From its discovery in ancient times, humans have both been in awe of, and frightened by, fire. An uncontrolled fire is a risk to which we are all exposed and manage daily. Steam trains, of course, depend on a well-controlled fire for their very operation, although their emissions often cause uncontrolled fires on the lineside.

As with much of railway safety, lessons have been learned from accidents over the years. There have been many fires, often with tragic consequences, that have led to significantly improved fire safety. However, we could still be caught out and we must never be complacent.

I recall my introduction to fire safety in the early 1970s when many people still carried lighters. As a young engineer, I would find a new component that might be suitable for use on a train, and a colleague might whip out his lighter and subject it to the flame; a very basic fire test that separated the ‘awful’ from the ‘possible’.

There were also occasional incidents where a train would catch on fire several hours after coming out of service. Smouldering cigarettes, hardboard panelling and oil-based paints were not a good combination.

Lessons learned

In one incident in particular, a terrible choice of materials led to a devastating fire, fortunately without serious injury. On 23 June 1949, in Penmanshiel Tunnel, two carriages of an express train from Edinburgh to London were destroyed by “a fire of great ferocity which spread very rapidly,” to quote from the inquiry report.

These coaches, only two years old, had steel underframes and exterior panelling with wooden frames and interior cladding. The wood on its own was not the cause of the fire. The inspectors discovered that parts had been painted with a cellulose nitrate varnish that was shown to be extremely easy to set alight and to have exceptionally rapid rate of flame spread. It was thought that a discarded cigarette ash set the varnish alight.

There was some consternation from the inspecting officer that this varnish had been selected in the first place, given that the dangers of cellulose nitrate were well known at the time.

More recent incidents have led to big changes in the industry. On the main line, the Taunton sleeper fire in 1978, which killed twelve people, was caused by imported material (bedding) catching fire when stored against an electric heater. The Taunton fire led to fundamental changes to the design of the mk3 sleeper coaches which were being considered at the time.

For London Underground, it was the King’s Cross fire in 1987 which led to changes to underground station legislation, management process, improved fire systems and the upgrade of materials for improved fire performance on stations and over 3,000 metro cars.

In terms of standards, these fires also led to engineers and chemists developing standards that drove the supply industry to develop better materials. LU developed its Code of Practice for the Fire Safety of Materials, the main line industry developed GM/RT 2130 – Vehicle Fire Safety and Evacuation and BS6853 – Code
of Practice for Fire Precautions in the Design and Construction of Passenger Carrying Trains followed. In 1989, an international committee was set up to develop a European standard that finally gained approval in 2013 and brings us to the subject of this article.


So much for a micro-history, the story was taken up by David Tooley in his recent lecture to the Institution of Mechanical Engineers, “Rolling Stock Fire Safety – Where Next?” David’s paper concentrated on EN 45545 revisions, FCCS systems, PHRR calculation methods and CFD.

Baffled by standard numbers and acronyms? Don’t worry, all will become clear!

Firstly, some recent background. We all know that passenger numbers are increasing and that train movements and mileage are also increasing. Against this expansion, however, the number of reported train interior fires has fallen by 95 per cent since 2004, with just 10 fires being reported in 2015.

Technical fires on trains (engine fires, electrical fires and so on) have also fallen, but not to the same extent, giving an overall total of just 30 fires on the UK main line railway reported in 2015, two thirds of them technical fires. Most of them are of very low power and flashover fires are very rare. This is good news, but, as ever with safety, one cannot be complacent.

EN 45545

David moved onto the topic of the current standards applicable to train fire safety. He mentioned EN 50553 – Requirements for Running Capability in Case of Fire on Board of Rolling Stock and EN 45545 Railway Applications: Fire Protection on Rolling Stock. The latter is the Euronorm that is replacing BS 6853 as the requirements to control the risk of ignition, development, consequences and management of fire on trains. It is a suite of standards in seven parts covering:

  • General
  • Fire behaviour of materials and components
  • Requirements for fire barriers
  • Requirements for rolling stock design
  • Requirements for electrical equipment
  • Fire control and management systems
  • Requirements for flammable liquid/gas installations.

EN45545 took a long time to develop, mainly because of very different approaches to fire safety in EU member states, but finally appeared in 2013. There was much concern in the UK that some requirements were being significantly diluted, especially those for upholstered seats, and the UK’s concern was such that it voted against its adoption. The main problem was believed to be the seat test, which seemed to be inadequate to eliminate poorly performing products.

RSSB commissioned research projects T843 and T1012 to explore formally the weaknesses and strengths of the new standard. These reports showed that the transition to EN 45545 was a risk to the continued improvement to fire safety compared with BS 6853. David illustrated the results of tests on a number of seat designs that would fail the BS 6853 criteria, but would pass the EN 45545 test.

These results were instrumental in persuading CEN to set up a new working group to introduce revisions to EN45545. This work started in 2015 and needs to be completed by 2018 to coordinate with changes to some of the TSIs, which must become mandatory by that date.

The principal change is to increase the heat source applied during the seat test from 7kW for three minutes to 15kW for the same time. Moreover, the assessment will, in future, include both the heat release rate and smoke produced by the sample. Other changes have been made to improve test repeatability. David reported that this new test does indeed separate good seats from bad. David also touched briefly on work to transpose the results of recent research (the Transfeu programme) into standards.

Fire Containment and Control Systems (FCSS)

The presentation moved onto the control of the situation once a fire has started on a moving train.

FCCS is the process of making sure that it is tenable for passengers to survive in a fire event until they can reach a place of relative safety (another acronym explained!). Often, this is achieved by provision of physical fire barriers and by moving people away from the fire beyond the fire barrier. Indeed, the Safety in Rail Tunnels TSI requires fire barriers every 30 metres for some types of train unless an alternative ‘equivalent’ FCCS protection is provided.

It often possible to have a fire barrier between coaches as they are rarely longer than 30 metres although, increasingly, suburban trains are provided with open wide gangways between coaches. This is also an issue on metro trains (where TSIs are not applicable). Clearly, where open wide gangways are provided, fire barriers are impracticable.

A variety of techniques have been used instead to control smoke and/or fire on such trains including airflow management systems on Thameslink and Crossrail trains and use of water mist systems on Italian trains. A standard under development for FCCS will set the requirements to ensure that conditions are tenable for passengers and staff on ‘adjacent’ cars. This will set the requirements for managing the development of a small suitcase fire. The aim is to be able to suppress or extinguish fires so that there is no effect outside a 30-metre zone around the fire with no physical barriers.

David added, however, that a colleague had carried out a survey and found that the only recorded event of carry-on luggage causing a fire was when someone took a motorcycle fuel tank on an aircraft. That said, deliberate acts cannot be ignored.

Peak Heat Release Rate (PHRR)

The next acronym, PHRR, – ‘the rate of heat energy released during a rail vehicle flashover fire’ is used to help define infrastructure safety, structural, and ventilation requirements. There is a hierarchy of needs when developing a system strategy for fire protection (thinking particularly about tunnels):

» The rolling stock engineer is concerned to design a train where its materials are really difficult to set alight but, if they do ignite, they should burn slowly with low smoke
and toxic products. This enables passenger evacuation and represents the short term – a few minutes.

» The station manager has a somewhat longer timescale to manage. The passengers are off the train, but have to be evacuated from the station, then safe access has to be provided for fire fighters.

» Infrastructure managers must ensure that the structure can withstand the worst case credible fire and the PHRR is used to define infrastructure ventilation and structural requirements – basically to ensure that the structure is not compromised by the fire.

This has probably been one of the most difficult areas to nail down. There are differing views about how to calculate and apply PHRR. Calculations have typically been carried out on spreadsheets and make several significant assumptions which are not valid for modern trains:

Illustrating the benefits of modern train design, David outlined the recent research project carried out in Sweden where they set fire to an older train in a tunnel by setting simulated luggage alight. Fire spread rapidly with a PHRR of 70MW after 10 minutes.

They then repeated the test with a simulated modern vehicle (panelling over all the old vehicle’s cladding with metal and fitting modern seats). The PHRR was much the same at 70MW but it took 100 minutes to develop. It was suspected that the old materials which had been hidden by the metal cladding were eventually involved in the fire, but the principle of slowing down fire development had been demonstrated.

David also highlighted an exercise carried out by London Underground in late 1990s where they demonstrated the fire performance of upholstered seats on a 1992 Central line car. Rather than use the normal test involving one standard no.7 wooden crib, eight of these, amounting to one kilogramme of timber, were piled on a seat and set alight. The estimated PHRR of the wood was 1MW, significantly in excess of the test specification. David illustrated the test with photos showing ignition, burning furiously, and the remains when all the wood had burnt which is when the fire went out without any fire suppressant being used.

The repair required a new single seat, a new melamine panel and some paint.

David mentioned calculation techniques developed for use in Singapore and the UK and applied elsewhere which, essentially, involves calculating the PHRR per unit area, HRRPUA (yet another acronym), which is calculated for each type of material taking account of ignition source and carry-on luggage and then summed for the overall vehicle. From work of this nature, a typical metro car is usually assessed as having a PHRR of approximately 8MW and a main line car of approximately 13MW. Is this a real or credible value, David asked? That takes us onto the next section.

CFD Modelling

Computational Fluid Dynamics (last acronym!) is a process that is used to model how fluids (liquids, air, smoke) move over time in a given space. It has been used to assess PHRR by some train builders and they have been able to establish PHRR for metro cars of approximately 3MW, less than half the value calculated by other means.

So, which is the right value, 8MW or 3MW? If the lower value is the right answer it could significantly reduce system cost as the infrastructure controls would be less onerous. David explored whether this is a valid process for rolling stock, whilst recognising that it has been used in infrastructure projects for many years. He highlighted that a realistic model requires significant processing power to calculate the output. However, having such a model would allow both the FCCS and vehicle PHRR to be assessed, along with modifications to the vehicle and changes to the fire scenarios.

If CFD is to become an accepted technique for certifying designs it needs to be developed into a standardised process.

Drawing his fascinating presentation to a close, David explained how train fires are shown on the RSSB risk model (approximately the same risk as a two-train collision in a station from permissive working) and reminded everyone that, whilst fires are low probability events, they have a high potential for injury or fatality, so there is no room for complacency.

Written by Malcolm Dobell

Thanks to David Tooley, principal rolling stock engineer at Mott MacDonald, for his support in developing this article.