article

Tunnel fire safety systems

Posted: 7 December 2007 | Dr. Ricky Carvel, Assistant Director, BRE Centre for Fire Safety Engineering, University of Edinburgh | No comments yet

In the past decade, over four hundred people worldwide have died as a result of fires in road, rail and metro tunnels. Fires in tunnels have destroyed over a hundred vehicles, brought vital parts of the European road network to a standstill – in some instances for years – and have cost the European economy billions of Euros. Tunnels are being upgraded, research is being carried out and new technologies are being developed, but are our tunnels becoming safer?

The recent fire in the Burnley Tunnel in Melbourne, Australia (23 March 2007) involved a collision of two heavy goods vehicles (HGV) and two cars. The fuel tank on one of the HGVs ruptured in the crash, resulting in a small explosion and a fairly large fire was produced instantly. This fire was probably more severe than the fire which resulted from the collision of two HGVs in the St Gotthard Tunnel, Switzerland, in October 2001. That spread to involve a long line of vehicles, claimed 11 lives (nobody died in the actual crash) and destroyed a significant length of the tunnel interior. Yet the fire in the Burnley Tunnel did not spread much and no lives, beyond the three who died in the collision, were lost.

In the past decade, over four hundred people worldwide have died as a result of fires in road, rail and metro tunnels. Fires in tunnels have destroyed over a hundred vehicles, brought vital parts of the European road network to a standstill – in some instances for years – and have cost the European economy billions of Euros. Tunnels are being upgraded, research is being carried out and new technologies are being developed, but are our tunnels becoming safer? The recent fire in the Burnley Tunnel in Melbourne, Australia (23 March 2007) involved a collision of two heavy goods vehicles (HGV) and two cars. The fuel tank on one of the HGVs ruptured in the crash, resulting in a small explosion and a fairly large fire was produced instantly. This fire was probably more severe than the fire which resulted from the collision of two HGVs in the St Gotthard Tunnel, Switzerland, in October 2001. That spread to involve a long line of vehicles, claimed 11 lives (nobody died in the actual crash) and destroyed a significant length of the tunnel interior. Yet the fire in the Burnley Tunnel did not spread much and no lives, beyond the three who died in the collision, were lost.

In the past decade, over four hundred people worldwide have died as a result of fires in road, rail and metro tunnels. Fires in tunnels have destroyed over a hundred vehicles, brought vital parts of the European road network to a standstill – in some instances for years – and have cost the European economy billions of Euros. Tunnels are being upgraded, research is being carried out and new technologies are being developed, but are our tunnels becoming safer?

The recent fire in the Burnley Tunnel in Melbourne, Australia (23 March 2007) involved a collision of two heavy goods vehicles (HGV) and two cars. The fuel tank on one of the HGVs ruptured in the crash, resulting in a small explosion and a fairly large fire was produced instantly. This fire was probably more severe than the fire which resulted from the collision of two HGVs in the St Gotthard Tunnel, Switzerland, in October 2001. That spread to involve a long line of vehicles, claimed 11 lives (nobody died in the actual crash) and destroyed a significant length of the tunnel interior. Yet the fire in the Burnley Tunnel did not spread much and no lives, beyond the three who died in the collision, were lost.

The most striking difference between these two tunnels is that the Burnley Tunnel has a sprinkler system – or more precisely a deluge system – which operated automatically shortly after the fire was detected. Over four hundred people safely evacuated from their cars and escaped from the tunnel because the deluge system was operational.

It has long been known that, in buildings at least, sprinklers save lives. But can this reasoning be applied to tunnels? Certainly in the Burnley Tunnel it would appear that lives were saved as a result of the deluge system, but was that a unique situation, or are sprinklers the solution to tunnel fire problems?

For many years now, the advice of the regulatory bodies, specifically the World Road Association (PIARC) and the National Fire Protection Association (NFPA) in the USA, have taken a stance against the use of sprinklers in road tunnels (although this is now changing). One of the primary reasons for this was that sprinklers tend to hinder the escape of people from the tunnel; visibility will be reduced, stratification of smoke will be destroyed, steam may be produced and there may be other risks associated to extinguishing certain fuel types with water. As a consequence of this, it became commonly accepted that sprinklers were not a life safety installation and should only be used to suppress or contain a fire once all the people had made their escape.

In most life safety strategies, the objective is to enable the escape of all people from the incident location to a place of safety in the shortest timescale. In order to achieve this, safety systems must be designed to ensure that all escape routes are free from smoke, heat, water sprays or anything else which would hinder egress, wherever possible. Thus, historically, the primary life safety system for tunnels has generally been the ventilation system; intended to control smoke and keep escape routes free of smoke.

However, the reasoning used in the fire strategy for the Burnley Tunnel (and several other road tunnels in Australia) as well as an increasing number of tunnels in Europe, is that the advantages of containing, suppressing or hopefully even extinguishing a fire outweigh the disadvantages of impeding the escape of the people by drenching them with water and reducing their visibility.

The old advice of PIARC and the NFPA appears to have been based largely on expert opinion, supported with only a little experimental or operational evidence and only slight consideration for the underlying science. Over the years, research has shown that several of the assumptions built into the old advice were flawed or were only related to specific scenarios and not generally applicable to all tunnel situations. Recent publications by PIARC have given a more open view of suppression systems, highlighting both the benefits and the disadvantages, although their use is still not promoted.

The decision to use a suppression system in the Burnley Tunnel and elsewhere was also a decision based largely on the opinions of experts, with a little (but not much more) experimental and operational evidence to support it. While a lot of European funding has gone into the area of tunnel fire protection in recent years, much of this has been diverted towards product development rather than fundamental research into tunnel fire phenomena. By promoting product development, the industry obtains technologically advanced systems rapidly, but without research how do we know which systems to develop, or if the systems developed will actually work in practice?

The development of the new ventilation system in the refurbished Mont Blanc Tunnel was carried out with a clear knowledge of research into the influence of ventilation on fire development. Research had shown that the severity of a HGV fire could be greatly exacerbated by the careless use of longitudinal ventilation.1

Anyone who has ever lit a wood fire knows that blowing on the fire helps to fan the flames and grow the fire. However, until recently it had not been known to what extent ventilation could enhance the burning of fires in tunnels. For HGV fires in particular, the influence is quite severe; in a small tunnel a longitudinal ventilation velocity of only 3 metres per second (m/s) is likely to enhance the Heat Release Rate (HRR) – a good measure of the severity of a fire – by a factor of about three compared to a fire with minimal longitudinal ventilation.2

This research appears to have been heeded in the design of the new ventilation system in the refurbished Mont Blanc Tunnel.3

The new system incorporates a fully transverse ventilation system with dampers on the openings to the extract ducts and an intelligent longitudinal ventilation system. In conventional longitudinal ventilation systems, the jet fans are used to increase the airflow to a prescribed velocity in the required direction, whereas in the Mont Blanc tunnel, the jet fans are used primarily to reduce the longitudinal airflow at strategic locations. Thus, in normal operation, the jet fans can be used to counter the influence of the local weather. In fire conditions, the jet fans are configured to control the ventilation such that there is negligible flow of air past the fire location, so the severity of the fire is not enhanced. To complement this, the transverse ventilation is configured to provide maximum extraction from the openings closest to the fire location. In the event of a vehicle fire, this unique combination of systems should keep the tunnel smoke free on both sides of the incident and will hopefully minimise fire size as well.

The experimental tunnel fire experiments in the 1990s and early 2000s, particularly the four HGV trailer fire tests carried out in the Runehamar Tunnel in 2003, have shown that HGV fires in tunnels can grow to a magnitude far greater than had been generally expected before. It is now clear that a fire on a HGV carrying non hazardous goods, in confined tunnel geometry and subject to a longitudinal airflow, can reach peak HRR to the order of 200 MW. Fires of this severity are beyond the capabilities of most, if not all, types of suppression system. Tunnel fire safety systems must therefore seek not to attempt to suppress such fires, but to prevent any fires from growing to this magnitude. This is only possible with early detection and early intervention. Unless a detection system is capable of identifying a fire in a tunnel and accurately pinpointing its location, it cannot be used to activate a fixed suppression system. Indeed, it is important that the detection and suppression systems are not considered as independent systems, but as part of one integrated tunnel safety system.

The same is true for ventilation systems. These must be designed and configured to work hand-in-hand with the other parts of the tunnel safety system. There also needs to be a change in our thinking of how to use the ventilation system in a tunnel. The mode of operation with a suppression system should be totally different to the mode of operation in a tunnel without one. In the forthcoming PIARC document, Road Tunnels: An Assessment of Fixed Fire Fighting Systems, it is suggested that one of the requirements of a fixed fire fighting system is that it should be ‘designed to handle air velocities in the range of 10 m/s’. The implication of such a statement is that a suppression system should be designed to operate with the ventilation system running at maximum capacity. However, it is highly unlikely that maximum ventilation flow will be the optimum operating conditions for any water based suppression system. While this may not be a problem for deluge systems like the one in the Burnley Tunnel, it could be a particular problem for water mist systems such as those recently installed in the M30 project in Madrid and the tunnel sections of the A86 near Paris.

Water mist droplets are tiny compared to conventional sprinkler droplets. It is claimed that this makes them more efficient at suppressing certain types of fires, but it is also clear that this makes them far more susceptible to the influence of longitudinal ventilation. Droplets with a diameter of 100 mm (reasonably typical for water mists) have a terminal velocity in air of about 0.2 m/s which, with no other forces acting on them, will fall from the ceiling to the roadway of a 6m high tunnel in about 30 seconds. With a longitudinal airflow of 10 m/s this would mean that they would hit the deck about 300m downstream of the nozzle location. Smaller droplets will be carried even further. Of course, the dynamics of an ensemble of millions of mist droplets are different to a single droplet in air, but it is clear that high airflow velocities will still carry the droplets a significant distance downstream.

If designing a suppression system for an existing tunnel with an existing ventilation system, it makes little sense to design the system to operate within a ventilation strategy which was originally devised to provide smoke control in a tunnel without a suppression system. In such circumstances we need to go back to the drawing board and design the best integrated detection, suppression or ventilation system that we can; a unified tunnel safety system. The requirements for operation of the ventilation system will, almost certainly, be different for a tunnel with a suppression system compared to one without.

It is at this point that someone will usually raise the issue of critical ventilation velocity to prevent backlayering of the smoke from a fire. Backlayering is the phenomenon, often observed in tunnel fires with insufficient ventilation flow, where the smoke from the fire advances upstream of the fire location against the prevailing flow. However, it must be noted that all studies of backlayering carried out to date have been for fires without sprinkler systems. The flow dynamics of a tunnel fire without a suppression system will be completely different compared to one with such a system. Therefore the ventilation required to control backlayering will be completely different also. Research on this is required and at present, it is not possible to say what changes to ventilation strategies will be required.

Indeed, research is required on a great many aspects of tunnel fire behaviour. Tunnels are complicated environments and the behaviour of a fire in such an environment is particularly complex. While a number of large scale tests have been carried out over the years (and future analysis of the existing data may yet increase our knowledge of tunnel fires), this data is not sufficient to address all the unknowns in tunnel fire behaviour. The answers are not to be found in modelling, either. Current computer fire models are versatile tools which incorporate some of the current knowledge and understanding of flow and combustion dynamics, but they don’t, and indeed cannot, hold the answers to many of the unresolved issues.

The only way to answer such questions adequately would be to carry out systematic detailed research programmes not tied to the development of specific product types. Research programmes need to be comprehensive in nature, involving both experimentation and model development. In the case of tunnels, the required experimentation covers all scales. There are answers that may be found using small and medium scale testing, but some questions will require full scale experiments. Furthermore, instrumentation and modelling associated with these tests needs to be detailed enough to allow for the development and validation of the models. This is essential because the models will be the tools used for the performance assessment of real tunnels. Given the complexity of the physical interactions in tunnels, if experimental validation is to be used for performance assessment, the studies should ideally be carried out in the tunnels the systems are intended for. Changes in geometry, gradient, fan positions and other factors will lead to different flow behaviour in different tunnels. In most cases this will not be possible, thus extrapolation of experimental information will be necessary. This extrapolation requires robust and well validated models. Existing models provide some extrapolation capabilities but are by no means robust, well validated or comprehensive.

The main goal of tunnel fire safety research is to define tools and mechanisms by which performance assessment of tunnels and tunnel fire safety systems can be made in a robust manner. Many tools are available, but none of them cover all the relevant physical phenomena, nor are they robust enough that they can be used as engineering tools for design purposes. The advancement of these tools will require experimentation at all scales, development of better diagnostic techniques and especially the development and validation of models that will represent the ultimate design tools. Only well funded overarching research programmes that integrate the efforts of all members of this scientific community will be capable of delivering such tools and help the designers address such a complex problem as tunnel fire safety.

References

  1. Carvel et al., Variation of Heat Release Rate with Forced Longitudinal Ventilation for Vehicle Fires in Tunnels, Fire Safety Journal, Volume 36 (2001) pp. 569-596.
  2. Carvel and Beard, The Influence of Tunnel Ventilation on Fire Behaviour, The Handbook of Tunnel Fire Safety, (ed Beard and Carvel), Thomas Telford, 2005, ISBN: 9780727731685, pp. 184-198.
  3. Brichet et al., The new ventilation systems of the Mont Blanc Tunnel active smoke control: From simulation to successful operation, Proc. 4th International Conference on Tunnel Fires, Basel, Switzerland, 2-4 December 2002, pp.95-113.

Bibliography

  • Beard and Carvel (Eds), The Handbook of Tunnel Fire Safety, Thomas Telford, 2005.
  • H. Ingason (Ed), Proceedings of the International Symposium on Catastrophic Tunnel Fires, Borås, Sweden, 20-21 November, 2003.
  • Fire and Smoke Control in Tunnels PIARC report 05.05.B, ISBN: 2-84060-064-1, 1999.
  • Systems and equipment for fire and smoke control in road tunnels PIARC report 05.16.B. ISBN: 2-84060-189-3, 2007.
  • Road Tunnels: An Assessment of Fixed Fire Fighting Systems PIARC report, forthcoming.
  • NFPA 502: Standard for Road Tunnels, Bridges, and Other Limited Access Highways, Revised 2004.
  • Fogtec Water Mist: www.fogtec.com
  • Marioff Water Mist: www.hi-fog.com
  • Aquasys: www.aquasys.at