Planning and Design for Fire and Smoke Incidents in Underground Passenger Rail Systems (2017)

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

7 IntroductIon This chapter summarizes findings from a literature review related to planning and design for fire and smoke incidents in underground passenger rail systems. A Transportation Research Infor­ mation Database search was conducted to aid the literature review, using key words such as “rail tunnel fires,” “rail fires,” “transit fires,” and “tunnel fires.” Reports are grouped under the following headings: • Rail Fire and Smoke Standards, • Railway and Rail Car Studies, • Passenger Behavior, • European Studies, • The METRO Project, and • Related Tunnel Studies. raIl FIre and Smoke StandardS NFPA usually publishes an updated version of a standard for the fixed guideway transit and passenger rail systems every 3 years (NFPA 2014). The current version of NFPA 130 was published in 2014. The purpose of NFPA 130 is to establish minimum requirements that will provide a reasonable degree of safety from fire and its related hazards in fixed guideway transit and passenger rail system environments. The scope of NFPA 130 is defined in Section 1.1.1 of the standard: This standard shall cover life safety from fire and fire protection requirements for fixed guideway transit and passenger rail systems, including, but not limited to, stations, trainways, emergency ventilation systems, vehicles, emergency procedures, communications, and control systems. The scope of the technical committee responsible for the preparation of NFPA 130 is: This Committee shall have primary responsibility for documents pertaining to fire safety requirements for under­ ground, surface, and elevated fixed guideway transit and passenger rail systems including stations, trainways, emergency ventilation systems, vehicles, emergency procedures, communications and control systems and for life safety from fire and fire protection in stations, trainways, and vehicles. Stations shall pertain to stations accommodating occupants of the fixed guideway transit and passenger rail systems and incidental occupancies in the stations. The goal is to provide an environment for occupants of fixed guideway and passenger rail system elements that is safe from fire and similar emergencies to a practical extent based on the following measures: (1) protection of occupants not intimate with the initial fire development; and (2) maxi­ mizing the survivability of occupants intimate with the initial fire development. Individual chapters of NFPA 130 address stations, train ways, emergency ventilation systems, vehicles, emergency procedures, communications, control systems, and wire and cable requirements. Enhancements anticipated for the 2017 version of NFPA 130 (Levitt 2016) include a requirement that closed stations be protected by a fire alarm system, new considerations on acoustics, sound level requirements revolving around emergency communications for first responders, protection of enclosed combustible insulation at stations, inclusion of additional electronic equipment in heat load analyses, chapter two lIterature revIew

8 and testing of adhesive and sealants with respect to fire heat release rates (HRRs). Additional needs to be addressed in the next edition include: • Standardized methodology for calculating fire HRR; • Incorporation of onboard fire suppression systems; • Issues associated with high­speed rail; • Additional methods for safe egress from deep stations; and • Acoustical measures associated with emergency communications by first responders. The International Code Council has developed a model building code (the International Building Code or IBC) and the International Fire Code (IFC), which have been adopted throughout most of the United States (International Code Council 2015). These codes establish minimum regulations for building systems as well as fire protection and life safety systems using prescriptive and performance­ related provisions. The codes establish provisions that adequately protect public health, safety, and welfare and do not unnecessarily increase construction costs; restrict use of new materials, products, or methods of construction; or give preferential treatment to particular classes of materials, products, or methods of construction. IBC/IFC and NFPA 130 are not always in agreement, and when they conflict, the international code usually supersedes the NFPA standard. APTA (2010) prepared a white paper that establishes guidelines for emergency ventilation of tunnels in the event of a fire. During emergency operation, ventilation is needed to influence the flow of smoke and combustion products to create a safer environment for tunnel users to escape and emergency services to intervene. The emergency ventilation system design requires the identification of the potential fire and smoke threat in terms of visibility, temperature, and toxicity effects. The design must consider characteristics of rail traffic (flow, type of vehicles, and combustible load); means to safeguard tunnel users and support responding emergency services; other safety facilities, such as egress routes, doors, passageways, and various equipment; and meteorological conditions between the portals that may create pressure differences and affect air flow velocities inside the tunnel. The white paper provides an overview of ventilation system types and the objectives for control, strategies for smoke management, and how they are affected by the design of the tunnel and its ventilation system. Finally, operational strategies for rail and road tunnels are described. The document is intended for use by transit agencies and emergency responders. raIlway and raIl car StudIeS The Volpe National Transportation Systems Center developed recommendations for improving practices for selecting vehicle interior materials during the procurement of new vehicles and the retrofit of existing rapid rail transit (RRT) and light rail transit (LRT) vehicles (Volpe National Transportation Systems Center 1984, 1998). The document provides recommended fire safety prac­ tices for testing the flammability and smoke emission characteristics of materials used in the con­ struction of RRT and LRT vehicles. The Volpe Center also analyzed transit system fire statistics to learn how often fire and smoke incidents occur on rail transit systems (Hathaway et al. 1992). Although fire accounts for only a small percentage of all rail incidents, the potential exists for loss of life and significant damage to property. The report identifies the countermeasures needed to prevent and reduce the severity of transit fires. The system safety approach was used to identify the necessary countermeasures. This method used (1) fault trees that graphically represented in a sequence of events how a fire develops, (2) an expert in transit safety who examined each sequence of events, and (3) another expert who examined the countermeasures for reducing and preventing transit fires. The system safety approach allowed an examination of the relationships between the various physical components and operating pro­ cedures of the entire transit system. In addition, potential problems relating to the construction and operating stages of the transit system could be identified. The report identifies five major areas of countermeasures: vehicle/equipment, procedures, human factors/training, environment, and infor­ mation management/data analysis. Subway tunnel fires often result in catastrophes with heavy casualties. Gao et al. (2012) described an opposite­double air curtain ventilation­assisted tunnel evacuation system (OTES) to help people

9 evacuate from a tunnel fire. It can be used to create a safe evacuation passageway that is free of smoke throughout the length of the tunnel. The performance of the OTES is compared with that of traditional ventilation systems: longitudinal ventilation and natural ventilation. The effect of the HRR, fire source location, and fire detection time are also discussed. The study also shows that compared with natural ventilation and longitudinal ventilation, the carbon monoxide (CO) concentration with OTES is significantly lower. Given the same HRR, the CO concentration values with OTES are only 0.58% to 2.41% of natural ventilation or 0.52% to 3.12% of longitudinal ventilation at the back end of the tunnel. OTES creates an obviously clear evacuation passageway. The effects of changes in fire source location and fire detection time on tunnel ventilation were tested and the effect of these changes on the formation of an evacuation passageway by OTES reported. Y. Z. Li et al. (2012) studied the provision of rescue stations in long railway tunnels. In the report, the stations already constructed or under construction worldwide are reviewed, and the basic pattern of smoke control during a rescue station fire is identified. Fifty­four model scale tests were carried out to investigate smoke control issues in rescue station fires. The effects of HRR, train obstruction, fire source location, and ventilation condition on smoke control in the cross passages of a rescue station were tested and analyzed. For the study, a critical smoke layer temperature near the fireproof door protecting the rescue station was investigated theoretically and experimentally, and a simple equation for the temperature obtained. A height of 2.2 m (7.2 ft) is proposed for the fireproof doors in cross passages of rescue stations. A number of serious tunnel accidents have put tunnel safety on the public agenda. Concerns have been directed toward the safety of road and rail tunnels. The choice of tunnel concept for double­ tracked rail lines has been given much attention. Anderson and Paaske (2003) discuss two alternative tunnel concepts from a safety perspective: single­bored tunnels and parallel twin­bored tunnels. The risk and safety arguments for various concepts are examined, and advantages and disadvantages for each of the concepts are discussed. An investigation of known tunnel and metro fires is used to assess how the choice of tunnel concept may have influenced the outcome of the accidents. In addi­ tion, typical results of computational fluid dynamics (CFD) calculations of tunnel fires and risk analy­ sis results are presented. Lupton (2001) considered the safety of long railway tunnels in light of current attitudes toward tunnel safety, with special regard to tunnel fires. Risk assessment should be considered an integral part of operating and constructing a railway. The primary safety protection function of a long tunnel is that of ventilation, correct use of which provides an area within the tunnel that will not be polluted and can sustain life. Another fundamental issue is evacuating passengers from the site of a fire, the length of time required to do so, and the capacity of the tunnel to accommodate passengers in walk­ ways and places of safety. Effective tunnel protection provides the chance of survival and gives the operator the ability to take control of the incident. In the report, reasons for not providing protection (probability, strategy, and cost) are not seen as sufficient. Social and economic issues are considered: acceptable safety is a variable dependent on the individual situation. Fire development inside a train car is a topic that has not been studied extensively owing to the complexity of the problem and the need for a real train car and the appropriate facilities to conduct such tests in a controlled environment. Lee et al. (2015) present detailed experimental data on fire development inside an intercity train car. The facility used for the tests provides an environment for conducting such large­scale fires, and one at which the HRR can be measured using oxygen con­ sumption calorimetry. Thermocouples and plate thermometers were installed inside the train car to provide insight into fire development. Cameras were placed inside the car and tunnel, providing live videos during the test. The peak HRR was 32 MW at 1,081 s after ignition, and the fire burned 83% of the initial fire load. Flame spread data and recordings of window breaking events are discussed and compared with the HRR data. A local flashover­type phenomenon, in which the fire involved all combustibles at the rear of the car, was found to occur. Lee et al. (2013) presented curve estimations of the HRR of an intercity rail car fire. Three estimation approaches were used, which were compared with a full­scale fire test of the car. Two of the estima­ tion approaches were based on the assumption of a specific burning rate of materials with the HRR

10 per unit area and burning area decision. The curve of the HRR of an actual rail car fire was measured by using the ignition scenario in the British Standards Institute report EN 45545­1 (British Standards Institute 2013). In the fire test, the surface temperature of every part of the interior was measured by using the burning area decision for summation method estimation. The third approach used combus­ tion and reaction heat to analyze microscopic­material pyrolysis. The pyrolysis model requires more sophisticated material property inputs to achieve the same degree of curve accuracy. The differences and similarities between the experimental fire curves and estimations were analyzed. Beginning in the mid­1990s, FRA funded the National Institute of Standards and Technology to develop a systematic approach to quantifying fire hazards in passenger trains that could form the basis for regulatory reform. An extensive literature review (Peacock et al. 1994) documented U.S. and European approaches to passenger train fire safety that rely primarily on individual small­scale test methods to evaluate material fire performance. This was followed by a three­phase research program. Phase I focused on the evaluation of passenger rail car interior furnishing materials using data from existing FRA­cited, small­scale test methods and from an alternative test method using the cone calorimeter (Peacock and Braun 1999). In Phase II, full­scale tests were conducted of selected interior material component assemblies using a larger scale furniture calorimeter; fire hazard analyses were then conducted for three types of intercity passenger rail cars using data from both types of tests (Peacock et al. 2002). Phase III compared the results of Phases I and II of the research program, with a series of full­scale fire tests conducted in an Amtrak coach rail car (Peacock et al. 2004). From the fire hazard analyses conducted, conditions in all three passenger rail car designs studied remain tenable long enough to allow safe passenger and crew egress for all but the most severe ignition sources. Comparison of times to untenable conditions for a range of fire sizes deter­ mined from the full­scale experimental measurements with those calculated by the Consolidated Model of Fire and Smoke Transport (CFAST) showed agreement that averaged approximately 13%. The range of ignition source strengths indicated that an ignition source size between 25 kW and approximately 200 kW is necessary to promote significant fire spread, which is consistent with con­ clusions from earlier research that the ignition source strength of passenger rail car materials is 2 to 10 times greater than that of typical office furnishings. Park et al. (2008) conducted a study to determine the optimum smoke­control ventilation mode in underground stations in South Korea. Numerical analysis was performed for the underground platform where the train with the fire was stopped. Three smoke­control modes were considered. Distributions of temperature, smoke, and visible range on the platform were analyzed for different smoke­control modes. The flow characteristics of smoke and heat on the platform of the underground station were also studied numerically. PaSSenger BehavIor Capote et al. (2012) presented a stochastic approach for modeling passenger performance during the evacuation process in passenger trains. The paper is divided into two parts. The first part describes the identification of variables and data collection. This process allows obtaining statistical samples of predefined random variables (personal responses, nonemergency actions, walking speeds, etc.). Statistical methods to determine the input and outputs for a stochastic analysis are proposed. In the second part, results from a stochastic model for passenger trains were compared with other evacu­ ation models and an announced evacuation drill. Results suggest that predicted evacuation times can be strongly dependent on the activities of individuals whose actions interrupt the continuous movement of other passengers within the aisle and the time spent by each passenger to negotiate the train steps. The advantages of using a stochastic approach for modeling passenger behaviors are discussed. Under the sponsorship of the FRA, the Volpe Center conducted a series of 12 commuter rail car passenger egress tests involving 86 passenger subjects, in cooperation with the Massachusetts Bay Transportation Authority, at North Station in Boston, Massachusetts, on August 25, 2005 (Federal Railroad Administration 2006). The egress experiment time data are intended to be used as an input

11 to a computer egress model to predict emergency evacuation times for various car configurations. Preliminary egress experiment results indicated consistent egress times by subjects for all trials with small learning effects and no fatigue effect. In addition, subject flow rates were less than those at previous train egress trials because no incentives were given and subjects were instructed not to push. No significant difference was observed between normal and emergency lighting conditions. The egress times for 84 passengers averaged 58 s using two side doors to a high platform, and 1 min 40 s using either a single side door to a high platform or an end door to the adjacent car. Fridolf (2010) reviewed and summarized literature related to fire evacuation in underground transportation systems as part of the METRO Project (described later in this chapter), and suggested areas for future research in the field. The literature can be roughly divided into three categories: (1) past accidents in underground transportation systems, (2) theories and models on human behavior in fire, and (3) empirical research related to evacuation in underground transportation systems. It was concluded that human behavior in fire is complex and sometimes can appear irrational in retrospect. However, instead of using “panic” to describe human behavior and the outcome of an accident, the adoption of a clear theoretical framework could aid the understanding of human behavior in under­ ground transportation systems. One of the major issues related to fire evacuation in underground transportation systems is that people often are reluctant to initiate an evacuation. This is explained by a number of factors: • That people tend to maintain their roles (e.g., as passengers); • The lack of fast, clear, and coherent information; • The ambiguity of the cues from the source of danger (e.g., a fire); and • The presence of others: that is, social influence. When an evacuation has been initiated, other factors affect the efficiency of the evacuation. Some of the problems identified are: • Door­opening mechanisms on trains; • The vertical distance between the train and the tunnel floor; • Tendency of people to evacuate through familiar exits; • Lack of lighting; and • Uneven surfaces inside the tunnels. It is suggested that occupants be provided with fast, clear, and coherent information. This infor­ mation could help people to initiate evacuation, find the way to safe locations without having to evacuate by means of familiar routes, and reduce the negative effects of social influence. However, providing users of underground transportation systems with this type of information demands an emergency organization in which staff members are educated and have clear areas of responsibility. When an evacuation has been initiated, technical installations are required if the evacuation is to proceed with efficiency. For instance, adding affordances (such as green flashing lights) that guide passengers to emergency exits inside a tunnel or station could help overcome the human desire to exit by a familiar route. Good lighting conditions and an escape path free of obstacles also are prerequisites for a smooth evacuation. The Fridolf report concludes with several suggestions for future research to improve safety for users of underground transportation systems. Nilsson et al. (2009) conducted an evacuation experiment in a road tunnel to investigate how motorists behave and emotionally respond when exposed to a fire emergency, how information and way­finding systems are perceived, and whether green flashing lights can influence exit choice. The participants believed that they were taking part in a study about driving behavior. Approximately 1 km inside the tunnel, participants encountered an accident: that is, cars and smoke. The fire alarm, which consists of a prerecorded alarm and information signs, was activated, as were green flashing lights at emergency exits. The results show that it was difficult to make out what was said in the prerecorded alarm. However, the use of an acoustic signal was a positive approach because it alerted motorists and made them look for additional information. The information signs were important in

12 the participants’ decision to leave their vehicles. Social influence was found to be essential in regard to the decision to leave the vehicle and the choice of exit. The results suggest that arousal level influ­ ences the amount of information noticed by motorists, which implies that technical installations, such as way­finding systems, should be tested under stressful conditions before they are relied upon in a real tunnel fire. euroPean StudIeS Several major tunnel fires in Europe in the 1990s prompted extensive research efforts, many of which specifically addressed passenger rail tunnels. These are summarized here. One difference between Europe and the United States is that few transportation tunnels in the United States have firefighting sprinklers in the tunnels. There are reports of two cities with sprinklers in passenger rail tunnels and less than a dozen U.S. road tunnels with sprinklers. In Europe the numbers are similar. How­ ever, the European tunnels often use a water mist system versus a deluge system. The efficacy of the two systems has not been directly compared. The U.S. deluge type system uses standard, readily available components, whereas the European mist systems use “proprietary” systems for which all of the components, design, and so forth are controlled by the mist system manufacturer. Cafaro et al. (2005) describe an automatically operated fire suppression system for protection against road and rail tunnel fires and fires in enclosed or underground spaces. The Indoor Protection System does not employ water mist: it controls and suppresses fire by using an additive called Tun­ nel Aqueous Film­Forming Foam (T­AFFF). It is a film­forming foam, based on fluorosurfactant, and suitable for use in Class B fires. In the presence of special wetting additives for enhancing its effectiveness, it is suitable for use in Class A fires. Cafaro et al. report the results of testing to demonstrate that the Indoor Protection System is able to limit the growth phase of the fire, extinguish and prevent reflashing pool fires, provide a restriction of the fire source, obtain an active protection for structures and equipment, mitigate and control a fire event generated by solid materials, and extinguish fires of flammable or combustible liquids. Haack et al. (1995) report results of a German initiative to identify the safety problems of tunnel fires. Full scale tunnel fire tests, dealing mostly with rail tunnel fires, were performed in Norway and the information gathered into a database. Calculations were made to evaluate the HRR associated with specific vehicle fires. A video camera network recorded the behavior of the smoke. Data analysis provided computer models of thermal fields. Various methods were used to control the database. The influence of longitudinal motion was evaluated. These tests are regarded as valuable because real vehicles were used as fire loads. Measurements related to these fire tests allow correct characterization of the physical phenomena. Following the spate of tunnel fires that occurred in Europe at the turn of the century, the safety of European tunnels was assessed and found to be generally poor. Tunnel safety can be improved only if the lessons of past incidents are learned properly. Carvel (2008) identifies some of the lessons learned from the Kaprun funicular tunnel fire of 2000, the King’s Cross underground station fire of 1987, the Baku subway fire of 1995, and the Channel Tunnel fire of 1996. Some recent advances in technology, specifically state­of­the­art ventilation and water suppression systems, are discussed. The 2007 fire in the Burnley tunnel in Australia is highlighted as an example of an incident in which technology prevented the initial fire from growing into a catastrophe. Fire safety in tunnels has become a major international issue after the catastrophic tunnel fires that have occurred worldwide. The fundamental principles that emerge from this research form the bedrock for decision making on how tunnels can be designed or upgraded and operated in an accept­ able way. Beard and Carvel (2005) bring together contributions from international experts in all areas of the field of tunnel fire investigation and firefighting. Their book spans the spectrum of current knowledge available in the field of tunnel fire safety, covering a diverse range of topics, including (1) fire safety management and human behavior, (2) fire prevention and protection, (3) tunnel ventila­ tion, (4) tunnel fire dynamics and fire investigation, (5) emergency services and emergency procedures, and (6) tunnel fire safety and the law.

13 Hohnecker (2000) describes the history of tunnel safety, German legislation, fire progression in tunnels, the development of safety plans, tunnel construction and operation (including emergency exits, lighting, routes and plans), and the transport of hazardous goods in tunnels. A list of tunnel incidents is provided. Van Weyenberge et al. (2015) describe the development of a risk assessment methodology to quantify the life safety risk for people present in a rail tunnel fire; the risk assessment methodology was produced in the context of the creation of a fire safety design. A bow­tie structure represents the risk assessment model, starting from major contributing factors leading to disastrous events. Using past accidents for the construction of the event tree part of the bow­tie, the most important factors are determined to be human behavior, fire growth, ventilation conditions, safety system (e.g., smoke and heat exhaust, detection, and voice communication), and population density. These factors are incorporated into the event tree using pathway factors. Frequencies are calculated for each branch outcome based on data from research projects, fault tree analysis, and engineering judgment. For the determination of the consequences, the method makes use of three integrated models: the smoke spread, the evacuation, and the consequence model. The models can take into account all types of geometry and materials, human behavior, and different susceptibilities of people to smoke. Together, they determine the possible number of fatalities, by means of a fractional incapacitation dose value, in the event of a fire in a rail tunnel. The final risk is presented by the expected number of fatalities, the individual risk, and the societal risk. The societal risk is demonstrated by means of a frequency/ number of casualties curve. Ferrazzini et al. (2011) focus on the main fire safety objectives and the concept for smoke control in the stations for the future rail link Cornavin­Eaux­Vives­Annemasse (CEVA), which will connect the Swiss to the French railway system in the urban area of Geneva. The total length of the project is approximately 16.5 km (14.7 km of which is in Switzerland) and includes a single­tube, double­track, 10­km long tunnel with four underground stations (Carouge­Bachet, Champel­Hôpital, Genève­Eaux­Vives, and Chêne­Bourg) and two bridges. The CEVA project is an example of an urban­suburban, heavy­rail underground system. For fire safety, the stations are equipped with ventilation systems. These are based on simple concepts to reduce the complexity and increase the robustness of the systems. During the design phase, one­ and three­dimensional numerical analyses of the ventilation concepts for each individual station were carried out to confirm the required performance during an emergency. Fulfillment of design objectives is shown by results from CFD simulations. The ventilation concepts of CEVA are compared with other contemporary projects with stations of similar size in Switzerland and elsewhere in Europe. In 2002, Markos and Shurland conducted a comparison study between U.S. and European approaches to passenger train fire safety. This led to FRA’s regulation for new and existing equipment on fire safety to now require passenger railroads to (1) conduct fire hazard analyses to assess and resolve hazards and (2) comply with flammability, smoke emission, and fire endurance criteria for certain interior materials (seats, walls, and curtains). the metro Project METRO is a Swedish research project related to infrastructure protection. The focus of the proj­ ect is on the protection of underground rail mass transport systems, such as tunnels and subway stations. Both fire and explosion hazards are studied, and such aspects as evacuation, rescue operations, and smoke control are important parts of the project. The project involves everything from human behavior during evacuation to fire and explosion tests and simulations, and is closely linked to practitioners, such as Stockholm Public Transport, to ensure that research results will be directly applied and tested in real underground rail mass transport systems, namely the Stockholm Metro. The METRO Project website (http://www.metroproject.se/index.html) includes specific subject reports. Summaries of a few of the relevant reports are included here. The website https://www. youtube.com/user/METROprojectSE has videos produced as part of the METRO Project.

14 The final report for the METRO Project compiles the results (Ingason et al. 2012). The project included six parts: • Design fires, • Evacuation, • Integrated fire control, • Smoke control, • Extraordinary strain on constructions, and • Fire and rescue operations. The most complicated and expensive part of the project was the performance of the large­scale fire and explosion tests in the Brunsberg tunnel. The maximum HRR measured from the metro carriage was 77 MW. The maximum ceiling gas temperature was 1,118°C. These values are high and should be put into a perspective of the situation and the type of carriages used. Figure 3 shows the interior of the train carriage. Figure 4 shows the fire in the Brunsberg tunnel. Figure 5 reproduces a figure from the METRO Project final report that shows a greater fire size during the Brunsberg tunnel test. Because one of the aims of the fire tests was to study the effects (condition in the tunnel, radiation, etc.) of a fully developed fire, it was decided to assist the fire development by igniting some additional pieces of luggage. However, when the first of these pieces of luggage (close to door 1) was ignited approximately 110 min after the original ignition, the firefighter igniting the luggage saw flames on the ceiling and had to exit the carriage without igniting the other prepared pieces of luggage. The fire had spontaneously spread to the driver’s cabin. The flashover of the driver’s compartment did not occur until the time 105 min: that is, after the temperature had started to increase in the passenger compartment but before the passenger compartment was fully involved in the fire. The maximum HRR in test 2 was calculated to be 76.7 MW (12.7 min after ignition), whereas the corresponding value for test 3 was 77.4 MW (117.9 min after ignition): that is, in both tests the maximum HRR was calculated to be approximately 77 MW. The egress study conducted as part of the METRO Project confirmed that one of the major issues related to fire evacuation in underground transportation systems is that people often are reluctant to FIGURE 3 Interior of converted C20 carriage (Source: Per Rohlén, the METRO Project).

15 initiate an evacuation. New data show that participants moved an average of 0.9 m/s in the smoke­filled environment (average visibility of 1.5–3.5 m). A way­finding installation at the emergency exit, which consisted of a loudspeaker, was found to perform particularly well in terms of attracting people to the door. Other installations consisting of light combinations were not as good, and some even repelled the participants because they were perceived as trains, gearshifts, or other track­related installations. Two smoke control systems were simulated for a single­exit metro station. The systems con­ sisted of a pressurizing supply air system and mechanical exhaust ventilation system with and with­ out platform screen doors. The results show that both the pressurizing supply air system and the mechanical exhaust air system provide effective smoke control for a one­exit metro station. The significance of the platform screen doors was shown to be important in relation to smoke control. Study recommendations include: • A design fire of 60 MW with a fast fire growth rate is proposed for use in the design of the ventilation system at metro stations. If the fire resistance of the interior lining material, seats, FIGURE 5 HRR in Brunsberg Tunnel test (Source: The METRO Project Final Report, Figure 9, p. 30). FIGURE 4 Fire test in the Brunsberg Tunnel. Fire took longer to reach flashover (Source: Per Rohlén, the METRO Project).

16 and windows is proven to be high, the designer can consider the use of a lower value, such as 20 MW. The definition of high quality resistance is left to regulators to define. No effects of fixed firefighting systems are assumed. • A design fire of 20 MW with a medium fire growth rate is proposed for use in the design of a tunnel system connected to a metro station. • In a performance­based design, a method presented by L. Li et al. (2012) is recommended. As an alternative to that method in a prescriptive fashion, a time temperature curve using the European thermally stimulated depolarization curve is proposed. • A walking speed of 0.9 m/s (for a visibility of 1.5–3.5 m) is proposed as an average value. Design values below 0.9 m/s are to be chosen for this visibility range. • A positive­pressure supply air system or a mechanical exhaust system is recommended as a smoke control solution for single­exit metro stations. • Platform screen doors are recommended in one­tube underground stations as a part of a technical fire safety solution. • High fire resistance of interior lining materials and windows in passenger and driver’s cabins is recommended. • New guidance for risk assessment before rescue operations in tunnels should be developed. • In metro tunnel systems, it is important that station personnel have routines to ease access for fire and rescue services, for example, through ticket gates. • Stations or the fire and rescue services should be equipped with trolleys or similar vehicles to transport equipment or injured persons in the event of fires or explosions inside the tunnel. • Robots or remotely operated vehicles for scouting and searching beyond the fire scene are useful and should be developed further. • There is need for the development and evaluation of different types of search patterns for tunnels. Normal infrared image search methods used in compartment fires are not applicable, so other search options need to be developed and evaluated. Several reports provided additional details on aspects of the METRO Project. Selected reports are summarized here. Claesson et al. (2012) present and discuss the results of six fire tests that were conducted in a mock­up of a subway carriage that is about one­third of a full wagon length. The tests were carried out in a laboratory environment with the use of a large­scale calorimeter. The aim of the tests was to investigate the initial fire growth in a corner scenario using different types of ignition sources that could lead to a flashover situation. The ignition sources used were a wooden crib placed on a corner seat or 1 L of petrol poured on the corner seat and the neighboring floor with a backpack nearby. The pieces of luggage and wooden cribs in the neighborhood of the ignition source were continuously increased to identify the limits for flashover in the test setup. The tests showed that the combustible boards on parts of the walls had a significant effect on the fire spread. In the cases where the initial fire did not exceed a range of 400 to 600 kW, no flashover was observed. If the initial fire grew to 700 to 900 kW, a flashover was observed. The maximum HRR during a short flashover period for this test setup was about 3.5 MW. The time to reach flashover was highly dependent on the ignition type, wooden cribs, or backpack and petrol. related tunnel StudIeS Two recent or ongoing NCHRP studies addressed ventilation issues and design fires in road tunnels. This section also includes a study focused on the behavior of highway users when confronted by a tunnel fire. Other studies on tunnel fires that may have application to rail tunnels are summarized at the end of this section. Currently, the design and operation of emergency smoke control varies from project to project in the absence of consistent and standardized practices. There is a need to identify the most effective operational practices for emergency ventilation smoke control in roadway tunnels. NCHRP Project 20­07/363 (Maevski 2016) developed recommended AASHTO guidelines for emergency ventilation

17 smoke control in roadway tunnels, to improve human evacuation and emergency responder safety. The guidelines consider: 1. Relevant conditions for application of various tunnel ventilation systems and configurations (e.g., full transverse, partial transverse, and longitudinal systems); 2. Fan utilization or placement based on tunnel geometrics and gradient; 3. Tunnel length and directional traffic flow (i.e., unidirectional or bi­directional flow); 4. The relationship between vehicle types and heat release and ventilation requirements; 5. Effects of ventilation on tunnel fires and fire size; 6. Fire detection and warning systems; 7. Actual fire smoke stratification duration and length of stratification in tunnel; 8. The interaction between firefighting operation and ventilation systems; 9. Applicability of tunnel vehicle fire suppression system and its applicable conditions; 10. Practicality of one­button or error­proof emergency ventilation (i.e., closed loop ventilation control); and 11. Applicable regulatory standards and guidelines (national and international). A 2011 NCHRP Synthesis by Maevski reviews the state of the practice of design fires in road tunnels, focusing on tunnel fire dynamics and the means of fire management for design guidance. Basic information is provided for tunnel operators, first responders, and tunnel agencies to better understand their tunnels and train their personnel. It includes statistical data for fire incidents in road tunnels since 1949 through the last decade, as well as statistical data documents for several tunnel fire safety projects in the United States and Europe. Extensive appendices offer more details about tunnel safety projects, fire tests, and national and international standards requirements, as well as descrip­ tions of past tunnel fires. The report does not address fires in underground passenger rail systems. Caroly et al. (2013) show that risk­management behaviors of highway users in tunnel fire situa­ tions depend on their knowledge of safety devices and their danger­handling behavior. The authors hypothesized that the unpredictability of the circumstances in which fires start, as well as drivers’ lack of knowledge about safety devices, are likely to have an impact on their behavior. The study is a detailed analysis of actual fires that have occurred in tunnels, with a close examination of users’ evacuation strategies and procedures. In the analysis of 11 tunnel fires, driver behaviors and the strategies they use to cope with a fire were studied. The tunnel users in the fires encountered dif­ ficulties in perceiving signs of danger and receiving warnings of the danger. The analysis showed that they engaged in a variety of evacuation behaviors and implemented few collective strategies to protect themselves. The problems were related to poor design or equipment, difficulty using safety devices or processing information, or a lack of emergency signals. Some recommendations are made regarding ways to modify existing prevention and warning devices to promote safer choices among the available options. The ratio of the cross­sectional area of the fire source to that of a tunnel is defined as tunnel block­ age ratio. These fire sources correspond to those that have considerable cross­sectional area, such as trains or heavy goods vehicles, which are common. Li et al. (2013) analyze the effect of tunnel blockage ratio on the maximum temperature under the ceiling in tunnel fires using experimental data from three previous studies. Results indicate that the maximum temperature decreases with the increase in the blockage ratio for small fires but does not vary with the blockage ratio for large fires. Previous models are modified based on the previous analysis by introducing a factor that accounts for the blockage effect. The modified models are more generally applicable. L. Li et al. (2012) studied the effect of tunnel blockage ratio on critical velocity in tunnel fires. A proposed empirical consideration that accounts for the blockage ratio effect has been verified. Based on the previous consideration, a new formula for predicting the critical velocity in blocked tunnels has been presented. To examine this formula, data from current numerical simulations and previous full­scale and reduced­scale tunnel fire experiments that have considerable tunnel blockage ratios are used. The formula shows good agreement with these data, which can be used appropriately to calculate the critical velocity for tunnels with blockage.

18 In the past decade more than 400 people worldwide have died as a result of fires in road, rail, and metro tunnels. Colella et al. (2011) apply a novel and fast modeling approach to simulate tunnel ventilation flows during fires. The complexity and high cost of full CFD models and the inaccuracies of simplistic zone or analytical models are avoided by efficiently combining mono­dimensional and CFD three­dimensional modeling techniques. A simple one­dimensional network approach is used to model tunnel regions in which the flow is fully developed (far field), and a detailed CFD representation is used where flow conditions require three­dimensional resolution (near field). This multiscale method previously has been applied to simulate tunnel ventilation systems, including jet fans, vertical shafts, and portals, and it is applied in the study to include the effect of fire. Both direct and indirect coupling strategies are investigated and compared for steady state conditions. The methodology has been applied to a modern tunnel with a diameter of 7 m and length of 1.2 km. Different fire scenarios are investigated with a variable number of operating jet fans. Comparison of cold flow cases with fire cases provides a quantification of the fire throttling effect, which is seen to be large and reduce the flow by more than 30%. The article places emphasis on the discussion of the different coupling procedures and the control of the numerical error. Compared with the full CFD solution, the maximum flow field error can be reduced to less than a few percent and can provide a reduction of two orders of magnitude in computational time. The much lower computational cost is of great engineering value, especially for the parametric and sensitivity studies required in the design or assessment of ventilation and fire safety systems. Summary The literature review has summarized a number of studies related to planning and design for fire and smoke incidents in underground passenger rail systems. The studies focusing on planning and design for fire and smoke incidents in rail tunnels are a small subset of the literature on tunnel fires. Many of the studies presented in this chapter are highly technical in nature. Others review past incidents or focus on passenger behavior. The literature review has informed the survey instrument used to gather input from transit agencies. An early question on the survey asks about sources used in planning and design for response to fire and smoke incidents in underground portions of the rail system. The conclusions in chapter six reflect the literature review as well as the survey and case examples. Additional research needs have been developed where unclear or conflicting information is identified. The next two chapters present the results of a survey of transit agencies. The results provide a snapshot of the state of the practice as it exists today.