A series of reports on accidents and incidents that took place over the last year on the railway all point to human error being the primary cause.
A flurry of Rail Accident Investigation Branch (RAIB) reports came out as the end of the year approached, giving safety expert Colin Wheeler plenty to mull over. The incidents they describe were diverse, including derailments, signals passed at danger (SPADs), a fire and, sadly, an injury and a death. But was there a common theme?
Depot operating beyond its capacity?
RAIB report 09/2020 follows its investigation into the fatal accident that occurred at 20:00 hours on Saturday 14 December last year. The train driver who died was “passing between two closely spaced trains when one was moved towards the other as part of a coupling operation.”
It happened at Tyseley depot in Birmingham, where carriage sidings 1 to 15 can be accessed from both ends and the depot is well lit. The trains were two Class 172 Diesel Multiple Units (DMU’s). The gap between the gangways of the two DMU’s was 540mm or just 350mm between the horns on the auto-couplers.
The report states that “West Midland Trains (WMT) had not adequately considered the risks faced by drivers on depots”. Tyseley depot was operating “at or beyond its capacity at night”, and “management assurance processes had not promoted safe working practices”. Workload is specifically referred to in the published report.
“RAIB found no evidence that WMT had undertaken any assessment of the hazards faced by train drivers when they were working or walking within maintenance depots, even though the presence of train drivers within such depots is a routine and regular part of their duties.”
Recommendations and learning points for train operators
WMT is recommended to assess the risks to those walking or working in depots, yards and sidings, so that suitable measures are put in place to control risks. The train operator is to “ensure that unsafe working practices within the company can be effectively identified and their causes addressed”.
The report also lists five learning points:
- The dangers of being close to rail vehicles;
- The need to make authorised walking routes known to staff;
- The importance of assessing the capacity of depots to ensure they can operate safely;
- The need to ensure brake tests and safety critical examinations are routinely checked and understood by staff;
- The importance of screening staff involved in accidents for the presence of drugs or alcohol.
Near miss at Trewern Mill UWC
This features on Network Rail’s own Safety Central website. At 13:53 on 20 July 2020, a signaller working at Whitland signal box was asked for permission to use the Trewern Mill User Worked Crossing (UWC). He checked the Box diagram and, seeing that the track circuit was clear, gave permission, believing the train shown in the next track circuit area had already passed the UWC. At 13:56 the user confirmed having crossed, but advised that, after opening the gates, a train had travelled over the crossing!
The reason for this was that the signal box diagram showed the wrong position for the UWC. During 2019, various signal box diagrams were altered, with mileages and level crossings added to them following an RAIB recommendation. Three of the four user worked crossings added to the Whitland Signal Box diagram “were depicted incorrectly-all have now been rectified.”
RAIB also found that some distances were quoted in yards whilst others used chains.
Sheffield station freight train derailment
On 11 November at 02:45, a 34-wagon cement-carrying freight train travelling from Hope, Derbyshire, to Dewsbury derailed at Sheffield station (see above). Significant damage was caused and there were days of delay as wagons were unloaded using heavy duty vacuum pumps before they could be lifted out using a rail crane.
The train derailed at the north end of Platform 1 at just 12mph. The leading 10 wagons and the rear eight stayed on track. One wagon tipped over and spilt onto the tracks and RAIB has already said that a number of rail fastenings were broken before the accident happened. They are looking at the design and maintenance of both track and wagons involved.
Bognor Regis passenger train derailment
This derailment on 22 October was reported as a news story by RAIB on 27 October. At 05:11 that morning, the 05:05 Bognor Regis to Littlehampton service derailed as it departed from Platform 4 with just 15 passengers on board. They were evacuated from the rear part of the train which was still on the platform.
There were no injuries, but a set of points was damaged. RAIB has decided to publish a Safety Digest “in the next few weeks”.
Llangennech derailment and oil train fire update
On 5 November, RAIB published its Urgent Safety Advice 02/2020 relating to the oil tanker train derailment and subsequent fire at Morlais junction, Llangennech, on 26 August. It focusses on wagons that carry dangerous goods and is aimed at those charged with their maintenance.
Such organisations, described as “Entities in Charge of Maintenance” or EMCs, are told to have “appropriate arrangements in place to manage the safety risk associated with malfunction of the braking system, including facilities, tools and equipment, systems for assuring the competence of those involved, and instructions for assuring work quality, methods for initial and ongoing assurance of the security of fastenings and processes for the identification and tracking of safety critical components.”
The Safety Advice adds that “the assessment should take into account the particular hazards associated with the conveyance of dangerous goods”.
The Urgent Safety Advice includes a picture from the Pen-y-Bedd CCTV camera, 12 miles from Morlais junction, where the train derailed after the leading axle of the third wagon stopped turning. It states: “There is no record of any check on the tightness of the relay valve fastenings ever having been made and there was no process requiring such checks or provision of any measures, such as witness markings, which would have indicated that the fastenings were becoming loose.”
This report will be important for all operators of rail freight services.
Unpowered new units with isolated brakes
RAIB’s report 10/2020 was published on 9 November on the signal passed at danger on 26 March at Loughborough South. Around 10:57, train 5Q26 was making a rolling stock transfer move from Old Dalby to Worksop when it passed red signal LR507 at 20mph and stopped 200 metres past it. The safety overlap of the signal was exceeded, but it came to rest 600 metres short of the point where conflict with other train movements could occur.
The train was formed a four-car Class 710 EMU (Electrical Multiple Unit) with Class 57 locomotives at each end (see above). New Class 710s were being moved unpowered with brakes isolated, with the brake piping between the 57’s providing braking.
The driver with the shunter prepared the locomotives and carried out a brake test. After departing Melton, the train was held in the loop outside Melton Mowbray station for 45 minutes but departed on time. It travelled at 75mph. Two miles before reaching signal LR 507, on seeing LR 503 showing a single yellow aspect, the driver applied half full-braking which he increased to full as he passed that signal. Signal LR 519 was replaced to danger before 2L58 departed from platform 3 at Loughborough in response to an alarm resulting from LR57 being passed at danger.
How the train was driven, its speed approaching the yellow signal, the train driving sheet incorrectly specifying a speed of just under 75mph and the amount and timing of the braking all contributed to the incident. RAIB commented that the Rail Operations Group’s (ROG) management did not detect lack of compliance with its safety systems; the driver did not conduct an adequate braking test, and ROG did not adequately manage the retrieval of information from locomotive OTDRs (On-Train Data Recorders). Both the RAIB and ORR (Office of Rail and Road) investigations were hampered as a result.
Previous SPADs (Signals passed at danger)
Back on 24 March 2016, at Ketton, Rutland, a Class 47 locomotive hauling a four-car Class 321 passenger unit passed a signal at danger. Since then, there have been five more ROG SPAD incidents:
- 10 July 2017, Droitwich Spa;
- 26 March 2018, Potters Bar;
- 25 November 2019, Moreton-on-Lugg;
- 8 February 2020, Bristol Barton Hill;
- 26 March 2020, Loughborough South.
The Office of Rail and Road (ORR) served an improvement notice on 28 April with an initial compliance date 28 July, this was later extended to 31 August.
Actions taken include the introduction of a weekly electronic notices system, with new documents issued to drivers at the start of each journey, prepared by shunters. On 18 June, a safety brief was issued explaining the new system. The company has procured more OTDR downloading equipment and trained staff in its use.
Over-speeding at Dauntsey Wiltshire
On 10 November, RAIB published preliminary results of its investigation of over-speeding that took place on 12 August. The 16:07 Great Western Railway (GWR) London to Bristol service travelled at 117mph over a section of track with a 20mph ESR (Emergency Speed Restriction) on it!
The restriction was imposed on 25 June due to a cyclic top track defect affecting a 100-metre length of track using 30/125 (30mph restriction for freight but 125mph for passenger trains).
In the heat of 12 August, rail temperatures reached a critical level, trains were stopped and cautioned with a 20mph imposed. This began at 13:55 and, by 15:52, revised local signage replaced the stop and caution procedure.
Subsequently, the first train over the affected section was the London to Bristol train whose driver had booked on at 14:36 hours. He had been issued with a hard copy notice with ESRs including the Dauntsey cyclic top one, over which he had driven in previous days.
Network Rail did not advise GWR of the hot weather ESR until 16:33, but, by the time the train reached Dauntsey, new signs had replaced the stop and caution. ATP (Automatic Train Protection) had not been modified to include the 20mph heat restriction ESR, since it was likely to be lifted within hours.
On seeing the ESR Board the driver assumed the 125 part had been displaced. He saw the 20mph at the ESR commencement whilst travelling at 117mph. A full brake application reduced his speed to 105mph. He reported to the signaller that some signage was missing.
The ESR was removed at 19:07
RAIB decided not to investigate further, referring back to its recommendations following a similar incident at Sandy on 19 October 2018. Recommendations made then included operating companies to “review their practices of making drivers aware of ESRs” and the industry was “to consider and review options for a safe and suitable means of providing drivers with warnings of ESRs on the route ahead through the use of suitable technologies”.
At Dauncey, the 30/125 mph restriction had been allowed to continue for some time without being converted to a TSR (Temporary Speed Restriction).
On motorways, car satellite navigation devices flag up emergency lane and other restrictions ahead. Surely, we can do the same for railways? Sandy happened in 2018, so why the delay?
Waybeamed bridge derailment
On 16 November RAIB issued report 12/2020 into the freight train derailment at Wanstead Park, London, on 23 January. Just before 06:00, a wagon in the rear half of a heavily loaded freight train derailed on a small radius curve whilst crossing a waybeamed bridge. Two adjacent wheel sets from different wagons derailed. One re-railed shortly afterwards, but the other caused significant track damage over a distance of two and a half miles (see below).
The remainder of the train stayed on track. Severely rotted longitudinal timbers at the point of derailment were concealed by superficially good exteriors. RAIB says “inspection methods and tools used by Network Rail were not sufficient to detect poor internal timber conditions”.
Track gauge widening had been identified on six occasions between March 2019 and January 2020. On three occasions, work was done, but in nearby locations due to “erroneous GPS data”. On three other occasions, the recording software failed to report any location of the fault!
Following a longitudinal timber failure in the same area in August 2017, Network Rail updated its standard for the inspection and management of long timbers. This was not published until March this year.
One wagon pushed the rails apart, resulting in the derailment. Wagon maintenance records show rapid wear over some years resulting in high lateral forces.
RAIB makes a single recommendation to Network Rail to improve the provision of track recording data to maintenance staff. The recommendation to the wagon owner concerns maintenance of wagons and reducing the risk of defective ones entering service. Learning points cover management of longitudinal timbered bridges and wagon loading in freight terminals.
Grove Ferry, Kent Route Southern Region
This was the location of a serious injury accident on 24 October, reported on Network Rail’s Safety Central website. Track measurements were being taken to align sleepers from the adjacent line on which an engineer’s train was standing.
The train was authorised to move and ran over the operative’s hand amputating two fingers.
Two hours elapsed before the accident was reported to Route Control, which delayed reporting of the incident. RAIB, the Office of Rail and Road, Network Rail and the contractor are now investigating.
Leeds Neville Hill Depot report
RAIB published report 13/2020 on 18 November, just over a year after the accident. A new LNER express train, travelling at 15mph, collided with the back of an HST (High Speed Train) which was travelling at 5mph. There were no injuries.
The report states that the driver, “focussed on reinstating the on-board system that he had recently isolated instead of driving”, had “commanded too much acceleration”. “Ambiguous documentation” from Hitachi led to LNER misunderstanding the setting up of the train management system when developing the driver training programme. The RAIB report also added that “crash-worthiness at speeds lower than 36kph and derailment performance were not specifically considered.”
Two recommendations to LNER relate to setting up the train management system, whilst Hitachi is to reassess the design against the crashworthiness standard and RSSB (Rail Safety and Standards Board) is to consider modifying its crashworthiness standard.
Bromsgrove, blame Boris?
On 23 March, at Bromsgrove in the West Midlands, a passenger train collided with a derailed locomotive. RAIB published its report 14/2020 on 19 November.
Before leaving home for work, the locomotive driver had watched the Prime Minister on television detailing the COVID-19 lockdown. He was to drive a Class 66 locomotive as a banking locomotive for a freight train ascending Lickey Incline at around midnight.
During his drive to work, he received a telephone call about child care arrangements.
He reached Bescot Yard, Walsall, and departed in the locomotive at 21:32 hours, travelling via St Andrew’s junction and Kings Norton. He received three text messages on his mobile in 11 minutes and sent four replies. A picture message, received at 22:38, showed advice from the school.
Approaching Bromsgrove Neck, the locomotive was coasting at 18mph, increasing to 23mph on the downhill gradient.
When just 40 metres from the siding buffer stop, the driver realised where he was and made a full brake application. Four seconds later he ran into the buffer stop at 21mph. The locomotive derailed foul of the adjacent main line.
Less than a minute later, a passenger train collied with the corner of it. The passenger train did not derail, but stopped between the locomotive and Bromsgrove Station.
RAIB’s recommendation is for the revision, as necessary, of processes and standards applying to buffer stops and ensuring adequate management of risks arising from buffer stop collisions.
Putting safety first, or get out look and listen
“Putting safety first” is the mantra with which all railway people are familiar. Doing a good job, one of which you can be proud, is satisfying.
Arguably, almost all of the incidents and accidents in this article are the result of human error. I suggest this is due to reduced commitment, motivation and acceptance of personal responsibility. This situation has grown due to remote management and, in particular, lack of face-to-face listening management.
Management via email and hiding behind rules never works. “Get out, look and listen” is a better slogan.
