Evacuation of a Metro Train in an Underground Rail Transportation System: Flow Rate Capacity of Train Exits, Tunnel Walking Speeds and Exit Choice

Abstract

Including in total 135 participants in the ages 19–69 years (recruited from the general public), an unannounced full-scale field evacuation experiment was performed in the Stockholm underground metro system on the night between October 17 and 18, 2014. The purpose was to collect data on the flow rate of people in train exits during the evacuation of a train in a tunnel and on the walking speed of people when moving long distances on an uneven surface in a tunnel, and to study exit choice and behaviour during an evacuation. Consequently, the experiment involved the evacuation of a rail car (a Bombardier C20 train) in a tunnel as well as the subsequent evacuation of the tunnel itself; the latter meant that the participants either could evacuate to the closest station (~400 m) or to an available emergency exit (~200 m). Among other things, the experiment demonstrated that the averaged flow rates of people in the train exits varied between 0.19 p/s and 0.22 p/s (0.14–0.16 p/m s when considering the train exit width of 1.4 m) and that the averaged walking speeds in the tunnel varied between 1.1 m/s and 1.2 m/s (no smoke present). Furthermore, all 135 participants found and used the available emergency exit, which had been equipped with a technical system consisting of a loudspeaker that broadcasted a combined alarm signal and a pre-recorded voice message.

Appendices

Appendix 1: Tunnel Environment

The tunnel design during the experiment is illustrated in Figures 20, 21, 22 and 23, and was identical to the environment during normal operation. This means that, in essence, the floor consisted of macadam (size 32–64 mm) and that the rail track was installed on wooden sleepers. The electric rail, which in the Stockholm metro is used to provide the trains with electricity, was situated on the tunnel floor (protected by a wooden plank). On most walls, cable ladders were available with cables providing, for example, the light installations with electricity. The tunnel inclined toward the Rinkeby underground station, corresponding to about 3% inclination (i.e., a 3 m rise per 100 m travelled from the rear end of the train).

Figure 20

Illustration of the tunnel environment at a location close to the emergency exit, which is located some 50 m ahead to the right (direction towards Rinkeby). Note that not all the lights in this picture were active during the actual evacuation experiment

Figure 21

A general illustration of the environment, representative for the setting in the tunnels where the evacuation experiment was performed. Note that not all the lights in this picture were active during the actual evacuation experiment

Figure 22

Illustration of the tunnel leading up to Rinkeby station (photographer standing at RInkeby station, with station behind). Note that not all the lights in this picture were active during the actual evacuation experiment

Figure 23

Illustration of the tunnel environment at a location near the rear end of the rail car that was evacuated (direction towards Rinkeby). Note that not all the lights in this picture were active during the actual evacuation experiment

Emergency signs, such as illustrated in Figure 24, were installed on both sides of the tunnel in order to provide evacuees with information about distances to the closest exit. These signs were available continuously throughout the tunnel, with a reoccurrence of approximately 8 m. A lamp was installed above each sign, which was turned on during the experiment. In addition to these lamps, fluorescent lights were also installed on both tunnel sides, most often a couple of meters above the tunnel floor, with a reoccurrence of approximately 4–8 m. The fluorescent lights were primarily used during maintenance work in the tunnel, but could as well be activated by a traffic controller during an evacuation (they were, however, not standardised emergency lights, which meant that the power supply were less resilient than for the emergency sign lamps). During the evacuation experiment, every fourth of the fluorescent lights were activated. This provided the evacuees with the illumination levels in the tunnel ranging from as low as 2.8 lux at the worst position to 32 lux at the best. Thus, not that there is a difference between what is seen in Figures 20, 21, 22 and 23, in which all lights are lit, and what was the case during the experiment.

Figure 24

Emergency exit signs were available on both sides of the tunnel used in the experiment, and above each sign a small lamp was attached in order to illuminate the information on the sign. The signs were installed every 8 m

Appendix 2: Procedure

The sequence of events during the night of the evacuation experiment is described in Table 11. It should be noted that any of the preparations prior to the experiment, performed by the researchers and other assistants, have been excluded from this presentation.

Table 11 The Evacuation Experiment Started at 24:00 when the Participants Started Arriving to the Vreten Underground Station on the Blue Line

Appendix 3: Recruitment of Participants

The recruitment of participants for the evacuation experiment can be separated into three phases. In the first phase, a one-page recruitment ad was published online on a number of web pages, in social medias (e.g., Facebook) and on an online portal used by researchers who want to get in contact with potential human test participants in the Stockholm region. The recruitment ad was published approximately 6 weeks before the date of the experiment, and it included brief information about the experiment. Among other things, the purpose of the experiment was included. However, the communicated purpose of the experiment deviated somewhat from the true purpose; in the recruitment ad, it was stated that the main purpose of the experiment was to study what type of information and which technical aids that passengers of the underground metro system in Stockholm would be helped by during an accident.

The first recruitment ad also informed that the experiment would be performed in closed areas of the underground metro system in Stockholm during night time, that the participants at some point may have to walk shorter distances on track level, that they would have the possibility to at any time abort the experiment and that a shorter fire safety training would be given to all participants after the experiment. Information was also given about the compensation for participation; all participants that took part in the experiment were compensated with 600 SEK for their participation (approximately 60€). In the end of the recruitment ad, it was stated that interested participants should contact one of the researchers in charge either by email or telephone. It was highlighted that this did not guarantee a position in the experiment.

Approximately 2 weeks after the recruitment ad first had been published it was removed, and from this point it was no longer possible to submit an interest to take part in the experiment. At this time, a total of 785 people had stated that they were interested in taking part in the experiment. However, the research budget only allowed for 150 people to take part in the experiment. All 785 people were contacted by email and instructed to fill out an online questionnaire within 2 weeks, which consisted of two parts; (1) one part on general information about the test participants (e.g., birth date, gender, name and contact information, current occupation, and possible disabilities), and (2) one part including a so called Hospital Anxiety and Depression (HAD) test [21]. The motivation of the HAD test was also described in some detail to the potential participants; it was mentioned that it was included and obligatory in order to minimize the risk of negative side effects related to psychological health. It was also explicitly stated that by answering the questionnaire, the potential participant was still not guaranteed a place in the experiment and furthermore that the information left in the questionnaire would not be distributed to third parties. Still in the second phase of the recruitment process, no further information about the evacuation experiment (e.g., information about the location and date) was presented. The reason for this was to minimize the risk of having uninvited participants showing up at the test location during the experiment date.

In total, 599 of the initial 785 potential participants filled out and submitted the questionnaire. Out of these, 150 people were selected based on the criteria to ensure a wide age distribution among the participants and to have men and women equally represented in the experiment (e.g., 50/50). In order to be selected for the experiment, the person were also required to have passed the HAD test (only persons that received a score of less than eight for both anxiety and depression were considered. The 150 participants were contacted by mail approximately 2 weeks prior to the experiment, and were informed that they had been selected to take part in the experiment, would they still be interested after having read the attached documents; (1) a document termed test participant information, (2) and a document including the rules of conduct in the underground metro system. They were finally asked to confirm their participation, and also asked to state if they were no longer interested in taking part. In that case, another individual who had completed the questionnaire with the same demographic characteristics were contacted.

The document termed test participant information, included in the final mail to the selected participants, was a requirement according to the ethical review board. It included information about the experiment and the experiment location, but also the background and purpose of the study, the procedures during the date of the experiment, the risks related to participation, benefits for taking part, treatment of data, where the results about the experiment were going to be published, and compensation for participation as well as insurance coverage during the experiment.

Appendix 4: Ethical Considerations

All research involving procedures with the aim of influencing participants physically or psychologically, or that involves an apparent risk of physical or psychological injury, must be subject to a review by a regional ethics board according to the Swedish ethics act [22]. Because these criteria were deemed applicable, the present study was submitted for ethical review by the regional ethics board in Lund and subsequently approved [23]. The important ethical issues discussed below were identified and addressed within the application.

4.1 Preparation and Precautions

A number of precautions were taken to avoid both psychological and physical injury during the experiment. The risk of psychological injury was minimised by preventing individuals who received a high score for both anxiety and depression according to the HAD questionnaire [21] from taking part. In addition, preparations were in place to take care of any participants who displayed signs of acute anxiety during the experiment.

Clear and specific safety instructions were given to participants before the experiment in order to minimise the risk of physical injury. Participants were also informed that they could terminate their participation at any time, e.g., if they were worried that they might get injured. They were also told that they would receive their reimbursement even if they decided to end their participation early.

One of the main risks identified in the ethical analysis was that participants might decide to evacuate the train before the electricity of the third rail had been turned off and the third rail had been grounded. This would have meant that several participants would have been at significant risk of being electrocuted. In order to avoid electrocution, participants were specifically informed not to open the train doors until told to do so by the train driver. In addition, observers were present on board the train and their main job was to stop any spontaneous evacuation. Some observers were wearing yellow vests, but others were concealed in the crowd, i.e., played the role of participant.

Another major risk was the risk of losing participants in the metro system during the experiment. Observers were therefore placed in the metro system at all possible branching points. The task of these observers was to hinder any participant who was heading in the wrong direction. Checks were also made to ensure that no one was left behind. More specifically, observers searched the tunnel and all participants were registered both before and after the experiment.

The experiments were filmed with video cameras at a number of locations both inside the train and in the tunnel. The films have only been analysed by researcher involved in the project, and they are stored at a safe location when they are not used. No sensitive personal information, such as name or identity, is connected to the videos, which makes it extremely difficult to link any sensitive information about the participants to the films.

4.2 Debriefing and Follow-up

Debriefing was performed when the participants had re-boarded the train after termination of the experiment. In the debriefing, detailed information about the experiment was given to the participants through the PA-system of the train. In addition, a short fire education was provided. This education was focused on fire and evacuation safety in metro systems, and the participants were given instruction of how to behave in case of a fire in the metro. The aim of the education was to guarantee that participants would directly benefit by taking part in the experiment by making them better prepared for potential future fire incidents.

Two months after the experiment, the participants were contacted by telephone in order to investigate if they had suffered any negative effects as a consequence of the experiments, e.g., anxiety, nightmares, etc. None of the contacted participants expressed any negative effects. However, it was not possible to get in contact with all participants; 128 out of the 135 participants were contacted.