Recent Advances in Engineering

Modelling the Evacuation of People from a Train on Fire in a Railway Tunnel P. KUČERA, I. BRADÁČOVÁ Faculty of Safety Engineering VŠB – Technical University of Ostrava Lumírova 13/630, Ostrava - Výškovice, 700 30 CZECH REPUBLIC [email protected], [email protected] Abstract: Fire safety of each structure consists in the creation of conditions for rapid and safe people evacuation, in the limitation of spread of fire and combustion products in the structure and the surroundings of the structure, and in the creation of conditions for rapid, effective and safe response actions of rescue units. In case of fire on a passenger train standing in a tunnel, a top priority task is to evacuate a relatively high number of passengers. The article presents the simulation of conditions in the tunnel during the fire using a program FDS. The program makes it possible to verify the conditions of evacuation in relation to tunnel geometry, to take into account the specifics of the movement of people, to assess the temperature and smoke hazards to people, and to carry out the check of visibility along escape routes. Key words: railway tunnel, fire modelling, evacuation, simulation, safe conditions infrastructure with reference to the specific features of the tunnel structure are then usually more severe than those of an open traffic road accident. Human lives and the environment are at risk, material damage to the means of transport and also to the goods being transported, damage to the tunnel – structure, installations and security equipment occur, and transport is interrupted, which affects the wider environment and other entities.

1 Introduction The developing trans-European road and rail networks with effects on regional transport routes require increasingly the construction of tunnels both in the rural and built-up areas. In a case of railway tunnels, the construction of tunnels is connected with the modernisation of railway corridors. For the design and construction of them many reasons exist. They are related to a necessity of locating a transport route in a geographically complicated area, passing below settlement units and watercourses; tunnels make it possible to shorten transport routes and speed up traffic, and provide protection to the environment especially by noise and gas emission reduction. To negatives, above all risk factors manifesting themselves during the operation of transport tunnels (transport of dangerous substances and dangerous goods, exceeding of design traffic intensity and/or traffic speed) should be added besides high time and financial requirements of the construction and equipment. They lead, under certain conditions, even to the occurrence of an incident in the tunnel (vehicle collision, fire, explosion, serious technical faults, and others). The consequences of an incident in a tunnel as an element of critical

ISBN: 978-1-61804-137-1

2 Fire Safety in Railway Tunnels The evacuation of people from a tunnel in case of incident (e.g. fire) is affected by many factors that depend, among other matters, on the behaviour and physical as well as mental abilities of human beings. The reactions of people in tunnel fires are influenced by panic, poor visibility due to spreading smoke and by insufficient knowledge of the environment. Evacuation is assessed for the case of a socalled “cold” incident not involving a fire or for the case of a “hot” incident involving directly a fire or a subsequent fire. By comparing the probability of a fire and the probability of a hazard to people in road and railway tunnels, a conclusion can be drawn that the probability of a fire in road tunnels is of

196

Recent Advances in Engineering

escape shifts in a sufficiently short time. In contrast to the standard procedure, some of simulation models (e.g. program FDS+Evac), in which evacuation is interconnected with a fire and smoke spread model, can be used, and thus the real evacuation of the given number of people from a specific railway tunnel can be proved. The evacuation time influences the requirement for ensuring the conditions for fire safety in the tunnel tube. It is required to let people leave the tunnel before spreading the smoke into escape routes; simultaneously the danger of a radiation heat hazard will be eliminated. The first task of the detailed assessment is the modelling of spread of fire and combustion products. Subsequently, the evacuation under conditions corresponding to the given tunnel will be assessed.

order higher than that in railway tunnels. Nevertheless, the level of hazard to people and fire consequences are higher in passenger train fires in railway tunnels. In the evaluation of statistical data, the fact that only the relatively small number of railway tunnels are operated in the Czech Republic and that the observation of tunnel incidents has been carried out for a relatively short time has to be considered. However, some incidents that happened in railway tunnels [1998, 2008] contradict the statement that thanks to the system protection of transport in railway tunnels and to the decreased influence of erroneous human actions during the passing of a train set through a tunnel, it is not necessary to take into account a failure in the transport system. The most tragic ICE train accident

happened in the vicinity of the town of Eschede in June 1998. An express train running from Munich to Hamburg was derailed here at the speed of 200 km/h; in the accident, 101 people died and eighty people were injured. That time, 14 coaches of the train set hit a bridge near the track. The cause was the broken frame of a train wheel. The third generation of ICE express trains can travel at a speed of up to 300 km/h. The trains run not only throughout Germany but also to Zurich, Vienna, Brussels, Amsterdam, Paris, etc. A German ICE high-speed train was derailed near the town of Fulda in the year 2008. When exiting the tunnel, the train collided with a herd of sheep. The train hit the herd at full speed, 10 of 12 coaches were derailed, 23 people were injured, 3 of them seriously. At the time of the accident, 170 passengers travelled on the Hamburg to Munich express train. The accident took place before exiting the 11 km Landrücken Tunnel, which is the longest railway tunnel in Germany.

4 Modelling the Spread of Fire and Combustion Products The analysis of the spread of fire and combustion products in a railway tunnel was carried out by means of a fire model. It is used for the verification of conditions of the evacuation of people from a train on fire (locomotive on fire) to the place of safety – in the presented example, only tunnel portals can be used for exiting to an open area. For the proper analysis of the spread of smoke the program FDS (Fire Dynamics Simulator program), utilizing the method CFD for modelling the dynamics of flow of hot gases and combustion products, was used. As an example a bi-directional railway tunnel having a length of 700 m, inner width of 12.2 m and clear height of 8 m was selected for modelling. The tunnel is double-tracked with bi-directional traffic. A simplified tunnel lining cross-section is illustrated in Fig. 1. The model considers a standing train set near the exit portal, in the point of the smallest height and the greatest distance from the other portal. Evacuating people can escape along both walkways. As the place of origin of the fire lasting 20 minutes the locomotive was chosen; owing to a moderate slope gradient in the longitudinal direction (about 5 ‰.), the movement of combustion products will be in case of fire supported by the stack effect. Other input parameters of the assumed fire scenario are the flow velocity in the tunnel of 1 m/s and the initial temperature of 10 °C.

3 Detailed Assessment of the Evacuation of People from Railway Tunnels The aim of the detailed assessment of evacuation is to identify evacuation critical points in any railway tunnel and to verify that proposed evacuation routes (escape walkways) will make it possible to leave the fire-affected tunnel tube through portals, cross-passages and

ISBN: 978-1-61804-137-1

197

Recent Advances in Engineering

Fig. 2: Train set geometry

Fig. 1: Simplified cross-section of bidirectional tunnel tube The passenger train set has the following technical data: train set

8 coaches + locomotive

length of train set:

225 m (1 coach ≈ 25 m)

external width of coaches:

2.8 m

height (pantograph locked down): material of body:

4.3 m steel welded body

In the model, the shape of the train set is simplified to one elongated rectangle divided into the locomotive and 8 passenger coaches divided from each other by partitions (Fig. 2). Along the sides of the passenger coaches, window openings are distributed evenly. The front of the train (locomotive) is placed 15 m from the exit portal. As already mentioned above, the locomotive fire is assumed and is defined by means of a heat release rate – determined for electric and diesel locomotives on the basis of experimentally determined values. For both the types of locomotives, these values are comparable and in the 20th minute of the fire correspond to the output of 12 MW. For the calculation of fire development thus the electric locomotive fire at the height of 1.5 m above level of the walkway with the duration of 20 minutes is considered.

ISBN: 978-1-61804-137-1

The computing results obtained by the program FDS can include a wide scale of observed values (temperature, thermal flow, flow velocity, visibility, and others). In this case, especially temperatures at the heights of 2 and 2.5 m above level of the escape walkway, the height level of a boundary between the upper hot and the lower cooler layer and smoke production in the tunnel in the course of fire development were recorded. In Figs. 3 and 4 are illustrated selected isotherms and smoke layer movement in the tunnel at the flow velocity inside the tunnel of 1 m/s using the visualization program Smokeview.

Fig. 3: Isotherms just behind the train set at tunnel internal temperatures of 40 °C (violet), 50 °C (grey) and 60 °C (green) in the 20th minute Note: These limit temperatures will not occur at heights less than 2.5 m on the escape walkway and will not endanger in any way escaping people.

198

Recent Advances in Engineering

• distance between the first coach and the entry portal 660 m • width of the escape walkway 1.5 m • length of the train set 225 m (length of one coach is about 25 m + locomotive) • number of people in the train set (8 passenger coaches) 8 × 80 persons = 640 persons. In all passenger coaches, a ratio between adults and children is identical, namely 85 % adults, 15 % children, i.e. 68 adults, 12 children. The distribution of people in individual passenger coaches is even. The space is divided into quarters; each of them has its own exit from the coach, which makes the directing of evacuation easier. The speed of movement of people is based on the proportional representation of adults and children in each coach and is variable.

Fig. 4: Smoke production in the tunnel near the locomotive in the 20th minute

5 Simulation of Railway Tunnel Evacuation A sample situation is as follows – the evacuation of people from a train set is to be simulated; the train set stands in the vicinity of the exit portal, from where the movement of combustion products is supported, owing to a gradient of the tunnel in the longitudinal direction, by the stack effect, and escaping people thus may be exposed directly to the combustion products. People from the train set will escape merely along escape walkways leading to the exit tunnel portals. For this variant, it is assumed that the escape of people along the walkway in the immediate vicinity of the locomotive on fire will not be possible. For modelling, an evacuation simulation module FDS+Evac for the program FDS can be used, by means of which not only the effects of fire action can be taken into account, but also the positions and physical conditions of individual evacuees can be set. In this module, every person is regarded as an independent unit and moves according to the person-specific equation of motion. The speed of movement of escaping people is influenced decisively, from the point of view of orientation and visibility, by the action of smoke.

5.2 Description of Evacuation from the Train Set People escape from the train set in both the directions, and for this reason, the evacuation is proposed in the program FDS+Evac according to a diagram illustrated in Fig. 5 as follows: • exit A – a smaller portion of the passengers from the right half of the train set (120 persons) are evacuated towards the closer (exit) portal along the farther escape walkway to avoid the thermal effects of the locomotive fire, • exit B – a larger portion of the passengers from the right half of the train set (200 persons) are evacuated towards the farther (entry) portal along the farther escape walkway, • exit C – all the passengers from the left half of the train set (320 persons) are evacuated merely towards the farther (entry) portal along the escape walkway that is closer to the train set. The escape towards the closer portal is impossible owing to the locomotive on fire from the very beginning.

5.1 Input Data for Simulation of People’s Evacuation • distance between the portals 700 m (= tunnel length) • distance between the exit portal and the front of the locomotive 15 m

ISBN: 978-1-61804-137-1

199

Recent Advances in Engineering

Human response to toxic exposure is determined in the FDS+Evac program by means of a concept of fractional effective dose, so-called FDS, which uses only the concentrations of CO, CO2 and O2 for the calculation of total FED. The CO2 concentration is considered here exclusively with regard to faster than normal breathing (hyperventilation). However, it is not assumed that the CO2 concentration would be so high that this gas would be unbreathable and would cause acidosis in the human body. The results of simulation of evacuation according to the time dependence of the number of people evacuated through specific exits were as follows: exit A - 5:04 minutes, exit B - 16:44 minutes and exit C - 18:32 minutes. The resultant dependence of the number of people evacuated through all the exits on time is shown in Graph 1.

Graph 1: Resultant time dependence of the number of people evacuated through all exits (A, B and C)

Fig. 5: Evacuation diagram in program FDS+Evac Before starting the evacuation of people, a 110 to 120 s time delay, which expresses the time required for informing the passengers by the train crew about the necessity of initiating the evacuation and the time required by people to react to the alarm, is considered. Into this delay the time required for going down the steps from the coach is partly included. The distribution of people in individual passenger coaches is even. The space is divided into quarters, each of which has its own exit from the coach, which makes directing the evacuation easier.

ISBN: 978-1-61804-137-1

Fig. 6 – Graphical representation of evacuation situation at the farther portal in the 10th minute Note: Blue little figures (agents) represent adults, white ones represent children.

200

Recent Advances in Engineering

the 1 m/s flow velocity in the tunnel to the value of about 25 m. The limit of visibility along the escape walkway thus was not exceeded. The acquired data on the evacuation thus was immediately confronted with the development of the fire. It was found that in the whole duration of the evacuation, none of the critical conditions of evacuation, under which the ability of people to escape from the structure would have been reduced (e.g. temperature effects on people, decreased visibility, action of toxic combustion products) had been exceeded.

6 Evaluation of the Simulation of Evacuation from a Selected Railway Tunnel The evacuation was carried out in two directions – along the proposed escape walkways through the portal exits designated, according to the diagram in Fig. 5, A, B and C. Owing to the uneven distribution of people and different distances from the emergency exits from the tunnel, the evacuation took place markedly differently in these two directions. The evacuation of a part of people towards the closer exit A (right part of the exit tunnel) lasted only about 5 minutes, whereas the evacuation towards the farther exits B and C in the farther entry portal took more time and was a factor decisive of the determination of total evacuation time – 18:32 minutes (Graph 1). The cause of the time delay was above all the distance that had to be overcome by people along the unprotected escape route. This proposed escape variant can be designated as the most unfavourable variant. On the basis of simulation using the program FDS for the selected fire scenario (12 MW output in the 20th minute at 1 m/s flow velocity), it is possible to state that the isotherm of 40 °C below the upper hot layer will not shift during the evacuation below the height of 2.5 m above the escape walkway even for the selected most unfavourable evacuation variant. The temperature of 40 °C was chosen as acceptable temperature for the short-term stay of people in the course of evacuation in the affected tunnel tube. Toxic combustion products were accumulated mainly in the upper hot layer; on the basis of data acquired by the module FDS+Evac, it can be stated that during the evacuation, any influence of toxicity of combustion products on escaping persons was not observed. With regard to a gradual increase in considered fire intensity and with regard to the tunnel dimensions, the observed parts of escape routes were filled with smoke only after the escape of all people. By means of the program FDS, visibility was verified as well. The acceptable limit of visibility is 10 m. In the course of simulation, visibility at the height of 2 m above level of the escape walkway changed minimally. The initial value of visibility of 30 m decreased merely at the end of simulation of the fire at

ISBN: 978-1-61804-137-1

7 Conclusions The simulation-based demonstration of escape route safety makes it possible to consider more factors – tunnel geometry, escape route widths, number of people and their physical and mental abilities, fire heat output and development, temperature effects on people, smoke generation, action of toxic combustion products, and conditions to maintain visibility. Various situations can be relatively quickly solved and the proposed solution can be optimized.

Acknowledgement Acknowledgements ledgements: This work was supported by the project of the Ministry of the Interior of the Czech Republic No. VG 20122014074 – “Specific Assessment of High Risk Conditions for Fire Safety by Fire Engineering Procedures“.

References: [1] MCGRATTAN, K., et al. Fire Dynamics Simulator (Version 5) User´s Guide. Washington: U.S. Government Printing Office, 2007. NIST, Special Publication 1019-5. [2] HOSTIKKA, S., et al. Development and validation of FDS+Evac for evacuation simulations: Project summary report. Finland: VTT Technical research Centre of Finland, 2007. ISBN 978-951-38-69823. [3] KORHONEN, T.; HOSTIKKA, S. Fire Dynamics Simulator with Evacuation: FDS+Evac: Technical Reference and User´s Guide. VTT Technical research Centre of Finland, 2009. ISBN 978-95138-7180-2.

201