HomeLong ReadsThe qualities of a world-beating railway

The qualities of a world-beating railway

Colin Wheeler looks at the latest rail safety incidents including November’s collision at Salisbury, the dangers of abandoned half trollies, and the derailment of a freight train at Sheffield station People skills, knowledge, motivation.

The use of railway engineering and operational judgement. These are the qualities I believe are needed to make the services and safety of our future railways − the Great British Railways – the best in the world. Recent reports of accidents, incidents, and near misses indicate the size of the task, but it will be worthwhile.

Unfortunately, while we await an announcement about the appointment of Simon French’s successor as Chief Inspector for the Rail Accident Investigation Branch (RAIB), the organisation’s workload has been substantially increased by recent events.

Salisbury passenger trains collide

On 3 November, Anna Witherington, on behalf of the RAIB gave an update on the causes of the collision between two in-service passenger trains that occurred on 31 October. At around 18:45 that evening, the 17:20 South Western passenger train from London Waterloo to Honiton collided with the side of the 17:08 Great Western passenger service from Portsmouth Harbour to Bristol Temple Meads at Salisbury Tunnel Junction. The junction is near Salisbury in the immediate approach to Fisherton Tunnel. The South Western train was using the junction to join the Down Main from the Down Dean line as the Great Western train approached the junction from the Basingstoke direction.

Low wheel/rail adhesion

The collision impact caused the front two coaches of the South Western train and the rear two coaches of the Great Western train to derail. Both trains continued for some distance into Fisherton Tunnel. The RAIB’s preliminary investigation established that the South Western train driver had braked on the approach to signal SY 31 and, 12 seconds later, made an emergency brake application. A second emergency brake was applied by the Train Protection Warning System (TPWS) but the train still passed the signal and reached the junction where the collision occurred.

The On Train Data Recorder analysis showed that wheel slide occurred “almost certainly as a result of low adhesion between train wheels and the rails.” Doubtless we will learn more when the RAIB has completed its investigation.

Injuries at Grosmont

On 21 September there was a 10mph collision on the North Yorkshire Moors Railway at Grosmont Station. It happened around 10:32 when a Class 20 diesel locomotive ran into the back of a passenger train. Five passengers suffered minor injuries and there was some damage to the train. Following a preliminary examination, the RAIB has decided that they will publish a Safety Digest.

Another UWC incident

On 18 September − soon after I reported on the mid-August incident at Kisby on 19 August − there was a near miss on Forestry User Worked Crossing (UWC) between Thetford and Brandon in rural Norfolk. This occurred in the early morning at about 06:05 when two cars were involved in a near miss with a passenger train. Again, following their preliminary examination the RAIB has decided to issue a Safety Digest. Local press reports have suggested that the level crossing itself may have been faulty.

Passenger train crashes into buffer stop

This accident at Enfield Town station involved 75 passengers though fortunately only two required treatment for minor injuries. It occurred at 08:21 on the morning of 12 October when a passenger train ran into the platform two buffer stop and ended up in a raised position, balanced on top of the buffer stop, which was also detached by the impact. On 25 October the RAIB announced that its investigation will “seek to identify the sequence of events leading to the collision,” as well as the “facts influencing the driver’s actions, the training, supervision and management of drivers by Arriva Rail London, design and configuration of the buffer stops, and the assessment and risk control of terminal platform overruns at Enfield Town Station”.

Half trolley struck at speed

At 06:10 on 21 October, a Great Western train struck a half link trolley at Challow, between Didcot and Swindon, at a speed of around 100mph. Thankfully there were no injuries, but the train was damaged and ended up with the trolley wedged beneath it. One hundred and thirty-five passengers were stranded, and the Down Main line was closed for three and a half hours. Network Rail’s Safety Advice comments that “the trolley had been left behind after the work”.

The Controller of Site Safety (COSS) had signed out with the Engineering Supervisor by telephone and the worksite was then handed back by the Person in Charge of Possession (PICOP) as “Safe for the passage of trains”. The website item adds that “this was the second occasion in recent weeks that a trolley has been left behind after work and then missed by the “clear of line verification arrangements”. Once upon a time the PICOPs actually walked through to check for themselves!

Derailment of freight train carrying cement powder

The RAIB report on this accident that occurred on 11 November 2020 was published on 5 October this year. The RAIB’s 50-page report is titled “07/2021 Freight Train Derailment at Sheffield Station”. Its contents are both relevant and significant in railway operating, and the responsibilities and competencies needed for track engineering inspections, maintenance and renewals. It is recommended reading for those designing the future track inspection and maintenance organisation of Great British Railways.

This incident saw 16 of the freight train’s 34 wagons derailed at the north end of Sheffield Station. The train was travelling from Hope in Derbyshire to Dewsbury in West Yorkshire. The wagons were loaded with cement powder carried in PCA tank wagons, hauled by a Class 66 locomotive. PCA cement tank wagons have two axles, a wheelbase of 4.88 metres, and weigh 13 tonnes when empty and 51 tonnes when loaded.

The train had been “coasting” through the station at just 12mph when the derailment happened. A number of the wagons were damaged and there was also significant damage to the track. This resulted in the partial closure of the station.

The leading right-hand wheel of the 12th wagon had dropped between the two running rails which were too far apart due to gauge widening. The train was stopped when the signaller saw a number of equipment failures on a display screen and alerted the driver.

Track screws broken

The report says that a number of track screws had broken “several weeks or perhaps months before the derailment.” These had not been identified by Network Rail’s maintenance and inspection activities, despite being a “location with a potentially high risk of derailment”. The RAIB found that “Network Rail’s guidance for identifying such risk had not been applied”.

The report clearly states that the train driver did not contribute to the derailment. The track maintenance engineer (TME) had worked on the railway for 12 years, six of them as assistant TME at Sheffield and was appointed as acting TME just five months before the derailment. The Sheffield Infrastructure Maintenance Engineer (IME) had 29-years’ experience with six of these as TME at Sheffield. His appointment as acting IME was also made just five months before the derailment.

The derailment occurred between the crossing and switch rail of 4062 points. The left-hand wheel then climbed over the left-hand rail because the outer face of the right-hand rail was constrained as the gauge tightened approaching the switch toes of 4062 points. When the derailed wagon 12 reached the points of 4068A it did not follow the diverging right-hand route but attempted the straight route into the shunt neck siding, causing it to fall onto its side and wagon 11 etc to derail.

There were no injuries, but the north end of the station remained partially closed for five days. There was damage to both wheels and couplings on wagon 11 and wagons 13 to 26. Wagon 12’s suspension was severely damaged and largely detached. Its tank was ruptured, resulting in cement powder being spilt onto the tracks.

Causes of the derailment

The RAIB found that the derailment occurred due to: (i) multiple track screws securing the rail baseplates to the bearers had failed prior to the derailment; (ii) failed track screws had not been identified by the inspection regime; (iii) there was no additional mitigation at 4062 points as they had not been identified as a high-risk location; and (iv) the design of 4062 points did not include a check rail on the sub-200 metre radius curve between the crossing and switch heel.

Visual inspections, shuffle and gauge widening

The RAIB found that in the weeks before the derailment all “Basic Visual Inspections” had been carried out weekly and in daylight hours by track patrollers and defects had been identified, although on the last four record sheets no defects had been recorded. The record for 7 October noted “baseplate shuffle at the heel of 4062 points” and a review of earlier records found a note of shuffle back on 5 June 2019.

he requirement is for three-monthly supervisor’s inspections and two-yearly engineer’s inspections. During their investigation RAIB inspectors were told that there were “long standing issues with gaining access to some of the lines at Sheffield Station to perform routine inspections and maintenance.” In September 2020 the gauge at the point of derailment was 13mm wider than that measured a year earlier.

A probable factor

The RAIB found that there had been no dynamic measurement of track geometry for 15 months and states that this was “a probable factor” in the derailment. Network Rail uses a single multi-purpose vehicle (MPV) to measure dynamic geometry and was scheduled to do so at Sheffield every three months. However, no recording runs had been made since November 2019. They also found that of the 19 three-monthly runs only five delivered data on the Down Through line. The RAIB report also observes that “the format of dynamic geometry data output from the MPV track recording vehicle makes it difficult for it to be analysed by maintenance staff”.


The RAIB recommendations relate to: (i) improved management of derailment risk at high-risk locations; (ii) improved implementation of safety critical track maintenance activities; (iii) alignment of standards and practice for the use of check rails; and (iv) improving the ability of track maintenance staff to detect changes in track geometry.
In the preface to the reports published by the RAIB they state that “it is inappropriate that RAIB reports should be used to assign fault or blame or determine liability since neither the investigation nor the process has been undertaken for that purpose”.

However, Network Rail and the Office of Rail and Road, (and, in the longer term, those who will design the management organisation for Great British Railways) would do well to review their plans having studied this report by the RAIB.

The questions I am left with after studying it include:

Is it good practice to fill vacancies by stepping up a number of safety responsible engineers and inspection staff rather than promoting them to other areas to widen their experience and bring a fresh pair of eyes to bear?

Do local inspection staff have the delegated power and authority to close the railway when they have track safety concerns and if not, why not?

Where maintenance or renewal work has to be deferred or delayed, and track conditions deteriorate, do the local TME and IME have the authority to apply speed or axle weight restrictions or close the railway for safety’s sake and without retribution?

Surely, if the standard requires inspections and dynamic measurement runs every three months, missing more than one should result in line closure?

You may gather that I consider this report to be as far reaching as any I have read. The lessons from it should be remembered when the organisation of Great British Railways is being developed. Perhaps the most important attributes for a safer railway are the commitment and team spirit aimed at always doing the best for the railway.

Whatever the organisation is, there will inevitably be potential conflict between commercial, operating and infrastructure engineering people. But, at the end of the day, local infrastructure engineering staff backed by their professional engineers must take responsibility for the safety of the line and be prepared to stop traffic when running trains becomes unsafe.

Image credit: RAIB