July 2015
This is a collation of some of the world’s railway formal inquiry reports. It includes a brief incident
synopsis, along with the main causes and recommendations from each investigation.
Readers may find some of the actions and recommendations useful to their own operations.
Co-ordinated by Greg Morse, Operational Feedback Specialist, RSSB
Contents: (Click to navigate)
US: Fatal shunting accident at Colorado Springs,
8 October 2014
UK: Near miss with track workers near Hest
Bank, 22 September 2014
Latvia: Fatal shunting accident at Skirotava, 7
September 2014
UK: Electrical arcing and fire and parting of the
overhead line at Walkergate, 11 August 2014
Walkergate
Some of the key issues raised and/or suggested by the stories in this edition:
Lookout sighting
Possible fatigue
Failure to implement recommendations
Shunting – clearance between wagons
Coupling issues
Distraction
Hump shunting
Operations during construction work
Distraction
Safety management system (shared risks)
Knowledge sharing
Produced by RSSB
Author: Dr Greg Morse
Email address: Greg.Morse@rssb.co.uk
14 July
US: Fatal shunting accident at Colorado Springs, 8 October 2014
For the full report, click here: LINK
At 12:30 (local time) on 8 October 2014, a Burlington Northern Santa Fe (BNSF) guard died while a
freight train was shunting a private siding on the Pikes Peak Subdivision, near Colorado Springs.
On the day before the accident, the driver and guard had tried to shunt four wagons on private
sidings owned by Thompson and Sitz Construction, Inc. They were unable to do so because the
reception line was occupied by track maintenance equipment that could not be moved. After
consulting with their supervisor, the crew secured the wagons on an adjacent siding and were driven
to Denver, where they went off duty. Before departing, they walked the tracks, noted the positions
of the wagons and prepared a plan for how they were going to perform the shunt the next day.
At 08:00 the following morning, the driver, guard and a shunter went on duty and were driven back
to Colorado Springs.
On arrival, the driver prepared the two locomotives and the guard and shunter released the
handbrakes on the four wagons left on the siding the day before.
The guard coupled the four loaded wagons
and two locomotives to six empty flatbeds
that were on tracks 813 and 815. She then
saw all the vehicles off track 815 and
changed the points for track 813. The
wagons were then propelled to the end of
that track.
The guard uncoupled three of the wagons
and had the driver pull forward, leaving
them on track 813. According to the
shunter, the handbrakes were applied before the move was made and were never released.
The guard told the driver to stop when the wagons passed the connection to track 816. She then set
the points for track 816 in preparation for coupling the train to two empty flatbeds thereon. The
guard told the driver to propel back towards track 816. While positioned on the north side of the two
empty flatbeds, she tried several times to couple the train to the wagons on track 816, but was
unsuccessful. The driver told her that road traffic had backed up on a level crossing blocked by their
train and that he wanted to clear it. The guard agreed and the driver pulled the entire train out of the
siding.
After the road traffic had cleared, the train returned. The guard, now on the south side of the train,
but still north of track 813, continued trying to couple up to two flat railroad cars on track 816.
The shunter told the National Transportation Safety Board (NTSB) that the guard celebrated making
the coupling: ‘She let out a big hurrah and exclaimed “you know, we got it. We coupled the cars”.’
She then instructed the driver to pull the wagons out of track 816. Soon after starting the movement,
the driver heard a strange noise on the radio and stopped.
The guard was between track 816 and track 813 when the wagons were finally coupled. When the
train pulled out of track 816, she became caught between them. The shunter told the driver to stop
Produced by RSSB
Author: Dr Greg Morse
Email address: Greg.Morse@rssb.co.uk
the movement. On finding the guard, he instructed the driver to back up gently, in order to try to
free his colleague. The shunter realised this was not working and stopped the movement altogether.
The driver radioed an emergency call to the signaller. Meanwhile, a nearby Thompson and Sitz
employee, who also witnessed the event, called the emergency services. By the time the ambulance
arrived, the guard had died.
The NTSB determined the probable cause of the accident to be the guard leaving wagons on track
813 with insufficient clearance to the adjacent track and then instructing the driver to move the
wagons before moving to a position of safety. The guard’s focus on successfully coupling the wagons
was a contributing factor.
Recommendations

None listed.
Back to top
16 July
UK: Near miss with track workers near Hest Bank, 22 September 2014
For the full report, click here: LINK
During the afternoon of Monday 22 September 2014, a group of nine track workers repairing a
section of the West Coast Main Line south of Hest Bank level crossing narrowly avoided being struck
by a southbound passenger train. Their site of work was located on a bend that restricted the
visibility of approaching trains. Warning of approaching trains was intended to be given by lookouts,
located remotely with good visibility of the track, using a radio-based lookout operated warning
system (LOWS). The system had been working normally prior to the incident, but the workgroup did
not receive a warning for the incident train.
The track workers saw the approaching train with just enough time to clear the track before it passed
them at 98 mph. They were shaken, but not physically injured. All work on the site was stopped for
the remainder of the shift.
The incident was caused because a lookout did not give a warning, either because he operated the
wrong switch on his radio transmitter by mistake, or because he forgot about the need to send a
warning during an intended delay period between seeing the train and operating the switches. This
delay was because he was positioned on a long section of straight track and could see approaching
trains for significantly longer than the time required for the workgroup to move to a position of
safety.
It is also probable that the lookout’s vigilance had degraded as he had been working continuously for
almost two hours.
A previous RAIB recommendation – relating to managing extended warning times due to extended
sighting distances (see Cheshunt Junction Recommendation 2) – had not been implemented due to
administrative errors. It is possible that the incident would have been avoided had this been dealt
with more effectively. Incorporating recent human factors research into the lookout’s role may too
have prevented the incident.
Although unrelated, the RAIB identified that a safety-critical element of the LOWS circuitry was not
subject to routine testing.
Produced by RSSB
Author: Dr Greg Morse
Email address: Greg.Morse@rssb.co.uk
Action taken
Following this incident, Network Rail’s Lancashire & Cumbria delivery unit prohibited further use of
the LOWS equipment involved and trialled a new version which gives improved user-feedback and
has the vigilance switch replaced by a movement sensor.
Network Rail has also introduced ‘non-technical skills’ training for lookouts as an e-learning course in
July 2014, with a requirement for all lookouts to complete it by July 2015. The training includes
modules on ‘attention management’, which cover the ability to remain alert and focussed, the ability
to manage distractions, and the ability to keep an awareness of the overall situation. The training
includes scenario-based discussions. The training mentions the 30-minute vigilance degradation, but
other areas of the training recommend lookouts to ask for breaks every 1.5 hours and if they are
aware their vigilance is degrading. RAIB notes that this is inconsistent with the two-hour interval
suggested by Network Rail’s ergonomics team and included in COSS training. It is also inconsistent
with the 30-minute period over which significant vigilance degradation will occur, according to most
research.
RAIB has concluded, on the basis of existing research, that vigilance degradation is a ‘hard-wired’ part
of the human condition and is not a ‘skill’ that can be trained because:

All individuals are susceptible; and

Self-monitoring for vigilance degradation is unreliable16.
Although training is useful to raise awareness of these risks, it is unlikely to mitigate them
completely. This means that the vigilance problem is best addressed through design of the task or
equipment; issues addressed by the recommendations in this report.
Network Rail modified its recommendation handling process in October 2014 because it recognised
the previous process required improvement. This involved a revision to its Investigations Handbook
which now states:
‘Recommendations are made to improve our safety performance so should only be closed when
either:
a) the risk that the recommendation seeks to address has been eliminated; or
b) improved risk mitigation has been implemented.
When considering if the above have been achieved the wording of the recommendation and its intent
should be considered.’
All closure statements are now subject to an independent review by Network Rail’s Corporate
Investigation and Assurance Manager to check that they meet the criteria above.
Further actions may be found in the full report.
Recommendations

Network Rail should reassess the working time limits and duration of breaks applicable to
lookouts and provide staff with appropriately updated instructions and guidance based on
these findings. This reassessment should make use of current research into vigilance activities
akin to railway lookout duties.

Network Rail should reassess the safe system of work hierarchy, taking account of evidence
from LOWS related incidents and the risk associated with using unassisted (flag) lookouts. If
Produced by RSSB
Author: Dr Greg Morse
Email address: Greg.Morse@rssb.co.uk
justified by the results of current tests of Semi-Automatic Track Warning Systems, where
workforce warnings are initiated using automatic train detection, Network Rail should include
such equipment within the hierarchy. If appropriate, Network Rail should specify any
circumstances in which this should be used in preference to LOWS.
Back to top
29 July
Latvia: Fatal shunting accident at Skirotava, 7 September 2014
For a summary in English, click here: LINK
At 06:17 (local time) on 7 September 2014, a
shunter was struck and killed by a runaway rake of
tank wagons in the hump marshalling yard at
Skirotava.
At the time of the incident, the yard was
undergoing reconstruction, meaning that its
automatic switches and crossings were being
operated manually. This gave the shunter the
additional responsibility of operating the point
levers.
The investigation concluded the immediate cause of the accident to be that the shunter did not
check that the wagons were moving and walked on to the line just in front of them before he was run
over. The report added that fatigue – as a result of his increased workload – may have caused his
‘temporary loss of concentration’. There is also a suggestion that personal circumstance may also
have distracted him.
Recommendation

The infrastructure manager should examine the possibility of reducing shunters’ working
environment risk factors, where there are changes in the workload and responsibilities of staff
members. It should also provide training opportunities regarding drowsiness and fatigue.
Back to top
30 July
UK: Electrical arcing and fire and parting of the overhead line at Walkergate, 11
August 2014
For the full report, click here: LINK
At around 18:56 on Monday 11 August 2014, a Tyne and Wear Metro train at Walkergate station
developed an electrical fault in equipment under the rear car, which tripped the power supply to the
overhead wire. About a minute later, the power was remotely restored by a power controller and a
fire started in the faulty equipment. The fault drew a high current from the overhead wire through
the car’s pantograph, causing localised overheating and, after about 18 seconds, the wire parted. The
loose ends of the live wire flailed around the train roof, showering sparks, and one end fell on the
platform. A second power controller happened to observe events on station CCTV and power to the
overhead line was disconnected. The parted wire had remained live for approximately 14 seconds.
Produced by RSSB
Author: Dr Greg Morse
Email address: Greg.Morse@rssb.co.uk
At the time of the accident, the train doors were closed for departure and there were no passengers
on the platform. About 30 passengers got out of the train after the power was turned off. There were
no reported injuries, although passengers were evidently distressed.
A sustained high current was drawn because the electrical fault occurred in a part of the train’s
power circuit, which was not protected by on-train equipment, and which could only be detected by
the overhead power supply protection equipment. However, because of the way the power was
switched back on, the level of current drawn by the fault was not sufficient to activate that
protection immediately. The underlying cause of the accident was the ineffective management of
risks created at the interface between system components, comprising the power supply, power
control and the train.
RAIB observed that the arrangements between the infrastructure operator, Nexus, and the train
operator, DB Regio Tyne and Wear, did not effectively facilitate the sharing of relevant health
and safety information on risks created within the system. It also observed deficiencies in the
competence management of the power control function and in the regime for assuring the
continued integrity of some electrical protection equipment.
Recommendations

Nexus should review its safety management system to ensure that it provides an adequate
framework for the management of its shared risks. Nexus should also review the effectiveness
of current arrangements with DBTW with a view to reaching a more effective arrangement on
the exchange of relevant safety information to facilitate the management of shared risks.

Nexus together with DBTW should identify (or review) and assess jointly created risks that
occur at all interfaces between the infrastructure, power operations and trains. This should
include the use of suitable risk assessment methodologies appropriate for identifying potential
failure modes and their consequences, and a recognised technique for assessing the extent to
which additional mitigations are required to reduce the risk as low as reasonably practicable.
To this end, Nexus and DBTW should ensure that they have access to, and utilise, competent
advice on conducting assessments of system-wide risks.

Nexus and DBTW should together complete the on-going review of procedures and practices
followed by power controllers, with a view to providing a codified set of procedures, which
have been appropriately risk assessed. Such procedures should be briefed out to power
controllers and linked to the power controllers’ training and competence management
systems.
Back to top
Produced by RSSB
Author: Dr Greg Morse
Email address: Greg.Morse@rssb.co.uk