion euro. To get an indication of the cost of a ticket, it is assumed that this investment has a payback period of 25 years
without any government subsidies. Dividing the investment costs by the total passenger kilometres travelled results in a
price of €0.074 per passenger kilometre travelled. However, this would not be a completely fair way of pricing tickets. Even
though London is closer to Amsterdam than Paris is, passengers travelling from Amsterdam to London need to travel via
Paris and travel therefore more kilometres in a hyperloop than the passengers from Amsterdam to Paris. Another way of
getting an indication of the ticket price is to use the actual great circle distance between every OD-pair that was travelled
instead of the total distance travelled over the network. This would result in a price of €0.116 per passenger kilometre.
In comparison, train fares in the Netherlands were €0.17 per passenger kilometre [Gleave, 2016] and full service airline
fares were €0.08 per passenger kilometre [McKinsey & Company, 2017]. It should be noted that this hyperloop ticket price
is just an indication and no final conclusion in terms of pricing of a hyperloop ticket. Ticket prices can be reduced by
government funding, increased demand and reducing infrastructure costs.

46
Chapter 11. European Hyperloop Network

Figure 11.2: Cities in the European Hyperloop Network

11.3 Societal Impact of a Hyperloop Network

Many parties across the globe are dedicated to realising a hyperloop system because of the positive benefits it could bring
to society. In addition to passenger transportation, the main arguments for a hyperloop system focus on the its positive
environmental impact. By providing an efficient and sustainable alternative for high-speed transportation, hyperloop can
transform the transportation industry to help meet the sustainability goals for the future as stated in the Paris Agreement.
A hyperloop network can help reach these goals by replacing the short-haul air transportation, and thereby reducing the
dependency on fossil powered transportation.

However, the advantages of hyperloop stretch further than just replacing fossil dependent short-haul air travel. Due to
the easy accessibility and high operational speed of a hyperloop, it provides European society with better connectivity.
This connectivity provides citizens with an increased scope for living. For example, with travel times around 30 minutes
from Amsterdam to Paris, commuting for work over these distances becomes a possibility. Studies also suggest that new
infrastructure leads to positive welfare effects [Knaap & Oosterhaven, 2011][Chen et al., 2016].

Even though the proposed hyperloop network has mainly positive impacts on society, some side-effects might arise. For
instance, in the proposed network by Delft Hyperloop only the largest cities are included. This design would lead to the
exclusion of cities that are not included in the network. Furthermore, by only providing such limited destinations, the
expectation is that citizens will concentrate around these central nodes and thereby increasing urbanisation. This will
have implications for the first- and last mile transportation system. However, as airports cope with the same side-effects,
these are not expected to be major issues.

When aiming to maximise the societal benefits of a hyperloop system, the risks of region exclusion and the population
concentrating around hyperloop stations need to be taken into account. These side-effects can be reduced by optimis-
ing inter-modal connectivity or providing more destinations along a link. However, these solutions result in a decreased
financial and geographical feasibility of a hyperloop system and therefore do not counteract these side-effects by default,
and should be investigated further.

47
Chapter 11. European Hyperloop Network

11.4 Recommendation
This network design provides a good structure for a European hyperloop system by connecting major cities. It is expected
that this network is able to take over two-thirds of all expected short-haul flights between the connected cities, while min-
imising the required investment costs. Each link individually should be designed in more detail to work towards actual
implementation of such a network.

It should be noted that the future is always uncertain and this analysis is based on real data of 2017 with a realistic growth
scenario determined by Eurocontrol. Furthermore, high-speed switches are a key factor to make the operations of this
network feasible but are currently not developed yet. The exact implementation of intersections has not been considered
either. Next to that, the cost of the high-speed switches were not considered in the presented network analyses and should
be included future research.

By connecting cities and allowing for easy high-speed connections, new demand will most definitely be created. It is
currently not a reasonable option to go out for dinner in Paris while living in Amsterdam. Likewise, living in Frankfurt and
working in Berlin is unfeasible nowadays. A hyperloop could and will change human behaviour in terms of transportation,
a phenomenon which is very difficult to predict and is therefore not taken into account in this analysis. This phenomenon
will likely increase the demand and therefore the profitability of the network. Furthermore, it can be expected that a Eu-
ropean hyperloop network has positive effects on welfare by increasing the connectivity between cities.

In order to proceed with the implementation of a European hyperloop network, additional market research is required
in order to make better estimations on expected demand. This will change the network design and especially the ex-
pected cost per travelled kilometre. Furthermore, links need to be evaluated and designed in more detail on individual
basis. This includes placement of the tubes in respect to existing infrastructure and urban areas both above- and under-
ground, placement of the stations with respect to the cities and potentially increasing connectivity by adding intermediate
stations. High-speed switches need to be developed in order to guarantee point-to-point connections. Moreover, the op-
erations of a hyperloop system are not yet determined. Decisions need to be made regarding time schedules, frequency
of routes and intermediate stops which will influence the design of the infrastructure. Finally, it is important that the
implementation of the first link and or first connections are determined.

48
12.Safety Analysis
This chapter discusses the safety of a hyperloop system on a top-level, in order to determine the largest safety risks. First,
the importance of safety in transportation is highlighted. Afterwards, the method, scope and top-level system description
of the safety analysis are given. A Hazard Analysis is conducted to come up with the most important risks per subsystem,
and the overall largest risks are described afterwards. Hazard mitigation methods are described for each risk, in order
increase the safety of the system. The complete Hazard Analysis can be found in Appendix I. The largest risks lead to
a conceptual Safe Haven design: a new method to guarantee safety whilst minimising the investment cost. Finally, a
recommendation is given on how to mitigate hazards and to pinpoint the major safety risks of a hyperloop system.

Figure 12.1: Flowchart representing the structure of the safety analysis.

12.1 Importance of Safety

For a hyperloop system to be realised, the safety of the system must be guaranteed. The safety level of a hyperloop needs
to be at least at the same level as other high-speed modes of transportation, to be accepted by governments and used by
passengers. When designing the system, many lessons can be learned from other modes of transportation. In the past,
large accidents have caused transport innovations to be slowed down or even to be withheld from implementation. An
example is the accident with the Transrapid magnetic levitation train in 2006 in Lathen, Germany. During a trial run, a
train collided with a maintenance vehicle, causing 23 casualties.1 Although the accident was caused by a human error, a
Maglev train has never been realised in Germany ever since. To avoid a similar situation, it is crucial that the hyperloop
technology is tested thoroughly before the first passenger ride takes place and that possibilities of human error are min-
imised.

Another accident, a train derailment in Santiago de Compostela in 2013 with 80 casualties2 , shows the importance of
track curvatures at high speeds. With hyperloop speeds multiple times higher than a train, it is important to design all
trajectories with safe track curvatures. Accidents can also signify the end of fully operational modes of transportation, as
can be seen with the Concorde. In 2000, an aircraft ran over debris during take-off, which was lost by the previous aircraft
taking off. This caused a blown tyre and a punctured fuel tank, leading to a fire and ultimately 113 fatalities.3 Although
the Concorde had been used for decades, after the accident, passenger numbers diminished and the Concorde ceased
operation in 2003. These accidents show that it is important to focus on safety at all times: during design, testing and
operation.

1 Transrapid accident - https://en.wikipedia.org/wiki/Lathen_train_collision, consulted on 31 May, 2019

2 Santiago de Compostela accident - https://en.wikipedia.org/wiki/Santiago_de_Compostela_derailment, consulted on 31 May, 2019
3 Concorde accident - https://en.wikipedia.org/wiki/Air_France_Flight_4590, consulted on 31 May, 2019

49
Chapter 12. Safety Analysis

12.2 Research Approach

This section describes the method, scope and assumptions, and a top-level overview of the system. These factors are used
as input for the safety analysis.

12.2.1 Method Description

For the safety analysis, a combination of methods is used. Elements of a Failure Mode and Effect Analysis (FMEA) and a
Hazard and Operability study (HAZOP) are used in the Hazard Analysis. For the analysis, the hyperloop system is classified
into multiple subsystems: pod structure, tube structure, vacuum pumps, airlocks, propulsion, levitation, braking, wheels,
high-speed switches, pod equipment, communication, human operations and external influences. The Hazard Analysis
lists the main risks of all subsystems with their causes and consequences. Hazard mitigation methods are introduced, and
afterwards each risk is scored for its likelihood and severeness, leading to a risk score.

12.2.2 Scope and Assumptions

As the hyperloop system is very broad and still quite conceptual, it is impossible to conduct a detailed safety analysis.
Therefore, not all aspects are taken into account. The scope of the safety analysis is as follows:

• Only top-level subsystem hazards are determined, so no safety analysis on component or part level is conducted.
• Only failures caused during standard operational conditions are taken into account, so not during emergency or
maintenance situations.
• Severeness scores are only based on passenger injuries and casualties, not on financial effects.
• The likelihood and severeness of each hazard is determined for a connection between Amsterdam and Paris.
• The likelihood and severeness of each hazard is determined for the complete rolling stock.
• External human influences, such as terrorism and other vehicles crashing into the hyperloop infrastructure, are not
taken into account because it is expected that this is a similar situation as current high-speed rail infrastructure.
• Security is not taken into account, as this is primarily relevant for external human influences, instead of safety of
infrastructure and pod design.
• Hazards at stations, except from airlocks, are not taken into account, as this is expected to be similar to train stations.
• Only first order hazards are taken into account, so no accumulation of hazards, to avoid over-complexity.
• Only single hazards are taken into account, so no combination of hazards.

Additional to the scope, various assumptions were made:

• When determining the consequences, worst-case scenarios are considered.

• People are not able to enter the tube from the station or from emergency exits.
• Deterioration can be any sort of intrinsic failure.
• The track is directly attached to the tube.
• When a tube deforms, a pod can no longer pass safely.
• All pods are 50% occupied according to average turn-up-and-go systems[Preston, 2009].
• It is expected that the effect of magnetic fields on passengers is within the limits of human safety guidelines [Kircher
et al., 2018], as this is comparable to Maglev trains that are already in operation.

12.2.3 Top-level System Description

When conducting the safety analysis, all system parameters taken into account are assumed by Delft Hyperloop, many of
which described in detail in Chapters 3 to 8 of this report. A short description of all subsystems is given in this subsection.

Pod Structure: A hyperloop pod is a pressure vessel with similar characteristics to the fuselage of aircraft. The material of
the pod is assumed to be mainly aluminium. The outer diameter is 2.70 m, of which 0.20 m is for structural components,
including isolation and stiffeners. The passenger compartment has a length of 22 m. Including the aerodynamic cones,
the total pod length is 30 m. Each pod is bidirectional and can accommodate 50 passengers who can (dis)embark through
four doors, of which two on each side of the pod.

Tube Structure: The inner tube diameter is 3.50 m. To withstand the pressure difference, the tube wall thickness is 25
mm of steel. Above-ground tubes are supported by concrete pillars, with a spacing of 30 m.

Vacuum Pumps: Both roots pumps and backing pumps are used to create and maintain the near vacuum environment
of below 50 P a. Clusters of vacuum pumps will be distributed evenly over a link.

50
Chapter 12. Safety Analysis

Airlocks: To transfer pods from atmospheric pressure in the station to the near vacuum pressure in the tube and vice
versa, airlock chambers are used. In these chambers, the pressure will vary, depending on which direction the pod goes.
This is the same principle that is used in the International Space Station.

Levitation: The levitation of the pod is performed by Electrodynamic Suspension (EDS), as this is considered to be the
safer option compared to Electromagnetic Suspension (EMS) as discussed in Chapter 3.

Propulsion: Propulsion of the pod is performed by a Linear Synchronous Motor (LSM) and is continuous throughout
the whole track. Secondary propulsion to accelerate the pod to speeds over 20 km/h is done by electric motors and wheels
on board of the pod.

Braking: Braking of the pods is performed by the LSM.

Wheels: As EDS is only viable at speeds above 20 km/h, wheels are required to suspend the pods at low speeds.

High-speed Switches: In order to realise point-to-point connections efficiently, high-speed switches are essential. To
reduce length and costs, these high-speed switches are not designed for speeds over 1000 km/h, but several hundreds of
kilometres per hour lower. A high-speed switch is conducted by a attractive force created by the active lateral guidance.

Pod Equipment: To increase passenger safety, safety equipment such as fire extinguishers and oxygen tanks are installed
in each pod. Furthermore, electronic devices such as display screens and information panels are present in each pod to
increase passenger comfort. The ceiling of each pod will simulate the time of day by matching the colour of the skylight
to the outside world.

Communication: The communication is divided in two systems: the telemetry data and the pod sensor data. The
telemetry data is essential for safety, as this determines the location and velocity of the pod. Pod sensor data is less crucial,
since it is mostly focused on the status of the pod.

Human Error: Each pod operates fully autonomous during its entire journey. However, human errors can still happen,
in particular during or after maintenance of a subsystem.

External: Due to its enclosed environment, the hyperloop encounters significantly less external influences compared to
other modes of transportation. However, external influences are still present in the system, mainly affecting the tube and
the pillars.

12.3 Hazard Analysis

In the Hazard Analysis, the most important hazards of all subsystems are listed. For each hazard, the cause and its conse-
quences are determined. Furthermore, mitigation methods are decided upon to decrease the likelihood and severeness
of hazards. After determining the mitigation methods, each hazard is scored for its likelihood and severeness, leading to
an overview of the largest risks. The most important hazard mitigations and largest risks are described in further detail in
this section. The complete Hazard Analysis can be found in Appendix I.

12.3.1 Hazard Mitigation for Hyperloop Design

For each subsystem, the main hazard mitigation methods are described in this subsection.

Pod Structure: For the pod structure, it is important to design the pod with a safety factor, such that it can withstand
higher loads and corrosion. To detect that the pod is in a bad state, maintenance and monitoring is essential to lower
the likelihood of a hazard. Furthermore, track monitoring is needed to detect any anomalies on the track and thereby
preventing puncture by an object.

Tube Structure: Designing with a safety factor is essential for the tube, in order to withstand excessive loading, corro-
sion and soil settlement. Furthermore, a safety factor ensures that puncture by an object is less likely. Maintenance and
monitoring of the tube is needed to ensure the tube is in a good state.

51
Chapter 12. Safety Analysis

Vacuum Pumps: To reduce the consequences of a vacuum pump failure, a redundancy in the number of vacuum pumps
is desired. Maintenance and monitoring of the pumps ensure a lower deterioration rate.

Airlocks: Besides maintenance and monitoring of the airlocks, it is required that only one airlock door can be opened at
a time. This ensures that the near vacuum environment of the tube and the atmospheric pressure at the station are always
separated. At each station, there will be multiple airlocks, reducing the impact on operations if one of the airlocks fails.

Levitation: The EDS system should be maintained and monitored properly to maintain sufficient lift forces. To ensure
a safe gap height, a safety factor in maximum pod weight is needed. In addition, if the weight of each pod is monitored at
the station, it can be detected before departure if a pod is too heavy.

Propulsion: To ensure that a pod can still propel when the LSM fails, a secondary propulsion system is needed, which
uses electric motors and wheels that allow the pod to travel at low speeds. The LSM system needs to be designed to
withstand overcharging. Furthermore, the LSM will be segmented, reducing the severeness of a LSM failure.

Braking: As the braking subsystem is one of the most critical subsystems, it is important that the braking system is not
dependent on a single communication system. A secondary emergency braking system is also beneficial. Furthermore, a
redundant power supply for the LSM is required to guarantee that the pod can come to a safe standstill in case of a power
outage.

Wheels: As the wheels are only used at low speeds (< 20 km/h), failure severeness is minor. Maintenance and monitoring
of the wheels ensure a reduced likelihood for wheels not retracting or deploying.

High-speed Switches: Maintenance and monitoring is also important for the high-speed switches, to ensure the attrac-
tive force of the lateral guidance magnets is sufficient to switch. A safety factor applied to the length of the switch ensures
that each pod has sufficient time to reach a high speed after departure from a station, in order to be able to merge safely
into the pod flow. A switch design that avoids collisions ensures that when the switch mechanism is not working, the pod
will always continue straight ahead, instead of crashing into point where the two tubes will split up.

Pod Equipment: Similar to aircraft, safety equipment must be present in hyperloop pods, sufficient to save all 50 pas-
sengers in emergency situations. This includes, but is not limited to, fire extinguishers, oxygen tanks and masks, dehu-
midifiers, carbon dioxide and carbon monoxide scrubbers, additional batteries, a first aid kit and an automated external
defibrillator (AED). All safety equipment needs to be maintained and monitored frequently. It is important that high qual-
ity equipment is used that can withstand the near vacuum environment.

Communication: For the communication of the pods, it is crucial that all systems are not dependent on a single method
of communication. In case of a power outage, a redundant power supply must take over the communication. To reduce
the consequences of failure, the sensors responsible for the telemetry data of the pod are spread out equally through the
tube. Based on the values of other sensors, the missed values can be approximated.

Human Error: Although the system will be highly autonomous, human errors can still happen. To prevent most of the
human errors, clear and strict maintenance protocols and guidance is necessary. After maintenance, the tube and track
must be monitored to ensure that no objects are left behind. Additionally, monitoring of the pod, tube and track after
maintenance will detect anomalies. To prevent trespassing at the station, obstacles should be placed at stations to avoid
people accessing the track.

External: To withstand extreme external influences, safety factors are needed in designing the tubes and pillars. During
trajectory planning, earthquake prone areas need to be inspected to determine whether safety can be guaranteed in these
areas. In addition, lightning rods need to be installed and large trees close to the infrastructure need to be removed. As
pods are always in a closed environment, no additional hazard mitigations are needed for the pods.

52
Chapter 12. Safety Analysis

12.3.2 Outcomes
The extensive Hazard Analysis can be found in Appendix I. This includes all top-level hazards per subsystem. For each
hazard, the cause and consequences are given, together with a score for its likelihood and severeness, resulting in a total
risk score. The hazards with the highest risk scores are identified as the largest risks and are displayed in Tables 12.1 and
12.2. A division is made between ‘regular risks’ and ‘black-swan risks’. Black-swan risks are risks with an extremely small
likelihood, but catastrophic consequences.

Most of the regular risks are related to the communication subsystem. When travelling at speeds over 1000 km/h with a
30-seconds headway, location detection of all pods needs to be extremely reliable. As described in the hazard mitigations,
back-up communication systems are essential. Furthermore, communication technologies need to be tested thoroughly
in a high-speed test facility, to ensure their reliability. This is preferably also tested underground, as this brings additional
challenges compared to above-ground infrastructure. Furthermore, the analysis showed the importance of a redundant
power supply system to avoid pod crashes. This applies both to the tubes for communication, and to the pods for braking.
Another large risk is the deterioration of pod doors, which can expose passengers to the vacuum environment. Therefore,
the doors need to be opened inwards, such that the pressure difference ensures a better door sealing, minimising the like-
lihood of failure. This is comparable to the door design of passenger aircraft.

For the black-swan risks, it can be concluded from the analysis that they are all related to the tubes. Tube deformation
due to excessive loading of the tube or the connections in between the tube segments will have major consequences. Next
to that, severe external influences, such as an earthquake, flood or extreme temperatures, can also cause the tube to de-
form. When designing the system, probabilities of these events need to be examined thoroughly for different geographical
locations. Safety factors for tube and pillar design need to be determined for each location specific.

53
 Chapter 12. Safety Analysis
Table 12.1: Largest regular risks

Subsystem Cause Hazard Description Consequences

Communication Location sensors in tube fails Unreliable pod location Pod crashes into other pod or airlock door
Communication Signal of location sensor to operational No data of pod location Pod crashes into other pod or airlock door
center broken
Communication Power outage in tube No data of pod location Pod crashes into other pod or airlock door
Communication Radio interference Unreliable pod location Pod crashes into other pod or airlock door
Pod structure Deterioration of pod door Pod door does not seal properly Passengers get exposed to vacuum environ-
ment
Propulsion Communication failure between LSM Pod is not propelled properly Pods travel at different speeds which might
and pod lead to a collision
Braking Power outage Braking out of service Pod crashes into other pod or airlock door
Human error Bad maintenance of pod Crucial part of pod failing Pod crashes, passengers get exposed to vac-
uum environment
Human error Bad maintenance of tube Crucial part of tube failing Tube deformation and track misalignment

Table 12.2: Black-swan risks

Subsystem Cause Hazard Description Consequences

Tube structure Excessive loading of tube Ultimate Limit State (ULS) bending of the Tube permanently deforms
tube
Tube structure Excessive loading of tube connections Weld or bolt connection failure Tube deformation and track misalignment
Tube structure Soil settlement Settlement of the tube Tube deformation and track misalignment
External Plate tectonics Heavy earthquake ≥ 5.0 Richter Tube and pillars deformation
External Flood More soil settlements and erosion of pillars Tube deformation and track misalignment
External Extreme frost Tube shrinking Tube deformation
External Extreme heat Tube expanding Tube deformation
54
Chapter 12. Safety Analysis

12.4 Safe Haven Design

When a large risk occurs, it is important that the safety of all passengers in the system is ensured as much as possible.
The Safe Haven concept is based around the principle that in case of an emergency, a vehicle can always make its way
to a ‘Safe Haven’. Pods can stop at a Safe Haven where it will provide safe exit to the passengers of the pod in emergency
situations, thus functioning as an emergency exit station. This safety concept is comparable to that of an airplane. If an
airplane which is airborne ends up in an emergency, it must first make its way to an airport and make an emergency land-
ing, before the passengers can safely exit the airplane.

This section will provide a first insight into the meaning and design of the Safe Haven concept for a hyperloop system,
starting with an explanation of the design concept. After which, the requirements and design parameters for the design
of a Safe Haven system are given. Due to the complexity of a thorough safety analysis, the design approach is limited by
assumptions. From these assumptions and requirements, a framework is proposed to determine the values of the design
parameters for a Safe Haven system.

12.4.1 Design Concept

The goal of the Safe Haven concept is to guarantee a sufficient level of safety whilst minimising the investment cost for
emergency exits. This is important as both factors influence the feasibility of a hyperloop system. Firstly, the guarantee
of safety is crucial for the realisation of a future hyperloop system. Secondly, minimising the cost for emergency exits is
crucial to increase the economic feasibility of the system.

In this consideration lies a balance, as more safety exits improve the level of safety, but decrease the economic feasibility
of the system. On the contrary, a system with fewer safety exits decreases the level of safety, but increases the economic
feasibility. Therefore, the Safe Haven system must be designed to minimise the cost by decreasing the amount and size of
Safe Havens, whilst still guaranteeing safety in case of emergencies.

To design the system for a sufficient level of safety, the largest risks listed in Tables 12.1 and 12.2 are considered. These risks
can be generalised into two emergency situations. Firstly, ‘in-pod emergencies’, which cover all emergencies inside of a
pod that threaten the safety of passengers but do not affect the functionality of the rest of the system. Secondly, ‘brick-wall
emergencies’ are considered. These emergencies consider emergencies that make the safe passage of a pod past a certain
part of the tube impossible.

12.4.2 Requirements
The design of a Safe Haven system must meet different requirements to provide safety, operability and economic feasibility
of the system. The requirements are used as input for the conceptual design of a Safe Haven and the configuration of the
design parameters. Below, a top-level list of requirements for a Safe Haven is given. A Safe Haven system must:

• provide a safe evacuation time in case of an ‘in-pod emergency’.

• provide a safe evacuation time in case of a ‘brick-wall emergency’.
• minimise impact on operability of the system in case of an emergency.
• must be designed to be as small as possible to minimise the investment cost.

12.4.3 Design Parameters

To design a Safe Haven system, the requirements mentioned must be translated into design parameters. Based on these
design parameters, a framework can be approached for the design of a Safe Haven system. The design parameters are
listed below and are visualised in Figures 12.2 and 12.3.

• Intermediate Safe Haven distance.

• Safe Haven pod capacity.
• High-speed switch length.

55
Chapter 12. Safety Analysis

Figure 12.2: Conceptual design of a Safe Haven system on a hyperloop link

Figure 12.3: Conceptual design of a Safe Haven, where the dotted line represents the track

12.4.4 Safe Haven Assumptions

For the approximation of the design parameters the system is analysed based on the top-level safety hazards. From these
safety hazards, a first list of requirements for the Safe Haven design has been established. Due to the high complexity of
a thorough safety analysis, the approach for a Safe Haven design has been limited by certain assumptions. Below, a list of
assumptions is given.

• A Safe Haven is always operational.

• No hazards occur inside a Safe Haven.
• The design is only based on an analysis of stand-alone risks, meaning that the accumulation of hazards is not con-
sidered.
• The headway between pods must always remain above a safe braking distance.
• The survivability inside of a pod, in case of a ‘brick-wall emergency’, is in the order of hours, comparable to aircraft.
• In case of an emergency, pods are always able to propel forwards or backwards on primary or secondary propulsion
systems.

12.4.5 Design Framework

In Figures 12.2 and 12.3 a conceptual design for a Safe Haven system is given. In this section, a conceptual framework is
given to approach the values of the design parameters that will determine the design of a Safe Haven system. The concep-
tual design framework is visualised in Figure 12.4.

Intermediate Safe Haven Distance

To determine the intermediate distance of the safe havens, the acceptable evacuation time for passengers, in case of an
‘in-pod emergency’, is considered to be governing. This is based on the assumptions that pods are always able to travel
forwards and backwards in emergency situations, and that the survivability of the pod in case of a ‘brick-wall’ emergency
is in the order of hours.

Thereby the intermediate distance between Safe Havens can be approached as a function of the acceptable time for pas-
sengers to be evacuated in case of an ‘in-pod emergency’. This time must be determined as an industry standard. Due to
the variance in evacuation time and procedures in different modes of transportation (e.g. trains and airplanes) a detailed
approximation cannot be given at this moment.

56
Chapter 12. Safety Analysis

Safe Haven pod Capacity

To facilitate any form of evacuation, the minimum capacity of a Safe Haven must at least be sufficient to evacuate one pod
at a time. Furthermore, in case of an ‘in-pod emergency’, the effect on the operability of the system must be minimised
by configuring the Safe Haven as a separate tube section, such that other pods can still pass by. This can be realised by
incorporating a high-speed switch at the beginning and end of the Safe Haven 12.3.

In case of a hazard which results in a ‘brick-wall emergency’, the Safe Haven system must be able to evacuate all pas-
sengers within an acceptable amount of time. Based on the assumption that the survivability of the pod is in the order
of hours, the pod capacity in a Safe Haven can be approached as a function of the acceptable time for pod evacuation in
case of a ‘brick-wall emergency’, and the number of Safe Havens on a link. Based on the link length and the intermediate
distance of the Safe Havens, the number of Safe havens can be determined.

High-speed Switch Length

To ensure system operability in case of an ‘in-pod emergency’, a switch must be incorporated in the Safe Haven design, as
visualised in Figure 12.4. Via this switch, a pod can stop at a Safe Haven without blocking the tube for incoming pods. The
switch length must be determined as a function of the pods cruising speed and the acceptable delay for other pods.

Since the switch length must enable a pod to come to a complete standstill at the Safe Haven, higher cruising speeds
result in a higher switch length. By incorporating the option of lowering the cruising speed of a pod in an emergency,
before taking a high-speed switch, the length of a switch can be significantly reduced. This reduction of cruising speed in
the main tube must be determined based on the acceptable delay for other pods, as lowering the cruising speed of a pod
in ‘in-pod emergency’ will affect the speed of incoming pods.

Figure 12.4: Conceptual design framework for a Safe Haven system

57
Chapter 12. Safety Analysis

12.5 Recommendation
The goal of this safety analysis is to highlight the main risks of a hyperloop system. Hazard mitigation methods are deter-
mined to manage these risks. Furthermore, a conceptual design framework for a Safe Haven is given.

To increase passenger safety of the system, hazard mitigation methods can be incorporated during design. A list of the
most important hazard mitigations is as follows:

• Maintenance and monitoring of all subsystems ensures a reduced likelihood for failures.
• Using proper safety factors for design enables the system to withstand unexpected situations.
• Different safety factors according to geographical location of infrastructure to account for external influences.
• Clear and strict maintenance protocols and guidance diminishes human errors.
• A secondary communication system is essential for a reliable location detection of all pods at all times.
• A redundant power supply system ensures safe operation in case of power outage.
• Safety equipment needs to be installed in every pod to enlarge passenger safety in emergency situations.

Communication is a critical subsystem. Current technologies do not suffice for hyperloop speeds and need to be devel-
oped further. Next to that, the tube is the most sensitive subsystem for black-swan risks: risks with a low likelihood and
severe consequences.

Furthermore, it is important to test all subsystems thoroughly. A test facility where pods can reach speeds over 1000
km/h is recommended to ensure safety and reliability of a hyperloop system. It is recommended that a part of the test
facility is located underground, amongst others to be able to determine if the communication system is reliable.

This safety analysis only takes into account passenger consequences; cost effects are not directly taken into account.
Ultimately, for designing a hyperloop system, cost optimisation must be taken into account.

To minimise the cost for emergency exits, whilst still guaranteeing safety, a conceptual Safe Haven design is presented.
These Safe Havens are intermediate emergency exit stations. In emergency situations, pods can stop at a Safe Haven,
where it will provide safe exit to the passengers. For further research in the feasibility of a Safe Haven system it is recom-
mended to integrate a more in-depth safety analysis where the accumulation of hazards and other emergency protocols
are also considered. Furthermore, the performance of a safety analysis with a cost-benefit analysis could give a better
insight into the feasibility of the Safe Haven concept.

58
13.Regulatory Implications
As the hyperloop is a new mode of transportation, regulation does not yet exist. Many lessons can be learned from other
transportation modes, and various standards and laws can be easily adjusted and applied to the hyperloop. For example,
large similarities with (Maglev) trains and aircraft are present. However, the hyperloop system has some major differences
as well. Therefore, suited legislation is required, as novel issues will arise. It can take a long time to set up and implement
legislation, and that is the reason why it is important to take regulatory implications into account early in the process
of implementing a new mode of transportation. Regulatory implications include all aspects where governments are in-
volved. In particular, standardisation, legislation and certification are important aspects, which are heavily correlated.

This chapter highlights why standardisation is important in hyperloop development. Afterwards, multiple factors to take
into account for legislation and policies are listed, followed by a description of challenges and considerations regarding
certification. In the end, a recommendation is provided on the most important factors related to regulatory implications
of a hyperloop.

13.1 Standardisation
Standardisation is a challenge for hyperloop. As explained in Chapter 2, multiple companies are working on the hyperloop
concept, with different ideas. In the end, it is important that a European hyperloop network has a single standard. There-
fore, hyperloop companies should eventually converge to a standardised concept. Although it might sound logical, system
parameters such as tube diameter must be the same to increase interoperability between countries. For comparison, the
width of European train tracks differs between countries, which used to make it complex and expensive to have trains
operating internationally. It is important that in the end, companies working on hyperloop converge to a consensus on
important design parameters. These parameters lead to standards that have to be determined together with governments.
However, standards must not be decided upon too early in the process, as this constrains the development of innovative
technologies or ideas. Multiple technologies have to be researched and developed first in order to determine what the
best option is to use in the eventual standardised hyperloop system.

13.2 Legislation and Policies

To be able to operate a hyperloop and guarantee its safety, legislation is required. At the moment, regulation does not exist
yet and it is hard to determine on final legislation, as there are still multiple uncertainties in hyperloop design. However,
various aspects of existing modes of transport are similar to a hyperloop. The pressure vessel of a pod has great similarities
with the fuselage of an aircraft, although the pressure difference that a hyperloop pod has to withstand is larger. Next to
that, the propulsion system is very similar to existing Maglev trains, however the hyperloop operational speed is signifi-
cantly higher. These examples both show that hyperloop legislation can be derived from other modes of transportation,
but has to be adjusted.

Before a complete legislation framework can be decided upon, the technology has to be proven. According to Leibow-
icz [2018], it is advised to support adequate infrastructure provision early in the technology life cycle, in order to support
a transition to sustainable mobility. In current transport systems, it is shown that infrastructure provision always pre-
cedes the adoption of vehicles. Findings suggest that early public policy support for infrastructure can be essential for the
hyperloop, in order to speed up the development process.

13.3 Certification
Similar to legislation, various norms and standards for certification can also be derived from railway and aircraft certifi-
cation. For example, the V-model approach that is used in the EN 50126 railway certification standard1 , could be applica-
ble to hyperloop certification. Furthermore, railways have multiple Technical Specifications for Interoperability (TSIs) 2 ,
amongst others related to the infrastructure, energy, rolling stock, control command and signalling, and maintenance and
operation. It is expected that the following TSIs are most useful for determining hyperloop certification, as these share the
most similarities:

• Safety in Railway Tunnels.

• Control Command and Signalling.
• Persons with Disabilities and with Reduced Mobility.
1 V-model EN 50125 - http://en50126.blogspot.com/2008/09/v-model.html, consulted on 31 May, 2019
2 Technical Specifications for Interoperability - https://www.era.europa.eu/activities/technical-specifications-interoperability_en,
consulted on 31 May, 2019

59
Chapter 13. Regulatory Implications

As hyperloop legislation does not exist yet, it is hard to determine certification standards. It is advised to start with cer-
tification as early as possible: during the design process, certification can already be taken into account. Furthermore,
as multiple stakeholders are involved in the certification process, it is important to include them early to speed up the
process. Another option to accelerate certification, is by proving the safety of the system. To prove that a hyperloop is safe
at high speeds, it is essential that a test facility is constructed where the system can be tested thoroughly on speeds over
1000 km/h.

Certification can be arranged in two ways. The first option is similar to trains, where multiple organisations are allowed
to carry out the certification. The second option is comparable to aircraft, where a single organisation is responsible for
certification. In Europe, this is the European Union Aviation Safety Agency (EASA). For the hyperloop, the latter option
is recommended, as this will ease the certification process on a European level. It is advised to found an agency that is
responsible for hyperloop certification. Multiple stakeholders should be involved in this agency.

13.4 Recommendation
This chapter briefly discusses the regulatory implications of a hyperloop system, focusing on standardisation, legislation
and policies, and certification. For standardisation, it is important that there will be a single European standard to improve
interoperability. However, it is necessary that this does not happen too early in the process, as multiple techniques have
to be developed first in order to research their potentials. To accelerate development of these techniques, policy support
for infrastructure is essential.

Legislation can be derived from (Maglev) trains and aircraft, as these share various similarities with the hyperloop. How-
ever, further investigation of these similarities is essential to determine what can be incorporated in hyperloop legislation
and what needs to be adjusted for or complemented on. Especially regulations concerning safety standards for hyperloop
should be determined at European level, as hyperloop’s unique vacuum and high-speed characteristics will require spe-
cific safety regulations.

The same holds for certification: multiple Technical Specifications of Interoperability of railways can be used as a base
for hyperloop certification. To speed up the process, it is important to start early with certification, which can already
happen during the design process. It is advised to found an agency that is responsible for certification on a European
level. Furthermore, a test facility, allowing pods to be tested at high speeds, is recommended to be able to guarantee and
prove the safety of hyperloop.

60
14.Future Points of Notice
In this chapter, the most important points of notice are listed. These points are either barriers, preventing the hyperloop
from realisation, or challenges, making the hyperloop less feasible. The purpose of this chapter is to provide an overview
of these barriers and challenges in order to determine focus points for the development of the hyperloop.

14.1 Main Barriers for Hyperloop Development

In this section the main barriers for the hyperloop are discussed. Barriers are defined as the main obstacles for the suc-
cessful implementation of a hyperloop system right now. These barriers have to be removed or solved in order to realise a
hyperloop system.

14.1.1 Standardisation
There should be one standard for the hyperloop that is used throughout Europe. This ensures that all links can eventually
be connected to create a large hyperloop network. The challenge here lies in determining the design parameters that will
be used. Currently, there are multiple parties spread out over the world with different visions for the hyperloop. These
parties want their vision to be realised and thus it is hard to converge to one standard. First, in order to be able to make
properly considered choices for standardisation, it is important that different concepts are developed and tested. Setting
up a standard too soon can constrain the development of innovative concepts.

14.1.2 Financing the Hyperloop

As the hyperloop infrastructure is expensive, it is impossible to find a single party that is able to finance the construction of
a hyperloop network. A public-private partnership is most likely needed, as it is not expected that governments will fully
finance the system. However, they will contribute to the project, as a hyperloop offers socio-economic value to society.
Furthermore, as many hyperloop trajectories will be international, it is challenging to determine the amount of investment
that is coming from each stakeholder that is financially involved.

14.2 Challenges
Challenges are seen as problems that do not stand in the way of realising the hyperloop. However, solving these challenges
will contribute to the feasibility of a hyperloop system.

14.2.1 Political Aspects of the Hyperloop

The hyperloop aims to be an infrastructure network connecting countries and large hubs. This means that different local
governments, national governments and other stakeholders need to have close communication. Large projects with many
involved parties take a long time to be realised. All stakeholders need to be satisfied and governments on different levels
have to be aligned at all times. The location of stations is also closely related to politics. A consensus has to be reached
which cities will be included in the hyperloop network, and which cities will not.

14.2.2 High-speed Switches

In a hyperloop network, all stations are connected with links. However, there are multiple ways to connect to a station
using the links. The first option is to directly connect all stations similar to a metro system. All links pass through the
stations where the pods have to stop. The second option is to use a highway system, using on- and off-ramps to connect to
the station. This allows pods to cruise past the station and thereby create direct links throughout the whole network. This
increases the efficiency of the system and reduces the average travel times. However, pods need to switch to an on- or off-
ramp at high speeds. These high-speed switches are crucial for efficient operation of the system. Current developments
on high-speed switches are in the early stages and the technological feasibility is yet to be proven.

14.2.3 Data Communication

The communication system of a hyperloop depends on multiple design aspects. The choice for the levitation mechanism
defines the responsibility for data collection, whether in the pod or at the infrastructure. This decision still has to be
made. Another decision relates to the communication between a moving pod and the tube. A potential concept is based
on photoelectric sensors located in the tube that will read information displayed on the pods. A decision on what type
of photoelectric sensor is best to use, and what data should be presented at the surface of the pod are two important
decisions that have yet to be made.

61
Chapter 14. Future Points of Notice

14.2.4 Safe Haven Design

To guarantee safety inside a hyperloop pod, an emergency system must be designed. A concept discussed in Section 12.4
provides a first vision on how a safety system could take shape to maximise passenger safety, whilst minimising emergency
exit costs. However, due to the uncertainty of the hyperloop design and a lack of regulation, it is still unknown how safety
can be guaranteed. Emergencies in which passengers need to be evacuated are impossible to avoid. Designing emergency
exits in a way that both accommodates sufficient safety and acceptable costs is a challenge that must be overcome.

14.2.5 Business Case Towards Implementation

Besides the technological challenges, it is important to determine a business case for a hyperloop. The first link will be very
important and crucial for the success of a hyperloop system. Many stakeholders will be involved including different levels
of governments. It is desired that this first link creates sufficient passenger demand to increase the economic feasibility of
the link.

14.2.6 Technology Costs Reduction

Parts of the hyperloop rely on new innovative technologies that have not been optimised for costs. Since these technolo-
gies are going to be used in large quantities, the costs for these systems will quickly add up. Reducing the technology costs
for example for levitation and propulsion could greatly reduce the total costs for the hyperloop infrastructure. Therefore,
cost reduction for these technologies must be realised to increase the economic feasibility of the system. For a first cost
estimation of the hyperloop system, see Chapter 10.

14.2.7 Integration with Current Modes of Transportation

Hyperloop is a means of public transportation and does not accommodate for door-to-door travel all by itself. The hyper-
loop needs to connect to other modes of transportation to allow passengers to travel further to their destination. This is
important for connectivity and will increase the accessibility of the hyperloop. Currently, there is not always space near
the stations of existing infrastructure for the implementation of hyperloop stations. This complicates the intermodal con-
nection of hyperloop to other modes of transportation. Finding smart ways of connecting to other public transportation
or spatial planning will be challenging.

14.2.8 Crossing Waterways

In order to connect destination overseas or destinations surrounded by large waterways, such as fjords, hyperloop tubes
must be able to cross waterways. Due to the large curvature radius, it is unlikely that the hyperloop can make the correct
horizontal and vertical curves to follow the irregularities on the seabed. A bridge is not always the solution as the spans
or water depths can be too large. Therefore, connecting these destinations that are separated by large waterways will be a
challenge.

A concept currently being developed is submerged floating tunnels. This concept would allow for more freedom in net-
work design which leads to a more optimal network. However, this technology is far from ready, expensive and requires
more fundamental research.

14.2.9 Tunnel Boring Machine Speed

Large operations speeds of hyperloop pods result in large curvature radii. Therefore, in densely urbanised and mountain-
ous areas the hyperloop is required to travel underground. It is currently estimated that 50% of hyperloop infrastructure
will be build underground. Current Tunnel Boring Machines (TBM) have a boring speed of approximately 15 metre per
day in soft soil. To construct a single link of several hundreds of kilometres long with one TBM, multiple decades of digging
is needed before completion of the link. Multiple TBM’s can be used for the construction of one tunnel to significantly
reduce the time needed for this process but with the current speed for TBM’s this is still a long and costly process. It is of
great importance to speed up the digging speeds of TBM’s to both decrease the construction time of the hyperloop and to
reduce the costs of the hyperloop, as this would be positive for the feasibility of the hyperloop.

62
15.Conclusions
The hyperloop provides a sustainable solution to the growing demand for high-speed travel within Europe. Besides be-
ing significantly more energy efficient than aviation, a hyperloop is powered by electricity coming from both solar panels
and other renewable energy sources. The turn-up-and-go system combined with speeds faster than airplanes will com-
pletely change human perception of travel and connect cities throughout the entire continent. In order to successfully
realise a hyperloop system, two main focus points can be identified: working towards standardisation and setting up the
foundation for implementation.

Standardisation
At the time of publishing, several organisations are working on the development of the hyperloop concept. These organi-
sations all have different ideas and stakes. However, in order to realise a hyperloop system on a European level, standard-
isation is required on system parameters including but not limited to tube size, operating pressure, operating speed and
guideway layout. It is recommended to already start setting up a framework for standardisation in the future. The three
main steps will be continuing research and development, sharing knowledge and investing in a long test track.

Continue Research and Development of Hyperloop Technologies: Companies are making progress on developing cur-
rent technologies for a hyperloop system, but the technology is not completely ready for implementation. In order to
be able to make properly considered choices for standardisation, it is important that different concepts are developed
and tested. Specifically, multiple concepts for the propulsion and levitation subsystem exist. The energy efficiency of the
levitation concepts and reliability and safety of propulsion concepts needs to be improved. Furthermore, the hyperloop
pod should be designed on subsystem level based on a system engineering approach. For the infrastructure, new promis-
ing materials and innovative designs for tubes can decrease costs. Top-level parameters of a hyperloop system such as
operation speed and pressure must be analysed in more detail with simulations.

Share Knowledge and Bring Parties Together: Multiple companies and governments need to collaborate on a European
level to be able to realise a hyperloop system. Engineering, certification and implementation are elaborate, time consum-
ing and expensive processes for which different experts and parties are required. Sharing knowledge will accelerate the
development and create public awareness and acceptance. An independent knowledge platform where various parties
contribute, such as Hyperloop Connected1 , is essential to achieve this.

Invest in a Long Test Track: A long test track is required to test subsystems at velocities over 1000 km/h which has not
been performed at the time of publishing. The different concepts mentioned before have to be tested on full-scale level
in order to evaluate their behaviour. Besides concept and subsystem tests, a long test track serves as a prototype for the
infrastructure. Testing the vacuum system and the structural behaviour of a full-scale tube will be crucial before standard-
isation can take place. Besides, communication systems must be tested at high velocities in the enclosed environment of
the tubes and potentially even underground. Finally, a long test track will be required for certification purposes as well.

Implementation
Standardising the technology and general system parameters is not sufficient for a successful realisation of a hyperloop.
Planning for actual implementation should already start in the early phases of the project in order to create maximum
benefit for future passengers and European citizens in general. The three main steps will be to found a European agency
for certification, secure financing and design links.

Found European Agency for Certification: Legislation can be derived from (Maglev) trains and aircraft, as these share
various similarities with the hyperloop. However, further investigation of these similarities is essential to determine what
can be incorporated in hyperloop legislation and what needs to be adjusted for or complemented on. Especially regu-
lations concerning safety standards for hyperloop should be determined at a European level. It is advised to found an
agency that is responsible for hyperloop certification. Multiple stakeholders should be involved in this agency.
1 Hyperloop Connected - www.hyperloopconnected.org

63
Chapter 15. Conclusions

Secure Financing: As the hyperloop infrastructure is expensive, it is nearly impossible to find a single party that is able to
finance the construction of a hyperloop network. A public-private partnership is most likely needed, as it is not expected
that governments will completely finance the system. A business model is required to attract private investors. The in-
vestment costs required for the infrastructure must be determined with a greater accuracy and revenue models should be
created.

Analyse and Design Individual Links: In order to proceed with the implementation of a European hyperloop network,
additional market research is required in order to make better estimations on expected demand. This will change the
network design and especially the expected cost per travelled kilometre. Furthermore, links need to be evaluated and
designed in more detail on individual basis. This includes placement of the tubes in respect to existing infrastructure and
urban areas both above- and below ground, placement of the stations with respect to the cities and potentially increasing
connectivity by adding intermediate stations. Moreover, the operations of a hyperloop system are not yet determined.
Decisions need to be made regarding time schedules, frequency of routes and intermediate stops which will influence the
design of the infrastructure. Finally, it is important that the location of the first link and first connections is determined.

64
Bibliography
Ai, B., Cheng, X., Kürner, T., Zhong, Z.-D., Guan, K., He, R.-S., Xiong, L., Matolak, D. W., Michelson, D. G., & Briso-Rodriguez,
C. (2014). Challenges toward wireless communications for high-speed railway. IEEE transactions on intelligent trans-
portation systems, 15(5), 2143–2158.

Bird, J. (2007). An investigation into the use of electrodynamic wheels for high-speed ground transportation.

Bird, J. & Lipo, T. (2019). A study of the effect of using electrodynamic wheels in series.

Bird, J. & Lipo, T. A. (2003). An electrodynamic wheel: an integrated propulsion and levitation machine. 3, 1410–1416
vol.3.

Campos, J. & De Rus, G. (2009). Some stylized facts about high-speed rail: A review of hsr experiences around the world.
Transport policy, 16(1), 19–28.

Cassat, A. & Bourquin, V. (2011). Maglev – worldwide status and technical review.

Chen, Z., Xue, J., Rose, A. Z., & Haynes, K. E. (2016). The impact of high-speed rail investment on economic and environ-
mental change in china: A dynamic cge analysis. Transportation Research Part A: Policy and Practice, 92, 232–245.

Cutnell, J. D. & Johnson, K. W. (2005). Essentials of physics. Essentials of Physics, by John D. Cutnell, Kenneth W. Johnson,
pp. 694. ISBN 0-471-71398-8. Wiley-VCH, March 2005., 694.

D. Van Wie, F. K. & Walsh, R. (1996). Starting characteristics of supersonic inlets.

Edenbaum, J., Brinckerhoff, P., Cox, S., & English, G. (2015). Cross-passageways vs. emergency exit stairways in rail tunnels.

European Environment Agency (2018). Passenger and freight transport demand.

Eurostat (2017a). European aviation in 2040.

Eurostat (2017b). Transportation database.

Fischer, G., Zapotowski, B., States., U., & Corporation., G. A. (1974). Tracked levitated research vehicle : Periodic test sum-
mary report : primary suspension-aeropropelled / G. Fischer, B. Zapotowski ; Grumman Aerospace Corporation, Bethpage,
New York. U.S. Federal Railroad Administration, Office of Research, Development and Demonstrations : for sale by the
National Technical Information Service Washington : Springfield, Va.

Gleave, S. D. (2016). Study on the prices and quality of rail passenger services.

Goeverden, K. v., Milakis, D., Janic, M., & Konings, R. (2018). Analysis and modelling of performances of the hl (hyperloop)
transport system. European Transport Research Review, 10(2), 41.

Han, H.-S. & Kim, D.-S. (2016). Magnetic levitation, maglev technology and applications.

Hauviller, C. (2007). Design rules for vacuum chambers.

He, R., ai, b., Wang, G., Guan, K., Zhong, Z., F. Molisch, A., Briso-Rodríguez, C., & Oestges, C. (2016). High-speed railway
communications: From gsm-r to lte-r. IEEE Vehicular Technology Magazine, 11, 49–58.

He, R., Ai, B., Wang, G., Guan, K., Zhong, Z., Molisch, A. F., Briso-Rodriguez, C., & Oestges, C. P. (2016). High-speed railway
communications: From gsm-r to lte-r. Ieee vehIcular technology magazIne, 11(3), 49–58.

HM Treasury and Infrastructure UK (2010). Infrastructure cost review: Technical report.

Kaye, R. J. & Masada, E. (2004). Comparison of linear synchronous and induction motors.

Kircher, R., Palka, R., Fritz, E., Eiler, K., Witt, M., Blow, L., & Klühspies, J. (2018). Electromagnetic Fields of High-Speed
Transportation Systems. Maglev Technologies in Comparison with Steel-Wheel-Rail.

Knaap, T. & Oosterhaven, J. (2011). Measuring the welfare effects of infrastructure: A simple spatial equilibrium evaluation
of dutch railway proposals. Research in Transportation Economics, 31(1), 19–28.

65
Bibliography

Leibowicz, B. D. (2018). Policy recommendations for a transition to sustainable mobility based on historical diffusion
dynamics of transport systems. Energy policy, 119, 357–366.

Lever, J. H. (1998). Technical assessment of maglev system concepts.

Liu, Z., Long, Z., & Li, X. (2015). Maglev Trains: Key Underlying Technologies. Springer Tracts in Mechanical Engineering.
Springer Berlin Heidelberg.

M, S. (2009). Optical Fiber Communications: Principles and Practice. Pearson Education.

M. Agarana et al. (2016). Application of dijkstra algorithm to proposed tramway of a potential world class university.

M. Dikaiakos et al. (2009). Cloud Computing Distributed Internet Computing for IT and Scientific Research.

Mankins, J. (1995). Technology readiness level – a white paper.

McKinsey & Company (2017). A better approach to airline costs.

Moore, G. E. (1965). Cramming more components onto integrated circuits. Electronics, 38(8).

Musk, E. (2013). Hyperloop white paper.

NRC (2016). Station delft pakt nog eens 39 miljoen duurder uit.

Ocean Concrete (2019). Concrete price list.

Opgenoord, M., Merian, C., Mayo, J., Kirschen, P., O’Rourke, C., Izatt, G., Monahan, G., Paxson, D., Wheeler, C., Zhang, S.,
Zhang, C., Sakhibova, N., Vancea, G., Aggarwal, R., Ball, S., Caplan, P., Chamberlain, P., Chen, J., Chen, S., & Vaish, S.
(2017). Mit hyperloop final report.

Peach, M. (2012). Nec and corning achieve petabit optical transmission.

Pfeiffer Vacuum GmbH (2013). The vacuum technology book.

Preston, J. (2009). The case for high speed rail: A review of recent evidence. Paper, 9, 129.

Rose, C., Peterson, D., & Leung, E. M. (2019). Implementation of cargo maglev in the united states.

Russell, S. J. & Norvig, P. (1995). Artificial Intelligence: A Modern Approach. Prentice Hall.

Saha, S. (2018). A comprehensive guide to convolutional neural networks.

Thornton, R. D. (1993). Linear synchronous motors for maglev.

UK Government (2015). High speed two. a guide to tunnelling cost.

UK Government (2016). Plans and sections of hs2 phase one as amended in select committee.

Wang, C. & Zong, G. (2010). A contrastive study on sustainable development of maglev and high-speed wheel-rail.

Yang, Ying, & Deng (2018). Study on the optimization of linear induction motor traction system for fast-speed maglev
train. Transportation Systems and Technology, 4(3), 156–164.

66
Glossary
Aerodynamics The study of air and the interaction of air and solid bodies.
Antenna An antenna is the interface between radio waves propagating through
space and electric currents moving in metal conductors, used with a
transmitter or receiver.
Artificial Intelligence Artificial Intelligence (AI) is the theory and development of computer
systems able to perform tasks normally requiring human intelligence,
such as visual perception, speech recognition, decision-making, and
translation between languages.
Autonomous Able to operate without being controlled directly by humans.
Bandwidth Bandwidth is the difference between the upper and lower frequencies
in a continuous band of frequencies.
Black-swan Risk A risk with a small likelihood, but large consequences.
Centralised Data Processor The centralised data processor is a station where computing stations
will be present. In here, computations take place. The centralised
data processor is constantly connected with pods and tubes.
Certification The process of providing official approval of a technology.
Cloud Computing Cloud Computing is the on-demand availability of computer system
resources, especially data storage and computing power, without di-
rect active management by the user.
Complexity Compromising parts in an intricate arrangement which makes the
system difficult to predict and introduces uncertainty.
Computational Power Computational power is the amount of useful work accomplished by
a computer system. Outside of specific contexts, computer perfor-
mance is estimated in terms of accuracy, efficiency and speed of exe-
cuting computer program instructions .
Computer Science Computer science is the study of processes that interact with data and
that can be represented as data in the form of programs. It enables
the use of algorithms to manipulate, store, and communicate digital
information. A computer scientist studies the theory of computation
and the practice of designing software systems.
Conductor Material that is able to conduct electricity.
Control Loop A computer algorithm that continuously checks if values are as de-
sired, it will send a signal to perform a certain action if this is not the
case.
Controllable The ability to dampen oscillations and irregularities in movement to
guarantee stable movement.
Convolutional Neural Network A convolutional neural network (CNN, or ConvNet) is a class of deep
neural networks, most commonly applied to analysing visual im-
agery.
Cryogenic Production and behaviour of materials at very low temperatures.
Data Data is any sequence of one or more symbols given meaning by spe-
cific act(s) of interpretation.
Datapoint The smallest unit of data.
Doppler Effect High velocities lead to a shift of the received frequency, called the
Doppler Effect, which might lead to communication problems .
Eddy Current Loops of electrical current induced within conductors by a changing
magnetic field.
Electromagnets Type of magnet in which the magnetic field is produced by an electric
current.
Ferromagnetic The phenomenon that certain materials form permanent magnets or
are attracted to magnets, iron is an example.
Fuselage The main body of an aircraft.

67
Glossary

Guidance Makes sure that the lateral forces are stable and that the pod always
stays on the guideway.
Guideway The track that is positioned in the hyperloop tube. Levitation, propul-
sion and other system will be integrated into the guideway.
Induction The creation of currents due to changing magnetic fields.
Legislation The process of making and enacting laws.
Lift-Off Velocity The velocity at which the lift force is larger than the gravity and levi-
tation is created.
Lift-over-Drag The amount of lift created divided by the amount of drag, an indica-
tion of the efficiency.
Mean time to failure (MTTF) Mean time to failure (MTTF) denotes the expected time to failure for
a non-repairable system.
Optical Fibre An optical fibre is a flexible, transparent fibre made by drawing glass
(silica) or plastic to a diameter slightly thicker than that of a human
hair. Optical fibres are used most often as a means to transmit light
between the two ends of the fibre and find wide usage in fibre-optic
communications, where they permit transmission over longer dis-
tances and at higher bandwidths (data rates) than electrical cables.
Paris Agreement An international agreement within the United Nations, signed in
2016, to deal with climate change by limiting global warming.
Passive No systems present that requires power.
Pattern Recognition Pattern Recognition is the automated recognition of patterns and reg-
ularities in data.
Photoelectric Sensors A photoelectric sensor, or photo eye, is an equipment used to dis-
cover the distance, absence, or presence of an object by using a
light transmitter, often infrared, and a photoelectric receiver. There
are three different useful types: opposed (through beam), retro-
reflective, and proximity-sensing (diffused).
Pressure vessel A container designed to hold gases or liquids at a pressure substan-
tially different from the ambient pressure.
Primary The part of a Linear Induction Motor that creates electromagnetic
fields and induces a current in the secondary.
Regenerative Braking Energy recovery mechanism which slows a vehicle by converting its
kinetic energy into electrical energy.
Regulation All aspects where authorities are involved, such as standardisation,
legislation and certification.
Reliability The ability of a system or component to perform its required func-
tions under stated conditions for a specified time.
Robot An artificial Intelligence system, based on living organisms.
Safety Condition of being protected from or unlikely to cause danger, risk or
injury.
Safety Margin Used in engineering to make sure a calculation or conclusion is safe
and reliable.
Scalable The ability of the system to be mass producible or adjustable.
Secondary The conductor part of a Linear Induction Motor, in which currents
are induced that created electromagnetic fields.
Sensor A sensor is a device, module, or subsystem whose purpose is to de-
tect events or changes in its environment and send the information
to other electronics, frequently a computer processor.
Shanghai Transrapid The Maglev that is currently operating in Shanghai, developed by
Transrapid.
Span A span is the unsupported distance a between two points in a con-
struction. The construction should be strong and stiff enough to hold
up across the entire span.
Stable The passengers experience a smooth journey, resonance oscillations
should be avoided.

68
Glossary

Standardisation The process of determining parameters conform a standard through-

out the whole industry.
Stator The non-moving part in a Linear Induction Motor.
Superconductivity A phenomenon of zero electrical resistance and expulsion of mag-
netic flux fields.
Through Beam A Through Beam sensor is a photoelectric sensor. A through beam
arrangement consists of a receiver located within the line-of-sight of
the transmitter. In this mode, an object is detected when the light
beam is blocked from getting to the receiver from the transmitter.
Tube Cylindrical element through which the hyperloop pods travel. The
tube holds the low pressure environment and protects the pods from
external conditions.

69
Appendices

70
Table of Contents
A Environmental Analysis 72

B Levitation Design Criteria Description 73

C Levitation Post-Chapter Remarks 74

D Technology Readiness Level 75

E Pod Characteristics Methods 76

F Tube Characteristics Tables and Calculations 78

G Vacuum Analysis Background 80

H Network Analysis 81

I Safety Hazard Analysis 82

71
A.Environmental Analysis
This appendix gives an environmental analysis of a hyperloop system, used to determine the boundaries for the levitation
and propulsion subsystems (Chapter 3 and 4 respectively).

The environment is determined by parameters such as:

• Pressure; the pod goes through pressure cycles since the pod has to change from the tube pressure environment to
the atmospheric pressure at the station.
• Temperature; in the tube there can be high temperature fluctuations, the hyperloop should be compatible with the
most extreme values. The tube should be able to handle the compression and expansion as the result of temperature
fluctuations.
• Accelerations; the subsystems could be subject to high acceleration forces.
• Weather conditions; the outside of the tube is subject to all weather conditions. It should be able to operate in rain,
under snow, wind and be equipped with lightning arresters.

The limits of top-level environmental parameters are represented in the enumeration below.

I The system will be exposed to pressures ranging from 0.01 mbar to 1.05 bar .
II The system will be exposed to accelerations ranging from 0 g 1 to 4 g , the latter is based on transportation without
passengers. Bumps in the track could cause higher accelerations but only for a short amount of time, therefore they
are considered not relevant for this study.
III The system will be exposed to temperatures ranging from −20 °C to 100 °C . This corresponds to 253.15 K to 373.15
K;
IV The pod mass will be around 1500 kg per meter.
V The cruising speed will be approximately 300 m/s, which corresponds to 1080 km/h [Musk, 2013] (top-level system
requirement).

1 1g = 9.81m/s 2

72
B.Levitation Design Criteria Description
This appendix describes the design criteria used to assess the potential levitation and propulsion technologies in Chapter
3 and 4.

• Reliability: focused on reducing the frequency of failures. The reliability shall be as high as possible. A part of
reliability engineering that can cause downtime of the system is the effort to perform maintenance. Reach-ability of
the levitation system, availability of spare parts and repair equipment is important to take into account.
• Power consumption: shall be low, if possible it should be lower than high speed rail and air passenger transport. The
power consumption is closely related to the vehicle mass, cruising speed and resistance forces (aerodynamic and
magnetic).
• Stability: shall be comparable to the stability for high speed rail, closely related to passenger comfort.
• Cost: the capital and operational cost shall be reduced as much as possible.
• Safety: focused on preserving lives and shall be high. Preferably, the safety indicators are similar or higher than
those of high speed rail and air passenger transport. Also take into account magnetic fields around and inside the
pod. A thorough safety analysis of the complete hyperloop system is given in Chapter 12.
• Complexity: indication of the ability to mass produce the system. Moreover, scalability is preferred.

73
C.Levitation Post-Chapter Remarks
This appendix provides post-chapter remarks of Chapter 3.

• Switches have not been extensively discussed in this chapter. The only levitation mechanism that hinders non-
mechanical switching is bottom-EMS.

• Fall-back mechanisms were not discussed in this chapter. What happens when the levitation system fails?
EMS: for bottom-EMS, skid-pads can be installed to land on if the levitation fails. For top-EMS, the pod will drop a
distance until it hits the bottom of the tube, resulting in a crash. A mechanism has to be develop to catch the pod in
case of a power outage.
EDS: the pod keeps levitating until it has reached speeds below the lift-off speed, after which the pod lands on its
wheels and can continue its low speed movement.
Air bearings: wheels have to be installed for the case a compressor fails and the air bearings disappear, on high
speeds this can lead to high impact on the passengers.
EDW: the lift force will quickly decrease and the electrodynamic wheels will hit the ground. This leads to high impact
on the passengers and a high risk for a crash is present.

• Compression and expansion of the guideway due to temperature fluctuations were not considered in this chapter.
However, this should definitely be taken into account when developing a guideway.

• The theoretical limit for the velocities to be achieved with a specific levitation mechanism was not discussed in the
chapter. This is because no reliable information was found about this and related tests have never been performed.

• Cost of the guideway were not shared quantitatively in this chapter since it was primarily focused on the functioning
of levitation systems. Often the cost are not represented realistically. E.g. if the guideway cost are high because a lot
of material has to be used, it could result in a safer guideway, reducing the cost required to ensure the safety of the
system.

74
D.Technology Readiness Level
Initially, Technology Readiness Levels (TRLs) were conceived for NASA space activities [Mankins, 1995]. The TRLs used in
this report are listed in Table D.1.

Table D.1: Technology Readiness Level Descriptions

TRL Description
1 Basic principles observed and reported
2 Technology concept and/or application formulated
3 Analytical and experimental critical function and/or characteristic proof-of-concept
4 Component and/or breadboard validation in laboratory environment
5 Component and/or breadboard validation in relevant environment
6 System/subsystem model or prototype demonstration in a relevant environment (ground or space)
7 System prototype demonstration in the eventual environment
8 Actual system completed and ‘flight/operation qualified’ through test and demonstration (ground or space)
9 Actual system ‘flight/operation proven’ through successful mission operations

75
E.Pod Characteristics Methods
This appendix provides the theoretical background for the sizing of the pod and the aerodynamic analysis that is used to
determine the blockage ratio and tube diameter in Chapter 5.

As for the inside-out approach, the pod will be sized first. The high speed aerodynamics play an important role in the
efficiency of the system which needs to be understood before continuing to size the tube.

E.1 Fuselage sizing

For aircraft, Equation E.1 is used to determine the amount of seats in one row. This is an empirical relation based on
historical data of aircraft fuselages. The number of seats abreast is highly dependent on the total number of passengers in
the plane. Whereas adding more seats abreast increases the total diameter of the fuselage, it will decrease the total length
of the fuselage.
p
n sa = 0.45 n pax (E.1)
In which n pax , the total amount of passengers is given. To calculate the inner diameter of the pod (the passenger com-
partment) equation E.2 is used. This is simply the sum of all the lateral components in the cross section of the fuselage.

d f ,i nner = n sa · w seat + (n sa + n ai sl e + 1) · w ar mr est + n ai sl e · w ai sl e + 2 · s cl ear ance (E.2)

To obtain the outer diameter, 20 cm for structural components including isolation is added based on aircraft design guide-
lines.

E.2 Aerodynamics
According to the drag equation shown in Equation E.3, the aerodynamic drag increases with the square of the velocity
in the subsonic (M<1) region.

1
D = ρV 2C D S (E.3)
2
Here ρ is the air density, S the frontal surface area of the pod and V the operating speed. The drag coefficient C D cap-
tures the aerodynamic effects related to the geometry and orientation of the object. This coefficient is either computed in
CFD (Computational Fluid Dynamics), measured with real world experiments, or estimated using statistical data. The low
speed drag coefficient of a hyperloop pod is conservatively estimated to be 3 based on trains which have a drag coefficient
of about 2.

The drag equation is only valid in the subsonic region as stated above. However, since the pod travels at close to sonic
speeds, the flow around the pod will accelerate and shock waves will eventually occur. An important parameter for
analysing the high speed aerodynamics is the blockage ratio, which is defined in Equation E.4.
1 2
A pod 4 πD pod D 2pod
β= = 1
= (E.4)
A t ube 4 πD t ube
2 D 2t ube

The area around a hyperloop pod act as contracting-expending nozzles which are scenarios that have been studied thor-
oughly for wind-tunnels and supersonic jet inlets. The flow around the pod accelerates as the bypass area decreases to a
maximum of M=1 at the throat (the location where the bypass area is minimum). The airflow will become choked if the
speed of the pod increases beyond this point. This causes a large increase in pressure drag and shock waves will start to
appear at the tail. This is known as the Kantrowitz limit, and is referred to in literature as the isentropic contraction limit.
Equation E.5[D. Van Wie & Walsh, 1996] is used to determine this limit.
¶ γ+1 µ ¶ γ+1
A t hr oat γ + 1 2(γ+1) γ − 1 2 − 2(γ−1)
µ
=M 1+ M (E.5)
A t ube 2 2

Where M is the Mach number which is defined as the velocity (V∞ ) divided by the local speed of sound. The local speed of
sound can easily be determined as a function of the temperature. γ is the ratio of specific heats, which always equals 1.4

76
Appendix E. Pod Characteristics Methods

for air. When using the isentropic flow relations for total temperature and total pressure, the drag for chocked flow can be
estimated. These results are presented in Figure E.1. Increasing the blockage ratio has two significant effects:
• The jump in drag due to the Kantrowitz limit occurs at a lower velocity;
• The total aerodynamic drag increases.

8
= 0.6
7 = 0.7
= 0.8

5
Pod drag [kN]

0
0 200 400 600 800 1000 1200
Pod Speed [km/h]

Figure E.1: Estimated aerodynamic drag for different blockage ratios

The aerodynamic drag before the flow becomes chocked is completely negligible due to the low pressure. However, the
increase in drag due to the chocked flow is significant. From an aerodynamic perspective, the blockage ratio should be
kept as small as possible. However, according to Equation E.4, a lower blockage ratio increases the tube diameter which
results in other challenges.

77
F.Tube Characteristics Tables and Calculations
In this appendix, supplemental information on the material properties are given as addition to Chapter 6. Furthermore,
calculations for the tube thickness and thermal expansion are explained.

F.1 Material Properties

In Table F.1 the main properties of the materials are summarised. The properties listed here might differ slightly from
other stated numbers, as the materials are not perfect. Fibre reinforced composites are left out, since the properties highly
depend on the design and the techniques used. Concrete and steel type are chosen because of their wide and frequent
application.

Table F.1: Material Properties

Concrete (C25/30) Steel (S235) Aluminium Acrylic

Density (ρ) 2,400 kg /m 3 7,860 kg /m 3 2,700 kg /m 3 1,180 kg /m 3
Thermal expansion 10 · 10−6 K −1 14 · 10−6 K −1 24 · 10−6 K −1 75 · 10−6 K −1
coefficient (αL )
Young’s Modulus (E) 30 GPa 200 GPa 70 GPa 3 GPa
Yield stress 16.67 MPa (compression) 235 MPa 100-600 MPa 70 MPa
(Dependent on alloy
type)
Price €100 /m 3 €157/m 3 Fluctuates, but higher Highly depends on process-
than concrete and steel ing, but more expensive than
concrete and steel

F.2 Tube Thickness Calculation

The vacuum buckling criteria can be seen in Equation F.1. [Hauviller, 2007]
µ ¶3
E t
p cr = (F.1)
4(1 − ν ) R
2

The Young’s Modulus (E ) for steel used is 200 ·109 P a, 1 −ν2 is taken to be 0.9 (which is generally true for metals according
to Hauviller [2007]), R (radius) is 1.75 m and p cr is the buckling pressure difference on the tube. This pressure is the
atmospheric pressure of 1 AT M or 101325 P a reduced by the tube pressure of 3 P a.

Equation F.1 is plotted in Figure F.1 to show the thickness on the vertical axis and the tube radius on the horizontal axis.
From this, it follows that the thickness t should be 21.4 mm to prevent buckling due to vacuum. The design of stiffening
elements is too detailed for this stage of hyperloop development and is out of the scope of this research.

The value found for the thickness is theoretically found without applying any form of safety. To build in safety, a safety
factor of 1.5 is applied to the pressure difference. This results in a final tube thickness of 25 mm.

Whether the assumption of the thin-walled approach is valid, should be checked.

t 0.025 1
= = 0.007 <
D 3.5 20
The requirement is met and thus the assumption is justified.

F.3 Thermal Expansion

Equation F.2 is a basic equation used to calculate the elongation under the influence of a temperature difference.

∆L = αL ∆T L (F.2)
In which ∆L is the expansion in [m], αL is the material coefficient (for steel, αL = 12 · 10−6 K −1 ), ∆T is the change in
temperature and L is the original length of the object.

78
Appendix F. Tube Characteristics Tables and Calculations

0.035 Tube wall thickness versus tube diameter

0.030

0.025
Tube wall thickness t [m]

0.020

0.015

0.010

0.005

0.0000 1 2 3 4 5
Tube diameter D [m]

Figure F.1: Tube wall thickness plotted against tube radius for a steel tube

For the goal of this calculation it is not required to achieve great precision. Therefore, the maximum temperature fluc-
tuation (∆T ) in The Netherlands is taken. The maximum temperature difference since 1942 is 66 °C 1 . The maximum
elongation is the length difference between the tube length at lowest and highest temperature.

1 Weather records - http://www.meteolink.nl/weer-records-nederland-en-de-wereld/, consulted on 31 May, 2019

79
G.Vacuum Analysis Background
This appendix covers the physics involved with vacuum pumps including the governing equations. These equations were
used to determine the optimum operational tube pressure and to analyse the vacuum installation for a hyperloop system
in Chapter 7.

The pump capacity relates to the total volume, time to pump and pressure difference as Pfeiffer Vacuum GmbH [2013]:

V p1
µ ¶
S = ln (G.1)
t p2

Where S is the pump capacity (in m 3 /h), V is the total volume (in m 3 ) that one pump has to pump to a lower pressure
level. t is the time (in h) it takes the vacuum pump to go from a pressure level p 1 to p 2 . p 2 is the final (desired) pressure
level (in mbar ).

The second governing equation is the specific heat equation given as [Pfeiffer Vacuum GmbH, 2013]:

Q = c · m · ∆T (G.2)

With Q the added heat (in J ), m the mass of the molecules (in kg ), c is the specific heat (in J /kg ) and ∆T indicates the
change in temperature (in K ). The power consumption of the pump is the pump capacity S multiplied by the pressure
difference ∆p at which the pump is operatingPfeiffer Vacuum GmbH [2013]:

P = ∆p · S (G.3)

The required pump capacity for water and air molecules at a certain temperature and pressure is given byPfeiffer Vacuum
GmbH [2013]:
h i
ma mW
Ma + MW · R · T
S= (G.4)
p
With R being the specific gas constant, T the temperature (in K ) and p the pressure (in P a), m a
M a the fraction of air mass
and the molar mass of air, and m w
M w the fraction of water vapour mass and the molar mass of water. Since the molar mass of
water is lower than that of air, the fraction of the water mass m W
MW is much more sensitive. Therefore if the amount of water
molecules increases, the required pump capacity increases faster than if it was only the air molecules. It is crucial for the
operation of the vacuum pumps that the amount of water particles is kept as low as possible.

The power required to overcome the aerodynamic drag follows from:

P r = D aer o · V (G.5)

In which P r is the power, D aer o is the aerodynamic drag, and V the operational velocity. The aerodynamic drag is deter-
mined in Appendix E and can be estimated for velocities above the Kantrowitz limit as:

21.65 · (V − 100) + 1777

D aer o = ·P (G.6)
100
In which P equals the operational power. Using Equations G.5 and G.6, the power for one pod to overcome aerodynamic
is determined. It should be noted that this is not the complete power required per pod since this does not include the
magnetic drag which is a result from the magnetic levitation.

80
H.Network Analysis
As a result of the heuristic Network optimisation, the network design of Chapter 11 consists of 48 stations and 51 links.
Table H.1 gives an overview of the expected amount of passengers travelling over each link.

Table H.1: Link overview European Hyperloop Network

Total Link Pax Total Link Pax

Bidirectional Links Bidirectional Links
Distance [km] [millions] Distance [km] [millions]
Paris Frankfurt 536 50.5 Manchester Dublin 318 12.4
London Paris 455 45.0 Madrid Porto 526 10.7
Frankfurt Berlin 520 44.5 Moscow St Petersburg 762 10.6
Paris Toulouse 727 42.1 Vilnius Warsaw 476 10.5
Barcelona Toulouse 305 40.8 Prague Vienna 334 10.0
Frankfurt Munich 360 39.4 Riga Vilnius 321 10.0
Vienna Budapest 257 33.0 Tallinn Riga 338 9.8
Munich Vienna 426 29.9 Porto Lisbon 332 8.9
Barcelona Madrid 579 29.7 Tallinn St Petersburg 380 8.9
Budapest Belgrad 384 28.1 Copenhagen Gothenburg 274 8.4
Belgrad Sofia 393 27.6 Nice Toulouse 565 8.0
Berlin Hamburg 318 27.1 Edinburgh Manchester 357 7.7
Zurich Frankfurt 342 26.6 Antalya Izmir 428 7.5
Frankfurt Dusseldorf 226 24.3 Madrid Malaga 519 7.5
Hamburg Copenhagen 336 22.4 Stockholm Helsinki 478 7.4
Manchester London 291 22.4 Milan Nice 301 7.2
Zurich Milan 245 20.7 Athens Thessaloniki 359 7.0
Amsterdam Dusseldorf 214 18.8 Sofia Thessaloniki 276 6.8
Sofia Istanbul 584 16.7 Gothenburg Oslo 347 6.2
Brussels Amsterdam 189 15.1 Ankara Istanbul 439 5.4
Berlin Warsaw 628 14.7 Sofia Bucharest 354 4.0
Copenhagen Stockholm 656 13.8 Rome Naples 238 3.1
Izmir Istanbul 397 13.8 Glasgow Edinburgh 80 2.8
Milan Rome 613 13.5 Belfast Dublin 169 2.4
Prague Berlin 338 13.2 Helsinki Tallinn 121 1.9
Paris Brussels 302 13.2

81
I.Safety Hazard Analysis
In this appendix, the elaborate Hazard Analysis is provided as an addition to the safety analysis of Chapter 12. For each
subsystem, the most important hazards are listed in Tables I.2 to I.14. For each hazard, the cause and its consequences
are determined. Hazard mitigation methods are decided upon to decrease the likelihood and severeness of hazards. Af-
terwards, a 1 to 5 score is determined for the likelihood and severeness of each hazard, based on Table I.1. The number of
casualties is based on the assumptions that all pods have a capacity of 50 passengers and all pods are 50% occupied. For
example, if a hazard causes one single pod to crash, the score for severeness is 3.

The risk score is determined by multiplication of the likelihood and severeness. Hazards with a risk score of 8 or higher
are identified as the highest risks and indicated in red. Furthermore, each hazard with a severeness of 5 is also identified
as a high risk and therefore indicated in red.

Table I.1: Score description for likelihood and severeness

Score Likelihood Severeness

1 Occurs once per 50 year or less No casualties, only injuries
2 Occurs once per 10 year <5 casualties
3 Occurs once per year 5-25 casualties
4 Occurs once per month 26-50 casualties
5 Occurs once per week or more frequently 51+ casualties

82
 Appendix I. Safety Hazard Analysis
Table I.2: Communication Safety Hazards

Cause Hazard Description Consequences Hazard Mitigation Likelihood Severeness Risk score
Location sensors in tube Unreliable pod location Pod crashes into other pod Redundant communica- 2 4 8
fails or airlock door tion system
Signal of location sensor to No data of pod location Pod crashes into other pod Redundant communica- 2 4 8
operational centre broken or airlock door tion system
Power outage in tube No data of pod location Pod crashes into other pod Redundant power supply 2 4 8
or airlock door
Power outage in pod On-board communication Passengers cannot report Redundant power supply 3 2 6
system fails dangerous situations inside
pod
Electromagnetic interfer- Unreliable pod location Pod crashes into other pod Design communication 1 4 4
ence or airlock door system to withstand EMI
Radio interference Unreliable pod location Pod crashes into other pod Use multiple frequencies 2 4 8
or airlock door
Pod ID damaged Pod ID cannot be recog- Pod routing unknown Redundancy in pod ID & 3 1 3
nised leading to unreliable pod ID maintenance and
pod data monitoring

Table I.3: Pod Structure Safety Hazards

Cause Hazard Description Consequences Hazard Mitigation Likelihood Severeness Risk score
Corrosion of the pod Leak in pressure vessel Passengers get exposed to Design pod with safety fac- 1 3 3
vacuum environment tor & pod maintenance and
monitoring
Puncture by object Leak in pressure vessel Passengers get exposed to Design pod with safety fac- 2 3 6
vacuum environment tor & track monitoring
Excessive loading of the Crack in pressure vessel Passengers get exposed to Design pod with safety fac- 1 3 3
pod vacuum environment tor & pod maintenance and
monitoring
Excessive loading of the Excessive bending of the Track and pod damaged, Design pod with safety fac- 1 3 3
pod pod, leading to pod touch- passengers get exposed to tor & pod maintenance and
ing the track vacuum environment monitoring
Deterioration of pod door Pod door does not seal Passengers get exposed to Design pod with safety fac- 3 3 9
properly vacuum environment tor & pod maintenance and
monitoring
83
 Appendix I. Safety Hazard Analysis
Table I.4: Tube Structure Safety Hazards

Cause Hazard Description Consequences Hazard Mitigation Likelihood Severeness Risk score
Excessive loading of tube Serviceable Limit State Passenger comfort be- Design tube with safety fac- 2 1 2
bending of the tube comes worse tor & tube maintenance
and monitoring
Excessive loading of tube Ultimate Limit State bend- Tube permanently deforms Design tube with safety fac- 1 5 5
ing of the tube tor & tube maintenance
and monitoring
Corrosion of the tube Leak in tube Airflow into the tube Design tube with safety fac- 2 1 2
tor & tube maintenance
and monitoring
Puncture by object Leak in tube Airflow into the tube Design tube with safety fac- 3 1 3
tor
Excessive loading of tube Weld or bolt connection Tube deformation and Design tube with safety fac- 1 5 5
connections failure track misalignment tor & tube maintenance
and monitoring
Soil settlement Settlement of the tube Tube deformation and Design tube to withstand 1 5 5
track misalignment soil settlement
Track misalignment Collision pod with track Track and pod damaged, Track maintenance and 2 3 6
passengers get exposed to monitoring
vacuum environment
84
 Appendix I. Safety Hazard Analysis
Table I.5: Vacuum Pumps Safety Hazards

Cause Hazard Description Consequences Hazard Mitigation Likelihood Severeness Risk score
Deterioration of vacuum Pump failure at tube Pressure in the tube in- Redundancy in number of 3 1 3
pump creases a few Pascals, more vacuum pumps & pump
drag for pods maintenance and monitor-
ing
Vacuum pump overheating Pump explosion at tube Tube deformation and toxic Pump maintenance and 1 3 3
gasses flowing into the tube monitoring
Deterioration of vacuum Pump failure at airlock More airflow from airlock Redundancy in number of 3 1 3
pump into the tube, more drag for vacuum pumps & pump
pods maintenance and monitor-
ing
Vacuum pump overheating Pump explosion at airlock Both airlock doors dam- Pump maintenance and 1 4 4
aged, leading to more air- monitoring
flow from station into the
tube, toxic gasses at the sta-
tion
Complete power outage No working pumps Pressure in the tube in- Redundant power supply 2 1 2
creases to atmospheric
pressure, more drag for
pods
85
 Appendix I. Safety Hazard Analysis
Table I.6: Airlocks Safety Hazards

Cause Hazard Description Consequences Hazard Mitigation Likelihood Severeness Risk score
Mechanical door failure Door unable to close (sta- Pod cannot leave the sta- Only one door can be 2 1 2
tion side) tion opened at the same time &
airlock door maintenance
and monitoring & multiple
airlocks at station
Mechanical door failure Door unable to close (tube Pod cannot enter the sta- Only one door can be 2 1 2
side) tion opened at the same time &
airlock door maintenance
and monitoring & multiple
airlocks at station
Mechanical door failure Door unable to open (sta- Pod cannot enter the sta- Only one door can be 2 1 2
tion side) tion opened at the same time &
airlock door maintenance
and monitoring & multiple
airlocks at station
Mechanical door failure Door unable to open (tube Pod cannot leave the sta- Only one door can be 2 1 2
side) tion opened at the same time &
airlock door maintenance
and monitoring & multiple
airlocks at station

Table I.7: Propulsion Safety Hazards

Cause Hazard Description Consequences Hazard Mitigation Likelihood Severeness Risk score
Power outage in LSM LSM out of service Pod comes to standstill in- Redundant power supply & 2 3 6
side the tube back-up propulsion system
Communication failure be- Pod is not propelled prop- Pods travel at different Redundant communica- 3 3 9
tween LSM and pod erly speeds which might lead to tion system
a collision
Overcharging LSM Destroying part of the LSM Pod temporarily deceler- Design to withstand power 3 1 3
ates abundance & segmentation
of LSM & Safety margin in
LSM design
LSM failure during acceler- No sufficient speed at the Not able to merge safely Monitoring acceleration 2 3 6
86

ation end of an on-ramp in the pod flow after high- trajectory & safety margin
speed switch in the length of the on-
ramp LSM & Segmentation
of on-ramp LSM
 Appendix I. Safety Hazard Analysis
Table I.8: Levitation Safety Hazards

Cause Hazard Description Consequences Hazard Mitigation Likelihood Severeness Risk score
Track misalignment Levitation magnets collid- Track and pod damaged, Track maintenance and 1 3 3
ing with bump passengers get exposed to monitoring
vacuum environment
Magnets decaying over Lower lift forces leading to Higher chance of pod hit- Design levitation systeem 3 1 3
time smaller gap heights ting the track with safety factor & EDS
system maintenance and
monitoring
Pod too heavy Smaller gap height Higher chance of pod hit- Use safety factor for full 3 1 3
ting the track pod weight & monitoring
pod weight
Deterioration of secondary Secondary suspension fail- More vibrations inside the Secondary suspension 2 1 2
suspension ure pod, passenger discomfort maintenance and monitor-
ing

Table I.9: Braking Safety Hazards

Cause Hazard Description Consequences Hazard Mitigation Likelihood Severeness Risk score
Power outage Braking out of service Pod crashes into other pod Redundant power supply 2 4 8
or airlock door
Communication failure be- Braking too early Coming to standstill in the Redundant communica- 2 2 4
tween LSM and pod tube tion system & secondary
propulsion system
Communication failure be- Braking too late Pod crashes into other pod Redundant communica- 2 3 6
tween LSM and pod or airlock door tion system

Table I.10: Wheels Safety Hazards

Cause Hazard Description Consequences Hazard Mitigation Likelihood Severeness Risk score
Mechanical or electrical re- Wheels do not retract Higher chance of wheels Wheel maintenance and 2 2 4
tractable wheel failure hitting the track monitoring
Mechanical or electrical re- Wheels do not deploy Pod hitting the track at low Wheel maintenance and 2 1 2
tractable wheel failure speed (±20 km/h) monitoring
87
 Appendix I. Safety Hazard Analysis
Table I.11: External Safety Hazards

Cause Hazard Description Consequences Hazard Mitigation Likelihood Severeness Risk score
Plate tectonics Light earthquake <5.0 Tube and pillars deforma- Design tube and pillars 2 3 6
Richter tion with safety factor & inspect
earthquake prone areas
Plate tectonics Heavy earthquake ≥ 5.0 Tube and pillars deforma- Design tube and pillars 1 5 5
Richter tion with safety factor & inspect
earthquake prone areas
Lightning Damaged electrical sys- System non-operational Install lightning rods 1 1 1
tems and power outage
Heavy storms Tree falling on the tube Tube deformation and so- Design tube with safety fac- 1 4 4
lar panels broken tor & remove large trees
close to infrastructure
Flood More soil settlements and Tube deformation and Design pillars with safety 1 5 5
erosion of pillars track misalignment factor
Extreme frost Tube shrinking Tube deformation Design tube for extreme 1 5 5
weather situations
Exreme heat Tube expanding Tube deformation Design tube for extreme 1 5 5
weather situations

Table I.12: Pod Equipment Safety Hazards

Cause Hazard Description Consequences Hazard Mitigation Likelihood Severeness Risk score
Oxygen tank failure Oxygen leak in pod Passengers experience oxy- Design oxygen tank with 2 2 4
gen toxicity safety factor & oxygen
masks in pod
Oxygen tank failure Oxygen tank explosion and Pod deformation, passen- Design oxygen tank with 1 3 3
fire in the pod gers in danger safety factor & oxygen
masks in pod
Oxygen tank failure No oxygen supply Passengers cannot breath Design oxygen tank with 2 2 4
safety factor & redundant
oxygen tanks & oxygen
masks in pod
Short circuiting of electron- Fire in the pod Smoke development, pas- Incorporate fuses in elec- 2 3 6
ics sengers in danger tronics design
Battery failure Battery explosion and fire Pod deformation, passen- Battery maintenance and 2 3 6
88

in the pod gers in danger monitoring

Deterioration of HVAC sys- HVAC system not working Passengers get exposed to HVAC maintenance and 2 2 4
tem properly toxic gasses monitoring & oxygen
masks
 Appendix I. Safety Hazard Analysis
Table I.13: Human Error Safety Hazards

Cause Hazard Description Consequences Hazard Mitigation Likelihood Severeness Risk score
Obstacle in tube after Collision pod and obstacle Passengers get exposed to Design pod with safety fac- 2 3 6
maintenance leading to pod deformation vacuum environment tor & track monitoring
Bad maintenance of pod Crucial part of pod failing Pod crashes, passengers get Maintenance protocols & 3 3 9
exposed to vacuum envi- pod monitoring
ronment
Bad maintenance of tube Crucial part of tube failing Tube deformation and Maintenance protocols & 2 4 8
track misalignment tube monitoring
Bad maintenance of air- Crucial part of airlock fail- Pod cannot enter or leave Maintenance protocols & 2 1 2
locks ing the station airlock door maintenance
and monitoring & multiple
airlocks at station
Bad maintenance of vac- Crucial part of vacuum Pressure in the tube in- Maintenance protocols & 2 1 2
uum pumps pumps failing creases a few Pascals, more vacuum pump monitoring
drag for pods
Bad maintenance of LSM Crucial part of LSM failing Pod temporarily deceler- Maintenance protocols & 2 1 2
ates LSM monitoring & segmen-
tation of LSM & Safety mar-
gin in LSM design
People trespassing Collision pod with human Human injury Installation of obstacles to 3 2 6
prevent people to access
the track

Table I.14: High-speed Switches Safety Hazards

Cause Hazard Description Consequences Hazard Mitigation Likelihood Severeness Risk score
Magnets decaying over Active lateral guidance Pod crashes into switch Design high-speed 2 3 6
time not generating sufficient switches with safety factor
pulling force & magnets maintenance
and monitoring
Communication failure be- Active lateral guidance acti- Pod crashes into switch Redundant communica- 2 3 6
tween guidance and pod vating too late tion system & safety factor
in switch length
Power outage Active lateral guidance not Pod not able to make the Redundant power supply 2 1 2
activating switch & collision avoiding switch
89

design