blog
James Lloyd
James Lloyd

BLOG: Off the rails: RAIB publishes its report on the November 2019 Neville Hill incident

On 18 November 2020, the Rail Accident Investigation Branch published its report into the collision and derailment incident involving two trains at Neville Hill on 13 November 2019. James Lloyd summarises the report’s findings and looks at the lessons to be learned for rail safety and regulation in the UK.

Shortly before 10pm on 13 November 2019, outside the Neville Hill maintenance depot at Leeds, a nine-carriage Class 800 Intercity Express Train collided with the rear of a High Speed Train (‘HST’) set comprising nine carriages and a Class 43 locomotive at each end. Both trains suffered significant damage, and some of the carriage wheels of the Class 800 set derailed.

The incident occurred almost ten years to the day since the Neville Hill derailment of 17 November 2009.

On 18 November 2020, the Rail Accident Investigation Branch (‘RAIB’) published the report of its investigation into the incident. Though comprehensive, the report makes makes clear that:

“[the] purpose of a Rail Accident Investigation Branch (RAIB) investigation is to improve railway safety by preventing future railway accidents or by mitigating their consequences. It is not the purpose of such an investigation to establish blame or liability.”

The investigation is of particular importance, as the Class 800 and its variants are new additions to the rail network and now operate across most of the UK.

 

The incident

The Neville Hill maintenance depot, 3.5km east of Leeds Station, has existed in one form or other since 1899. Currently, the depot is operated by East Midlands Railway (‘EMR’) and Northern, used for maintenance and train storage. London North Eastern Railway (‘LNER’) also uses the depot for train storage.

At 9:41pm on 13 November 2019, two train sets operated by LNER were headed for overnight storage at the depot. Neither had passengers aboard. On the same line, a Class 800 caught up and collided with the rear of a Class 43 HST set.

Alongside damage to the front of the Class 800 and rear Class 43 locomotive, the Class 800 suffered derailment of the trailing bogie of the second and third carriages and the trailing wheel set of the fourth carriage, with all derailed wheels moving towards the adjacent rails. As the report concludes:

The extent of the derailment (up to 1.25 metre of lateral displacement on the second vehicle) would have been sufficient to significantly infringe the swept path of a vehicle travelling on an adjacent line.

 

Background – the Class 800

The first LNER-operated Class 800 entered service on 15 May 2019. The Class 800 is a relatively new train, developed by Hitachi as part of the Department for Transport’s Intercity Express Programme. The programme sought a replacement for aging Class 43 and Class 91 stock, introduced in 1957 and 1988 respectively, which have served every part of Great Britain since.

The Class 800 and its variants rocketed to fame and ubiquity from 2014, and are now in service across the network, albeit differently branded by their respective Train Operating Companies (TOC’s): LNER [Azuma]; GWR [IET]; Hull Trains [Paragon]; and TransPennine Express [Nova]. Variants are also scheduled to enter service with EMR [Aurora], East Coast trains and Avanti West Coast.

Being at least 30 years younger than the trains it is designed to replace, the Class 800 marks a significant technological advancement and change for drivers accustomed to Class 43 and Class 91 operation. Most significantly, in respect of this incident, the Class 800:

  • is a bi-modal train capable of running on either electricity from overhead wires, or on-board generators powered by diesel engines;
  • is significantly more powerful, with greater acceleration than the older Class 43 and Class 91 HSTs; and
  • has a sophisticated on-board Train Management System (‘TMS’), a computer-based system controlling most aspects of train operations, which drivers access and control from a console in the cab.

 

What caused the incident?

Permissive working

Safe railway practice requires that only one train should be able to occupy a given section of line (the space between two signals) at any one time. Generally, multiple trains are not permitted to occupy the same section of line. Thus, a train should not be able to catch up with a train in front.

The section of line immediately before the Neville Hill depot is one of many across the country in which it ispermissible for multiple trains to occupy the same section of line, in a practice known as ‘permissive working’. Permissive working usually applies to stations (allowing multiple trains to occupy the same platform) or entrances to depots, where trains may ‘queue’ on in-lines (essentially, slip roads) to avoid congestion on main lines. Wherever permissive working is in place, rules require that drivers must proceed at caution, ready to stop short of any obstruction, keeping sufficient distance from the train in front.

Additional safeguards often include strict speed limits.

Speed

The speed limit for the section of line on which the incident occurred was 5mph. The driver told the RAIB that he knew this, and had no intention of travelling faster. At the time of the incident, the Class 800 was travelling at 15mph.

The Class 800’s Automatic Speed Limiting (ASL) equipment, when activated by a driver, prevents a train being accelerated to above the set speed for a section of line. That equipment was not activated at the time of the collision. Had the ASL equipment been activated prior to the incident, the train would have applied its brakes and reduced power automatically when it accelerated above 5mph.

The RAIB report details on-board data, showing that the driver of the Class 800 demanded approximately 20% of the available tractive effort on the train immediately after a brief stop before the collision. Despite the rising gradient towards the Neville Hill depot, this was sufficient for his train to reach a speed of 15mph within 27 seconds.

The report concludes that with the equivalent input, an HST (of which the driver had some 30+ years experience), on the same gradient, would have reached 7mph in the same time. The report concludes that:

At this rate of acceleration, it is probable that the gap between the two trains would have remained sufficient for the driver to realise that his train had started to catch up with the train ahead by the time he refocused on the driving task, and an accident would have been averted.”

Driver distraction

The driver of the Class 800, the RAIB concludes, was distracted in the seconds before the collision; looking at the TMS display, and unaware that his train was gaining ground on the slower HST. The report notes that:

Although he was experienced, the driver’s familiarity with driving [the Class 800] was limited. This was only his third unaccompanied turn driving [a Class 800] (a normal event in itself). Evidence indicates that he felt unfamiliar with the new technology, in particular, the TMS computer interface and the use of a combined power brake controller. The new technology made the [Class 800] very different to the trains he had previously driven. He had also not driven any trains for a significant portion of the previous two years.”

The report considers the source of distraction at length, and a number of difficulties posed by the TMS display, ease of use, and LNER’s training materials.

Two days before the incident, the driver reported difficulty with a different Class 800 on entry to Neville Hill depot. On leaving Leeds station for Neville Hill, trains pass a signal and a balise which interacts with the Class 800’s Automatic Power Changeover system (APCO). On the previous occasion, the APCO intervened when the train crossed the balise and the train automatically fired its diesel engines. This was not an intended consequence, and resulted in the engines starting cold, which can result in damage. The APCO had intervened, it transpired, as the driver had not changed the train’s headcode on the TMS to identify the train as a service running into the electrified depot, rather than a service running onto a non-electrified portion of mainline. 

On the night of the incident, as the report then concludes:

“[the driver] attempted to change the train’s headcode from 1D29 to 5D29 using the TMS touch screen. However, he was unable to do so and the TMS continued to display that it was using headcode 1D29.

[…]

To avoid a repeat of what had happened on the Monday night, the driver decided to isolate APCO using the TMS touch screen. This was to prevent APCO from intervening when travelling over the balise.

[…]

the [Class 800] came to a brief stop (just over one second) behind the HST, which by now was on the move. Knowing that his train had passed the APCO balise and keen to reinstate APCO as soon as possible, the driver turned his attention towards the TMS screen. At the same time, he realised that the HST was on the move and decided to follow it. He moved his power brake controller slightly to demand a low level of tractive effort, while continuing to focus his attention on the TMS.”

Less than thirty seconds later, the Class 800 collided with the rear of the HST.

Class 800 design

The report includes detailed testing and modelling of the Class 800 design, and crash-test simulations. The RAIB explains in the report that couplers between carriages of the Class 800 were designed to help the train to withstand high-velocity impact and collisions. The couplers are designed to deform on impact and absorb energy, reducing movement of wheelsets and bogies between carriages.

The report’s analysis concludes that none of the couplers of the Class 800 had started to deform on this occasion: insufficient energy had been absorbed; wheelsets had lifted up from the rail, veered out to the right and derailed. This is both a surprising, and extremely dangerous outcome, with great effort having been expended in the design process to avoid the risk of derailment after head-on collision. The report concludes, following digital modelling in co-operation with the Class 800 manufacturer, Hitachi, that “the critical speed at which the wheels started lifting off the rail was therefore likely to be between 5 and 7.5 mph”.

The report notes, crucially, that the design standards set under the Department for Transport Intercity Express Programme specified four “design collision scenarios” involving front-end impact. None of those scenarios, however, involved speeds below 22.5mph. Of such speeds, the report concludes:

At 22.5 mph (36 km/h), the inter-vehicle couplers were predicted to collapse significantly, which reduced the angle that the couplers can freely pitch and yaw […]

RAIB concluded that it was therefore probable that the risk of derailment decreased as the speed of collision increased (because the inter-vehicle couplers collapse and limit the pitch and yaw movement of the couplers). Overall, the risk of derailment appeared to be greatest at low speeds: very low speeds do not generate enough energy to lead to a derailment and high speeds create enough energy to collapse the inter- vehicle couplers which appear to reduce the risk of derailment.”

The RAIB’s analysis also highlights significant differences between the results of Hitachi’s 3D simulation models, and observed reality.

 

Lessons to be learned

Recommendations

The report makes five key recommendations:

  • LNER, with support from Hitachi, should review the Train Operation Manual and TMS documentation that Hitachi has provided;
  • LNER should review its procedures, training, and associated materials to prepare drivers to correctly interact with the TMS;
  • Hitachi should revisit its assessment of the performance of the Class 800 using a refined 3D model, including modelling changes demonstrated by the investigation to be necessary to predict train behaviour with sufficient accuracy. The assessment should be subject to a review by an independent third-party and, if applicable, Hitachi should declare any area of non-compliance;
  • LNER, with support from Hitachi, should assess the risk associated with derailment of a Class 800 following collision at low speeds, and take any necessary actions to demonstrate an acceptable risk; and
  • The Rail Safety and Standards Board should consider whether the findings of the investigation warrant proposing revisions to the design standards.

The report therefore identifies potential deficiencies at the design, manufacture and operational stages of the (so far, short) life of the Class 800; a significant range of factors which need to be addresses and which may result in further incident if ignored.

The driver

The extenuating circumstances of the driver of the Class 800 (health issues and time spent away from work, resulting in limited experience of the Class 800) are clearly appreciated by the RAIB in its report. 

It is clear from the report that the driver’s candour was well received; a useful lesson to those facing regulatory investigation of any sort.

Corporate response

The report lists a number of improvements to training already implemented by LNER, and notes the co-operation of the Class 800 manufacturer, Hitachi. Those firms’ co-operation with the investigation, again, was well received. Nonetheless, the RAIB makes stark and important recommendations.

As at the conclusion of any significant regulatory investigation, a period of introspection, consideration and remediation will now have to be demonstrated by those firms involved, alongside other TOCs operating Class 800s and variants.

Given its popularity as a replacement for Class 43 and Class 91 stock, the Class 800 is here to stay. The failings identified from both manufacturing and operational perspectives require considered and comprehensive remediation. If lessons are not learned, further serious incidents are likely to result. As the industry well knows, incidents involving collision and derailment carry potentially fatal consequences.

 

Broader lessons for firms and individuals under investigation

The RAIB was formed in 2005, following the recommendations of the Cullen Inquiry into the Ladbroke Grove crash of October 1999. The RAIB adopts a ‘blame-free’ investigative approach, similar to that of many safety investigation processes. Its investigations are notable for the co-operation and candour of stakeholders; a phenomenon which has led other sectors to consider carefully their approach towards accident investigation. The RAC, for example, continues to investigate the lessons to be learned for highways accident investigation.

Individuals

For those facing investigation, this report is an important example of the ‘credit’ to be gained by candid, constructive engagement with regulatory, or indeed disciplinary investigations. Few investigations, however, are designed to be blame-free; those under investigation occasionally face a difficult balancing act between co-operation and protection of their own interests, or future liabilities – factors which are often compatible, but not clearly so.

Given that a careful and considered approach toward engagement with a regulatory investigation can have such a significant positive impact, those facing investigations by regulators are well advised to seek legal advice and assistance at the earliest possible stage. Such assistance can provide peace of mind and a wealth of experience for what are often daunting and foreign processes, affording registrants and individuals under investigation the expertise and resources to navigate the pitfalls of regulatory investigations.

Firms

Those firms who have been subject to, or are concerned about regulatory investigation are always best placed to seek legal assistance. Such assistance need not only be reactive, focusing on specific investigations, but should be proactive. The use of past business reviews and cultural assessments of risk-taking are tools available to all regulated entities, which can mitigate the risk of regulatory investigation.

Where adverse findings or recommendations are made, as in the RAIB’s report here, a clear and committed response to regulatory criticism can make a significant difference if concerns re-emerge or are repeated: repetition of identified failings is likely to result in significant liability, unless the regulated entity can demonstrate clear and reasonable steps taken to address concerns.

 


James Lloyd is a barrister at Carmelite Chambers with significant experience of regulatory investigations. James practises in regulatory and professional disciplinary law, representing regulators and individuals. He also maintains a general criminal defence and extradition practice.