Rolling stock crashworthiness is often discussed in the media after a serious train accident, usually accompanied by graphic images of the wreckage. Collisions between trains and with obstacles dissipate large amounts of energy and it is almost impossible to forecast the outcome given the unique combination of vehicles and accident site.
As such events are now exceedingly rare, Rail Engineer examined serious accidents over the last 40 years to show how their frequency has been reduced and consequences made less severe, with reference to the role of crashworthiness.
Railways learn from accidents and adopt inquiry recommendations with the aim of preventing recurrence. Following serious accidents in the 1980s and 1990s, the industry formalised a five-part strategy based on good risk control principles:
- Eliminate the risk by preventing derailment or collision.
- If an accident occurs, vehicles should be capable of absorbing impact energy in a way that minimises risk to occupants.
- Motion needs to be controlled to keep the vehicles upright and in line so that the forces can be managed.
- Interiors must be designed to protect occupants as effectively as possible.
- If all else fails, vehicles and procedures must facilitate escape and rescue.
Eliminate the risk
Of the 19 incidents, nine would have been prevented by an effective train protection system. Thanks to TPWS – introduced from 2002 – there have been no incidents of these types since the 1999 Ladbroke Grove accident. Broken rails or defective points caused the accidents at Bushey, Hatfield, Potters Bar and Grayrigg. Today, the number of broken rails per year is 10% of that in 2000 and there have been significant improvements to point design and maintenance. Five accidents were caused by obstructions on the track, involving a cow (Polmont), road vehicles (Ufton level crossing and Great Heck) and landslips (Watford and Carmont). Significant work has been done to minimise landslip risk. Clapham Junction, in 1988, was caused by defective signalling equipment.
Minimise the consequences
In an accident, rail vehicles should be capable of absorbing impact energy in a way that minimises the risk to occupants. Casualty figures in the table of accidents show that the Mark 1 coach was unable to do this. These coaches – with their 1930s design – had been introduced in the 1950s and had a strong underframe supporting a relatively light superstructure. By the late 1980s, they had been replaced by more modern designs except for a few brake and catering coaches, and many Southern electric multiple units. The latter were involved in the accidents at Clapham Junction, Purley, Cannon Street and Cowden where almost all fatalities occurred due to the loss of survival space.
The Clapham Junction report noted that compression of the passenger space caused most of the 35 fatalities. The Cowden inquiry stated that a feature of accidents involving Mark 1 rolling stock was “one train or part of a train overriding another, the frame of one coach slicing through the bodywork of another.” In the 15km/h bufferstop collision at Cannon Street, fatalities were caused as one coach overrode the buffers of the adjacent coach and destroyed its superstructure.
Mark 2 and 3 coaches, introduced in the 1960s and 1970s, were of monocoque construction (i.e. the chassis is integral with the body), with a welded steel stressed skin, in line with the standard approach to vehicle design adopted since the 1930s by most European railways and in the United States.
In the 1980s, such vehicles were involved in four accidents at speeds of around 140km/h. The 1980 Bushey derailment was caused by a broken rail; the 1984 Morpeth derailment resulted from overspeeding whilst the 1986 Colwich collision between two trains resulted from a red signal being passed. Although a driver died at Colwich, all passengers survived these three accidents, despite coaches overturning, falling down embankments or piling into each other. Sadly, the 1984 Polmont collision with a cow resulted in 13 fatalities when the shallow cutting side forced the derailed leading coach into a vertical position during the accident.
More recent fatal accidents involving Mark 3 coaches were those at Southall, Ladbroke Grove, Ufton level crossing and Carmont. These were particularly violent accidents involving severe rates of deceleration. The inquiry reports (noting that Carmont is still awaited) commented on how the survival space of the Mark 3 coaches had generally been well maintained.
However, there was concern about the number of casualties. The Ufton level crossing report highlighted the requirement for bogie retention as two coaches lost survival space due to an impact from a detached bogie. It also noted that some of the fatalities were of passengers ejected through broken windows and recommended that laminated glass – which was already mandatory for new trains – should be fitted to existing coaches when undertaking refurbishment.
Eliminate the risk
Research in the 1990s attempted to understand these issues better so that suitable requirements could be mandated for new trains. For crashworthiness, these are covered by two British Standards.
BSEN12663 specifies the longitudinal compression and tension forces that locomotives and passenger coaches must withstand. BSEN15227 concerns the dissipation of a train’s kinetic energy in an accident – in a much shorter distance than usual – by such features as anti-override devices, provision of survival space and crumple zones.
BSEN15227 includes four collision scenarios for different types of railway operations:
- front-end impact between two identical trains
- front-end impact between different types of railway vehicle
- front-end impact with a large road vehicle on a level crossing
- impact into a low obstacle (car on a level crossing, animal, rubble etc).
EN15227 recognises that there are limits to what any structure can withstand. It requires main line trains to withstand a collision with another train at the low speed of 36km/h and with a 15-tonne deformable object at a level crossing at a speed between 50km/h and 110km/h that must be determined by the railway authority, based on the railway’s characteristics. Serious train accidents often occur at much higher speeds, with a considerable effect on the energy to be dissipated.
The standard defines “withstand” to mean:
- reduce the risk of overriding
- absorb collision energy in a controlled manner
- maintain survival space and structural integrity of the occupied areas
- limit the deceleration
- reduce the risk of derailment
- limit the consequences of hitting a track obstruction.
Keeping vehicles upright
BSEN15227 assumes that the train is on the track at the time of the collision and aims to retain it on the track during and following the collision. This affects the design of couplers and the effect of ‘weight transfer’ or ‘unloading’ arising from the shock of a collision.
The standard accepts that no rail vehicle can be expected to withstand a particularly severe collision without some loss of survival space as some collisions are not within its scope. It notes that collisions following derailment are relatively uncommon events and it is not possible to predict the behaviour of a derailed train. However, the standard’s protection measures are considered to mitigate the effects of such incidents.
Even low-speed collisions can have unexpected outcomes, such as the derailment of a new Intercity Express Train at Neville Hill depot during a circa 20km/h collision with a High Speed Train in November 2019 where some bogies unexpectedly derailed. This was the first accident in the UK involving a train designed to comply with BSEN15227.
Research into what happens in a train accident identified that people might be ejected though broken windows, injured by outside objects entering a vehicle, flying fixtures and fittings, or that people might be injured when hitting unyielding fittings. The inquiries into the Hatfield, Great Heck, Ufton level crossing and Grayrigg accidents made recommendations to address these issues, which included fitting laminated glass to the windows and interior divisions, preventing seats coming adrift, shaping tables to minimise injury and preventing interior fittings turning into missiles in an accident. Many of these requirements are incorporated in Railway Group Standard GMRT2100 for new vehicles and they are also often implemented on existing vehicles during refurbishment.
Inquiry recommendations have also been made about bogie retention and designing couplers to keep vehicles upright and coupled. However, in extreme conditions, vehicles may become uncoupled and bogies become detached – as they were at Hatfield – despite their securing method being to specification. Furthermore, in extreme circumstances, people inside trains will be subject to forces that are, sadly, not survivable, such as in the 1984 Polmont accident.
New standards, existing vehicles
Introducing a new crashworthiness requirement does not make the existing vehicles unsafe. UK railways generally apply new standards in the expectation that improvements applied to new products will gradually filter though the network and, where feasible, to existing vehicles when they are refurbished.
In addition, older vehicles in front-line service tend to be transferred to less arduous duties as they age. HSTs and their Mark 3 coaches are now mostly used on secondary Intercity services with a top speed of 160km/h – rather than the former 200km/h – and have been upgraded with laminated exterior glass, more sturdy interior fittings and improved means of escape.
An exception was the Mark 1 coach design – which was considerably less crashworthy than modern designs – and the 1999 Railway Safety Regulations mandated that they should be withdrawn from main-line service by 2005, subject to certain exemptions, such as charter and heritage fleets.
There have been suggestions that the almost 50-year-old Mark 3 coach design should similarly be withdrawn as it lacks modern crashworthiness features. Yet various inquiries have commented favourably on the Mark 3 coach’s crashworthiness. For example, the Ladbroke Grove report stated: “Other accidents involving mark III (sic) vehicles have demonstrated good crashworthy performance, providing excellent protection for passengers.”
Furthermore, it is difficult to assess the actual safety benefit of modern crashworthiness features such as crumple zones and override protection due to the relatively small number of accidents since 2000. Of these, three involved modern but pre-BSEN15227 stock – a Class 365 at Potters Bar in 2002, a Pendolino at Grayrigg in 2007 and a Class 350 in Watford Tunnel in 2016 when a comparatively low speed (55km/h) two-train collision occurred after one vehicle derailed on a landslip.
At Potters Bar, the bogies of the rear-most vehicle of a four-vehicle unit travelling at 160km/h took different tracks due to defective points. This vehicle came to a halt on the station platforms at 90 degrees to the track. The coach remained intact. Its crumple zones did not collapse as there were no high longitudinal forces induced on the vehicle. Its bogies became detached and the inquiry concluded that it was unreasonable to design for bogie retention in such an accident.
Energy is half the mass multiplied by the square of the speed. Using an example of a typical four-vehicle train travelling at the BSEN15227 specified 36km/h, hitting an obstacle will dissipate its kinetic energy of circa 7.5MJ very rapidly. For a buffer stop collision at circa 15km/h, the kinetic energy is circa 1.4MJ. With modern stock, it is reasonable to expect such a collision to be survivable compared with the Mark 1 stock involved in the 1991 Cannon Street buffer stop collision which offered little protection.
At higher speeds, the energy to be dissipated increases significantly. A train travelling at 160km/h has over 20 times the kinetic energy of its 36km/h value. For heavier and faster trains, the value is higher still. For example, the kinetic energy of an 11-coach Pendolino travelling at 200km/h is nearly 900MJ. The front end of the Class 390 can absorb 3MJ and an intermediate carriage end about 1MJ.
In normal operation, this large amount of energy is dissipated gradually as the train is braked to a halt. In a collision, the damage suffered is a function of the rate of energy dissipation which, in turn, is a function of the deceleration rate. For a train collision at high speed, the deceleration rate depends on the stopping distance. Whereas the normal service stopping distance from say 100km/h might be 750m, a collision might lead to a train stopping in just 5m and the deceleration would be in the order of 8g, leading, sadly, to damage and injury.
The coaches of the Pendolino train involved in the 150km/h Grayrigg accident suffered little loss of survival space and their rigid couplers generally kept vehicles attached and in line with each other as the train came to a stand within 13 seconds in a flat muddy field at the bottom of an embankment. Only two of the 18 bogies on the train detached completely. The report noted that features such as crumple zones and override protection were not required in this event and that research was needed into the way multi-axis accelerations cause injuries in such derailments.
By contrast, in the recent fatal Carmont accident, the Mark 3 coaches were all detached from each other and some suffered severe damage. The preliminary report’s photographs show that the locomotive did not travel further after it hit the embankment adjacent to the bridge. Of all the accidents considered in this article, Carmont was particularly violent as it was the only one where a train was stopped by an immovable object.
This shows that it is unhelpful to make specific comparisons between derailments. However, the historical record shows that Mark 3 coaches have ensured the survival of all their occupants in similar high-speed impacts. Furthermore, as collisions between trains and with obstacles are rare events, crumple zones and override protection have yet to prove their worth, noting again that these features may well be ineffective if a collision follows a derailment.
In the 13 years between the fatal Grayrigg and Carmont accidents, Britain’s train fleet travelled around ten billion miles. So, whilst this article has of necessity focused on the awful consequences of high-energy train accidents, the reality is that rail travel is extremely safe in the UK. During this period, the average passenger or train crew member would have had to make the equivalent of roughly 10 million return journeys from London to Glasgow before their train was involved in a serious accident.
It is also a testament to the railway’s high safety standards that a high priority is given to ensuring that such rare incidents are survivable, as far as possible. In its interim report into the Carmont accident, RAIB advised that the final report will consider this issue by investigating the “crashworthiness of rail vehicles in high energy accidents”. Rail Engineer will return to this topic after publication.
With grateful thanks to many colleagues who have contributed to the debates that led to this article being written and for generously reviewing the text.