Fires in tunnels can lead to many deaths and injuries in a single incident or, at the very least, significant disruption and financial loss. Jeremy Fraser-Mitchell unveils recent research into sustainable and low-cost improvements in tunnel fire safety.
The safety of tunnels has been put in question in the past decade after a number of incidents involving significant loss to human life. In particular, the fires in the alpine tunnels at Mont Blanc in 1999 (39 fatalities), Tauern in 1999 (12 fatalities) and Gotthard in 2001 (11 fatalities) all highlighted shortcomings in the levels of safety inside existing road tunnels. Rail and metro tunnels have also come under scrutiny, with the fire in the metro in Taegu in South Korea (197 fatalities), and the metro fire in Baku, Azerbaijan (289 fatalities) being of particular note. Furthermore, the cost of repair and the economic penalty incurred while a major tunnel is closed can be very high.
As part of its 5th Framework Programme, the European Commission has supported a 4-year research project called UPTUN (UPgrading TUNnels) to develop tunnel safety upgrading methodologies. The project, which began in September 2002, involved some 40 European organisations including BRE Global. The Commission covered 50% of the project costs, with the participating organisations providing the remaining funding. BRE Global’s activities in UPTUN were supported by the BRE Trust.
The primary objective of the research was to develop innovative, sustainable and low cost measures to improve tunnel safety, by reducing the risk and consequences of fire in existing road and rail tunnels. Examples of issues that were addressed included improved means for safe evacuation, better transfer of information (e.g. signage), enhancements to traditional means of fire control and better protection (e.g. detection, mechanical ventilation and structural protection). Furthermore, attention was also given to newer and more innovative measures for fire and smoke control, e.g. water mist fire suppression and balloons that inflate to prevent the passage of smoke and heat.
UPTUN was divided into seven work packages:
1. Prevention, detection and monitoring
2. Fire development and mitigation measures
3. Human response
4. Fire effects, tunnel performance and structural response
5. Evaluation of safety levels and upgrading of existing tunnels
6. Full-scale experimental analysis
7. Promotion, dissemination, education and socio-economic impact
BRE Global contributed to work packages 2, 3, 5 and 7. This included the design fire specification, modelling smoke movement and alternative ventilation strategies, studying means of escape, performing egress calculations and developing training material. Findings from various work packages were consolidated into a holistic tunnel upgrading methodology, which allows the benefits of alternative measures to be evaluated in terms of both life safety and economics.
Design fire and CFD calculations
One of the tasks of working on fire development and mitigation measures (work package 2) was to provide quantified measures of the time-dependent development of hazardous conditions within the tunnel environment for selected design fire scenarios. This was achieved by reviewing previous studies and by undertaking new experiments. To complement this, numerical simulations were performed using CFD (computational fluid dynamics) to provide detailed predictions of thermal and other hazard levels within the tunnel space.
BRE Global’s CFD fire model, JASMINE, was used for simulating a set of parametric fire scenarios in longitudinally ventilated, two-lane road tunnels with rectangular and arch cross-sections. The fire source was indicative of an HGV fire, based on previous published information and from data from the Runehamar tunnel fire tests, conducted in 2003 in a disused road tunnel in Norway by Swedish and Norwegian transport bodies in collaboration with UPTUN. A peak heat release rate (HRR) of 200MW was achieved in one test, and even the least severe test achieved a peak HRR of 70 MW. Figure 2 shows the time-dependent thermal loads of the fire source terms used in the CFD simulations and Figure 3 shows the geometry used. The trailer space was represented by 20 ‘commodity packets’ from which the fuel was released.
With longitudinal ventilation, pollutants are forced along the tunnel in the direction of the air flow. In a fire emergency, the basic remit is to ‘push’ all the smoke in the downstream direction, and to maintain clear conditions upstream of the fire source, allowing means of escape in this direction as well as firefighting access. There is generally a minimum longitudinal air speed required to prevent fire gases from travelling in the upstream direction, referred to as the ‘critical velocity’. Where this condition is not met, the movement of smoke in the upstream direction is referred to as ‘back-layering’.
For a range of longitudinal ventilation levels, up to 5 ms-1 in the pre-fire condition, detailed predictions were made for the hazards associated with thermal and smoke exposure. Impulse (jet) fans were modelled by two (fixed value) momentum sources located approximately 100m upstream.
A comprehensive set of contour plots of gas temperature, smoke level (visibility distance) and carbon monoxide concentration was generated, using empirical correlations for soot particulate and carbon monoxide yields. Figures 4 and 5 illustrate two such contour plots for a 70MW peak fire inside the arch cross-section tunnel, where the vertical dimension has been stretched in the figures for ease of viewing. In one case, back-layering is present while in the other it is eliminated by ventilating above the critical velocity. Thermal and toxicity hazard exposure levels to people escaping have been calculated from the CFD data. Some of the findings from the CFD study were:
– Back-layering was eliminated at pre-fire longitudinal ventilation rates (critical velocity) of the order 3 ms-1 for all combinations of tunnel and fire size
– However, at ventilation rates above the ‘critical level’, the thermal hazard was significant in the first 100m downstream in all cases, and for the 70MW and 150MW fire scenarios the thermal hazard extended well beyond 100m in the downstream direction
– CO hazard was negligible for the 30MW fire, but was pronounced for the 150 MW fire due, primarily, to the total calorific value of the fire
-The CO hazard was mitigated by ventilating at a high rate, but this finding should be cautioned by the fact that this may then feed the fire with additional oxygen
Holistic tunnel upgrade model
An important element of UPTUN has been the development of a holistic tunnel fire safety evaluation and upgrading procedure for road and rail tunnels. This is being integrated with the work from another European project looking at pan-European guidelines for safety in tunnels, and in doing so extending the approach to cover non-fire accidents and incidents too. An important feature of the procedure is that it is fast to execute, enabling the benefits of various measures to be evaluated quickly.
The upgrade procedure involves smoke dispersion calculations using one- and two-dimensional models, egress and toxic dose calculations, risk assessment using event tree analysis and socio-economic modelling.
BRE Global contributed by enhancing its building fire risk assessment and egress model CRISP to include evacuation from road and rail tunnels (see Figure 6), and by coupling CRISP with tunnel-specific zone and network smoke movement models, including the multi-layer zone model FASIT. This enabled the combined influence of smoke movement and human decision making to be incorporated into the tunnel upgrading assessment. After examining reports of actual tunnel fires, new behaviour rules were added to CRISP, covering the movement of staff, passengers, drivers and the emergency services. Exit choice probability distributions were developed, and options for people moving in groups were also developed.
Figure 7 illustrates a typical graphical output from CRISP, where gas temperatures and smoke movement data has been integrated with the egress rules and the toxic dose model, enabling the response of people to be tracked in space and time. Note that at 15 minutes in this example there are two people (marked red to indicate a high level of accumulated thermal and toxic exposure) who are still relatively close to the source of fire and are in danger.
The upgrade procedure has been applied to tunnel example test cases to explore its capabilities. Here, various upgrade options were considered:
– A detection/early warning system, assumed to give a warning message three minutes after the fire has started
– Side exits and refuges, located 200m apart and where once a person has entered a side door, they are assumed to have reached a place of safety
– A suppression system, where the simplifying assumption was made that a 20MW fire would be extinguished and a 100MW fire contained once it had grown to 20MW
– Transverse ventilation smoke extraction
The benefits of the upgrading measures have been weighed against the cost, in terms of life safety and economics, of closing the tunnel for the required period and needing to re-route traffic. Depending on the measures adopted, a tunnel may need to be closed only at night, or it may need to be closed for some months or even for over a year.
The results have emphasised the importance of a holistic approach to the risk assessment. For some upgrade options, it is even possible for the risks associated with closing the tunnel to be higher than not upgrading the tunnel in the first place, e.g. due to diverting traffic onto other routes where the risks of an accident may be higher.
Finally, during tunnel upgrades, when specifying active and passive fire protection systems as an integral part of the design, it is recommended that third party certified products are used. Third party product certification schemes provide a means of identifying products which have demonstrated that they provide a requisite level of performance in fire, and also provide confidence that the products actually supplied are supplied to the same specification as that tested and assessed. In the UK, the Loss Prevention Certification Board (LPCB) operates various certification schemes for fire protection products. These are listed in the LPCB List of Approved Fire and Security Products and Services, known as the Red Book.
This article is based on a one day conference was held in November 2007 to disseminate the results of BRE Trust supported projects in the fields of fire science and fire safety engineering. For further information contact Jeremy Fraser-Mitchell or David Charters of BRE Global on 01923 664100, email [email protected]. For general information on BRE Global’s work go to www.breglobal.com
Fires in tunnels can lead to many deaths and injuries in a single incident or, at the very least, significant disruption and financial loss. Jeremy Fraser-Mitchell unveils recent research into sustainable and low-cost improvements in tunnel fire safety.
The safety of tunnels has been put in question in the past decade after a number of incidents involving significant loss to human life. In particular, the fires in the alpine tunnels at Mont Blanc in 1999 (39 fatalities), Tauern in 1999 (12 fatalities) and Gotthard in 2001 (11 fatalities) all highlighted shortcomings in the levels of safety inside existing road tunnels. Rail and metro tunnels have also come under scrutiny, with the fire in the metro in Taegu in South Korea (197 fatalities), and the metro fire in Baku, Azerbaijan (289 fatalities) being of particular note. Furthermore, the cost of repair and the economic penalty incurred while a major tunnel is closed can be very high.
As part of its 5th Framework Programme, the European Commission has supported a 4-year research project called UPTUN (UPgrading TUNnels) to develop tunnel safety upgrading methodologies. The project, which began in September 2002, involved some 40 European organisations including BRE Global. The Commission covered 50% of the project costs, with the participating organisations providing the remaining funding. BRE Global’s activities in UPTUN were supported by the BRE Trust.
The primary objective of the research was to develop innovative, sustainable and low cost measures to improve tunnel safety, by reducing the risk and consequences of fire in existing road and rail tunnels. Examples of issues that were addressed included improved means for safe evacuation, better transfer of information (e.g. signage), enhancements to traditional means of fire control and better protection (e.g. detection, mechanical ventilation and structural protection). Furthermore, attention was also given to newer and more innovative measures for fire and smoke control, e.g. water mist fire suppression and balloons that inflate to prevent the passage of smoke and heat.
UPTUN was divided into seven work packages:
1. Prevention, detection and monitoring
2. Fire development and mitigation measures
3. Human response
4. Fire effects, tunnel performance and structural response
5. Evaluation of safety levels and upgrading of existing tunnels
6. Full-scale experimental analysis
7. Promotion, dissemination, education and socio-economic impact
BRE Global contributed to work packages 2, 3, 5 and 7. This included the design fire specification, modelling smoke movement and alternative ventilation strategies, studying means of escape, performing egress calculations and developing training material. Findings from various work packages were consolidated into a holistic tunnel upgrading methodology, which allows the benefits of alternative measures to be evaluated in terms of both life safety and economics.
Design fire and CFD calculations
One of the tasks of working on fire development and mitigation measures (work package 2) was to provide quantified measures of the time-dependent development of hazardous conditions within the tunnel environment for selected design fire scenarios. This was achieved by reviewing previous studies and by undertaking new experiments. To complement this, numerical simulations were performed using CFD (computational fluid dynamics) to provide detailed predictions of thermal and other hazard levels within the tunnel space.
BRE Global’s CFD fire model, JASMINE, was used for simulating a set of parametric fire scenarios in longitudinally ventilated, two-lane road tunnels with rectangular and arch cross-sections. The fire source was indicative of an HGV fire, based on previous published information and from data from the Runehamar tunnel fire tests, conducted in 2003 in a disused road tunnel in Norway by Swedish and Norwegian transport bodies in collaboration with UPTUN. A peak heat release rate (HRR) of 200MW was achieved in one test, and even the least severe test achieved a peak HRR of 70 MW. Figure 2 shows the time-dependent thermal loads of the fire source terms used in the CFD simulations and Figure 3 shows the geometry used. The trailer space was represented by 20 ‘commodity packets’ from which the fuel was released.
With longitudinal ventilation, pollutants are forced along the tunnel in the direction of the air flow. In a fire emergency, the basic remit is to ‘push’ all the smoke in the downstream direction, and to maintain clear conditions upstream of the fire source, allowing means of escape in this direction as well as firefighting access. There is generally a minimum longitudinal air speed required to prevent fire gases from travelling in the upstream direction, referred to as the ‘critical velocity’. Where this condition is not met, the movement of smoke in the upstream direction is referred to as ‘back-layering’.
For a range of longitudinal ventilation levels, up to 5 ms-1 in the pre-fire condition, detailed predictions were made for the hazards associated with thermal and smoke exposure. Impulse (jet) fans were modelled by two (fixed value) momentum sources located approximately 100m upstream.
A comprehensive set of contour plots of gas temperature, smoke level (visibility distance) and carbon monoxide concentration was generated, using empirical correlations for soot particulate and carbon monoxide yields. Figures 4 and 5 illustrate two such contour plots for a 70MW peak fire inside the arch cross-section tunnel, where the vertical dimension has been stretched in the figures for ease of viewing. In one case, back-layering is present while in the other it is eliminated by ventilating above the critical velocity. Thermal and toxicity hazard exposure levels to people escaping have been calculated from the CFD data. Some of the findings from the CFD study were:
– Back-layering was eliminated at pre-fire longitudinal ventilation rates (critical velocity) of the order 3 ms-1 for all combinations of tunnel and fire size
– However, at ventilation rates above the ‘critical level’, the thermal hazard was significant in the first 100m downstream in all cases, and for the 70MW and 150MW fire scenarios the thermal hazard extended well beyond 100m in the downstream direction
– CO hazard was negligible for the 30MW fire, but was pronounced for the 150 MW fire due, primarily, to the total calorific value of the fire
-The CO hazard was mitigated by ventilating at a high rate, but this finding should be cautioned by the fact that this may then feed the fire with additional oxygen
Holistic tunnel upgrade model
An important element of UPTUN has been the development of a holistic tunnel fire safety evaluation and upgrading procedure for road and rail tunnels. This is being integrated with the work from another European project looking at pan-European guidelines for safety in tunnels, and in doing so extending the approach to cover non-fire accidents and incidents too. An important feature of the procedure is that it is fast to execute, enabling the benefits of various measures to be evaluated quickly.
The upgrade procedure involves smoke dispersion calculations using one- and two-dimensional models, egress and toxic dose calculations, risk assessment using event tree analysis and socio-economic modelling.
BRE Global contributed by enhancing its building fire risk assessment and egress model CRISP to include evacuation from road and rail tunnels (see Figure 6), and by coupling CRISP with tunnel-specific zone and network smoke movement models, including the multi-layer zone model FASIT. This enabled the combined influence of smoke movement and human decision making to be incorporated into the tunnel upgrading assessment. After examining reports of actual tunnel fires, new behaviour rules were added to CRISP, covering the movement of staff, passengers, drivers and the emergency services. Exit choice probability distributions were developed, and options for people moving in groups were also developed.
Figure 7 illustrates a typical graphical output from CRISP, where gas temperatures and smoke movement data has been integrated with the egress rules and the toxic dose model, enabling the response of people to be tracked in space and time. Note that at 15 minutes in this example there are two people (marked red to indicate a high level of accumulated thermal and toxic exposure) who are still relatively close to the source of fire and are in danger.
The upgrade procedure has been applied to tunnel example test cases to explore its capabilities. Here, various upgrade options were considered:
– A detection/early warning system, assumed to give a warning message three minutes after the fire has started
– Side exits and refuges, located 200m apart and where once a person has entered a side door, they are assumed to have reached a place of safety
– A suppression system, where the simplifying assumption was made that a 20MW fire would be extinguished and a 100MW fire contained once it had grown to 20MW
– Transverse ventilation smoke extraction
The benefits of the upgrading measures have been weighed against the cost, in terms of life safety and economics, of closing the tunnel for the required period and needing to re-route traffic. Depending on the measures adopted, a tunnel may need to be closed only at night, or it may need to be closed for some months or even for over a year.
The results have emphasised the importance of a holistic approach to the risk assessment. For some upgrade options, it is even possible for the risks associated with closing the tunnel to be higher than not upgrading the tunnel in the first place, e.g. due to diverting traffic onto other routes where the risks of an accident may be higher.
Finally, during tunnel upgrades, when specifying active and passive fire protection systems as an integral part of the design, it is recommended that third party certified products are used. Third party product certification schemes provide a means of identifying products which have demonstrated that they provide a requisite level of performance in fire, and also provide confidence that the products actually supplied are supplied to the same specification as that tested and assessed. In the UK, the Loss Prevention Certification Board (LPCB) operates various certification schemes for fire protection products. These are listed in the LPCB List of Approved Fire and Security Products and Services, known as the Red Book.
This article is based on a one day conference was held in November 2007 to disseminate the results of BRE Trust supported projects in the fields of fire science and fire safety engineering. For further information contact Jeremy Fraser-Mitchell or David Charters of BRE Global on 01923 664100, email [email protected]. For general information on BRE Global’s work go to www.breglobal.com
