Intelligent Infrastructure User Centred Remote Condition Monitoring 1st Edition Nastaran Dadashi Download PDF
Intelligent Infrastructure User Centred Remote Condition Monitoring 1st Edition Nastaran Dadashi Download PDF
Intelligent Infrastructure User Centred Remote Condition Monitoring 1st Edition Nastaran Dadashi Download PDF
com
https://ebookmeta.com/product/intelligent-
infrastructure-user-centred-remote-condition-
monitoring-1st-edition-nastaran-dadashi/
OR CLICK BUTTON
DOWLOAD EBOOK
https://ebookmeta.com/product/intelligent-building-health-
monitoring-and-assessment-jingzhou-xin/
https://ebookmeta.com/product/prometheus-up-running-
infrastructure-and-application-performance-monitoring-julien-
pivotto/
https://ebookmeta.com/product/prometheus-up-running-
infrastructure-and-application-performance-monitoring-1st-
edition-brian-brazil-2/
https://ebookmeta.com/product/prometheus-up-running-
infrastructure-and-application-performance-monitoring-1st-
edition-brian-brazil/
Urban Remote Sensing Monitoring Synthesis and Modeling
in the Urban Environment 2nd Edition Xiaojun Yang
https://ebookmeta.com/product/urban-remote-sensing-monitoring-
synthesis-and-modeling-in-the-urban-environment-2nd-edition-
xiaojun-yang/
https://ebookmeta.com/product/strategic-communication-and-ai-
public-relations-with-intelligent-user-interfaces-1st-edition-
simon-moore/
https://ebookmeta.com/product/remote-sensing-intelligent-
interpretation-for-geology-from-perspective-of-geological-
exploration-1st-edition-weitao-chen/
https://ebookmeta.com/product/human-centred-intelligent-systems-
proceedings-of-kes-hcis-2021-conference-244-smart-innovation-
systems-and-technologies-244-alfred-zimmermann-editor/
Intelligent Infrastructure
With the development of sensor technology, wireless communications, big data, and
machine learning, there is an increasing interest in technologies and solutions that
assess and predict the state of equipment and assets within various industrial set-
tings. These technologies aim to collect information from multiple sources about
infrastructure asset status. Then, through current and historical data analysis, this
configuration of technologies delivers intelligence on current and future asset sta-
tus to a maintenance operator or manager to inform optimal maintenance decision-
making. These technologies are known under different terms – remote condition
monitoring, e-maintenance, prognostic systems, predictive maintenance, and smart
or Intelligent Infrastructure. Despite the promise of remote condition monitoring and
predictive technologies, there is a growing concern with such technologies because
they can be difficult or impractical to use.
Understanding and mitigating potential human factors issues could ensure that such
vast investments are not wasted. This book considers, in depth, the challenges placed
on users of current and future condition monitoring systems. Its primary focus is to
understand the cognitive processes, including managing alarms, interpreting data,
and collaborating with automation. The book describes a range of human factors
methods that can be used to understand the current and future functioning of people
and technology in an enhanced maintenance and asset monitoring context. The book
also presents a framework for describing these issues systematically and presents the
resulting design considerations to increase the effectiveness of individual operators
and organisations as a whole.
Intelligent Infrastructure
User-Centred Remote
Condition Monitoring
Nastaran Dadashi
David Golightly
Sarah Sharples
Richard Bye
First edition published 2024
by CRC Press
4 Park Square, Milton Park, Abingdon, Oxon, OX14 4RN
The right of Nastaran Dadashi, David Golightly, Sarah Sharples and Richard Bye to be identified as the authors
of the editorial material, and of the authors for their individual chapters, has been asserted in accordance with
sections 77 and 78 of the Copyright, Designs and Patents Act 1988.
All rights reserved. No part of this book may be reprinted or reproduced or utilised in any form or by any elec-
tronic, mechanical, or other means, now known or hereafter invented, including photocopying and recording,
or in any information storage or retrieval system, without permission in writing from the publishers.
For permission to photocopy or use material electronically from this work, access www.copyright.com or con-
tact the Copyright Clearance Center, Inc. (CCC), 222 Rosewood Drive, Danvers, MA 01923, 978-750-8400. For
works that are not available on CCC please contact mpkbookspermissions@tandf.co.uk
Trademark notice: Product or corporate names may be trademarks or registered trademarks, and are used only
for identification and explanation without intent to infringe.
DOI: 10.1201/9781315587288
Typeset in Times
by Deanta Global Publishing Services, Chennai, India
Contents
Foreword...................................................................................................................vii
About the Authors......................................................................................................ix
References............................................................................................................... 131
Index....................................................................................................................... 141
v
Foreword
The discipline of Human Factors has at its heart the incorporation of the under-
standing of the capabilities and limitations of people in designing complex systems.
Complex infrastructures, such as those which enable our railways, roads, manufac-
turing industries, or logistics operations, often require vast, fixed assets, often over a
wide geographical scale. These fixed assets have a limited lifespan, contain embed-
ded carbon, and represent significant financial costs. Therefore, any technologies
or systems which enable the proactive, efficient, and effective maintenance of these
assets will deliver long-term economic benefits.
However, the nature of fixed infrastructures has undergone a step change in the
past decades. Whereas once the infrastructures were dominated by the physical –
bridges, roads, rails; and the mechanical – points or cranes; we now see an ever-
increasing role for the digital. This presents tremendous opportunities, enabling a
move from reactive or planned maintenance to proactive prediction of where inter-
vention is required, and for the feasibility of data integration, enabling a whole sys-
tems view to be presented to and understood by operators. This results in a complex,
multi-agent, socio-technical system. In addition, often, such systems have evolved
through a principle of availability – where a technological breakthrough has made it
possible to obtain and represent additional data from the system, it has been added.
This results in a system that combines novel information with that obtained from
legacy systems, and with gaps. The phrase “data, data, everywhere” comes to mind.
It has been long established that placing the user at the heart of design will result
in more effective outcomes – better decision-making, increased engagement and
situation awareness, and less need for re-work or post-hoc adjustments to designs.
When considering Intelligent Infrastructure, users may vary in their roles, from
onsite engineers, to remote system controllers, and to managers and policymakers.
All may have different needs for the representation of information, at different levels
of complexity and granularity.
This book proposes an approach which is “intelligence-led”. We recognise the
challenge presented by an evolving, complex, emergent system through a set of
explorations of these complex systems, mainly in the rail context. The examples
in this book consider a range of challenges. For example, the monitoring of age-
ing, heritage infrastructure can benefit from advanced sensing approaches, using
both embedded sensing technologies and external high-accuracy observation tools.
In new buildings or infrastructure, we can embed sensing as we are constructing,
aiding the construction process itself. There is also an increasing capacity to monitor
and control mechanical systems, such as points, robotics, or logistics technologies.
When we start to move towards integrated, real-time systems that enable both moni-
toring and control, we can describe these systems as “digital twins”.
We then propose approaches that consider how we can have appropriate frame-
works and design solutions that start with the problem and the challenge and iden-
tify what decisions are being made, what the cognitive requirements are for a task,
vii
viii Foreword
and what the goals of that task will be. The framework builds on a combination of
learnings from the cases presented in this book, as well as an established theoretical
understanding of design for automation and cognitive work. The aim of the frame-
work is to help designers, developers, and human factors specialists in recognising
the complexity of complex, distributed, Intelligent Infrastructure systems, and to
shift thinking from “what data are available and how can we integrate and use these
data” to “what do we need to understand in order to reach our goals more effectively,
how can data help, and how do we select, combine and present those data to help”?
A whole systems approach is key to this thinking. Too often we can be seduced by
technology, or we can see the people as the “weak links” in a complex system. The
thinking that underpins this book recognises the role of partnership – the importance
of an integrated consideration of how people, technologies, and infrastructure com-
bine to achieve a common goal. The presence and proliferation of technology within
what were previously physical systems will continue to increase. The examples in
this book help on a pathway towards a new way of thinking, towards integrating our
consideration of the physical, the digital, and people. We hope that readers enjoy the
examples and approaches presented in this book. As we see the evolution of systems
and data, it would be heartening to think that, alongside this, we can see a revolution
in thinking, and the successful embedding of an intelligence-led approach to system
development and design.
Sarah Sharples
October 2022.
About the Authors
Dr Nastaran Dadashi is Researcher, Educator, and a Chartered Human
Factors Specialist. She has a PhD in Human Factors (Future Railway Intelligent
Infrastructure) from the University of Nottingham, UK (Human Factors Research
Group), where she worked as a research fellow. As a researcher, she applies human
factors principles to a broad range of industries, including nuclear, health care, and
the creative sector. As an educator, Nastaran works with digital experience designers
at George Brown College, Canada (School of Design), to ensure that the next genera-
tion of digital innovations is accessible, ethical, and responsible.
Richard Bye is Head of Ergonomics and Human Factors at Network Rail. Richard
has been involved in a wide range of physical, cognitive and organisational ergo-
nomics projects, working on initiatives as diverse as the assurance of safety-criti-
cal operational processes, the introduction of smartphones and tablets for frontline
track workers, the development of decision support tools for the railway’s internet of
things, and the implementation of new technology to increase the resilience of real-
time rail operations.
ix
1 Introducing Human
Factors for Remote
Condition Monitoring
1.1 CHAPTER OVERVIEW
This chapter aims to set out the motivation and structure of the book, setting the
scene for the work that is to follow. This includes:
DOI: 10.1201/9781315587288-1 1
2 Intelligent Infrastructure
dry), bridges, passenger escalators, lifts, and ticket machines (e.g., Jing et al.,
2021; Vinberg et al., 2018; Balouchi, Bevan and Formston, 2021). In addition,
train-borne technologies can monitor both the train and the track it travels upon.
Condition monitoring is an important technology in making rail operations more
reliable and cost-effective, ultimately helping to deliver better value and better
service to passengers and freight customers (Grubic, 2018; Bernal, Spiryagin, and
Cole, 2018). Many rail assets are also safety critical, and ensuring they are func-
tioning correctly is, therefore, important to ensure the health of the asset and the
safety of the rail service. Predictive systems can also help extend the life of assets
where they might be prohibitively expensive to replace (e.g., major civil engineer-
ing structures such as bridges and tunnels). Finally, maintenance and inspection
of track and equipment are expensive and potentially dangerous, requiring staff
to go out at night, to remote locations, or work around moving trains (Golightly et
al., 2013). Reducing the need for staff to be present, or to maintain assets unnec-
essarily, can help keep them safe.
This approach is just as relevant to other forms of infrastructure. For example,
Baah et al. (2015) present a risk-based approach to sanitary sewer pipe asset manage-
ment, including a pipe condition grade prediction model, assessment of the conse-
quence of failure, and risk of failure. The synthesised information can consequently
inform future sewer pipe maintenance programmes. In this way, waste collection
systems and technologies to facilitate monitoring the condition of wastewater can
support better water management and prevent sewer-related groundwater contamina-
tion incidents.
Despite the promise of remote condition monitoring and predictive technologies,
there is a growing concern with such technologies because they can be difficult or
impractical to use. Operational staff may face huge volumes of data, which is in a
relatively raw form, presenting problems of interpretation and information overload
(Dadashi, Golightly, and Sharples 2017; Dorgo et al., 2018). Alarms may be complex
to interpret and duplicated many times. At the other extreme, the outputs are sim-
plified down to a simple “red-amber-green” alert with little explanation, where the
reasoning of algorithms may be very unclear, or there may be reliability problems.
While the shift from regular to predictive maintenance can save money through
a reduction in the need for visual inspection of assets, it requires operational staff to
have trust in the algorithms that are helping them to make decisions and, potentially,
to fundamentally change working practices that have been laid down for decades
(Ciocoiu et al., 2017; Kefalidou et al., 2018). Furthermore, there are challenges due
to the ongoing need for calibration. Once a system is installed, it is a new job to
maintain the calibration of existing sensors, react to changes on the ground that may
change what is being measured, and integrate new data sources so that the whole
picture remains coherent. This needs ongoing resource and knowledge management
(Golightly et al., 2018). In addition, the introduction of change in many industries
requires leadership, guidance, and, most critically, a sociotechnical system view of
change (Trist, 1981; Ottens et al., 2006 Wilson, 2014), where technology, people, and
organisational structures (including processes) are viewed as interdependent. All of
these challenges can be thought of as human factors challenges.
Introducing Human Factors for Remote Condition Monitoring 3
This book considers, in depth, the challenges placed on users of current and
future condition monitoring systems. Its primary focus is to understand the cog-
nitive processes, including managing alarms, interpreting data, and collaborating
with automation, if we are to design technologies that support, rather than hinder,
operational users. The book describes a range of human factors methods that can
be used to understand the current and future functioning of people and technology
in an enhanced maintenance and asset monitoring context. The book also presents
a framework for describing these issues systematically and presents the resulting
design considerations to increase the effectiveness of individual operators and organ-
isations as a whole.
FIGURE 1.1 Golden Gate Bridge, San Francisco, USA (left) and First Forth Road
Bridge, Scotland
4 Intelligent Infrastructure
FIGURE 1.2 Contrasting from data to intelligence with from intelligence to data
Introducing Human Factors for Remote Condition Monitoring 5
• Better data about asset behaviour and performance – at its core, these
technologies are about sensing the state of an asset. State information
may include material properties, environmental conditions, location, or a
detailed record of the movement of parts. New sensing and data collection
technology mean this data can be captured continuously with greater fidel-
ity than ever before, with multiple, accurate sensors. Not only can this sense
the state of the asset now, but it also supports an emerging picture over
time. The ability to measure multiple assets consistently means that devia-
tions from the norm can be seen either in an individual asset or potentially
regarding a whole suite of assets (e.g., across an entire estate of buildings
using the same air conditioning unit). This, coupled with machine learning,
enables learning about the behaviour of a whole class of assets in an exten-
sive range of different settings, which can inform not only maintenance but
the future design of products and technologies.
6 Intelligent Infrastructure
with different goals. This is particularly the case when technologies involve a high
degree of automation. Poorly designed automation can lead to confusion, a lack of
trust, over-reliance on technology, and degradation of skills (Bainbridge, 1983).
There is a need to ensure that any asset monitoring or predictive technology
is designed and deployed to meet the needs, capabilities, and limitations of the
people who are intended to use and benefit. Many successful implementations
exist, many with good interface design and strong design and deployment ratio-
nale (Golightly et al., 2018), but little systematic analysis across the paradigm.
Furthermore, many implementations fail, and it is acknowledged that a crucial
part of this is the poor match between technology design and the actual pat-
terns of usage that operators need to adopt (Golightly et al., 2018; Kefalidou et
al., 2018). For every success story, there are anecdotes of operators being over-
whelmed and confused, not trusting technology, and ultimately rejecting the new
ways of working.
One of the fundamental principles of a human factors approach is the principle of
user-centred design (ISO 9241-210 [ISO, 2010]) (see Figure 1.4).
User-centred design advocates the representation and, wherever possible, the
active participation of intended users at every stage of the design process. Methods
from human factors are often about capturing what users do, how they perceive
their activity, or about representing this for others to understand and apply (e.g., in
software design). Central to this is the idea of understanding the context of work, or
“Work as Done” (in contrast to “Work as Imagined” by senior management, software
Introducing Human Factors for Remote Condition Monitoring 9
developers, or project managers) (Hollnagel, 2017). In this book, our main aim is to
understand the context of work and “Work as Done” in asset monitoring situations,
using rail as a case in point.
The overall benefits of taking a human factors, user-centred approach are:
software development and procurement costs, this can lead to expensive re-
designs and change requests. By actively involving users from the outset,
technologies and processes can be designed right the first time.
• Return on investment – Ultimately, the benefits outlined above in terms
of better performance, reduced training, faster adoption leading to seeing
the anticipated benefits, and de-risking of the design process are a cost-
effective pursuit. Money and time spent on engaging users can pay for itself
many times over (Pew and Mavor, 2007).
range of human factors issues, and develop insights with regards to data
processing activities within Intelligent Infrastructure.
• To produce guidance for future development and implementation of
Intelligent Infrastructure systems in the railway, which will match and
complement human capabilities and needs (Chapter 9) – this guidance,
informed by the data processing framework, had two purposes: firstly,
it assists in designing and developing future Intelligent Infrastructure.
Secondly, the methodological approach adopted and practised in this book
provides guidance to human factors specialists in exploring cognition in
complex sociotechnical systems. Furthermore, applications of these rec-
ommendations were used to explore Intelligent Infrastructure within other
sectors.
1.7 CHAPTER SUMMARY
Having set out the aims and structure of this book, Chapter 2 presents detail on
remote condition monitoring concepts. It also gives more information on the rail
context that underpinned the case studies used in this book.
2 Remote Condition
Monitoring and
Predictive Maintenance
2.1 CHAPTER OVERVIEW
Chapter 1 introduced the motivation and structure of the book. This chapter aims to
give a more detailed introduction to remote condition monitoring, particularly for
people new to the domain (e.g., human factors specialists who are new to mainte-
nance technologies). This includes:
DOI: 10.1201/9781315587288-2 13
14 Intelligent Infrastructure
TABLE 2.1
Basic Elements of Implementing Remote Condition Monitoring
Element Description
Data collection Collecting raw data from assets through sensors and tracking
technologies
Intelligent analysis Systems that integrate raw data and provide insight into the asset
performance and its health condition
Access and presentation Visualise and present the health condition to respective operators and
decision-makers
Key performance indicators Ensure that health conditions are effectively monitored and facilitate a
proactive maintenance regime
Connections Reliable data distribution and management (e.g., phone network,
hard-wired, 5G)
published by the Royal Academy of Engineering in 2012 (RAE, 2012), four key
principles are mentioned:
• Data: turned into information, into knowledge, and then into value
• Analysis: synthesising and interpretation, based on specific contextual
needs
• Feedback: learning from experience and enhancing its effectiveness
• Adaptability: adaptive, resilient, and, consequently, sustainable infrastructure
In other words, condition monitoring aims to centralise data from existing technolo-
gies and analyse it before presenting it to the operator in an appropriate format that
can directly assist them with the decision-making problem at hand. Often, this is
used in a comparison or combination with historical data and involves predicting
asset state. In the most advanced conceptions, it may offer actual maintenance advice
and plans, though there are few practical examples.
Remote Condition Monitoring and Predictive Maintenance 15
Going beyond simply sensing and interpreting the current state of an asset, predic-
tive and proactive maintenance can be defined as “the deep embedding of sensing,
computing, and communication capabilities into traditional urban and rural physi-
cal infrastructure … for the purpose of increasing efficiency, resiliency, and safety”
(Mynatt et al., 2017). This supports a service that enables evidence-based operations
and decision-making in a way that it is:
FIGURE 2.1 IoT application for cleaning machine leasing– (taken from ISO/IEC TR
22417:2017).
turbine arrays and, consequently, to increase the efficiency of power plants (e.g.,
Coble et al., 2015; Sheng, 2017; Costello et al., 2017; Ramchurn et al., 2012). In
addition, connected sensors can provide the basis to explore the environmental ben-
efits for the environment of plants and processes associated with energy harvesting.
One crucial need for efficient and accurate health monitoring systems within the
energy sector is due to recent licence extensions. In the USA, 74 reactors have been
approved for first-round licence extensions (beyond 60 years), and therefore, real-
time monitoring of the health of a wide range of components within power plants is
essential (Coble et al., 2015).
Also, a wide range of research studies has explored the capability and connectiv-
ity of sensor-enabled networks (e.g., Akhondi et al., 2010). These capabilities have
provided opportunities to develop smart energy management capabilities for both
power suppliers and consumers. For example, McRoberts (2018) introduces a unified
infrastructure that intelligently manages packaged power within the mining industry
that is more resilient to outages and component failures.
A key contributor to creating a low-carbon built environment will be the inte-
gration of smart building systems. These systems improve asset reliability and
performance, reducing energy use, optimising how space is used, and minimis-
ing the environmental impact of the buildings. For example, Creative Energy
Homes is a seven-house development at the University of Nottingham that pro-
vides living test sites to inform the integration of energy-efficient technologies
into housing (see Figure 2.2). The building will turn into a living organism: net-
worked, intelligent, flexible, sensitive, and adaptable. Examples include Mogles
et al. (2017), who present smarter, building-aware, and more personalised, digi-
tal energy feedback.
TABLE 2.2
The Potential of Predictive Maintenance and Remote Condition Monitoring across Multiple Domains
Descriptive Prescriptive Predictive Proactive
Intelligent Real-time traffic congestion Reroute traffic; adjust Anticipate rush hour/large Suggest traffic patterns w/
transportation information dynamic lane configuration event congestion; anticipate intelligent stoplights
(direction) weather-related accidents
Intelligent energy Real-time energy demand Improve asset utilisation and Anticipate the demand response Suggest new market approaches to
management information management across required to ensure grid integrate production and
transmission and distribution reliability distribution capabilities
system
Intelligent public safety Real-time crowd analysis Threat detection; dispatch Anticipate vulnerable settings Suggest new communication and
and security public safety officers and events coordination response approaches
Intelligent disaster Real-time water levels in Timely levee management Anticipate flood inundation Inform National Flood Insurance
response flood-prone areas and evacuations as needed with low-cost digital terrain Programme; inform vulnerable
maps populations
Intelligent city systems Describe mobility patterns Adjust mobility management Anticipate changing needs for Inform future mobility capabilities
(pedestrian, cycling, to improve safety and reduce parking, charging stations, to drive economic development
automobile, trucking, energy usage bike, and ride share and reduce barriers to
electric, and autonomous programmes employment
vehicles)
Remote Condition Monitoring and Predictive Maintenance
Intelligent agriculture Characterise spatial and Advise based on Forecast crop yield; anticipate Customise management practices
temporal variability in soil, environmental stressors and seasonal water needs and seed selection to local
crop, and weather crop traits conditions
Intelligent health Block-level assessment of Inform asthma action plans Anticipate peak seasonal spikes Inform transportation plans to shift
current allergens/air based on local conditions in allergen and air pollutant road use away from “asthma
pollutant levels levels corridors”
FIGURE 2.2 Exterior of the Nottingham House, Green Close, University Park (Nottingham,
UK)
Having presented a general definition and purpose for remote condition monitor-
ing and predictive maintenance, the chapter now covers the specifics of one domain
– rail remote condition monitoring. The following section provides context to the
case study domain described in the rest of the book and research conducted explic-
itly for the Great Britain rail infrastructure provider, Network Rail.
change are all relevant to the railways. In many parts of the world, demand for the
rail system is at its highest and is only set to increase, with rail offering a low-carbon
alternative for private cars, short-haul flights, and road freight. This book focuses on
the rail system in Great Britain, but the challenges and solutions are just as relevant to
rail globally (ERRAC, 2017; IEA, 2019). While we primarily discuss mainline rail,
these issues covered are just as relevant to light and urban transit, which has a vital
role in moving people in and around cities as urbanisation increases (UN, 2018).
Rail infrastructure requires maintenance to ensure continuous availability and
safe performance. Track, overhead line equipment, and structures such as bridges,
tunnels, and embankments require routine inspection. Much of the work, both
inspection and maintenance, still relies on human effort. In 2012, an inspection of
Great Britain’s 16,000 km of track and 2560 stations required 1.3 million hours of
work. While new technologies such as the Plain Line Pattern have improved that
situation, staff still have a substantial physical need to go out on track. For example,
during the 2019 Easter weekend, over 13,000 maintenance staff worked to build and
repair Great Britain’s rail infrastructure.
This work occurs in a potentially hazardous environment, where track workers need
protection from nearby service or engineering trains. It can also occur in remote or
difficult-to-access locations such as tunnels, at night, and in poor weather (Farrington-
Darby et al., 2005; Schock et al., 2010; Golightly et al., 2013b). Inspection and mainte-
nance require consistent, time-intensive, high-quality work from a skilled but ageing
workforce and complex decision-making with an ever-increasing volume of asset data.
Failure to safely maintain the network can lead to significant delay and disrup-
tion (Szymula and Bešinović, 2020). Signal, track, and point failures are the most
critical and difficult events to manage (Golightly and Dadashi, 2017). The initial
steps associated with identifying and managing a fault are crucial (Belmonte et al.,
2011). Failure to rapidly identify and diagnose a problem can lead to “out-of-control”,
unmanageable situations, and significant rail disruption. When failures occur during
peak times at busy points in the network, the effects can be huge, with many tens of
thousands of delayed passengers and significant damage to passenger confidence and
the reputation of the railways.
More crucially, safety is the paramount concern for improving railway mainte-
nance, and maintenance failures can impair the safety of the railways (Kim and
Yoon, 2013). In practice, the railways in Great Britain (GB) and across the European
Union (EU) have an exceptional safety record (ERRAC, 2017), but outcomes can be
disastrous when failures occur.
As a case in point, a Virgin West Coast Pendolino service from London to
Glasgow derailed on February 23, 2007, with one passenger fatality (RAIB, 2008).
Deterioration of the point machines at Lambrigg was considered as being the imme-
diate cause of the failure. This was due to (1) the mechanical failure of a bolted
joint, (2) the incorrect setup of points, and (3) a track inspection that was missed on
February 18, 2007 (RAIB, 2008). Although images taken previously with the New
Measurement Train showed evidence of defects, the output did not lead to planned
maintenance. In other words, while technology theoretically generates the required
data, this only works if the required decision-making is in place. Accidents such as
20 Intelligent Infrastructure
this, the accident at Potter’s Bar in 2002, or Brétigny-sur-Orge in 2013, have funda-
mentally shaped how maintenance and inspection are conducted and organised. The
need to strive for better performance is a constant aim of all rail stakeholders.
reactive maintenance. This aims to integrate available data to allow for proac-
tive and predictive maintenance. An example of a point is shown in Figure 2.3.
In order to collect reliable and timely information regarding the wide range of
assets that are dispersed across the network nationwide, railway infrastructure has
increasingly relied on the use of remote condition monitoring. This is a capability
that enables the data collection from assets to identify functional failures. These
systems collate the acquired data from sensors and loggers and present the appropri-
ate information regarding the health status of the infrastructure to a control room
operator. Remote condition monitoring provides a platform to collect, synthesise,
and interpret data regarding individual assets and, in doing so, supports a human
operator in programming the maintenance process in order to mitigate the risk of
potential failures (Carretero et al., 2003).
Since these systems are costly to implement, it is important to carefully select and
prioritise which assets are monitored with remote condition monitoring equipment.
These are selected based on their safety criticality and the effect of their failure on
the efficiency of the service (i.e., delays and track availability). Remote condition
monitoring systems have been used on the rail network for many years. As early
as 2006, Network Rail had 3000 remote condition monitoring systems across its
infrastructure, monitoring a diverse range of asset types (e.g., signal, point machine,
overhead lines) (Ollier, 2006).
Rail RCM are divided into two groups: train-borne RCM and trackside RCM.
Train-borne RCM is facilitated by sensors installed on a train, while trackside RCM
has been made possible with the help of sensors directly installed on the infrastruc-
ture (Bint, 2008). However, there is a crossover in that trackside sensing may monitor
22 Intelligent Infrastructure
rolling stock (e.g., detecting wheel flats), and trains have the equipment to sense track
conditions.
Proactive maintenance goes beyond this, and it is mainly considered in terms of
centralising and integrating the support currently provided to infrastructure main-
tenance by monitoring the condition of assets remotely but providing greater levels
of automation, interpretation, and integration of data sources. Potential failure or
unnecessary fixed-term replacements will then be prevented by providing proac-
tive information to the maintenance function. Currently, maintaining assets is done
through fixed schedules. This approach is time-consuming and costly and involves
risk both to staff required to go out on track and the potential risk of asset failure
before a routine inspection picks up a fault. This routine inspection approach can be
replaced by analysing real-time information regarding the assets and attending to the
trackside equipment only when necessary. Most critically, however, it can predict if
and when an asset is likely to fail, thus ensuring that it is repaired or replaced before
it becomes an issue that impedes performance or impacts safety. Therefore, predic-
tive and proactive maintenance in rail will move the railway, especially its main-
tenance and engineering activities, from “find and fix” mentality to “predict and
prevent” and potentially to “design and prevent” (Bint, 2008). In Network Rail, as
with several other settings, this kind of predictive maintenance is called “Intelligent
Infrastructure”.
FIGURE 2.4 ISO 13374 strategic framework (taken from Network Rail, 2015).
In addition to standards that guide the design and development of RCMs, specific
guidelines such as Engineering Equipment and Materials Users Association (EEMUA)
191 also exist to inform best practices for designing alarms, particularly alarm man-
agement systems. These guidelines inform effective alarm systems within industrial
settings (e.g., chemical manufacture, power generation, oil, and gas) but may not be as
effective or as easy to follow when incorporating “smart” technologies within more
everyday socio-technical systems (e.g., smart home, health monitoring systems).
Standards and guidelines apply to specific safety-critical domains (e.g., nuclear
industry NUREG-0711: Rev. 1 Human Factors Engineering Programme Review
Model). As a highly regulated environment, control systems adopted within nuclear
power plants, often equipped with a wide range of asset management technology,
need to follow a specific range of standards and guidelines. These guidelines have
led to the need for early design planning and early incorporation of human factors
activities.
24 Intelligent Infrastructure
Finally, ISO11064 covers the design of control centres. Specifically, ISO 11064-
4:2013 specifies ergonomic principles, recommendations, and requirements for
workstations found in control centres. The standard covers control workstation’s
design, emphasising layout and dimensions, primarily for seated, visual display-
based workstations, although control workstations at which operators stand are also
addressed.
There are two major issues with design standards such as those described above.
First, despite the standards and guidelines, there is very little (if any) guidance in
terms of how information should be presented to users. For example, ISO 13374-1
explores communication and presentation but does not specifically guide how to
achieve this from a user perspective. EEMUA 191 discusses best practices within
designing alarm systems but does not explore data collection and analysis methods
within complex socio-technical systems to choose the right alarm configurations.
Also, as we cover later in the book, predictive maintenance systems may be as much
about searching and interpreting information (user pull) as sending alerts and alarms
to a user (system push). (Houghton and Patel, 2015). While ISO 11064:4 concentrates
more on the physical design of the workstations (for visibility, reach, comfort, and so
on), it does not directly address issues of the content and design of the information
itself. Also, it is a general framework that must be modified to the specific needs of a
workstation. In practice, many maintenance workstations already exist (see Chapter
5) – the key issue is integrating new displays and workflows into that environment.
There is a more fundamental issue. ISO11064:4 aside, these frameworks take
their starting point as data and data collection and then move through increasing
levels of complexity to deliver prognostics and predictive intelligence. While this
approach may make sense in terms of software and hardware architecture, it means
that the intelligent support offered to the end-user is often shaped by the nature of
data – this can sometimes mean the wrong data and often means too much data. As
discussed in this book, this approach has to be complemented by an approach that
puts operator decision-making at the heart of the process and derives requirements
for data and information presentation.
One ISO that is firmly focused on user-centred design is ISO9241-210 Human-
centred design for interactive systems. This general ISO introduces the main prin-
ciples of user-centred design for an interactive product or service. We discuss this
ISO and its relevance in greater depth in Chapter 4.
2.6 CHAPTER SUMMARY
This chapter has defined what we mean by remote condition monitoring and predic-
tive maintenance – a systemic approach to combining sensor data, algorithms, and
analysis to predict the state of assets. The application of predictive maintenance to
various domains has been discussed, and we present one domain, rail, in-depth. The
concern, however, is that this technology is data-driven, and there is little guidance
on how to present meaningful guidance to operators. In the next chapter, we map
out the human challenges of implementing predictive maintenance in greater depth.
3 Challenges of Remote
Condition Monitoring
3.1 CHAPTER OVERVIEW
Chapter 2 detailed the technical nature of remote condition monitoring and high-
lighted some of the challenges with successful deployment. This chapter aims to
give further details on the major human challenges of remote condition monitoring.
This includes:
Amongst these three topics, diagnosis and technology (items 1 and 2) have focused
on significant research and practice. There have been many deployments in multiple
sectors (some of the examples noted in Chapter 2 of this book). Topic 3 (i.e., people)
is more difficult to tackle and has received less attention.
DOI: 10.1201/9781315587288-3 25
26 Intelligent Infrastructure
Again, these issues, particularly the first, second, and fourth, have only received lim-
ited attention. In interviews with experts from across a number of predictive main-
tenance implementations, Golightly et al. (2018) found that there are concerns that
unless predictive maintenance technology can be designed in a way that meets the
needs of operators, it often goes unused and fails to deliver the necessary benefits.
Similarly, work in metro systems (Kefalidou et al., 2015; Ciocoiu et al., 2017) has
highlighted the problems of integrating predictive maintenance technologies within
pre-existing maintenance regimes.
There is also a wealth of anecdotal evidence, talking to operators who work with
condition monitoring and prognostics systems. They can be difficult to use, are
poorly integrated into the work environment, or have been subject to limited (or no)
acceptance. Consequently, millions of pounds of investments may be wasted, and,
more importantly, organisations have been discouraged from introducing innovation
and change to the workplace.
This is not a specific problem that is unique to remote condition monitoring sys-
tems. The types of innovation we have described in this book so far have several
features and characteristics common in the roll-out of the technology. The rest of
the chapter describes these challenges in greater detail, as they form the basis for the
human factors works in the rest of this book. Before discussing specific challenges,
we introduce the general principles of human factors.
3.3 HUMAN FACTORS
Human Factors or Ergonomics (often used interchangeably) has been a signifi-
cant scientific and practical field since the 1940s. There are many textbooks that
give a thorough introduction to the field and its concepts, theories, and methods.
In Chapters 3 and 4, we introduce some of the basic ideas of human factors and
ergonomics. However, the interested reader could find out more from professional
bodies such as the International Ergonomics Association (https://iea.cc/), Chartered
Institute of Ergonomics and Human Factors (https://www.ergonomics.org.uk/), or
the Human Factors and Ergonomics Society (https://www.hfes.org/). Introductory
texts include Shorrock and Williams (2016) or Wickens et al. (2004) and the first
chapter of Wilson and Sharples (2015), amongst many others.
Challenges of Remote Condition Monitoring 27
For Wilson and Sharples (2015), human factors/ergonomics (HFE) refers to the
“theory and practice of learning about human characteristics and capabilities, and
then using that understanding to improve people’s interaction with the things they
use and within the environments in which they do so” (Wilson and Sharples, 2015).
This definition suggests that both theory and practice should be in place, requir-
ing an understanding of human needs, socio-demographic characteristics, and physi-
cal and cognitive abilities that lead to tangible design recommendations to design or
improve a product or service. It is very much an applied discipline. Human factors
specialists are not merely evaluators (although a big part of their role is to critique) –
they, ultimately, influence or specify the design of systems, products, and/or services.
Physical HFE refers to the understanding of the physical aspects of action and
interaction and the environmental context (e.g., temperature, illumination level,
accessibility) to ensure that the design fits the environment and the design supports
the user’s physical characteristics. Physical aspects of HFE have been a major part
of human performance reviews, occupational ergonomics, and the design of work-
places. It also covers anthropometrics. Much of physical ergonomics also relates to
physical work – manual handling, upper limb injuries, working in extremes of cold
or heat, and the physical design of equipment such as personal protective equipment
(PPE).
Cognitive aspects of HFE refer to thinking, sensing, perceiving, responding,
and decision-making elements of interaction. All interactions require some form
of understanding of the context, information processing, and decision-making.
Reviewing cognitive aspects requires understanding user perceptions, mental mod-
els, problem-solving, decision-making, and contributing factors such as expertise
and learning. Despite decades of work in cognitive ergonomics and great break-
throughs in understanding and exploring complex control settings, there is still much
to be done in this field, particularly now as we see complex settings where mul-
tiple people, technologies, products, and systems work together as part of a Joint
Cognitive System (Hollnagel and Woods, 2005). Growing complexities due to recent
technological advancements raised by artificial intelligence (AI) and machine learn-
ing (ML) have only increased the need for suitable, feasible, and effective ways of
understanding how people and machines work together.
Organisational aspects of HFE are concerned with wider settings of interaction
and broader aspects that influence socio-technical systems. Some of the elements
that should be considered as part of organisational ergonomics include training,
human performance, safety culture, change management, job satisfaction, and lead-
ership. Also, it is essential to point out that work does not happen in a social vacuum,
and the social and personal elements that people bring to their work are critical in
our understanding of how to design effective technologies and systems.
Table 3.1 provides some examples of typical human factors issues that we could
come across in the major areas of physical, cognitive, and organisational ergonomics
in the context of railway maintenance and remote condition monitoring.
28 Intelligent Infrastructure
TABLE 3.1
Typical Topics in Ergonomics and Human Factors and Relevance to Remote
Condition Monitoring
Area Example topic Relevance to remote condition monitoring
Physical Manual handling Trackwork involves the lifting and positioning of
equipment, assets, and materials. Better planning
through remote condition monitoring can reduce the
need for redundant work and therefore reduce manual
handling exposure (Riley, 2006).
Environmental Inspection and maintenance often have to take place at
conditions night or in cold, wet, or enclosed conditions (e.g.,
working in a tunnel). These conditions can impair
human performance and present hazards. Asset
management can reduce the need for this work through
remote sensing of the asset state (Golightly et al.,
2013b).
Display screen Remote condition monitoring workstations must be
equipment (DSE) designed to comply with the needs of good DSE design
(e.g., http://www.hse.gov.uk/msd/dse/) and to ensure
that issues with glare, size, and portability are not
deterring from the effectiveness of remote condition
monitoring.
Cognitive Situation awareness Situation awareness supports maintaining an accurate
mental model of the present status of a system while
being able to predict its future state (Endsley, 1995).
While remote condition monitoring can support this
wide overview of many assets and their current and
future state, automation should also be designed to
keep the user “in the loop”.
Human–computer Asset management and remote condition monitoring
interaction should be designed to optimise information
presentation and apply principles of user-centred design
(Houghton and Patel, 2015). This is critical when
addressing the area of data visualisation.
Expertise and Expertise in remote condition monitoring comprises
decision-making knowledge about assets, their likely behaviour, and
coping strategies. These strategies influence decision-
making. Future design should support and complement
the most adaptive strategies (Dadashi et al., 2021).
Organisational Team coordination Asset management and maintenance involve the
coordination between multiple roles, particularly where
an unexpected failure stops the service. Processes and
technologies should be designed to reflect team
coordination.
(Continued)
Challenges of Remote Condition Monitoring 29
II
Stevenson’s poems are contained in four small volumes—
Underwoods, Ballads, Songs of Travel (a collection made by himself,
but published posthumously), and A Child’s Garden of Verses. Of the
four volumes the one that has enjoyed most popularity, as well as
most critical esteem, is A Child’s Garden of Verses, which book,
although, by Stevenson’s account, very easily produced, has the
value of being unique in scheme and contents. The other volumes
have less in them of wide interest, and so they are less generally
read. Certain poems, such as the Requiem (“Under the wide and
starry sky”) and The Vagabond (“Give to me the life I love”) arise
whenever the name of Stevenson is fondly mentioned; they are, as it
were, the stock-in-trade of the conversational anthologist, who, in the
same spirit, will have suggested to him by the name of Meredith the
words, “Enter these enchanted woods, Ye who dare.” These two
poems are not the best poems Stevenson wrote; but they are handy
for remembrance. That explains their frequent employment; that, and
their appropriateness to the conventional idea of Stevenson, which is
based upon a sentimental and mediocre marvel at the
unconventionality of the open road.
The best poems Stevenson wrote are his ballads. With a story to tell,
he was keener to represent truly the subject-matter upon which he
was engaged; and this engendered the “heat of composition,” if it did
not always spring from the native heat or intensity of inspiration. The
ballads, especially Ticonderoga, have a swift effectiveness and an
adherence to theme which is not so marked in the poems provoked
by occasional events. In these the rhyme and form sometimes lead
the way, and the poems become exercises in friendly versification,
without much feeling, and with only that Scottish affectionateness to
which reference has already been made. Examples of impoverished
emotion may be found in the two poems expressing gladness at
visits from Mr. Henry James. As cheerful little outbursts of pleasure,
such poems, in manuscript, would be interesting, even delightful: as
poems they fall short of complete success, even in their own class,
for the reason that they are as conversational and as fluent as
Stevenson’s letters, and are diffuse as his prose rarely is.
Better than these are some of the dryly humorous Scots dialect
poems, such as The Spaewife, with its refrain of “—It’s gey an’ easy
spierin’, says the beggar-wife to me.” These again are often purely
experimental versifications; but they are more than the casual
rhymings of the pleased householder, and they have more interest
as poetry. Far and away better even than these, however, because it
is the expression of a personal and, I think, a deep feeling, is that
poem, included in Songs of Travel, and quoted in The Master of
Ballantrae, which is untitled, but which is written “To the tune of
Wandering Willie.”
“Home no more home to me, whither must I wander?
Hunger my driver, I go where I must.”
In this poem there seems to be real emotion, as I think there is in the
dedication to Mrs. Stevenson of Weir of Hermiston. In other poems
there is a grace and the mellifluous flow of words which Stevenson
could always command; but the verses make a pattern, and a
pattern of only occasional significance. They are thus robbed of any
power to move us æsthetically.
The two long narrative poems, The Ballad of Rahero and The Feast
of Famine, are both well-sustained by a body of incident. They have,
in lieu of emotion, a certain vividness of excitement. One is excited
by what is going forward, one must read on for the story. In the
degree, therefore, in which one’s attention is removed from the
versification, these two narratives are good; and those other verses
based on legends—Heather Ale and Ticonderoga—would be
sufficient to emphasise the fact that Stevenson loved a story and
was always at his best with a tale to spin. When, however, we reach
poems in which no story is to be told, we are confronted with an
absence of emotion which robs the pages we read of all that
exceeds mere pleasurable line-scanning. Happy lines there are,
turns of phrase that perhaps have given rise to the poem into which
they are woven. But they are only, at best, the amiable pleasantries
of one who could handle with dexterity the words of whose music his
mind was full. “The bright ring of words” is not the phrase of a poet; it
is the phrase of a connoisseur, and of one who used words as a
connoisseur uses them. The poet is a singer first: he does not make
a poem out of his craft. And the tendency to diffuseness which mars
many of the longer lyrics is a curious instance of failure in a writer
who regarded compression as an essential of good style.
III
In A Child’s Garden of Verses Stevenson was doing a thing which
had never really been done before. There are nursery rhymes which
crystallise children’s ideas; but this book actually shows, in what we
must believe to be an extraordinarily happy way, the working of a
particular child mind over a great variety of matters. Its excellence is
due to the fact that Stevenson’s young days, lonely as some of them
had been, had never lacked interest, had always been full of those
simple and direct pleasures of incident and encounter and memory
which happy children enjoy. The world had been full of a number of
things; and the memory of those things had abided. It was the
memory of a fanciful rather than an imaginative childhood, a
childhood of superstitions and sports, of a buried tin soldier and of
the pleasant land of play; but we must not forget that such poems as
My Treasures, poor in some of their lines, are finely imaginative
reconstructions, the naïveté of which prevents many readers from
estimating their quality. So with The Unseen Playmate, which,
although it is a poem for grown-ups, reveals an understanding of a
most important fact in children’s games far more profound than are
the pretentious and unconvincing lines to R. A. M. Stevenson in
Underwoods. Even if the idea of The Unseen Playmate may be the
idea of a grown-up pretending, the writing of this, as of the other
verses, is almost without lapse, charmingly simple and natural. I
believe it is a fact that children appreciate and even delight in A
Child’s Garden of Verses, not merely at the bidding of their parents,
but as a normal manifestation of taste. This in itself would be a proof
that the book is already a secondary nursery classic. For our present
purpose, if that does not seem rather an over-bearing way of valuing
a book so slight in form, it is sufficient to say that Stevenson’s
success here was due to the fact that he was legitimately using the
memory of actual experience. Too many of his serious, or grown-up,
poems show their models; too many of them flow undistinguished by
any truly poetic quality; too many of them are experiments in metre
or rhyme, such as one may write for fun, but never for free
circulation. The Child’s Garden of Verses alone, then, of the four
volumes, exhibits a strict harmony of design with performance. Its
dedication to Stevenson’s nurse, Alison Cunningham, serves only to
make the book more complete.
IV
Implicit in the strictures upon Stevenson’s poetry which have
preceded this paragraph is the assumption that Milton’s
requirements of poetry—that it should be simple, sensuous,
passionate—is fundamentally true as applied to lyrical poetry. It
would be troublesome to apply such a test to many of the minor
poets; and it may be that a few of Stevenson’s poems would stand
the test. Not many of them, however, because none of them shows a
depth of emotion uncommon to the ordinarily sensitive person.
Stevenson was sensitive to many things; without sensitiveness he
could not have written A Child’s Garden of Verses or that very
excellent ballad Ticonderoga. But sensitiveness is only a poor
substitute for emotion; and Stevenson’s emotion ran in the few
ordinary channels of the normal Scotsman. He loved home; he loved
those around him; he desired to be loved, to be free of the fear of
poverty, to live in comfort and in health. Those things he felt deeply,
as Scotsmen, as most men, do. He loved truth; but it was a
conventional truth; a truth, that is to say, improvised from ordinary
usage, from hearsay, from the dogma of “that station of life”; a truth
such as any man who finds himself born in a little pit of earth may
harden his moral shell and his imagination and stultify his spiritual
curiosity by accepting; and it was a truth out of which Stevenson was
escaping towards the end of his life. But in all this love of virtues and
duties and usages there was never until Stevenson’s emergence into
the greater freedom of life in the South Seas the passionate love of
anything for its own sake. If he loved the open air it was with a
pleasant, “playing” love, a sort of self-indulgence. Over his heart he
kept the watchful guard of a Protestant Scotsman. It was unmoved, a
secret, not to be known. It did not inform his work, in which there is
sometimes a heat of composition, or even a heat of feeling, but
never the cold heat of profound and piercing emotion. That he was
capable of being easily moved, that he loved virtue and hated cruelty
and wrong, these things are true. That he could grow hot at a
calumny, as he did in the defence of Father Damien, is equally true.
But these things are the signs of a prudent man, eagerly interested
in life, rather taking pleasure in the thought that he is hot to attack
injustice; not of a profound thinker or of a poet. They warm us with,
perhaps, affection for Stevenson; they keep alive our admiration for
him as an attractive figure in our literary history. They do not thrill us,
because they appeal to the interest and excitement and honesty and
feeling in us, and not to those more secret, more passionate
reserves which we yield only to the poet.
VI
PLAYS
I
It is a commonplace of dramatic reporting, which in spite of its
frequently doubtful application has the truth of an old saw, that the
novelist cannot write plays. Certainly, it would seem that the qualities
which go to the making of good plays are not precisely those which
make good novels; for while it is possible to conceive a novel in
terms of narrative, descriptions of abounding nature, psychological
analysis, and tableaux, the play has rules more strictly objective and
more definitely rigid. Now if we, for the moment, pass over the
question of Stevenson’s collaborator in the four printed plays with
which his name is associated, and if we, for this occasion, treat them
as though they were his work entirely, we shall be better able to
distinguish certain remarkable characteristics of these plays, and,
anticipating certain general conclusions to be made later, of
Stevenson’s talent.
Stevenson, we are all aware, was never, strictly speaking, in spite of
Catriona and Weir of Hermiston, a novelist. He was a writer of many
kinds of stories; but they were not primarily, until we come to Weir,
domestic or psychological. Many of them were what no doubt would
commonly be called “dramatic,” in the sense that they contained
scenes of some violence; but for the most part they were narrative
interspersed with tableaux. They were “picturesque,” not because
they were startlingly visual, but because Stevenson had that flair for
the odd, the startling, or the vivid effect of contrast which is generally
described by the word “picturesque.” It was the oddness of Dr. Jekyll
and Mr. Hyde that allured him before he became oppressed by its
symbolism. It was, equally, oddness that always attracted him in
character: he had no profound sense of character, for this reason.
Passivity he never understood. His characters must forever be in
action. That, it might be supposed, was in itself a first reason for
turning to the theatre, since, according to modern dramatic reporters,
“drama” is a word synonymous with the word “action.” Action,
something doing—that, by the recipe, is the certain play. But while
action may give a play breathless suspense, while it may provide the
kind of play which, in a specifically theatrical sense, is called a
“drama,” action is not the whole battle. To action, or at least to the
psychological excitement created by a sense of action in progress
and a climax pending, must be added a very powerful sense of what
is effective in the theatre. A pause, a sound, verbal repetition, an
abrupt change—these things are crude examples, chosen at random
from among the obvious instances of what contributes to the sense
of the theatre. If we think of such things as the tapping of Pew’s stick
(in Admiral Guinea), and, in Deacon Brodie, the appearance of the
masked Deacon at the window by which Leslie is watching for him,
we shall realise that in some degree, in some very obvious and
primitive form, Stevenson was possessed of this attribute. But one
thing we shall infallibly discover him to lack, a thing which Mr. Henry
James missed in Catriona, a thing which has vital importance in
drama—the visual sense. These plays show no real power of
visualising a scene. Picturesque they all are; they all have qualities
which make them engrossing—as reading. But they are not focussed
for the eyes, and they are not well constructed for real dramatic
effect.
Deacon Brodie is in five acts and eight tableaux, and its effects are
indescribably broken, so that irrelevancies are numerous, distracting
side issues over-emphasised, and so that the Deacon is almost a
minor character. It is hard to realise that there are only a dozen
persons in the play, for their comings and goings are so frequent as
to give the effect of a confused number of straggling participants in
desultory action. The play itself centres round an historical figure—
Deacon Brodie—who was an honest man before the world by day,
and by night an expert cracksman. His name is familiar both in
criminal history and in the annals of Edinburgh, where his activities
became, after his death, notorious. In the play, Brodie at last is eager
for reform; but one of his cronies, tempted by a Bow Street runner,
and the only one of Brodie’s friends to yield to temptation, betrays
him. Though Brodie escapes, his absence from home has been
discovered in the excitement consequent upon his father’s death,
and, when arrest is imminent, he takes his own life. Stevenson had
found the details of Brodie’s life while he was preparing the sketches
collected under the title Edinburgh: Picturesque Notes; and it is
conceivable that in some measure the play’s technique was a little
influenced by a reading of some eighteenth-century episodic plays,
such, for example, as Gay’s “Beggar’s Opera,” which is similarly
broken in construction, though more permissibly so, because “The
Beggar’s Opera” is no more than a skein in which ballads and satire
may be found to provide our entertainment. This mention of “The
Beggar’s Opera” must not be taken too seriously, however, because
although that play deals with the life of highwaymen and pickpurses
and thief-takers in the eighteenth century, as Deacon Brodie does, it
is profoundly real, whereas Deacon Brodie is only too obviously
modern fake. Macheath and Polly Peachum are infinitely more real
than Brodie and his doxy. Moreover the ensemble in Deacon Brodie
is on the whole poorly conceived. The minor persons are mere
figures, introduced to stand here or there, or do this or that, and are
labelled with names and idiosyncrasies. The major persons, though
more detailed, have an equal lack of vitality. It is necessary to add
the further explanation that Deacon Brodie is the first of the plays,
and that it dates from 1880. It is easily the least coherent of them all.
Stevenson was to improve upon Deacon Brodie in that respect, at
least.
II
The two lightest plays—Beau Austin and Macaire—are experiments,
the one in manner, the other in bizarre or, as it is styled by the
authors, “melodramatic farce.” The manner of Beau Austin is the
manner of the costume play. It is highly sophisticated, and its
keynote is powder and patches. The beau is at his toilet, and one of
the women he has betrayed is in the town, still sick with despair at
her soiled virtue. Her true love hears from the lady’s lips the story of
her betrayal, and, on being forbidden to challenge the beau, contents
himself with demanding a marriage ceremony. His flatteries are
effective, the beau consents, and the formal proposal is made, only
to be rejected by the lady, whose hauteur is aroused. So matters
stand when the lady’s brother, learning by chance of the betrayal,
insults the beau before an important personage. As climax, the beau
proposes publicly, and is as publicly accepted. It will be seen that the
play could not claim, excepting in respect of verbal artifice, to be
more than a pretty jig-saw. It could have no effect of reality: the effect
desired by the authors was one purely of the stage. Verbally it is
exquisitely dexterous. That is its undoing. The attempt is made to
convey in words something more than the action of the piece would
successfully carry: words are to create an atmosphere of the
eighteenth-century fashionable life, to indicate the possibility that
calm picturesque heartless exteriors shielded even then hearts that
beat warmly beneath lace and brocade. The play was a pretence
that nothing was something, a pretty moving picture under the
perception of which, beating out in pianissimo airs from appropriate
music, and the faint throb of an unseen minuet, was the delicate
heart of the period. It was an æsthetic view of the eighteenth century,
the century of Fielding and of Smollett, tinkered about to make a
perpetual bal masque, or, shall we say, a picture by Watteau or
Fragonard. In point of fact the play is too slight to bear its weight of
intention: it remains verbal. As drama it is more negligible than
“Monsieur Beaucaire” or “The Adventure of Lady Ursula,” because
its literary pretensions are so much more elaborate. It has
sometimes fine shades of close verbal fence that are Meredithian: it
is better to read than it could be to see. But it is an attempt, one
might say an almost basely cunning attempt, to capture the theatre
as a place where costumes grace a barren play. It failed because its
authors were two conscientious literary men, bent upon a superficial
perfection undreamed of by practical dramatists. Just as Cowper, in
translating Homer, made an epic for a tea-party, so Henley and
Stevenson made about the rational and cynical eighteenth century a
sophisticated play for a boudoir. They concentrated upon the
superficial, and only said, but did not show, that the men and women
of the eighteenth century had hearts as true and passionate as those
of our day. The play lacked realism, and, more disastrously, it lacked
reality.
On the other hand, Macaire has a thin air of jocularity which almost
carries it through. It has a sententious cleric, a drunken notary, a
repetitious father for the bride, a courteous host, a little mystery of
the bridegroom’s nurseling days, the facetious Macaire and his
companion. It has all these things, and it has an idea, strong enough
for a single act, stretched to its thinnest over several acts which
demand cuts more severe than the authors allow.
Macaire escaping from justice, threatened each moment, in the face
of the audience, with instant arrest, carries himself with unfailing
sang-froid through all his difficulties but the last. Finding a chance of
sport, and possibly of profit, he impersonates an erring father. The
real father appears. Macaire still, after the manner of Mr. Jingle, is
imperturbable. Competition follows, until the desire for the genuine
father’s money becomes too strong for Macaire. Then only does he
show the blackness of his heart, which does not shrink, in such
desperate situations, from murder. So Macaire, still talking, still
watchful and unscrupulous, is brought to bay. Fiercely turning, in a
picturesque situation, upon the stairs, he is shot by a gendarme on
the stage. That is a skeleton of the play; but the play is again a
literary play, so that sensationalism will not redeem it. By repetitions
of catch-phrases and by trivial incidents which (e.g. the exchanging
of the wine-bottles) are not unknown to the humbler kinds of drama,
the story is continued until its idle joking can no longer be suddenly
stirred into flaming melodrama by the noise and zest of bloody crime.
It has many shrewd bids for theatrical effectiveness; but it faints for
want of a fabric upon which its devices might flourish and
triumphantly justify themselves.
III
The fourth play, Admiral Guinea, has fine qualities, both literary and
dramatic; it is the least literary and the most dramatically effective of
all the plays. It contains one figure, in Pew, which might have been,
as far as one may judge in reading, a hauntingly gruesome object;
and, in spite of Stevenson’s own subsequent contempt for this play
and for Macaire, shows a greater, if conventional, power of
simplification than does any of the other plays. Admiral Guinea, a
retired and penitent slaver, refuses his daughter her lover, on the
ground that the lover is ungodly. Pew, an old associate of Admiral
Guinea, become blind for his sins, and still full of vengeful
wickedness, arrives in the neighbourhood, catches the lover drunk,
leads him back to Admiral Guinea’s cottage, and tries, with his aid, to
rob his old captain of certain riches which he supposes to lie in a
brass-bound chest. The young man’s reaction, their discovery by
Admiral Guinea, the violent death of the unrepentant Pew follow;
whereupon the lovers are suitably blessed by Admiral Guinea.
It has been said, above, that this play shows a greater power of
simplification than the others; the action of it is certainly quicker,
more obvious, less choked with verbal expressiveness, than is the
action of the other plays; and in so far as this is so it would appear
that Admiral Guinea is a considerable advance, technically, upon
them.
The simplification is, to some considerable extent, effected by a
strange poverty of invention, and the play is likest of all to those
nondescripts which Stevenson as a little boy must have performed
upon his toy stage, with paper figures pushed hither and thither in tin
slides upon the boards. In spite of that, Admiral Guinea is the best of
the plays because, in a higher degree than its fellows, it is truly
actable. We cannot regard the confused cramped episodic Deacon
Brodie as theatrically effective. Equally it is impossible, from the
standpoint of public performance, to consider as satisfactory either
Beau Austin or Macaire. Admiral Guinea, however, even if it belongs
to a class of play which is associated in our minds with such titles as
“Black-Eyed Susan,” has its action very largely comprised in the
material put upon the stage; it has the obvious stage effects of
darkness and the dreadful tapping stick of Pew; and it has
picturesque struggles, death, wounded and reasserted honour, and,
for these plays, a minimum of soliloquy. More it would be impossible
to claim for Admiral Guinea without seeing it performed: again we
have types roughly “mannered” to serve as persons of the play: but
they are types clearly in accordance with tradition, and they preserve
their interest fully until they are done with and put away with the
footlight-wicks, and the tin slides, and the other paraphernalia of the
toy stage—paper figures, a penny plain, and twopence coloured.
IV
For that brings us to the pathetic final explanation of the failure of the
Henley-Stevenson plays. We may say that they are deficient in
drama, or that they are trivial in theme, or that they have no visual
sense to illumine them for our eyes; but the truth is that they fail
because they are false. The theatre has in it much that is false, much
to which we deliberately shut our eyes in order that we may accept
the dramatist’s formal conventions. We do not, in the theatre,
demand that “King Lear” shall be accompanied by a pandemonium
of crackling tin and iron and artificial whoopings of wind. Those
things we prefer to imagine for ourselves. But somehow the mixture
of legitimate convention and the basest imitation of reality has been
confused in the theatre. The exaggeration regarded as necessary by
an effete system of acting and production has created other
unpardonable falsenesses. The stage has been a place upon which
actors disported themselves. It was of such a stage that Stevenson
thought. In each case he hung a play upon a sensational figure—
Brodie, Macaire, Pew, and, in a much lesser degree, upon the
picturesque figure of Beau Austin. To him the drama was nothing but
play. It was an excuse—nay, a demand, for unreality. He supposed
that stage characters really were cardboard figures such as he had
known, moralising ranters, virtuous girls, spouters of Latin tags,
pious brands from the burning, handsome courageous puppet-like
juvenile leads, and so on. It never occurred to him to put a real figure
in a play: he never supposed that a character in a play had any end
but to be put back in the box with the other playthings. That is really
the cause of the shallowness of these four plays. As Stevenson
admitted to Mr. Henry James, he heard people talking, and felt them
acting, and that seemed to him to be fiction. But to hear people
talking and to feel them acting bespeaks a very unmaterial
conception of them: if a character in a play talks, however
monotonously, without developing any personality save that of verbal
mannerism, we are bound to feel that he has not been realised. And
just as Stevenson realised none of the characters in his plays, so we
are powerless to realise them. We find them, as Professor
Saintsbury pathetically found Catriona herself, bloodless. Professor
Saintsbury found Catriona full of sawdust, while of the characters in
the plays we have used the word “paper”: very well, the impression
of lifelessness is as clearly felt in each case. And such an
impression, carried to its logical end, explains why, in at least one
department of letters, Stevenson from the first mistook his ground.
Not one of the four plays has serious value as an example of
dramatic art; it is clear that not one of them so far has commended
itself to the public or to the actor-managers. Yet the plays were
obviously set to catch the popular taste, and their literary finish, a
confession in itself of an absence of dramatic impulse, does not
succeed in commending them to those who judge by more exacting
standards.
VII
SHORT STORIES
I
Stevenson himself establishes the fact that he found short-story
writing easier than the writing of novels. “It is the length that kills,” he
confessed. But length offered difficulties in the longer stories
because Stevenson, besides lacking the physical endurance for
continuous imaginative effort, had the experimental and inventive
mind rather than the synthetic or the analytical. It was easier for him
to see the whole of a short story. It could be compressed: it had not
to be sustained. And in the writing of a short story his confidence
never slackened. He was then not sailing in uncharted seas. It is for
this reason, in the first place, that Stevenson’s short stories are
better as works of art than his long ones. A little idea, a flash, it may
be, of inspiration; and Stevenson had his story complete, ready for
that scrupulous handling and manipulation which the actual
composition always involved. He did not greatly deal in anecdote; his
psychological studies are inclined to be hollow; but he was perfectly
effective in his not very powerful vein of fantasy, could tell a fairy tale
with distinction, succeeded once without question in picturesque
drama, and, when he fell to anecdote, as in The Treasure of
Franchard, Providence and the Guitar, and The Beach of Falesá, he
was pleasantly triumphant. Moreover, in two of his “bogle” stories,
the one inserted in Catriona, and the other famous to all the world as
Thrawn Janet, he seems to me to have risen clearly above anecdote
with matter which might have been left as unsatisfactory as it
remains in The Body-Snatcher.
In one of his reviews Stevenson speaks of “that compression which
is the mark of a really sovereign style.” Compression is no more the
mark of a sovereign style, of course, than it is of a suit of clothes.
Compression brings with it obscurity, and is a mark of self-
consciousness. What Stevenson meant was possibly a justification
of apophthegm and figure. He rather enjoyed what somebody once
called “minting the arresting phrase.” There is, at any rate, a
palpable connection between our two quotations. But it is certain that
precision, austerity, or, if I may use the word, chastity, of expression
is a sign of good style; and compression, where it takes the form of
heightening and intensification of effect, is the mark of a good short
story. It is the mark of Stevenson’s best stories. It is the mark of
Thrawn Janet, of The Pavilion on the Links, of The Bottle Imp.
Sometimes, after promising well, Stevenson abandons himself, it is
true, to his natural Scottish aptitude, and literally “talks out” such
tales as Markheim and A Lodging for the Night; but, quite as often,
his judgment beats his inclination, and the result is a classic short
story in a language not too brilliantly equipped with examples of the
craft.
For the short story is above all a matter of justesse, by which word I
mean to suggest delicate propriety of expression to idea. Mr. Henry
James can tell a short story, because Mr. Henry James writes, as it
were, with a very fine pen. Stevenson was not comparable as an
artist with Mr. Henry James; but he wrote in a less rarified
atmosphere; and it is still practically an unsettled question whether a
distinguished artist (one who perfectly expresses a fine conception),
such as Turgenev or Mr. Henry James, is the superior or the inferior
of the writer with more tumultuous sympathies whose sense of form
is less than his sense of life. So that when Stevenson wrote The
Pavilion on the Links, or The Bottle Imp, or Thrawn Janet, or
Markheim, he was writing particular stories of which only the last,
one supposes, could ever have occurred to Mr. James as a subject
for a short story at all. Conversely, one sees Stevenson blundering
into the bluntnesses and certainly the ultimate failure of Olalla, with
the knowledge that his delicacy of style was more marked than the
poignancy of his perception; and the psychological explorations of
Olalla are jejune stumblings compared with the finished delicacy of
“Washington Square.” One does not think, in reading, of Mr. James;
but one may perhaps be permitted to illustrate a point by a reference
to his work, which has no precise significance as a parallel. That
fact, I hope, will excuse a momentary comparison for the purpose of
showing that Will o’ the Mill, for all its stylistic accomplishment, is a
barren piece of moralising. Where Stevenson essayed profundity, as
all writers are drawn to essay profundity, whether it is from natural
profoundness or from the instinct of imitation, he was badly
hampered by his inexperience as an inductive philosopher. Both Will
o’ the Mill and Markheim are, as it were, appendages to that doleful
failure Prince Otto. They were experiments for Stevenson in a
particular genre for which talent and his mental training had lent him
no aptitude. It was on other work that he more successfully took his
stand as a writer of short stories. His success—considering that we
are now examining his position among the masters of our literature—
can only be attested where his work stands supreme or, at any rate,
is clearly distinguished, in its own class. It cannot be doubted for one
moment that Stevenson wrote some exceedingly fine short stories, fit
to be compared, in their own line, with any that have been written in
English. What follows must be read in the light of this claim. In their
own way, I regard The Suicide Club, The Pavilion on the Links,
Providence and the Guitar, Thrawn Janet, The Treasure of
Franchard, The Beach of Falesá, and The Bottle Imp as first-class
short stories. In a distinct second class I should place The Rajah’s
Diamond, some of The Dynamiter stories, The Merry Men, Will o’ the
Mill, Markheim, Olalla, The Isle of Voices, and Dr. Jekyll and Mr.
Hyde. The least successful short stories seem to me to be The Story
of a Lie, A Lodging for the Night, The Sire de Malétroit’s Door, The
Misadventures of John Nicholson, and The Body-Snatcher. I am
aware that one at least of the stories which I have placed in this third
division—The Sire de Malétroit’s Door—has given great pleasure to
many readers, and has even been not without its direct influence
upon Stevenson’s imitators, while another—A Lodging for the Night
—is greatly admired, and has been very highly praised; so that it
seems hardly necessary to say that the classification is roughly
made, and that it is only here attempted for reasons of convenience.
The stories will hereafter be grouped according to subject or
treatment, and will be examined individually. Those in the first
division are, I think, completely successful in their own conventions;
those in the second division are either incompletely successful or
successful in conventions which seem to me inferior in artistic value;
those in the third division are, as far as I can see, unsuccessful
either because they fail to impose their conventions upon the reader
or because they fail to convince the reader that Stevenson had
mastered the craft of short-story writing. But, upon the whole, I
believe Stevenson’s short stories to represent more successfully
than any other part of his output the variety and the brilliance of his
talent. It is for this reason that I shall endeavour in some detail to
justify the divisions indicated above, and to emphasise the fact that
such tentative distinctions, even if they prove inaccurate in the case
of some one or two stories, may yet have some value as providing a
basis for agreement or disagreement.
II
For that reason I shall add that the stories in the third division seem
to me to fail for these reasons. The Story of a Lie is obviously
prentice work. It is presumably based upon some experience of his
own in France; but the action, once transferred from the Continent, is
filled with sentimentality. Although written, apparently, much later
than The Story of a Lie, The Misadventures of John Nicholson is a
protracted anecdote which does not awaken very much interest by
its attempt to blend humorous exaggeration with bizarre incidents.
The Body-Snatcher is one which Stevenson had to supply in order to
satisfy a journal with which he had made a contract. It is meant to
shock us, but it loses power before the climax, which thereupon fails
to shock. The idea is horrible, and affords scope for much dreadful
detail: Stevenson, however, perhaps through ill-health, was
unsuccessful with it, and possibly the ugliness of the whole thing is
at fault. For The Sire de Malétroit’s Door I must confess to the
greatest distaste. It seems to me to have neither historical nor
human convincingness; and the phrase at the end of the story, “her
falling body” very significantly conveys the pin-cushion substance of
the demoiselle whose indiscretion gives rise to the sickly and cloying
tale. The last story in this division is one that enjoys great reputation,
first because it deals with Villon, second because there is an
outburst of Villon’s against the red hair of a murdered man, and last
because there is an elaborately written but entirely inconclusive
duologue between Villon and his host. The story seems to me to be
without point or form.
I believe that popular admiration for A Lodging for the Night is largely
founded upon tradition or imitation, like the popular admiration for
Shakespeare, without the basis of fact upon which the popular
admiration for Shakespeare rests. It is well known that popular
appreciation of great things is shallow, and that it rises from a
common attempt to emulate the enthusiasm of the apostles of Art.
Unfortunately, popular appreciation is more easily aroused by artifice
than by art. Accordingly, those who have been taught to cite “Put out
the light, and then—Put out the light” as a profundity are ready to cite
with equal conviction the saying of Villon in this story that the
murdered man had no right to have red hair. It is one of those
dreadful æsthetic blunders that quickly pass into unquestionable
dogma. If no protest is made, if those who detect an imposture
remain supine, the false continues to masquerade as the
magnificent; and common opinions are so impervious to proclaimed
fact that it is at length impossible to cope with them, save by some
such wearisome exposition as this. It should be remembered that
common appreciation of art is not guided by principles but by
intuitions and imitations. The decay of a thing once widely popular is
slow; and it is due, not to any native perception of mistake, but to the
sluggard realisation that the old enthusiasm is less ardently
canvassed than it was. A Lodging for the Night has enjoyed great
repute, because Stevenson “found” Villon at a time when other
young men were finding Villon; and now that Villon is quite settled
among the young men, Stevenson’s essay on Villon and his story
about Villon have reached the larger public that is always some
years after the fleeting fashion. The result is that, by imitation of
those who ought to have known better, and even by its muddled
acceptance of a bad play about Villon (called “If I were King”), the
public has been led to esteem A Lodging for the Night as something
more than the piece of laboured artifice that it always was.
In the second class I believe that The Rajah’s Diamond, The
Dynamiter, and Dr. Jekyll and Mr. Hyde are very efficient pieces of
craftsmanship, strong enough in invention to delight that typical
person called by Mr. H. G. Wells the “weary giant,” engrossing
reading to the accompaniment of cigars and whisky-and-soda, but
not, in the way of art, quite what we require from works of creative
imagination. The Merry Men, with one striking piece of
characterisation, has vigour, but poor form and several superfluities
of invention. The Isle of Voices is a pleasant enough fairy tale, but
clearly inferior to its companion piece The Bottle Imp. The other
three tales, Will o’ the Mill, Markheim, and Olalla are all
psychological studies of a kind that is nowadays called arid. That is
to say, they have greater elaborateness of treatment than their
intrinsic importance quite justifies. Will o’ the Mill is written with great
softness and delicacy, in a sort of slow and lulling drone very sweet
to the ear; Markheim has great virtuosity, is faint and exquisite in
manner, feeble in perception, and is sometimes, I believe, false in
psychology. Its plan and its manner would only be finally true if its
understanding pierced more sharply and finely to the heart of truth. It
lacks penetration. Olalla is, in many ways, fine, in some, beautiful. It
is, however, as Stevenson came to be aware, false. It is false, not
because it is insincere, but because Stevenson’s knowledge had not
the temper and the needle-like capacity to go ever deeper into the
subtleties upon which he was engaged. I suspect that he dared not
trust his imagination, that his imagination had more ingenuity than
courage or strength. The story does not produce æsthetic emotion: it
is as though the author had made a fine net to trap a moonbeam, as
though, when he thought to have come at the heart of the matter, it
had escaped him. He was perhaps not wise enough in the mysteries
of the human soul. Sensitiveness, and the desire to create a
passionate beauty, were not fit substitutes for that patient and
courageous, that fearless imagination which alone could have given
truth to so simple and so unseizable a problem. More, in his handling
of the conclusion of his tale, Stevenson’s emotion fell to a lower
plane, and his talent played him quite false. He became too intent
upon his rendering of the idea; his literary sense took command