Introduction

Fire is a serious, potentially catastrophic event. At the turn of the millennium, transportation tunnels experienced several fatal fires, including Baku1, Mont Blanc2, Daegu3, Kaprun4 and others. Every major fire, from Mont Blanc2 to Grenfell Tower5 to the future’s as-yet unknown major fire event, reminds us of the terrible consequences of fire if we fail to plan for it. We must continually be vigilant.

However, much has changed since the 1990s. Since those tragic tunnel fires, there have been significant improvements. The incidents incited a drive to improve tunnel fire safety by (i) reducing the occurrence of fire, and (ii) enhancing the survivability for occupants in event of fire. Preventative measures, such as fire-hardened rail vehicle construction, have aided in reducing fire occurrence. Provision of multiple layers of safety systems has improved survivability. Today, multiple systems provide a robust, resilient set of fire safety measures. Legislation, regulations, standards and guidance have been developed to provide appropriate minimum requirements. The Burnley Tunnel fire6 was an early signifier of improving tunnel fire safety.

New technologies are changing the usage of tunnels. In road tunnels, the increasing adoption of alternative fuel vehicles and the arrival of semi-autonomous and autonomous vehicles bring new hazards which challenge operations and incident response planning. In rail tunnels, challenges include the spread of Automatic Train Operation (ATO) and similar systems, and the adoption of driverless trains.

Tunnel assets are now reliable with good safety levels and may be safer than associated surface assets. This leads to some questions from operators. Are there efficiencies in tunnel safety that could be implemented to improve overall network safety? Does the industry continue to provide all current tunnel systems? What of the challenge of new technologies?

Sustainability

Sustainability is a critical criterion in project approvals. Tunnels consume significant money, resources and energy, both in construction and operations. Via the project approval process, society has decided the asset serves a need. The asset’s sustainability implications are accepted in return for societal benefit. However, sustainability remains relevant throughout the design, and engineers can significantly affect the sustainability of the asset through the design process.

Tunnel systems and safety objectives

Transportation asset operators must consider several interlinked, often conflicting, objectives. These include safety, business continuity, commercial viability, sustainability, risk management, and others. Assessing the relative importance of each objective is key to making efficient decisions about construction and operation of the asset.

Tunnel system requirements derive from the safety objectives and may be categorised by functional objectives. Normal operations and environmental needs may also have requirements that influence systems decisions. Table 1 shows a classification of systems in a tunnel. Selection of tunnel systems will depend on the facility’s usage. Facilities accessed by the public, such as transportation tunnels, will have a different safety profile than utility and special use tunnels where only trained, authorised persons have access. Examples of the latter facility include cable tunnels, pipeline tunnels and special usage facilities, e.g., the CERN Large Hadron Collider7.

Tunnel systems incur significant capital cost, and operating costs are a burden throughout the life of the facility. When time comes for refurbishment/ rehabilitation, systems costs may be the bulk of the total refurbishment costs.

The present

Requirements represent the current mandatory standard of practice and summarise current knowledge. Requirements exist in a hierarchy (Figure 1). Mandatory legislation sits at the top. Legislation is typically high-level and may lack implementation details. It is supported by regulations which provide this detail. Standards and codes may have a legal status equivalent to regulations. If not mandatory, they supply guidance and best practice documentation. Standards may be deemed mandatory at the project level, despite lacking the legal authority of the levels above.

Figure 2 illustrates the cycle of requirements development. Research generates new knowledge and over time gains industry approval. Where justified, this becomes codified in the requirements and is applied. Then feedback is received, and this may lead to new research and knowledge.

Figure 3 summarises factors that can influence the development of requirements. This includes not just history and innovation, but also includes societal attitudes, such as sustainability.

Figure 4 and Figure 5 show passenger rail incident data from the EU8 and the USA9, 10, 11. For both networks, train fires represent a small fraction of all incidents. Figure 6 shows data for the UK heavy rail passenger network12, 13. Since 2001, the incidence of train fire has declined by more than 80 per cent, with no serious fires recorded. The Transport Safety Board (TSB) of Canada has published safety data on the Canadian heavy rail network14 that covers the period 2001-2015. Until 2010 it recorded fire as a specific cause of casualties. Subsequently, the TSB stopped recording fire-related casualties as the incidence was deemed insignificant.

Road tunnel data reveals a similar picture. Figure 7 shows data for the incidence of fire in Austrian road tunnels15 between 2006-2012. Figure 8 shows regional road network data from the United Kingdom between 2009- 2015, including data for tunnels on dual carriageway highways. Fires are a minor cause of incidents.

The UK data also reveals that for tunnels on dual carriageway highways, the tunnels’ safety level is better than for the surface highway. The highest risk locations are single carriageway roads and junctions.

A survey of Norwegian tunnel fires17 also found a low incidence of fire. Most fires involving personal injury occurred due to accidents. The authors concluded that many of the injuries reported in these incidents were due to the traffic incident rather than the fire.

The key points from the available data are:

  • Tunnel fires are statistically rare, and the rate of occurrence is decreasing.
  • The occurrence of serious fires, leading to serious casualties and or loss of life, is extremely rare.
  • Most incidents that lead to casualties derive from causes other than fire.
  • Most incidents that lead to casualties occur on the surface assets of the network.

The future

Tunnels are now reliable and exhibit acceptable levels of safety. Some may exhibit a higher safety level than network surface assets. Consequently, operators have begun to challenge the current expenditure of both CO2 and funds on tunnel systems. They recognise that tunnels must be safe, but they must also consider overall safety. There is a desire to distribute safety resources more efficiently across the network.

Additional challenges arise from new technologies. In road tunnels, this includes hybrid, electric, and fuel cell vehicles, which bring new fire hazards with them. Another is the spread of semi-autonomous and autonomous vehicles, which bring new challenges to operations and incident response. In the rail industry this includes ATO/ ATC/ATP and driverless train operation, which may also revise operations and incident response. Ultimately, this could mean installing fewer safety systems, or different types of systems, and spending savings elsewhere on the network.

Thus, the effectiveness of tunnel systems should be estimated. Each system’s contribution to safety should be assessed to identify those with most value. There are many systems in modern tunnels. Some are strictly for fire safety, while others are multifunctional. To assess systems effectiveness, the safety objectives must be understood. Each functional group of systems has one or more safety objectives (Table 1).

All systems do not contribute equally to safety. Safety objectives may be satisfied when a system, although available, is not used. For example, the egress route, e.g., egress doors and a protected walking route, may not be used by occupants of a road tunnel. They may instead choose to walk along the roadway to the portal. In this instance the functional objective, egress from the tunnel, may be achieved even though not all occupants use the egress system. Safety objectives may also be met even if a system is not available. For example, train passengers evacuating to an emergency walkway may evacuate safely from the train even if communications between driver and passengers are not functional. In this instance, occupants could still evacuate safely despite this safety system’s unavailability. However, some systems are essential. For example, occupants would be unable to evacuate safely without tunnel lighting. Figure 9 shows a process for estimating the tunnel’s safety level.

Figure 10 shows an example process to estimate system effectiveness. Using expert knowledge and stakeholder consensus, success probabilities can be estimated when the system is available, and when it is not available. Failure probabilities can be estimated and a picture of systems effectiveness can be built up. From this, an assessment of the value of installing, refurbishing, upgrading, reducing or even deleting individual systems could be made. Reference18 details such a method applied to road tunnels.

This could be a powerful tool for regulators, owners, operators, insurers and others with responsibility for safety. Funding for infrastructure is perpetually squeezed, so available funds and carbon allocations must be spent efficiently. Figure 11 illustrates the dilemma faced by those responsible for funding decisions. As the fire risk profile in tunnels decreases (Figure 12), safety funding may be redistributed to parts of the network exhibiting higher risk.

Utility is a concept that is common in economics. Here, utility is used to assess the overall societal benefit. Safety is one aspect of an asset’s utility. Cost-benefit analysis can be used to assess the utility. The cost of providing safety may thus be estimated. The utility of safety may be expressed as:

U = F ( P, C ) where P = probability and C = consequences. Applying the concept of utility and using appropriate tools to assess the relative utility of different options may facilitate more efficient resourcing decisions (Figure 13).

With improved resource allocation, the network should experience improved overall safety and sustainability. The method should improve the ability to adapt to future challenges. This may lead to changes in the systems installed within tunnels, or adoption of new systems in response to changes in technology and usage. Systems that provide the most benefit may be retained or extended, while those of lesser benefit may be reduced or eliminated, provided the resulting tunnel safety level is acceptable. This could release resources to improve safety and sustainability across the network.

Conclusions

Advances in tunnel fire safety have been shown to reduce the risk of serious fire in tunnels. Operators find that their tunnel facilities may be safer than other parts of their network. They are asking if the expenditure on tunnel safety systems is fully justified. Engineers should respond to these queries and develop risk-informed methods to assess the evolving needs of tunnel safety as mobility technology evolves. Use of such risk-based methods could lead to more efficient use of resources across the wider mobility network, leading to an improved overall level of safety and sustainability in the future.