Digital twin

TOWARDS A MEANINGFUL FRAMEWORK

About Digital

Arup is committed to shaping the digital built environment. Our clients look to us to
combine deep domain knowledge and new technologies with strategies that sharpen
performance, uncover new business models, and ensure that they thrive in this new,
connected world. We seek collaborations with research partners in our mission to
shape a better world.

About Foresight, Research and Innovation

Foresight, Research and Innovation is Arup’s internal think-tank and consultancy

which focuses on the future of the built environment and society at large. We help
organisations understand trends, explore new ideas and radically rethink the future
of their businesses. Research has played a fundamental role in defining how we
anticipate and leverage emerging business opportunities. Applied research continues
to underpin our ability to address our clients’ greatest challenges.

Contact us at [email protected]

Download the report at www.arup.com/digitaltwinreport

Released November 2019

Cover Image: Arup / Florian Wehde on Unsplash

13 Fitzroy Street
London
W1T 4BQ
www.arup.com
© Arup 2019
Contents

Foreword 6

1. Current state of digital twins 9

Towards an evaluation framework 21

2. Value of digital twins 33

Cities 49

Energy 65

Property 85

Transport 111

Water 133

3. Challenges and conclusions 149

References 155

Contributors and acknowledgements 158

Features
74
C A S E S T U D Y:
TOKYO
The IoT Office —
enhancing office
well-being
FOREWORD

To fulfil its potential, the concept of digital twin(s) has to

evolve into a real and meaningful construct and not
just another idea on the never-ending digital hype cycle.

A digital twin, as a means to link digital models this research can inform and influence, so that we
and simulations with real-world data, creates new carefully consider the purpose of digital twins,
possibilities for improved creativity, competitive human factors, integrity and security. There is a
advantage and human-centred design. Digital twins clear need to collaboratively define and lead the
can help deliver on the grand challenges facing development of a digital twin narrative that works
society, including achieving the United Nations’ across the built environment.
Sustainable Development Goals and addressing
rapid urbanisation, population growth, and This report was developed through collaboration
escalating infrastructure costs. with global experts and their use cases, and we
are grateful for their ideas and contributions. Our
In their current state, digital twins can improve our research included both enthusiasts and critics
designs and the advice we give to our clients, yet in our aim to explore how we can best combine
challenges remain. Sharing data comes at a high human and machine intelligence, creativity, and
cost and with great tension, as digital knowledge, technologies into digital twins that can help our
practices and culture are not yet converging across mission to shape a better world.
the built environment. To realise the full value of
our work, we must develop partnerships through Our aim in sharing this work is to bring the
deep understanding and greater collaboration industry together into an important journey.
around a meaningful approach: an inspiring vision A journey towards meaningful developments of
supported by clear actions. digital twins, which drive value for our clients and
society as a whole. As we continue to explore and
In this report, we present the current state of digital collaborate, we encourage you to stay in touch
twins in the built environment and explore ideas with us via [email protected]. We welcome your
on what might become possible in the future. The feedback and interest in creating digital twins of
research examines a broad spectrum of digital today and for the future. ■
twins, both in use and in development, and finds
varying levels of progress and investment in early
prototypes, practice and governance. We hope that

6
VOLKER BUSCHER
Chief Data Officer
Arup

7
8 CHAPTER 1
CHAPTER 1

Current state of digital twins

Will digital twins be the technology that can help tackle
climate change, radically modernise water and transport
infrastructure, and help us achieve the United Nations’
Sustainable Development Goals?

9
INTRODUCTION C U R R E N T S TAT E O F D I G I TA L T W I N S

The digital twin concept offers great potential value The digital twin is becoming a widely familiar
for organisations in the built environment. The concept. In 2003, Michael Grieves coined the term
widespread use of digital twins holds the promise to ‘digital twin’ as part of his research into product
increase operational efficiency, allow for resource lifecycle management.1 On page 24, Dr Grieves
optimisation, improve asset management, deliver sheds light on the evolution of digital twins for the
cost savings, improve productivity and safety. The built environment. General Electric, Siemens and
digitisation of the built environment, enabled by Rolls Royce were designing rotors, turbines and
an increase in computing power, cheaper sensors, engines with the aid of simulations decades before
Internet of Things (IoT), advanced analytics and the term was coined. Similarly, the oil and gas
greater sophistication of 3D visualisation and industries have been working with simulations of
immersive environments, therefore has the potential fuel reservoirs since the 1980s. Prior to Grieves’
to actively contribute towards achieving the UN coinage, industry used terms as diverse as ‘digital
Sustainable Development Goals (SDGs). shadows’, ‘digital avatars’ and ‘digital models’. As
we look at the history of the digital twin, we must
Yet, while digital twins are clearly in vogue, we have make an important distinction between a digital
yet to define the extent to which they can drive value model, a simulation and a digital twin. A digital twin
for our clients. Digital twins have the potential to is not a static model but rather, a responsive system
disrupt markets, and once we address the barriers connected between the physical and digital systems.
of data exchanges, governance, and interoperability, In the following paragraphs, we provide examples
we expect the results to be profound. However, of how the industry is thinking and adopting digital
the industry is currently burdened by competing twins across the industry.
definitions, misunderstandings, and confusion
around what a digital twin is and what it can do. The At DHL, an international courier service, the
aim of this report is to align terminology, and to employment of digital twins is underway, and the
provide a working evaluation framework to clearly firm is exploring the challenges and opportunities
categorise the various levels of sophistication of of integrating them into their core operations.2 In
digital twins. automotive manufacturing, Siemens is creating
digital twin products to enable realistic simulations
The report was informed by interviews with 75 to optimise a car before it has been built.3
experts from across the industry, including thought Meanwhile, software companies such as SAP
leaders within key markets; cities, energy, property, and Microsoft are promoting the use of machine
transport and water. The report includes exemplary learning and IoT to help companies optimise their
case studies of emerging uses of digital twins in the operations.4,5 IBM Watson, on the other hand,
built environment, which are evaluated using our are providing services towards product lifecycle
working framework. improvements, and creating deeper insights between
design and operations.6 ►

10 CHAPTER 1
11
12 CHAPTER 1
The software industry is creating products to learning element, making decisions on behalf of its
integrate, analyse and manage high-throughput users with the aim of providing a more comfortable
data, targeted at smart cities, product manufacturing environment.9 A larger-scale example comes from
and asset management. At the city scale, the Smart Google DeepMind, who trained AI systems to
City Digital Twin paradigm has been introduced to manage the cooling of their data centres. The system
increase the transparency of human-infrastructure- uses reinforcement learning, determining by trial
technology interactions through the exchange of and error which cooling configurations consume
spatiotemporal information.7 At the national scale, the the lowest amounts of energy. The system was able
British government is looking into model federation to consistently achieve a 40% reduction in energy
and data sharing.8 Recent advances in artificial used for cooling, which equalled a 15% reduction in
intelligence (AI) have enabled better insight and overall power usage.10
analysis of physical assets, yielding a rich dataset of
previously inaccessible information. Advanced digital Self-driving technology is one of the most advanced
twins can be found in the property and transportation examples of applied AI twin in society today, with
markets, with some of the most prominent examples an important aim of reducing accidents caused by
coming from tech giants like Google. human error. Consortium partners such as the UK
Autodrive programme are collaborating to advance
As a first, small-scale example, we can take the this technology.11 Google Maps is another good
Google Nest learning thermostat. This device, example.12 While it may have drawbacks such as it
when paired with heating systems, has the ability to is not always 100% accurate and requires constant
programme itself by switching the system on and off updating, the maps is still best characterised as
or choosing between heating and air conditioning an advanced digital twin, providing a relevant
based on the temperature it detects in the room. The representation of the built environment and real-
Nest learns behaviour by understanding the patterns time information. The Google Maps API is now
and desired temperatures for certain days and times an essential substrate for many services such
during the week, and then creates a schedule for as ridesharing, restaurant reviews and business
heating, ventilation and air conditioning (HVAC). It information, resulting in a comprehensive and
is the first thermostat on the market to incorporate a unified platform of great value. ►

13
FIGURE 1
Below
Digital twins emerging
technology timeline

1990
Invention of
WWW by Tim
Berners Lee

1936 1970 1977 1990

First modern Apollo 13 Metretek, Augmented reality
computer by launched first smart coined by
Alan Turing by NASA meter Thomas P Candell

1930 1980 1990

1960 1981 1989 1992 1994

Simulations First IBM Virtual reality Smart First
personal coined by home smartphone
computer Jaron Lainer invented by IBM

At the city scale, the ambitious plan of Sidewalk Labs platform has now been expanded to include other
to build a state-of-the-art ‘responsive neighbourhood’ services such as Uber Eats (an online food ordering
is projected to draw £29.5b of private sector and delivery platform) and Marketplace Forecasting
investment by 2040.13 The project is aiming to (a spatiotemporal forecasting model able to predict
reimagine cities and improve quality of life. rider demand and driver availability).15,16 These are all
Among other things, the project will create a examples of technological advances bringing together
digital twin of a section of Toronto’s waterfront by the physical and digital worlds. They illustrate the
modelling and collecting data from a network of state-of-the-art in hybrid digital-physical systems with
sensors in public spaces. In addition, the project is high fidelity, increasing autonomy, and adaptability to
developing an underground freight delivery system changing environmental and operational conditions.
and tracking the delivery of packages, to avoid Given the advances we have seen in other markets,
clogging the streets with trucks. we believe the built environment is ready to
embrace digital twins, with new collaborations and
Lastly, consider Uber’s on-demand transportation partnerships in order to create new opportunities and
service, which uses a machine learning platform, value propositions.
Michelangelo. This platform is a digital twin of
Uber’s entire network that connects drivers and users
based on supply and demand.14 The success of the

14 CHAPTER 1
 2005 2010 2014
Launch of Google Amazon Alexa
Google Maps announced its invented
self-driving car
development

2000 2010 2020

2000 2003 2009 2011 2016

Smart city Digital twin Smart lighting NASA Airforce Launch of Google
coined coined by invented papers of Home
Dr Michael Digital twin
Grieves

This report presents case studies from a broad their development, and the societal needs calling for
spectrum of digital twin applications. They range this development, can help us define a clear role for
from existing, fully functional twins, such as those digital twins beyond the current hype cycle. In the
intelligently optimising the use of office space (as following section, we present the various definitions
highlighted in the Property market), to more future- of digital twins across the industry and present the
oriented projects (such as the Digital Patient case digital twin and its ecosystem around better decision
study on Page 62). Healthcare is forecasted to be a making, learning and autonomy. ►
fast-moving market, with a new focus on personalised
patient care and designs to provide better services
to patients including rethinking space and waiting
areas within hospitals.17 Progress made and lessons
learned in the redesign of healthcare facilities will be
instructive for the built environment.

The purposes and intentions behind the deployment

of digital twins found in this report vary across
industries and use cases; there is no one-size-fits-all
for digital twins. The technological trends enabling

15
D E F I N I T I O N S O F T H E D I G I TA L T W I N

A key motivation for our work has been to help the

industry compare the various definitions of digital
A digital twin is the
twins that includes the promises and value yet to combination of a computational
be achieved. Table 1 lays out existing definitions
of the digital twin from key players in academia model and a real-world system,
and industry. Differences in the characterisations
of digital twins are relatively minor, and tend to
designed to monitor, control
differ in focus (for example, the entity twinned as and optimise its functionality.
an ‘object’, a ‘product’, a ‘process’, a ‘system’ or
some mixture of these). While the characteristics Through data and feedback,
are relatively consistent, the stated purposes of
digital twins can vary widely. What is perceived
both simulated and real, a
as the main purpose is likely to vary depending on digital twin can develop
the interests of the researcher or business involved,
from increasing return on investment (Deloitte) to capacities for autonomy and to
ensuring safety (NASA).
learn from and reason about
Digital twins at first may appear to be an exact its environment.
replica, however, digital twins are not necessarily
realistic representations, but are rather relevant
abstractions of the physical asset. It is not necessary As the technology matures, so does the terminology,
for even the most advanced smart city to digitally hence the need to include autonomy, learning and
replicate the fluting of every column or the mortar reasoning within the definition. The proposed
of every brick. In other words, we need to develop definition acknowledges that the digital twin is still
digital twins that are fit for purpose, and the level a model, but it has the potential to evolve into an
of fidelity will vary depending on the primary use autonomous system with less human intervention,
cases. Digital twins need not attempt to mirror through AI-enabled design and control. Among the
everything about the original system.18 competing definitions, we agree with IBM’s definition
and the ability for the digital twin to “understand,
We can make the broad definitional consensus more learn and reason”. Over time, a digital twin will not
precise by stating what a digital twin is not. Neither only lead to better decision making, but it will make
a Building Information Model (BIM), for example, better decisions. In the following section, we discuss
nor a simulation operating in isolation from a physical various types of digital twins and their ecosystems. ►
asset can be termed a digital twin. Rather, there must
be a connection between the physical and the digital
systems. This requires data exchange, as well as
the inclusion of humans in the role of designers or
occupants. Digital twins must make ever-improving
predictions about our physical infrastructure; in their
more advanced instantiations, at least, they will be
increasing in intelligence continuously. Overall, the
missing idea across definitions, and which we stress in
this report, is that the twin controls and optimises the
physical asset. In other words, existing definitions shy
away from recognising the key role of AI.

TA B L E 1
Opposite right
Definitions of 'digital twins'
across the industry and academia

16 CHAPTER 1
SOURCE DEFINITION

CAMBRIDGE CENTRE A digital twin is a realistic digital representation of something physical.19

F O R D I G I T A L B U I LT
B R I TA I N
Academia

A “Virtual Twin” is a virtual representation of what has been produced. We can compare a Virtual Twin to its
D A S S A U LT engineering design to better understand what was produced versus what was designed, tightening the loop
SYSTÈMES between design and execution.20
Software

A digital twin is a near-real-time digital image of a physical object or process that helps optimise
DELOITTE business performance.21
Consulting

A digital twin is a digital representation of a real-world entity or system. The implementation of a digital twin is
GARTNER an encapsulated software object or model that mirrors a unique physical object, process, organisation, person
IT or other abstraction. Data from multiple digital twins can be aggregated for a composite view across a number
of real-world entities, such as a power plant or a city, and their related processes.22

A digital twin is a living model that drives a business outcome.23

GENERAL ELECTRIC
Conglomerate

A digital twin is a virtual representation of a physical object or system across its lifecycle, using real-time data
IBM to enable understanding, learning and reasoning.24
Software

A digital twin is a mirror image of a physical process that is articulated alongside the process in question,
M I C H A E L B AT T Y usually matching exactly the operation of the physical process which takes place in real-time.18
Academia

The digital twin is a set of virtual information constructs that fully describes a potential or actual physical
MICHAEL GRIEVES manufactured product from the micro (atomic level) to the macro (geometrical level). At its optimum, any
Academia information that could be obtained from inspecting a physical manufactured product can be obtained from its
digital twin.25

A digital twin is a virtual model of a process, product, production asset or service. Sensor-enabled and IoT-
MICROSOFT connected machines and devices, combined with machine learning and advanced analytics, can be used to
Software view the device’s state in real-time. When combined with both 2D and 3D design information, a digital twin
can visualise the physical world and provide a method to simulate electronic, mechanical, and combined
system outcomes.26

A digital twin integrates ultra-high fidelity simulation with the vehicle’s on-board integrated vehicle health
NASA management system, maintenance history and all available historical and fleet data to mirror the life of its flying
Government / Research twin and enable unprecedented levels of safety and reliability.27

A digital twin is a virtual representation of a physical product or process, used to understand and predict the
SIEMENS physical counterpart’s performance characteristics.3
Conglomerate

17
 COMPONENT

ASSET

TYPES OF PROCESS
D I G I TA L T W I N

SYSTEM

NETWORK OF SYSTEMS

PHYSICAL WORLD D I G I TA L T W I N
A real-world asset A digital twin is the combination of a
and its behaviour. computational model and a real-world system.

USER INTELLIGENCE D I G I TA L T H R E A D
The user interacting with AI enabling the digital twin to perform tasks Refers to information
the digital twin. with minimal or no human oversight, e.g. channels connecting the
visual perception, speech recognition, natural physical and the
language translation, decision making. digital asset.

18 CHAPTER 1
T H E D I G I TA L T W I N A N D I T S E C O S Y S T E M

The digital twin and its ecosystem may vary in As we noted above, any digital twin is modelled at
scale and complexity with respect to size and some level of relevant abstraction from its physical
scope. Size is the granularity of the representation counterpart, geared to a specific purpose. This can
(how accurate it is, spatial and temporal scale, for be viewed as intentionally relinquished information,
example), and scope is the portion of the real-world to streamline a digital twin to its bare necessities.
that the model considers (i.e. an engine or a city). However, we must also be aware of unintentionally
Digital twins in the built environment range from relinquished information: we will always be limited
basic to complex, and are organised into five types by some level of measurement error, and almost
as shown in Figure 2. The digital twin can have always by some level of human error. It is important
single or multiple stakeholders, and may make use of to limit both of these sources of error to non-
3D simulations, IoT devices, 4G and 5G networks, intrusive levels.
blockchain, edge computing, cloud computing, and
artificial intelligence. Depending on its complexity, With further research and development of digital
each digital twin may have access to past, present twins, we expect to overcome many of these
and future operational data and increased predictive challenges and provide our clients with a solution
capabilities. A digital twin can be viewed as a system fit for purpose. In the following section, we present
within an ecosystem. For example, in Figure 3, the a working evaluation framework which aims to
ecosystem depicts both a physical asset and its digital establish a common language around what a digital
twin. The user interacts with the digital twin through is, what it can do and what it can become. ■
applied intelligence while the digital thread connects
the physical and digital worlds.

A particularly important function of digital twins is

their ability to understand, learn and provide value
using a systems approach. For example, a digital
twin of an electrical network must simulate the core
behaviour of the network — generation, storage
and consumption — together with human inputs
to the system, as well as the capacity to deal with
unexpected occurrences. Only the simultaneous
processing of these various kinds of data will enable
efficient control of the network. Over time, the
digital twin would improve its knowledge
of and control over the electricity network,
from predictive maintenance to the autonomous
suggestion of network improvements and even large-
scale redesigns.

FIGURE 2
Opposite top
Types of digital twin

FIGURE 3
Opposite bottom
Digital twin ecosystem

19
20 CHAPTER 1
D I G I TA L T W I N

Towards an evaluation framework

We put forward a framework to articulate what a digital twin is,
what it can do, and what it can become. Our framework intends
to facilitate wider collaboration and discussion of digital twins
across the industry.

The objective of the framework is to evaluate the learning, and fidelity providing value against a
current state of digital twins across five key levels. backdrop of minimal human intervention.
The levels help the industry to use common language
when describing a digital twin and its capabilities. While we considered other metrics that may
The aim of the framework is to enable participation be applicable such as maturity, we chose not to
from our clients and collaborators at all stages of the include it as a key metric. Maturity refers to the
development. In Figure 4 we present the evaluation developmental stage of the digital twin, rather than
framework. The metrics of the framework can be its level of complexity.28 The stages of maturity run
found in Table 2. The four metrics are: autonomy, from the initial concept, through demonstration
intelligence, learning, and fidelity. While these four and development, and finally to commercialisation.
metrics are conceptually correlated, they should be Digital twins can be highly intelligent and highly
treated independently; as the framework evolves and autonomous, but nevertheless yet to attain maturity.
our understanding of digital twins grows, we can Therefore, maturity was excluded as a key metric.
revise the framework and metrics accordingly.
In the following chapter, we apply the evaluation
In Table 3, the framework moves through five framework using the key metrics through a series of
levels, beginning with a simple digital model. As Arup case studies to demonstrate how it is applicable
the model evolves, feedback and prediction increase across digital twin projects.
in importance. At higher levels, machine learning
capacity, domain-generality and scaling potential all
come into play. By the highest levels, the twin is able
to reason and act autonomously, and to operate at a
network scale (incorporating lower-level twins, for
example). Our vision for the built environment is to
work collaboratively towards the development and
adoption of ‘level 5’ digital twins.

Our research has revealed that the progression and

development of digital twins is far from reaching
level 4 or 5. We are still a long way from an industry
landscape populated by reasoning models, machine
consciousness, and full autonomy. As digital twins
evolve, however, they will control more and more
operations, increasing autonomy, intelligence,

21
 FIGURE 4
Left
Arup's digital twins metrics
framework, in terms of their
AUTONOMY autonomy, intelligence, learning
and fidelity
INTELLIGENCE

LEARNING 1

FIDELITY

2
TA B L E 2
Opposite right
Working levels of digital twins

4
TA B L E 3

3 Below
Definitions of digital
twin metrics

METRIC DEFINITION

Autonomy The ability of a system to act without human input. There are five levels of autonomy. At level 1, there is complete
absence of autonomy, with the user controlling all aspects of the digital twin. A level 2 can be understood as
user-assisted. At this level, prompts and notifications of system activity are expected, but autonomy is limited. A
level 3 has partial autonomy, the twin has the ability to alert and to control the system in certain ways. A level 4 has
high autonomy, the digital twin is able to perform critical tasks and to monitor conditions with little to no human
intervention. Finally, a level 5 can operate safely in the total absence of human intervention.

Intelligence The ability of digital twins to replicate human cognitive processes and to perform tasks. There are five levels
of intelligence. At level 1, the twin has no intelligence. At level 2, the twin has reactive intelligence (the twin only
responds to stimuli, cannot use previously gained experiences to inform their present actions). At level 3, the twin
uses learning to improve its response and are also capable of learning from historical data to make decisions. At
level 4, the twin understands the needs of other intelligent systems. Finally, at level 5, the twin is self-aware with
human-like intelligence and self-awareness.

Learning The ability of a twin to automatically learn from data in order to improve performance without being explicitly
programmed to do so. Through machine learning, a twin classifies aspects of the systems (objects, behaviours)
using reinforced learning. There are five levels of learning. At level 1, the twin has no learning component. At level
2, the twin is programmed using a long list of commands. At level 3, the twin is trained using a supervised learning
approach (using labelled data able to provide feedback and prediction performance). At level 4, the twin is trained
using an unsupervised learning (the twin uses no labels and tries to make sense of the environment on its own). At
level 5, the twin uses reinforcement learning by interacting with its environment. With reinforcement learning, the
twin learns from past feedback and experiences to find the optimal way to improve performance; the twin uses a
reward system for good performance.

Fidelity The level of detail of a system, the degree to which measurements, calculations, or specifications approach the
true value or desired standard. There are five levels of fidelity. At level 1, the twin has low accuracy and can be
considered as a conceptual model. At level 2, the twin has a low to medium range of accuracy and can be used
to extract measurements. At level 3, the twin has a medium range of accuracy and can be used as a reliable
representation of the physical world. At level 4, the twin can provide precise measurements and at level 5, the twin
has a high degree of accuracy and can be used in the case of life safety and critical operational decisions. Fidelity,
therefore, depends crucially on the requirements of a given asset operator, rather than constituting an absolute
property of a digital twin.

22 CHAPTER 1
LEVELS

LEVEL 1

1
A digital model linked to the real-world system but lacking intelligence, learning or
autonomy; limited functionality e.g. a basic model of a map.

LEVEL 2

2
A digital model with some capacity for feedback and control, often limited to the modelling
of small-scale systems e.g. building temperature sensors which feed information back to a
human operator.

LEVEL 3

3
A digital model able to provide predictive maintenance, analytics and insights e.g.
predicting the life expectancy of rail infrastructure, enabling repairs or replacements before
asset failure.

LEVEL 4

4
A digital model with the capacity to learn efficiently from various sources of data, including
the surrounding environment. The model will have the ability to use that learning for
autonomous decision making within a given domain e.g. the model can automatically
communicate real-time route recommendations through various modalities (app, signage,
radio), allowing drivers to better plan their journey.

LEVEL 5

5
A digital model with a wider range of capacities and responsibilities, ultimately approaching
the ability to autonomously reason and to act on behalf of users (artificial general
intelligence). Intuitively, a level 5 model, such as a model of a neighbourhood in a smart
city, would take responsibility for the tasks one would presently expect a human operator
to manage, as well as to react to previously unseen scenarios. Another hallmark of
this level would be the interconnected incorporation of lower-level twins e.g. take the
level 4 example of traffic updates across a network. In a smart city scenario, numerous
independent systems work in parallel to provide feedback to a central decision making
network to deliver value to city-level leaders.

23
E V O L U T I O N O F T H E D I G I TA L T W I N F O R T H E
B U I LT E N V I R O N M E N T

The digital twin concept, from

its origin over a decade ago
to today, has always relied on
progress in two very different
areas: technology and culture.

24 CHAPTER 1
DR MICHAEL GRIEVES
Chief Scientist
Florida Institute of Technology

Advances in modeling and simulation allow us to

see into the future. We can predict traffic patterns,
energy usage, building stresses, fire risks, and other
resource and risk profiles. We can move a step ahead
in solving issues before they occur.

Ever since the crystallisation of the digital twin
concept at the beginning of the twenty-first century,
I have always thought that architecture and buildings
— throughout their lives — presented a natural
use case for the digital twin. The architecture of
buildings, with their dynamic visual elements, has
evolved rapidly from ideas into 2D sketches and
through to 3D physical models. This has given
architects an improved sense of how their designs
will be physically realised.

Previously, once the design was complete, building

designs were translated into fragmented 2D
blueprints to inform both the construction and
maintenance of the building. This resulted in wasted
resources, as information had to be translated from
2D to 3D. There was further waste in the building
operation and maintenance phase, since much of the
infrastructure was now hidden behind opaque walls
and barriers.

This issue is especially pertinent to the urban

buildings of today, which operate within complex
systems. Modern buildings must integrate a wide
variety of subsystems, such as heating, ventilation
and air conditioning (HVAC), energy and electrical
systems, plumbing and water movement, and ►

25
DR MICHAEL GRIEVES: PERSPECTIVE

I TA L T W I N
DIG

COMPUTING
C A PA B I L I T Y

SENSORS V I S U A L I S AT I O N

complex elevator systems. These all make modern the massive amounts of information used to operate
buildings a fruitful area for the employment of modern buildings in previously impossible ways.
digital twins. The second technological advancement has come
through sensors, which have enabled the rise of
The ability to realise digital twins, from their smart buildings. Buildings can now be made to
conceptual origin over a decade ago to where they sense and communicate their exact status to allow us
are today, has always relied on two very different to know and manipulate exactly what is occurring
kinds of progress: technological and cultural. within, simply by interrogating their digital twin.
We can, for instance, dictate temperature throughout
A D VA N C E M E N T S I N T E C H N O L O G Y E N A B L E the building, minimise energy use in unoccupied
D I G I TA L T W I N S
areas, keep tabs on structural stress levels, and adjust
lighting on demand. We can predict elevator failures
On the technology front, there are three major
and automate maintenance schedules by aggregating
advancements that have enabled the digital twin;
information from different buildings across time. In
namely computing capability, sensors, and visualisation.
short, we can predict future performance, and thus
pre-empt and eliminate issues before they occur.
Digital twins rely on massive amounts of data
collection and information processing by historical The third technology advancement has been in
standards. During the past decade, as computing visualisation. We can now render our geometric
power has more than doubled ever eighteen months designs virtually, in breathtaking fidelity, including
(Moore’s Law), we have moved from dealing with the previously unfeasible perspectival changes. We can
low-end gigabit scale (109) up to petabits (1015), exabits move from a bird’s-eye view of the entire building
(1018), and will soon transcend these by further orders and its surroundings to zoom in on the smallest detail
of magnitude.29 of a room. With the advance of virtual reality, we can
now experience digital buildings immersively. We
While Moore’s Law deals only with computing can not only see our designs but also walk through
capacity, other advancements have occurred in storage them, experiencing the building long before shovel
and communications bandwidth. This has afforded meets earth. With augmented reality we can merge
us the ability to process, store, and communicate the digital twin with its corresponding physical

26 CHAPTER 1
building, in order to look at the walls of a building I M P L E M E N T I N G D I G I TA L T W I N S I N T H E
B U I LT E N V I R O N M E N T
and instantly visualise the infrastructure behind
them. We can look at building components such as
HVAC and see exactly how they are performing — We have been developing digital twin-based
and where problems may be developing. We can then concepts in other industries over the past decade,
repair those problems before any failures occur. but more intensely over the past few years. Digital
twins have the advantage that, unlike their physical
R E A L I S AT I O N O F T H E D I G I TA L T W I N W I L L counterparts, they are not an all-or-nothing
R E Q U I R E C U LT U R A L C H A N G E
proposition. We can choose to collect information
Technology continues to advance, with no signs of only about features that have value, such as the
slowing. However, cultural change requires that we amount of fuel at an airport, the tyre wear of an
actively engage, and help people to move from what articulated lorry, or the temperature at a certain point
they have always been comfortable with into this in a power turbine. There is no need to digitally map
new environment. and monitor the entire asset.

Building industry practitioners have traditionally Industries drive digital twin evolution by exploiting
been trained using two-dimensional blueprints specific use cases. We have aeroplane engines that
and physical models, whilst working in siloed and predict future failures. Farmers look at digital twins
disciplined environments. Digital twins, however, of their agricultural equipment that know the exact
require that we move to three-dimensional models location of the equipment, and can direct units to
that are integrated across a range of disciplines. work together seamlessly. Factory machines schedule
These new technologies are highly disruptive, and their own maintenance, while oil rig manufacturers
as with all major disruptions, people are hesitant to use digital twins of their platforms to monitor drill
adopt new ways and need to be brought along. bit wear.

There is also a generational issue at work. Younger The built environment must discover where the
generations, so-called digital natives, have information furnished by digital twins can capture
never known a world without computers and value. Relevant use cases can be found throughout
digital technologies.30 They adapt quickly to new the entire lifecycle of design and architecture,
uses. The older generation are often not as familiar construction, and building operation. The challenge
and need more time to adjust. However, they have is to identify the gaps which technology can fill, and
experience that the younger generations do not then to iterate and integrate digital twin solutions to
possess. The strengths of each generation must be produce comprehensive and useful tools. ■
combined effectively.

FIGURE 5
Opposite left
The three major advancements
that have enabled the digital twin
in manufacturing

27
S O , W H AT I S N E W A B O U T D I G I TA L T W I N S ?

The two game changers to

develop digital twins are: first,
data and its availability, and
second, computing power.

28 CHAPTER 1
PROFESSOR MARK GIROLAMI
Sir Kirby Laing Chair of Civil Engineering Cambridge
Royal Academy of Engineering Research
Chair in Data Centric Engineering

INTRODUCTION

The term ‘digital twin’ has recently entered our

vocabulary, bringing with it as many interpretations
as instances of its usage. There are two perspectives
one can take when considering the digital twin: the
first is a critical and dismissive response, the second
an enthusiastic embrace of the concepts behind
the expression.

Let us consider the first perspective. In July 2018,

the phrase digital twin was at the peak of the
Gartner Hype Cycle, nestled between ‘Deep Neural
Nets’ and ‘Autonomous Mobile Robots’.31 As of
September 2019, ‘digital twin’ has disappeared
completely, replaced by the peak hype of 5G.
Digital twins seem to have escaped the ‘Trough
of Disillusionment’, and they have now entered
mainstream usage, presumably via the ‘Slope
of Enlightenment’.32

A typical definition of digital twin is as follows:

Digital twins are realistic digital representations
of physical things. They unlock value by enabling
improved insights that support better decisions,
leading to better outcomes in the physical world.
Connecting digital twins to create a national digital
twin (NDT) will unlock extra value.’33 The critical
viewpoint maintains that this type of definition is
inaccurate in that there is little, if any, evidence to
support the claims of digital twinning delivering
extra value or better outcomes. In addition, critics
argue that a ‘realistic digital representation of
physical things’ is unachievable, and often ►

29
PROFESSOR MARK GIROLAMI: PERSPECTIVE

undesirable. These arguments have some merit, Data is the oil fuelling the latest AI revolution —
but are ultimately too simplistic and overlook the disrupting existing markets while creating new
potential of digital twinning. markets in all areas of engineering and related
professions. The levels of detailed data available from
A digital twin is in part a model of a real system emerging sensing technologies, from the nanoscale
that is coupled with the digital realisation of the to the global satellite scale, are breath-taking. Data
abstracted model via data generated by and collected provides the ability to strongly couple a digital avatar
from the real system. A simple linear regression, to a physical (or socio-technical) system in ways that
e.g. of regional power demand on time, may be all were unimaginable even a decade previously.
that is required to make useful forecasts. Coupling
this digital avatar with its socio-physical counterpart
yields the digital twin. The result is far from a The second difference is
‘realistic digital representation of physical things’;
it is, however, an appropriately abstracted digital
computing power, the revolution
representation of an observed complex physical in software development and
system. Furthermore, it does have the potential to
‘support better decisions, leading to better outcomes’. the democratisation of high-
So what, then, is new? What is different from what
performance, rapid development
engineers, designers, architects, physical scientists, computing power.
social scientists, and medical scientists have been
doing for decades — that is, building abstract
descriptive models of systems, making empirical This is making possible the development of
observations of their real-world counterparts, computation-hungry digital avatars, which range from
incorporating that data into the models, then using large-scale agent-based simulators of urban landscapes
those models to improve systemic performance (or to such as transportation networks at the city and regional
yield scientific knowledge)? scales, to complex multi-physics computational fluid
dynamics and finite element methods in aeronautics,
My personal view on digital twins is inclined to the geotechnics, civil engineering and architecture.
more positive second perspective. There are two clear The combination of big data and ‘Big Compute’ allows
differences to what has gone before, and these underpin us to make sense of digital twinning in a contemporary
the novel opportunities presented by digital twins. context, which I believe will ‘support better decisions,
leading to better outcomes’. The engineering industry
must, though, be conscientious In developing data-
The first difference is data and centric approaches to the full complement of possible
its availability. There is no applications for digital twins. We must heed our
environmental and social responsibilities as stewards
question that the increasing of the built environment and of civil infrastructure.

ability to gather, curate, The opportunity to systematically couple data-centric

manipulate, store and retrieve approaches such as digital twins that encompass
technical, environmental and social components, their
data is transforming all fields of competing criteria, and combined assessments of
impact provides one way of answering this challenge.
human endeavour, disrupting The technical modelling framework supports this
everything from the political endeavour, but working across disciplinary divides is
essential. For this, changes in working culture, and not
process to commerce to finding only technical capability, are required. The digital twin
concept provides a mechanism via its ability to fuse
one’s perfect partner. diverse models and data. We are not there yet, but the
direction of travel is promising.

30 CHAPTER 1
Finally, one can ask what ‘good’ looks like? What to existing practices, transformed inefficient practices,
do we want to achieve in defining, building, opened up new markets, created multidisciplinary
implementing, and using digital twins? What must teams working to deliver socio-economic benefit for
we include in our thought processes, and what all. Translating this into measurable KPIs is not a
assumptions must we include in our models? difficult undertaking, but achieving those KPIs is the
open challenge to us all. ■
In my view, ‘good’ looks like a strong evidence base
of exemplary successes of digital twins that have
made the impossible possible, brought efficiency gains

31
32 CHAPTER 2
CHAPTER 2

Value of digital twins

Digital twins have the potential to help deliver
on many of the grand challenges we face today,
including urbanisation, population growth,
climate change, escalating infrastructure costs,
and sustainable development.

33
VA L U E A N D B E N E F I T S O F D I G I TA L T W I N S

Digital twins hold promise to improve decision In the following chapter, we feature a wide range
making and investment for a broad spectrum of of digital twins varying in scale and levels of
stakeholders, from city-scale transport planners to complexity, and highlight their benefits, both
individual building owners. Potential benefits include quantitative and qualitative, across the markets. The
everything from better health and wellness in office benefits include the ability for the digital twin to
environments to improved air quality in our dense accurately describe and model our infrastructure in
urban environments. With the digital and social order to efficiently manage and operate it. As this
landscapes converging in terms of modelling, sensing ability develops, the digital twin will attain higher
and inclusivity, digital twins may be the technology degrees of predictive certainty, leading to more
to help deliver on these challenges. intelligent models which will co-evolve with our
infrastructure. We believe that significantly increased
Four emerging themes of digital twins are automation of our infrastructure is imminent, albeit
highlighted in Table 4. Some of the more advanced crucially enmeshed with human co-operation.
deployment of digital twins is currently found in the
manufacturing sector, with many factories already Our case studies include, for example, in the
using twins to simulate production processes. In the Melbourne office, we are collaborating with
automotive industry, the physical system of the cars architects at Hassell to create a digital twin which
covers millions of miles, and the digital twin of the will support optimal design and usage of office
car cover billions of miles to robustly enhance their space by leveraging the health and performance data
radar and image recognition and vehicle-to-vehicle of its occupants. We report a number of projected
communication capabilities.34 benefits from the adoption of the digital twin, such
as employee turnover, performance productivity and
In the aviation industry, engineers are able to reduction in energy usage. In Amsterdam, we are
pinpoint critical weak spots using digital twins, building a digital twin for the Dutch government’s
improving predictions of asset failure.35 In property, County Hall building in The Hague. The objective
built environment specialists are using sensors and of this project is to work with the authorities to
actuators to better understand how people are using determine optimal retrofitting and uplifting methods
the space in which they live and work.36 This enables of the existing structure, in order to achieve an
proprietors to give occupants more control over their energy-neutral building by 2040. Turning to the
own workspaces and environmental conditions, water industry, Sydney Water is preparing to invest
enhancing tenant satisfaction. The value in linking in a comprehensive digital twin to provide real-time
digital models and simulations with live feedback insights into their extensive infrastructure. This
to provide better design and control is an obvious digital twin would act as a network of smart meters,
choice; one that opens up opportunities to deliver and is aimed at saving water, energy and costs.
improved operations and maintenance, more accurate
predictions, more inclusive designs and more In the next section, we summarise some key
sustainable solutions. messages from the case studies, thought leadership,
and expert opinions gathered together in
While our research highlights issues of detail, of this report. ■
degrees of uncertainty, and of abstraction from
the real world, it also highlights the value for our
clients across a spectrum of sophistication, setting
out a clear framework for stages of improvement
and accuracy over time. By applying this evaluatory
framework to case studies from five major markets,
the research provides exemplars to learn from and to
build on.

34 CHAPTER 2
 A digital twin can make the entire supply chain transparent by tracking inventory in real-time and
then recommending or automating redistribution according to demand. Models of ridesharing
networks, for example, can identify demand and redistribute drivers accordingly. In this case it is a
S U P P LY A N D
twin of an abstract network of drivers, with location data of drivers. Each driver has a digital twin
DEMAND
at a lower level of the network, so in fact this can be seen as an instance of subtwins. A simpler
example is city parking networks, which can identify underused locations to drive decisions on the
reallocation of infrastructure.

Numerous examples have shown us how digital twins can continuously monitor operations and
identify abnormal behaviour, allowing human operators to react promptly and reduce downtime.
They can also apply machine learning for predictive maintenance. For example, we have a sewer
O P E R AT I O N A L system with a prescribed direction of flow. Predictive maintenance can be utilised to identify
PERFORMANCE blockages along the system by applying classification and anomaly detection algorithms. We are
able to predict where fatbergs are likely to disrupt the system by using active data such as current
rate of flow and historical data of non-biodegradable solid matter present along the system. Other
examples include automotive manufacturing plants, power plants, and wind farms.

Digital twins can help with the management of assets by keeping records of inventory, processes,
historical data and additional equipment including manuals and inspection data. This allows
owners to identify inefficiencies and ways to address them. For example, we have a physical
model of the road network to allow for inspectors to log inspection data to the twin. The data
may consist of water leaks, road repairs, underlying utilities and inspection photos. As the twin
L I V E D ATA
matures, we are able to optimise the network over time, adapting to various scenarios to facilitate
MANAGEMENT
better decision making, both in planning and operation. This information can help to establish
annual maintenance and operation budgets and more collaboratively, shared with emergency
services in the event of a road closure or diversion. This capability of digital twins holds great
promise for facility management and maintenance teams, portfolio and real estate owners, and
asset manager.

Designers and engineers can make use of models, which also fall under the digital twin rubric,
for quick, inexpensive prototyping of new ideas, particularly from the standpoint of user
experience. These twins can factor in anything from noise to weather, human interactions,
S I M U L AT I O N
lighting, and friction. Digital twins of transport hubs, for example, improve passenger experience
PURPOSES
by identifying the peak times and better understanding human flow, ultimately resulting in
reduced congestion. Simulated systems can include rotors, turbines, engines, trains, aircraft,
and autonomous vehicles.

TA B L E 4
Above
Four emerging themes
of digital twins

35
F U T U R E O F D I G I TA L T W I N S

A digital twin can be a powerful

tool to support diverse use cases
including efficient operations,
scenario planning, disruption
management, climate, carbon
and circular economy outcomes.

36 CHAPTER 2
ALAN NEWBOLD
Global Digital Aviation Leader
Arup

INTRODUCTION

In its simplest form, a digital twin describes a model

of an asset that has the engineering detail embedded
and coordinated within it that enables engineering and
construction firms to perform integrated design and
coordination. This is a simple and basic interpretation of
a digital twin considering only static data from a single
process or area or facility, level 1 if you like.

The outcome achieved here is that the construction

director has an excellent digital twin of what has been
built and accurately represents what will be handed
over to the maintenance team including all maintenance
information that could be used to track the assets from a
climate, carbon and circular economy basis.

Building on this model, now owned by a maintenance

business unit or director, a level 2 digital twin would
gather historic usage data and deploy analytics to start
to look at scenarios that could deliver improvements
to the physical asset or predict things that are going
to happen so that these risks and issues can mitigated
in advance. This would allow an asset owner to move
from a break fix scenario to a predict and prevent model
saving maintenance cost and potentially future capital
cost through engineering the solutions to actual needs
rather than perceived needs or historic engineering
standards as well as potentially extending the life of
the asset.

A level 3 model could add in real-time data from

a currently deployed situation to be able to look at
scenario planning in real-time to optimise the use of ►

37
ALAN NEWBOLD: PERSPECTIVE

People

Process
S

P Insights
N

L
IO

A
T

N Visualisation
A

N
R

IN
E
P

G Analytics
O

DATA
T
B

N
T
U

E
E
IL

M
S
D

E
S
IN

G
A
A
G

N
S

A
M

the asset making significant maintenance savings • Make recommendations as to how and
and adapting to the actual use of the building. This where to improve passenger, customer or
could all be fed back into the model to simulate any staff experience
asset replacement or other asset works in the model
and choose the optimised solution based upon the • Advise on how to increase revenue
customers prioritised requirements.
• Support decision making and optimisation of the
These three levels would then be the foundation to day to day operation
start to build a more complex and business realistic
model by starting to add linked data sets, including A level 5 digital twin could bring in more data from
weather data to predict how plant may need to alternative sources that place this digital twin in its
perform or occupancy data to look at optimising situational context and start to look at this as a system
environmental aspects within the asset. in real-time. This could include people and how they
behave, the interaction between surface access and
This scenario only considers the model from a the landside operation in airports or transport data,
building and maintenance point of view yet there logistics data and weather data.
are significant gains to be made in adding in the
operational data particularly when considering a retail Following this you could move a building or
mall, airport, railway station or hospital. operational asset into a set of connected assets and
then combine datasets across planning, operations,
Moving to a level 4 digital twin you could bring in building and maintenance to deliver a rich virtual
this operational data (built up in the same way as representation of the physical environment. This is
described previously) to: where a city wide model could really start to offer
value to understand the nuanced interactions between
• Support decision making and optimisation of the parts of a city and be used to offer optimised solutions
day to day operation to the impacts of significant decisions like a new
runway at a city airport or a high speed train line
• Make recommendations as to how best to bring a
disrupted operation back into normal operation

38 CHAPTER 2
 costs as well as being able to inform the future
and the potential knock on impacts to local transport
planning and designing of the estate on what has
infrastructure, retail, urban developments and health
actually been used as opposed to designed based
and wellbeing.
upon standards. In addition, scenario modelling can
be to demonstrate what would happen in the event of
An example of this could be built up in the aviation
x or y happening and then seeing how quickly you
context thinking about the following flow.
can bring back the asset to a normal state of operation
and hence model the disruption caused.
NOW

An airport terminal building that is being constructed The ability to be able to do this for a building as
using data from the construction model covering all significant as an airport would enable the airport
aspects of the static build data; cables, walls, structure, owner to save money through extending the life of
mechanical, electrical etc. The outcome would be a the asset through predict and prevent maintenance,
complete set of data on the terminal covering all of the reduce the cost of the maintenance through planned
assets that could be handed over and integrated into preventative maintenance and hence improving
the asset management system of the airport. the passenger experience where it is linked to the
performance of the asset.
The value of this would be the fact that the asset
manager would know exactly what they were getting, NEXT
how to optimise the maintenance of each asset and
details such as the operational lifecycle of the asset Taking this one stage further, this same digital twin
to know when it needed replacing. In the near-term it could be used in relation to the operation where the
would be expected that operational information will asset and maintenance model is combined with the
be added into these buildings or asset models so that baggage, ground handling and other critical operations
as the building is operated over time there would then to create the airport heartbeat.
be a data lake of statistics and operating information
of the asset based upon what actually happened Finally, once all of the assets and operations are added
historically across all assets. to a single model you could integrate this with all 3rd
party operations and passengers and their behaviour.
The outcome with this model is that it can inform the
planning, design and building of any future airport This is the final piece where the digital twin could be
terminal at the airport as well as provide information a real-time digital representation of the entire Airport
to the operating and maintenance team as to how to and effective run the day to day operation of the
improve the existing asset. This may also be able to be Airport potentially resulting in world class passenger
used to extend the life of the asset and reduce the cost experience, optimised non-aeronautical revenues,
of the maintenance. minimal costs of the operation, very low unplanned
disruptions whilst continually improving and learning.
NEW This would also give the airport operator a unique
ability to scenario plan across the entire operation
In the near future and with all of this historic data
before implementing and changes and minimise and
gathered and with the right analytics, predicted
understand the risks of any investment to ensure a
models of the building or asset in its operating
maximum return on that investment.
environment will be available that will enable
decisions to be taken to change maintenance regimes
This vision will become a reality over the next five or
to predict and prevent saving significant maintenance
more years as airport operators start to see the benefits
of their current proof of concepts in this space and
become more connected data driven organisations
across their planning, building, operations and
maintenance structures. ■
FIGURE 6 FIGURE 7
Opposite left Opposite right
Most digital twins currently are An overview of a single
addressing a single isolated use model using asset and
case in one of these four quadrants operational data

39
W H AT D O E S C Y B E R S E C U R I T Y M E A N F O R
D I G I TA L T W I N S ?

There are many IoT devices

being developed and rushed to
market, yet they are not being
designed, from the onset, with
security in mind.

40 CHAPTER 2
C H R I S LY T H
Global Head of Cyber Security
Arup

INTRODUCTION

What does cybersecurity mean for digital twins?

Cyber security is always a challenge in a fast
evolving digital landscape. The convergence
between the physical and the digital worlds means
that cyber security will need to underpin the digital
elements of the built environment. Cyber security
is a fundamental component of delivering business
resilience. In an increasingly connected world more
and more devices of all sorts are being connected
to the internet, thus leaving a huge potential attack
surface. There are a huge number of IoT devices
being developed and everyone’s rushing to bring
these things to market, and they’re not designed, from
the outset, with security in mind. This is an issue for
us in the security team and we’re going to need to
develop the right frameworks to enable IoT devices to
be used securely.

At Arup, there can be a perception that the security

team will turn away requests as they may be seen
as security threats, but we want to be in a position
to enable innovation and collaboration and facilitate
people to do what they want, albeit in a secure way. I
think there is definitely a place for digital twins, but
businesses will have to balance risk versus reward. As
with all of these things, we are living in and moving
towards an increasingly collaborative and open
environment. We need to think and build in from the
ground up how we plan to share, collaborate and work
with digital twins securely - be it in an open, closed,
shared or federated way. Core security principles
need to be carefully considered and a proper security
architecture put in place to develop safe and secure
digital twins.

When we are designing digital twins, we must

consider the CIA triad — confidentiality, integrity
and availability.37 Confidentiality of data must be
viewed from the perspectives of data sensitivity and
intellectual property. As you start to build up these
models, and you have more and more data in there,
and they become more federated, then there’s a huge
amount of intellectual capital that effectively sits in
those models. We will have to think about how we
protect our commercially sensitive stuff, because
that’s essentially our business. ►

41
C H R I S LY T H : P E R S P E C T I V E

We also need to ask ourselves what would happen Imagine another scenario where we have sensors
if the data were to get into the wrong hands (e.g. deployed in a bridge to predict maintenance
models of critical national infrastructure (CNI))? For requirements. Say that you wanted to intercept
example, terrorist groups or nation states; these days, and falsify the sensor data for malicious purposes:
we are increasingly seeing nation states using cyber intending to cause a failure in the bridge.
weapons to attack each other. If digital twins become
linked to control systems in the physical world you
may end up with potential compromise of physical We would set all that sensor
control systems — e.g. manufacturing, building
management systems, CNI control systems, etc. With
data, passing it on as if
digital twins, confidentiality of data is going to be everything were fine, while
important. It’s not just having the right data in there
to make decisions with, it’s also protecting that data in reality the bridge may
from unauthorised access. be experiencing some kind
When we think of open, shared and federated models, of engineering failure. The
we need to consider who is that open to and shared
with, how far does that go? For publicly available integrity of data and knowing
data, effectively, in these digital twins, there are
limits to what you’re actually going to be modelling,
that you’ve got the right data —
because you don’t want data to get into the wrong this is critical.
hands. There is a lot of sensitive data out there which
we will want to protect.
For availability, if you’re using digital twins to make
Considering integrity of information, we need to decisions in the real world in real-time, and you’re
make sure the data in the digital twin is correct. If we using that data to control things, then imagine the
plan on operating higher level digital twins this will scenario in which your internet goes down, or your
be critical since real-time decisions in the physical model goes down — what happens then? Look at the
world will be made based on the data within the WannaCry cyber-attack, which cost the NHS £92m
digital twin. For a level 1 digital twin, this is less — £73m of which was spent on IT costs to rectify the
critical. Increasingly the integrity of data will come problem.39 Over 19,000 appointments were cancelled.
under attack. Malicious actors will inject false data When we refer to availability, we need to consider
into data sets and systems and real-world decisions disaster recovery as a key component within the move
may become compromised. to digital twins. We need redundancy built in. These
days we tend to distribute our infrastructure to many
There was the widely known Stuxnet case in which different cloud platforms, and building in a distributed
a cyberweapon was created to cause substantial architecture is key. We cannot rely on a single model
damage to Iran’s nuclear enrichment facility.38 in a single place. When we distribute over a different
Stuxnet was a piece of code which intercepted data number of environments or cloud providers globally,
from the centrifuge sensors and reported back to then if one or two go down, we are still operating.
the control system that everything was operating
as normal. In reality, the code was speeding up and The level of engagement in each of these areas
slowing down the centrifuges beyond their tolerance comes down to the sensitivity of the data and
levels until they broke. The integrity of the data in the system involved. If you’re developing a digital
system was compromised. This attack set the Iranian twin, ask yourself the question: what is the impact
nuclear programme back by two to three years. With if it gets into the wrong hands or things go wrong
many of these types of cases, the facilities or assets from a confidentiality, integrity or availability
are controlled remotely. The sensors and actuators are perspective? If we are operating in environments
there to partially or fully replace human inspection. of higher sensitivity or greater risk then the
Depending on the level of interaction with the digital conversation about security investment in these
twin, we need to work out how much decision making areas is one to have. The key things that need to be
will be granted to it. in place can be found in Table 5. ■

42 CHAPTER 2
 SECURITY
Right from the start when we are thinking about digital twins, the models and the
architecture, we need to treat the security aspects as fundamental.

D ATA E N C R Y P T I O N
Data will need to be encrypted in the model, both at rest and in transit. ‘At rest’
means that data is not only encrypted as it travels from devices to the digital twin, but
also within the digital twin itself, under all circumstances.

I D E N T I T Y A N D A U T H E N T I C AT I O N
Identity is going to be at the heart of security for digital twins. We need to understand
who is trying to access what, not just from the user perspective, but also in terms
of who is sending us data. Authentication is a core part of identity: how do I know
that the right device is sending the right data to inform the digital twin, or vice versa?
Devices and users will need to authenticate in a secure way, and the whole concept
of identity will therefore be central to digital twins.

PRINCIPLE OF LEAST PRIVILEGE

This refers to applying the correct access rights: allowing users to access what they
need in order to perform their role (internally and externally), and no more. Sharing
the entirety of the model needs to be carefully considered as there may be highly
sensitive data to protect. Aggregation of data may also increase its sensitivity.

SECURITY AUDIT
We need visibility of all transactions, to be sure that the right devices and users have
the right level of access and what activities they have performed, etc.

MONITORING LIVE EVENTS AND RESPONDING TO INCIDENTS

We want to be in a position where we have visibility and monitoring in our live
environment. We want to be able to detect anomalies in real-time, and to provide
effective incident response. The live monitoring of environments is going to be
increasingly AI-driven, using behavioural analytics and anomaly detection.

MANAGEMENT OF DEVICES
When do my devices need to be updated and patched and how do I perform
this? The secure management of networks of millions of devices will need to be
considered carefully.

TA B L E 5
Above
Seven pillars of cyber security to
consider for digital twins

43
H O W D O W E E N S U R E T H AT T H I S I N F O R M AT I O N
REMAINS USABLE FOR DECADES?

The increasing amount of

data that buildings and
infrastructure will generate
in the future poses a new
challenge to the built
environment sector. Despite
many years of experience in
the security and longevity of
physical assets this industry
never really had to deal
with the same aspects for the
virtual counterparts.

44 CHAPTER 2
M AT T H I A S G E I P E L
Senior Consultant
Arup

We are increasingly making demands that we

have previously only made on the apps on our
smartphones allowing for connectivity, adaptability
and long-term support through function extensions
and security updates. This attitude extends
inexorably to other areas of our lives. From the
vehicles we use to the buildings we stay in.

But even without this recent demand of its users,

the high availability of sensors and information
processing technology is one of the main reasons
for an increasingly networked built environment.
A variety of interests and incentives play a role in
this, deciding who wants to install and use which
data sources. The interests differ widely. From the
user who would like to have a better overview of his
energy consumption or air quality, to planning offices
who would like to better understand the effects
of their planning on the operation, to the facade
manufacturer for whom data on the performance of
his installed facade solution can be valuable and last
but not least, of course, the building operator himself.
The vested interests, combined with the increasing
availability and decreasing acquisition costs of
suitable hardware and the associated platforms, will
ensure that our buildings, in addition to the typical
technical building equipment, will be more and more
strongly permeated with the technology resembling
the human nervous system. If you add up the existing
and the real-time data, they become a digital image,
the twin of the physical asset. ►

45
MAT T H I A S G E I P E L: PERSPECT IVE

This development raises two important questions: generated. Security against theft as well as security
against manipulation. The past has shown one
1. H O W C A N WE ENSURE T HAT T HIS INFORM AT ION thing, namely that data, whoever it is entrusted to,
R E M A I N S U SABLE FOR DECA DES? is not actually safe from any of these dangers. If
the size and therefore often the reputation of a data
2. A S WE L L A S THE INCREA SING LY IM PORTA NT processing company increases, hackers’ interest in
Q U E ST I O N O F DATA SECURIT Y? this “honey pot” also increases. In the recent past,
for example, neither state owned personnel service
For example, let’s assume an office building built providers nor Internet marketplaces nor global
in 2020 with an expected end-of-life in 30 years. consulting agencies were safe from being attacked.
Looking back, this corresponds to the challenge The motivation for such intrusions is multi-layered
to store data in 1990 in a complete, secure and and ranges from political-strategic reasons and
legible form to this day. How much of the digital economic incentives to pure vandalism.
information from 1990 is still available today? As
a reminder, in 1990 there were no cloud service What we are currently ignoring in this observation
providers in the current sense, the CD-ROM just is the fact that information from a digital twin will
started its victorious career (is anyone still using in future be attributed value to by its very existence.
optical storage media today?) and smartphones and Performance-related information of an office
social networks simply did not exist. This example building is directly valuable for the building operator
shows what a challenge we face in view of the flood and indirectly valuable for the operators of other
of data from digital building twins. The tools and buildings. Planning and operations information of
platforms we use today may not be the ones we will economically, infrastructurally or military highly
use in 30 years’ time. sensitive facilities is valuable for other reasons.

But let me address the second, equally relevant Systems that handle this data do not only move
challenge. It involves the security of the data that is pure data records but handle assets. The storage and

46 CHAPTER 2
transfer of such data therefore places more demands in the future and how many cultural and political
on a bank-like infrastructure than on any data room. interests will play a role in a single property, it is
This approach makes it clear why a focus should be good to know that one is not dependent on individual
placed on data security. For BIM data, the British actors when it comes to the transaction of valuable
Standards Institution has already published PAS building data. Private providers offering value
1192-5, a standard that addresses this exact concern.40 exchange services are not uncommonly forced to
More precisely, in the context of digital planning, discontinue their services for certain applications or
building and operation, it is a matter of securing the in certain jurisdictions. A dependency one would
following properties of each piece of information: like to do without in critical infrastructure and
safety, authenticity, availability, confidentiality, building operations.
integrity, possession, resilience and utility.
The data security of digital twins can furthermore
Until 2008, there were no major leaps in the area be increased with the help of blockchains, in which
of information security. Rather, we are still widely the networks are used to document unalterable and
using cryptographic methods from the 1970s.41 Not audit-proof snapshots of consumption or sensor data.
without good reason, because these concepts have Due to their decentralised nature, the data here is
proven themselves over many years and have been safer from manipulation than it could ever be in a
tested and improved many times. The author of the centralised platform.
bitcoin whitepaper published in 2008 also made use
of various existing ideas, but was able to answer a But also the storage of the actual, complete data
fundamental question in information technology: model can now be left to decentralised networks.
“How do I create something digital that can’t be Analogous to the value transfer just described, a
copied?”42 A new milestone in data processing network of economically incentivised participants
has been set with the concept that today is largely distributed across the globe is used here to offer high
referred to as Blockchain Technology.43,44 availability and data security online storage.

The cases of data misuse and data loss from the

For the first time it is possible past are reason enough for us to think about the
to carry out a digital, global collection, processing and storage of building data,
especially considering the value the data around a
and instantaneous value digital twin will gain in the future. With blockchain
technology, we now have a tool at our fingertips
transfer without the need for a that allows us to follow a new path. A path that will
central instance. exclude some intermediaries but at the same time
enable new business models.

The processing is carried out by a physically and We only have to evade the assumption that
organisationally decentralised network. A censoring every blockchain, no matter what form it takes,
or manipulation of such a transaction is virtually leads to this goal. I therefore recommend taking
impossible. A fascinating possibility with regard a look at the already existing, open, well-tested
to the aforementioned exchange of data between a blockchains first. Just like in cryptography where
building and its digital twin and vice versa. If one we still use the well tested, open standards to secure
considers how global the real estate market will be our communication. ■

47
48 CHAPTER 2 CITIES
MARKETS

Cities
Cities have been described as humanity's greatest achievement.
Their growth is certainly accelerating: more of us live in cities
than at any other time in our history.

LÉAN DOODY
Smart Cities Leader
Digital

Rapid urbanisation offers opportunities for sustainable and city leaders are therefore key actors in delivering
development, as services such as healthcare, housing, a just transition to a low-carbon economy.47 We need
infrastructure, transport, energy can be more to make sure they are equipped with the data and tools
efficiently delivered in a more densely populated they need to make effective decisions. Smart cities and
environment. However, delivering these services to digital twins offer some possibilities.
a rapidly growing population will be challenging.
John Wilmoth, Director of UN DESA’s Population The concept of the smart city has existed since 2000.48
Division, says “Managing urban areas has become one The basic idea of the smart city is one in which digital
of the most important development challenges of the technology and data are used effectively to address
twenty-first century. Our success or failure in building urban challenges. This is no longer such a radical
sustainable cities will be a major factor in the success idea as in 2000, when digital technology was still
of the post-2015 UN development agenda.”45 very much the purview of the office, the iPhone had
only just launched and digital platforms like Uber
Recent research by C40 shows that urban policy and Airbnb were still seen as the wave of the future.
decisions made before 2020 could determine up to a Since then, we have seen the digitisation of city
third of the remaining global carbon budget that is not infrastructures providing real-time transport data,
already locked in by past decisions.46 So, the decisions energy data from smart grids and smart metering,
made by city leaders will determine whether or not the mobile phone data tracking the movements of whole
world is set on a high or low carbon pathway. Mayors city populations. ►

49
This proliferation of data creates the possibility of supported by The Ove Arup Foundation, has
using it to address our city challenges through better- established a project entitled 'Digital Cities for
informed decision making, improved performance Change', with additional funding from the Centre for
and optimised resources. But city challenges are Digital Built Britain.50 The research team is currently
multi-headed — transport congestion is linked to developing a digital twin pilot for the Cambridge
access to housing, air quality, and employment.49 sub-region in collaboration with local authorities. The
Cambridge Digital Twin pilot project will test how
isolated policies from transport, housing, environment
The promise of the city digital and energy can be bridged using the digital twin,
twin is to help provide a and quantify some of the interdependencies
among transport, air quality housing and
simulation environment, to energy infrastructure in relation to changes and
uncertainties.51 Importantly, the work will test how
test policy options, bring out the Cambridge Digital Twin can be combined into the
dependencies and allow for workflows of the local authorities and used with other
models, or whether it should remain a standalone tool.
collaboration across policy This will help us to understand how best to design
digital twins and their associated governance and
areas, whilst improving processes, so that they can be used on an ongoing
engagement with citizens and basis by city governments.

communities. This could Data ethics is another area which needs further
discussion. The ethical implications of widespread
be transformative. use of new technologies are not clear, although for
example, although there is increasing disquiet about
This all depends on having the right processes and privacy and use of personal data to target people for
governance around the digital twin. For example, a political ends. Privacy, trust and surveillance are key
digital twin could be used to help citizens understand issues, particularly with more widespread deployment
and propose development options and trade-offs, of IoT technologies — one question is how to get
and therefore build political and community support people’s consent and build trust. The Centre for
for beneficial infrastructure investment. Successful Digital Built Britain has published The Gemini
outcomes will not only require the digital twin to Principles, which provide high-level guidance for the
be designed for the purpose of engagement, perhaps development of digital twins and are a useful step in
incorporating user-facing visualisations and tools, this direction.33 There will need to be further debate
but will also need participatory processes such as and discussion at at the societal and political levels
events and meetings. about the benefits and tradeoffs of data ownership and
privacy before we see a consensus in this area.
However, the experience of smart cities has shown
that city governments have struggled to incorporate The city digital twin holds great promise to help
technology and data strategically. This has included city governments make more informed decisions
issues relating to leadership (including knowledge with citizens, identify systemic risks and enable city
and capacity), political boundaries, collaboration with infrastructures such as transport, energy, housing to
multiple stakeholders (public and private), enabling work in concert more effectively. In our enthusiasm
R&D and urban innovation. The same issues will for the technology, we must not forget that it is always
apply to the adoption of city digital twins. human beings who use it. We must design human-
centred digital twins that can fit with our existing
This is an area ripe for research. The Cambridge organisations, even as they help to change them. ■
Centre for Smart Infrastructure and Construction,

50 CHAPTER 2 CITIES
51
QUESTION AND ANSWER

The internet is becoming the Internet of Things;

a fusion of bits and atoms. As a result, every atom has a
digital counterpart — or a digital twin.

P R O F E S S O R C A R L O R AT T I
Director
MIT Senseable City Labs

TELL ME A BIT ABOUT W H AT I S YO U R D E F I N I T I O N

in The Atlantic some months
YO U R S E L F, A N D W H AT YO U R A N D V I E W O F A D I G I TA L T W I N ? ago. Starting from a compelling
LEVEL OF ENGAGEMENT WITH historical comparison with the
D I G I TA L T W I N S H A S B E E N ?
CR I see digital twins as reflecting spirit of the Enlightenment,
CR I am an architect and current broad technological he warns about the danger of
engineer, and I wear several trends: the spaces around us mistaking machines’ capability
hats, as the English would say. are becoming permeated with for “unprecedented memorisation
Firstly, I lead the design and digital technologies and data. The and computation” with actual
innovation firm CRA [Carlo Ratti internet is becoming the Internet human intelligence, which is
Associati], based in New York, of Things, a fusion of bits and based on thinking.53 “AI is
Turin and London.52 Secondly, atoms. As a result, every atom likely to win any game assigned
I direct the Senseable City Lab has a digital counterpart — or a to it. But for our purposes as
at the Massachusetts Institute of digital twin. humans, games are not only about
Technology, in Boston. Finally, I winning; they are about thinking.
am engaged with a few start-ups. HOW DOES THE HUMAN FIT By treating a mathematical
I N T O T H E D I G I TA L T W I N ?
Although each activity focuses process as if it were a thought
on a different area — research, The human must always be process, we are in danger of
projects, or products respectively at the centre. As advances in losing the capacity that has been
— the vision is always the same, digitalisation are probably the essence of human cognition.”
and it deals with the impact of inevitable, we should not forget
digital technology on the way we that it will be humans who H O W W I L L D I G I TA L T W I N S
TRANSFORM THE INDUSTRY?
experience the built environment. determine how they will affect us.
In this sense, even if I havenot I’m reminded of a very interesting CR I think that the consequences
much used the digital twin label piece that former Secretary of will be profound. For instance,
itself, it is central to what we do. State Henry Kissinger wrote I believe that the traditional jobs

52 CHAPTER 2 CITIES
of structural and M&E engineers urban dynamics ever since the
will disappear — AI can do them beginning of our work at the
very easily on a digital twin Senseable City Lab at MIT.54
model, at a fraction of today’s This was, for instance, the
time and cost. I believe that the focus of our Real Time Rome
only engineering that will survive project at the Venice Biennale
in the medium term will be the 2006, a world premiere in the
one with a creative angle — à la use of live cellphone data in a
Ove Arup… large city.55 In it, we aggregated
data from cellphones (obtained
Construction itself will also using Telecom Italia’s innovative
change radically, with digital Lochness platform), buses and
twins allowing components to taxis in Rome to better understand
converge on a site just-in-time, urban dynamics in real-time.
before being assembled in a
similar way to how we assemble This project shows how one of
I believe that the
cars or airplanes today. And the the most important technologies only engineering that
monitoring of a building will for the future of our cities may
continue within the same digital be our cellphones, which already will survive in the
twin framework. Basically, we
will see changes across the
have very evolved sensors — so
evolved, in fact, that in a project
medium term will
whole chain. at the SCL, we used them to be the one with a
measure the vibrations coming
W H AT A R E D I G I TA L T W I N S
A B L E TO S O LV E , M E A S U R E , O R
from Harvard Bridge. We placed creative angle.
DO BETTER? smartphones on two cars crossing
the bridge and used the data they
CR Beyond what we were just collected to measure the bridge’s
saying about the construction structural health. Smartphones,
industry, digital twins will have in fact, are equipped with three
a broader impact on our cities. accelerometers that are able to
Linking the physical and the measure the movement of the
digital would allow us to better phone on three axes: height, length
measure our built environment, and depth, and to register a series
and with the additional layer of data of varying importance.
of information we would gain, This is an example of the way a
we could respond to it in more mobile network could be a cheap
tailored and efficient ways. This way to supplement and improve a
could take the form of digital network of fixed sensors. Digital
sensors dispersed throughout twins would need to leverage
the physical world, capable similarly dynamic sensors.
of supplying the necessary
information to update the digital Such measurements could give us F I G U R E 8
counterpart on a variety of an idea not only of the structural Above
elements. A digital twin would health of infrastructure, but of Carlo Ratti Associati equipped the
Agnelli Foundation HQ in Torino,
then be in a position to make the needs and desires of users. Italy, with hundreds of digital
calculations and combinations This is what we tried to do sensors that allow occupants
based on the data obtained from with our renovation (as design to choose their preferred
environmental settings, creating
these measurements, providing and innovation practice Carlo responsive micro-climates
information on how to transform Ratti Associati) of the Agnelli
the physical infrastructure. Foundation HQ in the heart of
Torino.56 In collaboration with tech
We have been interested in how company Siemens, we equipped
real-time (live) data can change the building with hundreds of ►

53
P R O F E S S O R C A R L O R AT T I : Q U E S T I O N A N D A N S W E R

digital sensors that monitor content so as to inform policy-

different sets of data, including makers, companies and citizens
occupancy, temperature, carbon alike. We have also been working
dioxide level and the status of extensively on the ethical and
meeting rooms. Users can access moral issues connected to big data.
this monitoring platform through In 2013, we launched an initiative
a mobile app that enables them called ‘Engaging Data’, involving
to interact with each other, check leading figures from government,
in, book meeting rooms and privacy rights groups, academia
regulate environmental settings, and business.57
all with an unprecedented degree
of personalisation. WHERE ARE THE
O P P O R T U N I T I E S F O R D I G I TA L
TWINS IN THE ARCHITECTURE,
For instance, once a building ENGINEERING AND
CONSTRUCTION INDUSTRY?
occupant sets her preferred
temperature and illumination
CR Apart from some of the
settings, the building management
system responds accordingly, streamlining issues we have
adjusting the levels of lighting, discussed before, I think that
heating and cooling. As the fan coil digital twins have the potential
units situated in the false ceilings to redefine the very meaning
are activated by human presence, of a building. Architecture has
the system can potentially been called a third skin that
follow occupants as they move envelops us atop our biological
around the building, just like an skin and our textile skin —
environmental bubble. When an clothing. It has often been
occupant leaves, the room returns the most uncompromising of
to standby mode and conserves the three, rigid and unable to
FIGURE 9
energy, just as a computer would adapt to the changing needs of
Above
do with a screensaver. different occupants. Yet, as the
MIT Senseable City Lab’s fleet
of Roboat units, which were first IoT revolution unfolds, with a
tested in Amsterdam’s canals, W H AT A R E T H E T E C H N I C A L widespread fusion between the
could be used as agile sensors
to gather data on the city’s
CHALLENGES OF DELIVERING physical and the digital, architects
D I G I TA L T W I N S ?
infrastructure, as well as air and are beginning to imagine how
water quality insights
CR I think that things are
this third skin can develop
coming together. However, I responsiveness, learning from
am concerned about the data people’s needs. ■
ownership issue. Today there is
a risk of data asymmetry, with
just a few companies and public
institutions knowing a lot about
us, while we know so little about
them. There is a real danger of
data monopolies and data misuse
by both large corporations and
governments. At our MIT lab, we
instead seek to investigate how
big data can be used to improve
life in cities for inhabitants. We
publish all the research and data
we collect from our projects
online, creating platforms that can
be expanded with user-generated

54 CHAPTER 2 CITIES
55
C A S E S T U D Y: A R U P C H I N A / H O N G K O N G

Digital twins for Neuron City

AUTONOMY

INTELLIGENCE

LEARNING

FIDELITY
1

LEVEL 2

2
4

B AC KG R O U N D A N D M O T I VAT I O N PROTOTYPE

China is one of the world's leading testing grounds The Neuron City prototype is shown in Figure 10.
for smart cities. Arup Hong Kong is building a city Our first-stage work was completed in 2018/19
scale digital twin platform to map space, people, where we built a city scale information model (CIM)
and activities in the physical city to a virtual city. platform operating system. The prototype included
The digital twin incorporates GIS, BIM, IoT, the following functionalities:
cloud computing and AI. Although many pilot
projects exist on the market, they tend to follow a • 3D modelling and spatial analysis
similar profile: they are often targeted at testing or • Visualisation of simulation data and statistics
showcasing computational capabilities, rather than • Building data dashboard
at solving real-world problems. They are often too • Parametric design module
fragmented, isolated systems such as transportation • Real-time data visualisation and analysis.
or security and in general, with a low degree of
artificial intelligence. Therefore, we built a platform P R O J E C T A P P L I C AT I O N S
that is able to overcome these challenges. With the
help of our city planners and design team, we are The project integrated an IoT data management
able to bring this to life. platform. We produced a demo of the complete
design scenario and implemented a machine
We built a digital twin of Hong Kong city called learning algorithm to provide predictions using
Neuron City. The platform maps the physical spaces city-scale data. ►
and people to a virtual city (i.e. the digital twin).
We are able to monitor, predict and control aspects
of the physical city by incorporating a closed-loop
data stream. This has lead us to solve rather complex
problems in the urban lifecycle.

56 CHAPTER 2 CITIES
Digital twins of cities map
space, people, and activities in
the physical city to a virtual
city. By building closed-loop
city-level data into the virtual
city, monitoring, prediction
and control of the physical
city can be achieved, in order
to solve complex problems of
the urban lifecycle.

PLANNING ECOLOGY T R A N S P O R TAT I O N BUILDING ARUP TOOLS

C I T Y S PAT I A L D ATA I N F O R M AT I O N D ATA M A N A G E M E N T A N D

P L AT F O R M P L AT F O R M S H A R I N G P L AT F O R M

C I T Y D ATA

I N F O R M AT I O N I N F R A S T R U C T U R E
CLOUD // IOT // NETWORK

FIGURE 10
Above
A holistic view of the Neuron City
framework

57
CAS E STUDY: A RUP CHIN A / HONG KONG

FIGURE 11
Opposite top
By building closed-loop data
into the virtual city, monitoring,
prediction and control of the
physical city can be achieved, to
solve the complex problems of the
urban lifecycle

VALUE

FIGURE 12
The value of the first-stage work consisted of building Opposite bottom
the CIM platform which was used in several projects Distribution of air pollution, as
to demonstrate the added value to our design. simulated by the digital twin
The platform adds value by integrating data for
collaboration and decision making, and by developing
innovative city prediction and control applications.
The value generated can be captured in the two
examples listed below:

Distribution of house prices

Online house price data was collected and visualised
by location as part of Neuron City. This provides a
quantitative basis for decision making and greater
insight into the housing market.

Multi-source spatial analysis

We used the ‘landsat’ multispectral remote sensing
image to achieve a specific land type distribution by
generating a Normalised Difference Moisture Index
raster image.58 This has allowed us to achieve remote
sensing to higher degrees of accuracy and precision.

FUTURE WORK

Future plans for this platform includes establishing

a data pipeline, collaborating across the wider
ecosystem and developing partnerships with
integrators and suppliers to continue to solve complex
problems in the urban lifecycle. In the next two years,
we plan on launching a market-ready commercial
product with the aim of incorporating more
intelligence and learning into our platforms. ■

MEET THE TEAM

ALLISON AN BRUCE CHONG

Analyst Associate Director

QIAN CAO Y WANG

Senior Engineer Associate Director

58 CHAPTER 2 CITIES
 SENSING A N A LY S I N G A N D F O R E C A S T I N G RESPONSE

IoT data City modelling Big data forecast Intelligent response

and simulation

AI
C I T Y I N F O R M AT I O N M O D E L
= GIS + BIM + IOT
CIM D ATA

A N A LY S I S A N D F O R E C A S T I N G D I G I TA L T W I N

59
QUESTION AND ANSWER

A smart city digital twin is an IoT-enabled virtual

platform that has some kind of data richness,
real-time information, and where we can see what is
happening in the city.

P R O F E S S O R J O H N E TAY L O R
Associate Chair for Graduate
Programs and Research Innovation &
Frederick Law Olmsted Professor
Georgia Institute of Technology

TELL US A LITTLE BIT W H AT I S T H E D I F F E R E N C E H O W D O E S A M A C H I N E D I G I TA L

ABOUT YOURSELF AND WHY B E T W E E N A D I G I TA L T W I N A N D TWIN DIFFER FROM A SMART
YOU ARE INTERESTED IN A S I M U L AT I O N ? C I T Y D I G I TA L T W I N ?
D I G I TA L T W I N S .

JT I’m a civil engineer who does JT For me, digital twins are about JT When we define the smart city
a lot of computational work. machines. You create a virtual digital twin, we are talking about
As I look at what is happening representation of the physical an IoT-enabled virtual platform
across universities, I see that a world by pumping in real-time data of a city that has the kind of data
lot of the discussion seems to about some system and monitoring richness and real-time information
centre around the computational it, so that you can anticipate a which allows us to see what is
and technological issues related problem before it occurs. You can happening in the city. We can
to smart cities, but I don’t see also use simulation to predict — start to analyse it to understand
as many civil engineers or given its current state — when a why things are happening the way
infrastructure experts at the table. machine may need maintenance, they are happening, and then we
Ultimately, IoT sensors are going or when it may wear out on can simulate what-if scenarios.
to be placed on infrastructure, so the assembly line. I think what One difference is what is being
I think architects, civil engineers differentiates this from simulation measured. When you move to the
and urban planners need to be a is that the answer it gives you city scale, you have to include
part of this — right in the centre of today will likely be very different humans in the loop. What I’ve
the conversation. to the answer it gives you a week learned from the manufacturing
from now. This is because it is assembly line is that humans
based on the flow of real-time certainly play a role in the degree
data. It is a continuously updating to which machines wear out. In
simulation that allows you to a city, if you’re going to study
predict future states. energy use, humans are the ones

60 CHAPTER 2 CITIES
who are making decisions about tools that put people in front of the
that energy use. If you’re going to data. We can put data in front of
study transportation, humans are users in contextually realistic ways.
the ones driving the cars. Human- We can put city managers at the
infrastructure interfaces become actual intersection, where they
an important part of the definition can then analyse the data. They can
of smart city digital twins. look at different aspects that are
occurring around that intersection
HOW WILL HUMANS INTERACT while they pull out different bits
W I T H D I G I TA L T W I N S ?
of data and do visual comparisons
and correlations.
JT We can boil it down to two
broad classes of interaction. On the
The way we interface with citizens
one hand we have what we call in
is a bit different. Just as we build
the built environment the facility
manager, who interfaces with
feedback systems for buildings,
we must engage occupants in
A digital twin is
a building energy management
system in a very different way than
the process too. We have been a continuously
gathering information about what’s
does the occupant of the building.
At the city level, we also have
happening in the city and allowing updating simulation
maintenance and operations for
occupants to input their needs and
preferences though smartphone
that allows you to
services in the city; there would be
an interface for those users to see
applications: this combines basic predict future states.
app development with augmented
what’s happening across all system
reality. They don't need full
in the city.
immersion in virtual reality; they
just want to contribute what they
The way we are developing
have to say about this building here
this type of interface is through
or that bike path there. ■
augmented and virtual reality

61
C E N T R E F O R D I G I T A L B U I LT B R I T A I N

National Digital Twin Standards —

Gemini Principles

The Gemini Principles report was published by the

Centre for Digital Built Britain in December 2018.33 It
aims to align approaches to information management
across the built environment, since establishing agreed
definitions and principles from the outset will make it
easier to share data in the future.

These principles are effectively the conscience of the

Information Management Framework for the Built
Environment and the National Digital Twin. To ensure
that these two initiatives are — and remain — for
the public good, they need strong founding values to
guide them.

Enshrined in these values is the notion that all digital

twins must have clear purpose, must be trustworthy
and must function effectively. All the Gemini Principles
flow from this. They are intentionally simple, but their
implications are far-reaching and challenging. They
are descriptive of intent, but agnostic on solutions, to
encourage innovation and development over time.

P H I L I P S H E A LT H C A R E

The digital patient in healthcare

Philips Research asks: if a digital twin of an MRI

scanner can help predict which part will need
replacing, could the same concept be applied to
discover and treat ailments in the human body before
they become apparent?59 By integrating different
measurements of a person, a digital model of a body
part, such as their heart, could be built. Eventually,
an integrated model of their anatomy and physiology
could be constructed.

The vision is to have a personalised, lifelong patient

model that can be updated with each measurement,
scan or examination, and that includes behavioural
and genetic data as well. The digital patient could
bring together all a patient’s information to support
diagnosis and treatment.60 In future, it might be
possible to predict diseases and provide treatment
and prevention plans unique to the characteristics of
an individual patient. Digital twins for the
whole human body could usher in an era of
personalised medicine.

62 CHAPTER 2 CITIES
VIRTUAL SINGAPORE

World's first city-scale digital twin

From bus stops to buildings, Virtual Singapore is a
data-rich, live digital twin developed to produce a
central platform for the modelling done by different
government agencies.

The digital twin draws on IoT sensors, big data

and cloud computing, combined with 3D models,
geospatial datasets and BIM. Virtual Singapore was
co-developed with the French firm Dassault Systèmes,
by leveraging its existing software platform.61

The interplay of map and terrain data, real-time

traffic, and demographic and climate information
show how a single change could affect the lives of
millions of people, and the systems they depend
upon. Virtual Singapore offers a platform that can
be used by urban planners to simulate the testing of
solutions in a virtual environment.

NEW SOUTH WALES (NSW) GOVERNMENT

A digital twin for Sydney

The Government of New South Wales, Australia,

is creating an open-source interactive platform on
which to capture and display real-time 3D data and
4D data on the urban environment.62 The project,
which enlists the help of DFSI’s Spatial Services and
CSIRO’s Data61, aims to stimulate development in
the region.63 The twin will assist with planning in,
design for and modelling of the city. The government
has already demonstrated the ability to render 4D
environments, with integration of live transport feeds
and infrastructure building models.

The project, currently at the proof of concept stage,

will eventually be publicly accessible through
standard internet browsers. Industry players and
government at all levels will collaborate to give users
access to a wealth of information to apply in their
work and lives. Meanwhile, the government will
invoke applications ranging from natural disaster
response, to transport scheduling, to security.

63
64 CHAPTER 2 ENERGY
MARKETS

Energy
Can we harness digital computing power to produce real benefits
in resource use and environmental protection?

JOHN BURGESS
Global Energy Skills Leader
Energy

The human race is at a crossroads on its pathway Digital twin modelling, sensing and reporting
to sustainable planetary balance. Will digital will help us to identify energy waste and nudge
technology save us from ourselves? Or will it people to adjust their behaviour through the use of
become a negative contributor to climate change, dashboarding, messaging and action tracking.
with exponential growth of energy consumed by
storage and computation? Let us look at a few examples to show how Arup
are already making inroads into various phases of
Since the eighteenth century, energy has been energy markets — from renewable energy resource
synonymous with belching fossil fuel emissions discovery to offshore asset inspections, from
which have ruined air quality wherever energy has transmission and distribution network maintenance
been consumed on an industrial scale. In the past 50 to renewable energy grid integration, from energy
years, these emissions have caused a temperature rise storage to network resilience planning, and from
equivalent to one which took 150,000 years in the energy economic policy modelling to supply- and
last major climate cycle. Hence, in the current push to demand-side management techniques. ►
wean our economies off those carbon-intense fuels,
major opportunities are arising in the use of the digital
twin concept to design and manage energy efficient
low-to-zero carbon (LZC) solutions for buildings,
communities, cities and regions.64

65
N E T W O R K O P T I M I S AT I O N F O R D I S T R I C T ELECTRIC VEHICLE CHARGING GRID
ENERGY SCHEMES
Arup have created a proprietary model that combines
Arup is currently developing the District socio-economic, housing, and transport data with
Information Modelling and Maintenance for Energy electric vehicle (EV) adoption rates and vehicle
Reduction (DIMMER) project, concerned with the performance data.68 This information is brought
interoperability of district energy production and together to provide a detailed picture of EV charging
consumption to provide comprehensive real-time demand through the day, across an entire city. With
feedback about the energy impact of user behaviour.65 the uptake of EV, we are exploring the possibility of
The aim of the project is to achieve enduring a digital twin to capture real-time charging demand,
reductions in both energy consumption and charge point placement and grid capacity. This
CO2 emissions. will allow us to test different scenarios and and to
understand how many charging points are needed,
OFFSHORE STRUCTURES ASSET MANAGEMENT the likely mix of vehicles and the projected energy
demand across the network.
Arup have secured a three-year partnership with
EnerMech to automate offshore platform inspections. Arup continues to invest substantial time and
Our digital approach (Arup Inspect MInteg; effort in R&D activities to expand our abilities in
AIMTM) significantly reduces the time and cost energy engineering advisory services. Our teams
spent undertaking mandatory inspections.66 are successfully harnessing the digital twin concept
It gives operators and asset owners instant access across the broad spectrum of our energy interests, in
to interactive inspection data at any time, allowing order to shape a better world. ■
them to make informed decisions about repairs,
replacements or improvements.

R E N E WA B L E E N E R GY G E N E R AT I O N
RESOURCE MAPPING

Arup have produced a ‘pumped hydro roadmap’ for

the government of New South Wales, Australia.67
Pumped hydro is a way of storing energy by pumping
water between two reservoirs — when the energy
demand is low, water is pumped from the lower
reservoir to the higher one. When the demand is high,
water is released back through the pumps, which
causes turbines to generate electricity that can then be
released into the grid. A digital twin of the reservoirs
is being considered as one of the technologies to
monitor the operational stages.

66 CHAPTER 2 ENERGY
67
C A S E S T U D Y: A R U P H O U S T O N

A digitised workflow for offshore inspection

AUTONOMY

INTELLIGENCE

LEARNING
1
FIDELITY
1
1

LEVEL 1

INTRODUCTION

A modern digital twin comprehends its input — gas industries (Figure 13). MInteg carries out onsite
inspection results, design information, condition inspection, surveying and data collection, while Arup
monitoring and so on — using data analytics and is responsible for data management. This includes
machine learning. The inputs and outputs are stored provision of the tablet-based inspection application
in a database which should ideally feed a 3D model- for MInteg to use in the field.
based front-end application used to communicate
data and analysis results about an asset to its operator. Integrity issues must be properly understood by
The digitised asset can be used to gain insight about qualitative or quantitative analysis of the effects of
the full sweep of operational data for that asset — degradation found. We are currently developing
everything from the environmental conditions within a module within the AIM tablet application that
a building to the response of an offshore wind turbine visualises structural analysis results. The next step is
to environmental loading. to provide clients access to this module, on top of
the existing inspection data review functionality,
AIM (Arup Inspect-MInteg) is a digitised workflow so that clients can visualise in one place how
for oil and gas asset inspection.69 At the workflow’s degradation affects the structural design performance
core is a tablet application for the recording of of their assets.
inspection data based on the AIM tablet application,
which was the first of its kind. The AIM tablet G E T T I N G B E T T E R DATA A N D U S I N G DATA
BETTER
application was developed within Arup over a
five-year period, and has been used in a variety of With AIM, data is captured on the tablet at the point
industries for improving inspection efficiency. of inspection, rather than by the traditional pen-
and-paper method. Where wireless connectivity is
A R U P I N S P E C T- M I N T E G ( A I M ) available, tablets synchronise with the cloud. Hence,
multiple tablets can be used at once to capture data
Arup has collaborated with an industrial rope on a single asset. Inspection includes photography,
access inspection and non-destructive testing the recording of conditions, and noting comments.
contractor, MInteg, to introduce AIM to the oil and Photographs can either be captured on ►

68 CHAPTER 2 ENERGY
FIGURE 13
Top
A screenshot from an AIM
tablet application

69
CAS E STUDY: A RUP HOUSTON

the tablet and automatically tagged to the model, database. AIM is hence lightweight, in that it can be
or an external camera used to incorporate images integrated into an existing workflow or management
into the application later on. Inspection condition system without replacing it entirely.
observations are recorded directly on the tablet.
Other data such as 360° panoramic images or laser TIME ON TOOLS
scan point cloud data can be appended to the asset
model afterwards. Conveying inspection results to clients traditionally
requires inspectors to take notes by hand, and
D E M O C R AT I S E D DATA later to type up those notes. However, the AIM
tablet application can automatically export reports,
Once the data is uploaded to the cloud, the office- including all relevant inspection data, removing
based client can access it in near-real-time using the need for the inspector to touch a computer and
a desktop version of the application. Normally thus maximising potential inspection time. We
the asset owner/operator receives relatively sparse estimate based on recent projects that the AIM tablet
information during the inspection, with only a terse application, with its user-friendly interface, can
report and limited context around any issues. The save up to 25% on inspection times compared with
client must then attempt to infer the inspectors’ tacit traditional approaches.
knowledge to fully understand the inspection results,
and to make use of them. The AIM tablet application NEXT STEPS
represents an improvement on this situation, by
providing clear, constant information on asset status. We are in the process of incorporating further
elements into the AIM workflow, and the AIM
P R E S E N T A N D F U T U R E DATA Q U A L I T Y tablet in particular.In the medium term, we plan to
visualise structural health monitoring data tagged
Traditional inspection is inherently subjective. Arup to relevant locations on the 3D model, to give a
Inspect 3D attains consistency by limiting free text complete picture of current structural performance
input to an absolute minimum. Data is instead entered as against design intent. We are also expanding AIM
using drop-down lists. The combination of the tablet to other safety-critical offshore systems, such as
application and and our 'Street View'-like panoramas pressure systems. By applying a deep learning layer
dramatically improves the onshore integrity to the back end, we will automatically screen the
engineer’s ability to query the data. images received from the inspections for defects or
degradation such as corrosion in a first pass, and flag
PA R S I N G H I S T O R I C A L DATA these to the responsible engineer as priorities
for review. ■
Using innovative machine learning and data
analysis techniques, Arup can parse historical MEET THE TEAM
data contained in both vector and raster PDFs
(either newly or historically scanned), as well as
B R YA N H O R T O N
in Word or Excel files. This data can be added
Senior Engineer
to the AIM model to create a record of historical
inspections. The intention is to liberate the value of
historical inspection data, and to make it maximally N AT H A N L E N N O X
Designer
compatible with data newly captured by the tablet
application. The AIM tablet application can also
store historical data in easily retrievable documents. TODD GRICE
Senior Engineer
EXISTING SYSTEMS
CAMERON DUNN
The AIM tablet application is compatible with Associate
existing systems, which is often an important
desideratum for our clients. We can, for example,
automate the process of exporting inspection findings
into a format suitable for an existing anomaly

70 CHAPTER 2 ENERGY
FIGURE 14 FIGURE 15
Top Bottom
Inspector at MInteg using An off-site oil rig project
the AIM tablet

71
CASE STUDIES

GENERAL ELECTRIC

For the digital power plant

GE offer solutions for the continuous optimisation of
power plants using digital twins.70 The twin represents
aspects of the physical asset or larger system using
thermal, mechanical, electrical, chemical, fluid
dynamic, material, economic and statistical data from
the plant, allowing operators to predict performance,
evaluate different scenarios, understand trade-offs and
operate at appropriate levels of risk. GE’s solution runs
on the industrial platform PredixTM, which ingests
data wholesale and implements analytical models
which can be tailored to the needs of executives and
managers to interact with the plant in real-time. In this
way, dispatch, efficiency, startup and asset life can all
be optimised.

GE's digital twins are integrated globally, to provide

them with over 44,000 operating hours of data each
day from gas turbines, steam turbines and generators.
This provides a glut of information to inform anomaly
detection software and to prevent forced outages
or damage to parts, reducing loss and maintenance
costs in the whole asset. This all adds up to a
comprehensive package for assisting and maximising
plant functionality and balancing needs with wants,
giving operators comprehensibility and visibility —
despite the complexity of plant operation — in their
decision making.

72 CHAPTER 2 ENERGY
AGILITY3, BRUNEL
UNIVERSITY LONDON,
DASHBOARD, ESI & TWI

Lowering wind turbine

O&M costs
Agility3, Brunel University
London, Dashboard, ESI and
The Welding Institute (TWI) are
collaborating on the WindTwin
project, developed specifically for
the wind turbine industry.71 It aims
to solve the problem of growing
maintenance and operational costs
associated with running wind
turbines. The consortium has
developed and integrated enabling
technologies that include high-
performance cloud computing,
system fault and degradation
modelling, data analytics and
visualisation for predictive
maintenance of wind turbines.
A sensor network system, using
optimised signal processing and
condition monitoring algorithms,
is applied to the live wind turbine
to collect operational data
which will then interface with
a replica, virtual 3D model, or
digital twin, of the wind turbine.
This collated and processed
data provides a description of
the wind turbine’s dynamic
behaviour, and physical state
during real-time operations. Wind
farm operators will be able to
diagnose performance variations
down to the level of constituent
individual components, allowing
them to predict failure and plan
maintenance, reducing costs
and downtime.

73
C A S E S T U D Y: A R U P T O K Y O

IoT Office — enhancing office well-being

AUTONOMY

INTELLIGENCE

LEARNING
1
FIDELITY
1
1

LEVEL 1

B AC KG R O U N D A N D M O T I VAT I O N by the onboard neural network to yield a real-time

headcount. Simultaneously, we can record voice
Recent developments in IoT technology have led data from those in the vicinity, analyse that data
to the accumulation of more detailed and more with a commercial cloud service, and finally classify
voluminous data. The Arup Tokyo office, with an sections of speech by their emotional characteristics.
area of approximately 700m 2, has recently been
outfitted with 12 environmental sensors and 16 D E R I V I N G VA L U E F R O M DATA
human count sensors to cover meeting rooms,
private rooms, and open space (Figure 16). In We first discovered that temperature and degree
addition, an AI camera and emotion analysis device of usage between various areas of the office varied
have both been installed in an open space used for widely. To communicate the current state of the
various events. We named this 3D-printed device office environment in real-time, we installed a
the ‘penguin sensor’ (Figure 17). monitor at the entrance (Figure 18). This installation
resulted in staff more easily being able to identify
Our environmental and penguin sensors taking malfunctioning air conditioning, for example.
readings every five minutes. Meanwhile, a further
sensor measures people count every minute. All data AI has recently begun playing a prominent role in
is sent to a single cloud database (InfluxDB), and our use of the office data. We took measurements
visualised using a dashboard.72 This dashboard is of temperature, humidity, CO2 concentration, and
displayed on a monitor installed at the office entrance, human count in our meetings rooms, before analysing
giving a rapid real-time overview of office status. correlations between all variables. We found, for
example, that CO2 concentration and the number of
Our commercial environmental sensors measure people are highly correlated (Figure 21). Using our
temperature, humidity, CO2 concentration, and results, we are now working to improve prediction
noise level. A further movement sensor detects the accuracy for each variable (Figure 19).
number of people within the sensing area using an
infrared camera. The penguin sensor, developed in- AI also plays a prominent role in our use of the office
house, comprises a Raspberry Pi linked to a camera data. For example, CO2 data is highly correlated with
and a microphone. The camera feed is analysed number of people for each day of the week. At the ►

74 CHAPTER 2 ENERGY
 GIANT
TABLE

E N V I R O N M E N TA L THERMO CAMERA /
SENSORS SENSORS MICROPHONE
FIGURE 16
Above
Layout of the Arup office in Tokyo

PERSON 2 PERSON 4

PERSON 1 PERSON 3

FIGURE 17
Sensors concealed as ‘penguins’
were used to gather environmental
(temperature, CO2), video and voice
data from occupants utilising the
meeting rooms

75
CAS E STUDY: A RUP TOK YO

FIGURE 18
Opposite right
Real-time monitoring
of sensor fluctuations in the
Tokyo office

FIGURE 19
Below
Temperature and CO2
concentration measurements
of the Tokyo office

beginning of the week, there is an upward trend in the

office hours with two spikes in the afternoon, but the
shape of this upward trend changes with each day of
the week. Closer to the weekend, CO2 concentration
can be further related to both staff who leave the
office immediately after working hours and staff who
work overtime, or who stay later to socialise.
Temperature changes [°C] (measurement data from
Various measurements have been conducted in August 1st to 31st is overlaid and displayed)
addition to those focusing on the office environment.
For instance, we introduced AI to classify human
emotions based on voice data, and analysed the 30
High
results from different events. Lecture-type events 29

show little emotional change. On the other hand, for 28

party-type events, we observed positive emotions 27

°C

initially, followed by a gradual rise in anger (Figure 26

24). It is likely that anger per se is not the true 25

underlying cause, but rather more enthusiastic 24

Low
communication styles as the night unfolds. 23
0:00:00 12:00:00 24:00:00

NEXT STEPS
CO2 concentration changes (measurement data from
August 1st to 31st is overlaid and displayed)
The Tokyo office continues to conduct various trials.
For example, pupillometry has been used to gauge
employee concentration levels, while stress level has 1500
High
been estimated from heart rate fluctuations. There 1300
are teething problems: for instance, concentration 1100
levels were significantly underestimated during
CO2(ppm)

900
creative meetings and talks. We think it is important 700
to test and evaluate continuously in the service of 500
shaping a workplace where staff can work creatively 300
and happily. We expect our continued testing to bear 100 Low
further fruit in the near future. ■ 0:00:00 12:00:00 24:00:00

76 CHAPTER 2 ENERGY
77
 CAS E STUDY: A RUP TOK YO

FIGURE 20
Right
Conference room in Arup Tokyo

FIGURE 21
Below
Correlations between the measured
variables in the conference room

FIGURE 22
Bottom
Target vs. measured CO2
concentrations in the
conference room

M1 M2 M3
1.0 1.0 1.0
CO2

CO2

CO2

1 0.22 -0.03 0.52 1 0.29 -0.037 0.48 1 0.2 -0.00094 0.56

0.8 0.8 0.8
Humid

Humid

Humid

0.6 0.6 0.6

0.22 1 0.25 -0.0076 0.29 1 0.17 0.17 0.2 1 0.15 -0.033
0.4 0.4 0.4
Temp

Temp

Temp

-0.03 0.25 1 -0.063 0.2 -0.037 0.17 1 -0.11 0.2

-0.00094 0.15 1 -0.066 0.2

0.0 0.0 0.0

Ppl

Ppl

Ppl

0.52 -0.0076 -0.063 1 0.48 0.17 -0.11 1 0.56 -0.033 -0.066 1

CO2 Humid Temp Ppl CO2 Humid Temp Ppl CO2 Humid Temp Ppl

Prediction
1400

1200

1000
CO2

800

600

400

2018-08-25 2018-08-26 2018-08-27 2018-08-28 2018-08-29 2018-08-30 2018-08-31

Dates

78 CHAPTER 2 ENERGY
79
MEET THE TEAM

K E N TA R O S U G A DAISUKE KAWAHARA
Associate Engineer

H A R U TA K A O E KENJI LIMURA
Engineer Engineer

YOSUKE KOMAI AT S U S H I M I YA Z A K I
Engineer Project Manager

80 CHAPTER 2 ENERGY
CAS E STUDY: A RUP TOK YO

FIGURE 23
Opposite
Arup, Tokyo office

FIGURE 24
Below
Voice-based emotion analyses of
two office events

Time series emotion analysis results of internal lecture event in April 2019

60

50

40

30

20

10

-10

17.30 18.00 18.30 19.00 19.30 20.00 20.30

Time series emotion analysis results of internal social gathering event in April 2019

60

50

40

30

20

10

-10
15:36:00 16:48:00 18:00:00 19:12:00 20:24:00 21:36:00

81
AV E VA

Savings for oil wells using digital twins

Block 18 is a group of undersea oil wells operated

by BP, off the coast of Angola.77 It was one of
the earliest projects to use AVEVA NET for the
handover of engineering and facility documentation.
AVEVA NET aggregates engineering and asset
information and makes it accessible from a single
application. Tag data from documents and drawings
can be automatically extracted and linked to an
information model of the digital asset. Several
major contractors were involved in the project and
as first oil approached, information from all parties
had to be collected and verified. Approximately
160,000 tags — individual component identification
numbers, 2,550,000 tag attributes and 180,000
documents — made up the data.78 Cross-referencing
documents to confirm the engineering information
associated with each tag would have been an arduous
clerical task. Using AVEVA NET, 96% of tags were
captured before first oil, as opposed to 65% using the
traditional process. AVEVA NET’s digital twin for
handover enabled a 25% reduction of engineering
man-hours, a 5% reduction in the cost of major
equipment thanks to streamlined documentation, a
10% reduction in commissioning costs through better
access to documentation, and a 10% reduction in
operational and maintenance services thanks to the
reduced need for site surveys.

82 CHAPTER 2 ENERGY
83
84 CHAPTER 2 PROPERTY
MARKETS

Property
Realising the ambitions of a digital twins for buildings will
come from embracing a framework of continuous improvement
and adopting an iterative approach to identifying and exploring
use cases.

MICHAEL STYCH
Director
Arup London

The construction industry is evolving rapidly, question the effort and expense of doing so without a
with unprecedented investment in technology set clear value proposition.
to transform the way buildings are constructed,
used and maintained. The digital twin concept is It is likely, then, that success will only come from
one prominent vehicle for the industry to realise conceptualising digital twins as a framework of
this transformation. continuous improvement and exploration. There
are two key principles to this idea. First, twin
The argument for having a digital twin of a built development must be seen as a foundational aspect of
asset is compelling — a virtual copy with which to the development of everyday processes and systems.
hone design and construction performance, optimise Second, we must adopt an iterative and experimental
the use of materials, improve user experience, and approach to identifying and exploring use cases.
predict component maintenance and replacement.
However, buildings are a mosaic of many parts A digital twin for a building should ideally be a
sourced in a fragmented way, and so unlike the by-product of the existing processes of design,
twin of a single component (a pump, a window), the procurement, construction and operations. Using
proposal can quickly become overwhelming unless well-structured design models taken through to
one concentrates on the value of specific outcomes. completion while taking a consistent approach to data
In other words, while it is technically possible to will provide a solid foundation for any digital twin.
construct a twin of any building, it is easy to call into To do this, it is important to conceive of data ►

85
apart from the specific software with which it with a wide range of building components from
has been processed, so that it can be shared and different systems. This initial step has provided
understood regardless of its source. positive answers to questions of scaling, integration
of legacy systems and asset databases. Such by-
products provide insight into occupancy patterns
In property, success in any use and building performance previously not easily
case relies on being able to accessible by operators. As geometric and asset data
is brought in from BIM, we are seeing the digital
integrate data from the various twin idea become a reality, not only for a single
building, but for an entire, data-enabled portfolio.
interested parties (builders,
owners, managers), to enable Optimising design using computational analysis is
normal practice, and we now have easy access to
effective analysis and avoid sophisticated techniques across multiple disciplines.
However, with our development of a prototype
rebuilding basic requirements building for the White Collar Factory in London,
each time. we demonstrated the potential for operation data as
instantiated in a digital twin to influence design.75
This project took a novel approach to ventilation and
Digital twins must, therefore, adhere to common data cooling, using an approach thus far untested in the
standards so that businesses and buildings can adapt London market, which resulted in the construction of
to changing technological infrastructure. a 200m² prototype.

IN PRACTICE We integrated separate systems and their associated

devices into a real-time database, which was
At our London office we have created a real-time data-, mapped to independent monitoring of environmental
analytics- and application-ready digital twin, which conditions to feed a predictive model. This allowed
forms the foundation of our Property Insight platform. the building designers to experiment with the
The team started with the question of how to optimise physical and virtual environments to optimise
the use of an eight-person meeting room. We trialled the design for the ventilation and cooling control
different technologies alongside existing installed systems, as well as to optimise the facade design,
systems, and integrated data such as room calendar improving daylighting while maintaining the solar
booking information. This quickly revealed that the shading requirements. Having a well-organised BIM
room was underused, providing evidence on which to model and operational data has allowed for direct
act on and which could be extended with other rooms. improvement to the design of the main building,
all while taking an agnostic approach to systems
Automating the retrieval of data from various selection and data.
relevant systems eventually resulted in a structured
data environment covering the entire building. The potential for digital twin thinking to transform
This included the integration of BIM data for rooms design and construction, as well as user experience
as well as for individual devices. Starting from this and operational efficiency, is an exciting prospect.
single use case with simple visualisations, we have The examples described here, along with the more
developed a system which allows the FM team to extensive case studies elsewhere in this publication,
make informed decision on building performance start to demonstrate the capacity of digital twins.
and space usage. They show that by adopting an iterative approach
while developing solid foundations for digital twins,
We are currently extending this approach to an entire we can achieve the longer-term goal of widespread
portfolio, working with a property developer on a adoption by the property industry. ■
programme to develop digital services for tenants.
As a proof of concept, we first gave tenants control
of their local lighting and temperature conditions.
While this is a relatively restricted functionality,
choosing this use case allows us to test interaction

86 CHAPTER 2 PROPERTY
87
QUESTION AND ANSWER

Our mission is therefore not merely to liberate data, but

to filter it. Our job is to extract meaning from the noise.
To turn data into insight and insight into action.

JOSHUA RIDLEY
CEO and Co-Founder
Willow

CAN YOU INTRODUCE WILLOW a digital twin by enriching the part of our business feeds directly
FOR US?
model with both static and live into our software platform,
Willow is the digital twin for the data. This includes the asset WillowTwin™, to ensure that at
built world. We are a technology register, maintenance logs, the end of construction we have
company revolutionising the warranties and O&M manuals the basis of a digital twin.
way we design, build, operate — as well as live data from the
W H AT A R E T H E PA I N P O I N T S
and interact with the built building management system. YOU'RE ADDRESSING?
environment. Through our
software and services, we are Today, we have an entire arm of For the owners of real estate,
converting bricks-and-mortar our business, WillowDigital™, the management of building
assets into living, learning, which works at the design data is a huge pain. Every day,
evolving digital twins. and construction phase of a the built environment produces
building’s life to enable its digital huge amounts of data, from the
CAN YOU TELL US ABOUT twin.76 This expertise in digital occupants who walk the corridors,
YO U R J O U R N E Y W I T H D I G I TA L
TWINS? architecture, digital engineering to the equipment that powers
and information structuring and measures operations. Sadly,
Willow’s origins were in the allows Willow to introduce digital most of this valuable information
design and construction of twins to the building owner very goes to waste, lost in fragmented
complex projects. We were an early in the project. This allows digital forms that are locked
early adopter of BIMs, and it was Willow to shape the client’s digital away in proprietary systems.
a natural progression for us to and technical requirements prior The mismanagement of this data
move from BIM to digital twins. to the builder being appointed; we has a negative impact on bottom
Once a BIM has been generated, work throughout the supply chain line performance, and owners
it makes sense to convert it into to deliver on this strategy. This

88 CHAPTER 2 PROPERTY
are throwing away inestimable want technology to be used
amounts of untapped value. for anything other than that
purpose. We want users to have
The real estate industry has a better experience, and we want
historically struggled to embrace owners to be able to deliver a
technological innovation. Major better experience. How this is
providers have fallen short in implemented in a hospital as
their ability to remain adaptable opposed to a high-rise building,
for future needs. This means that though, may differ. This is a
owners have zero data ownership, question that has to be answered
their information is isolated in the on a case-by-case basis, but
systems of individual buildings, with the user-centred philosophy
and they face the threat of cyber in mind.
breaches or great financial cost
every time they introduce a new
We estimate that a
W H AT A R E T H E B E N E F I T S O F
HAVING ACCESS TO THIS KIND
technology at scale. O F DATA?
15–20% reduction
Our customers receive so much Digital twins can increase
data that they feel overwhelmed the value of real assets in a in variable operating
and distracted. As Yuval Harari
put it in Homo Deus: A Brief
number of ways. We estimate
that a 15–20% reduction in
costs for office
History of Tomorrow: “In the variable operating costs for buildings is possible.
past, having power meant having office buildings is possible.
access to data.77 Today, having This reduction comes courtesy
power means knowing what to of big data analytics, improved
ignore.” Our mission is therefore visibility for facility managers,
not merely to liberate data, but optimised enforcement
to filter it. Our job is to extract of warranties, predictive
meaning from the noise. To turn maintenance, automated
data into insight and insight compliance with fire and
into action. In doing so, we are safety regulations and energy
making what came before us optimisation algorithms,
seem outdated, clunky, inefficient, as well as reductions in
costly and painful. machinery downtime.

W H O O W N S T H I S DATA? Additional benefits include an

estimated 10–15% improvement
It’s a very interesting question: in tenant retention for office
of course, we need legal clarity and industrial assets through
on who is the rightful owner of enhanced tenant feedback on
the data. The approach thus far building performance, more FIGURE 25
has simply been that whoever timely processing of tenant service Above
owns the building owns any requests, and improved visibility Willow's digital twin for tenants
data gathered within it. But it’s of the performance of facility
not that simple; just think of the managers. Organisations also see
example of tenants with medical benefits from big data analytics
patients whose data may be of space utilisation, use patterns,
highly sensitive. occupation growth, merger/
acquisition and key personnel
I believe in a free world, in changes in relation to the office and
which we empower people to industrial tenants. ►
make the world better, to have
better experiences. I wouldn’t

89
JOS HUA RIDLE Y: QUESTION A ND A N SWE R

WHY DO OWNERS NEED HOW WILL THE GROWING

D I G I TA L T W I N S ? N U M B E R O F D I G I TA L
TWINS AFFECT THE
PROPERTY MARKET?
Digital twins promote better
decision making across an The demand for digital twins is
organisation. With WillowRail, fuelled by behavioural changes in
our rail maintenance platform, we occupants and owners. Occupants
gather and integrate hundreds of are demanding more from their
sources of data from individual space, accustomed as they are
proprietary systems to provide to extensive personalisation and
a single view of a rail network, customisation in other aspects of
unlocking new and actionable their lives (smartphones, media
insights.78 This brings benefits to consumption, and even culinary
all levels of the business, from choices). Owners, meanwhile,
the CEO to Asset Managers and are eager to understand what is
maintenance teams, all in real- important to today’s tenants and
time. Through WillowRail, the their employees, and what the
organisation now has a single future will hold.
source of truth in a unified interface
across their business. Accurate Willow is developing a workplace
decisions can be made in shorter application to capture tenant
times, based on a complete picture feedback via surveys, which
of the problem. The platform has will allow owners to readily stay
the power to upskill and empower in touch with their tenants and
the workplace by providing access to measure tenant satisfaction
to richer information. (or otherwise) and attitudes to
continued occupancy. Owners
Digital twins are also key to currently find this difficult to do
FIGURE 26
driving savings throughout the in a cost-effective manner, as
Above
building lifecycle. Asset owners are tenants tend not to prioritise the
Willow's digital twin platform
having to spend more each year to provision of such information.
operate and maintain their assets. The manner and speed with which
For an average-sized building of facility managers respond to and
500,000 sq. ft, asset owners are address customer service requests
required to spend approximately can have a significant impact on
USD 2,985,000 on an annual basis overall tenant satisfaction and
on variable operating expenses, tenant retention. Our applications
including USD 1,075,000 on will allow owners to survey such
building repairs and maintenance, variables as frequency of tenant
USD 840,000 for cleaning and service requests, average time
USD 1,070,000 for energy. On a for closure of requests, facility
portfolio basis, assuming a portfolio managers responsible for each
of 40 buildings, this equates to item, tenant satisfaction following
more than USD 119 million per item closure and ratings for
annum of variable operating facility managers.
expenses. We estimate that the
net saving in a building’s variable Using this type of data, we can
operating expenses generated optimise the workplace, improve
by WillowTwinTM equates to satisfaction with space, and
approximately 3–9% of variable increase real estate performance.
operating expenses per annum. Most importantly, owners will
A reduction in variable operating be able to anticipate change and
expenses also leads to a substantial develop timely solutions to address
increase in asset valuation. that change. ■

90 CHAPTER 2 PROPERTY
I B M ’ S C O G N I T I V E D I G I TA L
T W I N ( I O T WAT S O N )

A platform for desk

availability

Connected IoT devices stream

vast quantities of data to the
cloud, have yielded unprecedented
levels of knowledge about how
these systems operate in the real
world. IBM's Watson® allows
us to learn from these sources
using powerful data modelling
techniques to aggregate data into
a representation that forms the
backbone of a so-called 'digital
thread' connecting digital twins,
digital models of physical assets,
and groups of assets. Watson
can absorb and analyse data in
virtually any format and from any
source, to correlate it with other
sources. Previously inaccessible
insights are made possible,
enabling the development of new
initiatives and the rejuvenation
of existing ones. Deloitte’s Edge
building in Holland analyses an
employee’s calendar and reserves
the necessary desks accordingly.79
With so much documentation
stored digitally, having an assigned
personal desk will become a thing
of the past. The proper collection
and analysis of data allows big
gains in productivity and space
usage to be made.

91
C A S E S T U D Y: A R U P A M S T E R D A M

Digital twin for the Dutch government, The Hague

AUTONOMY

INTELLIGENCE

LEARNING
1
FIDELITY
1
1

LEVEL 1

P R E PA R I N G F O R V I TA L C O M P L E X I T Y
a large amount of data collected over a long period
The inexorable drive towards sustainable energy usage of time. By running simulations on the model, the
means that in the near future, buildings will no longer digital twin can generate 100,000 times more data
be based on preprogrammed responses alone. Instead, than can be provided by the sensors alone. In this
they will need to be able to program themselves by way, the digital twin becomes smarter much more
monitoring the environment, different system states, quickly than the unaided physical building ever would.
occupancy and behaviour patterns, and then using Furthermore, a digital twin simulation can interpolate
this data to predict future states of the building. This many more virtual sensors than would ever exist in
goes far beyond what is possible with current building reality, thereby filling the information gaps which
controls, based on solitary sensor readings. Enabling could not be gleaned from the sensor data alone.
the energy transition therefore requires a new toolset, Beyond even this advantage, the digital twin will
enriching sensor data with far deeper insight: a already understand the relation between these points,
digital twin. whereas discovering this relation solely from real-time
building data would be extremely challenging, due to
The twin’s simulations use empirical and physics- the many non-linear and dynamic physical responses
based mathematics to model building performance, characteristic of the built environment.
and are thereby able to mimic dynamic and complex
behaviours, to increase the efficiency of the physical P R E PA R I N G F O R V I TA L C O M P L E X I T Y: T H E
P O W E R O F D I G I TA L T W I N S
asset. Each digital twin represents a unique one-to-
one correspondence with a physical asset. No two In the next decade, we foresee a large-scale
buildings are the same, and it is the digital twin that integration of (decentralised) electricity production
will provide the key capability of managing each from renewable energy sources. In many countries,
building’s particular environmental and spatial the the share of renewable energy sources will
conditions, user behaviour, and state of repair. grow in parallel with the extensive electrification
of demand, e.g. replacement of traditional cars with
The real power of the digital twin, however, arises electrical vehicles or displacement of fossil fuel
from its machine learning capabilities. To reasonably heating systems, such as gas or oil boilers, with
predict asset failure or to detect opportunities for energy-efficient heat pumps. Simultaneously, the
optimisation, a data-driven approach will require energy supply is developing, with low-temperature

92 CHAPTER 2 PROPERTY
district heating grids powered by various renewable D I G I TA L T W I N O F F I C E B U I L D I N G , T H E H A G U E
sources.80 These changes to both demand and supply
impose new challenges on the management of energy Arup is building a digital twin of the Dutch
systems, including, for example, the variability and government’s County Hall building in The Hague
limited control of energy supply from renewables — a replica of an office building of around 16,000
or increasing daily load variations, due to the high square metres.81 Over 30,000 data points from the
impact of additional electricity consumption. existing Building Management System were first
extracted. Next, another 350 IoT sensors were added,
The changes to the grid and on implementing energy specifically tailored to measure user interaction. The
storage systems will require major investments. dimensions of the physical building were translated
The more variability can be solved locally, the less to the virtual world through 3D scanning, and linked
influence it will have on the grid. Also, the more to a scientific simulation model which is fed the
demand and response can be matched locally, the sensor data.
less the integration will require energy storage. It is
expected that real-time variability management will The province aims that all buildings should be
become a major cost driver of reliable operation, energy-neutral by 2040.82 The strategy is twofold: it
which will result in variable rate structures and optimises current functioning of the building, while
financial incentives at all levels of the energy network. renovating, uplifting and transitioning where needed.
This will require buildings to become demand But, as with any other building, energy meters do
responsive and energy flexible. It will also require not at present indicate whether the current building
buildings to be able to relate to their context and is over-consuming. And as with any other building,
surrounding, forming part of a smart city. And it will the Province House is a unique building. There is no
require buildings that are able to base current control other building of exactly the same shape, at the same
decisions on future projections. location, experiencing the same weather conditions
— and no building is used in exactly the same way
as the Province House. There is no reference building
against which to compare it. But by building a virtual
twin of the building, we obtain an exact copy of the
building against which to compare real-time data. ►

Building
simulation
Building
sensors

FIGURE 27

6 630.720 Left
Number of data readings in
an hour from sensors as
compared to the simulation for the
readings per hour readings Hague building

93
CAS E STUDY: A RUP A M STE RDA M

This reveals not only whether the building is

consuming too much energy, but also how much FIGURE 28

more — and where exactly the energy leaks are to Opposite top
be found. The Hague’s County Hall and the
core components of the digital twin

Comfort requirements specific to the digital twin,

such as lighting, air quality, temperature, humidity
FIGURE 29
and acoustics, are likewise linked to specific office Opposite bottom
activities. These might include meetings, working The digital model for The Hague’s
behind a desk, relaxing and focussed working. County Hall
Each task comes with a specific set of preferences
and needs. By mapping actual performance
against comfort needs, opportunities for comfort
improvement or energy savings can be ascertained.

A N E N E R GY- N E U T R A L D I G I TA L T W I N

By creating a digital version of the building, different

alternatives for an energy-neutral county hall can
be digitally simulated and tested before being
implemented in real life. A better facade, more
sustainable energy generation, more comfortable
building services, energy exchange and storage are
all part of this future version. This makes it possible
to test different design alternatives in a context of the
exact user interaction and specifics of the building.
For example, knowing how the building heating load
will be affected by future improvements to the facade
can be very relevant information if part of the heating
system needs replacement before the renovation takes
place. Furthermore, certain system choices, or more
effective and targeted building utilisation, can have
such a large impact on the building energy balance
that the order of effectiveness of alternative measures
is changed. The projection of future scenarios with a
digital twin is therefore of great help in making future
investment decisions. ■

MEET THE TEAM

CHRISTA DE VAAN MATHEW VOLA

Associate Director

EWOUD RUIFROK VERONIKA HEIDEGGER

BIM Manager Software Developer

WAHIT KARAMALI SENNO KAASJAGER

Senior Engineer Programmer

94 CHAPTER 2 PROPERTY
 U T I L I S AT I O N
O P T I M I S AT I O N

11% 40%

Utilisation Occupation

C O M F O R T O P T I M I S AT I O N ENERGY
3D-scan TRANSITION
23% 60% 54% ADVICE

55%
IoT sensor
data 30% 60% 30% Energy savings
and production
E N E R G Y O P T I M I S AT I O N

D I G I TA L T W I N D I G I TA L T W I N
BMS data
Energy Savings
OF THE FUTURE

1 2 3 4 5 6 7

Comfort

95
GOOGLE

A personalised home assistant as your

digital twin

The concept of a digital twin can be extended beyond

replicating a building or a physical asset. What if you
replicated yourself? Google is developing its Google
Assistant technology to serve as your own personal
digital twin.83 The idea behind this is ultimately to
save time by carrying out the more mundane jobs
for you, such as booking appointments and making
calls. It can also make decisions to help improve
your personal productivity and wellbeing. In the
same way, a digital twin of a building will collect
data about room temperature and then suggest
ways that the building can be optimised to save
energy. The device learns about your routines and
your preferences and can then prompt you to make
decisions. Instead of the information being captured
through sensors in the building and then being
displayed on a 3D model, speech recognition is used
to collect data, enabling your device to talk to you.
This example highlights the variation of forms that a
digital twin can take, and the different ways in which
they operate.

96 CHAPTER 2 PROPERTY
MICROSOFT & THYSSENKRUPP
& WILLOW

Elevating the elevator

business
Willow and Microsoft are helping
ThyssenKrupp Elevators to
access the vast amount of data
in the central nervous system
of their new innovative Test
Tower in Rottweil, Germany.84
WillowTwin™ is a virtual
model of the built environment,
while Microsoft’s Azure Digital
Twins is an IoT service that
powers comprehensive virtual
representations of physical
environments and their associated
devices, sensors, and people.
Harnessing this combination,
Willow and Microsoft are
enabling ThyssenKrupp Elevators
to create new ways to manage
maintenance contracts, create
personalised experiences for
visitors, change the way security
is managed, and ensure that
their elevators are functioning
at all times. The virtual tower
encompasses 15,000 live data
points and over 57,000 static
data points — any of which can
be found at the click of a button.
This digital environment allows
checking of system behaviour
and simulations of the system
and people flow analytics, getting
robust results without the need
for physical testing. Detection
of possible failures will enable
service interventions before
problems even occur. This has
changed the ways that Thyssen
interfaces with the building
— whether this involves the
occupant, owner, or the operator.
Their belief is that it won’t be
long before every building will be
required to have a digital twin.

97
C A S E S T U D Y: A R U P M E L B O U R N E

Digital twins for building tenants

AUTONOMY

INTELLIGENCE

LEARNING

FIDELITY

LEVEL 2

2
2
2

INTRODUCTION

Arup Melbourne has recently moved to a new office, productivity. This qualitative information helps to
designed by the local team in collaboration with Hassell feed a digital twin to yield intelligent suggestions
architects. A central element of the design from the for alternate working locations and to provide user
outset was the deployment of sensors to gather data control over personalised environments. The result
throughout their tenancy. Now settled in the office, the projected an increase of up to 11% in productivity.86
team is using the digital fingerprint constituted by the Additionally, this functionality establishes the
collected data to create a digital twin. methodology of using AI derived from empirical
evidence to inform future designs.
The digital twin initiative aims to gather real world
data to simulate in digital space. Research continues 2. Spatial Optimisation
to support the importance of office space design, On average, rent constitutes 9% of business operating
and its link to the wellbeing and performance of the costs.87,88 With the aid of a camera-and-sensor IoT
people who work there. Following the Melbourne network, a digital twin could ascertain current space
team’s recent move, it has become clear that the usage. Pattern recognition could yield intelligent
design and optimisation of usable space will require suggestions on how a space is optimised (e.g. through
constant curation in order to get the best from our retrofitting or subleasing rooms or zones). Tenants
staff and business. The key benefits and value of this would benefit from such insight to optomise their own
work targeted five main areas, listed below: use of space and expenditure. This digital harvesting
methodology has the potential to expand to other
1. Improving Productivity typologies where the activity in the space is in flux
Research continues to demonstrate a close and the need for spatial optimisation is constant. The
relationship between worker comfort and team has identified public infrastructure, education
productivity. A recent controlled study reported a facilities and exhibition spaces as prime candidates to
productivity deficit of 4% at cooler temperatures, apply this type of thinking and approach.
and 6% at warmer ones.85 Building on this research,
the Melbourne team has developed an app to gather 3. Energy Reduction
information on worker comfort and perceived Approximately half of commercial energy ►

98 CHAPTER 2 PROPERTY
 I N F O R M AT I O N INTELLIGENT TENANT
G AT H E R I N G SUGGESTIONS VA L U E

from a sensor- from a digital twin derived from data-

enabled building driven by historical driven digital twin
tenancy data, algorithms, and intelligence and
machine learning visualisation

FIGURE 30 FIGURE 31
Top Bottom
A digital twin can be a feedback The diagram schematises how
and monitoring opportunity to value may be derived from
drive productivity gains, improve collected data when mediated
worker wellbeing and reduce by a digital twin
operational costs

99
CAS E STUDY: A RUP ME LBOURNE

consumption in advanced economies is due to offices. was the development in creating the backend of the
Due to imperfect control of HVAC and lighting digital twin to handle and process a large volume of
systems, a significant proportion of this energy is various data types from a wide range of sources. The
wasted maintaining unoccupied or inefficiently project dealt with many data silos and stakeholders
configured spaces. A digital twin can support efficient that required consistent and effective communication
automation of environmental control, based on data from technical IT infrastructure experts to general
gathered from actual usage of a space. Savings of up managers. We are at a crossroads, entering a new,
to 20% are projected to be made on energy costs with data-driven market. The industry does not fully
this kind of intelligent building management. understand the potential value that could be derived
from this technology. Navigating this space is
4. Security complicated, and solutions are unique to each client.
The team found that the digital infrastructure provided Our approach focuses on establishing projected
by a digital twin lays the ground work for proactive returns on investment for a particular user group, and
automated security systems. This can be achieved with facilitating discussions around the value of data and
AI image recognition technology in the IoT network. digital twin infrastructure while transcending the
Areas previously monitored by people could then be complexity of the project.
monitored by a digital twin, leading to a potential
reduction in labour costs by up to 50%.89 Insights from PROJECTED RETURN ON INVESTMENT
historical data could also be used for area-specific
threat detection. The project has revealed a number of significant
benefits delivered through the adoption of a tenancy
5. Wayfinding digital twin. One of the largest observed benefits is in
Current workplace designs rely on workers’ familiarity enhanced employee engagement, leading to reduced
with office fitout to find their way around. Using turnover. Employee turnovers are costly, and an
data from 3D as-built BIM models and occupant increase of 2.5% of employee satisfaction can result in a
geolocation, a digital twin can be used with an high-impact reduction in overall labour turnover costs.
augmented reality app to guide users to free rooms, From a productivity perspective, we are projecting
quieter or under-occupied spaces, or specific facilities. performance gains of up to 10%, based on industry
The introduction of two-way interaction systems research, translating to up to 182 hours of performance
in conjunction with wayfinding can also be used to uplift per person per year.
provided adaptive and real-time way finding in a
space. This type of digital twin provides facilities Additionally large revenue uplift was identified
managers with the ability to have real-time influence for building owners, who, when providing
over the usage of a space. This digital twin study appropriate infrastructure to facilitate DTs can bring
quantifies the benefits of wayfinding tools and a more attractive product to market. Providing offices
demonstration the extreme value in it's potential and spaces with enhanced digital infrastructure
to affect time reduction and adaptive management provides measurable improvement to businesses
applications. Additionally, wayfinding technology occupying them, as such rental yields of such spaces
can benefit the safety of the space by using it for the can be increased.
identification of evacuation paths in emergencies.
Given our commitment to the UN Sustainable
KEY CHALLENGES AND OUTCOMES Development Goals, energy reduction is also a key
focus. With ongoing building tuning unlocked through
The key challenges of this case study have been in the digital twin framework, we are projecting tenancy
data acquisition, data cleaning and communication. energy usage reductions of up to at least 20%, leading
Due to the fragmented nature of data ownership and to cost reductions of at least $5 per square metre per
creation in the buildings industry, acquiring data annum. Local electricity and building services will
continues to be a significant challenge. Once data is vary from region to region, and a digital twin approach
acquired, a significant proportion of time is typically allows for simulations to suit. Measuring spatial
spent on cleaning the data to be in a suitable format optimisation ROI continues to be a key challenge. As
for processing. This process can take up to 80% of the project develops, we are identifying drivers and
the component's delivery time based on industry metrics to measure and quantify investment returns for
standards. Key to successful delivery of the project spatial optimisation.

100 CHAPTER 2 PROPERTY

Much of the project’s findings and opportunities are CONCLUSIONS
now being applied to large-base building applications
as well as construction sites. For example, we Ongoing collaboration with industry and academia
see value in ‘micro digital twins’ for bespoke is key to the digital twin project. Through our strong
construction assemblies, which can incorporate AR relationships with local universities and augmented
and VR representations of building programmes, reality experts worldwide, we have developed a
drawings and 3D models to enable clients to mitigate cutting-edge visual platform to increase efficiency and
risk during the construction of complex systems. to boost worker productivity and satisfaction.
This could be achieved through the provision of
software platforms enabling workers to visualise and MEET THE TEAM
process physical tasks in the digital space, prior to
procurement and construction, driving significant SHAW KUDO BREE TREVENA
efficiency gains on site. ■ Engineer Senior Project Manager

TIM RAWLING KRISTON SYMONS

Engineer Principal

HOA YANG SEAN MCGINN

Designer Associate Principal

FIGURE 32
A screen shot of the temperature
gradient feature

101
C A S E S T U D Y: C A M B R I D G E

Digital twins for research and development facilities

AUTONOMY 0

INTELLIGENCE

LEARNING

FIDELITY
1
1
1

LEVEL 1

B A C K G R O U N D A N D M O T I VAT I O N

Arup is developing the entire Asset Management into this replica, enabling advanced analytics,
Service for the R&D facility of a large modelling and prediction of future behaviours.
pharmaceutical company. The new campus is
currently under construction within the UK’s leading Among other advantages, this will allow deeper
biomedical hub. The buildings and infrastructure integration and collaboration at each point along the
will not only support world-leading research, but asset management supply chain, in an economically
will also meet the most exacting of sustainability efficient manner. Successful service delivery
and energy targets. The highly regulated nature on campus will thus depend on the creation and
of the pharmaceutical industry means that robust maintenance of a digital twin, or asset information
information systems already exist: to support R&D model, capable of real-time behavioural responses
processes, for example, or to track the distribution of and prediction according to its surrounding
medicine to patients. However, the legacy IT systems ecosystem. The campus digital twin will play
used to manage physical assets often make difficult a key role in ensuring the highest standards of
or prevent the collation of information into a single safety, customer service and asset sustainability.
view of the asset. A comprehensive digital twin of It will enable the successful delivery of services,
the new campus appears the optimal solution for such as lifecycle planning, facility and laboratory
long-term operational benefit. maintenance, energy and utilities, provision of
scientific services, delivery of capital portfolio
SCOPE projects, and building monitoring and control.

Arup is working with the client to scope out the Arup and their client have worked closely to
work for the R&D campus, to develop a system of determine how customer needs can best be met
interconnected digital twins. The overall digital through the appropriate use of technology and of
twin, a system of systems, will furnish a competitive asset data. This was supported by an analysis of the
advantage, going beyond traditional R&D asset client’s IT and data landscape emerging from the
management practices. All primary asset data physical build of the R&D centre. We then looked at
sources, both static and dynamic, will be integrated several scenarios comparing the positive ►

102 CHAPTER 2 PROPERTY

 I N F O R M AT I O N SYSTEM A D VA N C E D
HANDLING SOLUTION INSIGHT

Collect Integration to Data

single truth enrichment

Control Centralised Business

and validate catalogue analytics

Maintain Maintenance Advanced

management analytics

Share User
visualisation

TA B L E 6
Top
Capability blocks of the digital twin

M L F O R H VA C
RISK PREDICTION

C O R E D T D ATA
I N T E G R AT O R

AR FOR AI ENERGY
MAINTENANCE O P T I M I S AT I O N

FIGURE 33
Right
An example of various applications
feeding into a single digital twin
data integrator

103
CAS E STUDY: A RUP CA MBRIDG E

S TA G E 1 S TA G E 2 S TA G E 3 S TA G E 4 S TA G E 5

RFI RFP Hybrid option PC + 5 Live

AIM Procurements live

Consolidated Market testing

requirements & supplier
conversations

SOLUTIONS/
DELIVERY

Procurements

PoC A
Implementation New capability DT
Increment
PoC B INFRASTRUCTURE
ENABLERS
ENABLED
Data strategy Implementation New capability
PoC C Increment New capability
Smart device Increments
/system strategy Implementation New capability
Increment
AI system architecture AI system architecture
Integration design & assurance
Implementation
sequencing Implementation approach
& business case Delivery oversight & assurance

Target operating model development & business change

outcomes of digital twins at different levels of user requirements. Armed now with a solid
maturity, as well as the feasibility of each level of understanding of the problem, and after iterative
maturity. The next steps of the planning process were refinement of the appropriate strategy, we have
then mapped out, clearing the way for successful collaboratively produced a clear roadmap towards the
project delivery: final digital twin infrastructure.

• RFI to understand market capabilities NEXT STEPS

• Data strategy
• System architecture Over and above the ongoing work described in
• Smart device and IoT selection guidelines this case study, Arup has been asked to expand the
• Implementation strategy. project to consult on further core capacities of the
R&D facility such as soft services including catering,
DEFINING THE SCOPE cleaning, waste etc. This work will help the client to
build a business case for an R&D-wide digital twin
We took care to define a high-level design for infrastructure, which would augment analytics and
the campus digital twin at this early stage of the simulation components of the R&D digital twin with
project. This design is built around capability blocks even richer information. ■
(Figure 34), which schematise the most important

104 CHAPTER 2 PROPERTY

 MEET THE TEAM

FIGURE 34
K ATE JAR R ITT ED R OUS-EY R E
Opposite left
Associate Director Senior Consultant
Roadmap towards the final digital
twin infrastructure
BEATR ICE NASSI ANDY K ER VE L L
Senior Consultant Associate

105
C A S E S T U D Y: H O N G K O N G

Neuron Digital Hub Platform: a data-driven approach to

achieve sustainability

AUTONOMY

INTELLIGENCE

LEARNING

FIDELITY

LEVEL 3

3
3

B A C K G R O U N D A N D M O T I VAT I O N

Modern buildings are complex amalgams of different the platform was expanded to include machine
technologies and systems. Those who run these learning and artificial intelligence capabilities, before
facilities face many challenges, including getting finally maturing into a comprehensive smart building
up-to-date information on building assets, verifying management solution.
the accuracy of operation and maintenance manuals,
checking equipment status, and reliably calculating The Neuron Digital Hub Platform is an intuitive and
building performance and energy costs. Although fully customisable visualisation tool to engage users,
building automation and energy management enhance building efficiency and optimise operational
systems have existed for some time, they tend workflows. One of the most innovative features of the
to focus on monitoring and providing alarm platform is the interaction between the building’s 3D
capabilities. But with the increasing convergence of BIM model and real-time data captured automatically
building systems, there is great value to be delivered from building management systems and HVAC
by a central analytics platform which can provide systems. This is achieved using open protocols
more insight from integrated data. including Building Automation and Control networks
(BACnet) and Modbus. By clicking on an item in
Arup’s Digital Services team has been researching the 3D BIM model, operators can visualise specific
and developing an integrated ‘BIM + IoT + parameters and statistics for the item with the help of
Analytics’ platform for smart buildings.90 The interactive and responsive dashboards, dramatically
outcome is the Neuron Digital Hub Platform, a cloud- rendering building performance more transparent.
based, centralised management platform which can
connect disparate building systems and equipment, 3D BIM also improves the capability of real-time
making them easily accessible and facilitating their monitoring of building system parameters. Capturing
operation and maintenance. The project began environmental changes, the integrated platform can
investigating the potential of IoT technologies to generate timely responses to threshold triggers in
provide a more comfortable and customisable office abnormal situations, reducing energy consumption.
workspace. Through subsequent development phases The Neuron platform also brings a new dimension to
asset management, operations and maintenance ►

106 CHAPTER 2 PROPERTY

FIGURE 35 FIGURE 36
Top Bottom
A view of the Neuron A view of a section
platform dashboard of a floor area

107
CAS E STUDY: A RUP HONG KONG

by creating a digital twin for each building. With this

dynamic real-time view of the integrated systems
of the building, operators can quickly retrieve asset
information for decision making, diagnose problems
remotely and test proposed fixes before having to
apply them to physical components. The Neuron Digital Hub Platform has already
been applied to several pilot projects, including the
The Neuron Digital Hub Platform also offers further iconic Water Cube in Beijing.91 Through energy
smart features. Precise measurement of indoor usage optimisation and predictive maintenance,
environment parameters such as temperature, substantial energy savings and increased operational
humidity, indoor air quality and air contaminant efficiency have already been achieved. Additionally,
levels are provided by a network of IoT sensors and air pollution is a major concern in Beijing, and
equipment gateways. By using the computational a network of IoT sensors deployed in the venue
power of AI together with machine learning, the continuously monitor the indoor air quality
platform boasts diverse functionality to transform while collecting data to help optimise building
building operations, as follows: system operations.

Trend discovery and energy forecasting The techniques and technology embodied by
Large historical data sets can be analysed to uncover the Neuron Digital Hub Platform represent a
hidden patterns, in order to better estimate future culmination of Arup’s expertise and innovation in
energy usage and allow better planning. the built environment. The system is constantly
learning to improve its own performance, and the
Building system optimisation data collected will help our clients create better
The building management system can be controlled buildings and facilities in the future. The Neuron
and adjusted automatically based on statistical Digital Hub Platform puts the concept of smart
insights retrieved from historical data, leading to buildings on a new level: it will change the way
automated workflows, energy usage reduction and buildings are designed and constructed, operated
sustainable building performance. and maintained. ■

Fault detection MEET THE TEAM

AI facilitates early detection and timely resolution by
identifying anomalies in system data and verifying
MAR K CHEN WILLIAM LO
fault occurrences.
Associate Engineer

Predictive maintenance
DAY I WANG
AI helps to monitor and estimate the condition of MAGGIE JI
Engineer Engineer
equipment and its components by analysing usage
patterns, frequency of maintenance and operational
parameters such as vibration and acoustics. System
maintenance can thereby be scheduled, in lieu of
disruptive unplanned downtime.

108 CHAPTER 2 PROPERTY

VINCI FACILITIES

Digital twins for facility

management
Vinci Facilities are using the large
amounts of data stored in a digital
twin to improve maintenance
efficiency.87 On a typical system,
a technician may receive an alert
which reads “Fan broken, Office
14, Building 3”. What seems like
a clear message is actually very
ambiguous. What type of fan is it?
What parts are required? Where
exactly is the fan located? Using
the digital twin for the building, a
vast amount of information can be
accessed by the technician before
they reach the job, allowing them
to prepare fully for the task in
hand. A virtual inspection can be
carried out showing the exact type
of fan, the most recent maintenance
report, a video tutorial and an exact
nomenclature. The correct tools
can then be prepared to carry out
the job. The technician is given an
accurate photo of the location as
well as the most efficient route to
take to reach the job. All of this
combined reduces the time spent
working out what is required and
keeping the actual maintenance
time to a minimum.

109
11 0 CHAPTER 2 TRANSPORT
MARKETS

Transport
The concept of a digital twin is not well-established in the
transportation sector, which means that there is an opportunity
to take the term and use it to drive progress in a well-defined
set of areas.

ISABEL DEDRING GERARD CASEY

Global Transport Leader Senior Consultant
Arup Arup

INTRODUCTION POTENTIAL

What can digital twins bring to the transport sector? A digital twin is a virtual model used to tighten the
Perhaps the most promising idea is the ‘city digital feedback loop between design and execution. As
twin’ — a realistic simulation of a city, including the an industry with modelling at its core, this sounds
transport network and population. The twin should familiar and reassuring language. But the historic
accurately model changes with the passing of time, context paints us a picture both challenging and
given initial conditions, while being reactive to changes exciting — on the one hand, we in the transportation
and interventions, such as new transport services, new sector deal with unique, boutique projects every
housing developments or changes to fares. Whereas time, which might suggest a preference for single-use
a more familiar kind of digital twin at present might, modelling. But equally, we see trends and patterns
for example, aid the delivery of a discrete piece of repeating themselves time and again, making the
infrastructure, the city digital twin would have the prospect of reusable models highly valuable.
greater scope of delivering broader social and transport
outcomes. Furthermore, the city digital twin can be This tension between boutique and more flexible
maintained, updated and extended — potentially modelling has effectively produced two polarised
remaining relevant for as long as the city is inhabited. camps. The latter is typified by tech start-ups who
That new models and new data will change how we claim to have ‘solved’ transport modelling for all
operate is undoubtedly true; the question is, what role cities using new data and new models, the latter by
will digital twins play in this change? those in established quarters who maintain that ►

111
we must continue to treat each case study in Our clients task us with challenges like these daily.
isolation, using slowly evolving models and This explains why Arup is investing heavily in
fragmented, small data. So, where is the reasonable our own internal capabilities, developing state-
middle ground? Do we think that years of training of-the-art models together with clients, academic
data and a black-box machine learning algorithm institutions and others. These include rapid
is a reasonable replacement for a well-understood assessment tools for quantifying the impact of
method with a sample size millions of times smaller? widespread autonomous vehicle deployment in
For example, is a machine vision method for cities, or building multimodal agent-based models
tracking cars on a junction better than a traditional of populations with City partners. A dedicated R&D
(human!) traffic count? Are these even meaningful team is currently developing a hybrid model that
comparisons? Usually they are not, as they are generates and trains synthetic agents (people) in
focussed on somewhat differing questions. In this an entirely digital transport model and compares it
case the machine vision method will likely give us against real-world behaviours.
more than just the traditional vehicle count metric,
adding information about how pedestrians, cyclists, RISKS
cars and other vehicles use this space.
The holy grail of transport modelling is to accurately
The sector is fairly aware of the limitations of current model future scenarios for example, autonomous
models and data sources — they are essentially the vehicles, widespread city-centre pedestrianisation,
devil we know, while the Silicon Valley models with or road pricing. There is a significant barrier to this,
slick user interfaces and amazing visualisations tend however — digital twins have only become viable
to fall into the category of the devil we definitely thanks to new data sources, but these remain heavily
don't know. Transport is a fairly conservative sector biased. Transport for London, for example, began
for good reason — infrastructure projects require publishing open data on their services ten years ago,
some of the biggest decisions cities and countries while decades-old GPS technology has only seen
ever make. Would we be able to credibly face a widespread adoption in the past decade.92 While
public enquiry or local community and justify our useful, this data only captures a relatively short
building on the basis that “the model said so"? We period in our history — even if we had complete
struggle to do this already, and although our models access to all of this data, would the training sets be
undoubtedly have flaws, they are significantly more adequate? A digital twin might also struggle to help
transparent than a neural network trained on a us with predicting rare or unprecedented events,
petabyte of sensor data. such as widespread autonomous vehicle deployment.
While these negative points will be mitigated over
CHALLENGES & OPPORTUNITIES time, as the twin continues to run, these fundamental
constraints must be acknowledged.
Arguments around the type of model to use are
of course somewhat abstract. The more important LOOKING AHEAD
question is: can digital twins improve project
outcomes? Consider large scale station developments The concept of the digital twin is not well-established
— how do we engage local communities in a project in the transportation sector, which means there is
with strategic national importance to ensure equitable an opportunity to commandeer it and use it to drive
outcomes for all stakeholders? More broadly, how progress in a well-defined set of areas. These areas
do we facilitate a shift away from polluting personal must be problem- and outcome-focussed; technology
vehicles towards public transport? We can schematise for technology’s sake is always a mistake. Ultimately,
some challenges and opportunities found in Table 7. there is incredible potential in this domain — the
potential to move away from a reactive approach
to problems, and towards a proactive exploration
of opportunities. ■

112 CHAPTER 2 TRANSPORT

 CHALLENGE OPPORTUNITY

Traditional transport models are big, Moving away from by-project model creation
slow and crude. It takes time to build and towards a digital twin model adaptable to
calibrate models. Scenario testing is often many scenario testing could democratise
limited by what our models can actually transport modelling, reducing delivery risk
reflect, while running a model can take up and increasing stakeholder engagement.
to weeks. All this deters interested parties
from model deployment.

Because of data and computing limitations, Rather than calibrating our models at
we tend to model systems in an isolated discrete time steps (as seldom as once a
fashion, and at incompatible levels of decade on some major models), we will
granularity. For example, different means of move towards real-time feedback. As real-
transport tend to receive different models. world data arrives over time, our models
A further divide is between operational and will become more accurate. Since we now
strategic models. These models thus rarely have the data and computing capability, our
interact with one another, reducing their strategic models and operational models
usefulness in scenario testing. can merge, with a strategic model simply
being a long-running operational model.

We don’t treat people as individuals, Individual-level data will sources enable

but rather as an aggregate. This often us to capture each unique personality.
entails assuming rational behaviour with Unprecedented detail is now available,
perfect access to information, given enabling a behavioural economic approach
some behavioural goal. Clearly, this is an to users. However, effective ethical and legal
idealisation too far. protections are required in this domain.

TA B L E 7
Top
Challenges and opportunities
of digital twins for the
transport market

11 3
WHY MODERN INFRASTRUCTURE CANNOT DO
W I T H O U T T H E D I G I TA L T W I N

We have had access to

increasingly sophisticated
modelling of networks like
these for years, but the
opportunity to connect users of
the network (as real-time data
streams) to the infrastructure
of the network itself (asset
performance information) is a
sea change.

FIGURE 37
Right
Modernisation of Big Bend Power
Station in Apollo Beach, Florida,
U.S.A. The digital twin was critical
to allow visualisation and access to
the physical and engineering data
throughout design and construction,
and will be key to future operation
of the facility. Image courtesy of
Sargent & Lundy.

114 CHAPTER 2 TRANSPORT

DR NABIL ABOU-RAHME
Chief Research Officer
Bentley Systems

Infrastructure networks evolve over time, and, being

complex systems, are increasingly interdependent.
Managing infrastructure today is about more than
just optimising demand in line with available
capacity — it’s about how that capacity is affected
by adjacent systems like weather, ground conditions,
and systemic efficiency. It’s about how failure in one
part has unexpected and unintended consequences in
another part, and how we might plan for and mitigate
such risk. It’s about going beyond the business case
for isolated projects and determining, in this age of
sustainability, which interventions deliver the greatest
downstream benefits.

In the early days of my career, I saw these

interdependencies emerging, through the prism of the
transport sector. I observed convergences with energy
networks on the question of accommodating electric
vehicles, with water networks for the challenge of
storm water mitigation, and with distribution networks
in the context of post-disaster emergency relief.

We no longer need to perform real-world validation,

with all its attendant risks, costs, and frustrations.
Instead, we can now envisage hundreds of scenarios,
leverage machine learning to generate designs, and
optimise for multiple parameters.

All this provides the context for digital twins, and

why they are essential for infrastructure. Looking
at the latest hype-cycle reports from industry, you’d
be forgiven for thinking that digital twins suffer
from highly over-inflated expectations. But in the ►

11 5
DR NABIL ABOU-RAHME: PERSPECTIVE

infrastructure sector today, things are turning out the buildings to going inside and checking detailed
differently, in that digital twins are turning out to be a sensor measurements spanning back over a year.
natural point of convergence, rather than an intrusive However, the real innovation lay in the rapid creation
change. This is clearly no unwanted consumer trend of a fully georeferenced and integrated dataset,
encroaching on our territory. Rather, it’s a lifeline conforming to the Gemini Principles, which could
which will ensure our viability. be read, interrogated, and reproduced by different
analytical and visualisation tools.33 But this is a time
Digital twins enable users to visualise the asset for further development, not for satisfaction with what
of interest, check its status, perform analysis and has already been achieved; many more examples will
generate insights in order to predict and optimise materialise in the coming months.
performance. An infrastructure twin incorporates
data from many sources to build its repositories
— including drawings, specifications, documents, In predicting the next big
analytical models, photos, reality meshes, IoT
feeds, enterprise resource data and enterprise asset
breakthrough in digital twin
management data. usage, we are faced with an
Digital twins span a plethora of use cases across embarrassment of riches.
the infrastructure lifecycle. A performance digital
twin, for example, may spawn many project digital How will our reality modelling improve, bypassing
twins, which are then federated or merged back in. or validating the sensor reading by taking thermal or
Similarly, a project digital twin may mature into a visual scans of the asset as part of the synchronisation
performance digital twin. A robust understanding of — guarding the guards, if you will? Human-operated
asset breakdown structure will help in determining at drones are giving way to autonomous swarms and
what point a system is well-bounded. Each individual other forms of robotics in research facilities, but will
digital twin can mature to provide ever-greater insight, these soon be buzzing around an asset near you? Or,
through continuous synchronisation, improved digital take enhanced images from open satellite data: if our
context, and a detailed chronology of change for the finest technology can see a black hole event horizon
digital twin itself. Not only can they be created rapidly when looking into space, then we can surely spot the
and to a high degree of accuracy from the outset, odd defect in our infrastructure when looking back
but they can also be maintained as evergreen in a from space.
cost-effective way over the lifetime of the asset. By
focusing on the integrated data set rather than the tools When will quantum sensors detecting gravity waves
used to manipulate the data sets, practitioners become become a cost-effective way of mapping underground
truly data-centric and adept at augmenting the twin infrastructure, and our digital twin ecosystems
with new data sources as they arise. double in size and usefulness overnight? Why stop at
immersive visualisation when we can take semantics
and start talking to our digital twins the way we
Bentley Systems participated do with chatbots? Or, think of the tokenisation of
in creating an area-wide value over the lifecycle of an asset through smart
contracts, oracles and cryptocurrency assets relevant
digital twin in the UK, to that model. Will such instant tokenisation mean
that early stakeholders, such as designers, are further
replicating the Institute of incentivised by the prospect of skin in the game
Manufacturing building and for years to come, as the real value of their design
intention materialises?
Cambridge University’s West
To answer these questions and others, Bentley is
Cambridge campus. investing in a collaborative research network with
academic and corporate strategic partners, to continue
The goal was to demonstrate the usefulness of exploring the art of the possible for digital twins.
a digital twin for facilities management as well We are excited to share our vision of the future with
as to improve productivity and well-being.93 The professionals who dedicate themselves to advancing
visuals were impressive enough, from a flyover of infrastructure, and thus society itself. ■

11 6 CHAPTER 2 TRANSPORT
FIGURE 38
Top
The West Cambridge digital twin
demonstrates the impact of city-
scale digital twins on infrastructure
management and productivity

117
C A S E S T U D Y: M A S S M O T I O N

Digital twins in the transport industry

AUTONOMY

INTELLIGENCE

LEARNING

FIDELITY
1

LEVEL 2

H O W D O W E A U G M E N T D I G I TA L T W I N S T O REAL-TIME CROWD CAPTURE WITH

U N D E R S TA N D T H E C H A N G I N G I M PA C T O F P E O P L E MASSMOTION LIVE
O N T R A N S P O R TAT I O N FA C I L I T I E S ?

Tracking and measuring passenger traffic flow We now inhabit a world where everything from
can inform design and optimise both performance sensor data to sentiment analysis to CCTV footage
and customer experience in domains as disparate is channelled wirelessly to provide real-time
as public transport and concert hall design. An information on people’s movements and motivations.
understanding of the qualitative and quantitative Transport facilities can make good use of this data
characteristics of crowds in specific spaces aids asset in operational and long-term planning, but only
owners to ensure safety, comfort and efficient transit. with appropriate intermediaries. MassMotion has
therefore launched an ambitious research programme
Arup’s MassMotion software has continually — MassMotion Live — to incorporate this data into
innovated over recent years to become the industry- a digital twin workflow.
leading crowd simulation package we know today.94
Its BIM-compatible workflows and native 3D MassMotion Live will significantly enhance the
environment allow rapid testing, providing existing product to enable near-real-time capture
feedback on congestion and thereby allowing and visualisation of observational data, as well as
optimised design. to develop new behavioural models which better
simulate pedestrian motion in particular spaces and
MassMotion is used for such eclectic building scenarios. This will crucially involve the capture of
typologies as railway stations, airports, commercial live data to produce detailed records on the range of
buildings, art galleries, and stadiums, and makes use movement types in transport facilities. We
of various classes of input data to shape designs from typically design these facilities for commuter
a human perspective. This commonly includes a 3D populations, but the ability to extract data on tourists,
model of the space under consideration, along with off-peak users, and so on will increase realism and
information on travellers’ origins, destinations, and thus aid simulation. ►
any intermediate nodes in their journeys.

11 8 CHAPTER 2 TRANSPORT
FIGURE 39 FIGURE 40
Top Bottom
A MassMotion model at MassMotion model of a
neighbourhood scale rail station

11 9
CAS E STUDY: M AS S MOTION

D I G I TA L T W I N S : I N C R E A S I N G T H E I R A P P E A L T O
THE TRANSPORT INDUSTRY

As well as their eponymous primary function, Melbourne Metro in Australia, and Istanbul Metro in
transportation facilities often serve as hubs for retail, Turkey.96,97,98 Major interchange stations are usually
commerce and leisure. As focal points of urban designed with lifespans of 50 to 100 years in mind,
stimuli, these facilities also tend to evolve over time based on limited population projections based on
to meet new use cases. They undergo upgrading limited population projections (e.g. using census
and expansion, as well as adapting to modern data) and with a focus on commuters. Once they are
standards in areas such as inclusive design for age accompanied by their digital twins, however, these
or disability. Pedestrian modelling tools already facilities can be thought of less as static investments,
attempt to simulate these ever-changing conditions, and more as evolving spaces in line with their
but are often fatally hamstrung by incomplete data. actual use.
This means that currently, they are not particularly
attractive to clients — there is thus great scope for M A S S M O T I O N : A K E Y C A T A LY S T F O R C O N F I D E N T
CITY PLANNING
improvement and increased uptake.
Current patterns of urbanisation and migration
Most pedestrian models, including MassMotion, suggest that 80% of the world’s population will
are based on a combination of academic research live in cities by 2050.99 Ambitious transportation
into human movement and validation against networks will be vital to ensuring that cities operate
industry standards. One example of this is the in an efficient and equitable manner for residents and
Fruin standard for density used in facility planning visitors alike. Similarly, ensuring that these facilities
in the transportation sector, where various crowd are responsive to changing use patterns, and can
density levels are ranked and colour-coded based quickly adapt to new ways of moving and altered
on movement restriction.95 While this combination demographics, will be critical in giving
generally produces accurate representations of cities the confidence to invest in these
human movement, those representations are transformational projects.
commonly based on historical information. However,
demographic and behavioural trends are changing Since its inception, MassMotion has been at the
rapidly, and designers must be on top of these forefront of mapping and understanding the impact
changes. MassMotion Live gives planners and of human movement on transportation design. By
designers the option to use contemporaneous data harnessing new sources of data, MassMotion will
as the baseline for modelling, to understand context be a key input to digital twins in transport, not only
and user types on a case-by-case basis. In this way, through its state-of-the-art forecasting of facility
existing generic models become digital twins of usage patterns throughout the design process, but
specific assets, greatly increasing their appeal to also by linking with other tools in order to
asset owners. contribute to shaping a better world for the end users
who live in cities. ■
S H O R T- A N D L O N G - T E R M B E N E F I T S
MEET THE TEAM
The aim of augmentation of pedestrian models with
near-real-time data is not simply to create standalone IAN MACKENZIE
digital twins. Instead, it is to empower our design Senior Software
tools to provide inputs for federated digital twins, Developer
such that all aspects of a facility are captured for
EDITH CHOW
optimal testing. The MassMotion team will develop
Software Developer
data pipelines to ensure that designers and planners
can engage with all available data sources according
to project requirements, all while keeping a close eye LACHLAN MILES

on the security and regulatory contexts. Associate

Transportation infrastructure is always a major MICAH ZARNKE

investment for cities, with major recent examples Technical Lead

including such as High Speed 2 in the UK,

12 0 CHAPTER 2 TRANSPORT
FIGURE 41
Top
MassMotion modelling of
a transportation hub and
public realm

FIGURE 42
Bottom
MassMotion Live research
initiative into computer
vision people tracking

121
C A S E S T U D Y: R E A L - T I M E E M O T I O N A L S E N S O R S

How can we use biometric data to determine user experience?

AUTONOMY

INTELLIGENCE

LEARNING

FIDELITY
1

LEVEL 2

2
2

INTRODUCTION performance. Understanding how citizens experience

their built environment can help us to develop and re-
Arup teams in Hong Kong and Canada are building develop a smarter city at every scale. This will allow us
a digital twin for airport planning and operation, to create dynamic and adaptable spaces, buildings, and
dubbed the Total Experience Lab (TEL): a tool which urban systems which are capable of accommodating
synthesises three existing modules and one original the changing behavioural patterns of users. This
technology. The TEL draws on various platforms to requires the definition and quantification of a person’s
generate, gather and analyse passenger behavioural experience, considerations often overlooked by
data. The platform combines MassMotion, Crowd- modelling systems which measure behavioural patterns
sourcing and Mobility Mosaic to create a real-world such as footfall, while ignoring the complex nature of
visual simulation. It then overlays data from the environmental experience. Current approaches leave
recent Emotional Mapping initiative, using real- us with potentially inaccurate data, which can lead to
time biometric data to map emotional states as users ineffective solutions.
interact with their environment.
TEL uses individual experience and the changing
The aviation industry has expanded dramatically behavioural patterns to inform dynamic environments,
over the past twenty years, and airport terminals are putting the user at the heart of design. This is
busier than ever. Passenger numbers rose from 1.4 particularly important in airports, spaces which
billion globally in 1998 to 4.6 billion in 2019.100 Such often deal with massive passenger flows within very
passenger traffic growth combined with evolving tight operational constraints, creating high-stress
passenger demands obviously has repercussions for environments which can even deter passengers from
the level of service provided by airport facilities. flying at all. A fuller understanding of passenger
Operators will need to adapt to the growth by experience can help to alleviate pain points, improving
expanding capacity, firstly by physically scaling up facilities and operations for the benefit of users, staff
airport facilities, but also by optimising operations — and airlines.
with the aid of digital twins.
With the Emotional Mapping initiative as its
In addition, design and operation considerations now centrepiece, the TEL can elucidate passengers’
see user experience as a primary factor in successful emotional status and cognitive states throughout their

12 2 CHAPTER 2 TRANSPORT
journey. Overall, operators can keep user experience
at the forefront of their consideration, with near-real
time feedback. TEL combines the following four FIGURE 43
technologies to achieve this: Above
Emotional Mapping:
Monitoring biometric data
MASSMOTION in response to the user
experience in the VR
MassMotion is among the most advanced pedestrian environment

simulation software packages. It allows designers,

operators and authorities to test design and
operation alternatives without limit, using crowd
modelling technology. In conjunction with an
experiential design component, MassMotion can be
used to optimise user experience, for example by
management of passenger flow.

CROWD-SOURCING DESIGN FEEDBACK

Crowd-sourcing is a virtual platform which allows

designers and operation managers to capture user
journeys in the virtual environment. It allows users to
explore design options in a game-like environment,
and simultaneously to capture users’ behavioural data.
This application is widely used to optimise design. ►

12 3
C A S E S T U DY: R E A L -T I M E E M O T I O N A L S E N S O R S

MOBILITY MOSAIC

Mobility Mosaic is an app that collects location The fusion of the operational digital twin and
information and intelligently identifies trips and the human-centric digital twin enables all
modes of travel.101 Combined with insight from stakeholders to understand the impact of any
Arup’s human factor specialists, the app presents an operational change to passenger experience. The
unprecedented opportunity to understand preferences focus is not only on optimising operation, but also on
and behaviours. Mobility Mosaic is heavily used by offering a dynamic multiple-criteria decision analysis
Arup to consolidate real passenger and pedestrian to all stakeholders, in order to reveal the impact of
experiences in order to design transport services that any decision.103
meet genuine needs.
This digital twin will allow designers and operation
EMOTIONAL MAPPING managers to discover previously unexplored solutions
without risk to operations or passenger experience.
The ultimate goal of Emotional Mapping is We want airport authorities to have access to the best
to understand individual users' subjective possible real-time data, putting them in a position
experiences of spaces. Since passengers’ emotional to face the ever-evolving nature of passenger needs.
responses constitute their true experience of their Further, new technologies will in future make it
environments, the project aims to render a real- possible to analyse the overall performance of an
time map of passengers’ emotional status as they airport terminal.
interact with a facility or process.102 This is achieved
by collecting biometric data and conducting post- VR technology, with its ability to simulate complex
survey and real-time survey analyses. Data can situations, also offers researchers an unprecedented
pertain to cardiovascular or respiratory systems, opportunity to investigate human behaviour while
electrodermal activities, muscular systems and brain trialling various structures and systems. This will
activities, as well as analysis of body language or dramatically increase the likelihood of optimal
behavioural responses. Information of this type results from the final built structures.
is superior to conventional methods which use
self-reported questionnaires, removing the user Finally, the project will also contribute to the rise of
from the environment and requiring them to take a ‘smart citizens’, connected interactively to responsive
retrospective evaluative stance, which is necessarily environments. The greatest challenge in all of this
inhibited by perception and experiential variability. is that people experience emotions differently, based
By failing to describe the unpredictable and ever- on factors including but not limited to culture,
changing nature of human experience, current age, gender and metabolism. EM must be honed to
approaches fall short of providing accurate and accurately identify emotional experiences in the face
reliable data upon with which planners, designers of individual variation. ■
and authorities can make decisions.
MEET THE TEAM
The TEL integrates all of the above platforms,
creating a human-centric digital twin which ALI JABBARI LACHLAN MILES
provides a real-time replica of passengers and their Aviation planner Associate
interaction with their environment. Such a digital
twin reflects passengers’ real-time experience,
J U S T I N T R E VA N MEI-YEE MAN O R A M
as well as generating new scenarios when the
Associate Principal Associate
environmental context changes, physically or
digitally. With this data, human-factor analysts can
address problems as they are identified, running SANKAR VS

them through the TEL to evaluate alterations from a Associate Director

user experience perspective.

124 CHAPTER 2 TRANSPORT

FIGURE 44
Top
MassMotion's pedestrian
simulation and crowd
analysis tools for a
transit station

FIGURE 45
Bottom
Crowd-sourcing
design feedback: a user
view of the interactive
3D environment

125
DNV GL

Determining the health of ships

DNV GL has developed a methodology for hull
condition monitoring incorporating the ship’s digital
twin, a virtual model prepared during the design
stage.104 As part of the Nerves of Steel Extraordinary
Innovation Project, DNV GL — Maritime has been
investigating ways of calculating wave damage to ship
hulls. Significant amounts of time and money are spent
by ship designers and shipyards in preparing analytical
models to document the design of ships and offshore
structures and to perform simulations.

The use of this virtual model as a digital twin during

operation of an asset would allow the visualisation of
key components and improve the understanding and
control of the long-term effects of operation on the
ship’s structural and functional components. Such
a digital twin could aid inspection and maintenance
planning, extending the lifetime of the ship through
prevention to hull damage.

WILLOW & TRANSPORT FOR NSW

Creating a data-centric network

Transport for New South Wales (TfNSW) is
implementing Digital Engineering (DE) to connect
emerging technologies with reliable structured data.
Willow is helping them to resolve variations between
current standards, provide gap analysis, specialist
advice and implementation on global best practice for
DE standards and advise on implementation.105 While
success with DE has been realised within industry,
implementation is more complicated when applied
to an intricate multi-modal transport network.
By building their assets twice — first virtually
then physically, TfNSW can create 3D models to
connect a massive amount of data in the planning,
detailed design and construction stages of projects
— improving coordination and oversight during
construction and efficiency during the operation and
maintenance of their assets.

DE brings together systems and related data sets into

a common data environment, connecting the static
data of asset information & documentation with live
data of IoT devices, control and visualising systems
them in a geospatial context (such as GIS mapping or
3D models in BIM).

12 6 CHAPTER 2 TRANSPORT
KONECR ANES

New standard for the Crane Industry

Konecranes, a Finnish company who manufacture and
service cranes and lifting equipment, has implemented
the Siemens digital innovation platform, including
MindSphere, an open, cloud-based IoT operating
system, and the Teamcenter® portfolio, the widely
used digital lifecycle management software, to
leverage the digital twin.106

Design, simulation and prototype testing organisations

often operate in their own silos, using out-of-
date processes. At Konecranes, a digital twin was
utilised to review data and provide feedback around
engineering, simulation and testing intent. Using the
Siemens platform, Konecranes has connected the data
to create a 360-degree view of how prototypes are
running and performing, correlating requirements to
real-world performance data. A closed-loop digital
twin framework, using IoT and PLM technologies, can
lead to quicker resolution of design issues. Leveraging
virtual sensor data in product simulations to provide
accurate results can lead to improved overall quality,
shorter prototyping phases and support downstream
for product lifecycle processes.

The Konecranes proof of value is one of the first

implementations of IoT to develop a framework that
connects and synchronises the virtual (engineering
design, analysis and simulation) and physical (testing
and operational reliability) worlds.

WILLOW & STRUKTON MOBILITIES

New level of smart rail maintenance

Willow and Strukton have come together to change

the way we view and manage rail systems.107 In the
heavy haul sector, disruptions and delays caused by
the condition of the network can result in hundreds of
millions of dollars in damage. Up until now the vast
amounts of data that are required to maintain and run
a railway are held up by individual silos, inaccessible
to the asset managers who are having to make
decisions without the full picture of information at
their disposal, and on critical decisions are having
to join the dots themselves. A complete digital twin
of the entire rail network can be viewed in real-
time, giving the asset manager a holistic view of the
system. Combining all the previously disconnected
information in one place allows the data to be turned
into valuable insight. Predictive maintenance and
network capacity improvements are just some of the
far-reaching impacts.

127
CASE STUDIES

APOLLO 13

NASA’s Original Digital Twin

Michael Grieves is commonly said to have coined

the term ‘digital twin’ in 2003.1 However, the
concept had become reality long before the modern
terminology. This realisation dates to the 1960s;
specifically to NASA’s Apollo 13 spacecraft.108 The
craft was designed and planned using simulations.
The models became vital in rescuing the three-man
crew, over 200,000 miles from Earth, following the
potentially disastrous explosion of an oxygen tank in
the craft’s service module.

The digital twin on Earth was used to model

solutions. NASA’s aim was to operate, maintain and
repair physical systems while they were in space.
The success of the mirrored system during the
Apollo 13 mission led the agency to develop more
advanced digital twins for all subsequent missions.

12 8 CHAPTER 2 TRANSPORT
ROLLS ROYCE & SVITZER

Operating full fleets of

autonomous vessels
Rolls-Royce and Svitzer - the
global towage operator - have
successfully demonstrated
the first remotely operated
commercial vessel in Copenhagen
Harbour in Denmark. Manoeuvres
by one of Svitzer´s tugs, the 28
metre long Svitzer Hermod, were
controlled remotely by the captain
of the vessel who was stationed
at the remote base at Svitzer
headquarters.109 The vessel was
berthed alongside the quay side in
Copenhagen Harbour, undocked,
turned 360°, and piloted it to
the Svitzer HQ, before docking
again. The Svitzer Hermod is
equipped with a Rolls-Royce
Dynamic Positioning System, the
key link to the remote-controlled
system. Using advanced software,
a range of on-board sensors
combine different kinds of data
which is transmitted to a Remote
Operating Centre (ROC) from
where the captain controls the
vessel. Input from experienced
captains went into the ROC design
to place the different system
components in the optimum
position to give the master
confidence and control. The aim
is to create a future proof standard
for the remote control of vessels.

12 9
O C A D O S M A R T P L AT F O R M

Automated storage
Ocado fulfils over 300,000 orders a week and delivers
these with 99.8% order accuracy. In its highly
automated warehouses, the most sophisticated of
their kind in the world, thousands of bots collaborate
like a swarm on a giant grid called ‘The Hive’ to
pick an average 50-item customer order in under 5
minutes. Ocado builds high-fidelity digital twins of
its warehouses and other parts of its business to de-
risk the physical world and uncover paradigms that
might otherwise remain out of sight. These digital
twins enable Ocado to evaluate new algorithms, test
production software before it is deployed, optimise
systems, and even generate data to train ML models
before the real data exists. Ocado is also working
on plans for “living labs” where they, and other
organisations, experiment before autonomous vehicles
and related technologies are widely adopted on the
public roads and in smart cities. This will enable testing
of the provision of smart services in real-life scenarios
such as offices, hospitals, shops, schools, universities,
and residential properties. These ‘living labs’ will
provide digital spaces for organisations to get their feet
properly wet before the Industry 4.0 tide rolls in.

MX3D & AUTODESK & ALAN TURING INSTITUTE &

ARUP

3D-printed smart bridge

An expansive collaboration of partners, encompassing
mathematicians, structural engineers, IoT specialists
and computer scientists, has come together in the
Netherlands to produce a true original. MX3D, with
help from Autodesk, the Alan Turing Institute and
the Amsterdam Institute for Advanced Metropolitan
Solutions, produced both a prototype and final version
of the world’s first 3D-printed bridge, spanning twelve
metres and printed with stainless steel.111 The bridge
was designed and installed with a fully functioning
digital twin, delivered by the Alan Turing Institute
and supported by Autodesk’s cloud services, which
support the bridge in collecting and processing
data. AMS will then apply the data, which concerns
structural factors such as strain, displacement and
vibration under load from pedestrians, cyclists and
motor vehicles. Simultaneously, environmental factors
such as air quality and temperature are measured.
These measurements inform maintenance scheduling,
as well comparison of the performance of real and
digital structures. This will allow designers to
create better 3D printed structures in future,
informed by the data collected. The bridge will
span one of central Amsterdam’s many canals, the
Oudezijds Achterburgwal.112

13 0 CHAPTER 2 TRANSPORT
I B M & A I R B U S & P L A N AT I R T E C H N O L O G I E S

Skywise digital platform at Airbus

As part of its digital transformation initiative, Airbus
and their partners have developed Skywise — a
platform built to make available all the company’s
data in one place.113 It is accessible to employees
and suppliers, and has several levels, ranging
from servers, to data management applications, to
applications tailored to specific processes. At the
core of the programme is its ability to extract and
link information from heterogeneous sources, which
were not originally intended to be compatible. Airbus
were able to demonstrate this to be possible within
three days. Next they created a system to discriminate
data types according to importance, accounting for
confidentiality levels. The system has yielded a 30%
reduction in ground time for aircraft at airports which
are not their usual bases of operation.

Skywise has swiftly provided the company with

a deeper understanding of factors such as part
durability; it has made maintenance scheduling more
efficient; it has allowed the company to anticipate
supply chain complications, and may even be able to
inform optimal cockpit design. Of particular note is
the decision by Airbus to make Skywise open to its
many suppliers, an openness that is paying off.

131
132 C H A P T E R 2 WAT E R
MARKETS

Water
The digital revolution is transforming the water sector and
technological developments are opening new possibilities, but it
is an integrated approach that will secure the future of water.

RICHARD BOWDEN
Digital Software, Insight and Operations Leader
Arup

INTRODUCTION expertise that enables us to assess risks and support

our clients in taking a strategic approach to water
Our most precious resource is at significant risk and usage and management.
we must act now. At Arup, we have been developing
our ‘Design with Water’ framework, which draws Underlying all of our work in water is the goal
on our experiences of designing, implementing and for the knowledge that we generate to both enable
maintaining water infrastructure around the world.114 and accelerate the realisation of the water related
Design with Water is strongly underpinned by our UNSDGs.115 The digital twin approach will be an
research activities. important element in building this knowledge.

The framework places an understanding of the local T H E C H A L L E N G E S I N T H E W AT E R B U S I N E S S

water cycle at the centre of responses to wider local
issues, such as economic development, food and There are multiple global challenges around the
agriculture, community, and energy use. In order to provision, quality and accessibility of water, as well
account for the whole water cycle, Arup conducts as the impact of human activity on this precious
research into multiple aspects of water management. resource. In most cases, these challenges are a
From flood modelling assisted by machine learning, mixture of digital, physical and biological factors, but
to working with partners to better understand water also social issues, driven by the advance of the Fourth
resilience in cities, we are continuously broadening the Industrial Revolution.116 ►

133
A strict, discipline-driven classification is no longer W O R K I N G W I T H A N D F O R N AT U R E
possible and an isolated approach to developing
solutions is to be discouraged. To ensure a sustainable We need to ensure that the impacts of our activities
future for water, we need multidisciplinary teams with on the environment can be minimised or even
broad expertise, and we must engage a range reversed. Design and development should be nature-
of stakeholders. aligned, nurturing biodiversity and ecosystems, and
simultaneously enhancing our health and wellbeing.
W AT E R F O R C I T I E S A N D C O M M U N I T I E S Looking for innovative solutions and technologies,
while taking inspiration from nature, can help us to
Growing demand for water in expanding cities calls develop more effective — but not necessarily more
for insights on the upgrade and future-proofing of complex — solutions.
existing assets, and on building new state-of-the-
art assets with an extended lifespan and which D I G I TA L I N N O VAT I O N S H A P I N G T H E W AT E R
INDUSTRY
support more sustainable water use patterns. Cities
and communities must also learn flexible water-use We need to continue investigating how digital
strategies, to enable themselves to stay afloat both innovation can support management in each of the
when there is too much water, and when supplies water cycle phases. The potential impacts will be
are limited. significant, and there are still many applications yet to
be discovered. Digital technologies can have associated
ENABLING FLOW IN VIRTUOUS CIRCLES risks; and digital twin technology can provide important
information on how best to mitigate these risks.
Water carries both valuable compounds and
pollutants, often at the same time. Closing resource D I G I TA L T W I N I N W AT E R
loops will help recover value and prevent pollution
from spreading. We need to understand how we can Working from the definition of digital twin in the
best implement the circular economy approach, which introduction to this report and applying it to the
technologies are required, and where the opportunities Water business sector, we need to be mindful of how
for cross-sector collaboration exist that will amplify modelling approaches have been used previously.
the potential benefits.
The water sector has achieved a high level of fluency in
C A P I TA L , G O V E R N A N C E A N D F I N A N C E digital modelling application as illustrated by:

The rise of megacities and the increasing power of • Hydraulic models, some extremely detailed and
local authorities is influencing decision processes, sophisticated, support planning decisions, problem
ownership models and investment capacity. solving, what-if analysis, forecasting and even
A knowledge of novel financing and funding regulatory reporting
mechanisms, methods for capturing the value
of social and natural capital, as well as approaches to • System-level models, representing bulk
community engagement are some of the water grids, major floodplains or water
crucial issues to be explored in the support of resource basins, are used to vet operational or
sustainable development. development decisions

TECHNOLOGIES TO ADDRESS MAJOR • Monte Carlo analysis applied in hydrological

CHALLENGES
assessments take advantage of now readily
Emerging pollutants, oceans full of plastics and accessible massive computing power117
microplastics, anti-microbial resistance being transferred
through water: these are just some of the challenges that • Smart networks are coupled with machine learning
require novel technological solutions. Many of them will to identify risks, predict failures and estimate
need to be applied in different industries, and as such, future performance.
cross-disciplinary collaboration will be vital.
With this firmly demonstrated success with Water
Authorities, the question is not where to go next, or
even how far, but only how soon can we get there ►

13 4 C H A P T E R 2 WAT E R
135
with assistance from digital twins? The ultimate
outcome would be a city-scale or regional-scale digital
twin which combine all the above elements into a
powerful tool for decision making by people at all levels
of society; government, service providers, communities,
businesses and individuals. This can be achieved by:

• Extending and connecting system models

to represent the interdependent effects in the
real world, such as the impact of a proposed
intervention or development on water regimes,
urban heat, ecosystem health, carbon footprint,
population movements and the economy,

• Using sophisticated hydraulic models as the engine

behind visualisation technology such as augmented
reality, virtual reality or 3D interactive displays as
part of a communication strategy,

• Making water information more personal, and

personally accessible, for smart consumers to
participate in water and wastewater network
transactions and operations.

With the appropriate planning, digital twins could

provide a platform for building social equity through
combining open access to system information with
crowdsourcing platforms to fund improvements that
increase equity. The digital twin can provide the
replica(s) demonstrating the societal benefits from the
funding investment.

CROSS-DISCIPLINARY OPPORTUNITIES

Digital twin enables cross-disciplinary opportunities

within the water sector. Within the water sector,
water authorities not only provide water for public
consumption, but also can provide services related to
water quality and pollution, water conservation and
fisheries management. In delivering these services,
energy is consumed by an authority and the related
operating costs and environmental impact is under
increasing focus in their efforts to improve efficiency.
The digital twin provides a capability to analyse the
energy usage under different scenarios by replicating
the assets and their energy profile that are managed by
the authority.

Related to water, roads authorities are responsible

for the maintenance and construction of roads and
motorways. From both construction and maintenance
perspectives, the management of surface water
runoffs and flooding is important for the resilience and
availability of the roads to the public. The digital twin
can provide a means to simulate and plan for different
rain and water flow patterns on a real-time basis and for
longer term water planning. ■

13 6 C H A P T E R 2 WAT E R
137
W AT E R T W I N S : W H AT C A N T H E Y D O F O R U S ?

The water sector is relatively

mature in the digital space, with
most entities using technology
in digitally fluent ways.

FIGURE 46
Right
Pool visualisation for
prospective freshwater pool
in Hudson Bay, New York

13 8 C H A P T E R 2 WAT E R
E L A I N E PA N G
Australasia Water Skills Leader
Arup

A wealth of asset and performance information

is commonly shared among multiple users;
sophisticated models are used to understand and
predict system behaviour, and interlinked databases
and models make possible all permutations of
analysis and interpretation. Software platforms are
GIS-based, with real world foundations for creating
insight. Continuing technological advances mean that
massive computing power is now readily available
to support advanced assessment techniques, whether
using parallelised desktop machines, supercomputers
or cloud-based computing.

When considering what is to come in this space,

we can first look at what has already been created:
models exist which represent the movement of water
through pipes, in channels, overland, and beneath
the earth’s surface. They may be detailed models
based on surveyed or as-built asset information, or
they may be conceptual models representing large-
scale movements of water, as in regional water grids.
Verification and calibration of models can be based
on disparate sources of information such as pump
operating cycles, rainfall radar imagery, river levels
and so on.

The models are used to support planning decisions,

solve performance and capacity problems, and
undertake what-if analyses. Multiple what-if scenarios
can be assessed in a relatively rapid manner to vet
important decisions on selecting the most robust
solutions for an uncertain future. Some pilot projects
have combined hydraulic models with gaming
concepts to test impacts on user behaviour, while
elsewhere, gamification has been tested for system
level decision making.

When combined with machine learning and sensor

data, the calibrated models support predictive
analytics to detect performance trends which
signal asset failure, as well as forecasting of future
performance. In some instances, the models, once
calibrated to real world data, become regulatory
instruments for reporting information such as sewage
overflows, as a proxy for in situ monitoring.

Comparing the water sector’s current technologies

with the received definition of a digital twin ►

139
E L A I N E PA N G : PE R S PECTI V E

— a digital replica of actual assets, processes, people, N E X T- G E N E R A T I O N C O M M U N I C A T I O N S T R A T E G I E S

places and systems — we can see that the water sector

is already there. Building upon current practices, and Visionary projects need visionary communication
considering global trends, some possible directions are tools. Sophisticated hydraulic models could drive
proposed for the reader’s consideration: visualisation technologies such as augmented reality,
virtual reality or 3D interactive displays as part of a
INTERCONNECTED, AGNOSTIC MODELS communication strategy for topics such as proposed
capital works, community education, or anticipated
Water influences life in so many ways. It makes changes. Typically, what may appear to be a new
sense to collaborate across industries to create a paradigm is built on the strength of that method being
city-scale, precinct-scale or regional-scale model demonstrated elsewhere, requiring an initial leap of
which represents the dynamic influences of one sort faith from the first adopter. Visualisation, coupled
of change on other aspects of the modelled area. with digital twin technology, can provide the required
For example, a proposed change of land use from confidence for decision making and agenda setting.
industrial to open space could be analysed in terms
of effects on flood resilience, urban heat, population These advances in digital water can be speeded by
movements, ecosystem health and the economy the further development of existing concepts such as:
through a series of linked models (a system of
systems), accessible from a central interface. Stakeholder collaboration
Acknowledgement of the potential shared value,
S O C I A L LY S M A R T D A T A S H A R I N G and standardised means of articulating that
shared value, would support better collaboration
Today’s consumers are sophisticated and expect more among stakeholders and thus the usefulness of
access to information. The water sector could adopt interconnected digital twin models. Work has already
the open data concept, adding value by personalising been done to monetise benefits, such as resilience
the experiences of each individual. Such a win-win impacts, blue green infrastructure outcomes and
situation would enable consumers to make smarter increasing service measures.
choices with more control, while water businesses
would gain improved insight into user demographics Open data
and more direct conversations on the issues that Further collaboration is needed on aspects such as data
matter, such as water usage, willingness to pay and standards, privacy issues, data warehousing logistics
environmental issues. and information security methods for making and
recording secure and transparent transactions.
The open data concept supports extension into many
areas such as building social equity. Combining open Model interconnection
access to system information with crowdsourcing It should be possible for models to be linked into a
platforms could be used to fund improvements that system of systems using heterogenous modelling
increase equity, leveraging willingness to pay. Crowd software, combining the strengths of each.
sourcing provides a mechanism to take the decision Standard IT architectural practices such as ReST
straight to the users; surveys have found that the (Representational State Transfer) were created for
public is willing to pay for environmentally beneficial this scenario and are widely used today.
actions not normally supported through government
funding, such as for cleaner waterways. Open data and Visualisation
crowd sourcing would short cut the political process Current techniques, originally drawn from computer
requiring the election of sympathetic politician to games such as third person shooters, may be
enact laws and policies which lead to funding of the complemented by alternative techniques, which can
desired outcomes. Present-day decision makers are be sourced from graphic design approaches used in
predicting or assuming what the community is willing modern marketing strategies. These include more
to fund, but some projects have been funded initially appealing graphics for information design such as
for feasibility studies and ultimately for the capital hand-drawn or painted media, use of typography
cost of construction through crowdsourcing. to enhance or emphasise information and linking
multiple layers of meaning through multiple layers
of visuals. ■

14 0 C H A P T E R 2 WAT E R
 FIGURE 47
Above
NSW digital twin

FIGURE 49
FIGURE 48 Right
Below Another +Pool visualisation
NSW Digital twin prospective in New York

141
W H AT A R E T H E D I G I TA L T W I N O P P O R T U N I T I E S
F O R S Y D N E Y W AT E R ?

Australia has a unique set of

challenges: water-poor and a hot
and dry climate.

142 C H A P T E R 2 WAT E R
MIKE WASSELL
Head of Operational Technology
Sydney Water

Being water-poor with a hot and dry climate,

Australia faces a unique set of challenges. Sydney
Water is the largest water utility by population
in Australia, supplying over five million people.
We have a long-term vision to be able to run our
entire integrated water services from a digital twin
platform. With digital twins, we will be able to plan
our assets more efficiently; which is particularly
important given Sydney’s population is growing at
2% per annum. If we can avoid building an extra
10% worth of assets because we can save 10% worth
of water, that is very significant for us.

At Sydney Water, we have a tradition of incorporating

all things data, automation, and increasingly IoT.
IoT includes a range of projects which are helping to
improve our customer experience and enable more
targeted decision making. One of the areas we are
actively exploring is building digital twins, with a
proof of concept developed for one of our largest
network assets.

For us, a digital twin is a dynamic virtual

representation of a unique physical object, process
or system across its lifecycle. It’s not just a static
digital representation.

Digital twins can provide us with real-time insights

into what is going on in our systems. Traditionally,
many water companies build hydraulic models based
on theoretical principles. Testing hydraulic models
using sensors on real-world assets, indicates there
can be anything between 10% to 35% inaccuracy.
With a digital twin, we have an opportunity to ►

14 3
MIKE WASSELL: PERSPECTIVE

significantly improve the accuracy of our models. P R E V E N T I N G W AT E R L E A K A G E

Using real-time data from sensors, we can enable
predictive analytics and visualisation to make more Water leaks are another big problem for us where
targeted decisions about our assets, improving our digital twins can help. Water networks leak for
ability to integrate services across our network. To many reasons – stress, age, drought and shifting
do this we are looking at a range of applications. soil conditions.

A VIRTUAL PROOF OF CONCEPT

We estimate that 20% of all
Our proof of concept virtualised a large water
pumping station – one of the largest in the southern
properties have a leak, which
hemisphere. We used 3D cameras to capture the means a great deal of water
physical infrastructure, and then mapped asset and
supervisory control and data acquisition (SCADA) is unnecessarily lost. This is
data to that scan. This gave us a 3D representation
which we can use for design and for real-time
where digital meters (smart
visualisation of our operations. Our next step will be meters) can provide much
to incorporate predictive maintenance.
clearer data.
T E A C H I N G S E N S O R S W I T H D I G I TA L M E T E R S
Where we install digital meters, we can measure
Another application is determining the optimal supply every 30 minutes and quickly inform
combination of sensors we should install in a customers who have leaks. We save water, and
network to enable accurate prediction. Part of the our customers save money. Digital meters are not
way we approach this is to start with far too many mandatory, but with an uptake of just 10%, our
sensors, before whittling down to understand the insight into reducing water leakage would be vastly
right balance of digital and analogue sensors. improved. Currently we do not know much about
individual usage patterns - how much water is used
We recently carried out a trial with 10,000 properties externally (e.g. watering gardens)? How much is
in Liverpool, New South Wales to look at the used on washing machines? How much is used for
number of sensors required to predict overflows. We showers? – insights into these usage patterns could
discovered to prevent overflows reliably, we require help us to be much more effective in how we manage
traditional float sensors in every second or third supply and demand.
maintenance hole to detect sewer blockages.
With digital twins, the biggest value for Sydney
Conservatively, we can install analogue sensors Water will be for our long-term planning and asset
in far less locations to achieve the same outcome. management. By using empirical system data to
The idea is for the digital float sensors, to teach the guide our design and investment decisions, we will
analogue sensor how to predict an event. In the past, ideally be able to reduce the overall cost of ownership
we’ve looked after our trunk networks using SCADA as the insights are fed back into city planning.
technology, and our customers were essentially our There are also shorter-term benefits for design, most
‘sensors’ in the distribution networks – notifying us obviously the ability to preview designs in 3D, to
when things went wrong. immediately identify construction problems. This
will be invaluable in both greenfield and brownfield
With digital twins there is a real opportunity to areas as our city grows.
change this – to proactively manage an overflow
before it happens. Predictive capabilities – if reliable
– will have a huge impact on how we can proactively
manage our network.

14 4 C H A P T E R 2 WAT E R
C O L L A B O R AT I O N W I L L B E T H E K E Y

Our long-term vision is to be able to run our Collaboration between utilities and customers
entire integrated water services from a digital twin will be incredibly important to get the most out of
platform. By exploring applications to improve the capabilities of digital twins. We are currently
our real-time insights, visualise a virtual reality of working with the NSW Government to put in place
our services and support our customers to make data sharing agreements across utilities, as well as
decisions around their water usage, we can set discussing privacy, security and the challenges
Greater Sydney up for a resilient, water these bring.
sensitive future.
We are moving into a new digital era and are excited
We’ve already started installing sensors and are for the opportunities it will bring for our city. ■
talking to the market about the appropriate platform
to build digital twins from single assets to capturing
whole integrated systems.

14 5
N E W C A S T L E & N O R T H H U M B R I A N W AT E R

UK’s Water Digital Twin to Help City

Plan for Disasters
A partnership between Newcastle University and
Northumbrian Water Group is developing a digital
twin of Newcastle — a computer replica of the entire
city.118 Much of the data comes from the university’s
Urban Observatory project, which has sensors all
over the city monitoring pollution, water quality
and biodiversity. Using the digital twin, real-time
resilience testing could be run to see how a city’s
infrastructure would perform in the face of climate
change and population growth. In 2012, a month’s
worth of rain fell on Newcastle in two hours during
evening rush hour, causing £8m worth of damage.
In future, the digital twin would allow the city to
react in real-time to such weather events and to test
large numbers of potential future emergencies. It
could predict which buildings would be flooded, the
infrastructure that would be closed, and the hospitals
that could be affected. In time, this model could be
applied to any city with accessible data. The project
is part of a pioneering initiative by the Rockefeller
Foundation to help cities become more resilient to the
challenges of the 21st Century.

SWAN H20

A Water Consortium for Digital Twins

The Smart Water Networks Forum has established
the Digital Twin H2O Work Group, with the
intention of reducing isolation and accelerating the
adoption of digital twin technology in the water
sector.119 Its strategy is to coordinate the efforts of
utility providers, solution providers, engineering
firms, government, and academia. This coordination
will give disruptive impetus to the industry,
increasing operational efficiency and cutting
prices for consumers, while increasing the global
availability of water. The consortium is organised
into three working subgroups: (1) Holistic Digital
Twin Technology Architecture, (2) Outcomes and
Applications, which determines priority areas for
application of the technology, and (3) Case Studies
and ROI Calculations — deriving useful knowledge
from digital twin success stories elsewhere, for
adoption by the consortium. Organisations may
also become observer members, which entitles them
to receive various updates from the Work Group.
SWAN H20 held its first Workshop in May 2019
and has so far been endorsed by prominent firms
including AngliaWater, AVEVA, KWR, GoAigua,
Jacobs and VCS Water Denmark.

14 6 C H A P T E R 2 WAT E R
INNOVYZE

Vision of Digital Twins for the Water

Industry
The water industry is currently waking up to the
potential of digital twins to orchestrate water
supply, wastewater, flood control, networked assets,
asset registry, and so on. These elements must be
aligned to function together effectively — especially
given the imminent supply challenges faced by the
industry. Putting operational data at the heart of
planning, design and operations will enable the best
possible decisions for consumers. For this reason,
Innovyze have created a digital twin to aid the
design, simulation, management and maintenance of
water assets.120 They aim to integrate their twin with
infrastructure assets, modelling and simulating using
real-world data streams to optimise performance. In
conjunction with InfoAsset Online, their new service,
they have opened information from utilities providers
to be shared with maintenance crews, contractors,
planning departments, regulators and consultants
to give them a more complete view of the asset
they are working with. The technology is already
used by Thames Water (UK), the Smart Tunnel
operators (Malaysia), East Bay Municipal Utility
District (California), and Sydney Water (Australia)
to provide guidance on leakages, flood risks, optimal
maintenance, and rehabilitation.

BENTLEY SYSTEMS

Smart Wet Infrastructure

Bentley Systems and Digital Water Works are
collaborating to provide digital twin solutions for
'smart wet infrastructure' to municipal and investor-
owned water and wastewater utilities worldwide.121
Wet infrastructure refers to water supply, treatment
and storage, water resource management, flood
management and hydropower. It also covers locks,
weirs, storm-surge barriers, guiding structures,
pumping plants, and related structures. Combining
real-time intelligence with spatial analytics,
the digital twin enables the simulation of water
distribution and sewer collection systems, providing
visualisation and analytics capability that can be used
to detect and isolate deficiencies, perform diagnostics
and troubleshoot problems. This geospatial platform
enables utilities to operate and sustain more resilient
wet infrastructure while maintaining compliance,
meeting appropriate levels of service at the lowest
total lifecycle cost.

147
14 8 CHAPTER 3
CHAPTER 3

Challenges and conclusions

Any large change to an industry requires overcoming
technical, legal, and cultural hurdles. This is particularly
true with emerging technologies like the digital twin.
New skills and talent must be brought in, new training
infrastructure must be created, and cultures must change.

14 9
CHALLENGES

Our research has revealed promising advances in the EU Generation Data Protection Regulation
the development of digital twins; notably, there in place, companies are now legally bound to
have been large investments from governments, greater accountability, and data collection must be
organisations, software companies and research planned accordingly.122 Location data, for example,
entities in recent years. Yet, as with any emerging is generally highly desirable, and when properly
technology, there are challenges we must overcome collected can be a major aid in smart city planning
– technical, legal, and cultural hurdles, for example. and development. However, data anonymisation
New skills and talent must be brought in, new remains a challenge.123 A recent study attempted
training infrastructure must be created, and cultures to gauge the ease of retrieving identities from
must change. Key risks in the current landscape supposedly anonymised data, and found that 99.98%
will need to be considered; these include big data, of Americans could be correctly re-identified from
security and economic returns. Furthermore, this their demographic attributes.124 These results clearly
must be done in the face of uncertainty around challenge current data protection practices. However,
long-term return on investment, which is where new methods such as synthetic datasets are targeting
experienced advisors and experts will play crucial the problem of anonymisation. True integration
roles. We need to be wary of solutions that promise remains a challenging proposition, and the successful
too much, and case studies with little evidence that and secure integration of digital twins will require
investments have yielded secure returns. Let us more research and development before the industry
briefly consider the challenges posed by big data and reaches a state of maturity and understanding.
by information security.
OPPORTUNITIES
We now live in a world where data affects every facet
of our lives, which raises the urgent question of what There are a number of technologies and trends we
kind of data, and in what amounts, we must collect must consider when thinking about the evolution
to achieve our desired outcomes. While data-driven and future of digital twins across our markets of
decisions are only as good as the data collected, interest. These include connected devices, cloud
the answer to this question is not always clear in and edge computing, 5G, and autonomous vehicles
advance. For instance, how do we collect emotional that may influence how we deliver new products and
data of employees and use that data to adjust the services. Connected devices are greatly benefiting
environmental parameters of the building, to promote the built environment, rapidly changing the way we
more productive and happier workspaces? How much design and offer services. Cloud computing, data
data do we need from each individual, and can we and analytics provide insights into assets and their
trust its fidelity once we have it? The designs of and performance, allowing us to take well-informed
inputs to our digital twins are ever changing, and and timely action. These continuous feedback loops
the twins must adjust to the changing parameters. lead to improved products and processes. A better
With regard to data storage, more organisations of all understanding of the relationships between devices,
sizes are moving to the cloud for strategic, financial people and civic infrastructure will be critical in
and operational flexibility and scalability. Given this advancing the digital twin.
shift, we need to consider how we use, build and
operate our digital twins, and how we handle their Edge computing is another important technology
vital role in the consumption, computation, storage gaining traction as digital devices continue to
and management of information. proliferate. It involves processing data on the
periphery (or ‘edge’) of a network, rather than on
Security and privacy are key for those operating centralised servers.125 The prospect of continuous
digital twins as well as for those interacting with feedback loops appears challenging at a time when
them. We need to consider data governance, open more and more of our data is stored in the cloud,
standards, APIs, privacy, intellectual property and but edge computing can allay these problems
cybersecurity. At present, the information assets of by processing data at or near the source of its
many organisations may not be properly managed generation. This approach is critical, for example,
and secured, i.e. in a state ready for a digital twin where connectivity is poor, or where network latency
to make use of them. Each organisation will need is an issue. ►
to develop appropriate policies and practices. With

15 0 CHAPTER 3
151
Another technology gaining importance is 5G, and feedback from model to real world.
whose integration into business will result in higher The research clarifies a series of challenges for
volumes of data processed at much faster speeds. 5G further research, including issues we have already
represents a step change in data transfer capacity, mentioned such as cybersecurity and privacy, but
and will introduce very high carrier frequencies with also less obvious issues such as marketing hype,
exponentially increased bandwidth.126 Autonomous or the continued challenges of data, which may be
vehicles are one prime candidate to benefit from both inconsistent and sparse datasets. In addition, digital
edge computing and 5G connectivity, which, taken twins require regular maintenance and upkeeping.
together, will facilitate instant reactivity in an ever-
changing physical environment. These technologies While the benefits that digital twins will bring to
will improve the safety and performance of the built environment are clear to most — predictive
autonomous vehicles. Remote assistance is another maintenance, cost control, and asset optimisation —
new service finding favour in the industry; the idea it is equally important to point out the positive role
takes aim at inefficient troubleshooting methods, they can play with regard to sustainability and other
allowing technicians to address the problem at hand societal challenges in the design and engineering
immediately such as indicator notifications on a of civil infrastructure. In order for this positive role
dashboard. We expect greater adoption of these to be realised, we must work and collaborate across
technologies over the coming years to advance the disciplines. The vision we emphasise for our industry
knowledge and development of digital twins, with is the widespread development and adoption of what
many more opportunities for growth and innovation. we term ‘level 5’ digital twin, whose hallmarks are
the abilities to autonomously learn from and reason
CONCLUSIONS about their environments. While we are still far from
a landscape populated by digital twins at levels 4 and
In this report, Arup presents the current state of 5, our case studies, interviews and thought leaders
digital twins and flagship examples of how they are give good reason for optimism in this direction.
changing the design, engineering, and operations of
the built environment. We discuss the definitions of Ultimately, the evolution of digital twins will help
digital twins across the industry and academia while asset owners, managers and society at large to take
including the notions of autonomy and learning. more informed decisions, on the basis of real-time
Furthermore, a working evaluation framework is data. While we currently have widespread autonomy
proposed to help the industry align in its ability of warning systems, this will turn into widespread
to evaluate the sophistication of individual digital predictive systems, and finally into reasoning twins.
twins; these metrics are used to evaluate each of
the case studies. We also present a collection of
interviews with global experts which give insight
into the complex issues around, for example,
security, privacy and data longevity as they relate to
digital twins.

By reviewing the current state of the digital twin

landscape, we are able to elucidate what is still
lacking in the typical twin – most obviously, the
ability to learn and reason. Our interviews make clear
that digital twins are a fundamentally data-driven
proposition, and in the world of modern computation,
this means that they will inexorably develop towards
learning and reasoning. Furthermore, it is clear
that digital twins must be user-centred, inclusive of
society — the ultimate beneficiary of a digital twin
landscape — but also inclusive of those who interface
with digital twins explicitly, the decision makers.
Simulations and static models alone are not digital
twins, as they do not take into account the direct link

152 CHAPTER 3
CALL TO ACTION
future, we envisage an ecosystem of digital twins
Digital twins have already enabled us to automate communicating across boundaries, sharing data
many redundant and tedious tasks using machine and learning from each other to form meaningful
learning, which has resulted in higher productivity. collaborations and solve global challenges. We
Soon, we will be able to alleviate the problem of must build partnerships and foster relationships as
fragmented and isolated data, further benefitting we tackle the technological, cultural and societal
our cities and infrastructure. The digital twin challenges of digital twins. As we continue to
concept is limited by our imagination, rather drive forward the development of the digital twin,
than by our technology. In the not-too-distant please stay in touch with us via [email protected].

153
15 4 CHAPTER 3
REFERENCES

1. Grieves M. Digital Twin: Manufacturing Excellence through Virtual Factory Replication. 34. Krasniqi X, Hajrizi E. Use of IoT Technology to Drive the Automotive Industry from Con-
Digit Twin White Pap. Published 2014. Accessed November 17, 2019. https://research. nected to Full Vehicles. Int Fed Autom Control. 2016;49(29):269-274.
fit.edu/media/site-specific/researchfitedu/camid/documents/1411.0_Digital_Twin_White_ 35. Etong N. Digital Twins in Commercial Aviation. Published 2019. Accessed October
Paper_Dr_Grieves.pdf 31, 2019. https://www.ifsworld.com/se/-/media/assets/2019/05/23/whitepaper-digi-
2. Dohrmann K, Gesing B, Ward J. Digital Twins in Logistics. Accessed November 17, tal-twins-in-aviation.pdf.
2019. https://www.logistics.dhl/content/dam/dhl/global/core/documents/pdf/glo-core- 36. Lau W. WeWork Takes on Design Research and the Internet of Things. Architect Maga-
digital-twins-in-logistics.pdf. zine. Published 2016. Accessed October 31, 2019. https://www.architectmagazine.com/
3. Siemens. Digital Twin. Accessed November 17, 2019. https://www.plm.automation. technology/wework-takes-on-design-research-and-the-internet-of-things_o.
siemens.com/global/en/our-story/glossary/digital-twin/24465. 37. Fenrich K. Securing your control system: the “CIA triad” is a widely used benchmark for
4. SAP. Digital Twin. November 17, 2019. https://www.sap.com/uk/products/digital-sup- evaluating information system security effectiveness. Power Eng. 2008;112(2).
ply-chain/digital-twin.html. 38. Langner R. Stuxnet: Dissecting a Cyberwarfare Weapon. IEEE Secur Priv. 2011;9(3):49-51.
5. Microsoft. Azure Digital Twins. Accessed November 17, 2019. https://azure.microsoft. 39. Department of Health & Social Care. Securing cyber resilience in health and care:
com/en-gb/services/digital-twins/. Progress update October 2018. Published 2018. Accessed October 31, 2019. www.
6. IBM. Watson Internet of Things. Accessed November 17, 2019. https://www.ibm.com/ nationalarchives.gov.uk/doc/open-government-licence/.
internet-of-things. 40. British Standards Institution. Introduction to PAS 1192-5 : 2015 A specification for secu-
7. Mohammadi N, Taylor JE. Smart city digital twins. 2017 IEEE Symp Ser Comput Intell rity-minded building information modelling , digital built environments and smart asset
SSCI 2017 - Proc. 2018;2018-Janua:1-5. management. 2015. Accessed October 31, 2019. https://www.cpni.gov.uk/system/files/
8. Enzer M. Developing the underlying framework for a national digital twin. National Infra- documents/18/6f/BIM-Introduction-To-PAS1192-5.pdf.
structure Commission. Published 2018. Accessed November 17, 2019. https://www.nic. 41. Tassel D Van. Cryptographic techniques for computers: Substitution methods. Inf
org.uk/developing-the-underlying-framework-for-a-national-digital-twin/. Storage Retr. 1970;6(2):241-249.
9. Google. Nest Learning Thermostat. Accessed November 17, 2019. https://store.google. 42. Nakamoto S. Bitcoin: A Peer-to-Peer Electronic Cash System. 2008:1-9.
com/gb/product/nest_learning_thermostat_3rd_gen. 43. Arup. Blockchain Technology. Published 2017. Accessed October 31, 2019. https://
10. Evans R, Gao J. DeepMind AI Reduces Google Data Centre Cooling Bill by 40%. www.arup.com/perspectives/publications/research/section/blockchain-technology.
Published 2017. Accessed November 17, 2019. https://deepmind.com/blog/article/ 44. Arup. Blockchain and the Built Environment. Published 2019. Accessed October
deepmind-ai-reduces-google-data-centre-cooling-bill-40. 31, 2019. https://www.arup.com/perspectives/publications/research/section/block-
11. UK Autodrive. Final Project Report. Published 2019. Accessed November 17, 2019. chain-and-the-built-environment.
https://www.arup.com/-/media/arup/files/pdf-downloads/uk-autodrive/ukautodrive_fi- 45. United Nations. World’s population increasingly urban with more than half living in urban
nalreport_may2019.pdf. areas. Published 2014. Accessed October 31, 2019. https://www.un.org/en/develop-
12. Google. Google Maps. Accessed November 17, 2019. https://www.google.com/maps/ ment/desa/news/population/world-urbanization-prospects-2014.html.
preview. 46. Erickson P, Tempest K. Keeping Cities Green: Avoiding Carbon Lock-in Due to Urban
13. Sidewalk Labs. Draft Master Innovation and Development Plan, “Toronto Tomorrow: A Development; Published 2015. Accessed October 31, 2019. https://www.sei.org/publi-
New Approach for Inclusive Growth.” Published 2019. Accessed November 17, 2019. cations/keeping-cities-green-avoiding-carbon-lock-in-due-to-urban-development/.
https://www.sidewalktoronto.ca/. 47. J.Foxon T. A coevolutionary framework for analysing a transition to a sustainable low
14. Jeremy H, Mike B Del. Scaling Machine Learning at Uber with Michelangelo. Uber carbon economy. Ecol Econ. 2011;70(12):2258-2267.
Engineering. Published 2018. Accessed November 17, 2019. https://eng.uber.com/ 48. Hall RE, Bowerman B, Braverman J, Taylor J, Todosow H, Von Wimmersperg U. The
scaling-michelangelo/. Vision of a Smart City. In: 2nd International Life Extension Technology Workshop. Paris,
15. Waliany R, Kang L, Murati E, Amin MS, Volk N. How Trip Inferences and Machine France; 2000.
Learning Optimize Delivery Times on Uber Eats. Published 2019. Accessed November 49. Government UK. A Time of Unprecedented Change in the Transport System. Published
17, 2019. https://eng.uber.com/uber-eats-trip-optimization/. 2019. Accessed October 31, 2019. https://assets.publishing.service.gov.uk/government/
16. Bell F, Smyl S. Forecasting at Uber: An Introduction. Uber Engineering. Published 2019. uploads/system/uploads/attachment_data/file/780868/future_of_mobility_final.pdf.
Accessed November 17, 2019. https://eng.uber.com/forecasting-introduction/ 50. Ove Arup Foundation. Digital Cities for Change. Published 2018. Accessed October 31,
17. Marketsandmarkets. Digital Twin Market, Global Forecast to 2025. Published 2019. 2019. https://www.ovearupfoundation.org/featured-grants/digital-cities-for-change.
Accessed November 17, 2019. https://www.marketsandmarkets.com/Market-Reports/ 51. SmartCitiesWorld. Cambridge digital twin to tackle congestion and air quality. Published
digital-twin-market-225269522.html?gclid=CjwKCAiA_MPuBRB5EiwAHTTvMa3E- 2019. Accessed October 31, 2019. https://www.smartcitiesworld.net/news/news/cam-
jzSmpCh_C7VPtIa7oaLFZ2RfKmZ8UYACbsSiX2SO5NV57pgg4hoCy_wQAvD_BwE. bridge-digital-twin-to-tackle-congestion-and-air-quality-3917.
18. Batty M. Digital twins. Environ Plan B Urban Anal City Sci. 2018;45(5):817-820. 52. Carlo Ratti Associati. Accessed October 31, 2019. https://carloratti.com/.
19. Centre for Digital Built Britain. National Digital Twin Programme. Published 2019. 53. Evolution News. Kissinger: A.I. Machines Won’t “Think.” Published 2018. Accessed Oc-
Accessed November 17, 2019. https://www.cdbb.cam.ac.uk/national-digital-twin-pro- tober 31, 2019. https://evolutionnews.org/2018/05/kissinger-a-i-machines-wont-think/.
gramme. 54. MIT. MIT Senseable City Lab. 2019. Accessed October 31, 2019. http://senseable.mit.
20. Grieves M, Vickers J. Digital Twin: Mitigating Unpredictable, Undesirable Emergent edu/.
Behavior in Complex Systems BT - Transdisciplinary Perspectives on Complex Systems: 55. MIT News. MIT Real Time Rome project to debut at Venice Biennale. Published 2006.
New Findings and Approaches. 2017:85-113. Accessed October 31, 2019. http://news.mit.edu/2006/realtime-rome.
21. Parrott A, Warshaw L. Industry 4.0 and the digital twin. Deloitte Univ Press. Published 56. Ratti C. Agnelli Foundation HQ. Accessed October 31, 2019. https://carloratti.com/
2017. Accessed November 17, 2019. https://dupress.deloitte.com/dup-us-en/focus/ project/fondazione-agnelli/.
industry-4-0/digital-twin-technology-smart-factory.html. 57. MIT News. The privacy risks of compiling mobility data. Published 2018. Accessed Octo-
22. Gartner. Gartner Glossary. Accessed November 17, 2019. https://www.gartner.com/en/ ber 31, 2019. http://news.mit.edu/2018/privacy-risks-mobility-data-1207.
information-technology/glossary/digital-twin. 58. ArcGIS. Landsat 8 Imagery: Normalized Difference Moisture Index Colorized. Published
23. GE Digital. Minds + Machines 2016, Meet the Digital Twin. Published 2016. Accessed 2016. Accessed October 31, 2019. https://www.arcgis.com/home/item.html?id=3750c-
November 17, 2019. https://www.youtube.com/watch?v=2dCz3oL2rTw 9c5799043978b32b45f789d75ad.
24. IBM. Watson Internet of Things. Accessed November 17, 2019. https://www.ibm.com/ 59. Philips. How a virtual heart could save your real one. Published 2018. Accessed October
internet-of-things/trending/digital-twin. 31, 2019. https://www.philips.com/a-w/about/news/archive/blogs/innovation-mat-
25. Grieves M, Vickers J. Digital Twin: Mitigating Unpredictable, Undesirable Emergent ters/20181112-how-a-virtual-heart-could-save-your-real-one.html.
Behavior in Complex Systems BT - Transdisciplinary Perspectives on Complex Systems: 60. Philips. Philips Digital Twin concept. YouTube. Published 2018. Accessed October 31,
New Findings and Approaches. 2017:85-113. 2019. https://www.youtube.com/watch?v=H6JzPCbyVSM.
26. Microsoft. The promise of a digital twin strategy. Accessed November 17, 2019. https:// 61. Dassault Systèmes 3DEXPERIENCity® - Virtual Singapore: Singapore’s Innovative City
info.microsoft.com/rs/157-GQE-382/images/Microsoft%27s Digital Twin %27How- Project. Accessed October 31, 2019. https://www.ge.com/digital/sites/default/files/
To%27 Whitepaper.pdf. download_assets/GE-Digital-Power-Plant-Brochure.pdf.
27. NASA. Apollo 13. Published 2009. Accessed November 17, 2019. https://www.nasa. 62. New South Wales Government. Emerging Technology Guide: Digital Twin. Accessed
gov/mission_pages/apollo/missions/apollo13.html October 31, 2019. https://www.digital.nsw.gov.au/digital-transformation/policy-lab/
28. Mankins J. Technology Readiness Levels–A White Paper (1995). Off Sp Access Technol- emerging-technology-guide-digital-twin.
ogy NASA, 2008:4-8. 63. New South Wales Government. Twinning! Spatial Services has created a Digital Twin
29. Schaller RR. Moore’s law: past, present and future. IEEE. 1997;34(6):52-59. of NSW. Published 2019. Accessed October 31, 2019. https://www.digital.nsw.gov.au/
30. Prensky M. Digital Natives, Digital Immigrants Part 1. Horiz. 2001;9(5):1-6. article/twinning-spatial-services-has-created-digital-twin-nsw.
31. Gartner. Hype Cycle for Emerging Technologies, 2018. Published 2018. Accessed 64. Arup. Transitioning to lower-carbon energy sources is underway. Accessed October 31,
November 17, 2019. https://www.gartner.com/en/newsroom/press-releases/2018-08- 2019. https://www.arup.com/perspectives/low-carbon-energy-the-future-is-now.
20-gartner-identifies-five-emerging-technology-trends-that-will-blur-the-lines-between- 65. Arup. Smart Cities District Information Modelling and Management for Energy Reduction.
human-and-machine. Accessed October 31, 2019. https://www.arup.com/projects/smart-cities-district-infor-
32. Gartner. 5 Trends Appear on the Gartner Hype Cycle for Emerging Technologies, 2019. mation-modelling-and-management-for-energy-reduction.
Published 2019. Accessed November 17, 2019. https://www.gartner.com/smarterwith- 66. EnerMech. Arup and EnerMech Join Forces to Revolutionise Oil and Gas Asset
gartner/5-trends-appear-on-the-gartner-hype-cycle-for-emerging-technologies-2019/. Inspection. Published 2019. Accessed October 31, 2019. https://www.enermech.com/
33. Centre for Digital Built Britain. The Gemini Principles. Published 2018. Accessed news/item/1003-arup-and-enermech-join-forces-to-revolutionise-oil-and-gas-asset-
November 17, 2019. https://www.cdbb.cam.ac.uk/system/files/documents/TheGemini-
Principles.pdf.

155
REFERENCES

inspection. https://www.arup.com/perspectives/melbourne-metro-2-a-major-city-shaping-project.
67. New South Wales Government. Pumped Hydro Roadmap. Accessed October 31, 2019. 98. Arup. Istanbul Metro Cekmekoy and Tasdelen Line. Published 2019. Accessed October
https://energy.nsw.gov.au/renewables/clean-energy-initiatives/hydro-energy-and-stor- 31, 2019. https://www.arup.com/projects/istanbul-metro-cekmekoy-and-tasdelen-line.
age. 99. Ritchie H, Roser M. Urbanization. Published 2019. Accessed October 31, 2019. https://
68. Arup. Hitting the accelerator: building an electric vehicle charging grid for all. Accessed ourworldindata.org/urbanization.
October 31, 2019. https://www.arup.com/perspectives/hitting-the-accelerator-building- 100. IATA. IATA Forecast Predicts 8.2 billion Air Travelers in 2037. Published 2018. Accessed
an-electric-vehicle-charging-grid-for-all. October 31, 2019. https://www.iata.org/pressroom/pr/Pages/2018-10-24-02.aspx.
69. Arup. Arup and EnerMech’s digitised inspection service aims to revolutionise the oil and 101. Arup. Mobility Mosaic. Published 2019. Accessed October 31, 2019. https://arup.io/
gas industry. Published 2019. Accessed October 31, 2019. https://www.arup.com/news- mosaic/.
and-events/arup-and-enermechs-digitised-inspection-service. 102. Jabbari Jahromi A, Man Oram M-Y, Shrisankaraan VS, Trevan J. Citizens as Real-Time
70. General Electric. The Digital Power Plant. Accessed October 31, 2019. https://www. Emotional Sensors in Smart Cities. In: International Conference on Smart Infrastructure
ge.com/digital/sites/default/files/download_assets/GE-Digital-Power-Plant-Brochure.pdf. and Construction 2019 (ICSIC): Driving Data-Informed Decision-Making; 2019:571-576.
71. UK Research & Innovation. WindTwin - Digital Twin of Wind Turbines for real time contin- 103. Campanella G, Ribeiro RA. A framework for dynamic multiple-criteria decision making.
uous monitoring and inspection. Published 2017. Accessed October 31, 2019. https:// Decis Support Syst. 2011;52(1).
gtr.ukri.org/projects?ref=103387. 104. DNV GL. Making your asset smarter with the digital twin. Published 2019. Accessed
72. Grafana Labs. Using InfluxDB in Grafana. Published 2019. Accessed October 31, 2019. October 31, 2019. https://www.dnvgl.com/article/making-your-asset-smarter-with-the-
https://grafana.com/docs/features/datasources/influxdb/. digital-twin-63328.
73. BP. Greater Plutonio. Published 2019. Accessed October 31, 2019. https://www.bp.com/ 105. Willow Inc. Transport for New South Wales. Published 2019. Accessed October 31,
en_ao/angola/home/where-we-operate/operated-by-us/block-18--greater-plutonio-. 2019. https://www.willowinc.com/customer-stories/tfnsw/.
html. 106. Konecranes. Digital twins - a new standard in industrial production. Published
74. Bole M, Powell G, Rousseau E. Taking control of the digital twin. Published 2017. 2019. Accessed October 31, 2019. https://www.konecranes.com/resources/digi-
Accessed October 31, 2019. http://www.polycad.co.uk/downloads/Taking_Control_of_ tal-twins-new-standard-in-industrial-production.
the_Digital_Twin.pdf. 107. Willow Inc. Willow and Strukton Build World’s First Digital Twin Platform for Rail Improv-
75. Arup. A model for a new type of office building. Published 2019. Accessed October 31, ing Uptime in The United States. Published 2018. Accessed October 31, 2019. https://
2019. https://www.arup.com/projects/white-collar-factory. www.prnewswire.com/news-releases/willow-and-strukton-build-worlds-first-digital-twin-
76. Willow Inc. WillowDigital. Published 2019. Accessed October 31, 2019. https://www. platform-for-rail-improving-uptime-in-the-united-states-300714430.html.
willowinc.com/willowdigital/. 108. McKie R. Apollo 13: celebrating the unsung heroes of mission control. Published 2017.
77. Harari YN. Homo Deus: A History of Tomorrow. Harvill Secker. 2016. Accessed November 1, 2019. https://www.theguardian.com/science/2017/apr/16/apol-
78. Willow Inc. WillowRail. Published 2019. Accessed October 31, 2019. https://www. lo-13-mission-control-unsung-heroes-jim-lovell-interview.
willowrail.com/. 109. Rolls Royce. Rolls-Royce demonstrates world’s first remotely operated commercial
79. Deloitte LLP. Making an impact that matters. Published 2015. Accessed October 31, vessel. Published 2017. Accessed October 31, 2019. https://www.rolls-royce.com/
2019. https://www2.deloitte.com/content/dam/Deloitte/global/Documents/About-De- media/press-releases/2017/20-06-2017-rr-demonstrates-worlds-first-remotely-operat-
loitte/gx-gr15-the-edge-of-tomorrow.pdf. ed-commercial-vessel.aspx.
80. EEA. Renewable Energy Sources. Published 2018. Accessed November 17, 2019. 110. McDonald C. How Ocado has disrupted its own model. Published 2018. Accessed
https://www.eea.europa.eu/airs/2018/resource-efficiency-and-low-carbon-economy/ November 1, 2019. https://www.computerweekly.com/news/252448108/How-Ocado-
renewable-energy-sources. has-disrupted-its-own-model.
81. Hill S. Re-Imagining Facility Management for the Digital Age. Published 2019. Accessed 111. Alan Turing Institute. A digital twin of the world’s first 3D printed steel bridge. Published
November 17, 2019. https://www.arup.com/-/media/arup/files/publications/d/reimagin- 2018. Accessed November 1, 2019. https://www.turing.ac.uk/research/research-proj-
ing-facility-management-for-the-digital-age.pdf. ects/digital-twin-worlds-first-3d-printed-steel-bridge.
82. The Hague Municipal Government. Climate Plan The Hague. Published 2017. Accessed 112. Block I. Robots complete span of 3D-printed bridge for Amsterdam canal. Published
October 31, 2019. https://www.gouvernement.fr/en/climate-plan. 2018. Accessed November 1, 2019. https://www.dezeen.com/2018/04/17/mx3d-3d-
83. Amanatidis A. Digital Assistants and the Future Role of Brands: Digital Twins (2/3). printed-bridge-joris-laarman-arup-amsterdam-netherlands/.
Published 2019. Accessed October 31, 2019. https://medium.com/voice-tech-podcast/ 113. Airbus. Airbus opens Skywise to global IT service leaders. Published 2019. Accessed
digital-assistants-and-the-future-role-of-brands-digital-assistants-2-3-6b7128fe50b3. November 1, 2019. https://www.airbus.com/newsroom/press-releases/en/2019/06/
84. Willow Inc. Willow and thyssenkrupp Elevator Take Building Smarts to a New Level. airbus-opens-skywise-to-global-it-services-leaders.html.
Published 2018. Accessed October 31, 2019. https://www.willowinc.com/thyssenk- 114. Arup. Water Research Review. Published 2015. Accessed November 17, 2019.
rupp2018/. https://www.arup.com/perspectives/publications/promotional-materials/section/de-
85. Lan L. Wargocki P. Wyon DP, et al. Effects of thermal discomfort in an office on perceived sign-with-water.
air quality, SBS symptoms, physiological responses, and human performance. Indoor 115. United Nations. The Sustainable Development Goals Report 2019. Published 2019.
Air. 2011 Oct;21(5):376-90. Accessed November 17, 2019. https://unstats.un.org/sdgs/report/2019/The-Sustain-
86. Loftness V. Hartkopf V., Gurtekin B, et al. Linking Energy to Health and Productivity in able-Development-Goals-Report-2019.pdf.
the Built Environment: Evaluating the Cost-Benefits of High Performance Building and 116. Government UK. Regulation for the Fourth Industrial Revolution.; Accessed November
Community Design for Sustainability, Health and Productivity. USGBC Green Build 17, 2019. https://www.gov.uk/government/publications/regulation-for-the-fourth-indus-
Conference. 2003. trial-revolution.
87. Terrapin. The Economics of Biophilia: Why designing with nature in mind makes financial 117. Wagener T, Kollat J. Numerical and visual evaluation of hydrological and environmental
sense. Published 2015. Accessed November 18, 2019. https://www.lbhf.gov.uk/sites/ models using the Monte Carlo analysis toolbox. Environmental Modelling & Software.
default/files/section_attachments/the_economics_of_biophilia_-_why_designing_with_ 2007;22(7):1021-1033.
nature_in_mind_makes_financial_sense.pdf. 118. Newcastle University. Newcastle’s “digital twin” to help city plan for disasters.
88. Clark D. What Colour is your Building? Measuring and reducing the energy and carbon Published 2019. Accessed November 4, 2019. https://www.ncl.ac.uk/press/articles/
footprint of buildings. Published 2013. archive/2019/01/digitaltwin/.
89. McKinsey&Company. McKinsey Global Institute ‘The Internet of Things: Mapping the 119. The Smart Water Networks Forum. SWAN. Published 2019. Accessed November 4,
Value Beyond the Hype’. Published 2015. Accessed October 31, 2019. https://www. 2019. https://www.swan-forum.com/.
mckinsey.com/~/media/McKinsey/Industries/Technology%20Media%20and%20Tele- 120. Fortune D. The Digital Twin in Water Infrastructure. Innovyze. Published 2019. Accessed
communications/High%20Tech/Our%20Insights/The%20Internet%20of%20Things%20 November 4, 2019. https://www.innovyze.com/en-us/about-us/blog/digital-twins-in-wa-
The%20value%20of%20digitizing%20the%20physical%20world/Unlocking_the_poten- ter-infrastructure.
tial_of_the_Internet_of_Things_Executive_summary.ashx. 121. Bentley. Digital Water Works, Inc. Receives Strategic Investment from Bentley Systems.
90. Arup. Living in a digital built environment. Published 2019. Accessed October 31, 2019. Published 2019. Accessed November 4, 2019. https://www.bentley.com/en/about-us/
https://www.arup.com/perspectives/digital-built-environment. news/2019/march/01/dww-receives-strategic-investment-from-bentley.
91. Archello. National Aquatics Center (Water Cube). Published 2009. Accessed October 31, 122. European Commission. Data protection in the EU. Published 2019. Accessed November
2019. https://archello.com/project/watercube-beijing#story-2. 18, 2019. https://ec.europa.eu/info/law/law-topic/data-protection/data-protection-eu_en.
92. Transport for London. Our open data. Published 2019. Accessed October 31, 2019. 123. Hern A. “Anonymised” data can never be totally anonymous, says study. The Guardian.
https://tfl.gov.uk/info-for/open-data-users/our-open-data?intcmp=3671. Published 2019. Accessed November 4, 2019. https://www.theguardian.com/technolo-
93. University of Cambridge. Infrastructure Digital Twins. Published 2019. Accessed October gy/2019/jul/23/anonymised-data-never-be-anonymous-enough-study-finds.
31, 2019. https://www.ifm.eng.cam.ac.uk/research/asset-management/research-proj- 124. Rocher L, Hendrickx JM, de Montjoye Y. Estimating the success of re-identifications in
ects/infrastructure-digital-twins/. incomplete datasets using generative models. Nat Commun 10, 3069 (2019).
94. Oasys. Crowd Simulation Software: MassMotion. Published 2019. Accessed October 125. Shi W, Cao J, Zhang Q, et al. Edge Computing: Vision and Challenges. IEEE Internet
31, 2019. https://www.oasys-software.com/products/pedestrian-simulation/massmotion Things J. 2016;3(5):637-646.
95. Fruin JJ. Designing for Pedestrians: A Level-of-Service Concept. Published 1971. http:// 126. Andrews JG, Buzzi S, Choi W, et al. What will 5G be? IEEE J Sel Areas Commun.
onlinepubs.trb.org/Onlinepubs/hrr/1971/355/355-001.pdf. 2014;32(6):1065-1082.
96. Arup. Arup team wins major (HS2) High Speed 2 Phase 2 contract. Published 2017.
Accessed October 31, 2019. https://www.arup.com/news-and-events/arup-team-wins-
major-hs2-high-speed-2-phase-2-contract.
97. Arup. Melbourne Metro 2: A major city shaping project. Accessed October 31, 2019.

15 6
IMAGE CREDITS

Cover: Adapted from Florian Wehde from Unsplash

P4 BG Fonagn
P7 (top) Paul Carstairs / Arup (bottom) Chuttersnap on Unsplash
P8-9 Fabio Roque on Unsplash
P11 Jayden So on Unsplash
P12 Chuttersnap on Unsplash
P20 Abigail Keenan on Unsplash
P24-25 Zhanhui Li on Unsplash
P28-29 Nick Fewings on Unsplash
P31 NASA on Unsplash
P32-33 Chuttersnap on Unsplash
P36-37 Jue Huang on Unsplash
P40-41 Lysander Yuen on Unsplash
P44-45 Fotolia
P46 Pang Yu
P51 JC Gellidon
P53 BG Fonagn
P54 MIT / AWS
P55 Owen Cl on Unsplash
P59 Arup
P61Airam Dato on Unsplash
P62-63 (upper left) Jorge Percival on Unsplash (upper right) Mike Enerio on Unsplash (middle
right) NIH Image Gallery, Flickr, Public Domain (bottom) Clément Falize on Unsplash
P64 American Public Power Association on Unsplash
P67 David Cristian on Unsplash
P69 Mario Babiera / Photosynthesis Inc
P71 (top) Arup (bottom) Arup
P72 (inset) General Electric
P72-73 Andrea Boldizsar on Unsplash
P75 Arup
P77 Arup
P79 Arup
P80 (top and bottom) Arup
P82-83 NOAA on Unsplash
P84 Christian Wiediger on Unsplash
P87 Vladimir Kudinov on Unsplash
P89 (inset) Willow¬
P90 (inset) Willow
P91 Lycs Architecture on Unsplash
P95 Arup
P96 (inset) Google
P96-96 Willow
P101 (left, right) Arup
P105 Rafael Vianna Croffi, Flickr, CC BY 2.0
P107 (top and bottom) Arup
P109 Kate Sade on Unsplash
P110 Dapo Oni on Unsplash
P114-115 Bentley Systems
P117 Bentley Systems
P119 (top and bottom) Arup
P121 (top and bottom) Arup
P123 Arup
P125 (top and bottom) Arup
P126-127 (top left) DNVGL (middle left) Willow (top right) Kone-
cranes (bottom) NSW Government CC BY 4.0
P128 (inset) NASA, attributed to History in HD on Unsplash
P128-129 Chuttersnap on Unsplash
P130-131 (top left) Ruchindra Gunasekara on Unsplash (top right) Adueck on Pixabay (bottom)
MX3D BV
P132 David Martin on Unsplash
P135 Keith Chong on Unsplash
P136-137 Petar Petkovski on Unsplash
P138-139 Family New York, Courtesy of Friends of + POOL
P141 (top) © NSW Department of Customer Service, Spatial Services
P142-143 Jamie Davies on Unsplash
P146-147 (upper left) Samuel Scrimshaw on Unsplash (upper right) Tommy Pequinot on
unsplash (middle right) American Public Power Association on Unsplash (bottom) Innovyze
P149-150 Rowan Huevel on Unsplash
P151 Laura Cros on Unsplash
P154 Monica Silva on Unsplash
P153 Jurriaan Snikkers on Unsplash

157
CONTRIBUTORS

Arup Guest

Allison An Prof Mark Girolami Anand Mecheri

Simon Blakeney Alan Turing Institute Andy Hamer
Richard Bowden University of Cambridge Invicara
John Burgess
Volker Buscher Dr Agnieszka Bert Van Hoof
Tom Butler Krzyzaniak Microsoft
Qian Cao Arcadis
Tristram Carfrae Prof. Carlo Ratti
Dr Gerard Casey Nicolas Mangon MIT Senseable City Lab
Dr Will Cavendish Autodesk
Mark Chen Ovi Cristescu
Isabel Dedring David Thompson Ocado
Léan Doody Aveva
Anni Feng Dr Ger Janssen
Dr Sarah Gallacher Dr Nabil Abou-Rahme Philips Healthcare
Matthias Geipel Bentley Systems
Senem Gulal Bruce Jackson
Ercan Gunay Michael Jansen Siemens
Ahmet Gunduz Andrew Penney
Josef Hargrave Cityzenith Mike Wassell
Stuart Hawkins Sydney Water
Stephen Hill (D) Patrick Mays
Bryan Horton Dassault Systèmes Ilka May
Ali Jabbari The Digital Twin
Kate Jarritt Dr Caroline Karmann Company
Maggie Ji Ecole Polytechnique
Fatema Karim-Khaku Fédérale de Lausanne  Dr Timea Nochta
Padraig Kenny Dr Li Wan
Chris Kinnaird Dr Michael Grieves University of Cambridge
Shaw Kudo Florida Institute of
Dr Chris Luebkeman Technology Joshua Ridley
Chris Lyth Gareth Stewart
Lachlan Miles Terence Lui Willow
Atsushi Miyazaki Gammon Construction
Marcus Morrell
Beatrice Nassi Prof John E Taylor
Alan Newbold Georgia Institute of
Harutaka Oe Technology
Kevin O’Grady
Elaine Pang Dr Claire Penny
Hilary Pherribo Wia (formerly IBM)
Alvise Simondetti
Paul Simpson Robert Saracco
Peter Slump IEEE Future Directions
Michael Stych
Kentaro Suga Jennifer White
Christa de Vaan Long Chen
Mathew Vola Imperial College London
Darren Wright

15 8
ACKNOWLEDGEMENTS

Global Research Team

Dr David Gerber
Global Research Manager

Dr Bella Nguyen
Senior Research Consultant

Dr Isabella Gaetani
Senior Research Scientist

Marketing

Annabel Rabbets

Design

Mark Pearsall
Emily Clements
Billy Searle

About Arup

Arup is the creative force at the heart of many of the world’s most prominent
projects in the built environment and across industry. We offer a broad range
of professional services that combine to make a real difference to our clients
and the communities in which we work.

We are truly global. From 88 offices in 33 countries our 15,000 planners,

designers, engineers and consultants deliver innovative projects across the world
with creativity and passion.

Founded in 1946 with an enduring set of values, our unique trust ownership
fosters a distinctive culture and an intellectual independence that encourages
collaborative working. This is reflected in everything we do, allowing us to
develop meaningful ideas, help shape agendas and deliver results that frequently
surpass the expectations of our clients.

The people at Arup are driven to find a better way and to deliver better solutions
for our clients.

159
WE SHAPE A BETTER WORLD

Digital twin
Arup has started the journey towards digital transformation.
Creativity with data at scale, across all aspects of our organisation,
is an integral part of our knowledge, practices and culture. As part
of this strategy, we are looking at research and collaboration to
shape our thinking and partnerships for the future. Digital Twin:
Towards a meaningful framework is part of our aim to share ideas
and develop meaningful relationships.

C O N TA C T
e: [email protected]
w: www.arup.com/digitaltwin
#digitaltwin

13 Fitzroy Street
London
W1T 4BQ
www.arup.com