Learning from
Operational
Experience
Annual Report 2014/15
Copyright
© RAIL SAFETY AND STANDARDS BOARD LTD. 2015 ALL RIGHTS RESERVED
This publication may be reproduced free of charge for research, private study or for internal
circulation within an organisation. This is subject to it being reproduced and referenced
accurately and not being used in a misleading context. The material must be acknowledged as
the copyright of Rail Safety and Standards Board and the title of the publication specified
accordingly. For any other use of the material please apply to RSSB's System Safety Director for
permission. Any additional queries can be directed to enquirydesk@rssb.co.uk. This publication
can be accessed by authorised audiences, via the RSSB website: www.rssb.co.uk
Written by Greg Morse D.Phil AIRO
Published: July 2015
If you would like to give feedback on any of the material contained in this report, or if you have
any suggestions for future editions, please contact:
Greg Morse
Operational Feedback Specialist
020 3142 5467
greg.morse@rssb.co.uk
Additional hard copies may be ordered at cost price by contacting the RSSB enquiry desk on
020 3142 5400.
Contents
Report summary..............................................................................7
Summary of key points arising during 2014/15 ............................................... 7
1 Introduction .................................................................................9
1.1 Purpose ...................................................................................................... 9
1.2 Scope and structure ................................................................................. 10
2 How do we learn?.......................................................................10
3 What can help us learn? .............................................................12
3.1 Measuring safety performance ................................................................ 12
3.2 Investigations ........................................................................................... 14
3.2.1 Accident investigation assistance ............................................................... 15
3.2.2 Incident Factor Classification System ......................................................... 16
3.3 CIRAS ........................................................................................................ 18
3.4 Right Track................................................................................................ 19
3.5 SMIS+ and Close Call ................................................................................ 19
4 Investigations – lessons learnt in 2014/15 ..................................20
4.1 RAIB Bulletins ........................................................................................... 22
4.2 RSSB analysis of key RAIB recommendation themes............................... 22
5 CIRAS – lessons learnt in 2014/15 ..............................................26
5.1 Who reports to CIRAS?............................................................................. 27
5.2 Why do people report to CIRAS? ............................................................. 27
5.3 Key issues of concern in CIRAS reports during 2014/15 .......................... 28
5.4 Positive outcomes from CIRAS reports .................................................... 29
5.5 Summary .................................................................................................. 31
6 Right Track – lessons taught in 2014/15 .....................................31
7 Lessons learnt in 2014/15 – train operations risk .......................32
7.1 Statistical overview .................................................................................. 33
7.2 Overseas accidents................................................................................... 33
7.3 Derailments .............................................................................................. 36
7.3.1 Lac-Mégantic – UPDATE.............................................................................. 37
7.3.2 Freight derailments on GB rail .................................................................... 40
3
7.3.3 Freight train derailment at Gloucester, 15 October 2013 (pub. 10/14) ..... 41
7.3.4 Freight train derailment at Camden Road West Junction, 15 October 2013
(pub. 10/14) ......................................................................................................... 42
7.3.5 Freight Train Derailments Working Group ................................................. 43
7.4 Signals passed at danger .......................................................................... 45
7.4.1 Passenger train SPAD at Greenford, 20 March 2014 (pub. 12/14) ............. 46
7.4.2 Incident Factor Classification System ......................................................... 48
7.4.3 SPAD ten-year strategy ............................................................................... 49
7.4.4 Train Protection and Warning System ........................................................ 54
7.4.5 TPWS – ‘reset & continue’ .......................................................................... 54
8 Lessons learnt in 2014/15 – people on trains and in stations .....55
8.1 Statistical overview .................................................................................. 55
8.2 Platform-train interface ........................................................................... 56
8.2.1 Wheelchair rolls on to the track at Southend Central (28/08/13; pushchair
rolls on to the track at Whyteleafe (18/09/13) (pub. 08/14) .............................. 57
8.2.2 Passenger trapped in a train door and dragged a short distance at Newcastle
Central station, 5 June 2013 (pub. 09/14) ........................................................... 58
8.2.3 Ongoing investigations ............................................................................... 62
8.2.4 Platform-train interface strategy ................................................................ 63
9 Lessons learnt in 2014/15 – infrastructure workers ...................66
9.1 Statistical overview .................................................................................. 66
9.2 Equipment design and modification ........................................................ 67
9.2.1 Runaway of a road rail vehicle and resulting collision in Queen Street High
Level Tunnel, Glasgow, 21 April 2013 (pub. 07/14) ............................................. 67
9.3 Track working ........................................................................................... 69
9.3.1 Track worker fatality at Newark North gate, 22 January 2014 (pub. 02.15)69
9.4 Machine use in possessions ..................................................................... 73
9.4.1 Engineering train collision at Kitchen Hill, near Penrith, 12 January 2014
(BULLETIN, pub. 10/14) ........................................................................................ 73
9.5 Infrastructure Safety Liaison Group ......................................................... 74
10 Lessons learnt in 2014/15 – road driving risk ...........................77
10.1 Statistical overview ................................................................................ 77
11 Lessons learnt in 2014/15 – level crossings ..............................81
11.1 Statistical overview ................................................................................ 81
11.2 Level crossings........................................................................................ 82
11.2.1 Near miss at Llandovery level crossing, Carmarthenshire, 6 June 2013 (pub.
05/14)................................................................................................................... 82
4
11.2.2 Near miss at Butterswood level crossing, North Lincolnshire, 25 June 2013
(pub. 06/14) ......................................................................................................... 83
11.2.3 Fatal accident at Barratt’s Lane No.2 footpath crossing, Attenborough,
Nottingham, 26 October 2013 (pub. 08/14) ........................................................ 85
11.2.4 Collision at Jetty Avenue level crossing, 14 July 2013 (pub. 12/14) ......... 86
11.2.5 Other level crossing initiatives .................................................................. 89
11.3 Road vehicle incursions .......................................................................... 91
11.3.1 Road vehicle incursion at Aspatria, 26 October 2013 (pub. 06/14) ......... 93
12 Lessons learnt – beyond the boundary fence ...........................94
Glossary.........................................................................................95
5
Report summary
If you’ve read our Annual Safety Performance Report, you’ll know the GB rail industry’s
performance remains consistent with the Railway Safety Directive’s intent to maintain
safety and improve it where reasonably practicable. One of the main contributors to this
is the learning that flows from accidents and near misses.
Arguably, the current era of railway safety began after the Clapham Junction multiple
train collision of 1988, the inquiry into which led to changes in signal testing procedures
and working hours for safety critical staff. Even in the relatively short period since then,
however, we’ve seen a huge amount of organisational change, and a variety of different
regimes and techniques for capturing (or not capturing) appropriate knowledge and
learning.
Where does RSSB fit in? The short answer is that – as both a listening and a learning
organisation – we help the rail industry put things right when something goes wrong,
and help it remember the route to the solution.
Through our research programme, periodic safety reports, strategic risk papers,
publications on incidents inside and outside the railway, facilitation of the RED DVD
series, Right Track magazine, and the analysis and support we provide to stakeholder
groups, we learn; through the groups, Close Call, CIRAS and SMIS, we listen.
In addition, we play a part in the accident investigation process by providing relevant
training and guidance, and by providing statistics to help the Rail Accident Investigation
Branch (RAIB) set incidents into context, by offering expert knowledge from staff with
extensive industry experience, and by bringing cross-industry groups together to tackle
industry-wide issues.
Finally, this Learning from Operational Experience Annual Report (LOEAR) looks at some
of the tools available to facilitate learning, captures some of the lessons learnt in the
fiscal year and considers specific issues affecting rail users and employees.
Summary of key points arising during 2014/15
Train operations
The railway industry continues to monitor overseas accidents, the current focus on
which stemmed in part from the four major incidents that occurred in July 2013.
However, with the industry working smarter in its response to such incidents, two
freight derailments on Network Rail metals – at Gloucester and Camden Road West
7
Junction (both 15 October 2013) – raised questions about loading, track condition and
compliance with standards in this country. A cross-industry working group has been
convened to consider the situation.
Similarly, a ten-year SPAD strategy has been launched to manage a risk which we have
never overlooked, but of which we were reminded by the double SPAD at Greenford
(March 2014) and the potentially catastrophic incident at Wootton Bassett Junction
(March 2015).
People on trains and in stations
Since 2007, RSSB has provided the industry with regular updates on risk and safety
performance. In 2010, exposure of a rise in platform-train interface (PTI) risk prompted a
focus of resource and expertise on this important area.
A number of incidents on which RAIB reported in 2014/15 highlighted the ongoing need
for this focus, with specific reference to train dispatch procedures, the operation of train
doors and the slope of station platforms towards the track.
Infrastructure workers
Investigations continue to highlight the dangers of track working, inadequate safe
systems of working, and culture. However, RAIB activity also raised questions of
equipment modification, possession length and permitted speeds in worksites.
Road driving risk
Many GB rail companies employ large road fleets and there is an increasing
understanding of the risks involved and the range of people that can be affected. A fatal
accident in June 2014 revealed fatigue as a factor, but also demonstrated gaps in our
safety management systems. The industry is now working to fill these gaps via a bespoke
Road Driving Risk Group.
Level crossings
Level crossings are key interfaces between the public and the railway. The industry has a
duty to ensure that both the crossing signs and controls in use are fit for purpose and
easily understood by users. Investigations revealed issues around crossing controls,
closure sequences, and signage and crossing instructions during the reporting year. An
investigation into a road vehicle incursion also highlighted issues around highway
authority risk assessments and the guidance for producing them.
Investigations and recommendations
During 2014/15, RAIB published 24 reports, 20 of which involved incidents on the
mainline railway. The 20 led to 70 recommendations (compared to 88 recommendations
8
from 22 investigations in 2013/14). The area of infrastructure asset management
received the most focus.
CIRAS
CIRAS received 978 contacts on a diverse range of topics in 2014/15. Of these, 216 (22%)
became reports after the screening process. Positive results included amendments to
the training given to banksmen, train dispatch arrangements, road vehicle driving and
safety critical (radio) communications (inter alia).
Beyond the railway
The industry is also mindful of the need to look beyond its own operations for insights or
initiatives, and knows that the key to success is not only about sharing lessons, but also
good practice and ideas. This section of the report links to RSSB’s summaries of some of
the major non-rail accident public inquiries, which can also offer suggestions for how
your own learning procedures might be finessed.
1 Introduction
The rail industry learns from operational experience by investigating specific events and
through the regular monitoring of trends. RSSB is here to help with that process:

This Learning from Operational Experience Annual Report (LOEAR) contributes by
summarising some of the learning points arising from investigations and other
sources of information.

The Annual Safety Performance Report (ASPR) – the ‘sister’ publication to the
LOEAR – contributes by providing wide-ranging analyses of mainline rail safety
performance data.
1.1 Purpose
The main purpose of LOEAR is to provide learning information to decision makers in
RSSB member organisations. Like the ASPR, however, it is also intended to inform rail
employees, passengers, the government and the public at large.
Since 2009, RSSB has worked with industry groups to shape the definition and objectives
of Learning from Operational Experience (LOE) to meet industry requirements. This
report describes the LOE processes and their evolution, while retaining the
recommendations analysis function of previous documents.
9
1.2 Scope and structure
The LOEAR considers a range of learning sources – like CIRAS, RAIB and Right Track
magazine – and identifies the key issues that arose between 1 April 2014 and 31 March
2015 in the following areas:

Overseas accidents

Derailments

Signals passed at danger

The platform-train interface

Track working

Road vehicle driving

Level crossings
Hyperlinks (blue text) have
been used throughout this
document to aid access to
relevant documents and
websites.
Boxes have been provided to
highlight learning points and
extra information that readers
might like to consider.
The report also takes a look ‘beyond the
boundary fence’, Chapter 13 considering lessons
from non-railway events, like the Fukushima nuclear accident, which came in the wake
of the Great East Japan Earthquake of 2011.
2 How do we learn?
The railway and its regulatory bodies have been learning
lessons from accidents since William Huskisson MP was
struck and killed by Rocket at the opening of the Liverpool
& Manchester Railway in 1830.
Early incidents like this led to the first Railway Regulation
Act (1840), which required all injurious accidents to be
reported to the Board of Trade. Within 50 years, block
signalling, interlocking and continuous braking on
passenger trains had been made mandatory. The twentieth
century saw further advancements, ranging from continuous welded rails and multiaspect signalling, through to automatic train protection systems.
The cycle of safety planning and performance reporting has become essential to
ensuring that this development continues, but much learning also comes from
investigations into accidents that have occurred, near miss data, reports to the
industry’s Confidential Incident Reporting and Analysis System (CIRAS) and Close Call.
In this report, LOE is defined as the process by which knowledge from the operation of
systems is gained, exchanged and used, leading to continuous improvement and the
development of a positive safety culture.
10
LOE is discharged through the rail industry’s national stakeholder groups, all of which
were established by the RSSB Board.
During 2013/14, a new meeting structure was progressively introduced. A System Safety
Risk Group (SSRG) now reports to the Board and looks at safety risks across the industry,
while supporting the development of strategies to address them. SSRG is supported in
turn by four sub-groups:

Level Crossing Strategy Group (LCSG)

National Suicide Prevention Steering Group
(NSPSG)

People on Trains and in Stations Risk Group
(PTSRG)

Train Operations Risk Group (TORG)
In addition, the Data and Risk Strategy Group (DRSG) looks at the industry’s mechanisms
for capturing and processing risk information, while the Infrastructure Safety Liaison
Group (ISLG) considers the risks involving the contractor community.
An example of learning in action may be seen in the work the industry has done on
reducing the risk at the platform-train interface
(PTI).
RSSB uses data from the Safety Management
Information System (SMIS) to provide the industry
with regular updates on risk and safety
performance.
In 2010, exposure of a rise in PTI risk to the (then)
Operations Focus Group (OFG) prompted a
poignant dramatization on the subject in RED 28,
numerous articles in Right Track and led OFG to
develop a Station Safety Improvement
Programme.
RED 37
RED 37 – released October
2013 – builds on the
dramatization covered in RED
28.
As part of this process, RSSB held a Station Safety
See Opsweb for details.
Improvement Workshop in 2012 to provide an
update on operational risk management
initiatives, promote sharing, encourage good practice and obtain the views of front-line
staff on the issues associated with the management of operational risk at stations.
11
This information (inter alia) was combined into a PTI strategy, the development of which
was approved by the RSSB Board in September 2013. More information on the fruits of
that work may be found later in this report.
Cross-industry groups review the outputs from a number of RSSB activities, including:

Safety performance reporting – information on the latest trends, updated on a
regular basis

Operational Feedback – learning from rail and other industry accidents

CIRAS – the rail industry’s Confidential Incident Reporting and Analysis System

Human Factors – the study of environmental factors, organisational factors, job
factors, human characteristics and their impact on health and safety

Safety Management Systems programme

R&D – RSSB’s management of research and development on behalf of
government and the railway industry
3 What can help us learn?
Just as businesses have to work as one to make a profit, and football teams play as one
to win championships, the rail industry needs to think and act as a cohesive unit to
maintain acceptable levels of safety and performance efficiently. This section describes
five tools and methods which will help you benefit from the loop of learning described in
the previous section:

Measuring safety performance

Investigations (including the Incident Factor Classification System)

CIRAS

Right Track magazine

Close Call
3.1 Measuring safety performance
If you’re on a diet, you might measure how much exercise you do and check your
weight. The exercise is an activity – or an input to your diet – whilst your change in
weight is an outcome of dieting.
12
Transporting the same principle to safety management, the activities and resources we
devote to improving safety are just as important as the outcomes – whether accidents
occur or not.
Focusing on outcomes means that
responses only come after
something has gone wrong. If the
balance is shifted towards
measuring inputs to safety
management, accidents can be preempted and avoided.
During 2013/14, RSSB concluded
research project T953 (Enhancing
and Promoting the use of Safety
Performance Indicators) – the
culmination of four years’ work, in
which we sought best practice from
around the world and interpreted it for the GB railway environment. To ensure that it
was a good fit, we supported a number of stakeholders in trialling the resulting guidance
document, Measuring Safety Performance.
VolkerRail was one of the stakeholders we worked with in developing safety
performance indicators (SPIs) to support a new initiative: Right First Time (RFT).
VolkerRail is a rail contractor engaged in major infrastructure projects, often working in
restricted hours when trains are not running.
The aim of RFT is to avoid accidents and disruption by doing a job once, at the right time,
in the right place with the right people and the right equipment. At the heart of the idea
is planning the works, so that everything runs smoothly on site.
By following the process set out in Measuring Safety Performance, VolkerRail quickly
identified potential activity and outcome indicators. The outcome of ‘not achieving RFT’
was ‘problems during weekend works’. This was recorded using a simple tick list, which
was in turn used as an outcome indicator.
To achieve RFT, the team identified work and task planning as a key activity, which was
measured through a ‘readiness review’ (an existing process that was not consistently
employed). An activity indicator was therefore set around their use, namely a series of
questions about key milestones before going to site. A readiness review score for a
weekend’s work was based around the total score at final completion, taking into
account scores at earlier milestones.
13
During the pilot stages, a clear correlation was evident between readiness review scores
and RFT scores. The higher the readiness review score, the more likely RFT was
achieved.
A recent presentation to the project teams within VolkerRail covered three projects: ‘the
Good, the Bad and the Ugly’.
The Good followed the readiness review process and featured detailed set-up, delivery
and management of the work, resulting in high RFT scores.
The Bad started without any structure, briefing and roles and responsibilities etc.
Through a robust re-vamp and briefing session, the Project Manager turned the project
round and started applying the readiness review process. As a result, the RFT scores for
the project went from a 50-60% success rate to over 90% each weekend.
The Ugly project team arrived on site unable to deliver the work as key assets/engineers
were not available to do it. On reflection, it was said that if the project team had used
the readiness review process, they would have been able to make informed decisions
weeks before and achieved RFT.
The process is now being incorporated into a company procedure to complement the
new Project Managers’ System.
An output of T953 is a number of supporting resources that have been made available to
the industry in the Measuring Safety Performance toolkit. This includes further details
on the approach that VolkerRail took, written as a supporting case study.
For more information, contact kevin.thompson@rssb.co.uk or jay.heavisides@rssb.co.uk
3.2 Investigations
Though the cycle of safety planning and performance reporting is essential to ensuring
that safety continues to improve, much of the industry’s learning comes from
investigations into accidents and incidents, much as it always has done.
The principal investigation of any safety event is conducted by the party immediately
responsible for the activity. To facilitate this, railway companies have their own
arrangements for carrying out internal formal and local investigations, as defined in
Railway Group Standard GO/RT3119 (Accident and incident investigation) and its
associated Guidance Note, GO/GN3519. The outputs are managed by the companies
concerned, with actions being picked up by their own tracking systems. The results of
duty holder-led formal investigations are also summarised in SMIS to give others the
opportunity to learn from them.
The more significant accidents (involving loss of life or potentially significant
consequences) are investigated by the Office of Rail and Road (ORR) as safety authority,
and the Rail Accident Investigation Branch (RAIB). The latter was set up after a
14
recommendation made by Lord Cullen’s inquiry into the Ladbroke Grove accident of
1999 (although a subsequent European Directive on rail accident investigation also
required Member States to create such bodies). RAIB was fully established in 2005, after
which RSSB ceased its own (interim) accident investigation role.
If an accident involves a derailment or collision which results in, or could result in, the
death of at least one person, serious injury to five or more people or extensive damage
to rolling stock, the infrastructure or the environment, then RAIB will lead an
investigation, draw conclusions and make recommendations.1
RAIB investigates incidents on UK railway infrastructure without apportioning blame or
liability. It is independent of the rail industry and the ORR, with the Chief Inspector of
Rail Accidents reporting directly to the Secretary of State for Transport. RAIB’s
recommendations on the rail industry are addressed to the ORR2, which must then
ensure that they are considered and that, where appropriate, action is taken.
3.2.1 Accident investigation assistance
RSSB can help companies with the investigation process and the monitoring that follows
on from it. Specifically, guidance and training is available on conducting the
investigations themselves, safety assurance, and dealing with the EU’s Common Safety
Method (CSM) for Monitoring.
Accident Investigation Guidance and Training (AIT) Programme
In 2011, RSSB produced a three-part guidance on accident investigation:

Part 1: The role of the senior manager;

Part 2: Development of policy and management arrangements; and

Part 3: Practical guidance.
Each was updated in 2014, being supported by material to help companies decide on
the proportionality of their response to an accident. Also included is a comprehensive
training programme, aimed at those undertaking all but the more serious accidents, in a
bid to raise the quality and consistency of investigations – and investigation reports –
across the industry.
1
RAIB may also investigate other incidents that have implications for railway safety,
including those which, under slightly different circumstances, may have resulted in an
accident.
2
RAIB can also address recommendations to other safety authorities and other public
bodies, such as the police, the Department for Transport and so on.
15
The course material can be delivered in a classroom environment or via a laptop. It
consists of 10 modules and takes around six hours to complete. Several companies have
already used it to train investigators or line managers who may undertake occasional
investigations. For more information, click here.
Safety Assurance and the CSM for Monitoring
Since June 2013, the CSM for Monitoring has required transport operators to include
problems identified via their monitoring strategies and plans3 – and the actions put in
place to address them – in their annual safety reports to the ORR.
As accident investigations are an important source of monitored information, RSSB has
published safety assurance guidance to help companies meet these requirements.
However, the guidance also emphasises the importance of learning lessons, including
from accidents, and the need to integrate the learning into safety management systems.
For more information, click here.
RSSB has also published a set of six guidance notes (GE/GN8640-8645) which give
practitioner level guidance on the application of the risk management process set out in
the Common Safety Method on Risk Evaluation and Assessment (CSM RA). The CSM RA
has applied since July 2012 to the implementation of all significant changes to the
railway system – ‘technical’ (engineering), operational and organisational – and the
latest revised version applies from May 2015. The notes may be found via
http://www.rssb.co.uk/railway-group-standards
Modernisation of Safety Cooperation
Over the last two years, RSSB has facilitated a restructuring of the main safety groups
across the industry. Significant progress has been made and this now allows an
improved systematic review of industry risks. The new hierarchical structure is intended
to link the industry’s management of risk across sectors, functions and geographical
bounds more effectively.
The remits of these new groups encourages the right information to go to the right
representatives to facilitate the best decisions on management of risk at the interfaces.
A key part of this information is the output from investigations.
3.2.2 Incident Factor Classification System
As mandated under GO/RT3119, RSSB receives investigation reports from all GB railway
organisations. Currently, around 4,500 are stored, dating back to the late 1990s. The
conclusions therein carry a great deal of valuable information about event causes.
3
That is, their safety assurance processes.
16
The trouble is, once their recommendations have been acted upon, there’s a danger
that some of their learning points will be lost or will not reach other parties who could
benefit from them.
What to do?
RAIB’s investigation into the Shap rollback incident of 2010 demonstrates why a new
way of capturing how accidents and incidents happen had to be devised.
In short, Shap was caused by driver fatigue, but it was not mandatory to flag this factor
up in SMIS. This meant that there was little accurate, tangible evidence available to
show the magnitude of the issue.
Clearly, some way of capturing this sort of information was going to be vital if we, as an
industry, were to get to real grips with common themes below the root causes.
RAIB recommended that RSSB improve rail industry information on fatigue-related
accidents and incidents, although the project to address this – and more – was already
under way.
What we did
In 2009, we developed a means to analyse accident reports through an Incident Causal
Classification System (ICCS), using a taxonomy developed by RAIB, to help us understand
(inter alia) what makes operators do certain things at certain times and what makes
certain equipment fail under certain conditions.
Previous editions of the LOEAR featured analysis using the ICCS. More recently,
however, RSSB and Network Rail have worked together with industry to combine the
ICCS, human error and violation taxonomies, and the Network Rail ‘10 incident factors’
within SMIS.
Software for a single module – the Incident Factor Classification System (IFCS) – was
commissioned in November 2012.4 Between January 2013 and March 2014, it was
populated with data by a team of specialists at RSSB to enable:

Cross-industry learning: Causal trends are being identified for all in the industry
using a consistent classification for key incidents. The information within the IFCS is
heavily used by RSSB and Network Rail learning from operational experience
functions in their central reporting and analysis.
4
As part of research project T994: Development of an incident factor classification
system module for SMIS
17

Incident investigation: The IFCS is being used by incident investigators to identify
past incidents with similar causes, which not only aids analysis, but also helps
prevent previous recommendations being duplicated or contradicted.
In time, the IFCS module will contain causal classifications for all RAIB reports and
Formal Investigations. The risk-based sample also includes causal classifications for some
local level investigations (such as for high-risk irregular working incidents). The IFCS
module will also include non-GB rail and non-rail investigation reports, in order to
ascertain how other industries learn from safety events that may have parallels with our
own.
In 2014, IFCS data was used to produce a report for rail industry managers on fatigue’s
contribution to a sample of high-risk incidents (collisions, derailments and so on). In all,
246 (covering 2011-13) were analysed. Fatigue was identified as a factor in 21% of them.
Furthermore, 40% of the incidents where fatigue played a part included fatigue from
home life-related issues and their management; 38% included work-related fatigue,
which covers factors such as shift pattern design and hours of work. Click here to access
the full report.
The IFCS has also been used to start informing an Industry Human Factors SPAD Review.
For the initial outputs, see section 8.4.2.
3.3 CIRAS
As the focus of Measuring Safety Performance suggests,
learning does not only occur after an event; many valuable
lessons are revealed by what might be termed ‘accidents
waiting to happen’.
Reports to the industry’s Confidential Incident Reporting and
Analysis System (CIRAS) focus mainly on such ‘near miss’
events or perceived deficiencies in safety systems and
arrangements, a better understanding of which provides a
solid foundation for shared learning across different industry
sectors.
Capturing this knowledge, which comes from workforce members who have daily
operational contact with the railway, makes it possible to identify issues before they
cause injury.
Maintaining confidentiality is a key aspect of CIRAS, but while we recognise that this
could restrict the information that can be disclosed, it allows reporters to state their real
concerns – and describe underlying causes – more openly than they might to their line
manager. This gives CIRAS the potential to provide unique insights into safety issues.
Lessons learnt via CIRAS during 2014/15 may be found in Chapter 6.
18
3.4 Right Track
In 2012, RSSB launched Right Track, a quarterly magazine aimed
at front-line personnel to capture, share and promote safety
learning and initiatives in a down-to-earth way.
This year, the magazine has focused on station safety, suicide
prevention, SPADs, track worker safety, train dispatch and
autumn adhesion, among many other subjects relevant to
drivers, guards, on-train staff, station staff, dispatchers,
signallers, shunters, depot workers and track workers.
Right Track is available to all RSSB members, London
Underground, and other companies and bodies who have a role in supporting railway
operations.
It is available as a pdf and in paper form. Hard copies are distributed in bulk by
arrangement with individual companies.
A summary of the main themes covered by Right Track during 2014/15 may be found in
Chapter 7.
3.5 SMIS+ and Close Call
At the request of Network Rail and its contractors, RSSB has developed a new internetbased Close Call System (CCS), which allows the industry to record and analyse ‘close call
incidents’ centrally.5
A ‘close call’ is defined as ‘an event that had the potential to cause injury or damage’,
like leaving a cable troughing cover where someone might trip over it, but not a near
miss with a train or on-track plant, both of which will continue to be reported into SMIS.
In 2015, a programme was launched by RSSB to expand SMIS to include the CCS (inter
alia). The overall ‘SMIS+’ plan will provide new functionality resulting in the
implementation of a national Safety Management Enterprise System. This will help the
industry manage its safety management system (SMS) responsibilities, give it more
control when making changes, and cut costs.
The first phase of implementation is due to begin in October 2016.
5
This was managed as research project T1015: Revision of the close call system.
19
4 Investigations – lessons learnt in 2014/15
RAIB published 24 reports between 6 April 2014 and 5 April 2015, covering the following
categories:

Heavy rail – on Network Rail managed infrastructure (NRMI) (20)

Heritage (1)

Northern Ireland Railways (1)

Metro (2)
Table 1 lists each of these investigation report. Note that:

70 recommendations were issued from 20 RAIB investigations involving incidents
on NRMI. This compares to:

88 recommendations from 22 RAIB investigations in 2013/14

82 recommendations from 19 RAIB investigations in 2012/13

90 recommendations from 23 RAIB investigations in 2011/12

76 recommendations from 15 RAIB investigations in 2010

167 recommendations from 27 RAIB investigation reports in 2009

127 recommendations from 18 RAIB investigation reports in 2008

158 recommendations from 22 RAIB investigation reports in 2007
RAIB investigations published 2014/15
Publication
Date
Report Title
Infrastructure
Owner
REPORTS PUBLISHED IN 2015
30/04/2015
Passenger train derailment at Paddington
station, 25 May 2014
NRMI
02/04/2015
Freight train derailment at Stoke Lane level
crossing, 27 August 2013
NRMI
16/02/2015
Track worker fatality at Newark North Gate, 22
January 2014
NRMI
REPORTS PUBLISHED in 2014
22/12/2014
Double SPAD ay Greenford, 20 March 2014
NRMI
15/12/2014
Collision at Jetty Avenue level crossing, 14 July
2013
NRMI
20
11/12/2014
Derailment at Liverpool Street station, 23
January 2013
NRMI
24/11/2014
Buffer stop collision at Chester station, 20
November 2013
NRMI
20/11/2014
Passenger train collision with trolley at
Bridgeway user-worked crossing, 16 January
2014
NRMI
13/11/2014
Rail breaks on the East Coast Main Line
NRMI
23/10/2014
Dangerous occurrence at Denmark Hill station,
1 August 2013
NRMI
23/10/2014
Passenger trapped and dragged by train at
Holborn station, 3 February 2014
Metro
14/10/2014
Freight train derailment at Camden Road West
Junction, 15 October 2013
NRMI
09/10/2014
Freight train derailment at Gloucester, 15
October 2013
NRMI
18/09/2014
Person trapped and dragged by train at
Newcastle Central, 5 June 2013
NRMI
21/08/2014
Fatal accident at Barratt’s Lane No.2 footpath
crossing, 26 October 2013
NRMI
14/08/2014
Falls from platform at Southend and
Whyteleafe
NRMI
28/07/2014
Uncontrolled evacuation of a passenger train
at Holland Park, 25 August 2013
Metro
17/07/2014
RRV runaway and collision at Glasgow Queen
Street, 21 April 2013
NRMI
26/06/2014
Road vehicle incursion at Aspatria, 26 October
2013
NRMI
16/06/2014
Steam locomotive failure near Winchfield, 23
November 2013
Heritage
16/06/2014
Near miss at Butterswood level crossing, 25
June 2013
NRMI
21
15/05/2014
Near miss at Llandovery level crossing, 6 June
2013
NRMI
12/05/2014
Accident at Balnamore level crossing, 31 May
2013
Northern Ireland
01/05/2014
Passenger train collision at Norwich, 21 July
2013
NRMI
Key:
Off NRMI
Source: RAIB website
4.1 RAIB Bulletins
When RAIB’s preliminary examination of an incident suggests that a full investigation
would not lead to further significant safety lessons for the rail industry, in some cases it
provides related information or advice in the form of a short bulletin.
Between April 2013 and March 2014, RAIB issued one such bulletin – on the engineering
train collision at Kitchen Hill, near Penrith, which occurred on 12 January 2014. This is
discussed in the Infrastructure worker section of this report.
4.2 RSSB analysis of key RAIB recommendation themes
Recommendations tend to reflect the nature of the incident from which they arise, but
the selection of incidents and the number of recommendations also indicate the
weighting given to the event by the investigating organisation. In other words, only the
tip of the accident/incident/unsafe act or condition ‘pyramid’ is represented by looking
at recommendations in detail.
It should be noted, therefore, that numeric analysis of recommendation trends has little
statistical validity. Indeed, a single report may generate multiple recommendations for
one category. In the interests of continuity, however, we have used the categorisation
process applied in previous years to RAIB recommendations. The results are presented
overleaf.
22
SMIS recommendation categories
A
Cat code
Recs category
Signalling system
B
Competence management
C
Rules, standards and instructions
D
Vehicle operation and integrity
E
Infrastructure asset management
F
Event mgmt/investigation/ reporting
G
Monitoring and audit
H
J
Research and development
Safety communications
K
Culture
Description
Lineside SPAD controls, signal sighting issues, train
planning and regulation, operation of the signalling
equipment.
Training and development, driver management,
competence systems, briefing, assessment, staff
selection procedures, drugs and alcohol, fitness for
duty, fatigue.
Modification /development of rules and predefined
standards for operation, standards/process change
management.
Train-borne safety equipment, fire protection, vehicle
maintenance, train data recorders, crashworthiness, incab ergonomics.
Managing contractors, track/signalling maintenance
operations, work planning, technical specifications,
method statements.
SPAD management, public accident investigation, site
investigations, post-accident management, formal
investigations, formal inquiries, public inquiries, fault
reporting, emergency procedures.
Monitoring activities, safety performance monitoring,
follow-up processes.
Suggested research topics/specific areas of research.
Defining and communicating safety responsibilities,
general safety related communications, meetings,
techniques, methods and equipment.
Management commitment, organisational change.
23
Chart 1.
Distribution of RAIB recommendations (%) – 2014/15
Research &
development
5%
Event
mgmt/investigation/
reporting
4%
Culture
2%
Signalling system
6%
Monitoring &
audit
15%
Competence
management
6%
Rules, standards
& instructions
20%
Infrastructure asset
managment
32%
Vehicle operation &
integrity
10%
RSSB’s figures suggest that, in 2014/15, the largest recommendations component was
infrastructure asset management (32%). This is in line with three of the last four
reporting periods.
Chart 2.
24
Recommendation categorisation – by year (%)
Comparing 2014/15 with 2013/14, reductions in the percentage of recommendations
can be seen for:
The ORR also keeps a

Signalling system
record of the status of all

Competence management
RAIB recommendations.

Infrastructure asset management
This is available on its

Event management/investigation/reporting
website.

Safety communications
However, there has been a rise in the percentage of recommendations which deal with:

Rules, standards and instructions

Vehicle operation and integrity

Monitoring and audit

Research and development

Culture
Network Rail’s biggest asset is obviously the track. Some of the issues in this area were
causal to the derailments discussed in the chapter on train operations. However, in
November 2014, RAIB published a report on three rail breaks that had occurred on the
East Coast Main Line (ECML) during 2012 and 2013 (none of which resulted in injuries or
damage to trains.)
A break at Corby Glen, near Grantham, was triggered by wear of the pad intended to
separate the rail from the underlying concrete sleeper. Breaks at Copmanthorpe and
Hambleton were due to movement at rail joints caused by inadequate support from the
underlying ground.
Statistics showed that, after allowing for differences in route length and the amount of
traffic, the ECML has more rail breaks than comparable main lines. After considering
both the types of break occurring on the ECML, and the measures being taken by
Network Rail to manage them, RAIB concluded that the most significant factor in the
relatively high number of incidents on the ECML between 2009 and 2013 was the
relatively high proportion of older track.
Network Rail is now replacing older track components on this line. It has also altered the
maintenance criteria on the ECML to increase the likelihood of replacing moving
(dipped) joints before they cause rail breaks. These measures appear to be reflected in a
recent reduction in the number of incidents.
RAIB also recommended:
25

Undertaking or commissioning research to identify opportunities for reducing the
size of cracks and defects which can be identified in rails in circumstances likely to
be associated with rail breaks.

Reviewing actions already being taken to reduce the number of rail breaks on the
ECML, in order to identify whether similar actions would provide significant safety
benefits elsewhere on its infrastructure.

Establishing a process for inspecting parts of rail pads beneath rails (on a sample
basis) and, if necessary, replacing rail pads outside rail replacement projects in
areas where this is justified.

Completing a test programme to establish the practicability of extending current
testing and analysis to identify defects throughout the full depth of a rail and/or
defects on the underside of a rail.

Modifying existing document preparation processes to ensure that markings
intended to show changes to standards and other safety critical documents clearly
indicate the change that has occurred.
5 CIRAS – lessons learnt in 2014/15
CIRAS has continued to implement its five-year strategic plan. Supported by clear
objectives, its governance was also strengthened, and its funding arrangements
restructured.
Furthermore, CIRAS membership is growing to include other UK transport modes, key
relationships being established in the bus, coach, and road sectors during 2014/15. This
expansion is supported by substantial IT improvements, both the CIRAS website and
database having been redesigned from the ground up to serve the needs of the wider
community more effectively.
With the right infrastructure in place to share learning across different industry sectors,
CIRAS aims to cement its role as a key driver of proactive transport safety culture.
26
5.1 Who reports to CIRAS?
Chart 3 shows the distribution of reports received in 2014/15, by reporter occupation.
Chart 3.
Report distribution 2014/15 – by job category

Train drivers submitted the largest number of reports – a trend consistent with
previous years.

Though more station staff reported this year than last, fewer track workers
contacted CIRAS during 2014/15.

It is, however, particularly positive to note that a diverse range of staff feel able to
raise safety issues, including conductors, station staff, managers, signallers, as well
as train crew.
5.2 Why do people report to CIRAS?
Of all issues reported to CIRAS in 2014/15, 27% came because reporters found it difficult
to raise concerns with their managers. However, most (73%) had previously been taken
through internal company channels; this figure remains consistently high from year to
year. Chart 4 provides a picture of why those within the 73% then went on to contact
CIRAS.
27
Chart 4.
Reporters’ views of company response after issue raised internally
Other
1%
No Response
15%
Adequate (but
not
implemented)
27%
Inadequate
57%
57% of reporters used CIRAS as they felt the response they had received from their
company was inadequate:

27% believed the company response to be adequate, but they had not seen any
changes or implementation at work.

13% claimed they had received no response whatsoever.
Note that concerns that get as far as CIRAS are likely to represent a small proportion of
all the issues that are pursued through a company’s internal processes.
5.3 Key issues of concern in CIRAS reports during 2014/15
In 2014/15, CIRAS received 978 contacts on a diverse range of topics. Of these, 216
(22%) became reports after the screening process.
A breakdown of reports per industry risk category is shown in Chart 5.
Reporters predominantly focussed on the potential for accidents or incidents. The
reports therefore represent the perceived risks identified by reporters in the course of
carrying out their duties.
28
Chart 5.
CIRAS reports against industry risk categories
Public safety Level crossings
5%
2%
Train accidents
17%
Passenger safety
20%
Workforce
safety
56%
5.4 Positive outcomes from CIRAS reports
CIRAS continuously monitors the outcomes from the reporting process to ascertain the
value and benefits delivered to the industry. Positive outcomes from CIRAS reports, such
as an investigation leading to action, briefing, review or change, are recorded. Here are
some of the best examples of a CIRAS report ‘nudging’ management behaviour in
2014/15:

Training: Two large infrastructure contractors investigated a valid CIRAS concern
about the training of new banksmen, whose safety behaviour was reported to be
below standard. The direct result of the report was that 17 staff were given more
training and a more effective assessment regime was put in place.
The contractors’ proactive response made sure not only that the problem was
tackled, but that the issue behind it was dealt with too.

Maintenance: CIRAS also makes a difference to infrastructure maintenance, as
these two examples demonstrate:

A signal was being obscured by overgrown vegetation, raising the potential
signal passed at danger (SPAD) risk at one location near Edinburgh. The CIRAS
report prompted swift, remedial action by Network Rail.

A long-standing Automatic Warning System (AWS) fault at a signal was
repaired soon after it was reported to CIRAS. However, Network Rail was not
29
content to leave it there, going on to amend its procedures to help it respond
more quickly to faults of this nature in future.

Train dispatch arrangements: When train dispatch arrangements were not being
followed closely enough at a station, the train operator undertook a thorough
review of competence monitoring and management before committing to a robust
action plan.
Another benefit of this report is that it is clearly encouraging further improvement
– in this case to the supporting infrastructure. The installation of a ‘Close Door’
indicator to supplement the existing ‘Right Away’ indicator is now being
progressed.

Equipment: A conscientious CIRAS reporter flagged up the inappropriate use of
concrete breaking equipment by 360 machines. The equipment had been modified
in a potentially unsafe manner, and then attached to the machines for piling work.
The report led to work on a site in Lincolnshire being stopped by contractors before
corrective action was taken.

Road risk: Every year, CIRAS receives reports on the risks involved with driving long
distances to and from work sites. Two infrastructure workers died last year in a
road traffic accident whilst on duty. Network Rail’s guidance is that, if at all
possible, the shift and travel time in both directions should not exceed 14 hours.
One contractor was grateful to receive one of our reports on driver fatigue and
long travel times. Its own investigation had revealed some exceedances and
concluded there was a need for more effective monitoring. More staff briefings on
this important issue were scheduled, with corrective action being taken.
In the long-term, IT investment for the more effective monitoring of travel times
and working hours is now on the contractor’s agenda. As was the case here, the
operational feedback from CIRAS reporters often plays a critical role in the reassessment of safety risk, and the drive for business improvement.

Health and well-being: A number of issues potentially affecting health and
wellbeing were highlighted in a CIRAS report about the train crew accommodation
at Bristol Temple Meads.
The responsibility for the various issues was subsequently established at a meeting
between the Network Rail, who manage the station, and First Great Western. A
total of nine faults were addressed to improve the quality of the accommodation,
proving that CIRAS can effectively assist in the resolution of issues away from the
running line.
In another case, a complete review of welfare site facilities was carried out by a
contractor after a CIRAS report about a particular location. This led to
recommendations for procedural improvements, extending far beyond the scope
of the original report.
30

Communications: A CIRAS report on safety critical radio communications at the
port of Felixstowe led to a full investigation. There was only one channel available
for each terminal, and the potential for communications to become confused, or
misunderstood, was high.
After the two freight operators using the port, and the port authority itself, agreed
to add an additional radio channel, the safety risk was effectively eliminated. This
is a good example of a CIRAS report triggering an investigation involving the
collaboration of several different parties, and a robust action plan to tackle an
obvious risk.
5.5 Summary
CIRAS continues to evolve, and continues to aid the industry’s safety efforts by providing
another line of defence for subscribing organisations. There’s a growing recognition
that, whilst internal channels of reporting should be used first, people are different, and
some may feel unable to use them. Mature organisations value all good intelligence,
whatever the source, and accept that confidential reporting can provide valuable
feedback and input to safety management systems.
6 Right Track – lessons taught in 2014/15
Starting with Issue 9, three issues of Right Track were published during 2014/15. All may
be found on Opsweb.
Issue 9: ‘coupling up’, published May 2014
The cover feature dealt with the challenge of shunting in the busy
unit train environment of Selhurst Depot. The issue also considered
passenger behaviour, and what one operator has done to improve
‘etiquette’.
The regular RAIB report summary focused on the yellow plant
collision at Arley in August 2013, while an extended SPADtalk feature
described what actually happens after a SPAD and how the industry
learns from events of this type.
31
Issue 10: ‘not playing trains’, published August 2014
Following on from the ‘Selhurst’ article in Issue 9, Issue 10 considered
shunting in the more traditional – freight – environment of
Westbury. Meanwhile, the Mid Hants Railway’s regime for dealing
with staff training and competence showed that the ‘big railway’
doesn’t necessarily have all the answers. In the run-up to the
autumn, ‘wet rail syndrome’ and an on-train gel applicator were also
described.
The work being done to tackle Platform-Train Interface (PTI) risk was
aired, while the driving task was covered in a summary of RAIB’s
report on the passenger train collision at Norwich on 21 July 2013.
Issue 11: ‘stepping up’, published February 2015
The final issue of the reporting period provided an update on RSSB’s PTI strategy, rail
industry road driving risk and the level crossing collision at Valhalla,
New York, which occurred on 3 February 2015.
Network Rail presented the results of its ‘deep dive’ into station risk,
while cross-industry research into mobile devices revealed dangers
beyond the cab environment.
Questions of competence management, safety culture and the role
the Controller of Site Safety (COSS) were considered in the summary
of RAIB’s report on the collision between a passenger train and a
trolley at Bridgeway user-worked crossing on 16 January 2014.
7 Lessons learnt in 2014/15 – train operations risk
The railway industry continues to monitor overseas accidents, the current focus on
which stemmed in part from the four major incidents that occurred in July 2013.
However, with the industry working smarter in its response to such incidents, two
freight derailments on Network Rail metals – at Gloucester and Camden Road West
Junction (both 15 October 2013) – raised questions about loading, track condition and
compliance with standards. A cross-industry working group has been convened to
consider the situation.
Similarly, a ten-year SPAD strategy has been launched to manage a risk which we have
never overlooked, but of which we were reminded by the double SPAD at Greenford
(March 2014) and the potentially catastrophic incident at Wootton Bassett Junction
(March 2015).
32
7.1 Statistical overview
Version 8.1 of the Safety Risk Model (SRMv8.1) shows the risk from train accidents to be
8.0 fatalities and weighted injuries (FWI)6 per year. This is 5.8% of the total accidental
risk profile (which includes accidents in possessions and in yards, depots and sidings).
In the last ten years, only one train accident has led to an on-board fatality: the
derailment at Grayrigg in February 2007, in which one passenger was killed.
Train accident risk at a glance
10
10.0
9.6
9.2
Level crossings
Not level crossings
8.5
7.4
8
8.2
7.4
7.9
7.6
6.7
6
4
2
2014/15
2013/14
2012/13
2011/12
2010/11
2009/10
2008/09
2007/08
0
2006/07
Train
accidents
(8.0 FWI; 6%)
12
2005/06
Other
accidental risk
(131.6 FWI;
94%)
Trends in PIM indicator
PIM modelled risk (FWI per year)
Risk in context (SRMv8.1)
For more statistical analysis on train accidents, see Chapter 7 of the ASPR.
Topics covered in this section:

Overseas accidents

Derailments

Signals passed at danger
7.2 Overseas accidents
July 2013 saw four major train accidents across the world: in Canada on the 6th (47
fatalities), France (12th, 6 fatalities), Spain (25th, 79 fatalities) and Switzerland (29th, 1
fatality).
6
Fatalities, injuries and shock and trauma are combined into a single figure, termed
fatalities and weighted injuries (FWI). For more details on the injury classifications and
their associated weightings, see the Annual Safety Performance Report.
33
The concentration of so many major incidents gave the industry pause and led to a
paper to the RSSB Board, which was also put on Opsweb and considered in last year’s
LOE Annual Report.
An expanded consideration of the Canadian incident may be found in Section 8.3.1, but
details of the other three are:
Bretigny-sur-Orge, France, 12 July 2013
This incident involved a passenger train derailing at high speed on a broken fishplate
that got jammed in a crossing. The unit divided and came to rest under the station
canopy, killing six and injuring 62.
In Britain, fishplates are reportedly used at around 20 sites on 100mph lines to secure
track to switches and crossovers. They are effective if well maintained and allow the rail
to expand and contract under prevalent temperature conditions.
Fishplate breaks increased by 30% between 2008 and 2011. Though the trend reversed
in 2012-13, Network Rail took a proactive step and reviewed bolted crossing risk,
briefing its inspection teams to highlight the issues that came out of the Paris accident.
Santiago de Compostela, 24 July 2013, Spain
An over-speeding passenger train derailed on a curve at high-speed, killing 79 and
injuring 94 passengers.
The driver was later charged with ‘79 counts of homicide and numerous offences of
bodily harm committed through professional recklessness’.
It transpired that he’d been distracted by a telephone conversation with the guard, who
was seeking to confirm the arrangements for stopping at a later station so that a group
of passengers could alight. This caused him to miss the braking point for the curve.
On 4 June 2014, the Spanish Transport Ministry published the final investigation report
into the accident, which focused on driver error as the sole cause.
Despite this, Spain’s National Investigating Body issued nine recommendations across
the entire rail sector.
Two were placed on the infrastructure manager, inviting it to install conventional fixed
signals to indicate the maximum permitted speed, and balises to control speeds using
traditional signalling at sharp curve locations. (Both were absent on the curve where the
accident occurred.)
Train operator Renfe received two recommendations, one of which called for better
coordination of its internal communications procedures. This relates to an alert raised
by a driver two weeks before the official opening of the line. The driver had expressed
concerns about the lack of conventional speed markers before the curve in question, but
his comments were not shared with those responsible for infrastructure safety.
34
Regarding the signalling on the line, the wayside block system along the entire route is
based on Level 1 of the European Train Control System (ETCS), while the final section
near Santiago, is fitted with ASFA, which relies on a series of beacons to communicate
with the cab, giving audio and visual warnings if speed limits are surpassed and applying
the brakes if there’s no response from the driver.
This means that the accident occurred in the transition from the speed-supervised
system to conventional signalling. In order to avoid similar situations on GB rail, the
Train Control Technical Sub Group of the Train Control and Communications Systems
Interface Committee has sponsored research into the transition arrangements between
the European Rail Traffic Management System (ERTMS) and the Train Protection and
Warning System (TPWS).
Granges-pres-Marnand, 29 July 2013, Switzerland
Two passenger trains collided head-on after a dispatch error, killing one driver and
injuring 35 passengers.
The accident was caused by the driver of a departing train self-dispatching and ‘notching
up’ without waiting for an express to pull into an adjacent loop. The exit signal is a
‘common signal’, referring to several tracks, and is sited beyond the conflict point. While
at danger in this case, the train hadn’t reached it prior to the collision, and in fact had
built up some speed.
In Britain, there are a number of these ‘common signals’ located in certain yards up and
down the country, and in June 2014 there was an incident in which a position light signal
was cleared for a stock move. However, the driver of a light locomotive on an adjacent
siding, to which the signal also applied, thought it was for him, and he set off. He was
asked to stop via GSM-R. As the locomotive was actually blocking the stock move in,
arguably there couldn’t have been a collision. But had the circumstances been slightly
different…
In summary
To recap, overseas incidents can provide us with vital food for thought. With a railway as
ostensibly safe as ours, there’s a very real danger of complacency, so measuring our own
controls against accidents from abroad means we constantly keep a check on any gaps
in our processes, rules and methods.
RSSB can help monitor the situation overseas. Our regular papers covering recent
incidents and investigation reports are sent to the PTSRG, ISLG, the National Freight
Safety Group, the Dangerous Goods Group and OFG’s successor: TORG – and the paper
we send to TORG gets sent out by them to the Operational Risk & Mitigation groups
(OPSRAMs) throughout the country.
35
Now we’re doing more, by working a bit smarter. While these papers continue to
include information on each incident, we’ve also started asking Group members to
decide whether an industry response – and further analysis – is required. These
responses will be tracked via a database on SPARK, which will create – in effect – an
audit trail for action taken.
International accidents provide a useful reminder that such things are always possible,
no matter how mature and technologically developed we are. They’re a useful source of
information and cross-learning, though we should note that we also need those less
severe local incidents to help expand our analysis of risk. Which is where Close Call and
CIRAS data comes in…
RSSB’s overseas accident summaries may be found on Opsweb.
7.3 Derailments
Derailments – specifically freight train derailments –
continued to receive much attention in 2014/15. The
current focus stemmed in part from the runaway,
derailment and subsequent explosion of an oil train in
Lac-Mégantic, Quebec, on 6 July 2013.
However, as the RAIB reports into Gloucester and
Camden Road West Junction (both 15 October 2013)
reminded us, freight train derailments can happen
here too.
Particular issues have included uneven loading, and
instances where track and wagon are compliant, but
at the outer limits of what’s considered acceptable.
There is – arguably – an intersection of these outer
limits where a derailment becomes more likely.
There have been no reported injuries resulting from
GB freight train derailments in recent years, but to
quote the Baker Panel inquiry into the BP Texas City
oil refinery accident of 2005:7
7
Thirsk derailment, 1967
On 31 July 1967, a passenger
train struck a wagon that had
derailed foul of the adjacent line
near Thirsk. Seven people were
killed and 45 were injured.
The problem stemmed from a
wagon ‘hunting’ in motion, which
led a coupling to break.
In general, the lateral wheel
oscillations characteristic of
‘hunting’ had been exacerbated
by new, rigid wagons running on
new, rigid continuous welded rail.
The speeds possible with diesel
traction added to the problem,
which had not been foreseen and
required specialist research to
solve.
On 23 March 2005, the BP Texas City Oil Refinery, a ‘process accident’ resulted in an
explosion which caused 15 deaths and injured more than 170 other people.
36
‘The passing of time without a process accident is not necessarily an indication that all is
well and may contribute to a dangerous and growing sense of complacency.’8
With this ever in mind, work is being done by the industry (see section 8.3.5), taking
risk-based action before we have another ‘Thirsk’ (see panel).
7.3.1 Lac-Mégantic – UPDATE
In last year’s report, we featured the Lac-Mégantic runaway, derailment and explosion
of July 2013, and though nothing on this scale has been seen since, dangerous goods
accidents have continued to occur across North America.
On 30 April 2014, for example, defective track saw a train carrying crude oil derail and
burst into flames in Lynchburg, Virginia. There were no reported injuries, but some 300
people were evacuated from nearby buildings.
8
Baker Panel, The Report of the BP Refineries Independent Safety Review Panel (BP,
2007), p. i.
37
In August, the Transportation Safety Board of Canada (TBSC) published its final report on
Lac-Mégantic, which identified 18 distinct causes and contributing factors, as
summarized on the following diagram:
The report contains 16 specific findings as to risk. Although these did not lead directly to
the accident, they are related to unsafe acts, unsafe conditions, or safety issues with the
potential to degrade rail safety. Some of the risks that need to be addressed are:

The continuing risk from leaving trains unattended.

The risk from implementing single-person train operations.

The risk from not systematically testing petroleum crude oil.

The risk from not planning and analysing routes on which dangerous goods are
carried.

The risk from not having emergency response assistance plans in place.

The risk from Transport Canada not ensuring that safety management systems
work effectively.
The TSBC conclude that the accident ‘was not caused by one single person, action or
organization. Many factors played a role, and addressing the safety issues will take a
38
concerted effort from regulators, railways, shippers, tank wagon manufacturers, and
refiners in Canada and the United States’.
Although its investigation is complete, the Board says it will continue to monitor the five
recommendations, and report publicly on any progress – or lack of progress – until all
the safety deficiencies have been corrected.
To read RSSB’s full summary of the report, click here.
For the TSBC’s full report, click here.
What was done?
Since Lac-Mégantic, a number of initiatives have been rolled out, including a move to
harmonise Canadian and US rules on how dangerous goods are identified. This will
provide cross-border consistency, and will give emergency teams a clear understanding
of the risks posed by dangerous goods in the event of an accident.
On 9 December 2014, North Dakota's Industrial Commission also approved an order
requiring Bakken crude to be conditioned before shipment by rail. Under the order,
crude oil cannot have a vapour pressure that exceeds 13.7 psi (1 psi below the national
standard of 14.7).
More importantly, the Federal Railroad Administration (FRA) proposed measures to
prevent trains transporting specific hazardous materials from being left unattended on
the main line and called for secure locks to be installed on locomotive cab doors to
prevent unauthorized access.
The TSBC also recommended that tougher standards be placed on all DOT-111 class tank
wagons (as featured in the Lac-Mégantic accident). The DOT-111 is considered the
workhorse of the North American fleet, one US government-commissioned report
noting that there are about 228,000 in service, about 92,000 of which carry flammable
liquids.
In May 2015, however, the final specification for a new non-pressurized tank wagon –
the DOT-117 – was released. The new wagons will adopt jacketed and thermally
insulated shells of 9/16-inch steel, full-height half-inch-thick head shields, sturdier, recloseable pressure relief valves and rollover protection for top fittings.
The timeline for the upgrade or withdrawal of existing DOT-111s and the safer, industrysponsored CPC-1232s (constructed since 2011) is 1 May 2025.
39
Furthermore, new ruling requires HHFTs (high-hazard flammable trains9) to have in
place a two-way end-of-train (EOT) device or a distributed power (DP) braking system.
HHFTs are limited to 50 mph, with a conditional 40 mph maximum in densely populated
urban areas.
Trains meeting the definition of a HHFUT (high-hazard flammable unit train10), must be
operated with an electronically controlled pneumatic (ECP) braking system by 1 January
2021, or reduce maximum speed to 30 mph. All other HHFUTs must have ECP braking
systems installed after 2023.
The urban speed limit will be lifted for trains consisting entirely of new or upgraded
wagons meeting the DOT-117 requirements.
The requirement to equip tank wagons and locomotives with ECP brakes is intended to
slow a train with braking force applied simultaneously along its length. Conventional
train braking relies on air pressure releases that occur serially along the length of the
train, rather than instantaneously at each wagon.
Rail companies objected to compulsory ECP after it was proposed in the DOT’s ‘notice of
rulemaking’ last August, saying braking distributed via mid-train locomotives and end-oftrain devices would be just as effective.
A summary of the US Department of Transport’s new ruling may be found here.
7.3.2 Freight derailments on GB rail
As noted above, RAIB published two reports on freight train derailments in 2014/15,
which highlighted issues with:

Track maintenance

Track geometry – absence of check rail

Vehicle acceptance

Vehicle loading

Staffing levels

Compliance with standards
9
Defined as a continuous block of 20 or more tank wagons or 35 or more wagons
dispersed through a train loaded with a flammable liquid.
10
A single block with 70 or more tank wagons loaded with Class 3 flammable liquids,
with at least one tank wagon carrying Packing Group I materials.
40
7.3.3 Freight train derailment at Gloucester, 15 October 2013 (pub.
10/14)
At about 20:15, a container train derailed
about four miles south-west of Gloucester. It
was travelling at 69 mph when the rear
wheelset of the last wagon came off on
defective track. The train continued to the
station, where it was stopped by the
signaller.
As a result of the derailment, a container fell
from the train, and the rear wagon, four miles of track, signalling cables, four level
crossings and two bridges were damaged.
Causes
RAIB found the immediate cause of the accident to be regularly spaced dips in the rails –
a defect known as cyclic top. The dips had formed due to water flowing underneath the
track and, although the local maintenance team had identified the problem, the repairs
they carried out were ineffective.
The severity of the dips required immediate action by Network Rail, including the
imposition of a speed restriction, but no such restriction had been put in place. In fact,
speed restrictions had repeatedly been imposed since December 2011, but were
removed each time repair work was completed; on each occasion – as with the most
recent occurrence – such work proved ineffective.
The susceptibility of the wagon to these dips in the track, especially when loaded with
the type of empty container it was carrying, was not identified when the wagon was
tested or approved for use on the main line.
RAIB also observed that the local maintenance team had a shortfall in its staffing
resources.
What was done?
In March 2014, Network Rail renewed the track where the derailment occurred. It also
replaced the steel sleepers with concrete through the cutting, allowing a wider range of
track repair methods, including measured shovel packing and stone blowing, to be used
to correct any defects that might form in the future.
Measures were taken to improve the drainage through the cutting, while the ballast
formation was adjusted to intercept water flowing from the bottom of the local cutting
towards the track bed.
41
Network Rail’s Western Route also implemented a programme to survey all its off-track
drainage. Any missing drainage assets will be added to its inspection and maintenance
management system.
In addition, Network Rail has recruited staff to fill the five vacancies in the track
maintenance team based at Gloucester. The management team within the Bristol
Delivery Unit has also carried out a review of its resources across all its teams.
RAIB recommended reviewing the drainage in the area where the train derailed, revising
processes for managing emergency speed restrictions for cyclic top, providing track
maintenance staff with a way of measuring cyclic top after completing repairs, and
investigating how cyclic top on steel sleepered track can be restored effectively.
The train operator is recommended to mitigate the susceptibility of the type of wagon
involved to cyclic top, while the methods for assessing a vehicle’s response to defects of
this type should be reviewed. (Note that cyclic top is also featured in the freight train
derailment at Heworth on 23 October 2014, the investigation into which is ongoing).
7.3.4 Freight train derailment at Camden Road West Junction, 15
October 2013 (pub. 10/14)
At about 02:40, a freight train travelling from Birmingham to Felixstowe derailed close to
the site of the former Primrose Hill station in north-west London. There were no
reported injuries, though both train and infrastructure were damaged.
The North London route, which carries London Overground services as well as freight
trains, was also closed for six days.
One wagon ran derailed until the train reached Camden Road West Junction. At this
point, an empty container toppled off and damaged the overhead line equipment
(OHLE).
42
Causes
The derailment was caused by a combination of the track geometry and condition, as
well as the longitudinal and lateral asymmetric loading of the wagon, which reduced its
resistance to derailment on twisted track. The rules on the loading of FEAs had been
relaxed following a derailment at Duddeston Junction in 2007, allowing greater
longitudinal asymmetry, although it is possible that wagons of this type are particularly
prone to flange-climbing derailments on twisted track when loaded asymmetrically.
Furthermore, the effect of asymmetric loading on the resistance of FEAs to derailment
on twisted track was not considered as part of the process for accepting them for
operation on the main line.
Following previous similar derailments, Freightliner, Network Rail and RSSB did not fully
quantify the risk from operating FEAs with asymmetric loading or determined whether
measures were required to mitigate the risk.
There was insufficient awareness of the ongoing poor condition and classification of the
North London lines among the managers of the Euston Maintenance Delivery Unit.
In addition, no check rail had been provided on the tight-radius curve to provide lateral
restraint to wagon wheels and prevent flange-climbing.
As a learning point from this accident, RAIB also identified that Network Rail should give
particular attention to the possible consequences of a high turnover of responsible staff
during reorganisations. (Note that staffing was also a theme of the Gloucester
derailment.)
What was done?
Network Rail has since lifted and packed the track in the vicinity of the point of
derailment to correct the twist faults.
RAIB has also recommended providing guidance to managers responsible for track
maintenance on the actions to be taken if measurements by track recording vehicles do
not take place as planned. It has called for the factors that contribute to the derailment
of unevenly loaded container wagons to be considered and the case for additional
measures to mitigate the risk to be evaluated. Furthermore, the requirements for the
design and acceptance of freight wagons should be clarified.
7.3.5 Freight Train Derailments Working Group
The standards and processes the industry has in place – coupled with the significant
amount of time, effort and money spent on infrastructure inspection, renewals, rolling
stock design, approvals and maintenance – mean that the residual risk from freight train
derailments caused by track twist and uneven loading is actually relatively low.
43
Indeed, RSSB’s Safety Risk Model (SRM) shows it to be 0.242 FWI/year (not including
possessions or incidents in yards, depots or sidings). This is 27% of all train accident risk,
which is itself 8% of the total railway system risk.
However, in light of the events at Gloucester and Camden Road, the Freight Technical
Committee proposed that a cross-industry group be established so that rolling stock and
infrastructure experts could consider the issue jointly. The resulting Freight Train
Derailments Working Group includes representatives from Network Rail, freight
operators, RSSB, Interfleet, Huddersfield University, Lloyds Register Rail and the ORR.
The Group intends to look closely at data relating to derailments, focusing initially on
track twist-related incidents, but also considering cyclic top. In addition, it will review
the origin of current requirements, with a view to ascertaining what could be done
differently to manage the situation in light of the low level of risk, and in light of the
changes the industry has seen over the last decade.
So what has changed?
A decade of development
Our industry has never stopped evolving, improving, innovating, since Richard Trevithick
ran the world’s first locomotive in 1804. Since 2005, however, we have enjoyed
increases in traffic – both on the freight and passenger sides – which in turn has meant
an increase in our hours of operation. On the other hand, the track recording regime is
more extensive and track quality has improved.
The wagon fleet has fallen by a quarter over the same period, though bogie wagons –
one-third of the total fleet in 2005 – have doubled in number and now carry more that
90% of the total tonnage. Around 70% of the more ‘twist-prone’ two-axle wagons have
also been withdrawn, but changes to the standard covering a wagon’s structural
strength has resulted in vehicles which ‘flex’ less and may therefore be less tolerant of
track twist.
At the same time, we have seen a 50% rise in container traffic, while the containers
themselves are getting larger both in height (8’ to 9’ 6”) and length (20’ to 40’, 45’ and
50’). These increases may have a bearing on stability, although the two recent incidents
involving containers being blown from freight trains appear to implicate the reliability of
the retaining spigots.11
To take the matter forward, RSSB has proposed to:

11
Update its risk assessment criteria.
Scout Green (07/03/15) and Deeping St Nicholas (31/03/15) are still under
investigation by RAIB.
44

Identify changes to the railway over the last 10 years and ascertain those that may
come in the future.

Assess the benefits that could be gained from amending the track twist
measurement criteria.

Review existing loading practice and provide guidance for vehicle testing.

Assess the need for and feasibility of testing and computer simulations of existing
wagons to establish limiting offset loading conditions and related wheel unloading
limits.

Analyse the existing wagon uneven load profiles.

Review the rules and regulations around the loading of containers as a wider
transport issue.
Learning points:

Is your freight train properly loaded?

Is the track in your area in good condition?

As a rolling stock engineer, do you ever discuss issues with your
infrastructure counterpart (and vice versa)?

Do you have enough staff to complete maintenance tasks to the required
standard?
7.4 Signals passed at danger
In the aftermath of the high-profile SPAD accidents at Southall (1997) and Ladbroke
Grove (1999), the rail industry took a closer look at the causes of SPADs, the precursors
to SPADs and the risks that surround them. Groups were set up nationally and locally to
monitor the situation and implement various initiatives to bring the risk down. When
this work began, there were over 500 SPADs a year; now there are fewer than 300.
The professionalism of drivers, the relevance of driving policies and practices and the
success of the TPWS have combined to achieve a situation where:

Only one SPAD occurs for around every 50,000 red signals approached

The vast majority of train journeys are SPAD-free

Only a small minority of drivers are ever involved in a SPAD
45
However, although SPADs are relatively low in frequency, they still have the potential
for high-consequence loss. This is why they continue to be monitored closely (as in
Network Rail’s ‘deep dive’ review of 2013, which we featured in last year’s LOE Annual
Report).
7.4.1 Passenger train SPAD at Greenford, 20 March 2014 (pub.
12/14)
At around 11:55, a Paddington–West Ruislip
service passed two consecutive signals at
danger near Greenford. It was stopped
when a signaller sent an emergency radio
message to the driver. Although no-one was
hurt, the unauthorised entry of any train on
to a single line creates the potential for a
serious collision.
A freight had passed the junction shortly
before the passenger train was due.
Because the freight was still occupying the
line between Greenford and South Ruislip,
the signaller kept the signal at danger. The
passenger train, travelling at about 20 mph,
passed both it and the next one, which was
also at danger. It also passed over the
junction and on to the single-track section
towards South Ruislip, which was still
occupied by the freight. The train had
travelled about one mile beyond Greenford
by the time that the driver received the
emergency radio message.
Causes
The reasons why the driver did not react to
the two signals are not certain. It is possible
that he had formed the impression that the
train had been given clear signals through
Greenford, because of his interpretation of
the meaning of the signal preceding those
that he passed at danger, and he had not
been stopped by signals at Greenford in the
recent past.
46
SPAD at Wootton Bassett
At the end of the reporting year, RAIB
had started investigating an incident at
Wootton Bassett Junction, involving
steam locomotive no. 34067 Tangmere.
At 17:25 on 7 March 2015, a special
from Bristol Temple Meads passed a
signal at danger and came to a stand
across the junction.
RAIB’s preliminary findings show that, at
around 17:24, the train was approaching
SN43 signal at 59 mph, when it passed
over an AWS magnet in place for a
temporary speed restriction. This gave
an audible and visual warning in the cab.
However, as the driver did not
acknowledge within 2.7 seconds, the
train’s brakes were automatically
applied. This should have resulted in the
train being brought to a stand. In these
circumstances, the Rule Book requires
that the driver immediately contact the
signaller.
RAIB says that evidence shows the
driver and fireman to have taken an
action which cancelled the effect of the
AWS brake demand after a short period
and a reduction in train speed of only
around 8 mph. The action taken also
made subsequent AWS or TPWS brake
demands ineffective.
TPWS was fitted to the train and to both the signals, but it did not intervene to apply the
brakes of the train, as it was intended to do. This was because the on-train equipment
had self-isolated when the driver prepared the train for departure from Paddington
while it was standing over a TPWS trigger loop. The isolation of the equipment was
indicated by a flashing light in the cab, but the driver still drove the train.
Although the signaller at Greenford wanted to stop the train by sending an emergency
call via GSM-R12 radio, he did not try because believed it wouldn’t reach the driver, the
train having ‘vanished’ from the system soon after it passed the box. Instead, he
contacted Marylebone signal box, which was able to send a message to the train.
RAIB noted that the driver management process within Chiltern Railways did not
address the driver performance issues which contributed to this incident. Furthermore,
the signaller’s training on GSM-R had not equipped him adequately to deal with
emergency situations.
RAIB identified the following learning points for the railway industry:

At locations where the delayed clearance of signals is used to warn train drivers
about the state of the line ahead, signallers must be confident that they know
enough about the position and speed of the train to judge accurately when the
signal should be cleared. This may mean waiting until they are certain the train has
stopped.

Train operators are reminded of the need to assess periodically whether it is
reasonably practicable to upgrade on-train TPWS to address known shortcomings
in the Mark 1 equipment (such as the equipment self-isolating when a cab is
opened with the receiver directly over an active loop, and the readiness with which
it can be reset after an intervention).
What was done?
On 24 March 2014, the operator issued a briefing notice to its drivers on ‘Checking and
Responding to TPWS Indications on Cab Mobilisation’, describing how TPWS should
behave when a driver is setting up the cab, highlighting the meaning of the flashing
yellow light, and setting out the action to take if the flashing or steady yellow light
appears.
Network Rail modified the GSM-R system at Greenford, so that trains travelling between
there and Northolt Junction will remain the signal box’s terminal for the whole time they
are on the single line.
12
GSM-R stands for Global System for Mobile Communications – Railway.
47
RAIB also recommended that the operator review its driver management processes to
confirm that the training and briefing given to drivers is comprehensive with regard to
the equipment and systems that drivers use, and that assessment of drivers covers the
identification of, and response to, TPWS fault warnings.
Network Rail reviewed its implementation of GSM-R, and the training given to signallers
in its use.
7.4.2 Incident Factor Classification System
As noted in section 4.2.2, RSSB and Network Rail have developed a SMIS-based Incident
Factor Classification System (IFCS) to capture the accident causes in RAIB reports and
Formal Investigations. In 2014, the
The IFCS SPAD report also provides:
system was used to start informing an

Analysis of human errors/violations and
Industry Human Factors SPAD Review.
incident factors that have the potential
The Review aims to provide more insight
to increase risk after a SPAD.
into the human factors issues associated

Examples of how the data can be used
with SPADs, in order to identify the
to look at different error/violation and
strengths and any weaknesses in SPAD
incident factor profiles for different
management processes. Its first output
contexts (eg driver years in service or
was a report that summarises the human
light locomotive/ECS movements).
error types and underlying causes (using
Network Rail’s ‘10 incident factors’) for a
sample of 257 SPADs that occurred
between 2011 and 2013.

Data on quality ratings for investigation
reports from a human factors
perspective.
From the investigation reports on these
257 SPADs, 1651 factors were identified,
of which 924 were causal/contributory.
This highlights the multi-factor nature of
SPAD incidents, with an average of six
factors being identified for each report.

A summary of a short survey on
industry SPAD investigation processes
for passenger operations, freight and
Network Rail.
The dominant driver error types that cause or contribute to SPADs may be grouped
together as slips/lapses. This includes cases where the driver had the correct intention,
but their behaviours didn’t go as planned. Examples include pressing the wrong button,
forgetting to implement routine signal observation tasks and misperceiving signal
aspects.
Slips/lapses are causal/contributory in 70% of SPADs. The key mechanisms underpinning
them are distraction/pre-occupation and expectation errors (eg, a caution signal is
perceived to be green rather than at caution because it is usually at green in the driver’s
experience).
48
At least one of Network Rail’s ‘10 incident factors’ feature in 58% (150) of the 257
incidents. Table 3 shows the ‘Top 5’ factors for passenger operations, freight and
Network Rail, based on the percentage of SPAD incidents to which they contribute:
‘Top 5’ 10 Incident Factor types identified as causal / contributory
No.
1
Passenger
Freight
Network Rail
(n=197 incidents)
(n=54 incidents)
(n=257 incidents)
Personal
12%
Knowledge, skills and
22% Equipment
experience
15%
2
Knowledge, skills
and experience
9%
Equipment
17% Communications
7%
3
Supervision/
9%
Communications
13%
Practices/Proces
ses
5%
Communications
8%
Personal
13%
Supervision/
Management
5%
Workload
5%
Supervision/
Management
13%
Work
environment
2%
Management
4
5
The rankings and review of the details of these categories can inform industry SPAD
management strategies, allowing us to identify (for example) the importance of fatigue
or communication to SPAD incidents.
Overall equipment factors contribute to 22% of incidents, while personal factors,
communications, knowledge skills and experience, and supervision/management each
contribute to 13-14% of incidents. This diversity emphasises that SPAD management is a
broad church encompassing competence management, communications and driver
welfare management (inter alia).
The findings on error categories and the ‘10 incident factors’ will be reviewed and
debated with industry during the next phases of the Review.
7.4.3 SPAD ten-year strategy
In January 2015, the RSSB Board agreed the need to develop a forward strategy for the
risk management of SPADs over the next ten years, covering the period before
widespread installation of ERTMS.
49
The strategy will be delivered through three key phases, consider the need for new
controls and mitigations in the short, medium and long term, and capture existing good
practice.
Note, however, that existing management activity to address SPAD risk will continue
during the development phases of the strategy though established industry
mechanisms, local tactical groups (like OPSRAMs), and at national level through the
System Safety Risk Group (SSRG) and its expert sub-group, TORG. These existing
mechanisms will help inform Phase 1 of the strategy.
The SPAD Risk Reduction Strategy, developed through industry collaboration will form
one component of the new Rail Industry Health & Safety Strategy and will comprise
three key phases:

Phase 1 – Review of existing management
activity, tactical improvements and good
practice

Phase 2 – Next generation of risk mitigation

Phase 3 – Delivering the long-term plan
(Phases 1 and 2 will run in parallel.)
CIRAS activity – SPADs and
vegetation
Sometimes vegetation can
obscure signals and increase the
chance of a SPAD. Find out what
happened in one specific case in
Issue 55 of the CIRAS magazine.
A SPAD Risk Reduction Strategy Steering Group will
provide executive level commitment and support,
establish and set the remit for expert sub-groups,
Click here for more details.
make strategic decisions, monitor progress with
individual projects or work streams, and assist with
any cross-industry issues. It will develop the ‘vision’ for SPAD risk management, endorse
business cases, set the critical success factors, and (where appropriate) set specific
targets for risk reduction over time, aligned to the phases and various controls
proposed.
Phase 1 – Review of existing management activity, tactical
improvements and good practice
As recognized by the Human Factors review, the issues that affect SPAD performance
and risk are varied and multi-faceted and have been the subject of considerable analysis
and review over many years. This has resulted in cross-industry improvement initiatives,
including:

Identifying fit-for-purpose and continuously improving Safety Management
Systems (SMS).

Optimising the existing infrastructure, including signal risk assessment and
management of multi-SPAD signals.

Highlighting the opportunities and risks posed by infrastructure renewal schemes.
50

Recognising the role of recruitment, competence, fitness and the human
performance of train drivers.

Recognising the effects of timetabling and performance on SPADs.

Ensuring TPWS remains an effective and reliable system.

Delivering and managing the ERTMS fitment programme.

Making sure we have robust SPAD and signal risk assessment tools that people can
understand and use to make informed decisions.

Undertaking comprehensive investigations and learning from identifying underlying
causes.
The industry continues to invest considerable resources and efforts to mitigate SPAD risk
and although these efforts have been successful in driving improved performance over
the last 15 years (mainly through the implementation of TPWS), this improvement has
slowed since 2006 and in late 2012 started to rise again.
These improvements have been achieved through mitigations introduced in a
fragmented way and the challenge facing the industry now is how to further improve
risk control performance without devoting disproportionate resources considering the
relatively low risk that SPADs now present.
This phase of the strategy will therefore consider the ‘as is’ situation, through
understanding the existing SPAD mitigations in place within the industry, reviewing
previous research and literature, identifying patterns and trends, and capturing existing
good practice.
Phase 2 – Next generation of risk mitigation
Work has begun to develop phase by following the principles set out in Taking Safe
Decisions, which are aligned to overarching legislative requirements.
Taking Safe Decisions presents a framework comprising three related activities:
monitoring safety performance; analysing and selecting possible options to implement,
and implementing agreed change. Changes made are then monitored and the cycle
begins again. Implementation of Phase 2 therefore requires the need for initial analysis
to underpin the selection of options and initiatives. This will include not just data and
expert judgement, but also the collation of existing good practice through Phase 1.
RSSB has recently made significant progress in its ability to understand and analyse
SPAD risk. In particular, it is starting to identify the SPAD rate per demand on a signal by
signal basis. This analysis should drive a more targeted appreciation of risk; therefore
the ability to understand and implement more targeted, cost-effective control strategies
can be proposed.
51
This phase of the strategy will be delivered through the three expert groups for
operations, human factors and infrastructure & engineering, discharging their activity
through the following staged approach:
The expert sub-groups will each work to a specific terms of reference developed by the
steering group, which will set out the scope of their activity. Early nominations for the
groups have already been sought for the operations, and infrastructure and engineering
groups.
The existing Human Factors SPAD Review group will continue with its existing
membership, but to a revised remit once set.
The scope of the three expert groups will cover the following areas within their
individual areas of expertise:
52
On-going development and delivery of the first two phases will be communicated to key
industry groups, conferences, roadshows and so on. Findings will also be shared with the
UIC13-sponsored SPAD Task Force. This will provide learning on a European scale.
Phase 3 – Delivering the long-term plan
Phase 3 of the strategy will be delivery of the longer-term mitigations to reduce the risk
from SPADs through CP6 and beyond, supported by the business cases developed as
outputs from Phase 2.
Many of these mitigations will be aimed at engineering solutions. Success will require a
step-change in the industry’s approach to such strategies and their funding mechanisms.
A new approach will need wide industry ownership and support in order to seek
appropriate levels of funding in future control periods.
13
International Union of Railways.
53
In particular, funding for engineering-led mitigations will often fall to the infrastructure
manager to fund, but with the majority of the benefits being realised by train operators
in reduced industry costs, rather than in terms of risk reduction.
7.4.4 Train Protection and Warning System
TPWS was implemented in Great Britain as an interim measure to reduce the
consequences of SPADs, pending the implementation of full protection through systems
that monitor driver performance continuously.
It was envisaged that this higher level of protection would be delivered by the roll-out of
ERTMS. In the intervening period, it has become clear that this roll-out will take
considerably longer than planned, leaving TPWS as the primary means of mitigating
SPAD risk for some time to come.
TPWS is installed on around 13,000 main aspect signals protecting junctions, 650 buffer
stops and 1150 permissible speed reductions, along with a number of other locations
where its fitment has been considered to be reasonably practicable.
Since the completion of installation at the junction signals in 2003, and in conjunction
with other risk reduction measures, SPAD risk has reduced by 90%. Over this period,
TPWS has proved to be a very reliable system, with the probability of it failing to operate
on demand after a SPAD (due to an undetected fault) being around 0.005% (or 1 in 2000
SPADs).
However, an incident in which a passenger train passed two consecutive signals at
danger near Greenford on 20 March 2014, and an incident involving a charter train at
Wootton Bassett Junction in March 2015, reminds us that we should not be complacent.
7.4.5 TPWS – ‘reset & continue’
A TPWS ‘reset & continue’ incident occurs when a train passes a signal at danger, is
stopped by TPWS, but the driver then resets the equipment and continues without the
signaller’s authority. When this happens, the protection provided by TPWS is obviously
reduced.
There was one TPWS ‘reset and continue’ during 2014/15, in which a depot driver
shunted an empty coaching stock formation past a signal on the main line.
What was done?
In November 2009, the TPWS Strategy Group was set up to help ensure the long-term
viability of TPWS.
Part of the remit of the group is to monitor the reliability of the system with a new
programme of work being recently established to ensure that the industry understands
what the reliability requirements of TPWS should be going forward and how it should
monitor against the requirements.
54
Completing and promoting the industry response to reset and continue has also been a
key part of the group’s work.
A TPWS strategy action plan may be found on Opsweb, but for more information,
contact: Chris Harrison at chris.harrison@rssb.co.uk
8 Lessons learnt in 2014/15 – people on trains and in
stations
Since its formation, RSSB has provided the industry with regular updates on risk and
safety performance. In 2010, exposure of a rise in PTI risk prompted the focus of
resource and expertise on this important area.
A number of incidents on which RAIB reported in 2014/15 highlighted the ongoing need
for this focus, with specific reference to train dispatch procedures, the operation of train
doors and the slope of station platforms.
8.1 Statistical overview
There were four fatalities occurring in stations during 2014/15:

Two passengers who fell from the platform edge: in one event, the person was
electrocuted and in the other event the person was hit by a train entering the
station.

One passenger fell while running down a flight of stairs.

One member of the public, a young boy, was hit by a train during an incident where
his mother accessed the track with the apparent intention of committing suicide.14
When non-fatal injuries are taken into account, the total level of harm to passengers
and public in stations was 41.6 FWI (40.6 FWI in 2013/14). When normalised by
passenger journeys, the level of harm in stations may be seen to have fallen by 2%.
More than half the harm in stations involves slips, trips and falls. Incidents tend to occur
while people are moving around the concourse and other areas – and while using stairs
and escalators. In fact, around 20% of all station harm over the last five years has arisen
from slips, trips and falls on stairs. The fatality during the current year highlights their
potentially serious nature.
14
In this incident, the boy’s mother was also fatally injured, but this event is counted in
the suicide/suspected suicide category.
55
Train and station safety at a glance
Passenger &
public in stations
(35%)
Passenger &
public on trains
(6%)
Weighted injuries
Fatalities
13.7
11.9
12.6
12.1
11.2
10.2
12.0
9.4
9.3
8.5
Other
accidental
risk (51%)
90
80
70
60
50
40
30
20
10
0
42.4
41.3
38.9
39.6
39.6
46.7
47.2
48.7
45.1
47.9
Workforce on
trains (3%)
FWI
Workforce in
stations (5%)
Trends in harm
2005/06
2006/07
2007/08
2008/09
2009/10
2010/11
2011/12
2012/13
2013/14
2014/15
2005/06
2006/07
2007/08
2008/09
2009/10
2010/11
2011/12
2012/13
2013/14
2014/15
Risk in context (SRMv8.1)
Passsengers/public
Workforce
For more statistical analysis on passenger risk, see Chapter 4 of the ASPR.
Topics covered in this section:

The platform-train interface (PTI)
8.2 Platform-train interface
There are many overlapping factors which affect the occurrence of accidents at the PTI,
including the age and gender of the passenger, whether or not they are intoxicated, and
their familiarity with rail travel.15
The accident rate varies throughout the year, with changes to the passenger
demographic (and the weather), and also changes according to the time of day or week
that the journey is taking place.
RAIB published two mainline PTI-related reports in the year, covering incidents in which
wheeled conveyances rolled off the platforms at Southend and Whyteleafe, and an
incident in which a person was trapped in a train door and dragged a short distance at
Newcastle Central.
15
A commuter who is a more experienced rail user may be able to deal with hazards
unique to rail travel better than a tourist who has little experience.
56
8.2.1 Wheelchair rolls on to the track at Southend Central
(28/08/13; pushchair rolls on to the track at Whyteleafe (18/09/13)
(pub. 08/14)
On Wednesday 28 August 2013, a wheelchair user and her carer were waiting at
Southend Central station for the arrival of a train when the wheelchair started to roll
towards the edge of the platform, before falling to the track. Although the passenger
and her wheelchair were recovered before the train arrived, the passenger was seriously
injured.
On Wednesday 18 September 2013, a mother and her two young children entered
Whyteleafe station, near Caterham in Surrey. As the mother was purchasing a ticket
from the machine on the platform, her baby’s pushchair started to roll away. The
mother was unaware of this until it was too late to stop the chair rolling off the edge of
the platform. The baby suffered minor injuries.
Causes
RAIB found the immediate cause to be that the platforms at Southend Central and
Whyteleafe sloped towards the railway. (Note that when built – in the nineteenth and
early twentieth centuries – there were no specific requirements or standards re
platform slopes.)
RAIB also found that the individuals in charge of the wheelchair and pushchair had not
applied the brakes and had not noticed that the platform sloped towards the track.
There was nothing to alert users of either station to the presence of the slope.
The railway industry had generally recorded previous incidents of a similar nature as due
solely to errors by the individuals concerned. As a consequence, it had not recognised
the part that sloping platforms played in the incidents.
RAIB has identified the following learning point for the railway industry:

All station operators are reminded of the importance of providing a means for their
staff to call the signaller in case of emergency. Measures for consideration include
the provision of mobile phones with a quick dial facility, or alternatively, posting a
current and legible list of emergency numbers at prominent positions throughout
the station.
What was done?
A number of companies took action after these incidents. c2c, for example, issued a
poster to all its stations to brief staff and highlight the actions they should take to stop a
train in an emergency or make contact with the signaller before going on to the track to
deal with an emergency.
57
c2c has implemented posters/ platform
transfers asking for wheelchairs and pushchairs
to be positioned parallel to the platform,
reviewed its station notices on how to stop a
train in an emergency to ensure contact
telephone numbers are more prominent and
legible, and introduced stencils on platforms
that slope towards the railway (inter alia).
Southern surveyed its stations, focusing in
particular on the location of platform furniture
(seats, ticket machines, help points, vending
machines). It also introduced a poster to warn
passengers of the risks (see opposite).
RED 28
RED 28 – released November
2010 – covered the risks at the
platform-train interface.
See Opsweb for details.
Network Rail incorporated the slope risk to
wheelchairs/pushchairs into the joint safety
working group (Network Rail and RSSB) remit to review the risk associated with the PTI.
RSSB included consideration of platform cross fall in the development of its national
strategy for managing risk at the platform-train interface and is working on research into
gradient profile risk
RAIB also recommended (inter alia):

Implementing processes for managing the risk of wheelchairs and pushchairs
rolling onto the track, including the factoring in of platform slopes.

Determining when a slope towards the railway could become a significant hazard,
publishing ways of mitigating that risk in a guidance document.

Implementing a process to improve the investigation and recording of roll-off
incidents and the way in which data is shared.
8.2.2 Passenger trapped in a train door and dragged a short distance
at Newcastle Central station, 5 June 2013 (pub. 09/14)
At 17:02, a passenger was dragged by a train leaving Platform 10 at Newcastle Central.
Her wrist was trapped by an external door and she was forced to run beside it to avoid
being pulled off her feet.
The train – a ‘Desiro’ Class 185 diesel multiple unit (DMU) – reached a maximum speed
of around 5 mph and travelled around 20 metres before coming to a stop. Its brakes
were applied either by automatic application following a passenger operating the
emergency door release handle, or by the driver responding to an emergency signal
from the conductor.
58
Causes
RAIB found that the conductor had not carried out a safety check before signalling to the
driver that the train could depart. Platform 10 at Newcastle Central is curved and safe
dispatch is particularly reliant on following the correct dispatch procedure.
RAIB also found that, although the doors complied with the applicable train door
standard, they were, in certain circumstances, able to trap a wrist and lock without the
door obstruction sensing system detecting it. Once the doors are detected as locked, the
train is able to move.
In 2004, although the parties involved in the train’s design and its approval for service
were aware of this hazard, the risk associated with it was not formally documented or
assessed. The train operator undertook a risk assessment in 2010 after reports of
passengers becoming trapped. Although they rated the risk as ‘tolerable’, the hazard
was not recorded in such a way that it could be monitored and reassessed, either on
their own fleet or by operators of similar trains.
RAIB identified the following learning points:

Those responsible for train dispatch should ensure that they undertake a thorough
train safety check as defined in module SS1 of the Rule Book and should never
solely rely on any indication given by the train door interlock circuit.

Train designers and operators should make themselves aware of public perceptions
of how train doors operate when they come into contact with obstructions and
take account of this in both the design and testing of doors, and dispatch
procedures. Train operators should also seek ways in which they can positively
influence the behaviour of passengers to minimise door trappings.

Those who work for train operating companies who receive information from their
customers should be trained to identify safety issues within customer
communications and alert those responsible for safety in a timely manner.

Train operating companies should share with other operators, designers and
maintainers of similar fleets, details of train door trapping incidents and precursor
events that may indicate shortcomings in safety equipment or systems, either
existing or emerging, which could lead to injuries or fatalities.
What was done?
Since the accident, the operator (First TransPennine Express) has completed station
operational risk assessments for all of the stations at which it calls, including Newcastle
Central. This has been done following the guidance within the Rail Industry Standard
RIS-3703-TOM. It has also created risk reduction action plans where issues have been
identified.
59
RSSB ensured that passenger behaviour at train doors was considered by the crossindustry Platform Train Interface Strategy Group.
RAIB issued an Urgent Safety Advice on 2 August 2013 once the method of overcoming
the detection of the sensitive edges had been confirmed. This was sent to operators of
Desiro UK fleets with electrically operated sensitive door edges.
The notice advised that a wrist could become trapped and to gather data on the
frequency of trapping incidents. It also advised them to consider the need for additional
operational and/or technical measures to manage the risk of passengers being trapped
and dragged, paying particular attention to curved or higher-risk platforms.
First TransPennine Express:

Re-briefed its conductors on the need to undertake thorough checks when
dispatching trains.

Reassessed the risk of a fatality and injury from a trapping and dragging accident
and, together with Siemens, is appraising options of modifying the door sensitive
edges.

Added signs, assisted by Siemens, to all the passenger doors of its Class 185 and
350/4 units warning passengers not to obstruct the doors when they are closing.

Revised its procedures to improve the reporting and assessing of safety related
matters including guidance to its customer relations department to raise its
awareness of possible safety issues that may be within communications from its
customers.

Stopped operating services from Platform 10 at Newcastle Central station.

Increased the frequency of its planned unobtrusive conductor monitoring visits by
30%. The choice of locations is now based upon risk and these have been
supplemented with additional monitoring visits by the On Board Service and
Operations Teams.

Begun to deliver an accident investigation and risk management course to its team
leaders, health and safety representatives and managers.

Took disciplinary action with regard to the conductor.
London Midland:

Issued an operational notice to its senior conductors and train dispatchers
highlighting this incident and the hazard exposed.

Ran a campaign at stations (involving a mix of posters, station and train
announcements, and passenger information system display messages) to influence
passenger behaviour away from trying to board/alight trains when the doors are
closing.
60

Fitted warning signs on all doors of its current Class 350 fleets (with the assistance
of Siemens). It is planning to fit these to its new Class 350/3s by the time they come
into service.
Heathrow Express:
RAIB report – Holborn LUL

Completed a review of its station dispatch
risk, including changes to the competency
management system for dispatchers in
order to improve monitoring, assessment
and knowledge.

Added signs, assisted by Siemens, to all
passenger doors of its Class 360/2s,
warning passengers not to enter the train
when the doors are closing.
On 3 February 2014, a passenger was
dragged about 10 metres along the
platform at Holborn by a departing
train after her scarf had got caught
between the closing doors of one of
the carriages.
Siemens has reported that its latest Desiro units
(such as the Class 380 and Class 700) have
electrically operated doors, unlike those on the
older 185s, 350s and 360/2s, which are
pneumatically operated. Furthermore, the
doors on these newer trains have stiffer leading
edges; it is not possible to deflect them in the
same manner as those on the 185s.
RSSB included passenger behaviour in relation
to train doors in the development of its national
strategy for managing risk at the platform-train
interface.
RSSB also reported that the Railway Group
Standard for passenger doors, GM/RT2473, will
be superseded by the new European standard
EN14752. The obstruction tests specified within
it are the same as the Railway Group Standard
tests. RAIB has consulted with RSSB as to how
best capture the learning from this accident
with respect to obstruction detection.
RAIB also recommended:

Re-assessing the risk from injury and
fatality due to a trapping and dragging
incident and take appropriate action to
reduce the risk.
RAIB found that the force required to
remove the trapped scarf is likely to
have been less than the maximum
specified in the relevant standard,
though it may have been hard for a
person taken by surprise like this to
exert such a force.
LUL modified its training courses to
raise awareness of the ways staff can
take rapid and effective action to tell
an operator not to start the train
should any emergency or out-of-course
event occur.
RAIB also recommended:

Providing station assistants with
an effective way of alerting train
operators to dangerous situations
that arise after they’ve given the
signal to start the door closing
sequence, but before the train
has started to move.

Reviewing how the role of the
station assistant is described in
the rule book (inter alia), so that
the duty to respond rapidly to
dangerous dispatch events is
given appropriate emphasis.
61

Redesigning the doors, as used on the Class 185 and other similar units, for future
vehicles supplied to the UK, to reduce the probability of a passenger being trapped
in them but not detected by the door control system.

Reviewing design processes to ensure that hazards associated with the design of
trains are recorded and assessed.

Reviewing safety management processes to ensure the presence of a system for
the identifying and recording hazards, assessing the risk associated with each, and
managing the implementation of any necessary control measures.

Urging the British Standards Institution (BSI) that, in the forthcoming BS EN version
of European Standard EN 14752 (Railway applications – Body side entrance systems
for rolling stock), the UK National Foreword informs readers of the possibility of
entrapment even on correctly adjusted doors that comply with the specified
obstruction tests.

Identifying any additional data that should be captured within SMIS from incidents
of persons trapped by train doors, who are outside the train which subsequently
moves, whether this results in injury or not.
8.2.3 Ongoing investigations
At the end of the year, there was one ongoing RAIB investigation covering the PTI.
Like Newcastle Central, this is another ‘trap and drag’ incident. Unlike Newcastle
Central, it involves not a passenger’s wrist, but her bag strap, which was trapped by the
closing doors of a Class 465 as she alighted at West Wickham on 10 April 2015. As the
train departed, she was dragged under the unit, suffering life-changing injuries as a
result.
At West Wickham, drivers are responsible for checking that it is safe for the train to
depart after the doors have been closed. As is normal at many such stations, no other
members of staff are provided on the train or platform to assist with dispatch. To help
drivers view the side of the train, CCTV monitors are positioned next to the stopping
position of the cab.
RAIB will consider the circumstances of the accident, including the design and operation
of the doors, the associated control system and the actions of those involved.
Preliminary tests have revealed the potential for passengers to be misled – by the ‘open
doors’ button remaining lit after the driver has initiated the door closure sequence –
into thinking the doors will open for enough time for them to safely join or alight from
the train (particularly where the hustle alarm is not sounding because no doors have
been opened in that coach). In such cases, the door can then suddenly close, with
considerable force and without warning, onto a passenger.
62
For this reason, RAIB issued advice to all train operating companies urging them to check
for the presence of this design feature in their own rolling stock. Where the same design
feature is identified, RAIB has advised that consideration be given to ways of reducing
this risk, including the potential to change the door control system. It has also advised
that operators re-brief their train crew and station dispatchers of the need for a final
check that the side of the train is clear before the train starts its journey.
Key issues raised by recent RAIB activity:

The importance of fully checking train doors before trains depart;

The need for drivers to stop trains immediately if the passenger alarm is operated
when any part of the train is within a station;

The need for the passenger alarm to be fully functioning;

The possibility of providing a warning to people on the platform that a train is
about to move; and

The possibility of providing means for platform staff to remove the RA indication
after it has been given.
8.2.4 Platform-train interface strategy
In September 2013, the RSSB Board approved the development
of a PTI strategy for the GB mainline railway and endorsed the
creation of a cross-industry PTI strategy group to support its
production.
The core aim of the strategy is to reduce safety risk and
optimise operational performance and availability of access in a
manner that promotes the long-term best interests of the
mainline railway system.
The first issue, published in January 2015, presents a
comprehensive overview of the industry’s consolidated
approach to identifying causal factors and mitigations in this
area. It considers research (planned or underway) that supports these activities and
proposes timetables for immediate (first year), short-term (CP5), medium-term (CP6 and
CP7), and long-term (CP8 and beyond) implementation.
Like many issues within the whole system railway the PTI affects many areas of design
and operation that are not always compatible:
63

Platform clearances for passenger, freight, and plant vehicles.

Platform and passenger vehicle floor heights.

Optimal step and gap configurations for passengers with and without mobility
issues, and those using wheelchairs.

Passenger train designs, including door configurations, train capacity, provision for
luggage.
And how these might affect overall performance.
The broad range of issues covered in this strategy means that there are areas of overlap;
where an issue in one area impinges on another. For ease of approach, the document
looks at six specific areas of activity describing, where relevant, the activity plans for the
short- and long-term targets:

Data and intelligence gathering

Passenger movements through the station and across the PTI

Train stopping positions, dispatch, monitoring the dispatch corridor, and stopping
once dispatched

Optimising the step and gap

Accessibility

Performance and capacity
The PTI strategy will be a living document, being updated as activities are progressed,
knowledge gained, lessons learned, and targets refined.
Research and development
‘T’ no.
Title
T426
Minimisation of accidents at the platform-train interface – This
research investigated PTI accidents, exploring their primary causes
and the extent to which they can be reduced in number and severity.
It examined public (and staff) behaviours and made
recommendations on how minor changes to procedures or designs
could make cost-effective improvements.
R649 (in
Identifying platform gradient risk – This project will consider when the
development) platform gradient presents a hazard to wheeled transport.
T743
64
A review of passenger train dispatch from stations – This project
reviewed passenger train dispatch arrangements in the light of
current operations, taking risk and human factors principles into
account.
T866
Investigation of platform edge positions on the GB network – This
project built on previous RSSB and Department for Transport
research to improve the platform/train interface on the existing
network, for the benefit of passenger accessibility and compliant
stepping distances.
T1029
Train dispatch tool. This project will provide a train dispatch risk tool
that will be adopted by passenger train operators and widely utilised.
Work is underway with Network Rail to establish whether completed
risk assessments can be fed into a national database to establish an
overview of total dispatch risk.
T1037
Train passenger footsteps investigation to support research into the
reduction in passenger stepping distances and gauging constraints.
This project examines the range of passenger footstep positions for
vehicles in GB in order to inform options for how to best improve
stepping distances in terms of infrastructure works, vehicle fleet
deployment/cascade/modification and new vehicle design.
T1063
Platform recess. An investigation into whether the minimum recess of
300 mm for new or altered platforms is suitable for people to move
into and be clear of a train.
RSSB’s station safety booklet covers slips, trips and falls, PTI risk and mitigation actions,
and related research into issues such as crowd management, tactile edges, wayfinding
and signs, and how to deal with winter conditions. A new edition is in preparation.
The current edition is available from RSSB, or can be downloaded here: LINK
65
9 Lessons learnt in 2014/15 – infrastructure workers
Investigations continue to highlight the dangers
of track working, inadequate safe systems of
working and culture. However, RAIB activity
also raised questions of equipment
modification, possession length and permitted
speeds in worksites.
CIRAS activity –
infrastructure worker
safety
Issue 55 of the CIRAS magazine
includes an article on safety
glasses, their practicality and
issues around their use.
9.1 Statistical overview
There were no workforce fatalities involving
infrastructure staff working on the running line.
However, two members of the workforce were
fatally injured in separate road traffic collisions.
Click here for more details.
The total level of harm arising from track work was 10.2 FWI during 2014/15 – a
decrease of 11% on the previous year.
The total harm comprised 80 major injuries, 1,470 minor injuries and seven cases of
reported shock/trauma.
Infrastructure worker safety at a glance
Risk in context (SRMv8.1)
Trend in harm
16
14
Weighted injuries
Fatalities
13.8
11.4 10.9
10.4
12
FWI
Other
accidental risk
(129.4 FWI;
93%)
Working on or
about the
running line
(10.1 FWI;
7%)
10
8
9.2
10.8
9.1
9.3
9.7
8.1
6
4
2
0
For more statistical analysis on infrastructure worker risk, see Chapter 5 of the ASPR.
Topics covered in this section:

Equipment design and modification

Track working

Machine use in possessions
66
9.2 Equipment design and modification
9.2.1 Runaway of a road rail vehicle and resulting collision in Queen
Street High Level Tunnel, Glasgow, 21 April 2013 (pub. 07/14)
At about 03:00, a road rail vehicle (RRV) ran away as it was being on-tracked on a sloping
section of line north of Glasgow Queen Street High Level Tunnel. The RRV ran through
the tunnel and struck two scaffolds that were being used for maintenance work on the
tunnel walls. A person working on one of the scaffolds was thrown to the ground and
suffered severe injuries to his shoulder. The track levelled out as the RRV ran into
Glasgow Queen Street station. After travelling around 1.1 miles, it stopped in Platform 5,
about 20 metres short of the buffer stop.
Causes
The RRV was a mobile elevating work platform that was manufactured for road use and
then converted by Rexquote Ltd for railway use. Its road wheels were intended to
provide braking in both road and rail modes. This was achieved in rail mode by holding
the road wheels against a hub extending from the rail wheels.
RAIB found the RRV’s design meant that, during a transition phase in the on-tracking
procedure, its road wheel brakes were ineffective because the vehicle was supported on
the rail wheels, but the road wheels were not touching the hubs.
Although instructed to follow a procedure which prevented this occurring
simultaneously at both ends of the RRV, the machine operator unintentionally put it into
this condition. He was (correctly) standing beside the RRV when it started to move, and
the control equipment was pulled from his hand before he could stop it.
The RRV was fitted with holding brakes acting directly on both rail wheels at one end of
the vehicle. These were intended to prevent a runaway if non-compliance with the
operating instructions meant that all road wheel brakes were ineffective. However,
shortcomings in Rexquote’s design, factory testing and specification of maintenance
activities meant the holding brake was insufficient to prevent the runaway. The lack of
an effective quality assurance system at Rexquote was an underlying factor.
Furthermore, the design of the holding brake was not reviewed when the RRV was
subject to the rail industry vehicle approval process, because provision of such a brake
was not required by Railway Industry Standards.
RAIB identified one learning point to remind the rail industry that the rail vehicle
approval process does not cover all aspects of rail vehicle performance:

Although the vehicle approvals process assesses compliance of plant design with
the relevant mandatory requirements of railway industry standards, it does not
67
provide assurance concerning all aspects of the design and manufacture of that
plant. For this reason, it’s important that designers, manufacturers and convertors
of plant for use on railway infrastructure apply established principles of
engineering safety management to the specification, design, manufacture and
testing of plant.
Reliance should not be placed on approvals bodies identifying that new plant may prove
to be unsafe when placed into service. Guidance on applying good practice in this field is
to be found in RSSB’s The principles of the safe management of engineering change.
What was done?
The contractor briefed machine operators about the circumstances of the accident, and
about the need to use the documented procedure when on-tracking all high ride RRVs.
Network Rail now audits engineering safety management systems of rail plant
manufacturers and converters who do not directly supply Network Rail, but whose
equipment is used its infrastructure.
The ORR served a Prohibition Notice on Network Rail, forbidding RRVs of the type
involved in the accident to be used where there is a risk from runaway during ontracking. Network Rail complied by imposing a ban until each machine had been shown
to meet improved pull test criteria.
Rexquote revised the relevant operation and maintenance manual to adopt the
increased pull test figures required by Network Rail. This revision also incorporated a
requirement to adjust the brake cylinder positions if required to meet the pull test
figures, along with instructions for carrying this out.
The operator involved (and others) tested the rail wheel holding brakes on all examples
of this type of RRV, adjusting them where necessary, to meet the revised performance
criteria. It also incorporated the higher 750 kg brake pull test into the three-monthly
maintenance checks on the RRV, instead of testing the rail wheel holding brakes
annually, as specified in the operation and maintenance manual.
Rexquote made available a redesigned rail wheel holding brake cylinder, with the same
maximum available brake retardation, but which used a longer spring assembly, and so
had a lower rate of reduction of brake retarding force with brake pad wear. These
cylinders were only fitted to RRVs that had been unable to meet the higher brake
retarding force pull tests with the existing cylinders.
Rexquote modified the rail wheel holding brake release interlock circuit to prevent the
possibility of the rail wheel holding brake being released when the road wheel braking is
not effective through contact with the rail wheel hubs.
Network Rail committed to extending the direct rail wheel braking retrofit programme
to cover all remaining high-ride RRVs, including the type involved in the accident. This
68
will replace the rail wheel holding brake on this type of RRV with a direct rail wheel
service brake, which will provide higher brake retardation.
RAIB also recommended:

Implementing a quality assurance process commensurate with good practice in
engineering safety management.

Extending processes for auditing the engineering management system of rail plant
suppliers, so that it includes auditing the engineering safety management
processes of all organisations manufacturing and/or converting rail plant likely to
be used on Network Rail managed infrastructure.

Reviewing the requirements for RRV lighting in standard RIS-1530-PLT, with the
objective of reducing the risk from RRVs running away without active lights.

Reviewing and improving the requirements and guidance for testing of RRV parking
brakes so that such tests reliably demonstrate that the brake will be effective in all
foreseeable operating conditions.
9.3 Track working
9.3.1 Track worker fatality at Newark North gate, 22 January 2014
(pub. 02.15)
At around 11:34, a track worker was struck by a passenger train as it approached
Newark North Gate station. He was part of a team of three carrying out the ultrasonic
inspection of two sets of points at Newark South Junction and was acting in the role of
lookout.
A few minutes before, the lookout and two colleagues arrived at the adjacent yard in a
van. One was in charge of carrying out the inspections; the other, the COSS, was in
overall charge of the team’s safety. They had planned to carry out the inspections on
lines that were still open to traffic in accordance with a pre-planned Safe System of
Work (SSoW). All three had many years of relevant experience in their respective roles
and were familiar with the work site.
Upon arrival, the lookout and tester proceeded to the track to start the inspection work;
the COSS remained in the van. Shortly after they had begun, the 10:08 King’s Cross–
Newark North Gate approached. It was due to stop in Platform 3, which required it to
negotiate two sets of crossovers. The train blew a warning horn, which the two staff on
site acknowledged before moving to the nominated place of safety.
However, just before the train moved onto the first crossover, the lookout turned away
from the train, walked towards the station and then out of the position of safety. He
moved to a position close to where he had been before the train approached, most
69
probably to check for trains approaching in the opposite direction, having decided that
the approaching train was proceeding straight into Platform 1. Although the train braked
and blew a second warning horn, the lookout did not turn until it was too late for him to
take evasive action.
Causes
RAIB confirmed the immediate cause of the accident to be that the lookout was in a
position on the track where he could be struck by the train. The report also lists these
causal factors:

The lookout probably moved into the path of the train to check for traffic
approaching on the up line, having incorrectly decided that the train was going into
Platform 1.

The lookout did not take evasive action when the train sounded the horn a second
time, probably because he believed the approaching train was on another line.

There was a breakdown in safety discipline and vigilance at the work site.
A possible underlying factor was that Network Rail’s process for assessment in the line
was not being followed as envisaged by the procedure, which may have led to a
deterioration in the safety attitude and discipline of individuals and teams going
unaddressed.
Although not linked to the cause of this accident, RAIB observed that the SSoW planning
at Doncaster Marshgate Depot was not being implemented in accordance with the
hierarchy of risk set out in the appropriate Network Rail standard. The intention of the
standard is that the planner should choose the safest practicable system of work.
However, in the case of the rail testing and lubrication teams, this was being disregarded
in favour of the most flexible and convenient system of work for the availability of staff
and equipment, or one that fitted with custom and practice.
The amount of ‘red zone’ working undertaken by the rail testing and lubrication teams
based in the Doncaster delivery unit was significantly higher than the target for the
LNE&EM route.
70
What was done?
During 2013, Network Rail began a major review
of the way work activities on the track are
controlled (the ‘Planning and Delivering Safe
Work’ programme) in order to improve track
worker safety. As a result, training for a new role
of Safe Work Leader (SWL) is currently being
introduced. This role is intended to provide
better safety leadership on site. Initial operation
is currently scheduled for mid-2015, in the East
Midlands region.
CIRAS activity – COSSs
and SWLs
Issue 52 of the CIRAS magazine
highlights a concern raised
about a lack of clarity
surrounding the proposed
changes to the COSS role.
Click here for more details.
All SWLs will have selection, training and
mentoring requirements which will include nontechnical skills and other safety leadership requirements. There will be three types of
SWL, ranging in responsibility from that of a current COSS (SWL1) to managing complex
work sites (SWL3).
New processes for planning and implementing work activities on the track will also be
introduced, including the use of an electronic work permit system, linked to electronic
maps. Network Rail also intends to introduce a Safe Task Leader (STL) role to replace the
COSS within engineering possessions.
On 10 February 2014, Network Rail issued a safety bulletin to all staff about this
accident. It encouraged internal discussion between staff about what could be done to
reduce risks associated with working on or near the line under lookout warning red zone
conditions and posed the following questions:

Safe System of Work – Can your work be done other than ‘red zone’ with lookouts?
What would you need to do to plan it at a higher level in the hierarchy next time?

Positioning site lookouts – When your COSS positions your site lookout, is
consideration given to positioning them in a permanent position of safety?

Staying vigilant – If you are doing routine work you have done numerous times
before in that location with the same safe system of work how do you make sure
that you stay focussed on the risks?

Recommencing work – When you stop work for a train to pass, do you always wait
for the permission of the COSS before you leave your position of safety?
On 2 and 3 April 2014, Network Rail held a national safety briefing event for all its track
maintenance staff, including COSSs and lookouts. A video reconstruction of the accident
was shown at the briefings, followed by discussion on the learning points and areas of
improvement.
71
Network Rail has indicated that it is progressing the following actions in its LNE&EM
route as a result of this accident:

A trial of a non-technical skills programme for lookouts, which aims to assess the
individual’s capabilities and aptitude for effectively and consistently performing the
role of lookout (completed in June 2014).

An assessment of briefing sessions for lookouts and COSSs on challenging the safe
system of work (completed in June 2014).

A trial of the use of trained safety mentors to develop the non-technical skills of
existing lookouts (ongoing).

The use of standard, fenced, lookout ‘stations’ at track locations such as junctions,
where lookouts are required on a frequent basis (ongoing).

The development of co-ordinated maintenance work programmes in junction areas
so that multiple maintenance activities are carried out during one visit rather than
several separate visits (ongoing).

Investment in the procurement of semi-automatic track warning systems to
provide early warning of approaching trains (to be started early 2015).
RAIB also recommended:

Briefing COSS/SWLs that they must be on site at all times, even when working with
experienced staff, and that they must provide a full site based safety briefing once
the safe system of work has been verified by them as being appropriate for the
conditions at the time of the work.

Re-briefing lookouts about not leaving the position of safety until the COSS has
given permission.

Actively monitoring the degree to which work site discipline is being maintained,
and take appropriate corrective action if any issues are found.

Investigating how best to maintain vigilance and safety discipline for cyclical and
repetitive tasks and implement any practicable measures into its working
procedures.

Introducing sufficient managerial supervision and audit checking to confirm that
the standards governing the safety of track workers are being correctly
implemented by its delivery units in the planning of SSoWs, particularly in those
areas where staff regularly work on lines that are still open to traffic.

Taking steps to strengthen any weaknesses it finds, including the re-training of staff
involved in planning safe systems of work.
72
9.4 Machine use in possessions
9.4.1 Engineering train collision at Kitchen Hill, near Penrith, 12
January 2014 (BULLETIN, pub. 10/14)
At 13:25, two freight trains, being operated in
connection with engineering work, were
involved in a collision at Kitchen Hill access
point on the West Coast Main Line, around five
kilometres north of Penrith station. Both trains
were within a work site contained inside an
engineering possession.
Train 6L43 was stationary and waiting for
permission to pass the work site marker boards
at the north end of the work site.
Train 6L42 was also preparing to leave the site
and was making a move, authorised by the Engineering Supervisor, towards its north
end. It ran into 6L43 at 17 mph, and pushed it forward around 20 metres.
Around 20 metres of rail fastenings unclipped as a result of the collision, allowing the
track gauge to spread and causing four wheelsets of 6L42 to derail.
Wagon buffers overrode at two locations: between the leading wagon on 6L42 and the
locomotive, and mid-way along 6L43. The driver of 6L42 jumped from his cab just before
the collision. He sustained abdominal injuries and was detained in hospital overnight.
No-one else was injured.
The West Coast Main Line remained closed between Penrith and Carlisle for recovery of
the trains and repair to the track until 14:04 on Monday 13 January 2014.
RAIB noted that this accident emphasises the importance of Recommendations 1b and
2ai from its investigation of the collision at Arley with respect to the rules for permitted
speeds in work sites (and possessions) and the related information given to drivers.
RAIB did not conduct any further investigation as it believed it unlikely that this would
lead to the identification of any new recommendations.
However, the accident raised the following learning points:

All movements of engineering trains (and on-track machines and plant) in work
sites (and possessions) should be made ‘at caution’ (meaning that trains should
always be capable of stopping in the distance ahead that the driver can see the line
to be clear).
73

Engineering Supervisors (and Persons in charge of Possessions – PICOPs) should be
aware of the risks from misunderstanding railway place names and that they take
the most appropriate measures to check that drivers (and machine controllers)
clearly understand the locations associated with movements in work sites (and
possessions).
Key issues raised by recent RAIB activity:

Equipment design and modification

Adequacy of the vehicle approvals process

SSoW – planning

On-site safety vigilance and discipline

Overspeeding

Worksite length

Unfamiliarity with area
9.5 Infrastructure Safety Liaison Group
ISLG – the Infrastructure Safety Liaison Group – represents the contractor community.
Part of its remit is to establish and implement arrangements to address the ‘duty of
cooperation’ across both the mainline and non-mainline rail networks, in order to help
infrastructure managers meet their responsibilities.
ISLG is self-governing, but the chair is a member of the System Safety Risk Group, which
allows it to contribute to the management of the whole railway.
ISLG also has two operational sub-groups: the Rail Infrastructure Assurance Group
(RIAG) and the Rail Industry Environmental Forum (RIEF). These are consultation groups
in their own right, but are also tasked by ISLG to address problems and advise on
recommendations for ISLG to take forward.
Integrated Health and Safety Improvement Plan
At the start of 2015, all maintenance contractors with an operating certificate agreed to
the development of the Rail Industry Health and Safety Strategy to help improve specific
risks and management capability. ISLG developed the integrated plan through its
members to:
74

Identify and maintain a common ‘Top 10’ risk profile associated with ISLG member
company operations and their interface with infrastructure managers and other
transport operators.

Develop suitable practical risk control measures to address identified risks in a
consistent way leading to continuous improvement of safety performance and
management arrangements.

Develop a suite of performance indicators to allow continued measurement of
positive or adverse trends in safety performance.

Develop a statistical model specific to infrastructure contractor operations.

Establish arrangements for learning from operational events across the UK,
European and worldwide rail industry.
ISLG will continually develop the Integrated Plan to provide a consistent approach to risk
reduction and improvements across the ISLG Infrastructure Contractor community.
Key objectives
The specific delivery objectives follow the output of ISLG workshops to identify the ‘Top
10’ key risk areas, which will be ‘managed’ using the following model:
75
The key risk areas determined by the workshops are listed below and will be validated
by RSSB data:

Competence

Fatigue

Management of Change

CDM Application

Road Driving

Communications

Competent Resource

Possession Management

Working at Height

Commercial Pressures / Delivery / Performance / Interface

Each risk area will be reviewed to determine the following:

Specific area identification

Detailed issue description

Current / required research / justification
76

Current practice identified

Lead organisation or group identified

Interfaces required

Dependencies to achieve success

Desired outputs

Test and evaluation criteria
10 Lessons learnt in 2014/15 – road driving risk
Many GB rail companies employ large road fleets and there is an increasing
understanding of the risks involved and the range of people that can be affected. A fatal
accident in June 2014 revealed fatigue as a factor, but also demonstrated gaps in our
safety management systems. The industry is now working to fill these gaps via a bespoke
Road Rail Steering Group.
10.1 Statistical overview
There were no workforce fatalities in train accidents during 2014/15, although there
were two workforce fatalities in two separate road driving incidents:

On 1 May 2014, an infrastructure worker was fatally injured on the A7 at Craighall
in Scotland, when the tractor he was driving was involved in a collision with a lorry.

On 10 February 2015, an office-based worker, who was travelling by motorbike to a
meeting near Waterloo – a location different from his normal place of work – was
involved in a road traffic accident.
77
Road driving safety at a glance
Risk in context (SRMv8.1)
Trend in harm
2.8
2.7
2014/15
Weighted injuries
Fatalities
3.0
2013/14
3.5
0.3
1.4
1.3
0.6
2010/11
0.3
2009/10
0.0
<0.1 <0.1 <0.1
2007/08
0.5
2008/09
1.0
2012/13
1.5
2011/12
2.0
2006/07
Risk to the
workforce
from driving
whilst on duty
(1.2 FWI; 1%)
2005/06
Other
accidental risk
(138.4 FWI;
99%)
FWI
2.5
For more statistical analysis on road driving risk, see Chapter 6 of the ASPR.
Many GB rail companies employ large road fleets to support their operations. Indeed,
there are over 75,000 vehicles out there being used to keep the network running.
The issues
Whether it’s getting train drivers into position for an early start, Mobile Operations
Managers (MOMs) out to an incident, maintenance teams between sites, or passengers
on to replacement buses, we’ve come to rely on what was once seen as our biggest
rival.
As an industry, we’ve been quick to learn from accidents, and quick to adopt designs
with safety in mind – from signalling, methods of working and management systems, to
recruiting the right people and training them in the optimum fashion.
Yet roads do not offer the same level of discipline, and the risk they offer car occupants
is 20 times greater than the risk to rail passengers.
It could be argued rail-related road driving risk has grown because of the large numbers
of people now required to travel across country to engineering worksites, to maintain
assets, or to respond rapidly to operational and / or safety incidents.
Either way, the issue prompted an industry steering group (including Network Rail,
Balfour Beatty, the ORR and the HSE) to do some research through RSSB to get a better
understanding of the subject.16
16
78
T997: Managing occupational road risk associated with road driver fatigue.
Network Rail also ran focus groups with road
vehicle drivers, while RSSB undertook a survey,
which revealed that fatigue was the major risk
to road vehicle drivers, that over 80% of staff
want more information on road safety and
that up to 100,000 rail workers could be at
risk.
What’s being done?
A Road Vehicle Project Steering Group was
established in 2013 to improve our
understanding of road driving risk to workers
and the business as a whole.
The Group includes members from Network
Rail, the TOCs, the ORR, the Railway Heritage
Trust, RoSPA, the trades unions and RSSB. Its
scope covers:

All rail duty holder employees,
contractors and subcontractors

The consequences of their actions (third
parties and damage)

All travel (including commuting)

All modes of transport (including cycling)

‘Door to Door’ taxi provision

Bus or taxi replacement services

NOT passenger journeys to and from
stations where not provided by a duty
holder
RED 35
RED 35 showed a young, keen rail
worker played by James Redmond
(Hollyoaks, Casualty) involved in a
road traffic accident. His line
manager, played by Ben Hull
(Hollyoaks, Crime Stories), knows he’s
a committed family man, a
conscientious team player for the
railway, so how could he possibly end
up involved in an accident?
The film helps increase understanding
of the ways a wide range of things
can line up so easily and tragically,
but also what can be done by
everyone – staff and managers alike –
to combat things like fatigue and
make sure the workforce get to work
and home again safely every day.
RED 35 is still available. See Opsweb
for more details.
The Group’s objectives for 2014/15 include the
development of reliable arrangements for
reporting and analysing Road Traffic Collision,
a survey of rail organisations in 2013 having
revealed 500 road traffic collisions, 100
injuries and 5 fatalities in one year, based on a
response representing about a third of the
industry. These figures are higher than the official data in SMIS.
79
Work is now under way to reform processes and
approaches to extend the reporting scope.
CIRAS activity – road
The Group also intends to provide a resource
centre on road driving risk to help rail managers
understand and share good practice; in addition, it
will evaluate and develop work-related road driving
principals for measuring SMS performance across
the rail industry.
vehicle driving
In March 2015, key experts in rail safety from
across the industry gathered at Network Rail’s
training centre in Westwood for a major
conference on road driving risk.
See section 6.4 for more
details.
During 2014/15, CIRAS activity
led to positive industry action
to combat road driving risk.
Organised by the Institution of Occupational Safety and Health (IOSH) and RSSB, the
conference saw first-hand the obstacles to collecting and sharing the right data, and
how the issues affect specific sectors.
The momentum on road driving safety is being harnessed through a dedicated steering
group in RSSB, which neatly connects directly to other existing sector groups with
representatives from across the railway, including train and freight operators,
infrastructure contractors, agency staff suppliers and plant suppliers.
Want to know more?
Check out the Road Rail Steering Group’s pages of the RSSB website.
Learning points:
80

Are you already briefing your staff on road driving risk?

Check your company’s processes for reporting road traffic incidents

Talk to your line manager, colleagues or safety representatives about any concerns
or experiences of your own, or talk to CIRAS in confidence

Find out if your company is tapping in to the work of the Road Driving Risk Steering
Group
11 Lessons learnt in 2014/15 – level crossings
The majority of the risk to members of the public arises from their own behaviour,
although this does not diminish the industry’s need to consider reasonably practicable
measures to deal with it. Level crossings are key interfaces between the public and the
railway. The industry has a duty to ensure that both the signs and controls in use are fit
for purpose and that its operations allow the users to understand and follow them.
11.1 Statistical overview
Excluding suicides, eight pedestrians, including one cyclist, and two road vehicle
occupants died in accidents at level crossings in 2014/15. There were five major injuries,
52 reported minor injuries and 27 cases of shock or trauma. This equated to a total FWI
of 10.7, which is higher than the 2013/14 figure and just above the average for the last
ten years.
Furthermore, there were seven collisions between trains and road vehicles at level
crossings, which is three fewer than recorded in 2013/14. There has been an average of
13.7 accidents per year since 2005/06; there is evidence that the underlying rate of
collisions at level crossings has reduced over this period.
Level crossing safety at a glance
Risk in context (SRMv8.1)
Trend in harm
16
14
13.6
13.2
14.0
Weighted injuries
Fatalities
10.7
9.9
8.8
11.0 10.9
12
7.4
8
5.2
6
4
2
2014/15
2013/14
2012/13
2011/12
2010/11
2009/10
2008/09
2007/08
0
2006/07
Level crossing
risk (11.4 FWI;
8%)
2005/06
Other
accidental risk
(128.2 FWI;
92%)
FWI
10
For more statistical analysis on public risk, see Chapter 8 of the ASPR.
Topics covered in this section:

Level crossings

Road vehicle incursions
81
11.2 Level crossings
The majority of the risk to members of the public arises from their own behaviour,
although this does not diminish the industry’s need to consider reasonably practicable
measures to deal with it. Level crossings are key interfaces between the public and the
railway. The industry has a duty to ensure that both the signs and controls in use are fit
for purpose and that its operations allow the users to understand and follow them.
Investigations revealed issues around crossing controls, closure sequences, signage and
crossing instructions.
11.2.1 Near miss at Llandovery level crossing, Carmarthenshire, 6
June 2013 (pub. 05/14)
At around 05:56, a passenger train traversed Llandovery level crossing while it was open
to road traffic. As the train approached, a van drove over immediately in front of it. A
witness working in a nearby garage saw what happened and reported it to the police.
Causes
RAIB ascertained that the crossing was still open to road traffic because the guard had
not operated the controls on the station platform – possibly because he lost
concentration, possibly because he had been distracted by a person walking a dog,
whom he thought may trespass on the railway. The driver didn’t notice that the crossing
had not been operated because he may have been distracted by the same person, or
distracted by an indicator light in the cab. RAIB also note that the positioning of the
operating equipment at Llandovery was sub-optimal.
RAIB identified that an opportunity to integrate the operation of Llandovery level
crossing into the signalling arrangements (which would have prevented this incident)
was missed when signalling works were planned and commissioned between 2007 and
2010. RAIB also identified that there was no formalised method of work for train
operations at Llandovery.
What was done?
Network Rail placed a ‘car stop’ marker eight metres on the approach to the level
crossing stop board at Llandovery station to improve the visibility of the white flashing
light associated with the level crossing stop board.
RAIB also recommended:

Identifying all locations where train crew carry out operational activities (eg token
exchange and level crossing operation) in addition to train dispatch, and develop
risk assessed methods of work for each.

Reviewing the positioning of platform equipment and signage used by train crew at
unmanned stations.
82

Improving the processes for designing new and altered signalling (including a
consideration of reasonable opportunities to improve the control of risk beyond
the immediate scope of the proposed works, including identifying where operator
errors could lead to unsafe conditions).

Reviewing operational risk management arrangements, with a view to improving
the process for assessing the risk associated with station duties (eg the application
of route risk assessments).

Reviewing the current arrangements for providing an indication to the train driver
of the status of the crossing at Llandovery.

Reviewing and improving the training and guidance given to its duty control
managers on the steps to be taken when train crew are involved in a serious
operating incident where their actions directly contributed to it.
11.2.2 Near miss at Butterswood level crossing, North Lincolnshire,
25 June 2013 (pub. 06/14)
At around 07:35, a passenger train was involved in a near-miss with a car on
Butterswood level crossing in North Lincolnshire. The train passed over the crossing with
the barriers in the raised position and the road traffic signals extinguished. There were
no injuries or damage to rolling stock or infrastructure.
Normally, the train’s approach would have initiated the closure sequence automatically.
However, the crossing was not working normally, as the power supply had been
interrupted. The crossing was of a type where train drivers are required to check that it
is not obstructed as they approach and that it has operated correctly. A flashing light is
provided for this purpose, with a flashing white light displayed if it has correctly closed
against road users, and a flashing red light shown at all other times (including when the
crossing has failed to close on the approach of a train). The driver of the train did not
notice until it was too late to stop that the flashing light was indicating that the crossing
was not working normally, and was still open to road traffic.
Causes
RAIB found that the train driver expected that the crossing would operate normally as
the train approached and that he had not focused his attention on the flashing light at
the point where he needed to confirm that the crossing had operated correctly for the
passage of his train. Although the crossing had probably failed around nine hours before
the incident, the fact of its failure was not known to any railway staff.
The investigation also found that the crossing was not protected with AWS and that the
maintenance arrangements at the crossing were not effective in ensuring the reliable
83
performance of the equipment. In addition, the train operator’s briefing material did not
clearly explain to drivers their role in respect of failures at this type of interface.
RAIB identified the following learning points for the railway industry:

Railway Group Standard GE/RT8075 (AWS and TPWS interface requirements) states
that AWS shall be fitted on all signalled lines, except where a train protection
system provides a level of protection equal to, or better than, that provided by
AWS and TPWS. This investigation identified that AWS equipment had not been
provided at the level crossing warning board as required by Railway Group
Standards and that no derogation had been granted. The industry must comply and
risk assess any proposals for non-fitment of AWS on any line of route.

The records associated with the testing of the equipment at Butterswood crossing
were not completed in accordance with the required process. This meant the
asset’s poor condition was not formally recorded and an opportunity to identify
repeat failures was lost. Where an asset fails to meet the required test or
inspection criteria, it is important that the matter is recorded in accordance with
company procedures.

When storing replacement batteries for safety related equipment their age should
be recorded and their condition monitored to ensure that they are in adequate
condition when they are eventually brought into use.

When carrying out investigations into accidents and incidents, it is important that
those investigations involve people with the required technical expertise so that
safety lessons can be effectively identified, and recommendations addressed
appropriately.
What was done?
The driver managers at the depot concerned now analyse the on-train data recorded
(OTDR) fitted to Northern Rail’s Class 153s following training in the downloading and
analysis of data.
The equipment at Butterswood has also been replaced with a new system.
Furthermore, Network Rail is installing AWS on the approach to Butterswood level
crossing as part of its upgrade of the North Lincolnshire routes.
RAIB also recommended:

Conducting a human factors and technical review of the indicators provided on the
approach to automatic locally monitored level crossings, and evaluating alternative
means (eg audible and visual) of indicating to train drivers that the level crossing
has not operated as intended.

Reviewing the arrangements in place at all types of automatic locally monitored
level crossings, and making improvements to the reliability of those crossings.
84

Evaluating the practicality of remote condition monitoring of the power supply
system, and key sub-systems whose failure can have the same effect as a loss of
power supply, at all locally monitored level crossings, so that prompt action can be
taken to manage the failure.

Reviewing and enhancing briefing techniques and guidance material for train
drivers to explain the role of the driver at locally monitored crossings (inter alia).
11.2.3 Fatal accident at Barratt’s Lane No.2 footpath crossing,
Attenborough, Nottingham, 26 October 2013 (pub. 08/14)
At 14:48, a pedestrian was struck and fatally injured by a train on Barratt’s Lane No.2
footpath crossing, at Attenborough near Nottingham.
The train was travelling from Nottingham towards Birmingham. At the same time, a
Nottingham train was slowly approaching the crossing from the other direction. It is
likely that the pedestrian had concentrated her attention on the London train and did
not notice the train approaching from the Nottingham direction.
Causes
Both trains were fitted with forward-facing CCTV. The recording from the Nottingham
train showed that the pedestrian approached the crossing and waited at the gate for 17
seconds before opening it; she started to cross the line nine seconds later (the train was
stopped at a red signal for part of this time). It is most likely that, having seen the
London train stopped at the signal, she waited until she had determined that the train
was not moving before deciding to cross the line. The sighting distances in both
directions were adequate.
Network Rail had assessed the risk at the crossing, in accordance with its standard
procedures, and, because the rating was relatively high, discussed the options for
reducing this risk at a meeting with the highway authority. The chosen option was to
divert the footpath and close the crossing. This had not been implemented at the time
of the accident as the route of the proposed diversion was obstructed by an equipment
room. The room contained signalling equipment that did not become redundant until
completion of the Nottingham station resignalling project at the end of August 2013.
The equipment room was demolished and the footpath diverted after the accident.
RAIB identified the following key learning point for the railway industry:

For double track lines, kissing gates arranged with the hinge on the right-hand side
encourage footpath users to face towards the oncoming traffic on the nearest line
as they exit from the gate. This is particularly relevant where the gate is close to
the track.
85
What was done?
Network Rail applied to Nottinghamshire County Council for a formal diversion order for
Beeston footpath 66 on 22 November 2013. An order was duly made on 17 January
2014 and confirmed on 28 March 2014. This diversion was on the originally proposed
route via Attenborough Lane level crossing. Construction of the diversion was
completed in March 2014.
RSSB conducted research into the causes of pedestrian accidents at level crossings
(project T984), which included findings relevant to the design and installation of kissing
gates. It identified a design principle that the alignment of a kissing gate (or chicane)
should, if practicable, enable the user to look in both directions and encourage the user
to look in the direction of oncoming trains on the nearest line.
This information has also been included in the recent update of the level crossing risk
management toolkit for consideration as part of the ongoing assessment of risk at
footpath crossings. These points were incorporated into the research report at the
suggestion of RAIB and by agreement with all concerned.
11.2.4 Collision at Jetty Avenue level crossing, 14 July 2013 (pub.
12/14)
During the early evening, a passenger train approaching Woodbridge station in Suffolk
struck a car at Jetty Avenue user worked level crossing (UWC). The accident occurred in
daylight and at low speed. The train was not derailed, but the car driver suffered minor
injuries.
The car driver was using the level crossing to access a private boatyard situated between
the railway and the River Deben. He was a volunteer, assisting in removing equipment
following a local regatta which had been held partly on land owned by the boatyard
earlier in the day. The car driver had used the crossing on previous occasions, but had
not been briefed on its use.
There were no telephones or warning lights at the crossing so safe use depended on
vehicle drivers looking for approaching trains.
The car driver, who was an occasional user of the level crossing, normally relied on
checking for trains by looking up and down the railway when swinging open the
vehicular gates on foot. He did this because he was aware that his view of the railway
would be obscured as he returned to the car and drove it towards the crossing.
A curve in the railway meant that the train was not visible to the car driver when he was
at the crossing, and could only be seen from this location after the driver had begun to
return to his car. The driver did not become aware of the train until he had driven his car
into its path.
86
Causes
RAIB’s investigation found that instructions given to car drivers using this, and similar,
level crossings were inadequate. It also found that Network Rail’s method for ensuring
that vehicle drivers have an adequate view of approaching trains was incompatible with
the characteristics of both the car involved in the accident and many of the vehicles
expected to use crossings of this type. Furthermore, RAIB notes that the train driver did
not sound the horn at the whistle board. If he had done so, it is possible this would have
prevented the accident.
RAIB also listed the following underlying causes:

Network Rail standards did not adequately specify the locations at which sighting
distance should be measured.

Network Rail’s staff may have sometimes incorrectly interpreted the ORR’s related
guidance to mean that a three-metre decision point is normally adequate at UWCs.
What was done?
Telephones were provided at Jetty Avenue, and the signage has been changed to
require all road vehicle drivers to telephone the crossing operator before using the
crossing.
From mid-2012, the management arrangements for level crossings changed. The
position of Level Crossing Manager was introduced to replace the Operations Risk
Control Co-ordinator (ORCC). This role brings together the duties of undertaking site
visits and managing the crossing, and is intended to clarify responsibility and eliminate
problems caused by poor communication between different departments within
Network Rail.
Level crossing inspection checklists have been replaced by an ‘app’ for use on a
smartphone.
The Route Level Crossing Manager informed RAIB that additional resources are now
being provided to allow work identified to be done.
Anglia route has introduced an events register to enable it to identify recurring events,
like the regatta, which may increase the road traffic over certain crossings. It is also
trialling the provision of ‘sighting marker’ signs at five crossings, including Jetty Avenue.
These markers are installed adjacent to the line at the required sighting distance + 20%
to assist with checking sighting distances from the decision points.
The Anglia route has restarted its programme for installing level crossing information
signs (showing the crossing name and telephone details) at UWCs where such signs had
87
not been provided. Note that this will not include Jetty Avenue UWC, which now has
telephones.
RAIB wrote to Network Rail, the ORR and Northern Ireland Railways to inform them that
sighting criteria based on a three-metre decision point at UWCs are not sufficient to
provide a safe crossing method for many cars in the UK.
RSSB has researched signs at private level crossings under project T983 (similar to the
T756 project on signs/signals at public crossings). The findings from T983 were published
in early 2015. (RSSB has also undertaken other research projects relating to level
crossings which, although not targeted at vehicular use of UWCs, include findings which
could assist in identifying measures which would improve the safety of such use.)
Network Rail is developing relatively low-cost systems to give crossing users audible and
visual warnings of an approaching train. The technologies include fitting GPS equipment
to trains and mounting train detection devices (eg axle counters, treadles and radar)
near crossings.
RAIB also recommended:

Implementing a time-bound plan for the re-assessment of the sighting of
approaching trains at all UWCs where safe use depends on vehicle drivers sighting
approaching trains. The time-bound plan should also cover implementation of any
mitigation needed to permit safe use of such crossings.

Commissioning research into ways of improving the safety of UWCs where
vehicular users are reliant on sight to detect the approach of trains.

Identifying the need for any modification to the legal requirements relating to level
crossing signage requirements, and make suitable representations to government
that this be done.

Reviewing and, if necessary, modifying its processes so that staff checking level
crossing signage have a practical and easily used means of establishing the signage
required at each crossing they are inspecting.

Reviewing and, if necessary, amending the criteria used to calculate crossing times
with reference to vehicle speed, the time taken to reach a decision when to start
crossing and vehicle length.

Providing enhanced guidance to remind duty holders that, when determining the
position of decision points at UWCs, they must take due account of the
characteristics of vehicles likely to use the crossing and recognise that a minimum
dimension of three metres from the nearest rail is insufficient for most vehicles.
88
Key issues raised by recent RAIB activity:

Equipment failure

Driver error

User error

Sighting

Crossing instructions
11.2.5 Other level crossing initiatives
As the industry recognises that level crossings represent a unique interface between the
railway and other infrastructures (highways and public or private rights of way), it has
formed a Level Crossing Strategy Group, reporting to SSRG, which includes members
from the following companies and bodies:

Network Rail

Train operators (two passenger, one freight)

RSSB

British Transport Police

Department for Transport (1 member)

ORR

Trades unions

ADEPT (local highway authorities)
The purpose of the group is to understand and review the risks within its scope, by:

Monitoring the effectiveness of current control arrangements

Identifying and sponsor improvement opportunities including research

Identifying, seeking learning from, and promoting good practices

Reviewing progress against objectives and targets

Facilitating cooperation

Reviewing the effectiveness of risk mitigations

Responding to requests from SSRG and other cooperative forums
RSSB published project T984: Research into the causes of pedestrian accidents at level
crossings and potential solutions on behalf of the industry. An early output was advice
on decision points, which may be found here: LINK
89
Network Rail has put a level crossing risk reduction programme in place, and has met
the targeted reduction in level crossing risk of 25% over the course of CP4. A further
substantial safety improvement in is planned for Control Period 5 (CP5), which runs from
April 2014 to March 2019.
Furthermore:

The 100+ dedicated Level Crossing Managers and Route Level Crossing Managers
continue to support asset inspections, as well as data collection for risk assessment
and modelling. Their role includes building relationships with authorised users and
in the wider local community in a bid to understand local risks.

The programme of level crossing closures closed 804 crossings and downgraded a
further 48 in CP4. A total of 256 crossings are scheduled be closed during CP5.
There are 383 crossings planned for renewal over the period and an additional 360
wig wags upgraded to LED lights.

High-risk footpath crossings are being replaced by footbridges, in line with a policy
decision to remove the need for pedestrians to cross high-speed main lines
unprotected by barriers. Across the network, nine modular footbridges were
erected in 2013/14, allowing nine crossings to be closed.

Work has continued for the development of red light safety cameras (RLSE) with
number plate recognition technology. The aim is to deter users from traversing the
crossing when they are not permitted to do so. One system has been ‘Home Office
Type Approved’ and has been installed at 10 crossings. A further two RLSE systems
are currently under review.

Fifteen mobile safety vehicles continue to operate. The vehicles are staffed by BTP
and have detected and prosecuted more than 1,500 motorists responsible for red
light violations since 2012.

Audible warning devices have been developed to provide a spoken warning to
pedestrians when ‘another train is approaching’. They have been installed at 117
level crossings and aim to ensure that pedestrians understand that it is not safe to
cross when the crossing sequence continues after a first train has entered, or
passed through, the station.

Network Rail continues to develop a range of technologies to locate trains in long
signal sections, including GPS and sound wave based solutions. A new overlay
system, Vamos, is currently operating in ‘shadow’ trial mode and is awaiting
product approval in 2015/16.

Power operated gate openers are being installed at 82 user-worked crossings.
These devices avoid the need for users to leave their vehicles and make multiple
traverses over a crossing on foot. Training is to be provided to ensure users are
aware of how they work.
90

Half-barrier overlays (AOCL+B) have been installed at 64 AOCL crossings. A further
two installations will be completed in 2015. The provision of additional half-barriers
enhances user safety.

Covtec technology which uses way side train horns which trigger on the approach
of a train. Radar picks up the train in the vicinity and provides a localised warning at
the crossing.
Research and development
Road-Rail interface research is conducted in nine main areas:

Understanding the risk at level crossings to enable prioritisation of remedial
actions;

Identifying and sharing good practice in Britain and overseas to facilitate the
adoption of appropriate solutions;

Identifying new technical and operational solutions to prevent errors and misuse of
crossings;

Understanding the costs of level crossings and the benefits of adopting alternatives
to optimise societal benefits;

Working in collaboration with highway and planning authorities to design out
safety risk and reduce the overall cost to society;

Understanding the needs of vulnerable users at level crossings to facilitate social
inclusion;

Review and overhaul of the legislative framework for level crossings to identify
legal requirements and consolidate disparate regulations;

Research into bridge strikes and vehicle incursions; and

Research to support inquiry recommendations,
government and regulatory policies, proposed and new
legislation.
RSSB’s Road-rail interface safety guide pulls together
summaries of this research in one document to provide a
useful resource to assist industry with accessing relevant
information. A new edition is in preparation.
11.3 Road vehicle incursions
On 28 February 2001, a Land Rover and trailer came off the M62 just before a bridge
over the East Coast Main Line near Great Heck, North Yorkshire. The Land Rover and
trailer went down the embankment and came to rest obstructing the southbound line.
The vehicle was then struck by an express passenger train travelling close to 125 mph.
91
The leading vehicle of the train was derailed, though the consist continued for some
distance before running into the path of, and colliding with, a northbound freight. Ten
people were killed and 82 required hospital treatment.
The driver of the Land Rover was subsequently convicted of ten counts of causing death
by dangerous driving and sentenced to five years’ imprisonment.
The railway industry conducted a Formal Inquiry into the accident, resulting in 27
recommendations (all of which have been closed out).
The (then) Deputy Prime Minister asked the Health and Safety Commission (HSC) to
examine the obstruction of railway lines by road vehicles. He also asked the Highways
Agency (HA) to review its standards for near side road safety barriers. Both the HSC and
HA reports presented their findings in February 2002.
The Department for Transport (DfT) produced a report, Managing the accidental
obstruction of the railway by road vehicles, in response to the recommendations in
these two documents. It set out the steps to be taken jointly by railway infrastructure
authorities and highway authorities to manage the risk from the accidental incursion of
road vehicles onto the railway. It also included a protocol for apportioning responsibility
and costs of mitigation measures.
Approximately 11,000 potential public road to rail incursion sites have so far been
identified, of which:

9,400 were eliminated as ‘low risk’ after assessment;

1,250 were identified as ‘medium’ or ‘high risk’, of which:

450 have been remediated;

150 have been deemed to be ‘remediation not practicable’; and

650 sites are still awaiting remediation.
(In addition, 400 assessments are still outstanding.)
The Level Crossing Strategy Group seeks to track progress with the mitigation plan and
help facilitate solutions. Since the first report in March 2010, the number of sites
requiring initial assessment has been reduced from 400 to 80. However, new sites
continue to be identified (although they are less likely to be high scoring and in need of
mitigation).
There were 57 road vehicle incursions in 2014/15. This is below the ten-year average of
61, but an increase on the previous year’s total of 43.
Most of these vehicles accessed railway property via fences, often as a result of a road
traffic accident. Those incidents categorised with a level crossing being the access point
relate to road vehicles which have moved off the crossing, along the line, to some
extent.
92
None of the incursions in 2014/15 resulted in a train accident. However, an incident in
October 2013 highlighted the need to retain focus in this area.
11.3.1 Road vehicle incursion at Aspatria, 26 October 2013 (pub.
06/14)
At 10:01, an unattended commercial vehicle ran away on the B5299 at Aspatria. It
crossed the main A596, broke through a wooden fence and rolled down the side of a
cutting onto the railway. Although a passenger train was approaching Aspatria at the
same time, prompt action by those concerned resulted in this train being stopped about
1.5 miles from the incident site. There was therefore no collision with the commercial
vehicle on the track, and none of the passengers or crew on the train were actually put
at risk.
Causes
The management of road vehicle incursions onto the railway is described in guidance
published by the Department for Transport (DfT). RAIB found that the guidance does not
explain how to assess the risk of a vehicle that has lost control on a side road (eg a
runaway on a side road with a downhill gradient towards the railway). Also, the
guidance does not describe how this risk should be combined with the risk of road
vehicle incursion from the corresponding main road to give an overall risk ranking score.
What was done?
At the time of publication, Network Rail was working on the design of a safety barrier to
be implemented to reduce the risk of vehicle incursion opposite the B5299/A596
junction. The funding for this will be shared between Network Rail and the highway
authority.
RAIB also recommended:

Reviewing and amending the current guidance (Managing the accidental
obstruction of the railway by road vehicles, 2003), so that it adequately takes into
account in the risk ranking process for neighbouring sites the risk of road vehicles
on side roads, including those that are unattended, running downhill onto a
railway.

Using the updated guidance to implement a time-bound plan to review the risk
ranking scores for sites where there is a significant risk from side roads, in
particular with respect to road vehicles running downhill onto a railway. Additional
risk mitigation measures justified by increased risk ranking scores should be
considered and implemented.
93
12 Lessons learnt – beyond the boundary fence
Our industry does not exist in a vacuum. Clearly,
there will be accidents and incidents in other
sectors that carry lessons from which we can all
benefit.
It is widely recognised, for example, that the Baker
Panel review into the BP Texas City oil refinery
disaster of 2005 and the Haddon-Cave report on
the Nimrod accident of 2006 contain useful lessons
for the rail industry. Indeed, the Baker review led us
to take a closer look at safety performance
indicators and the proactive management of risk.
Did you know?
The Federal Aviation
Administration runs a website
devoted to the lessons learned
from aircraft accidents across
the world.
Accidents can be selected
according to three
‘perspectives’: airplane life
cycle, accident threat
categories, and accident
common themes.
However, learning is a line signalled for bidirectional running, as the Buncefield Oil Depot
explosion of December 2005 demonstrated.
Follow this link for details.
Included among the investigation report’s
recommendations was a proposal that the oil
industry ‘develop incident databases that can be
shared across the entire sector, subject to data protection and other legal
requirements’. It added that examples ‘exist of effective voluntary systems that could
provide suitable models’, naming RSSB’s National Incident Reporting System, NIROnline, as a worthy model.
This shows both how rail learned from oil and oil learned from rail. It’s something that
other industries would do well to consider. Had the Environment Agency (EA) been
aware of the Hatfield derailment of October 2000, for example, it might have
understood the implications that ‘maintenance holidays’ can have on infrastructure.
Railtrack stepped back on track repairs and gauge corner cracking was the result; the EA
failed to dredge the Rivers Parret and Tone and flooding was the result during the
storms of early 2014.
Yet the key to success is not only about sharing lessons, but also best practice and ideas
– that is, we need to learn just as much from what we do right as what we’ve done
wrong.
Occasionally, RSSB will produce an ad hoc report to highlight issues raised by inquiries
into non-rail events to promote pan-industry learning and offer suggestions for how
learning procedures might be finessed.
All may be found on the LOE resources page of the RSSB website.
94
Glossary
For a full list of definitions, see the Annual Safety Performance Report.
AHB
automatic half-barrier crossing
AOCL
automatic open crossing, locally monitored
ASPR
Annual Safety Performance Report
BTP
British Transport Police
CCS
Close call system
CIRAS Confidential Incident Reporting and Analysis System
COSS
controller of site safety
ERA
European Railway Agency
ERTMS European Rail Traffic Management System
FWI
fatalities and weighted injuries
GB
Great Britain
GSM-R Global System for Mobile communications – Railway
HSE
Health & Safety Executive
IFCS
Incident Factor Classification System
ISLG
Infrastructure Safety Liaison Group
LOE
Learning from operational experience
LUL
London Underground Ltd
MOM mobile operations manager
NRMI Network Rail managed infrastructure
OFG
Operations Focus Group
OHLE
overhead line equipment
ORR
Office of Rail Regulation (now Office of Rail and Road)
PHRTA potentially higher-risk train accident
PICOP person in charge of possession
PIM
Precursor Indicator Model
95
PTI
platform train interface
RAIB
Rail Accident Investigation Branch
RGS
Railway Group Standard
RIDDOR Reporting of Injuries, Diseases and Dangerous Occurrences Regulations 1995
RRV
road–rail vehicle
RSSB
Rail Safety and Standards Board
SMIS
Safety Management Information System
SPAD
Signal Passed At Danger (without authority)
TOC
train operating company
TORG Train Operations Risk Group
TPWS train protection and warning system
UK
United Kingdom of Great Britain and Northern Ireland
UWC
user-worked crossing
96
Email enquirydesk@rssb.co.uk
Tel
+44 (0) 20 3142 5400
Twitter@RSSB_rail
Webwww.rssb.co.uk
Rail Safety and Standards Board
The Helicon
One South Place
London EC2M 2RB