- 284 - EFFECT OF AUTOMATIC FIXED FIRE FIGHTING SYSTEMS ON TUNNEL SAFETY Andreas Mühlberger, Prof. Dr., Universität Regensburg, Regensburg, D Anne Lehan, Dipl.-Wirt.-Ing., Bundesanstalt für Straßenwesen, Bergisch Gladbach, D Alexander Wierer,, Dipl.-HTL-Ing. Mag.(FH) ASFINAG, Innsbruck, A AndreasPlab,, MSc Psych., Universität Regensburg, Regensburg, D ABSTRACT Measures applied for the safety of tunnels are intended to ensure the protection objectives (1) self-rescue of the tunnel user, (2) support of the rescue by emergency personnel, and (3) protection of the structure. The prioritization of protection objectives may vary by country, but synergy effects between the protection objectives are always desirable. While in Austria decisions regarding the additional application of automatic Fixed Fire Fighting Systems (FFFS) may be taken with regard to the protection of the structure and as a consequence of the availability, the installed systems in Germany are primarily to support the rescue by the emergency personnel. The self-rescue of tunnel users must not thereby be affected. Regarding the use of FFFS in tunnels, different studies have been carried out by the Federal Highway Research Institute of Germany (BASt) concerning the effectiveness and efficiency of FFFS. It has been shown that the efficiency of using a FFFS is more given, the more protection objectives are incorporated in the analysis. In addition to the technical effectiveness, the effect of FFFS on the human behavior regarding the evacuation behavior of tunnel users in case of an incident is another essential aspect to the safety assessment. Therefore, several studies were conducted concerning the effect of installed FFFS (type compressed air foam and water mist) on the reaction and escape behavior of the tunnel users. These experiments with tunnel occupants were conducted in virtual reality (VR) and in real tunnels, and analyses were focused on the behavior of the participants during an evacuation situation. In all experiments, behavior with and without activated FFFS were compared. The results of a study carried out in 2014 in the Tunnel Jagdberg (Germany) regarding the influence of compressed air foam-based systems, as well as the results of a study carried out in 2015 analyzing the effect of water mist systems in the Bregenz City Tunnel (Austria) are presented. By comparing the results of studies in real tunnels and in VR important insights with regard to the interpretation of the results can be generated. The results show that the decision of using FFFS has to be made by taking into account the different protection objectives and the interaction with other safety infrastructure (e.g., evacuation instruction). A critical view is essential to achieve an optimized solution with consideration of safety-related and economically perspective. Keyword: tunnel, FFFS, Human Behaviour, Virtual Reality, Field Test
- 285-1. INTRODUCTION Compliance with a defined level of safety with the focus on the user safety is a top priority in the evaluation of tunnel safety. The measures for tunnel safety are used to ensure the safety objectives self-rescue of the tunnel users, support of the rescue by emergency personnel and protection of the structure. The prioritization of protection objectives may vary by country, but synergy effects between the protection objectives are always desirable. The focus, however, is basically to ensure efficient self-rescue. It is the first phase after an event and the behavior of tunnel users often decide on the consequences arising from the event. The serious tunnel accidents in the Alpine countries have shown this in particular. Figure 1: Classification of psychologic-behavioral investigation in the context of the evaluation of safety devices in road tunnels If conditions change or new factors have to be considered, which affect the level of safety of a tunnel or the self-rescue of users, these have to be evaluated holistically. In addition to riskanalytical and economic issues, the psychological evaluation of the tunnel user in its entirety, especially in relation to his behavior is an essential object of investigation (see Fig. 1). This applies equally to technical innovations as well as for changes in the composition of the userscollective (e.g. demographic changes). If the level of safety cannot be adhered by the new environment or a change in the user-collective, hereinafter measures have to be taken subsequently to ensure the highest protection objectives. The main objective of safety measures in tunnels is to ensure and to further enhance the selfrescue of the road users in case of events of an incident. To understand the behavior of road users in the event of an incident and to support self-rescue, it is necessary to consider various aspects of human information processing. A central component is the perception because of the various sensory channels the necessary information is included, which first include an understanding of the current situation as well as the perception of opportunities and goals of behavior. Major channels for communication of safety-related information are the visual and the acoustic channel. In addition to perception, the evaluation of the information and the
- 286 - decision making are of particular importance. Also motivational and emotional processes have be taken in consideration. Several models attempt to integrate the relevant processes in case of an incident [1,2]. 2. FIXED FIRE FIGHTING SYSTEMS Automatic fixed fire fighting systems (FFFS) in road tunnels are ment to ensure the containment of a vehicle fire in a road tunnel as early as possible to prevent the development of a major fire. They are used to limit the development of a fire and the fire spread and thus the effective fire power. Thereby the conditions for foreign rescue and also for firefighting should be supported as well as the protection of the structure from high temperature stress. Regarding the installation of FFFS the envisaged support of the protection objectives for each country varies, even if synergy effects between the protection targets are detected. For example, while in Austria decisions for the additional use of FFFS are conductive to the structural protection and thus to availability, in Germany the installed FFFS serve primarily to support the foreign rescue. In Germany, to date there are two tunnels with installed FFFS from the type foam, the Pörzbergtunnel and the Tunnel Jagdberg (both in Thuringia). In Austria, the City Tunnel in Bregenz, the Mona Lisa Tunnel in Linz and the Felbertauerntunnel feature a FFFS of the type high-pressure spray system. The equipment of the Arlberg tunnel and the Tunnel Liefering in Salzburg are in implementation. In Germany, there is a tendency with the updating of the guidelines Richtlinien für die Ausstattung und den Betrieb von Straßentunneln (RABT) [3], that when there are certain criteria (e.g. a tunnel length greater than 3000 meters (unidirectional traffic) or length greater than 1200 m in case of two-way traffic tunnels and a designed heat release rate major 100 MW), in individual cases and in addition to the safety systems the use of FFFS can be taken in consideration. Other criteria are also that a strengthening of the structure or operational safety systems does not bring sufficient benefits or they are uneconomical. There has to be done an evidence about the expediency of FFFS for the defined specific protection objectives in a holistic evaluation. In Austria's regulations, there are no defined guidelines for the use of FFFS, depending on certain criteria tunnel. However, in the case of the establishment of FFFS the RVS 09:02:51 - Fixed firefighting systems [4] gives detailed requirements which technical and operational criteria are to be observed. The protection objectives are similar to those of the RABT and are solely expanded to the reduction of contamination of equipment and construction of fire products, as well as to the reduction of environmental impact expanded. However, only highpressure spray equipment has been authorized in Austria. For a holistic consideration of the effects of a FFFS on the safety level, the interaction of the tunneling equipment to each other and their influence on user behavior must be validated. It should be noted that on the one hand an early activation of a FFFS brings the best chance of controlling the fire, on the other hand in case of a previous activation, tunnel users may be in the effective range of the system. For a review of the safety level, the criteria that make up the respective protection objectives have to be known. However, so far there were no systematic studies of the effect of an activated FFFS on the behavior of tunnel users, because a FFFS is currently activated only after the self-rescue phase. Therefore several controlled randomized studies were carried out on the initiative of the Federal Highway Research Institute (BASt).
- 287 - BASt devoted himself primarily the study of the influence of an activated FFFS to the selfrescue of tunnel users. Particular attention was directed to the investigation of the effects of the FFFS in terms of perception and the evacuation behavior of tunnel users. The important parameters here are the response- and escape velocity, which were considered in relation to the case "without" FFFS. Different investigations were conducted: one study about the influence of water mist systems in Virtual Reality as well as two field studies in real tunnels to investigate both types of FFFS (water mist and compressed air foam). It was expected that an activated FFFS possibly delaying the evacuation behavior or even prevented because the visibility can be seriously impaired and people avoid leaving their vehicle in the area of the activated FFFS, due to the water or foam. 3. VIRTUAL REALITY RESEARCH Besides examining tunnel user self-rescue behavior in real trials with limited possibilities and high availability requirements it is only possible during maintenance periods virtual reality (VR) offers an increasingly popular alternative. VR strongly evolved as a research method in recent years and offers multiple advantages like repeated presentation of standardized scenes and presentation of dangerous situations, which wouldn t be researchable in reality due to both ethical and logistical reasons [5]. Thus experimental approaches to various questions can be conducted with reasonable effort. In a first attempt within a research project the influence of an activated FFFS on tunnel users was examined ( Wirksamkeit automatischer Brandbekämpfungsanlagen in Straßentunneln (15.0563/2012/ERB)). Within this study 50 participants randomly divided in two groups were examined. One group was presented the activated FFFS (water mist), the other group was not. All participants took part in a virtual tunnel drive entering a tunnel as driver and encountering an accident involving a burning truck, three cars stuck behind it and smoke slowly spreading towards the participant. After stopping the vehicle an announcement instructed the participants to evacuate the tunnel. In the activated FFFS group, the FFFS was activated during the announcement playing for the first time. Virtual cars of participants came to hold directly in the FFFS s area of influence and the car was entirely covered in water mist (see image 3). Measurements were taken on whether participants left the vehicle or not and on latencies occurring until participants got out of the car. After they left the vehicle the experiment was paused. Upon resuming the simulation, participants had the opportunity of continuing their escape via gamepad until they reached an escape goal. Goal and duration of the escape were measured, too. This research was conducted in a 3D multisensory laboratory (Cave System) [6].
- 288 - Figure 2: The participants view immediately upon the end of the announcement in the FFFS condition (on the left) and in the no-fffs condition (on the right) [7] This VR experiment provides first valuable insights on how a FFFS affects experience and behavior of tunnel users. While the activated FFFS had distinct impact on the participant visual perception, hence a considerable reduction of vision inside and outside the car, the FFFS activation had only little effect on escape behavior. Both, participants with activation and those without activation of the FFFS left their vehicle within half a minute upon the beginning of the announcement and mostly chose the nearest by safety infrastructure as escape goals. Participants however differed in their escape routes to the emergency exit. Whereas participants in the FFFS condition kept rather close to the tunnel walls, those in the no-fffs condition evacuated right through to the middle of the tunnel. The announcement, which asked to leave the tunnel, was equally well understood in both conditions and was not drowned out by the sound of the FFFS. 4. RESEARCH IN REAL TUNNELS WITH FFFS To gain further insights considering humidity and cold, (real life) trials with test participants and both types of FFFS (high pressure foam and water mist) were conducted. Even if the basis for the decision on usage of FFFSs differ in both countries (Austria and Germany), the studies at hand provide clues on how to improve their application. Objective of this first of two research project (Einfluss einer aktivierten Brandbekämpfungsanlage (DLS) auf das Reaktions- und Fluchtverhalten der Verkehrsteilnehmer (FE 89.0299/2014)) was to determine the influence of an activated, water mist based FFFS on behavior and experience of tunnel users in respect to haptic stimuli like humidity and cold (which can t be simulated in VR) in a field study and to validate the findings of the VR experiment. This study was conducted in the Jagdberg tunnel (A4 near Jena), which was mostly completed when the study was done but not yet opened to the public. In a randomized, controlled design in the Tunnel Jagdberg the escape behavior of the participants inside and outside the car was measured. Participants drove a car into the tunnel and were confronted with a simulated accident with smoke proliferation. After stopping the vehicle in front of the accident an announcement asked them to evacuate. Additionally the FFFS was activated for one half of the participants (see figure 3, on the left). A special focus was put on analysis of the participants reaction and escape behavior and also on the choice of the escape goals [7].
- 289 - Figure 3: Car immediately after activation of the FFFS, Foam (on the left) and Water mist (on the right) In accordance to the VR experiment s findings participants in the FFFS condition report restricted vision due to the foam (especially when sitting in the car without activating the windshield wiper). Compared to the VR experiment the loudspeaker announcement in the tunnel was less well comprehensible. Participants hardly feared any negative effects of the foam and reported no irritations on mucous membranes. They felt slightly influenced in their escape behavior, because they assumed the foam to be slippery and thus had watched their step. Behavior analysis shows most participants to have left their vehicle even with the FFFS activated. Both groups did not show any meaningful differences respective to escape goal and required time to exit the car. In both groups most participants evacuated to the nearest emergency exit. This finding suggest that most participants complied with the loudspeaker announcement s request to leave the tunnel. After the study of the foam based FFFS a second research project about the analysis of tunnel users reaction and escape behavior upon activation of a water mist based FFFS in realistic conditions (FE 15.0607/2014) was conducted. The objective was similar to the previous study changing the FFFS type to water mist. With support by the ASFINAG, this study took place in the Citytunnel in Bregenz, which was closed to the public during the research period (nighttime). Here too the escape behavior of the participants inside and outside the vehicle was analyzed in a randomized, controlled study. In the same way as in the study conducted in the Tunnel Jagdberg, participants drove a car into the tunnel and encountered a simulated accident with smoke proliferation. After stopping their car next to the accident a loudspeaker announcement asked them to evacuate. One half of the participants additionally experienced activation of the FFFS (see image 4, on the right). Again, a special focus was put on analysis of reaction and escape behavior of the participants as well as on the choice of escape goals. Within this study the amount of participants was further extended to 64 to conduct additional couple tests to determine tendencies concerning the interaction of passengers and possible delays or improvements regarding the escape behavior. In accordance to the previous findings participants in the FFFS condition report the water mist based FFFS to considerable restrict their vision inside and outside the vehicle. As seen in the Tunnel Jagdberg study, loudspeaker announcements were less well comprehensible due to the FFFS activation. A particularly meaningful finding of this study concerned about 40% of the participants in the FFFS condition not disembarking their vehicle. As a result the number of participants reaching the emergency exit was considerably smaller whereas the number of other escape goal choices did not seem to be influenced (see figure 4, on the right). The couples, which entered the tunnel together, showed a tendency towards faster and more consequent reactions in comparison to individual drivers. This becomes especially apparent considering none of the couples stayed in the vehicle.
- 290 - Figure 4: Behavior of participants and goals of escape behavior in %. Numbers over the boxes indicate absolute numbers 5. SUMMARIZING EVALUATION Even though these three FFFS studies only covered one possible scenario and although the transfer of the findings at hand on real fire scenarios has to be handled very carefully, it seems that activating a FFFS in a tunnel might indeed limit vision considerably, but has no decisive negative impact on tunnel users, as long as relevant parts of tunnel infrastructure are being adjusted to the FFFS activation appropriately. In doing so, a well comprehensible announcement, when indicated with reference to the FFFS activation, is very important, so tunnel users finding themselves in the FFFS s immediate area of influence can understand the announcement and disembark their vehicle. Interestingly in all studies at hand a considerable percentage of participants, which also got prompted to evacuate via loudspeaker announcements (real and VR), initially located an emergency call system. In the case of fire and strong smoke production, in which appropriate authorities are already informed, this is no constructive behavior. Having heard the announcement, tunnel users should be able to deduct that authorities are already involved. This raises the question of whether measures should be taken to change this behavior. Moreover, the VR and field study comparison suggests that virtual reality research is an appropriate means for researching behavior of tunnel users in case of an incident and gaining valuable insights on infrastructure requirements for tunnels with specific safety systems. Furthermore these findings show the importance of researching user behavior from psychological perspective to extensively evaluate safety standards of tunnels. Psychological user studies should be established as a fundamental pillar for evaluation of tunnel safety, especially with regard to a synchronized overall safety concept.
- 291 - REFERENCES [1] Kinateder, M., Ronchi, E., Nilsson, D., Kobes, M., Müller, M., Pauli, P., & Mühlberger, A. Virtual, Reality for fire evacuation research. Proceedings of the 2014 Computer Science and Information Systems (FedCSIS), 2. 313 321, 2014. [2] Kuligowski, E., The process of human behavior in fires: US Department of Commerce, National Institute of Standards and Technology, 2009. [3] Forschungsgesellschaft für Straßen- und Verkehrswesen: Richtlinien für die Ausstattung und den Betrieb von Straßentunneln (RABT), Entwurf 2015, unveröffentlicht [4] Österreichische Forschungsgesellschaft Straße-Schiene-Verkehr: Richtlinien und Vorschriften für das Straßenwesen 09.02.51 Ortsfeste Brandbekämpfungsanlagen, Ausgabe 1. Juli 2014 [5] Kinateder, M., Ronchi, E., Nilsson, D., Kobes, M., Müller, M., Pauli, P., & Mühlberger, A. Virtual, Reality for fire evacuation research. Proceedings of the 2014 Computer Science and Information Systems (FedCSIS), 2. 313 321, 2014. [6] Kinateder, M., Ronchi, E., Gromer, D., Müller, M., Jost, M., Nehfischer, M., Mühlberger, A., & Pauli, P. (2014). Social influence on route choice in a virtual reality tunnel fire. Transportation Research Part F-Traffic Psychology and Behaviour, 26, 116-125. [7] Lehan, A., Mühlberger, A., Gast, P., Pauli, P.: Selbstrettung der Verkehrsteilnehmer: Menschliches Verhalten in Wechselwirkung mit der Tunnelausstattung [Self-Evacuation of Road Users: Human Behaviour interacting with the Tunnel Furnishings], Tagungsband zur STUVA-Tagung 2015, Stuttgart, 321-325.