Evacuation and Intervention Management

SafeT Work package 3 Task 3.1, 3.2 and 3.3

D3 integrated report

Final report

Evacuation and intervention management

Version: December 2005 Author: Nils Rosmuller (Nibra) Gernot Beer (SiTu) Roberto Gomez (EBSCC)

Table of contents

Outline of the integrated report...... 5

Overview of FIT products relevant for the development of fire guidelines for tunnels (SiTu)

1. Introduction to FIT network...... 8 1.1 Short description of FIT network ...... 8 1.2 Relevant data bases and documents ...... 9

2. Summary of data base information...... 10 2.1 Database 1: RTD on fire safety in tunnels ...... 10 2.2 Database 4: Data on safety equipment in tunnels...... 10 2.3 Database 5: Assessment reports on fire accidents in tunnels ...... 10

3. Summary of FIT reports relevant to SafeT ...... 11 3.1 Report: General approach to tunnel fire safety ...... 11 3.2 Report: Best practice for fire response management.12

4. Discussion...... 13

5. Reference ...... 14

6. Appendix I: RTD projects relevant to SafeT ...... 15

7. Appendix II: List of safety equipment...... 16

8. Appendix III: Fire accidents in Tunnels...... 17

9. Appendix IV:Sample of accident report sheet ...... 18

Computer supported training tools for rescue personnel in tunnel accidents (EBSCC)

10. Abstract...... 21

11. Objectives ...... 22

12. Introduction ...... 23

13. Data collection ...... 24

14. Data analysis and practical examples ...... 25 5.1 Virtualfires ...... 25 5.2 Virtualtraining- Advanced disaster Management systems (ADMS) ...... 32 5.3 GAMMA-EC...... 36 5.4 Fire simulators comparison...... 44

15. Proposal for EU guidelines...... 48

16. Limitations...... 49

17. Recommendations...... 50

18. References...... 51

Cross border incident management (Nibra)

19. Abstract...... 54

20. Objectives ...... 56 1.3 Problem statement...... 56 1.4 Objective ...... 56

21. Introduction ...... 57 1.5 Background ...... 57 1.6 Outline...... 60

22. Data collection ...... 61

23. Data analysis...... 62 1.7 Literature...... 62 1.8 Case studies: Eurotunnel and Mont Blanc Tunnel 63 1.9 Procedures in some TERN tunnels...... 64

24. Practical examples ...... 69

25. Proposals for EU guidelines...... 70

26. Limitations...... 72

27. Recommendations...... 73

28. References...... 74

Appendix 1: Cross-border tunnels in Europe...... 76

Appendix 2: Cross-border incident management aspects...... 81

Outline of the integrated report

This report is the integration of the three WP 3 task reports: 3.1 Overview of FIT products relevant for the development of fire guidelines for tunnels (SiTu) 3.2 Computer supported training tools for rescue personnel in tunnel accidents 3.3 Cross border incident management

The integration was done by the workpackage 3 leader, Nils Rosmuller (Nibra). To this end, he copied the submitted reports by each of the three partners in workpackage 3. There have no changes made to the submitted contents of the documents.

WP task 3.1 provides the relevant links for all SafeT partners to the results of EU-research programme Fire in Tunnels (FIT). These links are provided by SiTu. SiTu is a partner who is involved in both FIT and SafeT. All the SafeT partners could use these FIT results for their activities in SafeT. The FIT results have been assessed and the report indicates why which FIT results are relevant for which SafeT purposes. As will be clear from this description, this workpackage does not results in proposals for EU guideliens. The appendices are the links to the FIT database on www and therefore not incoporated in this report.

WP task 3.2 assesses the possibilities of computer supported training tools for tunnel intervention management. EBSCC conducted the assessment and was involved in an earlier EU research project that involved a computer supported training tool for rescue personnel. Based upon the assessment of three existing computer supported training tools for rescue personnel, the critical succes items of these training tools have been revealed. Seven guidelines are proposed for member states in case they intend to deal with computer supported training of rescue personnel with regard to tunnel intervention management.

WP task 3.3 assesses the typical intervention issues in case of cross border or transboundary tunnels. Nibra conducted the assessment and was involved in various tunnel safety studies in which emergency response aspects were of primary importance. Based upon tunnel accidents, literature study and the study of existing (cross border) tunnel safety procedures, intervention barrieres because of the cross border aspects were derived. Fourteen guidelines dealing with intervention management are proposed for member states that inhabite cross border tunnels.

Task 3.1: Overview of FIT products relevant for the development of fire guidelines for tunnels (SiTu)

Executive Summary

The report gives an overview of the data bases and reports produced by the thematic network Fire in Tunnels (FIT) that are relevant to SafeT. This should ensure that no duplication of work is made and that all information developed in FIT is used as input to the documentation prepared in SafeT.

Contents WP 3.1

1. Introduction ...... 8 1.1. Short description of FIT network ...... 8 1.2. Relevant data bases and documents ...... 9

2. Summary of data base information...... 10 2.1. Database 1: RTD on fire safety in tunnels...... 10 2.2. Database 4: Data on safety equipment in tunnels10 2.3. Database 5: Assessment reports on fire accidents in tunnels ...... 10

3. Summary of FIT reports relevant to SafeT ...... 11 3.1. General approach to tunnel fire safety...... 11 3.2. Best practice for fire response management...... 12

4. Discussion...... 13

5. Reference ...... 14

6. Appendix I: RTD projects relevant to SafeT ...... 15

7. Appendix II: List of safety equipment...... 16

8. Appendix III: Fire accidents in Tunnels...... 17

9. Appendix IV: Sample of accident report sheet ...... 18

10. Appendix V: FIT reports

1. Introduction to FIT network

The objective of report 3.1 is to present a collection of data form the FIT web page relevant for the preparation of guidelines. The relationship of the document to the SafeT objectives is to ensure that work is not duplicated, that ideas of FIT are considered in drafting the guidelines and that network members are informed about FIT products without having to get into the FIT webpage.

1.1 Short description of FIT network The European Thematic Network on Fire in Tunnels was introduced in the Growth 2000 call. FIT was created out of the idea of co-ordination and ex-ante clustering of different RTD1 initiatives on fire & tunnels. These RTD initiatives were set-up following some major tunnel fire accidents in recent time. FIT is a group of stakeholders focusing on the exchange of knowledge, integration and creation of know-how. FIT will have strong links with actual and future European and national research projects. The FIT Thematic Network starts on 1 March 2001 and runs for 4 years.

The following main objectives have been identified for the FIT Thematic Network:

• The network has as main objective the dissemination of RTD and design results obtained in European and National RTD projects. The aim is to optimize research efforts, to reach critical mass and to enhance impact at European level by combining the results of the different projects.

• FIT will establish a set of consultable database with essential knowledge on fire in tunnels.

• A third common objective of the network members is to realize recommendations on design fires for tunnels.

• Consequently FIT has also the objective to develop European consensus for fire safe design on the basis of existing national regulation, guidelines, code of practices and safety requirements.

• The last objective is the definition of best practices for tunnel authorities and fire emergency services on prevention and training, accident management and fire emergency operations.

1.2 Relevant data bases and documents The data bases and reports which were identified as being potentially relevant to SafeT tasks are listed in tables 1 and 2.

Table 1: Overview of relevant data bases

Task Description Relevant data bases Reason why

1.2 State of the art DB 4: Data on safety Gives an overview of the detection, prevention equipment in tunnels available safety equipment and traffic management 1.3 State of consequence DB 1: RTD on fire safety in Completed and current RTD mitigation tunnels 2.1 Systems and methods DB 4: Data on safety Could be part of the safety for incident detection equipment in tunnels equipment 2.3 Best practices in DB 5: Assessment of reports Information on current accident detection and on fire accidents in tunnels practices prevention 3.2 Training of rescue DB 1: RTD on fire safety in Contains also RTD on personnel tunnels training tools 3.3 Evaluation of tools for DB 5: Assessment reports Past accident reports to evacuation/intervention on fire accidents in tunnels show where improvement is management required 4.2 Recommendations for DB 5: Assessment reports Past accident reports to the collection of data on fire accidents in tunnels show how data are collected on tunnel incidents 6.3 Integrated ICT based DB 1: RTD on fire safety in Data base also includes tools for tunnel safety tunnels RTD on this topic management

Table 2: Overview of relevant reports

Task Description Relevant report Reason why

1.1 State of the art General approach to tunnel Helps in the literature search safety approach fire safety related to fire safety 1.6 State of the safety Best practice for fire Includes a review of current management response management practice systems 2.3 Best practices in General approach to tunnel States the general approach accident detection fire safety with respect to fire and prevention 3.2 Training of rescue Best practice for fire Includes training of fire men personnel response management

2. Summary of data base information

2.1 Database 1: RTD on fire safety in tunnels Here the RTD projects that are relevant to the SafeT objectives are summarized (details see Appendix I)

• SAFE TUNNEL (Innovative systems and frameworks for enhancing of traffic safety in road tunnels)

• UPTUN (Cost-effective Sustainable & Innovative Upgrading Methods for Fire Safety in Existing Tunnels)

• TP 10: Priority research project on safety of users in case of a fire

• TP 11: Priority research project on safety policy in tunnels

• SIRTAKI (Safety Improvement in Road & Rail tunnels using Advanced ICT and Knowledge Intensive DSS)

• Emergency scenarios for tunnels on public transportation rail systems and how to master them

• DARTS (Durable and reliable tunnel structures)

• A-TEAM (Improve learning process in complex technical domains)

• VIRTUAL FIRES (Virtual Real Time fire emergency simulator)

2.2 Database 4: Data on safety equipment in tunnels The complete database-index is quite extensive and is shown in Appendix II. Most of the equipment is related to fire detection and mitigation

2.3 Database 5: Assessment reports on fire accidents in tunnels This contains a good and extensive list of tunnel accidents too voluminous to be included in the report. A complete database-index is shown in Appendix III. An example of how data were collected (template) is shown in Appendix IV.

3. Summary of FIT reports relevant to SafeT

3.1 Report: General approach to tunnel fire safety Summary The report gives a general overview of all consecutive safety aspects of a Tunnel: accident prevention-mitigation-suppression-reopening of tunnel. After a chapter on the causes and consequences of fires the report deals in separate chapters in more detail with each of the aspects mentioned. The tunnel system is then discussed with respect to infrastructure (new and existing tunnels) and operation. One chapter deals with the prescriptive vs. performance based research to tunnel safety. There is a short chapter on risk analysis, where the process and the main methods of risk assessment available are discussed. The issue of cost vs. safety is also considered. The report ends with chapters on human behaviour (especially for road users) and on organisational plans for emergency staff.

Contents 1 General approach to tunnel fire safety 1.1.1 Causes of fire 1.1.2 Consequences of fire 1.1.3 Objectives for fire design 1.2 The consecutive safety aspects 1.2.1 Principles 1.2.2 Pro-active measures 1.2.3 Prevention 1.2.4 Mitigation 1.2.5 Suppression 1.2.6 Reopening 1.2.7 Evaluation 1.3 Integrated approach to safety in tunnels 1.3.1 The tunnel system 1.3.2 Prescriptive vs. performance-based approach 1.3.3 Risk analysis 1.3.4 Cost versus safety 1.3.5 Dangerous goods transport 1.4 Human behavior 1.4.1 Road Users 26 1.4.2 Train and metro passengers 1.4.3 Operators 1.4.4 Emergency Staff

Relevance to SafeT and rating The report is well structured and written and should be read by all members of the network that are concerned with the development of guidelines, as some of the information in the report is general and not only limited to fire.

3.2 Report: Best practice for fire response management

Summary This report presents an overview of the ability of the Fire and Rescue Service to deal with major fires in road tunnels and also a basic discussion for other types of tunnels. This is an area of common interest for both the Fire and Rescue Service, tunnel designers and tunnel owners, in order to create a common view of how rescue work should be interwoven with other aspects of a tunnel’s safety systems. The report describes how rescue work in connection with fires in road tunnels can be approached, and the problems and difficulties to be overcome in successful rescue work.

Contents 1 Introduction 2 A DESCRIPTION OF THE CONCEPT OF TACTICS FOR RESCUE WORK 3 TUNNELS AND FIRES IN TUNNELS 4 FIRES IN TUNNELS: FIRE AND RESCUE OPERATION 5 SUMMARY AND CONCLUSIONS

Relevance to SafeT and rating Although the report deals exclusively with the response to fire emergencies some of the critical comments made especially with respect to the coordination between the emergency services and tunnel operators will be of interest to the developers of SafeT guidelines.

4. Discussion

This report tries to summarize the data bases and reports of the FIT network that are relevant for tasks of the SafeT network. FIT deals only with fires whereas SafeT deals generally with safety aspects and because of possible overlaps with FIT fire safety is not considered explicitly. However, some of the FIT products were found to be relevant because they either contain general statements about tunnel safety and emergency response management or contain relevant data as for example for safety equipment. Some RTD projects that are listed in the data base are also relevant as they have a general safety aspect and the reports on fire accidents are also seen of interest not only because of their contents but also by the way that the information is gathered (template).

It can be seen that there are not many overlaps between FIT and SafeT, as the emphasis in FIT seems to be on fire safe design guidelines. Also there seems to be very little work done on risk assessment.

A few of the data bases and the two reports have been attached in Appendices.

5. Reference

FIT web page: http://www.etnfit.net/

6. Appendix I: RTD projects relevant to SafeT

15 Fire In Tunnels Seite 1 von 1

FIT info: You are logged in as a FIT - Network Member (normal rights)

Index / RTD Projects / Details: SAFE TUNNEL (Innovative systems and frameworks for enhancing of traffic safety in road tunnels) RTD Project general information Project Title: SAFE TUNNEL (Innovative systems and frameworks for enhancing of traffic safety in road tunnels) Type: European

Added by Martin Yves (martin) - BBRI

Funding Organisation: European Commission Timeframe: start stop 01-09-2001 31-08-2004 Contact Name: Paola Carrea Organisation Name: Postal Address: Strada Torino 50 Town: 10043 rbassano Country: Italy Phone: (39-011) 9083130 Fax: (39-011) 9083083 E-mailaddress: [email protected] Website:

RTD Project detail information Partners : Telecom Italia Lab Spa; Societe Francaise du Tunnel Routier du Fre; SITAF Spa - Societa Italiana Traforo Autostradale del Frejus; Fiat Engineering S.p.a; TLC Tecnosistemi S.A.; Ben-Gurion University of The Negev; ENEA - Ente per le Nuove Tecnologie, l'Energia e l'Ambiente; Renault VI; Tuev Kraftfahrt Gmbh Description : SAFE TUNNEL main objective is to contribute to reduce the number of accidents inside road tunnels by preventive safety measures. Main focus is to achieve a dramatic decrease of the "fire accidents". Basic ideas are to avoid the access into the tunnel to those vehicles with detected or imminent on-board anomalies and to introduce measures to achieve the control of the speed of the vehicles inside tunnel.

http://db.etnfit.net/db1/index.cfm?action=details&detail=10&print=TRUE 16.12.2004 Fire In Tunnels Seite 1 von 1

FIT info: You are logged in as a FIT - Network Member (normal rights)

Index / RTD Projects / Details: UPTUN (Cost-effective Sustainable & Innovative Upgrading Methods for Fire Safety in Existing Tunnel) RTD Project general information Project Title: UPTUN (Cost-effective Sustainable & Innovative Upgrading Methods for Fire Safety in Existing Tunnel) Type: European

Added by Martin Yves (martin) - BBRI

Funding Organisation: European Commission Timeframe: start stop 01-09-2002 31-08-2006 Contact Name: Kees Both Organisation Name: TNO Postal Address: PO Box 49 Town: 2600 AA Delft Country: The Netherlands Phone: +31 15 2763483 Fax: E-mailaddress: [email protected] Website: http://www.uptun.net

RTD Project detail information Partners : UPTUN represents 41 members (see http://www.uptun.net) Description : The Common Transport Policy is one of the original ingredients of the first European treaty (Treaty of Rome). Tunnels are important in the existing and envisaged transport system. However, the safety of tunnels was put in question after recent fires, which resulted in fatalities, casualties and economic damage. The fires also discouraged tunnel usage in some cases. This adversely affects a sustainable socio-economic development. The UPTUN project aims at solving these problems.

The UPTUN project aims at: 1) development and promotion of innovative, sustainable and low-cost measures to limit the probability and consequences of fires in existing tunnels and; 2) development and promotion of an integrated evaluating and upgrading procedure, incorporating the innovative measures, for existing tunnels to allow owners, stakeholders, designers and emergency teams to evaluate and upgrade the human and structural safety level.

http://db.etnfit.net/db1/index.cfm?action=details&detail=17&print=TRUE 16.12.2004 Fire In Tunnels Seite 1 von 1

FIT info: You are logged in as a FIT - Network Member (normal rights)

Index / RTD Projects / Details: TP 10 : Priority research project on safety of users in case of a fire RTD Project general information Project Title: TP 10 : Priority research project on safety of users in case of a fire Type: National

Added by Carlotti Pierre (carlotti) - CETU

Funding Organisation: Cetu Timeframe: start stop 01-01-2001 31-12-2005 Contact Name: P. Carlotti Organisation Name: Cetu Postal Address: 25 av. François Mitterrand Town: 69674 Bron Country: France Phone: Fax: +33 4 72 14 34 70 E-mailaddress: [email protected] Website: http://www.cetu.gouv.fr

RTD Project detail information Partners : Ministery of interior, CSTB, DREIF, University Lyon Claude Bernard, Imperial College London Description : The project aims to: a) explain the French regulations and if necessary suggest modifications to ithem b) suggest technical solutions to fulfill these regulations c) investigate new innovative methods and hypotheses compared to those of the regulations. In order to investigate these points in a consistent way, the project is organised in subtopics, called 'ARD' The main ARD of the project are: A1-07: guidelines for fire reaction of tunnel structures: the aim of this ARD is to provide guidelines to tunnel designers E1-02: use of ventilation during a fire: this ARD can be decomposed in 4 sub-subjects: 1) methods in longitudinal systems, including wall heating and optimal use 2) methods in transverse systems, especially the problem of longitudinal air flow control 3) physical understanding: theoretical aspects of tunnel fires 4) models: use and creation of computer models, 1D, zone models, 3D RANS, 3D LES E5-06: fire detection: understand the major needs for a correct detection, and analyse systems able to perfrom such a detection E5-09: fire extinguishing: analyse the cases where fixed devices such as water mist, are useful or dangerous G4-03: behaviour of materials and structures: 1) guidelines on materials to be used in tunnels 2) better understanding of concrete spalling Reference Publications : see Cetu publication list, http://www.cetu.gouv.fr Reference Article: DL_Washington_paper.doc MemoireTheseFD.pdf Eurotherm_2002.doc

http://db.etnfit.net/db1/index.cfm?action=details&detail=44&print=TRUE 16.12.2004 Fire In Tunnels Seite 1 von 1

FIT info: You are logged in as a FIT - Network Member (normal rights)

Index / RTD Projects / Details: TP 11 : Priority research project on safety policy in tunnels RTD Project general information Project Title: TP 11 : Priority research project on safety policy in tunnels Type: National

Added by Carlotti Pierre (carlotti) - CETU

Funding Organisation: Cetu Timeframe: start stop 01-01-2001 31-12-2003 Contact Name: P. Carlotti Organisation Name: Cetu Postal Address: 25 av. François Mitterrand Town: 69674 Bron Country: France Phone: Fax: +33 4 72 14 34 70 E-mailaddress: [email protected] Website: http://www.cetu.gouv.fr

RTD Project detail information Partners : Ecoles des Mines de PAris, AREA, Bonnard et Gardel SA, CSTB, DDE ses Bouches du Rhône, Docaligic Inflow, INERIS, Ligeron SA, Scetauroute, SETEC TPI, DDT (Ministery of Transportation), CESTR Description : The project aims to collect information on fires in tunnels and to follow the implementations of new regulations from a general policy point of view. It is broader than a strictly speaking 'fire safety' project, being concerned in all aspects of road tunnel safety. It is split into several sub-topics, called ARD, of which the following are concerned with fires: E5-01: feedback on actual accidents and fires: all incident, accident or fire which caused an unexpected closure on the French trunck network tunnels have to be reported to Cetu. This ARD aims to analyse these informations, and a report is issued each year. E5- 02: guidelines for tunnel safety files: tunnel safety files are compulsory in France. The aim of this ARD is to help tunnel operator to have high quality safety files Reference Publications : see Cetu publication list, www.cetu.gouv.fr see also documents available on Cetu website, http://www.cetu.gouv.fr Reference Article: Finalité_Protégé.pdf

http://db.etnfit.net/db1/index.cfm?action=details&detail=45&print=TRUE 16.12.2004 Fire In Tunnels Seite 1 von 1

FIT info: You are logged in as a FIT - Network Member (normal rights)

Index / RTD Projects / Details: SIRTAKI (Safety Improvement in Road & Rail tunnels using Advanced ICT and Knowledge Intensive DSS RTD Project general information Project Title: SIRTAKI (Safety Improvement in Road & Rail tunnels using Advanced ICT and Knowledge Intensive DSS Type: European

Added by Martin Yves (martin) - BBRI

Funding Organisation: European Commission Timeframe: start stop 01-09-2001 31-08-2004 Contact Name: Antonio Marqués Organisation Name: Postal Address: Calle Tres Forques 147 Town: 46014 Valencia Country: Spain Phone: (34-96) 3134082 Fax: (34-96) 3503234 E-mailaddress: [email protected] Website: http://www.sirtakiproject.com

RTD Project detail information Partners : Risoe National Laboratory; Regie Autonome des Transports Parisiens; Instituto Dalle Molle di Studi Sull'Intelligenza Artificiale; SITAF Spa - Societa Italiana Traforo Autostradale del Frejus; Sinelec S.P.a; Deutsche Bahn AG; FIT Consulting S.R.l.; SAFETEC NORDIC AS; Ajuntament de Barcelona; Athens University of Economics and Business; Servicios y Obras del Norte, S.A. Description : SIRTAKI plans to improve tunnel safety by the development and assessment of an advanced tunnel management system that specifically tackles (i) safety issues and emergencies and (ii) the integration within the overall network management. The benefits expected as result of the Project are measured in technical, social and economic terms, in particular by: Improving safety in tunnels: reducing the risk of accidents in tunnels and the severity of those taking place; Reducing stress in operators and citizens in front of an emergency; Managing tunnels and the rest of the transport network in a co-ordinated way and therefore improving the performance of the available transport infrastructures; Performing the integrated management of not only emergencies, but also other special situations -e.g. congestion, maintenance works, etc. The prototypes will be evaluated in four pilot tunnels in Spain, Italy, France and Germany. These tunnels will be road (urban, interurban), metro and railway which will validate the system developed in a wide range of situations.

http://db.etnfit.net/db1/index.cfm?action=details&detail=11&print=TRUE 16.12.2004 Fire In Tunnels Seite 1 von 1

FIT info: You are logged in as a FIT - Network Member (normal rights)

Index / RTD Projects / Details: Emergency scenarios for tunnels on public transportation rail systems and how to master them RTD Project general information Project Title: Emergency scenarios for tunnels on public transportation rail systems and how to master them Type: National

Added by Haack Alfred (haack) - STUVA

Funding Organisation: Federal Ministry for transport Timeframe: start stop 01-07-2002 30-06-2003 Contact Name: Alfred Haack Organisation Name: STUVA Postal Address: Mathias-Brüggen-Str. 41 Town: D-50827 Köln Country: Germany Phone: ++492215979510 Fax: ++492215979550 E-mailaddress: [email protected] Website: http://www.stuva.de

RTD Project detail information Partners : None Description : The aim of this research project is to develop a standard emergency scenario to act as the basis for the safety measures required for planning tunnels and operating concepts for public transportation rail systems in future. Further locally specific details must be decided on the spot e.g. so that the different types of rolling stock and stations can be taken into consideration. An orientation guide for cases of emergency and how to master them in tunnels on public transportation rail systems is to be provided within the scope of the research work. This code of practice is intended to assist planners, operators and fire brigades in resorting to suitable measures designed to avoid, reduce and master as many conceivable emergency situations as possible. First of all, different emergency scenarios e.g. technical defects, vehicle collisions, suicide, vandalism, criminal acts, fire, terrorism, hostage-taking, attacks using explosives are to be compiled and assessed with respect to possible consequences on services and rescue operations as well as potential damage to persons and property. The STUVA will propose one standard scenario from all those put forward. This proposal will be scrutinised by the working group accompanying the project. This group is made up of experts from the municipalities, the technical supervisory authorities, the Eisenbahn-Bundesamt, the Deutsche Bahn AG, the fire services, the BG Bahnen and transportation companies as well as the Federal Ministry for Transport, Construction and Housing. Reference Publications : tunnel 6/2002 page 52-54

http://db.etnfit.net/db1/index.cfm?action=details&detail=30&print=TRUE 16.12.2004 Fire In Tunnels Seite 1 von 1

FIT info: You are logged in as a FIT - Network Member (normal rights)

Index / RTD Projects / Details: DARTS (Durable and reliable tunnel structures) RTD Project general information Project Title: DARTS (Durable and reliable tunnel structures) Type: European

Added by Martin Yves (martin) - BBRI

Funding Organisation: European Comission Timeframe: start stop 01-03-2001 29-02-2004 Contact Name: Steen Rostam Organisation Name: COWI Postal Address: Parallelvej 15 Town: 2800 Lyngby Country: Denmark Phone: (45-45) 972935 Fax: (45-45) 972114 E-mailaddress: [email protected] Website: http://www.dartsproject.net

RTD Project detail information Partners : Bouygues Travaux Publics; Citytunnelkonsortiet i Malmoe; Ingenieurbuero Professor Schiessl; HOLLANDSE BETON GROEP NV; Rijkswaterstaat - Ministerie van Verkeer en Waterstaat; Netherlands Organization for Applied Scientific Research Description : The rapidly growing urbanisation within the European societies is leading to congestion of people and all types of associated traffic. Tunnels and underground structures are becoming indispensable when installing new infrastructure in congested area's as well as when raising the qualities of urban living within the existing urbanisations. Therefore the EU-countries will be investing 100-150 billion EURO the next 10 years. Disproportionate budget overruns, often > 100%, and serious construction delays have been the rule specifically for tunnels, due to their complexity in technical, organisational and environmental aspects. This is unacceptable in view of the projected investment. Integrated design methods and supportive tools are lacking with respect to life cycle tunnel optimisation including durability, environmental aspect, sustainability and safety. DARTS will develop integrated compatible methods of durability, environmental and hazard design and improve decision basis for owners.

http://db.etnfit.net/db1/index.cfm?action=details&detail=13&print=TRUE 16.12.2004 Fire In Tunnels Seite 1 von 1

FIT info: You are logged in as a FIT - Network Member (normal rights)

Index / RTD Projects / Details: A-TEAM (IST project) RTD Project general information Project Title: A-TEAM (IST project) Type: European

Added by Marlair Guy (marlair) - INERIS

Funding Organisation: CEC 5th framework Inf. Soc. DG Timeframe: start stop 01-06-2000 31-12-2003 Contact Name: Prof Nicolas Moussiopoulos Organisation Name: LHTEE Univ Thessaloniki Postal Address: PO Box 483, 54006 Town: THESSALONIKI Country: Greece Phone: + 30 31 996011 Fax: + 30 31 996011 E-mailaddress: [email protected] Website: http://www.ess.co.at/A-TEAM

RTD Project detail information Partners : Environmental Software & Service GmbH (ESS) (AUT), Donau Chemie AG (AUT), Chiron (P), DNV (UK), Universitat Politecnica de Catalunya (SP), SYRECO (I), Aristole University of Thessaloniki (GR), ASIT (CH) Description : Main Objective of the A-TEAM project is to improve learning process in complex technical domains, using the example of technological emergency management; Improved learning is integrating infomation technologies such as dynamic simulation, visualisation, GIS, Expert Systems reasoning and training case studies are including fires in rai and road tunnels Reference Publications : Dispersion of a passive pollutant in the vicinity of a U shaped Building by Goetting et al (1997), Int Jl of Environment and Pollution

http://db.etnfit.net/db1/index.cfm?action=details&detail=61&print=TRUE 16.12.2004 Fire In Tunnels Seite 1 von 1

FIT info: You are logged in as a FIT - Network Member (normal rights)

Index / RTD Projects / Details: VIRTUAL FIRES (Virtual Real Time Emergency Simulator) RTD Project general information Project Title: VIRTUAL FIRES (Virtual Real Time Emergency Simulator) Type: European

Added by Martin Yves (martin) - BBRI

Funding Organisation: European Commission Timeframe: start stop 01-11-2001 30-04-2004 Contact Name: Gernot Beer Organisation Name: Technical University Graz Situ Postal Address: Lessingstrasse 25/II Town: 8010 Graz Country: Austria Phone: (43-316) 8736185 Fax: (43-316) 8736185 E-mailaddress: [email protected] Website: http://www.virtualfires.org/

RTD Project detail information Partners : Stadt Dortmund; European Virtual Engineering, S.A.; Kungliga Tekniska Hoegskolan; University of Leoben; Fraunhofer Gesellschaft zur Foerderung der Angewandten Forschung E.V.; Alpetunnel; Ministere de l'Equipement, des Transports et du Logement Description : A Virtual Reality Real Time Fire Emergency Simulator (VIRTUALFIRES) will be developed. The VIRTUALFIRES system will be a unique system that can be used for assessing the fire safety of tunnels, for training of rescue personnel and for planning rescue scenarios and will be able to replace real fire tests. The end users of this system will be tunnel operators, government organizations concerned about tunnel safety and rescue organizations such as fire brigades. The system can be used for ascertaining the fire safety of existing European tunnels.

Objectives: The project has the following objectives: - To offer a cheaper and environmentally friendlier alternative to real Fire tests currently being carried out in tunnels by replacing them with virtual tests; - To develop and implement a virtual fire emergency simulator (VIRTUALFIRES) with complete user interaction; - To contribute to the improvement of the fire safety of tunnels by allowing to perform a large number of inexpensive fire tests for the evaluation of rescue scenarios and for testing fire safety and ventilation systems; - To make VIRTUALFIRES available for use by tunnel operators, designers and government regulatory authorities.

http://db.etnfit.net/db1/index.cfm?action=details&detail=12&print=TRUE 16.12.2004

7. Appendix II: List of safety equipment

16 Fire In Tunnels Seite 1 von 3

FIT info: You are logged in as a FIT - Network Member (normal rights)

Index / Equipment Insert Equipment

Legend: New: not treated Under Approval Accepted

All Equipment Information from all the categories. Monitoring Equipment Main Creation Sub category Mod. Date Detection name category Date Extinguishing infrared beam Communication Detection Smoke 2002-03-28 Ventilation detector Indication Linear resistance Detection Temperature 2002-03-28 2002-04-11 Traffic control heat detector Rescue scattered light Detection Smoke 2002-03-28 2002-04-11 Protection smoke detector

fire extinguisher Extinguishing Manual 2002-03-28 with ABC-Powder maximum heat Detection Temperature 2002-04-30 detector differential heat Detection Temperature 2002-04-30 detector fire extinguisher Extinguishing Manual 2002-04-30 with foam Photoluminescent safety guidance Indication Emergency signs 2002-04-30 system photoluminescent escape route Indication Emergency signs 2002-04-30 signs variable message Traffic control Indicators 2002-05-02 signs photoluminescent Indication Emergency signs 2002-05-02 emergency signs high pressure fire Extinguishing Manual 2002-05-02 extinguisher photoionization Detection Gases 2002-05-02 detector (PID) video identification Traffic control Counting 2002-05-06 system impavement Traffic control Indicators 2002-05-06 lights resistance Detection Temperature 2002-05-06 thermometer thermocouples Detection Temperature 2002-05-08 passiv-infrared- Traffic control Counting 2002-05-08 detector smoke hood Rescue Mobile rescue equipment 2002-05-08 2002-06-14 cage stretcher Rescue Mobile rescue equipment 2002-05-08 railbound trolley Rescue Mobile rescue equipment 2002-05-08 video identification Monitoring Visual 2002-05-14 systems

optical Monitoring Optical 2002-05-14 absorption http://db.etnfit.net/db4/?print=TRUE 16.12.2004 Fire In Tunnels Seite 2 von 3

measuring instrument passiv-infrared- Monitoring Visual 2002-05-14 detector illuminated signs Indication Emergency signs 2002-05-16 illuminated escaperoute Indication Emergency signs 2002-05-16 signs lamps for escape Indication Emergency signs 2002-05-16 route lighting fire protection Protection Structure 2002-05-16 building boards underground water tank for Extinguishing Installation for fire brigades 2002-05-16 fire fighting fire extinguishing Extinguishing Installation for fire brigades 2002-05-21 2002-05-22 pipe system fire exting. pipe Extinguishing Manual 2002-05-21 2002-05-22 system low location Indication Emergency signs 2002-05-21 lighting wall hydrant with Extinguishing Manual 2002-05-21 2002-05-22 semi-rigid hose push button Communication Alarms 2002-05-21 alarm electrochemical Detection Gases 2002-05-22 gasdetector infrared Detection Gases 2002-05-22 gasanalyser flame ionization Detection Gases 2002-05-22 detector (FID) heat conductivity Detection Gases 2002-05-22 sensor thermowire detector "twisto- Detection Temperature 2002-05-22 wire"/"Alarmline" 'VIP/I - Incident Traffic control Counting 2002-05-23 Monitor linear rate-of-rise temperature Detection Temperature 2002-05-31 2002-06-26 detector heat interference Detection Temperature 2002-05-31 sensor "Firant" sprinkler system Extinguishing Fixed installation 2002-05-31 IFEX railed Extinguishing Fixed installation 2002-05-31 2002-05-31 system IFEX fixed Extinguishing Fixed installation 2002-05-31 system water mist extinguishing Extinguishing Fixed installation 2002-06-07 systems fire extinguishing pipe system Extinguishing Fixed installation 2002-06-07 "wet" UV-flame- Detection Radiation 2002-06-07 detector thermowire- detector "twisto- Detection Temperature 2002-06-07 wire"/"Alarmline"

water-spray http://db.etnfit.net/db4/?print=TRUE 16.12.2004 Fire In Tunnels Seite 3 von 3

system Extinguishing Fixed installation 2002-06-07 standpipe Extinguishing Manual 2002-06-10 linear fibre-optic Detection Temperature 2002-06-14 detector semiconductor Detection Gases 2002-06-14 gas detector aspirating smoke detection Detection Smoke 2002-06-14 systemm heat toning Detection Gases 2002-06-14 detector ionization smoke Detection Smoke 2002-06-19 detector IR-flame- Detection Radiation 2002-06-19 detector shine-through- Detection Smoke 2002-06-19 detector shine-through detector / optical Detection Smoke 2002-06-26 smoke detector Pipe systems for Extinguishing Installation for fire brigades 2002-12-20 fire mains GRP (Fiberglass) cable Protection Installation 2003-05-16 management Suction Device Ventilation Fans 2003-07-23 Xenvi system Monitoring Others 2003-07-30 Multicomponent gas analyzer, Detection Gases 2003-08-13 GASMET HI-FOG Sprinkler Extinguishing Fixed installation 2003-08-15 system Temporary Rescue Chambers 2003-09-08 escape tunnel Fire limiting Rescue Doors 2003-09-08 screen Water curtain Rescue Doors 2003-09-08 partition fast response Extinguishing Installation for fire brigades 2003-09-09 vehicles linear fibre-optic temperature Detection Temperature 2003-09-09 measurement high pressure Extinguishing Manual 2003-09-09 wall cabinets Escape Breathing Rescue Mobile rescue equipment 2003-09-09 Apparatus compressed foam extinguishing Extinguishing Fixed installation 2004-04-03 system

http://db.etnfit.net/db4/?print=TRUE 16.12.2004

8. Appendix III: Fire accidents in Tunnels

17 Fire In Tunnels Seite 1 von 3

FIT info: You are logged in as a FIT - Network Member (admin rights)

Index / Fire Accidents Create new Fire Event for Road Tunnel Create new Fire Event for Metro/Railway Tunnel

Legend: New: not treated Under Approval Accepted

Brief overview of existing records of fire in tunnels Update date Excel sheet with short information about fire accidents in tunnel 23 November 2004 Comment Excel sheet with short information about fire accidents in tunnel 16 December 2004 Comment To update this list: click on comment and send your information to the database manager (an update every 3 months is foreseen).

Road Tunnel Creation Date Mod. Date App. Date 44 - France (France) 72 - France (France) 80 - France (France) BAREGG TUNNEL (Switzerland) Fire after vehicle crash (14-04-04 2004-07-22 2004-12-02 2004-07-22 Comment 14:02) 12:21:57.0 17:10:39.0 12:23:47.0 CASTELLAR (Nice) (France) Spontaneous fire on vehicle (14-04- 2002-10-18 2003-02-18 Comment 94 ??:??) 19:51:05.0 18:15:12.0 Caldecott Tunnel (United States) Fire after vehicle crash (07-04-82 2004-11-24 Comment 24:00) 11:29:33.0 FELBERTAUERN (Austria) Spontaneous fire on vehicle (01-06- 2002-10-09 2003-02-18 Comment 84 ??) 11:49:16.0 18:13:55.0 FREJUS road tunnel (France) Spontaneous fire on vehicle (05-05-83 2002-10-16 2003-02-18 Comment 17:46) 16:13:03.0 18:15:21.0 GUMEFENS (Switzerland) 2002-10-25 2003-02-18 Fire after vehicle crash (1987 ??:??) Comment 23:29:20.0 18:13:47.0 Homer Tunnel (New Zealand) Spontaneous fire on vehicle (03-11- 2003-09-05 2004-07-19 Comment 02 ??:??) 11:57:39.0 19:36:28.0 ISOLA DELLE FEMMINE (Italy) Fire after vehicle crash (18-03-96 2002-10-25 2003-02-18 Comment 02:30pm) 22:53:06.0 18:14:56.0 Kinkempois - Liège (Belgium) Spontaneous fire on vehicle (03-02-04 2004-06-28 2004-07-19 Comment 14h15) 16:52:49.0 19:37:20.0 L'Arme (France) Fire after vehicle crash (09-09- 2002-10-16 2003-09-03 http://db.etnfit.net/db5/ 16.12.2004 Fire In Tunnels Seite 2 von 3

1986 ??:??) 18:45:52.0 19:11:18.0 Comment MONT-BLANC (France) Spontaneous fire on vehicle (10-11-04 2004-11-16 Comment 00:00) 17:03:32.0 Spontaneous fire on vehicle (24-03- 2002-10-19 2002-10-22 2003-09-03 Comment 1999 10:53) 21:56:37.0 11:19:05.0 19:10:20.0 Spontaneous fire on vehicle (11-01- 2002-10-22 2003-09-03 Comment 1990 10:42) 11:45:23.0 19:10:20.0 MOORFLEET (Hamburg) (Germany ) Other type of fire event (31-08-1969 2002-10-22 2003-02-18 Comment 01:10) 22:56:51.0 18:15:03.0 Nihonzaka-Tunnel-Japan (Japan) PECORILA GALLERIA (Italy) 2002-10-22 2003-09-03 Fire after vehicle crash (1983 ??:??) Comment 23:47:09.0 19:03:46.0 PFANDER (Austria) Fire after vehicle crash (10-09-1995 2002-10-09 2003-03-14 Comment 08:41) 13:56:42.0 13:36:55.0 PORTE D'ITALIE (Paris) (France ) Spontaneous fire on vehicle (11-08-76 2002-10-18 2003-02-18 Comment 20:00) 19:01:45.0 18:13:36.0 Prapontin (Italy) Rheinufertunnel (Germany) SELJESTAD TUNNEL (Norway) Fire after vehicle crash (14-07-00 2002-09-29 2003-06-05 Comment 20:50) 23:50:15.0 17:58:15.0 SERRA RIPOLI (Italy) 2002-10-25 2003-09-03 Fire after vehicle crash (1993 ??:??) Comment 22:32:34.0 19:10:53.0 ST GOTTHARD road tunnel (Switzerland) Fire after vehicle crash (24-10-2001 2002-06-20 2003-02-18 Comment 9h39) 14:58:03.0 18:14:01.0 Spontaneous fire on vehicle (17-09- 2002-11-08 2003-06-05 Comment 1997 11:13) 10:15:40.0 17:57:46.0 Spontaneous fire on vehicle (31-10- 2002-11-08 2003-06-05 Comment 1997 07:20) 10:32:59.0 17:58:01.0 Tauern tunnel (Austria) Fire after vehicle crash (29-05-99 2002-10-16 2003-09-03 Comment 04:50) 12:28:04.0 19:09:57.0 Railway and Metro tunnel Creation Date Mod. Date App. Date Baku Metro (Baku) 2003-03-14 2003-09-03 Fire event on 28-10-85 at Comment 13:36:19.0 19:08:03.0 Channel tunnel (Coquelles -) 2003-03-13 2003-09-05 2004-07-19 Fire event on 18-11-96 at Comment 11:18:55.0 09:36:37.0 19:36:46.0 Daegu metro system (Daegu) 2003-02-18 2003-09-05 2004-07-19 Fire event on 18-02-2003 at Comment 18:30:56.0 14:58:04.0 19:37:32.0 Exilles Tunnel (near Turin) (Exilles) 2003-08-15 2004-07-19 Fire event on 1 July 1 at 13.30 Comment 11:08:24.0 19:37:42.0 Great Belt East Tunnel (Korsoer) 2002-11-21 2003-09-05 Fire event on 11.6.1994 at Comment 15:47:04.0 10:18:48.0 Guadarrama tunnel (80 km far from Madrid, near Segovia)

2003-09-05 2003-09-05 2004-07-19 http://db.etnfit.net/db5/ 16.12.2004 Fire In Tunnels Seite 3 von 3

Fire event on 06-08-03 at ??:?? 10:12:11.0 10:17:12.0 19:38:14.0 Comment Howard street tunnel Baltimore (BALTIMORE) 2002-10-17 2003-09-12 2003-09-06 Fire event on 18-07-2001 at 15:10 Comment 09:49:21.0 14:20:55.0 17:17:21.0 Kitzsteinhorn tunnel (Kaprun) 2003-03-13 2003-09-05 Fire event on 11-11-00 at 09:05 Comment 09:23:15.0 10:18:06.0 Leinebusch Tunnel (Goettingen) 2003-03-16 2003-09-06 Fire event on 01-03-19 at 13:00 Comment 09:50:47.0 17:17:42.0 Salerno tunnel (Salerno) 2003-02-18 2003-09-03 Fire event on 23-05-1999 at Comment 18:12:35.0 19:07:26.0 Summit Tunnel (Todmorden (2 miles)) 2003-03-14 2003-09-03 Fire event on 20-12-84 at 06:00 Comment 12:58:33.0 19:36:33.0 Tunnel Km 23.2 Line: Genova Nervi-Pisa Centrale; (Genova Nervi) 2003-09-22 2004-07-19 Fire event on 03-05-02 at 15.20 Comment 17:41:07.0 19:37:55.0 Tunnel on Line: Nodo di Napoli (Pozzuoli-Napoli) (Napoli) 2003-09-23 2004-07-19 Fire event on 29-03-02 at 22:05 Comment 10:34:51.0 19:38:04.0 Tunnel on line: Ozieri Chilivani - Decimomannu (Cagliari) 2003-09-23 2004-07-19 Fire event on 01-08-02 at 07:07 Comment 11:19:54.0 19:38:23.0 Weesperplein Metro Station (Amsterdam) 2003-08-14 2004-07-19 Fire event on 12.0799 at Comment 12:36:41.0 19:38:41.0 tunnel on Line: Arquata Scrivia-Mignanegro (Genova) 2003-09-23 2004-07-19 Fire event on 19-11-01 at 22:45 Comment 10:12:53.0 19:38:32.0

http://db.etnfit.net/db5/ 16.12.2004

9. Appendix IV:Sample of accident report sheet

18 Fire In Tunnels Seite 1 von 3

FIT info: You are logged in as a FIT - Network Member (admin rights)

Index / Fire Accidents / Show Fire Accident

Road event added by Marlair Guy (marlair) - INERIS

Available to the public: Public fire event (visible for FIT members and FIT Corresponding WWW visibility of this report: Members) Source of information Tunnel: BAREGG TUNNEL - Switzerland Tunnel Tube: NORTH TUBE BERN

Fire event Type of event: Fire after vehicle crash Type of collision: Vehicles moving on opposite lanes in opposite direction Type of vehicles involved: Other (please give details) A truck couldn't stop on time to avoid stopped vehicle ahead and ran back Details: of then rode up over a Jeep Cherokee car

Positioning of the fire event in time and space: Date of occurrence: 14-04-04 Time of occurrence: 14:02 Accurate location of fire: in middle

Information regarding gravity of the fire event: Number of slightly injured people: 5 Number of heavily injured people: 0 Number of fire deaths: 1 Damage to tunnel structure: Minor damage expanding less than 10m of tube section Duration of tunnel unavailability for traffic: 6 to 40 hours

Costs induced by repairs and traffic rupture: Direct costs: unknown kEuros/na Indirect costs (operation losses,...): unknown kEuros/na

Causes and main circumstances: Main cause type: HGV on fire Spontaneous: no Details of spontaneous fire: --- Accurate location of fire origin: Tractor Fire burst out after crash: yes Due to involvement of flammable substance: yes Fire caught due to arson: no

Other major known circumstances: Vehicle driving on wrong lane: no Presence of object of obstacle on lane: yes Technical incident immediately preceding fire: no Lighning break-down: no Ventilation break-down: no Flammable liquid spill: no Unusual temperature or pollution level in tunnel: no http://db.etnfit.net/db5/index.cfm?action=details_road_event&eventId=141 16.12.2004 Fire In Tunnels Seite 2 von 3

Vehicles involved in fire event: Number of passenger cars: 1 Number of 2-wheeled vehicles: Number of HGV's: 1 Number of busses or coaches: Other vehicles: Number + description: Vehicles carrying dangerous goods involved: None Dangerous goods class: None Dangerous goods class details: Dangerous goods released during the fire event: None Main DG involved in release:

Detection of fire event: Primary fire detection mode: Close circuit television (TV) Other primary fire detection mode: end users Secondary fire detection mode: Other (please give details) Other secondary fire detection mode: dont know

Counter-measures taken to cope with the situation: Action of operating staff (including tunnel 2 min firemen): Police assistance: Public fire brigade: 11 min Break-down service intervention: 7 min Other action: Other action delay:

Use of equipment during fire scenario: Service traffic lights: Yes Ground marking, signaling: -- Tunnel extinguishers: -- Tunnel fire network: -- Videotape recorder: -- Other equipment:

Smoke venting: Smoke venting activation mode: None Delay time from first alarm:

Fire control mode: How was fire put under control (main): By public fire brigade How was fire put under control (secondary): ---

Information given to tunnel users: Equipments used for information of the public: Other equipment used for information of the public: Other equipment used for information of the public: Other equipment used for information of the public: Information given about vehicle breakdown: -- Information given about lane obstruction: -- Information given about fire event: -- Information given to stay in vehicle: -- Information given to leave vehicles and move to -- place of safety (shelters): http://db.etnfit.net/db5/index.cfm?action=details_road_event&eventId=141 16.12.2004 Fire In Tunnels Seite 3 von 3

Information given to evacuate tunnel: -- Other message given to the public: Languages used for the information: Official language(s) of country concerned only

Users behaviour in response to fire event: Use of garages: no Back turn in tunnel: no Escape on foot by one end of tunnel: -- Use of emergency exits to the open: -- Transfer to other tube: yes Use of gallery (service gallery): no Shelter into places of safety (refuges): no people all went through cross connections to other tubes in the middle of Other behaviour of the users: the tunnel

Aftermaths of the fire accidents: yes Official enquiries undertaken: if yes, official enquiry type: Administration Tunnel repair needed: yes Type of refurbishments: lots of electric needed replacement + road surface New safety measures implemented: no Did the accident modify consequently tunnel no operation/regulation: What are the main consequences: Has the accident had noticable impact on no national regulation:

Reference: Other known complementary information in open press releases at the time + web site of the tunnel operator literature: good report + pictures of scenario in www.feuerwehr-wettingen.ch/baregg04b.htm

Official report:

Other information not given elsewhere in this accident report: personal communication by Electowatt Infra Ltd (Zurich), Dec. 2004 about suggested corrections and amendments.

http://db.etnfit.net/db5/index.cfm?action=details_road_event&eventId=141 16.12.2004

10. Appendix V: FIT reports

FIT European Thematic Network FIT Report General approach to tunnel fire safety

November 2004 FIT European Thematic Network FIT Report

General approach to tunnel fire safety

November 2004

Report no. Contribution to FIT report Issue no. 0 Date of issue 22 November.2003

Prepared … Table of Contents

1 General approach to tunnel fire safety 2 1.1.1 Causes of fire 2 1.1.2 Consequences of fire 4 1.1.3 Objectives for fire design 5 1.2 The consecutive safety aspects 5 1.2.1 Principles 5 1.2.2 Pro-active measures 5 1.2.3 Prevention 6 1.2.4 Mitigation 10 1.2.5 Suppression 12 1.2.6 Reopening 13 1.2.7 Evaluation 14 1.3 Integrated approach to safety in tunnels 14 1.3.1 The tunnel system 15 1.3.2 Prescriptive vs. performance-based approach 17 1.3.3 Risk analysis 20 1.3.4 Cost versus safety 23 1.3.5 Dangerous goods transport 24 1.4 Human behaviour 25 1.4.1 Road Users 26 1.4.2 Train and metro passengers 28 1.4.3 Operators 28 1.4.4 Emergency Staff 29 1 General approach to tunnel fire safety In general it can be observed that the safety level in tunnels is comparable to similar open routes of the infrastructure. In fact it can be observed that fatality risk in tunnels is often lower per vehicle km than it is for a comparable open part of the infrastructure. However, some specific events are unique for tunnels or can result in much more severe consequences in a tunnel than for an open section. So, for tunnels it is of major importance to study these events. Exam- ples of tunnel specific events are explosions, release of toxic gases and other dangerous substances and for some tunnels also flooding. However, first and foremost the discussion of tunnel specific risks is related to fire in tunnels.

In spite of the relative few lives lost in tunnel fires compared to the number of fatalities in road and rail traffic in general fire in tunnels is the tunnel specific risk with the largest influence on the safety of the users.

The topic of tunnel fires has been in focus in the recent years. The fires and the media coverage have had an impact on the perception of the safety of tunnels. For this reason it is necessary from a tunnel industry point of view to take actions for improving the safety. Some improvements may certainly be justified from a technical and risk evaluation point of view. However, care should be taken not to introduce all conceivable measures without consideration of their cost implications for installation, operation, maintenance and impact on the op- erability. The aim must be to achieve safety with a balanced use of the resources and with convenience for the users.

1.1.1 Causes of fire A fire requires 3 basic elements: fuel, oxygen and an ignition source. For all practical infrastructure tunnels these elements will be available to some extent, so the risk of fire cannot be completely mitigated.

A fire safety analysis will include a study of the materials, which (potentially) are available in the tunnel. In addition it is studied how the materials can be ig- nited and how the burning process under given circumstances develops. Com- bustible materials behave quite differently depending on their physical condition and a number of circumstances and conditions in the tunnel will have to be considered. Also interaction with ventilation and other safety systems in the tunnel will have to be taken into account.

Road tunnels A number of unwanted events, combination of events and various faults can lead to a fire in a tunnel. Up to about 95% of all fires are caused by electrical and mechanical defects in the vehicles, for example: • Electrical defects • Motor overheating • Brake overheating

Other less frequent causes are:

D:\NPH_COWI_Activities\52604_FIT\CommonPart\FIT_CommonPart5_Rev.DOC 2 • Collisions • Technical defects of tunnel equipment • Maintenance works in tunnels

Collisions and other vehicle accidents are not the most frequent cause of fires, on the other hand most large fires are caused by accidents. The original cause of collisions and other traffic accidents is often in-attention of the driver or similar, see also section 1.4.

The probability of fires in heavy good vehicles is larger than for passenger cars, and when heavy goods vehicles is involved in the fire, there is a higher risk that the fire develops into a serious fire.

The duration of recorded serious fires in road tunnels ranges from 20 minutes to 4 days. Most of the serious fires are in the magnitude of 2 - 3 hours, however, 4 fires in road tunnels stand out as particularly serious: • Nihonzaka, Japan, 1979, collision, duration 4 days • Mont Blanc, France/Italy 1999, overheated motor, duration 53 hours • Tauern, Austria, 1999, collision, duration 15 hours • Gotthard, Switzerland, 2001, collision, duration 20 hours

Rail tunnels For fires in rail tunnels it is important to distinguish fires in passenger trains and fires in goods trains. Passenger trains do not transport goods other than baggage and in some cases mail. Goods trains can transport very significant quantities of goods also including dangerous goods, but the number of persons on the train is limited to the loco-driver and in some cases a staff of 1 - 2 persons. Only in twin track rail tunnels and in long rail tunnels there is a chance to have more than one trains in the tunnel at he same time. In case of an incident the train, which may be in the tunnel at the time of the incident but not directly affected will in most case be able to drive out. Hence, fires in goods trains may involve a significant fire load and has the potential for developing into serious and even extreme conditions, which is some cases can endanger the structure and devastate the installations. However, the number of persons affected by the fire is in most cases very limited. Fires in goods trains can be caused by e.g.: • Derailment, collision and other train accidents • Electrical defects in the wagons • Mechanical defects in the wagons, e.g. the axles and brakes • Fires starting in the goods

Fires in passenger trains involve in the majority of cases a rather modest fire load. This is the result of design requirements for the rolling stock. On the other hand a large number of passengers can be exposed in a limited space. Fires in passenger trains can be caused by e.g.: • Carelessness by passengers • Arson • Derailment, collision and other train accidents

D:\NPH_COWI_Activities\52604_FIT\CommonPart\FIT_CommonPart5_Rev.DOC 3 • Electrical defects or defects in the heating system • Mechanical defects in the wagons, e.g. the axles and brakes

A majority of fires inside the rolling stock will not develop seriously during the time the train uses to pass the tunnel. In addition the fires are better fought outside the tunnel, so for this reason there has been introduced a "drive-out- concept" in many tunnels, whereby the train will drive out of the tunnel also in case of a fire in the train. For rail tunnels there is generally more electrical and mechanical equipment in the tunnel, and fires in the installations contribute to the fire risk in rail tunnels. The probability of fire is dependent on the power consumption, the design standard and the maintenance of the installations. The installations may be located either directly in the tunnel or in separate rooms, for example cross passages or service galleries. Examples of particularly serious rail tunnel fires: • Summit tunnel, UK, 1984, derailment of goods train, 13 petrol tankers • Channel Tunnel, UK/France, 1996, arson (?), severe structural damage • Kitzsteinhorn, Austria, 2000, fire in heating system of funicular railway, 155 fatalities

Metro tunnels With respect to the causes of fires metro tunnels are in may respects similar to passenger train tunnels. One significant difference is the short distance between stations, the fact that most stations are underground and that the underground stations are often connected to shopping arcades and other buildings. For metro tunnels fires starting in stations and adjacent facilities and spreading to the metro area is a relevant hazard.

Another characteristic of metro tunnels is the often very high density of passengers.

Examples of particularly serious metro tunnel fires: • Kings Cross Station, UK, 1987, escalator fire, 31 fatalities • Baku Metro, Azerbaijan, 1995, electrical defect, 300 fatalities • Daegu Subway, Korea, 2003, arson, 194 fatalities.

1.1.2 Consequences of fire Fires cause generally heat, smoke and toxic products, which can lead to de- struction of values and loss of lives. Heat is the cause of damage to structure and installations, whereas it is rarely the original cause of death. The threat to humans is in first hand smoke (lack of fresh air, obscuration) and toxicity. Fi- nally, fires may potentially also represent a hazard to the environment caused by the toxicity of the smoke and substances in the drainage.

The main consequences of fires considered can be distinguished as: • Fatalities and injuries for - Tunnel users - Residents near the tunnel

D:\NPH_COWI_Activities\52604_FIT\CommonPart\FIT_CommonPart5_Rev.DOC 4 - Operating personnel - Rescue forces • Values lost by the fire and costs of repair of damage / reconstruction • Traffic disturbance due to closure or reduced service level of the tunnel after fire (for example: re-routing resulting in extra transport time, direct economic losses and possibly increased risk to the users) • Potential environmental damage due to the fire.

1.1.3 Objectives for fire design The objectives for fire design is to ensure the level of safety by evading critical events that may endanger human life, the environment and the tunnel structure and installations, as well as by the provision of protection in case of fire. It is also part of the objectives to provide the facilities for the rescue and to aim at a reasonable safety of the rescue personnel.

The risk of fires leading to loss of lives, injuries, damage to the structure and disruption of traffic may be reduced partly by preventive measures, partly by mitigation of the consequences of fires and by improving the possibilities for intervention and rescue in case of serious events.

1.2 The consecutive safety aspects

1.2.1 Principles A 100% safety against fire cannot be achieved, but actions can be taken to reduce the risk to a reasonable minimum. The actions taken to achieve fire safety in a tunnel are formulated as "consecutive safety aspects" as illustrated in Figure 1.1. The consecutive safety aspects are sometime referred to as a "safety chain", but it should be noted that the reduced measures in one "link" could to some extent be compensated by increased measures in another "link". Hence the aspects are not connected by a serial chain structure but are rather serial and parallel combinations.

Pro-active Prevention Mitigation Suppression Reopening Evaluation measures

Figure 1.1 Consecutive safety aspects

1.2.2 Pro-active measures The pro-active measures comprise all the actions taken in the planning phase to improve the safety. • Legislative initiatives and other actions, which highlight the awareness of the problem and contribute to an improvement of the standards related to designing and operating tunnels, count as pro-active measures.

D:\NPH_COWI_Activities\52604_FIT\CommonPart\FIT_CommonPart5_Rev.DOC 5 • Research project and similar gaining know-how about tunnel fire contribute to future tunnel safety. In this respect the FIT network itself is a pro- active measure. A number of RTD projects are listed in FIT database 1 • The pro-active measures may be actions and considerations made as part of the design process related to the tunnel lay-out and the design of structure and installations. • Also the preparedness and emergency planning is a pro-active measure, which can positively influence the safety in the tunnel. • In the anticipation of the conditions prevailing during a fire it may be relevant as part of the pro-active measures to carry out fire tests or to model the fire numerically. In FIT database 2 and 3 test sites and computer codes are presented. • In addition the general pro-active measures concerned with user behaviour in terms of more careful driving and correct behaviour in case of an incident may significantly influence the safety in the tunnel.

1.2.3 Prevention General Preventive measures are safety measures, which reduce the probability that the unwanted event occur. Preventive safety measures in tunnels can be related to • Organisation and traffic management • Structural or geometrical solutions • Safety equipment.

Preventive measures concerned with fire in a tunnel are related to • Removal of sources of ignition • Reduction of the likelihood of a fire. • Prevention of the development from the ignition to a severe fire

The preventive measures are related to the causes of fires (see section 1.1.1). The typical causes are • Electrical/mechanical defects • Accidents • Defects in tunnel equipment • Arson and other detrimental action

The preventive safety measures will have to be decided based on a study of the effect of the measures and their disadvantages and costs. It should be noted that many preventive measures can have effects on several unwanted events. The evaluation of the safety measure should take into account all these effects. A systematic study of causes of fire and the risk reducing effect of preventive measures can be carried out for example in terms of so-called "Fault tree analyses", see also section 1.3.3.

Some measures to prevent fire risk in road- rail- and metro tunnels are presented briefly in the following.

Preventive measures in road tunnels Electrical/mechanical defects in the vehicles

D:\NPH_COWI_Activities\52604_FIT\CommonPart\FIT_CommonPart5_Rev.DOC 6 Source. The sources of electrical and mechanical defects resulting in vehicle fires are the various parts of the vehicle, which are necessary for running the vehicle, for example the combustion engine, the brakes, the heating/ cooling system etc. The source cannot realistically be re- moved, but with improved design standards of the vehicles the probability of fires could possibly be reduced. Probability. The probability can in addition be influenced by the alignment of the road outside the tunnel and of the tunnel itself. Steep gradients tend to give a higher risk of fires. The probability of bringing a vehicle on fire or with a high risk of fire into the tunnel can be reduced by inspection of the vehicles outside the tunnel. However, this measure is not realistic in all cases. The chance that a vehicle on fire can drive out of the tunnel by it own force can be influenced by geometrical design and tunnel lay-out. Development. The probability of a development from the first ignition to a fully developed fire endangering structure and users can be influenced by the design of the vehicles, restrictions to the nature and quantity of goods in the vehicle, the quantity of fuel and the distance between the vehicles. Also the construction material of the tunnel shall be chosen so that it does not contribute to the fire load. Accidents Source. The most common source of accidents is different types of human error and misconceptions, see section 1.4. Human errors can in general not be completely prevented, but the awareness can be influenced by different means and the probability of accidents can be reduced Probability. The probability of road accidents and accidents in tunnels can be influenced with a variety of measures. To some degree the fact that the road located in a tunnel can be regarded as a preventive measure in itself; some accidents related to weather condition, crossing and turning, pedestrians etc are either not possible in the tunnel or much less frequent in a tunnel than outside. An important design issue with impact on the occurrence of accidents is the number of tubes. By arranging the tunnel with unidirectional tubes, the risk of frontal collision is prevented, however regarded as a safety measure the construction of an additional tunnel tube is expensive and normally the number of lanes and the arrangement of the road is the same in the tunnel as it is outside. The geometrical measures also include suitable design of alignment, longitudinal profile, cross section and provision of lay-bys etc. Other risk prevention measures in road tunnels are the speed limits, restrictions for overtaking and requirements for minimum distance between vehicles, which can effectively reduce the frequency of accidents and thereby also the risk of fires. Systems with information and warnings to the road users about and variable speed limits adapted to the actual conditions. also reduce the risk of accidents and the manned control centres and their actions to manage the traffic and inform the users are also part of accident prevention. Development. The measures, which can be taken to prevent a fire starting with an accident from developing into a serious fire, are the same as for the

D:\NPH_COWI_Activities\52604_FIT\CommonPart\FIT_CommonPart5_Rev.DOC 7 fire caused by electrical/ mechanical defects. However, the probability of the serious development is higher when the fire starts with an accident. Defects in tunnel equipment Source and probability. High standards for the equipment, installation quality and maintenance will be preventive measures with respect to defects in tunnel equipment. It is also a preventive measure to limit the equipment and power consumption in the tunnel to a minimum. Development. The installations shall be designed so that the probability of development from a defect to a fire and from a fire to a serious fire is as low as possible. Some materials should be avoided (for example PVC) and the fires should be confined to a limited area. Arson and other detrimental action Source and probability. Even though the risk of arson and similar cannot be completely prevented it can be relevant to study the security. An example of a preventive measure is the limited and controlled access to vital parts of the tunnel.

Preventive measures in rail tunnels Electrical/mechanical defects in the trains Source. The source of electrical or mechanical defects in the train are various parts of the traction unit or the wagons, for example the electrical system, combustion engine (for some trains), the brakes, axles, the heating system, and other facilities on the train for example the on-board kitchen. A substantial improvement has been obtained by introduction of design standards particularly for the passenger trains. Probability. Prevention of occurrence of fires is achieved by designing the critical component to a high standard and by high quality maintenance and control of the rolling stock. The probability of bringing a train with a defect, which can develop into a fire, into the tunnel is reduced by on- track control of the train, for example blocked brake, hot boxes, wheel defects, etc. Fire detectors can prevent the train from bringing a fire into the tunnel. The risk of having a fire inside the tunnel is also prevented by the "drive-out" concept, whereby the train is taken out of the tunnel if possible. The emergency brake neutralisation is part of this concept, which aims to maintain the movement capability. Development. The development of the initial fire to a serious fire is in passenger trains prevented by designing the trains with limited amounts of combustible materials. Risk of passengers can be prevented by operative measures, which segregate transport of goods trains and trains with dangerous goods from passenger trains. The risk of development of very serious fires can also be prevented by limitations to the type of goods and the quantity of goods transported on goods trains. Furthermore the design of the tunnel and the materials used in the tunnel can contribute to the prevention of development of serious fires. Also maintenance, control and cleaning procedures of the track and other parts of the tunnel can contribute to the prevention of fire development. Accidents

D:\NPH_COWI_Activities\52604_FIT\CommonPart\FIT_CommonPart5_Rev.DOC 8 Source. The accident rate of railways is very low. The sources of the few accidents are mainly human errors of the loco driver, the operational personnel and defects in rolling stock, tracks and equipment. The sources of accidents have to a large degree been reduced by an increased level of automatic control systems (for example ATC and similar systems). Probability. The prevention of occurrence of severe accidents and fires is achieved by a number of measures in the infrastructure, among others the lay-out of the tunnel in single track tunnels, arrangement of tracks and switches, inspection of the track and tunnel conditions. Also the monitoring of train location and speed as well as the communication system, signalling, train control and traffic management systems are parts of the preventive measures. Development. The measures, which can be taken to prevent a fire starting with an accident from developing into a serious fire, are in principle the same as for the fire caused by electrical/ mechanical defects. Defects in tunnel equipment Source and probability. Rail tunnels generally require a significant amount of equipment. High standards for the equipment, installation quality and maintenance will be preventive measures with respect to defects in tunnel equipment. It is also a preventive measure to limit the equipment and power consumption in the tunnel to a minimum. Development The installations shall be designed so that the probability of development from a defect to a fire and from a fire to a serious fire is as low as possible. Some materials should be avoided and measures should be taken to confine the fire to a limited area. Arson and other detrimental action Source and probability. A relative large percentage of fires are caused by arson and similar actions including careless deposing of cigarettes etc. A preventive measure can be restriction of smoking in the trains and on the stations. On the other hand the wish of the arsonist to act harmful can hardly be influenced by regulators, designers and operators. In theory control of passengers could be carried out similarly to the control in airports, but it is hardly realistic for train traffic. Access to vital parts of the tunnel shall be limited and controlled. Development. The development from the act of arson (or similar) to a serious fire can be prevented by monitoring, detection and control. Both automatic systems and supervision by train crew and personnel at the stations can contribute to this prevention.

Preventive measures in metro tunnels Electrical/mechanical defects Source and probability. The source of electrical or mechanical defects in metros are various parts of the rolling stock for example the electrical supply system, the engine, the brakes, axles, the heating system, and other facilities in the rolling stock. A substantial improvement has been obtained by introduction of strict design standards for metros. Development. Metros transport only people and the density of passengers is often very high. For this reason any fire may be regarded as serious. The probability of a development from the first ignition to a fully developed fire endangering structure and users can be influenced by the

D:\NPH_COWI_Activities\52604_FIT\CommonPart\FIT_CommonPart5_Rev.DOC 9 design of the compartments. The construction material of the tunnel shall be chosen so that it does not contribute to the fire load. Accidents Source, probability and development. The traffic control and management systems of metros are very strict and the accident rate of metros is low. The preventive measures are generally the same as for rail tunnels. Defects in tunnel equipment Source and probability and development. Metros generally require a significant amount of equipment and serious accidents have occurred due to defects in the equipment. The preventive measures count strict standards for equipment, installation and maintenance and are in general the same as for rail tunnels. Arson and other detrimental action Source, probability and development. Unfortunately a number of serious fires caused by arson, sabotage and carelessness have occurs in metros. Similarly to rail tunnels the preventive measure which can be taken are monitoring, detection and control in the metro trains and at the stations. In addition the access to non-public part of the metro should be restricted.

Preventive measures and the scope of FIT The major part of the activities in the European Thematic Network FIT (Fires in Tunnels) has been concerned with risk mitigation, i.e. the activities, which reduce the consequences, and on emergency response. Prevention of fires and prevention of risk in general are included as part of the discussions, but the systematic evaluation is mainly related to mitigation and response. For further discussion of preventive measures reference is made to for example DARTS, UPTUN and especially the SAFE-T network.

1.2.4 Mitigation Mitigation measures are safety measures, which aim to limit the consequences once the ignition has taken place and developed into a fire. The mitigation measure may be related to: • Reduction of the fire development • Reduction of consequences to structure and equipment • Reduction of the consequences to humans • Fire suppression can also be regarded as a mitigation measure, but this aspect is discussed separately below

The mitigation safety measures will have to be decided based on a study of the effect of the measures and their disadvantages and costs. It should be noted that even though mitigation measures often are directed to reduce one specific consequence, they may have positive or negative side effects for other wanted or unwanted events. A systematic study the risk reducing effect of the mitigation measures must take into account the characteristics of the specific tunnel and the possible fire scenarios. The study can be carried out for example in terms of so-called "Event tree analyses", see also section 1.3.3.

D:\NPH_COWI_Activities\52604_FIT\CommonPart\FIT_CommonPart5_Rev.DOC 10 Some measures to mitigate fire risk in tunnels are presented briefly in the following. For a detailed discussion of the safety measures and their application in road- rail- and metro tunnels reference is made to the chapter X3X Fire Safe Design. For discussion of the possible fire scenarios reference is made to chapter Y2Y. Furthermore, some of the safety measures are presented in detail in FIT database 4 "Safety Equipment in Tunnels"

Reduction of fire development Ventilation. The ventilation can to some degree influence and reduce the development of the fire. However, its main function during a fire is to control the smoke. Drainage. During or before a fire flammable liquids may be leaked. A suitable drainage system for flammable liquids reduces the quantity of flammable liquids from the seat of the fire and thereby mitigates a serious development of the fire. Reduction of consequences to structure and equipment Structural fire resistance and fire protection. One of the serious consequences of a fire is damage to the tunnel structure and ultimately collapse of the structure. By suitable design of the tunnel structure and by passive fire protection the tunnel can withstand the relevant fire scenario and the reinstatement costs can be reduced. Fire resistance of other installations, equipment, cabling. The installations are often the first part of the tunnel system to suffer damage in a fire. By suitable design and passive or active fire protection systems the damage to the installations can be reduced. The cabling and other installations should as far as possible be located so that they will not be affected by a fire, and should be design in compartments so the fire does not spread by means of cable ducts and similar. Fire resistance of safety systems. It is particularly important that the function of the installations necessary during an emergency is also ensured for a suitable duration at the conditions prevailing during a fire. Reduction of consequences to humans Tunnel lay-out. The highest priority is to mitigate consequences to humans. The geometrical lay out of the tunnel can contribute to the mitigation. For example, it is less complicated to ensure the majority of tunnel users with smoke free conditions, if the tunnel is operated in one-way traffic. Also the cross sectional area influence the chances of creating smoke free areas and giving conditions for escape from a fire. Location (spacing) of emergency exits. One of the most important mitigation measures for users exposed to a fire in a tunnel is the provision of escape ways. The safety will be influenced by the spacing and design of the emergency exits. Ventilation smoke control, Another crucial safety measure is the ventilations system, which in case of a fire shall be able to create smoke free escape ways. The ventilation system shall be able to control the smoke spread for example by blowing the smoke in one direction. Ventilation, smoke extraction. In addition to the normal ventilation, smoke extraction is a mitigation measure, which can reduce the risk to humans in the tunnel.

D:\NPH_COWI_Activities\52604_FIT\CommonPart\FIT_CommonPart5_Rev.DOC 11 Communication systems, messages and alarms. Communication systems are mitigation systems because the can influence a timely evacuation of the tunnel in case of a fire and thereby reduce the number of persons exposed to the fire. The communication systems can be radio contact message boards, load speakers, alarms or similar. Exit signs, exit route guidance and markings. Once the evacuation has been initiated it is important that the tunnel users reach the safe area as quickly as possible. Exit signs, exit route guidance and markings are mitigation measures, which can make the escape more efficient. Facilities for assisted rescue. In some cases the rescue and evacuation will have to be assisted by the rescue forces, tunnel operators, etc. Facilities in the tunnel can make the assisted rescue more efficient. For example the emergency exits can serve as access routes for the rescue forces.

1.2.5 Suppression The suppression measures are taken to fight a fire. The suppression measures can be part of an installation in the tunnel or it can be the response from the tunnel users, the operators and the emergency services.

More details about fire fighting and emergency response is discussed in chapter Z4Z, and the fixed installations are discussed in chapter Y2Y. The different technical solutions are included in the FIT database 4. In the following the different categories of suppressions measures are briefly presented.

Fixed installations. It both possible to have fixed installation in the traffic space in order to suppress fires in vehicles and also in equipment rooms and similar to suppress fires in installations. It is furthermore a possibility to have fixes fire suppression systems on board the rolling stock and in the vehicles. Often the aim of the fixed suppression installations is not to extinguish the fires but to control and limit the development. Different types of fixed systems can be relevant. The following examples are included in FIT database 4: • Sprinkler • Water spray systems • Water mist systems • Impulse fire extinguishing systems • Foam extinguishing systems User operated systems. The best chances of a successful fire fighting are in the initial phase of a fire therefore systems operated by the users may have a good chance of success. Fire suppression installations for the users should be easy to use as the tunnel users most likely will be unfamiliar with fire fighting and with the tunnel environment. Different types of user operated suppression systems can be relevant. The following examples are included in FIT database 4: • Fire Hose installed in cabinet in the tunnel

D:\NPH_COWI_Activities\52604_FIT\CommonPart\FIT_CommonPart5_Rev.DOC 12 • Extinguishers (powder or foam) installed in cabinet in the tunnel or in the passenger compartment Tunnel operator. The tunnel operator may be active in the fire suppression by implementing procedures, operating the technical equipment, and call in the emergency services. Tunnel operator staff may also be active in the first aid fire fighting. Emergency services The professional fire fighting is provided by the emergency services, but they face an unpredictable and difficult situation on their arrival at the fire site. Par- ticularly if the fire has already developed into a severe fire. Some of the aspects, which are important for the success of fire suppression by the emergency services, are mentioned below. More details can be found in chapter Z4Z. • Organisation of fire fighting, including the location of the fire brigades, • Equipment for fire fighting in the tunnel: water supply by fixed installation, for example pressure pipes water main, or dry pipes and mobile units • Emergency and rescue planning should be carried out envisaging all relevant scenarios. • Exercises can be a measure to achieve understanding of the tunnel details and knowledge of tunnel operational possibilities and hereby obtain a higher probability of successful fire fighting. • Means of communication for the rescue forces will improve the safety and success rate of the fire fighting. • Availability of necessary information for the rescue services when they arrive at the scene is important for their safety and for the success rate of the fire fighting.

1.2.6 Reopening The reopening aspects deal with post-accident situations. When the fire has been extinguished or has died out the consequences has to be assessed. Forensic investigation may have to be carried out by the Police and other parties to establish the legal facts of the accidents. Before the tunnel can reopen the site of the fire will at least have to be cleared.

Also engineering investigations may have to be carried out in order to assess the damage on the structures and the installations and to establish the actions necessary for the reopening of the tunnel.

The minimum criteria for operation and thereby for reopening should be established in terms of functionality of the installations and the degree of damage of the structure.

In order to achieve the shortest possible closing of the tunnel the post-accident activities should be planned beforehand based on scenarios. As part of the post- accident response the investigation methods can be planned. It may be possible to have spare equipment on stock. Various repair projects and temporary as well as permanent strengthening can be designed and prepared depending on the degree of the damage.

D:\NPH_COWI_Activities\52604_FIT\CommonPart\FIT_CommonPart5_Rev.DOC 13 Inspiration for the post accident planning can be taken in the upgrade activities, which have taken place in other tunnels. FIT database 6 contains information about a number of upgrade activities related to fire behaviour.

1.2.7 Evaluation A general evaluation of the fire event will have to take place. The accident may have revealed shortcoming in the actual tunnel system or in the understanding of the fires and the behaviour of material, structure, safety systems etc.

Also "near misses", i.e. events and incidents, which could have developed into severe fires, can be evaluated and lessons can be learned.

Information from previous fire accidents has been collected as part of the activities in the FIT network and can be found in FIT database 5. Reference is also made to the UPTUN project, where one of the activities is to study and learn from actual fires.

The general post-accident evaluation can be seen as a pro-active measure for the safe operation of tunnels. Thereby the consecutive safety aspects are closed into a circle.

1.3 Integrated approach to safety in tunnels Safety is a result of integration of the infrastructural measures, the operation of the tunnel, the user behaviour as well as the preparedness and incident management. The assessment of fire safety in tunnels is a complex issue, where broad multi-disciplinary know-how and application of different physical models are necessary in order to explore the causes of fire, the consequences and developments of fires and evaluate measures to prevent fire and reduce fire consequences.

Some safety measures for fire might be in conflict with the general safety and vice versa. For example reducing the fire risk but at the same time increasing the general risk. Introduction of safety measures based on their effect on fire risk and neglecting the impact on the general risk will in these cases tend to give sub-optimal solutions with too high costs and too high risk. In other cases, the safety measures will have a positive influence on general safety. Neglecting general safety in these cases will consequently lead to underestimation of the cost-efficiency of these safety measures. This may lead to a sub-optimal solution and too high risk.

It can be concluded that fire risk is an integral part of the general risk. Only for the cases where the safety measure only influence fire safety, these measures can be studied and decided on isolated from the general risk.

Hence, in general an overview of the entire complex is necessary in order to appreciate the best possible actions.

D:\NPH_COWI_Activities\52604_FIT\CommonPart\FIT_CommonPart5_Rev.DOC 14 The complex to take into account comprises: • The occurrences and physics of the fire (see chapter Y2Y) • The tunnel system, i.e. - Infra structure - Operation • Human behaviour of users, operators, emergency forces (see section 1.4) • Other factors influencing the safety

In the following the interaction between design options, the modelling of physical behaviour and support for the decisions are presented.

1.3.1 The tunnel system 1.3.1.1 General The decisions related to design and operation of tunnels should generally seek the best possible solution with respect to the objectives, and at the same time respect the relevant constraints, which may be of legal, economical and organisational nature.

In order to achieve the best possible solution all relevant measures should be identified and evaluated. An integral evaluation of the measures should take into account all benefits and disadvantages of the options.

Various institutions and persons may have different stakes in the decision. It is common at least to distinguish the points of view of society, the users and in some cases the tunnel as an economic unit.

In order to support the decisions proper studies with state-of-the-art techniques should be carried out. All relevant hazards should be identified; the effects of the hazards on the users, structure, economy, etc. should be estimated in terms of injuries, loss of life, structural damage, income losses, etc. The physical models to be used for these studies include for example models for the fire: combustion, smoke development, smoke spread, heat development, temperature development, fire spread etc, and models for the material subjected to fire: thermal stresses, damage, strength reduction, deformations, spalling etc.

1.3.1.2 Infrastructure The infrastructural measures cover structural components, ventilation and elec- tro-mechanical equipment as discussed briefly in section 1.2.3 and 1.2.4.

Interaction between safety, durability and other design goals The design process has different goals, which in some cases may be conflicting. The best possible solution is sought not only with respect to safety but also for durability, construction costs, environmental impact etc.

The best possible solution depends on the conditions for each specific tunnel. Furthermore, the infrastructure solutions should be evaluated in relation to its reliability, the possibilities of construction or installation, and the future process of operation and maintenance.

D:\NPH_COWI_Activities\52604_FIT\CommonPart\FIT_CommonPart5_Rev.DOC 15 Examples of safety measures with conflicting impacts could be: fire protection of the structure, which may prevent inspection and result in higher maintenance costs, various safety systems, which may have high running costs and the maintenance may disturb the traffic.

Integration of safety into other design and operational criteria has been dealt with in the research project DARTS "Durable And Reliable Tunnel Structures". The objective of DARTS has been to develop operational methods and support- ing practical tools for the best pro-active decision-making process. The aim is the optimal tunnel design and construction procedures regarding environmental conditions, technical qualities, safety precautions and long service life. The approach is based on a minimum total life cycle cost including operation and maintenance, and to optimise safety and reliability, to create the best environment and safety for users and best benefit for society and owner.

New Designs and existing tunnels For new designs a large variety of design solutions for geometrical lay out, operational concepts and installations are open. On the other hand no site-specific experience with the traffic, the operation and the tunnel has been achieved. The design decisions are based on models and assumptions for the conditions at future stages of the project.

For existing tunnels the possible actions to improve safety are limited. Some fundamental parts of the design can hardly be changed or the changes are associated with very high costs (for example the geometry of the tunnel). On the other hand the decisions can benefit from on-site experience, which can be taken into account. As it has been illustrated in a number of tunnels in the recent years, a significant improvement of the safety of existing tunnels can be achieved by upgrading of the safety installations. Reference is also made to the UPTUN project, which is dealing specifically with this problem.

1.3.1.3 Operation Influence of operation on tunnel safety The operational safety measures comprise accident / incident management, communication with tunnel users, tunnel closure, control of traffic, detection of traffic conditions, traffic management and control of transport of dangerous goods. The performance of installations (e.g. ventilation and lighting) is moni- tored and adjusted to the conditions also in case of an accident or a fire. The reliability of the safety systems is expected to be improved in a well-maintained tunnel. Good preparation of maintenance activities can have a positive effect as well.

The safety can be improved by restrictions to the traffic, i.e. by limitation of the traffic, and the speed, requiring distance between vehicles and restricting substances transported in the tunnel. These measures may have limited costs for the tunnel operation but may give costs and disadvantages for the users and the society in general. An integral approach to the safety should take into account also the external costs.

D:\NPH_COWI_Activities\52604_FIT\CommonPart\FIT_CommonPart5_Rev.DOC 16 The safety in the tunnels may also be improved by a better safety culture in the operation of the tunnel, i.e. awareness in the organisation that safety is important. The alertness and the knowledge about the correct reaction can be maintained and improved by instructions, training and fire drills.

Time and urgency of decisions. Improved fire safety is evaluated during design as well as in the operation of tunnels in situation where the design options can be evaluated without urgency. In this situation any given complexity can be taken into account in the evaluation. Other actions and decisions will have to be taken instantaneously e.g. the actions by the operator at the time of a fire alarm.

Therefore the amount of information given to operator during normal traffic and in case of incidents should be sufficient but not too complex or too redun- dant. The set-up can be studied based on scenarios or as part of exercises.

1.3.1.4 Other factors influencing the safety The design of the vehicles is another factor, which can significantly influence the safety in the tunnels. For rail tunnels and metros a significant improvement of the fire safety has been achieved by more strict design criteria for the rolling stock. For rail and especially metro tunnels the tunnel owner can control which types of rolling stock is using the tunnel and can take this into account in the design concept.

For road tunnels the designer and operator generally cannot control the type of vehicles using the tunnel, but also here an improvement of the safety can be achieved by vehicles design and equipment. The vehicle producers may contribute to improved tunnel safety by designing the vehicles to be less ignitable and reduce the fire load by the choice of materials in the vehicle. HGVs may be equipped with smoke detectors and video monitoring of the freight. Automatic extinguishers in the motor compartment may also be considered, and video surveillance of the goods from the drivers cabin can also improve the safety.

1.3.2 Prescriptive vs. performance-based approach Prescriptive approach Traditionally fire safety standards for tunnels and other structures have been prescriptive; they have contained minimum requirements, which must be fulfilled. These requirements have been established during years based on experience, tradition, and engineering/expert judgement. They apply in principle an absolute evaluation of safety: if the design is in accordance with the standard, the safety is acceptable; if not it is absolutely unacceptable. The advantage of a prescriptive standard is its uncomplicated use and the ensured minimum level of equipment etc in the tunnels. On the other hand prescriptive standards may not be able to handle unusual situations and may in some case not be able to reasonably take into account the interaction between the different parts of tunnel structure, installations and the local conditions.

D:\NPH_COWI_Activities\52604_FIT\CommonPart\FIT_CommonPart5_Rev.DOC 17 Performance bases approach In recent years fire safety engineering or performance based approach has found application to buildings and other structures. By application of fire performance concepts, fire safety is achieved based on a scientific appreciation of the fire phenomena, of the effects of fire and of the reaction and behaviour of people. Emphasis is given to the safety of life, but fire safety engineering can also be used to assess property loss, interruption of service, contamination of the environment etc. Furthermore, risk of fire and its effects are quantified and the optimum safety measures are evaluated.

The form of a performance based regulatory system can be presented as shown in the hierarchy in Figure 1.2. The goals objectives, functional statements and performance requirements are discussed and evaluated qualitatively. Based on the performance level, including the qualitatively formulated risk level quantitative criteria are established and applied in the verification of the system.

Figure 1.2 Hierarchy - objectives for fire engineering design By a performance-based approach, the regulatory requirements are given on a more general level specifying the safety of the users, economic values etc.

The fire safety engineering will normally involve the following steps: • Qualitative design review - Definition of objectives and safety criteria, with reference to performance based regulatory requirements or possibly in consultation with approval authorities. - Definition of the tunnel system - Identification of fire hazards - Selection and definition of fire scenarios - Identify design options - Identify methods of analysis • Quantitative analysis of design using the appropriate subsystems: - Initiation and development of fire and generation of heat and smoke - Spread of fire, heat and smoke - Structural response to fire - Detection, activation and suppression - Behaviour of tunnel users and the influence of the fire on life safety

D:\NPH_COWI_Activities\52604_FIT\CommonPart\FIT_CommonPart5_Rev.DOC 18 • Assessment of the outcome of the analysis and evaluation against criteria

The objectives and the associated acceptance criteria used in a performance based shall be clearly defined and established to the particular design under consideration. The acceptance criteria established the adequacy of the design can be according to the following approaches: • Deterministic (including, when appropriate, safety factors) • Probabilistic (risk based) • Comparative (comparison of performance with accepted codes of practice)

The deterministic and the comparative approaches are not far from the prescriptive approach used so far. The risk based probabilistic criteria can be formulated as evaluation of probability of harm or evaluation of an expected value of harm as discussed in section 1.3.3.

Discussion It could be relevant to specify that all tunnels should have the same safety level relating to the classification of the infrastructure.

However, from a user’s point of view it would be desirable if all tunnels had the same lay-out, equipment, escape facilities, emergency procedures etc. How- ever, in practice such homogeneity towards the users is not possible or is very expensive because the tunnels are built at different times and under different natural environments. If for instance the distance between the cross passages should be the same short distance for bored tunnels as it can be for immersed and cut and cover tunnels, the costs would be very high and in disproportion with the risk reducing effect. The safety standard will therefore have to respect the actual conditions.

A performance based approach taking into account the local conditions may give solutions, which deviate significantly from structure to structure. It may, for example, result in very little safety measures in tunnels with very low traffic volume. In this way the user may not find the equipment he would expect in a tunnel. For example the distance between the cross passages will generally turn out to be longer in a bored tunnel than in a cut&cover or immersed tunnel.

For tunnels a combination of minimum prescriptive requirements and performance based objectives and acceptance criteria would be relevant. The performance-based criteria need to be established and should be unified on national and European level for the various types of infrastructure. The performance-based criteria should be formulated so that it is possible without ambiguity to establish whether the criteria are met and the reporting and documentation should be clear and transparent.

For the designer a prescriptive approach is with respect to liability an advantage. If the designer has fulfilled the requirements given in the codes, he is in most countries free of liability. With application of a performance-based approach some degree of subjectivity may be included in the risk-based approach and the question of liability will have to be specified in the actual case.

D:\NPH_COWI_Activities\52604_FIT\CommonPart\FIT_CommonPart5_Rev.DOC 19 The overall goal of fire safety engineering is the optimum safety design For the tunnel owner and the designer a performance-based approach gives the possibility of adaptation to the local conditions and to take into account all parts of the tunnel system. More detailed studies may be necessary and it may be relevant to establish criteria for the quality of the studies. In totality it must be expected that a design using a performance based approach will be more expensive.

The use of performance-based design is a promising approach also for tunnels design regulatory initiatives will have to be taken in order to establish the framework and to solve the remaining problems.

1.3.3 Risk analysis Purpose Risk analysis of tunnels can have purposes, which can be grouped into the following areas: A. The result of risk analyses can demonstrate and document a sufficient safety level to public authorities or it can be part of internal documentation to demonstrate compliance with internal corporate policies. B. The risk assessment can serve as basis for risk communication to the public, to users, nearby inhabitants or to other stakeholders. C. The risk assessment can further be a basis for decision-making. The overall risk of undesirable events can be integrated in the planning, design and operation of a tunnel. This includes decisions concerning tunnel layout, structural decisions, preventing and mitigating safety measures, traffical and operational procedures and emergency planning. The information can help to balance risks against costs associated with risk reducing measures, both within certain constraints.

Risk analyses can be used both in tunnel projects where a prescriptive design approach is used and as part of the performance-based approach (point C above).

Process Generally, risk analyses are carried out using the process illustrated in Figure 1.3. The process includes the following: • Identification of hazards, events and scenarios: what can happen or what can go wrong and what are the causes • Estimation of probability: what is the probability that the unwanted events occur • Estimation of consequence: what are the consequences in case the unwanted event occurs • Assessment of risk: comparison and evaluation of the estimated risk in relation to safety objectives and to the possibility of introducing risk reducing measures.

D:\NPH_COWI_Activities\52604_FIT\CommonPart\FIT_CommonPart5_Rev.DOC 20 Acceptance Planning Criteria System definition

Hazard identification

Frequency / Consequence Risk Probability Analysis reducing Analysis measures Risk estimation Risk picture Risk anallysis Risk evaluation Riisk assessment Additional risk reducing measures Part of safety managementt and risk controll Figure 1.3. Risk Management Process Methods Proper studies address with state-of-the-art techniques, the potential threats and hazards, the effects of the threats on the users, structure, economy, etc. and the damage or results of these effects in terms of injuries and loss of life, structural damage, income losses, etc.

Most commonly used techniques to study the scenarios and the combinations of probability and consequence are: • Fault tree analysis: the evaluation of the so-called undesired or top event by a top-down analysis of identification of combinations of causes leading to the undesired event • Event tree analysis: a logical diagram of success and failure combinations of events, leading to all possible consequences of a given initiating event • Most credible accident identification: identification and study into the effects and consequences of the most likely accident, with a focus on the optimisation of the mitigation by emergency response teams • Worst case scenario identification: identification of an extreme event, which may have a very low or event remote probability of occurrence.

For illustration a simplified fault tree and a simplified event tree are shown in Figure 1.4 and Figure 1.5.

The structure of the analysis and the logical trees will have to be detailed to a level where all relevant characteristics are taken into account. By means of fault trees the causes of fires can be studied and the measures for prevention can be identified.

D:\NPH_COWI_Activities\52604_FIT\CommonPart\FIT_CommonPart5_Rev.DOC 21 Figure 1.4 Simplified fault tree for illustration, indicating logical combinations (and / or gates) of events leading to the top event "Fire in tunnel".

Figure 1.5 Part of a (simplified) event tree (for illustration only!), exploring the consequences of a fire. PV Personal car, HGV Heavy Goods Vehicle Risk evaluation The quantitative evaluation of risk can be carried out by: • Comparison of probability of risk with upper limit acceptance criteria • Evaluation of the risk in relation to the cost benefit of all relevant design options possible for reduction of the risk. For this approach weights of all consequences will have to be established.

D:\NPH_COWI_Activities\52604_FIT\CommonPart\FIT_CommonPart5_Rev.DOC 22 • A combination of an absolute upper limit and evaluation of all relevant design options possible for reduction of the risk

It is common to evaluate the risk by using a combination of acceptance limits and cost-benefit considerations. The acceptance limit ensures that the risk to people does not exceed a certain level and in the area below the risk is reduced depending on the advantages and disadvantages of risk reduction. This evaluation of the risk can follow the so-called ALARP (As Low As Reasonably Prac- ticable) principle, which is illustrated in Figure 1.6 below.

Unacceptable High risk Risk is intolerable and cannot be justi- region fied even in extraordinary circumstances Tolerable only if risk reduction is im- practicable or if its cost is grossly in dis- ALARP proportion to the improvement gained region

Tolerable if cost of reduction would exceed the improvements gained Broadly No need for detailed studies. Check that acceptable risk maintains at this level Negligible risk region Figure 1.6 ALARP region and upper limit

1.3.4 Cost versus safety General Safety measures are associated with direct and/or indirect costs and any decision of safety measures will include considerations of costs and other disadvantages of the measure and considerations of the risk reducing effect.

Prescribed safety measures The regulation may prescribe specific safety measures for all tunnels or for tunnels in different classes. The regulator has in this case evaluated that the measures in average are sufficiently effective to reduce the risk and has evaluated that the costs and disadvantages of the measures are reasonable.

Directly or indirectly in the prescription of the measures is an evaluation of the average effect of the measure and the average costs. Inherently this can be expressed as a valuation of safety and can be expressed as a general valuation of fatalities.

Prescribed safety measures will in most case not give a homogeneous valuation of safety for all measures.

Prescribed safety level A prescribed safety level can be regarded as boundary conditions on the allow- able design. It is in principle possible to find the valuation of fatalities inherent

D:\NPH_COWI_Activities\52604_FIT\CommonPart\FIT_CommonPart5_Rev.DOC 23 in a prescribed safety level, but it depends on the costs of all possible risk reducing measures in average for all tunnels covered by the regulation.

A prescribed level will ensure that the users are exposed to a defined maximum of risk for each tunnel. However, the prescribed safety level does not ensure that the risk is as low as reasonably practicable in all tunnels.

Direct weighting In order to ensure that the risk is as low as reasonably practicable in all tunnels, the risk must be evaluated and compared with all possible risk reducing measures in the individual case. The evaluation is an evaluation similar to a cost benefit analysis and requires that the benefits and disadvantages are brought into one scale for comparison.

The most common used scale for comparison of safety measures and their effects and disadvantages is a monetary scale. Using a monetary scale requires that for instance the weight given to fatalities must be expressed as "costs".

Associating lives with costs is always controversial, however it ought not to be more controversial to explicitly formulate the "value of lives" compared to an inherent valuation.

A common framework for the valuation of fatalities etc. should be found. Some countries have established capitalised costs to be used for various evaluations of safety measures. For evaluation of safety in road traffic the public authorities in European Countries like Denmark, Finland, Germany, Norway, Sweden and UK have used values between 1 and 2 million Euros per fatality.

As discussed in section 1.3.3, the direct weighting can be combined with prescribed safety levels for example as formulated in the ALARP principle.

1.3.5 Dangerous goods transport The fire design depends on decisions made with respect to the operational conditions of the tunnel. It is particularly important to discuss whether transport of dangerous goods should be allowed or restricted in the tunnel.

Road tunnels A serious accident involving fire in dangerous goods can be particularly costly in terms of human lives, tunnel damage, transport disruption and the environment. On the other hand needlessly banning dangerous goods from tunnels may create unjustified economic costs and may force the transport operators to use more dangerous routes. International treaties have been signed for transport of dangerous goods. For transport on roads the ADR and for railways RID define and regulate the transport. However, the rules and regulations for transport of dangerous goods in tunnels vary significantly from country to country. For this reason OECD and PIARC established a study with the purpose of harmonising the regulations of transport of dangerous goods through road tunnels. As part of the study dangerous goods was split into 5 groupings and a Quantitative Risk

D:\NPH_COWI_Activities\52604_FIT\CommonPart\FIT_CommonPart5_Rev.DOC 24 Assessment Model (QRAM) and a Decision Support Model (DSM) were established. The following policy recommendation was a main result of the study:

It is strongly recommended that administrations, which allow the transport of dangerous goods through road tunnels, implement the "groupings of dangerous loadings" system as a basis of regulations. This system should be implemented at both the national and international levels. … The adoption of the "groupings" system requires a systematic and scientific basis for decision-making. To this end, the QRAM and the DSM developed as part of the project is currently the state of the art in the field and are recommended for use in all countries to support the adoption of the proposed groupings system. Quote from "OECD/PIARC Safety in Tunnels/ Transport of dangerous goods through road tunnels".

The methodology used in the OECD/PIARC study is well in accordance with the approach mentioned in the above sections. The decision support model is based on mutual weighting of the various objectives.

Rail tunnels Railways are often considered a suitable means of transporting dangerous goods due to the low frequency of accidents on railways. The considerations related to the transport through rail tunnels are different from the case of road tunnels discussed above. For the rail transport there is in the majority of cases no alternative route by-passing the tunnel. The question for the rail tunnels is whether the safety transport through the tunnels is acceptable.

By segregation of the trains with dangerous goods from passenger trains the hazard of human casualties can be significantly reduced. The segregation is carried out by the train schedule so that trains with dangerous goods do not meet passenger trains in a twin track tunnel and does not enter a single-track tunnel before passenger trains have left the tunnel and vise versa.

The scheduled segregation can furthermore by followed up by actual control of the location of the trains.

1.4 Human behaviour A tunnel system is not purely a technical system; human behaviour highly in- fluences both the occurrence of unwanted events, the development of these events and the success of intervention. It is relevant to distinguish 3 groups of humans: • Users: - Road users - Train and metro users • Operators - Control room staff - Train staff (for rail tunnels) • Emergency staff

D:\NPH_COWI_Activities\52604_FIT\CommonPart\FIT_CommonPart5_Rev.DOC 25 The human behaviour can be distinguished based on the situation: • Routine behaviour before the incident and prevention of incidents • Behaviour causing unwanted events, accidents and fires • Human reaction to fire • Rescue and evacuation • Fire fighting

For a more general description and discussion of the physical and psychological models and the actions, which can be taken, reference is made to the UPTUN project, which has one work package WP3 dedicated to this topic.

1.4.1 Road Users Routine behaviour Technical solutions in a tunnel imply a set of rules that are important for the behaviour of tunnel operators and users. From the user perspective the rules are instructions that reduce the possible variability of the user behaviour. The rules define obligatory, permissible and “forbidden” behavioural repertoires.

The tunnel infrastructure, the interpretation of the rules of tunnel use, and the behaviour of other participants in the tunnel are the main sources of stimuli that determine user’s behaviour.

Each tunnel user has a set of goals, which are pursued in the traffic situation. During driving the user continually diagnoses the traffic situation, predicts the possible outcomes and decides about corrective actions that fit to the perceived situation.

Behaviour causing unwanted events, accidents and fires

Causes associated to human action

Direct Indirect

Improper Physical lookout High speed Physiological

Inattention Alcohol

Erroneous Mental / assumption emotional Improper Experience / manoeuvre exposure Internal distraction

Figure 1.7 Direct and indirect causes of accidents related to road users’ behaviour The cause of the accidents is in the majority of cases associated to human action of the drivers. The direct causes comprise e.g. high speed, inattention, er-

D:\NPH_COWI_Activities\52604_FIT\CommonPart\FIT_CommonPart5_Rev.DOC 26 roneous assumption, improper lookout or internal distraction. The human action of the driver may be an indirect cause of the accident due to e.g. physical or physiological condition and not least alcohol, see Figure 1.7.

A large number of accidents and fires could be avoided if all motorists would drive observant with modest speed, sober, using the seat belts and respecting suitable long distance, listening to the traffic radio in well-maintained vehicles.

Human reaction to fire Fire is experienced by the users as a complex, rapidly changing event, which, in its early stages at least, is usually highly ambiguous, providing little positive information to act upon. A person confronted with a fire needs a lot of information in order to understand the situation fully and to decide what to do. The general model of sequences can be described as illustrated in Figure 1.8.

Receive information

INTERPRET Ignore Investigate

PREPARE Instruct Explore Withdraw

ACT Evacuate Fight Warn Wait

Figure 1.8 Sequences of human behaviour in fires. It has been observed in a number of emergency situations that people behave rationally in these cases. The apparent "incorrect" behaviour is a result of mis- conception of the situation and the safety can be improved by communication in the tunnel and general information campaigns

Evacuation The human behaviour related to evacuation can be modelled in 3 stages: • Stage 1 Becoming aware • Stage 2 Hesitation to leave the car • Stage 3 Walk to the exit

It has been tragically observed that motorists are very reluctant to leave their vehicles in case of tunnel fires. The users tend to underestimate the situation and shall be directly instructed to leave their cars. Only in case of a clearly hazardous situation will motorists leave their cars. In some cases of the recent severe tunnel fires, the consequences in terms of human lives could have been significantly reduced if the users had evacuated in time and in the right direction.

D:\NPH_COWI_Activities\52604_FIT\CommonPart\FIT_CommonPart5_Rev.DOC 27 As part of the UPTUN project, modelling and tests of the human behaviour in relation to evacuation has been carried out also including a quantification of the time used in the stages of evacuation.

1.4.2 Train and metro passengers Behaviour causing unwanted events, accidents and fires For train and metro passengers no complex actions are required. The behaviour of train and metro passengers, which may cause accidents and fires, are limited to actions of carelessness, daring or sabotage. Particularly careless smoking and arson contributes to the risk.

Human reaction to fire For trains and metros the situation is more controlled than for road tunnels. The train staff has normally been educated in the correct behaviour in case of a fire and they will instruct the passengers in connection with fire in the tunnel.

The passengers are instructed by the staff and it is the experience that the passengers will follow these instructions. Therefore, it is essential that the instructions given are correct, because otherwise it can lead to critical situations for a large number of passengers. General information to the passengers about safety may be given in form of folders placed in the trains.

Passengers may react by activating the emergency brake, but the activation of the emergency brake is often not desirable in the tunnel and can be overruled by the loco driver.

Evacuation In case a train needs to be evacuated in the tunnel this may influence a large number of passengers and the design and capacity of the escape ways is important.

Normally nobody should or would be able to leave the train in the tunnel before instructions are given to evacuate. When the instructions have been given there will normally be no hesitation to leave the train. A group behaviour will govern and the time to leave the train and reach the exit is dependent on the door, walkways, lighting etc as well as the walking ability of the passengers.

Models have been established and tests and exercises have been carried out to quantify the evacuation behaviour.

1.4.3 Operators The human behaviour of operators of tunnels follows some general principles for operation of technical systems. Tunnel operator tasks are dependent on the various types of infrastructure and vary from tunnel to tunnel. Tasks may include: • Securing safety in normal conditions (prevention) and in the event of an incident;

D:\NPH_COWI_Activities\52604_FIT\CommonPart\FIT_CommonPart5_Rev.DOC 28 • Monitoring the efficient performance of all technical installations during normal operation and adjusting them as required during incidents; • Properly maintaining structural and electromechanical installations.

The cognitive task load model is used to describe the tasks and the mental load of operators in process control environments. In this model, tasks are catego- rized in four generic task types: • Situation assessment - Situation awareness: Continuous monitoring of traffic and critical parameters that represent the state of the tunnel and the traffic. - Disturbance assessment: The situation awareness information is used to assess incidents in the tunnel • Decision making and control - Decision making: When no standard procedure is available, the operator might have to make decisions to prevent escalation of incidents. - Direction and control: Intervention in the tunnel and participation in rescue operations if necessary. The operator is crucial in incidents, especially in the first stages, prior to the arrival of rescue services.

Unsafe acts can occur as unintended actions, where the basic error types are attentional slips (intrusion, omissions, reversal, mistiming) and memory fail- ures. Furthermore unsafe acts can be intended actions with rule-based mistakes (misapplication of rule, application of bad rule etc) or it can be violation.

Errors can be the result of both cognitive underload and cognitive overload. Safety will benefit from the right balance in the tasks of the operator.

In general the performance can be improved by the following: • Planning - Inventory of the tasks of the operator - Identification of bottlenecks, using the task lists, knowledge and experience - Interviews to check, complete and prioritise the bottlenecks - Identification and evaluation of solutions • Recruitment • Training and exercise • Personnel and organisation • Task support (procedures, guidelines and software) • Control room and interface design

1.4.4 Emergency Staff Tunnel emergencies requiring the action of the rescue teams occur infrequently and it can not be expected that the teams can obtain experience with the situation without training. Emergency planning will be necessary for efficient and successful intervention.

D:\NPH_COWI_Activities\52604_FIT\CommonPart\FIT_CommonPart5_Rev.DOC 29 There are two levels of emergency preparedness plans; the organisational plan, which remains constant whatever the emergency situation, and the response plan, which vary according to the emergency situation.

An emergency organisational plan includes five main tasks; assessment, prevention, preparedness, response and recovery.

Response management is divided into three levels: operational, tactical and strategic level. Incidents usually are managed at the operational level and only when the situation becomes critical management moves to the tactical and then to the strategic level.

In an emergency situation every actor involved must have a specific role. Per- sons in charge must have the proper training and experience. The overall person in charge needs to coordinate the leaders of the teams involved. Depending on the type of emergency, different people may take charge.

Leadership strategies are necessary for dealing with emergency situations. In- spiring confidence, calm, assessing the situation and handling difficult mo- ments is what makes a successful leader. Means and methods of communication are highly important during the emergency and have to be well planned beforehand.

Emergency preparedness needs to be tested regularly by involving both the operators and the emergency teams.

D:\NPH_COWI_Activities\52604_FIT\CommonPart\FIT_CommonPart5_Rev.DOC 30 Best Response guidelines

ABSTRACT This report presents an overview of the ability of the Fire and Rescue Service to deal with major fires in road tunnels and also a basic discussion for other types of tunnels. This is an area of common interest for both the Fire and Rescue Service, tunnel designers and tunnel owners, in order to create a common view of how rescue work should be interwoven with other aspects of a tunnel’s safety systems. The report describes how rescue work in connection with fires in road tunnels can be approached, and the problems and difficulties to be overcome in successful rescue work.

1. INTRODUCTION

Incorporating safety into the design of tunnels around the world involves very thorough investigation of the various safety features in terms of both design and construction, generally accompanied by extensive risk analyses. Work includes consideration and calculation of the various design rating events and their consequences, with very careful analyses of how persons caught in accidents are likely to behave. The knowledge derived from these investigations and analyses is used to design and determine the necessary capacities of safety systems in the tunnels. What, however, is generally missing from most analyses is a detailed description of the various pro-active and consequence-reducing actions to be taken once an accident has occurred, which are normally performed by the Fire and Rescue Service. The Fire and Rescue Service becomes actively involved in the event of an accident, which can and should be seen as an integral part of the tunnel’s overall safety concept. However, as opposed to most other aspects of the safety concept, which have been very thoroughly analysed, this particular part of the safety concept generally seems to have been very poorly analysed, with very few analysis results officially available. It seems as if those responsible for the safety aspects of the design of modern tunnels simply assume that the Fire and Rescue Service can safely, quickly and efficiently deal with any accidents that occur. However, it also seems as if the public rescue and fire service is not reacting or acting by notifying those responsible for the safety of a tunnel that they (the tunnel owners/operators) are assuming the provision of Fire and Rescue Services for a duty in which the service has had no hand in determining the necessary capacity. The rescue services are being asked to provide a service for which they do not have the necessary capacity, for which they do not have the necessary working methods and/or for which they do not have the necessary personnel or equipment resources. It seems, in other words, as if there is simply not sufficient communication between those responsible for the safety design of a tunnel and the public rescue and fire services. This lack of communication (or even of understanding?) between the various parties involved in the design of safety aspects of Swedish tunnels can be partly due to the fact that Fire and Rescue Authority does in fact have two roles to play. One role is to be that of a public authority, ensuring that the design of the tunnel fulfils the guidelines and various forms of practice that ensure a sufficient level of safety for individual users. The other role is that it is also the organisation that maintains and applies the various actions and responses in the event of accidents. I do not believe that it is made sufficiently clear that Fire and Rescue Authority 1(21) Anders Bergqvist, Stockholm Fire Brigade actually have these two roles, which means that both the rescue and fire services and the tunnel owners misunderstand each other. Tunnel owners have a major responsibility in terms of preparedness for accidents, which includes making it possible to perform rescue work to save lives and to limit damage arising from an accident. A study of the present legislation1 can show2 that a property operator or property owner has a relatively substantial duty to ensure that conditions are such that rescue work could be carried out in the event of an accident, and particularly if the property is large and complicated, such as a tunnel. By working closely with the Fire and Rescue Authority, a tunnel owner can arrive at a solution that integrates the Fire and Rescue Service with other safety aspects of the tunnel. The material considered in this report is based on Swedish conditions.

2. A DESCRIPTION OF THE CONCEPT OF TACTICS FOR RESCUE WORK In simple terms, fire and rescue operations can be said to consist of a number of different elements: 1. The working methods or active measures for which the personnel have been trained and equipped, with the results of the various operation methods depending largely on the ability of the personnel to make best use of the equipment under the conditions of the particular accident with which they are dealing. 2. Coordination of the various individual methods, so that they work together to produce an effective whole, the rescue action. 3. Selection of tactics for effectively fulfilling the objective of the rescue action.

Tactics in connection with fire and rescue operations have been, and are, thoroughly discussed, with Svensson3 providing the terminology used in this section. Most of the documents and reports produced in connection with this subject are relatively theoretical, and do not consider specific tactical patterns for how fire and rescue operations can or should be structured. The tactics in fire and rescue operations can be relatively simply described as being the ongoing decisions by the incident commander of the resources to be used, and actions to be carried out, in order to achieve the objectives of the action in the most efficient manner. The direction of the tactical thrust will be very dependent on the type of accident concerned, the local conditions, the consequences of the action that may have already occurred or which could occur, the objective of the work of the Fire and Rescue Services and the resources available for performing the work. The tactical execution of the work will be very dependent on the competence of the personnel available for the action. All these different elements combine to affect tactical performance of the fire and rescue operation. A fire and rescue operation dealing with a fire in a residential apartment should be carried out in a different way to that of tackling a fire in, for example, a barn. This depends not only on the particular objectives of each action, but also on the fire-fighting methods intended for different types of fires and different types of environments. From this, it should be relatively easy to see that fire and rescue operation in connection with a fire in a cable tunnel will be (or should be) carried out in a different manner to tackling a fire involving private cars in a tunnel. The biggest difference between these two can be seen to be the fact that the main thrust of the operation in the road tunnel would be to rescue persons involved, which is not the main purpose of operation

2(21) Anders Bergqvist, Stockholm Fire Brigade in the cable tunnel. There are also many other differences that should mean that fire-fighting operation is carried out in different ways.

Objective of the rescue action

Resources available for Tactical executtion of Results of the the rescue action the rescue action rescue action

The type of accident and the environment in which it has occurred

Figure 1 Tactic for rescue actions4

3. TUNNELS AND FIRES IN TUNNELS The term ‘tunnel fires’ is extensive, covering a very great variation in terms of different types of tunnels and different types of fires. There are considerable variations in the design of tunnels throughout the world, while fires can differ greatly in terms of intensity or extent, depending on what is burning. It is generally materials that are moving through, or which are in, a tunnel that burn. In road tunnels, it can be road vehicles that burn; in railway tunnels, it can be trains that burn; in service tunnels (cable tunnels etc.) it may be electrical cables or the thermal insulation of district heating pipes that burn, while in underground railways it can be the trains that burn, or fires can occur in accumulations of rubbish, such as collections of thrown-away free newspapers. In other words, it is very difficult to define a fire in a tunnel. However, as the objective of this report is to describe how rescue and fire services can deal with catastrophic fires in tunnels, the rest of this report will concentrate on extensive fires that can have serious results. From the point of view of the Fire and Rescue Services, the most important aspect is that accidents should not occur at all, which means that the most important safety measures are those intended to reduce the probability of accidents. However, as accidents cannot be ruled out, and as accidents in tunnels can have particularly serious consequences, tunnels must be designed acknowledging that accidents can occur. As far as the Fire and Rescue Services are concerned, the most important measures that can reduce the severity of accidents are that there should be short distances to, and simple means of reaching, escape routes for those escaping from a fire, that the firefighters can safely approach the fire and that the fire cannot grow excessively before fire-fighting work can start. These various conditions can be achieved in different ways, but there must be an overall safety programme that identifies all the parameters involved, makes them work together and creates the best conditions for high safety levels.

3(21) Anders Bergqvist, Stockholm Fire Brigade 3.1 Road tunnels Road tunnels are generally designed as twin-bore tunnels, carrying one-way traffic in each bore, as shown in Figure 2. This arrangement provides a relatively good ability to connect the two tunnel bores with each other, thus providing escape routes and approach routes for tackling the fire. Most of these twin-bore tunnels have ventilation systems that, in the event of a fire, can ventilate away the fire gases in the traffic direction of the tunnel, so that those upstream of the fire can be evacuated without being affected by the fire gases.

Connections between the

The arrows show the traffic direction and the ventilation direction in the event of

Figure 2 Schematic diagram of a twin-bore tunnel with one-way traffic, seen from above.

Another way of building road tunnels is as single-bore tunnels with two-way traffic, as shown in Figure 3. This normally creates relatively considerable difficulties in incorporating escape routes, and thus also in providing routes by which the fire brigade can approach the fire. Evacuation and fire-fighting are further complicated by the fact that these tunnels often do not have a ventilation system that can deal with the smoke from a fire, thus assisting evacuation or fire-fighting. Designing a ventilation system as a longitudinal-flow system means that there is a considerable risk that many of those caught in the fire will be exposed to the smoke from it. If the ventilation system is designed on some other principle, e.g. as transverse ventilation, it is important that it should have adequate capacity to deal with the fire effect, and with the quantities of smoke developed by the fire.

4(21) Anders Bergqvist, Stockholm Fire Brigade

The arrows show the traffic

Figure 3 Schematic arrangement of a single-bore tunnel with two-way traffic, seen from above.

There are several other designs of road tunnels, but the two described above can be regarded as representative of existing tunnels. This simplification is not so important here, as the theme of this report is that of describing rescue actions.

3.2 Fires in road tunnels The size of a fire in a road tunnel will have a very considerable effect on the ability of the Fire and Rescue Service to perform effective rescue and/or fire-fighting operations. When tackling fires in road tunnels, personnel and equipment should be capable of dealing with fires of the magnitude that can be encountered. This is important, as many persons can be involved in road tunnel fires.

Fires in private cars in twin-bore tunnels will almost certainly not be of capacity-determining rating for normal fire-fighting. Figure 4 shows that the likely fire outputs should be within the capabilities of a fire-fighting force. However, the same fire in a single-bore tunnel could lead to considerable difficulties, depending on whether there is any air flow through the tunnel or whether there is a ventilation system capable of evacuating the smoke from the fire.

Extinction begin 8 Renault - ref 6 Trabant - ref 7 Citroen - ref 7 6 Petrol tank burst Austin - ref 7 Austin Maestro - ref 8 4 HRR (MW) Citroen - ref 8

0 0 1020304050607080 tid (min)

Figure 4 Experimentally determined thermal outputs, HRR, for passenger car fires5

5(21) Anders Bergqvist, Stockholm Fire Brigade The curves in Figure 4 do not consider the risk of the fire spreading to other cars, and thus the potential for greater thermal power output and a larger area of fire. The factors that will probably set the capacity requirements for fighting a fire in a tunnel will be: 1. the number of persons whom the rescue and fire services must assist out to safe conditions 2. the size of the fire, and thus the temperature and thermal radiation power that will face the firefighters 3. the distance that the firefighters have to travel in a smoke-filled environment to reach the fire.

Fires in trucks and coaches, on the other hand, are likely to reach such output levels (see Figure 5) that it can be difficult effectively to tackle the fire. As the carriage of goods by road is generally increasing in Europe, this can also mean that the probability of fires in freight vehicles is also increasing. Many tunnels have been designed with capacities that are capable only of dealing with fires with outputs of up to 20 – 30 MW6.

140 HGV- Grant and Drysdale Simulated truck load - Ingason 120 Bus - Ingason et al Bus - Steinert 100

60 HRR (MW) 40

0 0 5 10 15 20 25 30 35 40 Time (min)

Figure 5 Experimentally determined fire powers, HRR, for trucks and buses4

Fires of this order of size will generate very high radiation levels, both from the smoke and from the actual flames. The table below shows the thermal radiation levels recorded at the fire trials in the Runehammar Tunnel in Norway in 2003.

6(21) Anders Bergqvist, Stockholm Fire Brigade Thermal output Radiation level Radiation level Radiation level (peak), 5 m upstream 10 m upstream of 20 m upstream Basis of the fire HRR (MW) of the fire the fire of the fire q” (kW/m2) q” (kW/m2) q” (kW/m2) 110 (peak) 10-12 (17 minutes) 2 (peak) Wooden pallets and 200 MW plastic pallets, 10 tonne 40-60 (17 minutes) 29-35 (5 minutes) 9-19 (6 minutes) 3 (peak) Wooden pallets and 170 MW mattresses, 6 tonne 20 (short period) 9 (short period) 2 (peak) Furniture, 7.7 tonne. 130 –140 MW Truck tyres, 0.8 tonne Cartons of plastic 40 (short period) 8 (short period) 4 (peak) beakers on wooden 70 – 80 MW pallets, 2.6 tonne

Table 1 Measured thermal outputs and radiation values from fire tests in the Runehammar Tunnel, September 2003.

4. FIRES IN TUNNELS: FIRE AND RESCUE OPERATION

Safety in tunnels depends on many different factors and conditions: the actual design of the tunnel itself, and the type of traffic using it, will have a considerable effect. Escape routes and approach routes for the public fire service are only parts of this overall safety, although they are the vital parts that must operate satisfactorily in the event of an extensive tunnel fire. The fire and rescue operations should be so structured, and matched to the design of the tunnel and the conditions likely to be encountered, as needed to suit each particular specific case: see Chapter 2. Today’s Fire and Rescue Service have developed many methods for dealing with different types of accidents. In general, it can be said that the more common the type of accident, the more developed are the methods for dealing with it. The concepts of upstream and downstream of the fire will recur in this report, and are to be seen as in relation to the direction of air flow in the tunnel. The area upstream of the fire is that away from the fire and against the direction of air flow. Downstream of the fire is the zone away from the fire itself, in the direction of the air flow. It is in this latter direction that most of the fire gases flow. ‘Backlayering’ is the phenomenon by which the fire gases flow against the direction of air flow in the tunnel. The distance to which this effect can occur depends on the size of the fire, the volume of the air flow in the tunnel and other parameters.

4.1. Methods available to the fire services for fighting a tunnel fire In principle, the Fire and Rescue Services in Sweden apply the following tactical approaches to tackling fires in tunnels: 6 1. working in the tunnel to extinguish the fire, thus eliminating the threat to those caught in it, 2. working in the tunnel to assist/rescue those caught in the fire, to get them out of the tunnel as quickly as possible, 3. ventilation of the tunnel in order to drive the smoke away from the fire in one direction, thus facilitating evacuation and fire-fighting, 4. fighting the fire from a safe position, in order to limit its consequences, 5. actively dealing with those escaping from the fire to safe conditions or outside the tunnel.

These various approaches must then be brought together to provide a suitable combination for dealing with each specific accident. 7(21) Anders Bergqvist, Stockholm Fire Brigade

One aspect which becomes clear when studying the information shown in Figures 4 and 5 is that of the time before the fire starts to grow rapidly. These results seem to indicate that fires tend to take hold fiercely after about the first 5 - 10 minutes. This can be compared with the ‘golden 60 seconds’ that are available for tackling aircraft fires, and which indicate the maximum time that can elapse before the first fire-fighting work starts. If the Fire and Rescue Services are to be able to start fighting fires within about ten minutes, much needs to be done in order to improve the efficiency of their work.

The working methods normally employed by Fire and Rescue Services have come to reflect the accidents and fires most commonly encountered today. In the case of major fires in tunnels, it is highly likely that it will be necessary to use very different methods from those employed in tackling fires in residential buildings or ordinary traffic accidents. Serious fires, such as those in the Mont Blanc Tunnel or the Tauern Tunnel7, show all too clearly that the methods employed by the Fire and Rescue Services are based on the necessary responses after the fire has been brought under control, and the assumption that the Fire and Rescue Services will be able easily to get to the fire. Working methods that can be considered for dealing with fires in tunnels are as follows. 1. Entry into the tunnel to ascertain conditions, i.e. to note the conditions at the accident site and to obtain an overall picture. This is done with the aim of providing information needed for further work. It may be necessary to do this in smoke-filled conditions, which means that those carrying it out must be appropriately protected. It must also be done immediately on arrival, and be fast and effective, in order not to delay the rest of the work. 2. Entering the tunnel to extinguish the fire and eliminate the threat to persons in the tunnel. This may have to be carried out under dangerous conditions, facing smoke and high thermal radiation levels, which means that those involved must be suitably protected. Actually extinguishing the fire may be very difficult, and can probably be done in a number of different ways, of which the following are examples of possible methods: - The use of ordinary hoses and nozzles. - The use of portable water monitor. - The use of vehicle-mounted water monitor. - The use of fans, with water being injected into the air flow. - Moving the burning object etc. out of the tunnel. - The use of remotely controlled fire-fighting equipment.

3. Work in the tunnel in order to guide those escaping from the fire, i.e. those capable of fleeing. This may also have to be carried out under dangerous conditions of smoke and high radiation levels, which means that those performing the work must be suitably protected. 4. Work in the tunnel physically to carry out victims, i.e. what is generally called life-saving. This may also have to be carried out under dangerous conditions of smoke and high radiation levels, which means that those performing the work must be suitably protected. 5. Work in the tunnel to rescue those involved in the fire and help them to survive in the vicinity. This may also have to be carried out under dangerous conditions of smoke and high radiation levels, which means that those performing the work must be suitably protected. 6. Ventilation of the tunnel in order to control the quantity and direction of flow of smoke in the tunnel. The purpose of this can be: - to ventilate the tunnel in order to assure the existing air flow through it, thus facilitating evacuation and rescue work.

8(21) Anders Bergqvist, Stockholm Fire Brigade - to ventilate the tunnel in order to start air flow through it, thus creating a possible escape route and a means of approach for the firefighters. - to ventilate the tunnel in order to reverse the direction of flow of smoke, thus facilitating the rescue of those in the smoke downstream of the fire site.

7. Advanced acute care for victims, under safe conditions in the vicinity of the fire. This method is likely to require very considerable resources if large numbers of persons are involved.

The prime aim of all these methods is to save lives, although they also represent different and, in many cases, equally important, methods of ensuring an effective input, depending on the particular conditions of the accident. The reason that several of these methods are often not considered as methods of saving life is probably because most fires occur in considerably less complicated areas, involving considerably fewer persons. Under such circumstances, it is not as apparent that, for example, information is one of the most important points when starting the work, or that quickly evacuating those who are first found is not necessarily the most effective way of saving lives.

4.2 Tackling fires in twin-bore tunnels When a fire breaks out in a twin-bore tunnel, traffic in the affected bore must be stopped upstream of the fire. The normal ventilation of the tunnel must ensure that those escaping upstream from the fire are not affected by smoke from it. All traffic downstream of the fire should be able to continue to drive out of the tunnel before it is filled with smoke. Traffic in the other bore of the tunnel must be stopped, so that the Fire and Rescue Service can enter the tunnel and reach the fire via the connections between the two tunnel bores.

Connections between the

Flow direction of the air in the

Figure 6 Tackling a vehicle fire in a twin-bore tunnel

Tackling this type of fire will involve the following elements. 1. Stopping all further traffic into both bores of the tunnel. Experience shows8 that this closure must be in the form of an actual physical barrier. 2. The fire tenders enter the unaffected tunnel in the normal traffic direction from the nearest entrance point. It is important that there should be clear access routes for the fire service vehicles, as there is likely to be traffic chaos outside the tunnel. 3. First personnel to arrive quickly tackle the fire in the vehicle. 9(21) Anders Bergqvist, Stockholm Fire Brigade 4. At the same time, the smoke-free part of the tunnel upstream of the fire site should be evacuated. 5. Depending on the size of the fire, the second and third crews to arrive should assist the first crew in putting out the fire and start to extinguish any fires downstream of the original fire and search the tunnel for persons trapped. The purpose of this action is to rescue anyone trapped in the smoke, and to remove the danger of the fire spreading to any vehicles in the smoke downstream of the fire: see Figure 6.

If this twin-bore tunnel is in a major urban environment, it is very likely that there will be queues in the tunnel. Dealing with a tunnel fire under such circumstances means that the most important step in protecting those already in the tunnel is unavailable: vehicles downstream of the fire will not be able to drive out of the tunnel, but will be trapped in the traffic queue and unable to move. This means that there will be a large number of persons trapped in the smoke downstream of the fire. Some of them will escape through the connections to the parallel tunnel bore, but is unlikely that all will do so. As a result, the fire service will suddenly be faced with the problem of dealing with large numbers of persons trapped in the smoke.

10(21) Anders Bergqvist, Stockholm Fire Brigade

Connections between the

Flow direction of the air in the

Figure 7 Fire and rescue operation dealing with a car fire in a twin-bore tunnel with queuing vehicles.

Tackling this type of fire will involve the following elements. 1. Stopping all further traffic into both bores of the tunnel. 2. The fire tenders enter the unaffected tunnel in the normal traffic direction from the nearest entrance point. 3. First personnel to arrive quickly tackle the fire in the vehicle. 4. At the same time, the smoke-free part of the tunnel upstream of the fire site should be evacuated. 5. Depending on the size of the fire, the second and third crews to arrive should assist the first crew in putting out the fire and start to extinguish any fires downstream of the original fire and search the tunnel for persons trapped. The purpose of this action is to rescue anyone trapped in the smoke, and to remove the danger of the fire spreading to any vehicles in the smoke downstream of the fire: see Figure 7. 6. Powerful ventilation after the fire has been extinguished will more quickly dilute the dangerous gases in the smoke, thus improving survival conditions in the smoke. 7. If the firefighters fail to extinguish the fire, it will be necessary to reverse the direction of air flow and smoke in the tunnel after the tunnel upstream of the fire has been cleared of persons and vehicles. The effect of this will be to create a safe environment for those trapped in the original downstream direction of the fire: see Figure 8.

11(21) Anders Bergqvist, Stockholm Fire Brigade

Connections between the Fan

Flow direction of the air in the

Figure 8 Fire and rescue operation dealing with a car fire in a twin-bore tunnel with queuing traffic, after reversal of the direction of air flow.

4.3. Tackling fires in single-bore tunnels

Fan

Flow direction of the air in the

Figure 9 Tackling a car fire in a single-bore tunnel

1. Quick reconnoitring of the tunnel in order to obtain a picture of the situation, and also to see how far the smoke has spread. 2. Ensure an air flow through the tunnel by starting ventilation in the most suitable direction. 3. If possible, those who were first into the tunnel to investigate the conditions should start to put out the fire in the vehicle. If not, they should take themselves out of the tunnel: see Figure 9. 4. Evacuate persons in the tunnel upstream of the fire site. 5. Enter the tunnel from the downstream end (against the smoke) with the aim of rescuing anyone in the smoke close to the tunnel mouth. 6. Once the fire has been put out, powerful ventilation of the tunnel in order quickly to reduce the toxic concentrations in the smoke, thus improving the prospects for survival of those trapped in the smoke. 7. If it is not possible to put out the fire, the air direction in the tunnel must be reversed in order to dissipate the toxic smoke from anyone trapped in the tunnel, after the upstream end has been cleared of persons and vehicles. 12(21) Anders Bergqvist, Stockholm Fire Brigade

Fan

Flow direction of the air in the

Figure 10 Fire and rescue operation dealing with a car fire in a single-bore tunnel after reversing the direction of air flow.

4.4. Fire and Rescue operation problems encountered in tunnel fires The rescue and fire services will be faced with several problems9 which must be considered when tackling the fire. Some of these problems are described in more detail below.

4.4.1. A picture of the fire scene It is difficult to obtain a picture of the accident scenario, as it is very difficult to see what is happening in a tunnel fire. There is therefore a considerable lack of information on what is happening, which makes it difficult to decide what to do. As a result, important time may be lost. This aspect must be tackled and resolved. Probably the most effective way is for the tunnel-owner to install surveillance equipment in the tunnel, and to ensure that there is good communication with the Fire and Rescue Service, in order to ensure that the information is available when needed. If there is no physical equipment to provide the information, the Fire and Rescue Service will have to solve the problem itself. Section 4.1 suggested that the very first step of the rescue and fire services should be to obtain information on the fire. This needs to be done quickly and effectively, and must provide an accurate picture of conditions. This is at present an area where further development is required, particularly as Swedish practice in connection with the use of breathing equipment seems to reduce the efficiency of work in tunnels, without increasing the safety of the firefighters themselves. More investigation is needed of combinations of rapid methods, either using vehicles or on foot, and with the assistance of aids such as IR cameras and illuminated tracking lines, laid by the firefighters, and without pulling fire hoses with them into the tunnel.

13(21) Anders Bergqvist, Stockholm Fire Brigade 4.4.2. Controlling the air flow in the tunnel Together with the tunnel operators, the firefighters need to know in what direction to drive the fire gases in order to facilitate evacuation, rescue and fire-fighting. As ventilation is likely to become the only effective method of tackling serious tunnel fires, there is a need for better knowledge of ways of ventilating tunnels, and for the development of standard procedures for doing so. Generally, today there is small possibilities for a Fire and Rescue Service to ventilate a tunnel in the event of a fire, nor knowledge of how this should best be done. However, a number of projects have been started with the aim of providing information on how this should be done 10,11. Ingason10 has provided the following results of CDF calculations for ventilation flow rates in the Manesse Tunnel (523 m long and 23.4 m2 cross-section) and the Käferberg Tunnel (2118 m long and 45.4 m2 cross-section) in Switzerland. The theoretical fan capacity was 37.5 m3/s, with a fan diameter of 1.22 m. Manesse Tunnel Käferberg Tunnel Air flow in the tunnel, Final air velocity after 3.7 2.2 powered by a mobile fan, reversing, (m/s) but without a fire Time to reverse air flow, 4 10 (minutes) Air flow in the tunnel Final air velocity after 3.6 2.2 established by a mobile fan reversing, (m/s) with a train in the tunnel, Time to reverse air flow, 1 3 but without a fire (minutes) Air flow in the tunnel, Final air velocity after 2.5 2.1 established by a mobile fan reversing, (m/s) with a train in the tunnel Time to reverse air flow, 1 3 and with a 15 MW fire (minutes)

Table 2 Results of calculations of air flow reversal with a tunnel fire

The method described above seems to work so well that it should be worth while continuing with more, and more accurate, investigations in order to provide better information for use by the public fire services. Ingason’s results indicate that it should be possible to develop this to a practical working method for dealing with fires in tunnels.

4.4.3. Large and complicated objects Tunnels can be large and extensive objects, with long and complicated routes from safe environments, sometimes outdoors and sometimes in the tunnel system itself, in order to get to the final point where rescue or fire-fighting is needed.

Investigation is needed in order to find ways for the Fire and Rescue Services quickly to make their way to a safe position in the vicinity of the fire and of those requiring rescue. On the international plane, this has been resolved in a number of different ways in long single-bore tunnels, using everything from special rescue vehicles to special trainsets in the tunnel, and also with special-purpose access tunnels in parallel with the main tunnels. This need not normally be a decisive problem in the case of twin-bore tunnels, as in most cases it is possible to provide relatively frequent connections between the two bores.

14(21) Anders Bergqvist, Stockholm Fire Brigade 4.4.4. Tackling the smoke Tunnels fill with fire gases, which make it necessary to use breathing equipment when entering them. This severely limits the scope for action, as well as the distance which can be covered in the tunnel.

What is generally meant by the use of breathing apparatus is that of the use of equipment for saving lives and tackling normal house fires. Such equipment and methods have definite limitations when used in tunnel environments. Bergqvist9 et al. carried out tests to investigate the effective working range of a firefighter wearing breathing apparatus in a tunnel environment. Some of the results of these tests are as shown in the table below.

Test conditions Movement speed (m/minute) Maximum range (m)* Smoke-filled tunnel, dry hose 4.3 58 Non-smoke-filled tunnel, pulling a 18 243 water-filled hose Non-smoke-filled tunnel, no hose 80 1080 Trials in an industrial environment in 6 80 another investigation * Based on 2400 l of air, an air consumption of 62 l/min and ability to retreat

Table 3 Movement speed and range of breathing apparatus groups in tunnel environments

‘Working with breathing apparatus’ is an umbrella name for all working methods involving the use of protective fire-fighting clothing and breathing apparatus in a smoke-filled environment. Further development and more detailed investigation of all these working methods is needed: today, not enough is known about the capacity and limitations of the various methods. This information is needed, as it is of decisive importance when planning fire-fighting work in tunnels.

4.4.5. Extinguishing the fire Actually extinguishing the fire can be very difficult. It has been found, in the cases of previously described fires, that fires in trucks and buses can be relatively extensive, reaching high temperatures and producing high levels of radiation and dense smoke. If the tunnel is ventilated, the Fire and Rescue Service can probably reach the seat of the fire without having to pass through excessive smoke or heat. Nevertheless, radiant heat at the site of the fire can be very considerable, imposing severe limitations on the ability to stay in the vicinity of the fire for a sufficiently long period of time to put it out and so reduce the amount of thermal radiation. If, on the other hand, there is no ventilation in the tunnel, tackling the fire will be very dependent on the choice of approach route, as there will only be the natural draught in the tunnel to determine the direction of flow of the smoke. It is also very likely that, under such conditions, there will be substantial backlayering of the smoke as a result of insufficient air flow, which will create very unpleasant temperature conditions in the fire gases and radiation from them and from the fire, regardless of from which side the fire is tackled. The firefighters will probably use water to tackle the fire: other methods of extinguishing it have not been sufficiently developed, or are not sufficiently widely used at present. How much water will be needed in order to put out the fire? This is an important question to answer, as it determines the use of a certain number of jets for a certain period of time. In turn, these jets require a certain number of firefighters, working under difficult conditions. 15(21) Anders Bergqvist, Stockholm Fire Brigade Results from Särdqvist12 have been used in order to obtain some idea of the quantity of water needed to extinguish a vehicle fire in a tunnel. Särdqvist’s work was based on the extinguishing requirements for fires occurring in non-residential buildings. In such fires, the firefighters had straightforward access to the fire. In this context, we need to remember that vehicle fires are particularly difficult to put out, which means that the following simplifications must be seen as an absolute minimum requirement in terms of water quantities.

Minimum Type of Fire area Fire power (MW) extinguishing Number of vehicle (m2) water requirement 360 l/min jets (l/min) Private car 10 5 226 1 Van 35 15 462 2 Truck 200 100 1250 4

Table 4 Absolute minimum water requirements for extinguishing a vehicle fire6

In order to be able to extinguish the fire, the water must reach the seat of the fire which, particularly in the case of fires in vehicles, means that the firefighters have to get very close to the vehicle, as it is particularly difficult to reach the seat of the fire in such cases. This water flow rate then has to be maintained for a significant period in order to put out the fire. According to Ingason6, it takes about 30 minutes, with at least the above quantity of extinguishing water, in order to put out a fire in a truck. Although firefighters are protected against toxic smoke and high temperatures, they cannot withstand high temperatures or high radiation levels for a long period of time. In experiments, Person13 showed that firefighters could withstand a heat load of 5 kW/m2 for at least seven minutes. However, for a firefighter to withstand a stay of 20 minutes, the radiation level cannot exceed 2 kW/m2. Experiments14 were carried out in Stockholm on the equipment normally used by Swedish fire services for putting out vehicle fires. The height of the tunnel in which the tests were performed was about 4.5 m, which is typical of European clearance heights above roads. The results of the trials were as follows.

16(21) Anders Bergqvist, Stockholm Fire Brigade Water Nozzle Pump Maximum Flow rate delivered to pressure distance from (l/m2, min) a 1 m2 fire (bar) the fire area (l/m2, min) Fogfighter (Handheld 5 27 320 40 nozzle, 300 l/min) Fogfighter (Handheld 10 32 431 17 nozzle, 300 l/min) TFT-jet (Handheld nozzle, 5 20 702 26 1300 l/min) TFT-jet (Handheld nozzle, 10 35 946 58 1300 l/min) Water monitor (Handheld 5 - - - nozzle, 1000 l/min) Water monitor (Handheld 10 - - - nozzle, 1000 l/min)

Table 5 Maximum throw length in a tunnel environment of fire fighting nozzles as used by Fire and Rescue Service

The collection area (i.e. the fire) was positioned at a height equivalent to 2 m above ground level. One of the points noted was that it was very difficult to target the application area, as a result of poor visibility. If the TFT jet is to be used, there should not be more than 25 m of hose between the pump and the jet, in order to ensure sufficient pressure and flow rate. The water monitor could not be used in a tunnel due to the fact that it was not possible to depress the angle of the jet sufficiently to avoid all the water being sprayed on to the roof of the tunnel. As far as the Fogfighter was concerned, the throw angel at a distance of 35 m was so shallow that very little water landed on the intended target.

Summarising the above, it can be seen that the Fire and Rescue Service will have considerable problems in tackling a fully established tunnel with a higher rate of heat release (HRR). The firefighters need to get sufficiently close to the fire to be able to get water on the flames and the fire. This is essential in order to reduce the amount of radiant heat, so that they can get closer to the vehicle and get water on to the seat of the fire. If they cannot do this, they will not be able to control the fire, or gradually to reduce its rate of heat release. The main obstacle in preventing firefighters from getting close to a fire is thermal radiation from the fire and from back-layered fire gases. It should be possible to deal with these fire gases, using fans to increase the air flow, but thermal radiation from the fire and from any residual backlayering will be difficult to deal with. Development of some form of protection against thermal radiation is needed in order to assist tackling fires of this type, perhaps through the use of water mist jets or water curtain jets.

4.4.6. Water transport to the site of the fire Getting fire-fighting water to the position where it is required. This can be extremely difficult if thought has not been given to this problem when designing the tunnel.

4.4.7. Facilitate evacuation and rescue those caught in the fire The Fire and Rescue Service may find itself faced with a major evacuation and rescue situation, involving large numbers of persons, for which it must be prepared.

17(21) Anders Bergqvist, Stockholm Fire Brigade 4.4.8. Assessment of risks while working It is very difficult to assess the risks to which the fire and rescue personnel are exposed when tackling a tunnel fire. These risks can range from failure of the air supply from a breathing apparatus to falling rock from the heat-affected tunnel roof.

4.4.9. Communication between various positions There must be a reliable communication system between those in safe positions and those in the tunnel.

18(21) Anders Bergqvist, Stockholm Fire Brigade 5. SUMMARY AND CONCLUSIONS It seems as if those responsible for the design of safety features and systems in present-day tunnels assume that the Fire and Rescue Authority will be able quickly, safely and effectively to deal with any accidents that occur. It also seems as if the Fire and Rescue Authority are not reacting, by notifying those responsible for the safety of tunnels that they are being seen as the potential providers of a service for which, in fact, they do not have the appropriate working methods, nor resources in terms of equipment or personnel. In future the Fire and Rescue Authority will have to investigate their abilities to provide the necessary evacuation and life-saving services in tunnel fires. This is important information, which needs to be considered in the overall planning and provision of tunnel safety systems. It will be very important that the Fire and Rescue Services can conduct a dialogue with tunnel operators concerning their rescue capacities, particularly as it ought to be the responsibility of the tunnel-owner to ensure that the tunnel can be safely evacuated, and persons rescued, if necessary. The tunnel-owner has important responsibilities in creating the right conditions so that rescue operations can be carried out if needed, but it is only when working closely with the local rescue and fire service that the tunnel-owner can draw up a solution describing how the work of the rescue and fire services can be integrated with other safety features and systems in the tunnel.

The way in which the Fire and Rescue Service performs its work will be very dependent on the type of accident that has occurred, on the environment in which it has occurred, the consequences that the accident has had, or is having, the objective of the work and the resources available. The tactical approach adopted will be very much affected by the competence of those available for the rescue action. Together, all these various aspects will affect the tactical approach to the rescue operation.

As far as the Fire and Rescue Authority are concerned, the most important consequence-reducing measures that can be incorporated in a tunnel to reduce the severity of accidents are that there should be short distances to, and simple means of reaching, escape routes for those evacuating a fire, that the fire-fighting personnel can safely approach the fire and that the fire cannot grow excessively before fire-fighting work can start. The size of the fire in a road tunnel will have a very considerable effect on the ability of the rescue and fire service to perform an efficient rescue action. Fires in private cars will not be the determining factor for a normal fire and rescue operation in a twin-bore tunnel, but they will be critical in a single-bore tunnel. What is likely to be the determining factor in the ability to deal with a fire in a tunnel is the number of persons who will need to be assisted out into safe conditions, the size of the fire (and thus the temperature and radiation level to which the fire-fighting personnel will be exposed) and the distance that the firefighters have to travel in smoke-filled conditions.

The tactical approach that the rescue and fire service can adopt in dealing with a tunnel fire consists of combinations of the following: • Working in the tunnel to extinguish the fire, thus eliminating the threat to those caught in it, • Working in the tunnel to assist/rescue those people caught in the fire, to get them out of the tunnel as quickly as possible, • Ventilation of the tunnel in order to drive the smoke away from the fire in one direction, thus facilitating evacuation and fire-fighting, • Fighting the fire from a safe position, in order to limit its consequences, • Actively dealing with those escaping from the fire to safe conditions or outside the tunnel.

19(21) Anders Bergqvist, Stockholm Fire Brigade It is clear that the thermal output of a vehicle fire seems to rise rapidly after about the first 5-10 minutes. Much has to be done in order to improve the efficiency of fighting tunnel fires if the Fire and Rescue Service is to be able to reach the fire and start work within about ten minutes of receiving the alarm. This shows that the method of tackling such fires has to have been very well planned if it is to be carried out efficiently and limit damage and injury before it is too late. If fire-fighting is to be carried out effectively, the rescue and fire service must obtain an overall picture of the situation. This means that obtaining relevant information is probably the first step to be taken in such a situation, which means that methods need to be developed in order to ensure that this can be done quickly and effectively.

It has also been found that the efficacy of the work is dependent on being able to get close to the seat of the fire, which emphasises the importance of being able to control ventilation at the site and establish the flow of air to carry away the fire gases. With present-day methods and equipment, the use of breathing apparatus is not an efficient method of dealing with fires in larger tunnels. The working methods involved need to be investigated, and new methods developed. Putting out a fire can be very complicated. In order properly to extinguish a vehicle fire, the extinguishing water must reach the seat of the fire, which requires the firefighters to get close to the vehicle, as it is exceedingly difficult to reach the seat of the fire from a distance. Under such conditions, thermal radiation from the fire is likely to make this very difficult, and so there is a need to develop methods to protect firefighters from thermal radiation

20(21) Anders Bergqvist, Stockholm Fire Brigade REFERENCES

1 The Public Rescue and Fire Services Act, SFS 1986:1102. 1986 2 Conversation with Key Hedström, lawyer with the National Board of Civil Defence, Rescue and Fire Services (SRV), concerning practical application of 41§ of the Public Rescue and Fire Services Act, SFS 1986:1102. 3 Svensson, S. The operational problem of fire control, PhD thesis, Report LUTVDG/TVBB- 1025-SE, Lund University. 2002 4 Widlund, P. Rescue tactics: influence on them and execution. [In Swedish.] National Board of Civil Defence, Rescue and Fire Services, Report U29-385/92. 1992 5 Ingason, H. An overview of vehicle fires in tunnels. Proceedings of the Fourth International Conference on Safety in Rail and in Road tunnels. Madrid 2-6/11. 2001. 6 Ingason, H, Bergqvist A, Frantzich H, Hasselrot K, Lundström S. Planning for Manual Firefighting and Rescue in Tunnels. Proceedings of the Fourth International Conference on Safety in Rail and in Road tunnels. Madrid 2-6/11. 2001. 7 Bergqvist, A. The fire in the Tauern Tunnel. [In Swedish.] SRV Report. 1999 8 The author's own experience from traffic in Stockholm, Sweden. 9 Bergqvist, A, Frantzich H, Hasselrot K, Ingason H. Rescue work in tunnel fires. SRV Report P21-391/01, 2001. [In Swedish.] 10 Ingason, H, Romanov, L. Use of mobile fans in tunnels. SP report no. 2002:06. SP Swedish National Testing and Research Institute. 2002 11 Kumm, M., Nyman, H., Hasselrot, K. Experiments with mobile fans in the Masthamn Tunnel, Stockholm. [In Swedish.] Research Report MdH Ist 2003:3. Mälardalen Institute of Technology. 2003 12 Särdqvist, S. Demand for extinguishing media in manual firefighting. PhD thesis. Report LUTVDG/(TVBB-1021). Lund University 13 Person, H. Basic equipment for the use of foam as an extinguishing medium. Experimental results and recommendations as a basis for determinations of design capacity and performance. SP report 1990:36. SP Swedish National Testing and Research Institute. 1990 14 Throw tests, performed by Anders Bergqvist and Ulf Lundström, Stockholm Fire Service, July 2003.

21(21) Anders Bergqvist, Stockholm Fire Brigade

Task 3.2: Computer supported training tools for rescue personnel in tunnel accidents (EBSCC)

Executive Summary

The report gives an overview of three virtual training systems for emergency personnel. This report should explore the possibilities for virtual training and the evaluation of them.

Contents WP 3.2

1. Abstract...... 21

2. Objectives ...... 22

3. Introduction ...... 23

4. Data collection ...... 24

5. Data analysis and practical examples ...... 25 5.1 Virtualfires ...... 25 5.2 Virtualtraining-ADMS ...... 32 5.3 GAMMA-EC...... 36 5.4 Fire simulators comparison...... 44

6. Proposal for EU guidelines...... 48

7. Limitations...... 49

8. Recommendations...... 50

9. References...... 51

10. Abstract

The purpose of this document is to provide recommendations on training of rescue personnel by the evaluation of computer supported training systems (multimedia, virtual reality tools, team training by gaming) for rescue personnel and disaster management organisations, with the objectives of increasing the effectiveness of rescue operations. The criteria used for the evaluation have been: user acceptance, user friendliness, and availability. The softwares evaluated have been the following: Virtualfires, advance Disaster Management System and GAMMA-EC. Based upon these three computer training tools, topics for EU guidelines have been derived.

Due to lack of enough information the evaluation realised has been limited and therefore recommendations have to be considered taken into account this situation.

Some recommendations are made being them general ones as the three software’s analysed are different in its objectives and therefore intended to be used by different kind of users. Recommendations on training of rescue personnel are difficult to be provided as these software’s are specific for particular situations and training teams. Therefore, the election of the software has to be realised for each particular situation and be evaluated in function of its expected results for the situation analysed, specific objectives to achieve and type of users.

Virtual reality based simulators offer a safe and cost effective alternative to study life safety in tunnels being actually the only method available to perform full scale real fine tests. The aim of these software’s is to develop and implement software components integrated into an inexpensive, environmentally friendly analysis and training tool for fire hazard studies.

In the EU Directive recommendations on the use of these kind of computer supported training systems should be mentioned and recommended for all the safety units in charge of safety in tunnels. Closing a tunnel for training is in most cases impossible due to the inexistence of non used tunnels and therefore training is not possible in the same conditions than when an accident/incident occurs in a tunnel. That is why computer supported training systems are essential to increase the effectiveness of rescue operations and general safety in tunnels.

11. Objectives

Objectives of this task are to provide recommendations on training of rescue personnel by the evaluation of computer supported training systems (multimedia, virtual reality tools, team training by gaming) for rescue personnel and disaster management organisations, with the objectives of increasing the effectiveness of rescue operations and to be in support of the implementation of the point 5 of the Annexe II (related to computer simulation exercises) of the EU Directive 2004/54/EC on minimum safety requirements for tunnels in the trans-European road network. The criteria used have been: user acceptance, user friendliness, and availability. The softwares evaluated have been the following: Virtual fires, ADMS-Virtual training and GAMMA-EC.

Comparison between the characteristics, capabilities and results of the three programs is provided.

12. Introduction

A simulator is a computer application that, with an operative system and a computer, generates similar results to the ones obtained in real practice.

The use of simulators has widely developed during the last decade and fire simulators and emergency management have been largely applied in public administration and business sectors training as very useful tools to improve the training of rescue personnel and disaster management organisations and therefore increase the effectiveness of rescue operations.

In Chapter 4 of this report the way of how data has been collated is presented. The information available to analyse the three softwares have been insufficient in order to make a proper evaluation. The evaluation made is quite poor in the sense that only three softwares have been available to be tested because of problems of availability, low input obtained from other countries experiences and licenses. The evaluation has consisted on testing a DEMO of Virtual fires and Virtual training-ADMS and Gamma EC and therefore most of the aspects to be evaluated have not been evaluated properly.

In Chapter 5 of this report “Data analysis”, information on the three software’s evaluated is presented. The data comes from the information placed on the Internet (websites, articles, leaflets, etc.). No information on Virtualfires has been found and ADMS Virtualtraining website it has not worked during the whole period of the project. Therefore the data gathered is quite limited. This section is divided in three subparagraphs each of one related to one of the software’s: Virtualfires, ADMS- Virtualtraining and Gamma EC as stated in the document of work

In Chapter 6 “Proposal for EU guidelines” general recommendations and specific ones are provided to the Corrigendum to Directive 2004/54/EC of the European Parliament and of the Council of 29 April 2004 on minimum safety requirements for tunnels in the trans-European road network after its deep analysis and its comparison with the data collected.

Limitations encountered during the data analysis are stated in Chapter 7. These limitations, such as reliability of data, insufficient data and other have been considered before establishing the recommendations.

From the analysis of the limitations encountered some recommendations for future work are provided in Chapter 8.

13. Data collection

As described in previous paragraphs information has been collected by looking for computer supported training systems available around Europe to be analysed. As mentioned before only three softwares have been possible to test due to lack of availability.

The information gathered in order to make the evaluation has been quite limited.

For Virtualfires and Virtualtraining the information used has been a DEMO provided by TNO and NIBRA. GAMMA-EC has been tested though the personnel of the EBSCC that has collaborated in the project. Literature referred to Virtualfires, ADMS Virtualtraining and GAMMA-EC has been obtained from the Internet.

User acceptance, user friendliness and availability are the parameters used for the comparison and evaluation of the three softwares.

Mention should be made to the low input obtained from other partners due to lack of specific information on their countries that makes us not to have information enough to provide much reliability on the sources as this is a very low number of analysed softwares to establish a general recommendation.

Thus, it is important to take into account that the recommendations produced on the EU Directive are made on this basis.

14. Data analysis and practical examples

In this chapter a general description and a comparison table between the three softwares is provided. Furthermore a detailed description of each particular software is provided.

5.1 Virtualfires

Introduction

The Virtual Fires (Virtual fire emergency) simulator is presented, that allows training fire fighters in the efficient mitigation of fires in a tunnel and in rescue operations, using a computer generated virtual environment. The simulator can also be used to test the fire safety of a tunnel and the influence of mitigating measures (ventilation, fire suppression etc.) on its fire safety level. The simulator is developed with financial support from the European community under the IST (Information society technology) program and combines the simulation of fires using advanced Computational Fluid Dynamics (CFD) software on parallel computers and the visualization of smoke, toxicity levels and temperature with Virtual Reality.

Background

Past serious fire accidents in tunnels have highlighted the problems that exist with respect to the safety of tunnels and the prevention of serious fatalities in the case of a fire. This required action on a European scale with respect to:

• Ascertaining the safety level of existing tunnels and retrofitted tunnels (i.e. Mont Blanc tunnel). • The specification of the required safety features and installations for new tunnels. • Training of rescue personnel in order to increase the efficiency of fire and smoke mitigation procedures. • Training of drivers with respect to correct behaviour in the case of a fire emergency.

With respect to ascertaining the safety level of existing and retrofitted tunnels much reliance is still placed on real tests using fire/smoke pans or vehicles set on fire. Such a test has been recently performed for example in the refurbished Mont Blanc Tunnel

Description

The disadvantages of real tests are that they are expensive, can only be carried out at certain times and are not environmentally friendly since toxic smoke is produced. In a virtual test the tunnel and the fire emergency only exists in computer memory. Using computational fluid dynamics (CFD) computations, the spread of fire and smoke in a particular tunnel is calculated and then visualized.

The tunnel including the safety installations, traffic signs, vehicles etc. is displayed together with the results of the CFD calculations using the method of virtual reality. Under the term virtual reality we mean total immersion in a three-dimensional data set. The simulator can be used as a training tool for fire fighters and for assessing the safety level of existing or retrofitted tunnels. It may also be used to check the design of a planned tunnel with respect to ventilation, exits, rescue chambers and other equipment.

The simulator can operate in two modes: One where the CFD simulations are carried out prior to the visualization (pre-calculated scenario) and one where the CFD simulations are carried out concurrent with the visualization. The advantage of the second type is that fire mitigation operations can be performed and that ventilation characteristics may be changed during a session and the resulting effect can be seen immediately (i.e. one may check the effect of reversing the ventilation on the spread of smoke and fire). The first type of system can be used for the training of fire fighters and drivers and for checking the fire safety of existing tunnels. In first type of system it is possible to pre-calculate several alternatives and the user can select which alternative to choose when the fire is visualized. This is very useful in an educational situation.

Visualization Hardware

The simulator is implemented in scalable way which means that the visualization can be done on a normal PC, a PC with a head-mounted display (HMD) or in a CAVE environment. A CAVE is a multi-person, room-sized, high-resolution, 3D video and audio environment. Graphics are projected in stereo onto the walls, the floor and the ceiling, and viewed with shutter glasses (see Figure 1). As a viewer wearing a position sensor moves within the display boundaries, the correct perspective and stereo projections of the environment are updated in real time by the rendering system, and the images move with and surround the viewer. Hence stereo projections create 3D images that appear to have a continuous presence both inside and outside the projection room. To the viewer with stereo glasses, the projection screens become transparent and the 3-D image space appears to extend to infinity.

Different display-systems have different advantages and disadvantages. A normal PC screen is something everyone has. The images are also normally of good quality but it may be hard to navigate due to the flat appearance. The HMD is portable and lightweight and requires a moderate computing power for image rendering. Only two different images of the scene need to be generated for a given frame rate. The disadvantage is that the user is not fully immersed into the scene due to the limited field of view. The impression is more like watching the scene through the glasses of a diver. Another disadvantage is the low resolution of the display. The CAVE has a wide field of view because the user is fully immersed into scene and there is a very high resolution of the rendered scene. The disadvantage is that it is a stationary installation and high computing power is needed for rendering. Depending on the number of projection walls 8 to 12 images are required at the given frame rate.

Figure 1 User in a CAVE wearing shutter glasses Software

Architecture

The layout of the software is depicted in Figure 2. At the heart of the system is the CFD simulation software ICE, which uses the Lattice Boltzmann method [1, 2] to compute the air velocity, temperature, pressure and smoke density at a cell point due to a fire. The “smoke density” is an artificial quantity, which varies between 0 and 1. The smoke production is taken to be proportional to the CO and CO2 standardized production curves provided as input. It must be pointed out that smoke is a result of fuel rich combustion and the modeling via standardized curves is only a very crude approximation. However, a real combustion model requires input data, which are normally not available and the calculation is very time consuming.

The storage and retrieval of the calculated CFD-data and the states of all objects that are involved in a simulation-run is handled by the Database Manager module. This component serves as the communication layer between the simulation front end and the database server back end. Currently it transparently supports the MySQL 4.0.15 open source SQL server, but is adaptable to any other SQL server.

The communication between the CFD solver and the storage layer is done by the data manager controller module. This module has been integrated into the Covise VR-environment [3] and also handles all requests from the user interface.

Figure 2 The VirtualFires software architecture

Visualization

Within the project also some new visualization techniques have been developed. This was necessary as the currently available where not sufficient for the system, mainly for 2 reasons:

1. They were too slow to handle the amount of the data produced by the CFD to update the rendering in real time 2. They were not capable of rendering photo realistic fire and smoke

These visualization methods where integrated as plug-ins for the Cover renderer and can be managed from the user interface. The photo realistic rendering of smoke is done by a fast volume rendering approach which takes advantage of the availability of programmable shader functions on modern graphics boards. This way frame rates around 25fps for the volume rendering of the CFD- results are possible on normal PC hardware. To achieve a photo realistic rendering of fire a fractal 3D texture is applied to the regions of the flames. As CFD results are to coarsely spaced compared to the fast visual fluctuations of a flame front, this behavior is interpolated by the fractal texturing process until the availability of the next CFD result. The following scientific visualization methods are currently available on most platforms:

1. Isosurfaces 2. Line integral convolution 3. Streamlines

Figure 3 Temperature isosurface in a tunnel fire

Figure 4 “Realistic” visualization of smoke and fire

User Interface

As there are normally limited capabilities inside a CAVE environment to input text and numbers, a new PDA-based graphical user interface (GUI) has been developed. This GUI allows the user to specify the mission he/she wants to examine, change simulation parameters and restart a simulation. The major advantage of this solution is that this navigation tool can also be used outside the CAVE with the PC-based VR environment without any changes to the simulator, because it is integrated into the network communication layer inside the simulator.

Navigation in space is supported by a space mouse device, by the navigations tools normally used in the Cave environment or by the PDA. Navigation in time is possible by a simple “VCR-like” graphical user interface.

Within this user interface the user can create and define new missions, edit existing ones and start new calculations. The visualization system shows these new results as soon as they are available on the database server.

Examples

For demonstration purposes some popular fire incidents have been calculated with the simulator. These data-sets also serve as a base for the verification of the whole system.

The calculated dataset consists of different ventilation scenarios for the Mt. Blanc tunnel in France and the Gleinalm tunnel in Austria. Both tunnels where examined with their former ventilation system and also with the improved ones after the reopening. Examples of fire simulations are given in Figure 3 and Figure 4.

As it can be seen in figure 7, a fire inside the station on a subway train is visualized in the Cave at PDC. Together with the calculation of smoke spread this kind of simulation is important for the firefighters to plan their missions inside these stations and to verify that their strategies are efficient.

Figure 5 Fire in a Dortmund subway station simulated in the Cave

Conclusions

A simulator was presented which allows firemen to perform virtual training exercises with a PC, a head mounted display or in a CAVE environment. The simulator also allows assessing the fire safety of existing tunnels and can be used as a tool for designing new tunnels.

The Virtual Reality Real Time Fire Emergency Simulator (VIRTUALFIRES) is a simulator that allows the observer to visualise the fire and smoke development and transport inside a space (like a tunnel or aircraft) and walk through the virtual scenario in the same way by means of virtual reality techniques.

The simulator can reproduce the geometry of the scenario (for example, a tunnel with all the elements that it involves, like structural elements, safety installations, etc.) as well as with the results of computer simulations of fires, in order to be used by government authorities, tunnel operators, and rescue personal.

The system is presented in two versions: a CAVE virtual environment, and a PC with a head-mounted display (HMD).

The graphical output is in general better and more inmersive and fire graphical modelling is more realistic. These features together with the simulation in real time make the experience with this simulator more realistic.

Other feature which is very useful is the possibility of obtaining fire simulation data from simulations carried out on different CFD packages.

Software evaluation

Fire modelling is considered as very realistic and accurate compared to a real situation. Models of people escaping are missing in the system. Loss of visibility caused by smoke has been represented, maybe not very accurately. Representing the temperature isosurfaces was very useful for having a better idea of safe areas in the emergency scenario. Real time simulation seems a very useful tool for planning different strategies to be applied in real situations. More extinguishing agents should be modelled. The possibility of using data from different CFD packages makes the system more flexible. The system is highly user friendly in terms of graphical modelling.

5.2 Virtualtraining- Advanced disaster Management systems (ADMS)

Introduction

ADMS™ is an interactive virtual reality-based team training system that provides emergency responders an opportunity to develop skills in command, control, coordination and emergency communication.

ADMS realistically simulates emergency incidents such as aircraft accidents, terrorist acts, hazardous material spills, airfield incursions, fires and natural disasters for the purposes of intra and inter-agency coordination, training, planning, testing and validating emergency plans.

Timely and properly directed rescue efforts can often make the difference between mishap and disaster. Interactive, targeted training is the key to successful incident response. The Advanced Disaster Management Simulator (ADMS™) is a product developed by the Environmental Tectonics Corporation under NIBRA indications.

This state-of-the-art virtual reality training system allows on-scene disaster response team commanders to gain experience and develop maturity that would otherwise take years of high risk, high cost training exercises and actual disasters to develop. ADMS is a high fidelity, fully interactive system that provides first responders an opportunity to develop skills in command, control, mitigation and emergency communication under extremely stressful yet safe conditions.

As said before ADMS can be used to simulate virtually any emergency which are not prescript and play out in real time as fires, injuries, spills and other elements progress according to dynamic and sophisticated algorithms based on real-world physics.

ADMS SmartModel™ technology enables the progression of incidents to be driven solely by the decision making process of trainees. With its industry-exclusive, Scenario Generator™ software, along with multitudes of dynamically interacting elements, ADMS provides infinite training possibilities.

For over a decade, ADMS has been used at US Airports and major firefighting and emergency response training facilities worldwide, and is an industry-proven and mature solution for optimal emergency preparedness.

System description

The Advanced Disaster Management Simulator (ADMS™) is an interactive virtual reality-based team training system that provides emergency responders an opportunity to develop skills in command, control, coordination and emergency communication. Provided to emergency response organizations, airports and government agencies, ADMS realistically simulates various emergency incidents and natural disasters for the purposes of intra and inter-agency coordination, training, planning, testing and validating emergency plans.

ADMS simultaneously trains Incident Commanders along with their Incident Command team members in disaster management skills. ADMS allows trainees to learn, retain and rehearse the four

C's of disaster management: Command, Control, Coordination and Communication while exercising their Incident Command protocols and action plans. ADMS effectively meets training goals by presenting realistic audio and visual representations and immersing trainees in the disaster environment. As students respond to different scenarios by commanding their virtual team, the simulator reacts and generates new cues in real time. Trainees learn by responding to various situational inputs, observing the results of their actions, and reacting to a new set of cues, while gaining the experience of decision-making under stressful conditions.

The flexibility and advanced features built into ADMS have resulted in a commercial, off-the-shelf training system that can be easily customized to present customer specified:

• Emergency Scenario Incidents • Locale Matched (Geospecific) or Locale Representative (Geotypical) Environments • Standard Operating Procedures/Emergency Operating Procedures • ADMS provides positive training by presenting authentic environments, combining high- fidelity 3D databases with people and functional vehicle models powered by ETC's SmartModel™ technology, in combination with dynamic physics-based fire and weather models. 3D environmental sound is incorporated to enhance the situational immersion. By combining these elements with the stress of a realistic, dynamically developing situation, ADMS can facilitate the following types of training: • Incident Command Training • Preparedness Validation • Resource Management • Multi-Agency Coordination Training • Facility & Vehicle Familiarization • Aircraft Rescue Firefighting • Natural, Terrorist & Hazmat Threats

Scenario generator™

An industry unique toolset, Scenario Generator enables users to create any type of disaster scenario and generate situational details such as people, weather, time of day and threats. Because ADMS operates in a truly authentic synthetic world, no two training sessions can ever be the same. Generation of a scenario only presents the initial incident and conditions. Action is driven solely by the decision- making process of the trainee responders, so no results are pre-scripted or "canned". Fires, gases and injuries will progress if left unattended, wind can spread smoke and fire into adjacent structures and cause explosions or hazmat leaks.

Scenario Generator also allows the training instructor to automate teams which are physically absent from the training session, allowing simulated inter-agency coordination to take place. These virtual teams operate by event-based action triggering. For example, when a fire is put out by present Fire/Rescue personnel, an automated EMS team can begin assessing patients.

Configuration

ADMS is available with generic or customized environments, vehicles, people and uniforms, resources and exercise scenarios. Environments can be created to be geospecifically accurate to the users' locale. A typical ADMS system includes one Incident Command Station and four Team Stations. Individual points of view are presented at each training station as students manoeuvre around the scene using joysticks. As would occur at a real incident scene, commands are issued via radio. Facilitators (usually Training Officers or other assigned proctors) then enter the commands into ADMS via a simple, Windows-like menu system. Although the menu system is easy enough to learn in a few minutes, conducting training with facilitators allows the technology to remain virtually transparent.

ADMS can be installed with a 180-degree surround view projection screen or a simple flat screen for the Incident Command Station. Team stations can be set up within canopied hoods, obstructing sharing of extraneous information between stations and forcing realistic radio communication. The 3D visualization is projected on a screen or displayed on a monitor or flat panel TV within each station. The team leader at each station uses a control pad to navigate their way around the scene.

Alternatively, ADMS can be delivered as a portable system which packs into 3 travel cases. In 2006, the complete ADMS system is available on one single laptop computer. This laptop-based system is ideal for training at large facilities or municipalities for sharing amongst multiple agencies. The portable ADMS system is also ideal for use by training institutions.

Training services

Nibra offers on-site training services utilizing ADMS to train and evaluate emergency responders. Nibra works closely with each customer in customizing exercises to ensure that the training program meets training objectives. Based upon Nibra indications, ETC provides the relevant environmental database, as well as the responding vehicles and personnel, and incorporates the Standard Operating Procedures or Emergency Plan of that specific authority.

Nibra provides qualified technical personnel to set up before, during, and after each training session. Nibra also provides a facilitator to assist customer personnel with running each training session. The facilitator's role is to input commands into the system as they are disseminated, allowing the technology to run transparently within the training while the communications process drives the action.

During the evaluation period this training services has not been available for the EBSCC that has installed only a DEMO of the ADMS-Virtualtraining.

Evaluation

Main features analysed of the ADMS-Virtual training software :

- The Nibra ADMS Virtual training system is a visual environment of 450 km2. It consists of two cities, an airport, an industrial estate and woodlands. - The infrastructure contains roads, tunnels, bridges, rivers and railway tracks. - For the handling of incidents hundreds of relief workers and dozens of vehicles are available. - Casualties can receive treatment in one of the four hospitals. - Weather conditions such as wind, rain, mist and snow are realistically reproduced. - Night times, as well as daytime scenarios are available. - Smoke and hazardous materials react with wind direction. - Sound effects are directional and realistic. - The number of scenarios is unlimited (accident in tunnels).

Main advantages :

- You train to achieve control of complex incidents which are difficult or impossible to simulate in reality. - The exercises are extremely realistic: train in real time. - The learning effect in massive and you can see the consequences of all actions. - You can try for diverse scenarios in a short time period. - The scenarios and training environment are easily adapted to your requirements. - You are not dependant upon the availability of other agencies, transport or people. - You can compare the performance of numerous participants: the scenario can be repeated. - You can stop a scenario at any moment for mid-training evaluation.

Applications:

- Mono disciplinary exercises of Fire Brigades, Police and medical services. - Multi disciplinary exercises for operational Command Teams. - Multi disciplinary exercises for Tactical Command Teams. - Multi disciplinary exercises for companies and governmental departments. - Testing scenarios and scripts of practical exercises.

5.3 GAMMA-EC

Gaming And MultiMedia Applications for Environmental Crisis Management Training

Introduction

Disasters do not happen every day. Therefore, disaster management will never be part of the daily routine of the fire brigade, police or health care units. In order to be prepared to these situations as good as possible, however, these organisations exercise regularly. These exercises are mostly limited to the own organisation, for, due to a lack of time, fire brigade, police and health care units do not often succeed in arranging multidisciplinary training programmes. This is due to the pressure of the daily operational process, and to the large amount of time and effort that is needed in order to prepare such an exercise. GAMMA-EC (Gaming And MultiMedia Applications for Environmental Crisis management training) is developed to reduce this amount of time and effort, and to promote education and training.

The GAMMA-EC Project

Context

The goal of the EC project GAMMA-EC is to develop a set of computer programmes to support the education and training of disaster control managers. The results will be a multimedia programme for the individual education of disaster control managers, and an interactive simulation programme or crisis game to support the team training of a disaster management staff. In addition to these tools generic pedagogical guidelines will be developed for designing scenarios and for assessing team member performance in decision making and exchanging information in crisis circumstances. All tools will be developed and validated in cooperation with future end-users (fire-academies, disaster control managers). During the project two applications will be worked out: large-scale forest fires and chemical accidents.

Multimedia

The multimedia application is intended to complete and freshen up the official’s knowledge. In addition to this, the application can be used to apply the acquired knowledge in simple cases (Fig.1). These cases, or scenarios, are the core of the multimedia course. The purpose of the scenario is to let the trainee test its knowledge or browse the knowledge base by trying to solve the problems he faces as he goes through the scenario. Such a training tool enables the official to determine himself the place and moment he wants to get to work with this application. In this way the official is able to prepare well for this team training.

Interactive game simulation

The interactive game simulation supports the training management with the preparation of the exercise (such as making scenarios), as well as with the real performance of the multidisciplinary team training. The interactive game simulation supplies the necessary feedback with regard to the development of the disaster, taking into account the events set in advance (scenarios) and the players’ actions and decisions. The actions and decisions will relate to evacuation, warning the population, combating the cause, etc. During the team training, the officials are in a room resembling the official’s natural surroundings as much as possible (Fig.2). Via the usual lines of communication the report concerning the development of the (simulated) disaster reaches the official.

With this interactive game simulation, skills such as coordination, communication, decision-making under time pressure, and decision- making based on unreliable and incomplete information can be trained. Elements from the exercises can be reused through the generic intention of the game simulation, which considerably reduces the effort for making new exercises. During the game the simulation calculates the consequences of decisions for the further development of the disaster, which makes the output consistent. Because of this, it is expected that the training management required will be reduced with respect to the present exercise method.

An existing exercise can be used several times to train different teams. For example, the exercise developed during the project will be available for fire academies willing to have a training session.

The game simulation tools also facilitate the creation of new exercises: it includes an exercise definition module.

Both the multimedia application and the interactive game simulation are developed from an instructional viewpoint.

Background

An emergency situation is characterised by the occurrence of a large-scale accident making it necessary to save people, to clear areas, to protect properties, etc. Due to the large-scale character of an emergency situation, many authorities are involved in controlling both the cause and the consequences of the emergency. The activities of the operational services of the fire-brigade, police and health-care units will have to be adequately co-ordinated in order to work quickly and efficiently. The larger the scale of the emergency situation, the more municipal, regional, and national services will be involved.

To be able to operate in an effective and efficient way the various organisations should be well- prepared. According to some studies (Netherlands Ministry of Interior (December 1995) and European Commission, DG-XII (June 1996)) it turned out, however, that the middle and higher echelons of these organisations are hardly or poorly trained for crisis circumstances. One of the main reasons is that most disaster managers have other jobs and activities. Therefore, they cannot or even do not want to spend much time on crisis management education and training, although they are aware of its importance. Whenever emergency managers train, they mostly do this by ‘exercises on paper’ or by role-playing using scenarios.

These types of training have some important draw-backs, e.g.:

• They need a lot of preparation time; so scenarios are hardly updated; • scenarios very often are situated in another area; so the emergency management staff does not train in its own area with its specific characteristics while training in the staff’s own area would be more stimulating and would increase the emergency preparedness; • scenarios are more or less pre-defined; if the disaster management staff decides to take a decision that was not foreseen by the training staff in advance, such a decision can normally not be effectuated during the exercise because it does not fit into the scenario.

In the past the military world faced similar problems with higher echelon preparation. Defence solved this problem successfully by introducing all kinds of computer supported education and training tools. Some examples are simulators used for task training and war-games used in so-called command post exercises. TNO has developed several of such systems. The basic idea of the project GAMMA-EC is to re-use the military experience in the domain of disaster management. The project aims to show that computer support in education and training is feasible and efficient in this domain too.

Project information

GAMMA-EC started in August 1998, and finished in the second half of 2000. The goal of GAMMA- EC is to develop a set of computer programmes which enable education and training of Disaster Management Staffs (DMS) in decision making, co-ordination and communication in emergency situations.

Therefore, the GAMMA-EC solution exists of: • individual education of DMS members by using multimedia tools; • DMS team training by making use of interactive simulation tools (crisis games).

Two applications of both the multimedia and the game will be worked out. These applications represent a natural and an industrial emergency situation:

• large-scale forest fires which is an important issue for the European Commission (DG XII) and especially for the Mediterranean countries;

• chemical accidents which cause a high risk for industrialised areas (e.g. Rotterdam harbour).

GAMMA-EC is developed by a consortium consisting of: • TNO (The Hague): crisis game development, chemical accident modelling, determination of guidelines for team training

• University of Barcelona: determination of educational guidelines for the multimedia and the crisis game

• Italsoft (Rome): multimedia development

• MTDA (Aix-en-Provence): forest fire modelling, determination of the contents for forest fire courseware

• fire-academies from The Netherlands (NIBRA), Catalunya (EBC), Toscany (UOC Foreste) and France (CEREN): knowledge of disaster management instruction, provision of domain expertise, representation of future users.

Target group

Disaster Management Staffs can be formed at different organisational levels. For reasons stated before the intermediate and higher echelons have a real need for training. The primary target group of GAMMA-EC consists of the intermediate echelons of the operational services who will act in an emergency situation, such as the fire-brigade, the police and the medical services. The highest echelons (ministers, mayors of big cities, etc.), however, will not be considered in the project. This because it is assumed that computer assisted training is less effective at these levels.

The intermediate levels that have been defined are:

• Regional Crisis Unit: a command centre at tactical level, mostly dislocated from the accident (in a regional command centre or a townhall);

• Field Command Centre: a command centre at operational level, mostly situated nearby the accident.

At both levels, hereafter called DMS, decisions have to be made in a short period of time, while the activities of the different operational services have to be adequately co-ordinated in order to work quickly and efficiently. So, on the one hand these commanders should be able to take decisions within their own organisations; e.g. a fire-chief should know how to extinguish a fire at a chemical plant. On the other hand they should be able to communicate important information (e.g. inform about the dangers of the fire) and to co-ordinate actions with their DMS colleagues (e.g. set-up an evacuation plan together with police and health-units).

The ability to carry out the abovementioned activities is in fact the main learning goal of the education and training of a DMS. The GAMMA-EC tools should facilitate realising this learning goal.

Therefore, a constructivist approach has been chosen.

…

Regional Crisis Unit Target Groups

Field Command Centre

Fire Police Medical Brigade Services …

Constructivism

The basic idea of constructivism is that the best way of learning is reached when the student is active himself (this in contrary to the traditional idea of education by knowledge transfer).

From a constructivist point of view the learning environment should:

• support multiple representation modes • enable the student to actively build up his knowledge from experiences • provide experience in, and appreciation for multiple perspectives • encourage ownership and responsibility for the learning process • embed learning into a realistic and relevant context • encourage self-awareness of the knowledge construction process • embed learning in a social experience (i.e. in contact with other students) • provide learners with the reasons for learning within the learning activity • encourage intentional learning and examination of errors.

In both the design and the development of the multimedia and the game these aspects will be taken into account as much as possible.

Multimedia programme

The problem of emergency managers having not enough time to learn or to follow lectures at a training location, can be solved by providing them with the possibility to individually prepare at any time or location by so-called distancelearning.

With help of the multimedia applications the DMS members will be instructed individually, and will be supported to learn their emergency tasks and organisations and to refresh their domain knowledge.

The multimedia programme contains several levels, varying from novice up to experienced. When defining a course, at each level the teacher or author is able to define a set of scenarios with pictures, questions, remarks, etc. Furthermore, he builds up a knowledge base with information (texts, pictures,

videos, etc.) about relevant topics on disaster management. The set of scenarios will be the core of the multimedia course.

Mistakes while solving scenarios should reflect knowledge gaps. Not optimal performance in the scenario should guide the student towards target positions in the knowledge base.

By using the multimedia tools the students should play an active role in their learning process.

The basic idea of introducing levels, scenarios, quizzes and scores in the multimedia course, is that these will motivate the students to continue. Furthermore, the use of scenarios and questions offers the opportunity to see whether there exist gaps in the students’ knowledge. Special attention can be put on specific topics, and thus making the course more personal and thus more attractive.

Crisis game

With help of the GAMMA-EC crisis game the DMS can be trained as a team. The crisis game is an interactive networked simulation programme that runs on PCs. It supports the training staff both during the preparation and during the execution of a DMS exercise.

In the traditional way of scenario building the course of the exercise has to be worked out into detail. All possible decisions and their consequences have to be looked at, messages have to be edited and have to be distributed amongst the training staff members, etc. So, setting up a scenario is a time-consuming job. For this reason there is an enormous lack of good training scenarios. With help of the crisis game it will be much easier to build a new scenario. Because GAMMA-EC will contain models for standard actions (e.g. move units from A to B, dispersion of a gas cloud with its consequences) not all effects have to be calculated in advance.

Besides, a lot of messages can be generated during the exercise by the system itself, which will save quite some efforts from the scenario-builder.

GAMMA-EC contains a scenario editor. Writing a scenario will consist of editing the script of the exercise and filling in the so-called event list with characteristic events (e.g. an accident, a message about a gas leakage, etc.).

Furthermore, it will be possible to re-use events from earlier defined scenarios or to apply them in other geographical areas. During an exercise the DMS will be located in its natural environment e.g. like one or more crisis rooms in a regional alarm centre or town-hall. The DMS members, who are the players of the game, should be able to use the same tools they would use during a real emergency situation as much as possible, such as maps, information systems and decisions support tools. Also communication from the DMS members with the outside world concerning decisions, orders, information reports, etc. will take place by the standard means such as phone and fax.

Player room(s) Staff room

Trainee 1 + Information Training Staff Observer requested or Command Manual Interpretation Posibility separated

Trainee 2 Inside game + Model? Observer N Information or situation Y Posibility separated uptdate

Trainee 3 Simalation model + (N users) Observer

The training staff will use the crisis game to determine the feedback they provide to the DMS members about the development of the disaster, taking into account the scenario, the players’ actions and their decisions. In the current, traditional way of training this is a very complicated and time- consuming task for the training staff. It is therefore expected that the GAMMA-EC’s interactive simulations will be of great help in future exercises, and that they will enable the training staff to focus more on the decision making processes and on the DMS team behaviour. In combination with the crisis game’s evaluation facilities, training sessions can be evaluated in a better and more structured way, and consequently will improve the quality of training sessions.

Conclusion

Within the GAMMA-EC project a new learning environment for education and training of disaster control managers has been built. It has been proved that GAMMA-EC tools can support the education and training of disaster control managers in an effective and efficient way, and that the use of these tools will lead to a better protection of the population and of the environment in the future.

The GAMMA-EC simulator is made up of two computer programmes:

- a multimedia programme for individual training - a programme for interactive simulation for team training

The aim is to improve the officers’ decision-making process in case of emergency (for instance: forest fires, incidents in petrochemical plants, tunnel fires.....)

The software develops in a windows environment and it does not need an advanced computer equipment.

Practical exercise:

This exercise took place in the EBSCC in order to evaluate the software with a team composed by different agents involved in the safety chain.

The ESBCC uses GAMMA-EC for officers training either being middle or high ranks . A working team is called periodically to put the simulator into practice in the field of tunnel intervention. All members have a good knowledge of the software and have got detailed information on the installations for the drill tunnels.

The simulator scenario is carried out in two adjacent rooms.

ROOM 1: It is made up of a fire-fighter, a policeman and a medical auxiliary. They are located in the Advanced Commandment Centre (CMA in Spanish), which is next to the tunnel and from which the strategies and the decision-making will be implemented according to the information requested and received through radio stations

ROOM 2: It is made up of fire-fighters, policemen, media, political representatives, etc..They have got a radio station to communicate with the CMA and a computer which is connected to the Intranet, which manages the practice.

In this room, the scenario is being modified according to the decisions that are being made at the CMA. This new situation simulated by the computer is mutually communicated to the CMA.

During the practice two observers take note of the relevant information regarding its development. This piece of information will be used to foster the participants’ feedback at the explanation and reflection sessions.

As the simulation takes place in real time (a tunnel fire lasts 3 hours approximately), it is split into two parts:

FIRST PART: Practice start, gathering of information, first decisions-making and immediate action.

SECOND PART: Before starting this part, the participants in both rooms get together to make a first evaluation. At the end of the simulation an evaluation and a value judgement are made and they comment the relevant aspects and mistakes made.

The result of the multimedia programme adapts to the decisions and applied methodology according to the answer given by the groups taking part in the emergency event. The resulting methodology for intervention will become part of the basic criteria for the training of emergency teams.

5.4 Fire simulators comparison

VS VS

Main differences between Virtualfires, Gamma-EC and Virtualtraining

First of all, it must be remarked that the three simulators pursue the same aim, although they follow different methodologies.

VIRTUALFIRES gives a three-dimensional interface which enables to monitor in real time what would happen inside the tunnel in case of fire. That is why it can be considered as a training simulator for operational working units: it enables them to see how fire, temperature and smoke vary in each particular case.

This is the reason why it requires an advanced hardware and also a three-dimensional vision system.

VIRTUALTRAINING- ADMS provides a wide visual environment including different situations and physical areas such as cities, airports, industrial areas, etc with its respective infrastructures (tunnels, bridges, rivers and railway tracks). Therefore it permits to have a more complete visualisation of how to act on different situations. As there are availability of hundreds of relief workers and dozens of vehicles are available for the handling of incidents it is more useful for planners or public organisations in charge of the general coordination of disasters and complex incidents.

GAMMA –EC manages the main variables that determine a disaster. The resulting output varies according to the decisions and commands given to the simulator. This way, it can be considered as a simulator directed to train the high-ranking officers and the executive teams who manage the disaster.

The table below shows the three computer tools in the first row. Vertically, the derived topics for EU guidelines are listed and the cells contains 'X' indicating which topic is derived form which tool. The guidelines will be described in chapter 6.

Topic Virtual fires Virtualtraining-ADMS GAMMA EC Real time X X All response fire all All organisations Different Scenarios X X Levels of personnel operational tactical Strategic involved Realistic models X X X Team training X Availability

These are the various characteristics that have been taken into account in each topic per softare:

Real Time

• Graphical modelling, with 3D video and audio environment • Graphics projection in stereo onto the wall. • Possibility of planning different strategies to be applied in real situations • Respond to different scenarios by commanding the virtual team • Generation of new cues in real time • Projections of the environment updated in real time • Computation of air velocity, temperature, pressure and smoke density • Action driven by the decision making process of the trainee responders • Event Based action triggering • Creation of a geospecifically accurate to the user’s locale • Allows the reproduction of the geometry of the different scenarios • On-scene disaster response • Meet training goals by presenting realistic audio and visual representations • Possibility of issuing the commands via radio • Casualties can receive treatment

All response organisations

• Training simulator for operational working units • Simulation for one user • Parallel simulation for several users • Built into a more commercial, off-the-shelf training system that can be easily customized • Training for Fire fighter, police and health staff management

Different scenarios

• Reproduction of the geometry of the scenario with the tunnel itself and all the elements involved inside.

• Creation of any type of emergency scenarios • Generation of situational details such as people, weather, time of the day and threats. • Natural, Terrorist & Hazmat Threats • Aircraft Rescue Fire fighting • The number of scenarios is unlimited • The availability of trying for different scenarios in a short period of time. • Facilitate the creation of new scenario/exercises • Re-use events from earlier defined scenarios. • Simulation for all kind of accidents

Levels of personnel involved

• Different existing agents modelled in the program.

Realistic model

• Combined with any virtual reality method. • Includes CFD simulations carried out prior to visualization • Includes CFD simulations carried out concurrent with the visualization. • Possibility of obtaining fire simulation data from simulations carried out on different CFD packages. • Combination with dynamic physics-based fire and weather models • Weather conditions are realistically reproduced • Smoke and hazardous materials react with the wind direction.

Team Training

• Provision of infinite training possibilities. • Training of incident Commanders along with their Incident command team members • Multi-agency coordination training • Simulation of inter-agency coordination to take place. • Possibility of training at large facilities or municipalities • Be able to stop the scenario at any moment for mid-training evaluation. • Training for different organisations such as fire brigades, police, medical services, operational command Teams, Tactical Command Teams, companies and governmental departments • The players of the game have the possibility of using the same tools (map, information systems and decisions support tools) they would use during a real emergency situations • Provides an opportunity to develop different skills for the emergency responders

Availability

• Not high computing power needed for rendering • Visualization in a flat appearance, with limited field of view • PDA-based graphical user interface (GUI) • Possibility of changing simulation parameters and restarting simulation • VCR or alike graphical user interface

15. Proposal for EU guidelines

As mentioned within the text these kind of software are very useful for training rescue personnel and disaster management organisations, with the objective of increasing the effectiveness of rescue operations. Nevertheless, it has to be mentioned that each software is dedicated to one purpose and therefore its use has to be indicated for the correct one and focused to the appropriate personnel.

Based on the results of the research done in this document, some recommendations for computer supporter training tools are listed below.

1. Member States should use computer supported training tools for training rescue personnel and management operations for tunnel accidents

2. Member States should use Computer supported training tools that are real time

3. Member states should use Computer supported training tools that consider all tunnel disaster response organisations

4. Member states should use Computer supported training tools that are based on scenario's that for the particular tunnel has been incorporated in the contingency plans.

5. Member states should use Computer supported training tools that consider all different level of personnel involved in mitigating the effects of tunnel accidents.

6. Member states should use Computer supported training tools that integrate realistic models (CFD computations, communication, ..)

7. Member states should use Computer supported training tools that can be applied to any computing power environments

16. Limitations

As it has been mentioned before, further analysis of the evaluation process is needed to bring it up to satisfy the required level regarding to the provide recommendations for the EU Directive.

The evaluation of the different softwares is quite limited in the sense that each software is focused on different safety teams and need to be chosen by the authority responsible for the rescue operations of the tunnels and the safety and of the personnel using the tunnel. In this case the softwares have been evaluated by one single team of the EBSCC without previous experience in these kind of softwares.

17. Recommendations

For future work a more detailed analysis taking into account cost effectiveness, user acceptance, availability and support should be realized in order to have more feasible criteria for evaluation.

In order to make a more accurate evaluation the following aspects should be taken into account:

- If softwares would have been evaluated by a wider user group involving different ranges and types of security services, the results would have been more accurate, if possible.

- More time, resources and more detailed information on the softwares here presented and on others similar ones should be provided.

Nevertheless, at this stage of the project we consider that SafeT network has done a very important and useful work which will result in an improvement of safety in tunnels around Europe. We consider that SafeT network or a similar consortium where different agents of all parts of the safety chain and international organisations discuss together the best practices and new technological developments and adaptations to establish recommendations should be kept in order to update the information here presented which is continuously changing. EU Directive should be a “live” document opened to changes.

18. References

- Instruction Manual GAMMA-EC Player Module, short version TNO-FEL (The Netherlands), November 2000 and other related documentation from the project.

- www.virtualfires.org

- Virtual training documentation and demo provided by NIBRA

- Implementing a CFD steering system for immersive environments. Kai-Mikael Jaa-Aro. Department of Numerical Analysis and Computer Science. Royal Institute of Technology. http://eve.hut.fi/cavews2003/CFD-steering.pdf

- GAMMA-EC:Gaming And MultiMedia Applications for Environmental Crisis Management Training (IFFN No. 23 - December 2000,p. 96-99) www.fire.uni- freiburg.de/iffn/country/fr/fr_4.htm - 9k

- VirtualFires a Virtual Reality Simulator for Tunnel Fires Gernot Beer1, Thomas Reichl1, Gunther Lenz1 and Gert Svensson2 1 Institute for Structural Analysis, Graz University of Technology, Austria 2 Center for Parallel Computers, Royal Institute of Technology (KTH), Stockholm. www.pdc.kth.se/~gert/VirtualFires3.doc

Task 3.3: Cross border incident management (Nibra)

Executive Summary

The report gives an overview of the aspects involved to cross border intervention and evacuation management. Literature, cases and tunnel specific procedures have been studied to this end. As a results, so far, 14 guidelines have been proposed to manage cross border evacuation management.

Contents WP 3.3

1. Abstract...... 54

2. Objectives ...... 56 2.1 Problem statement...... 56 2.2 Objective ...... 56

3. Introduction ...... 57 3.1 Background ...... 57 3.2 Outline...... 60

4. Data collection ...... 61

5. Data analysis...... 62 5.1 Literature...... 62 5.2 Case studies: Eurotunnel and Mont Blanc Tunnel 63 5.3 Procedures in some TERN tunnels...... 64

6. Practical examples ...... 69

7. Proposals for EU guidelines...... 70

8. Limitations...... 72

9. Recommendations...... 73

10. References...... 74

Appendix 1: Cross-border tunnels in Europe...... 76

Appendix 2: Cross-border incident management aspects...... 81

19. Abstract

Cross-border tunnels are tunnels that cross national borders or borders between autonomous regions within countries. The fact that these tunnels cross a border has several implications for incident management. The Eurotunnel fire and the Mont Blanc tunnel fire indicated that cross-border aspects hampered the interventions by the various national emergency response teams. The EU (2004) concluded:

"Insufficient co-ordination has been identified as a contributory factor to accidents in trans-boundary tunnels. Moreover, recent accidents show that non-native users are at greater risk of becoming a victim in an accident, due to the lack of harmonisation of safety information, communication and equipment".

Based upon literature study, case studies and analyzing existing procedures in cross-border tunnels, we proposed 14 guidelines to deal with cross-border aspects in intervention management. Some of the guidelines are specific for cross-border tunnels between countries with different languages (these are indicated with *).

1. the involved countries that are crossed by the tunnel should make appointments with regard to one sovereignty

2. at each Trans European Road Network tunnel (TERN-tunnel) and with regard to intervention aspects, tunnel users should be informed in English and language(s) of the two countries that are crossed by the tunnel

3. for all TERN-tunnel, the (intervention) information to the users should be provided using the same information format (harmonized)

4. for each TERN-tunnel, the contingency plans should be prepared in the language(s) of the countries that are crossed by the tunnel*

5. for a whole TERN-tunnel, there should be appointed one single primary official who is in the lead with regard to the emergency response intervention, including communication to the tunnel users, exploitation of tunnel installations, and emergency response tactics

6. communication between the responsible officials for the intervention management, should be in a predefined language*

7. countries' communication systems in use during intervention management should be compatible

8. the organisational structure of the emergency response teams (administrative as well as operational) should be matched between the two countries that are crossed by the tunnel

9. the emergency response teams in both countries that are crossed by the tunnel should use the same intervention scenarios

10. emergency response resources (material and equipment) should be useable in both countries that are crossed by the tunnel

11. emergency response training exercises should be held involving both countries that are crossed by the tunnel

12. in each country that is crossed by the tunnel, a casualty centre should be activated in case of an emergency. The coordination should be appointed to one single organisation

13. one single insurance procedure must cover the complete tunnel

14. actual emergencies in tunnels should be investigated involving parties of both countries that are crossed by the tunnel

These guidelines should be the basis for individual tunnels, which means that per tunnel, these recommendations should be specified according to specific tunnel and regional characteristics. For example, for the Mont Blanc tunnel, guidelines 2, 4 and 6 could be further specified as that tunnel users should be informed in English, French and Italian, that contingency plans should be available in French and Italian, and that for example French is the language to be used for communication between the Italian and French fire commander.

20. Objectives

1.3 Problem statement

Several large-scale tunnel accidents, such as the Mont Blanc and the Eurotunnel made clear that accidents in tunnels might have peculiar aspects in situation in which more countries are involved. Remember for example the language problem between the French and Italians respectively the differing administrative levels involved between the French and the British. However, apart from such incidental cases, poor structural knowledge is available to deal with cross-border aspects with regard to tunnels. As a result, we defined the following problem statement for this research1:

There is a lack of structural knowledge with regard to cross-border incident management in tunnels.

1.4 Objective The SafeT final proposal (2002) stated the objective of work package 3.3 as formulated below:

To evaluate tools to improve the efficiency of cross-border management procedures for evacuation and intervention and to make recommendations on its use in order to streamline tunnel emergency management procedures.

The EU (2004) concluded in the minimum safety requirements for tunnels in the European road Network that

To realize the above-described objective means that tunnel operators, emergency responders and other stakeholders involved in tunnel safety are provided with guidelines with regard how to translate the EU conclusion into practical solutions.

1 A brief survey of tunnels in the Netherlands showed that apart from the cross-border aspect, evacuation plans for tunnel have hardly been developed.

21. Introduction

1.5 Background

In Europe, an increasing number of road tunnels are realized. In 2002, 398 Trans European Network (TEN) tunnels existed, and between 2002 and 2010, additional 114 TEN road tunnels will be operated (EU, 2002). Several of these tunnels cross national borders, such as the Mont Blanc (France-Italy) or the Somport (Spain-France). In particular the cross-border aspect might hamper the evacuation of people in tunnels in case of an accident for example caused by language problems.

Several key terms need to be further specified.

Cross-border or trans-boundary:

A cross-border is any situation that cross national borders or borders between autonomous regions within countries, such as the Bundesländer in Germany, the Cantons in Switzerland, the counties in Norway the departments in France or provinces in Greece, Spain and the Netherlands.

border

country B country A

highway tunnel

Figure 1: Cross-border tunnel.

This definition of cross-border is broader than only trans-boundary situations because we argue that cross-border situations within countries but crossing border between autonomous regions, might

involve typical problems with regard to legislation and administrative levels during the emergency management.

Evacuation:

Evacuation is the process of an individual (in the tunnel) reaching a safe area (Persson, 2002; Boer 2002).

The evacuation process can be divided into several phases (Persson, 2002; Ashe and Hall, 1999):

• Awareness or wake up time2: the time for an individual to become aware of a threat. According to Boer (2002), this time can vary from several seconds to several minutes.

• reaction time3: the time, once the individual has realized that there is a potential threat, to make an action, normally to make a decision. Persson (2002) notes that this period will be relatively short.

• movement time: the time spent in direct movement towards a safe area (according to NFPA, 1997, the movement speed could vary between 0,7 m/s and 1,2 m/s) and depends upon the distance to a safe area ,

• rescue time: the time it takes emergency responders to move the individual to a safe area in case an individual cannot rescue him/herself.

In general, two types of evacuations are distinguished (Asch and Hall, 1999; Rhodes, 1999)

• self evacuation: a person reaches a safe area without the intervention of the tunnel operator or emergency response teams

• organised evacuation: a person reaches a safe area with the help of the tunnel operator or emergency response teams.

To reach a safe area, a person in the tunnel needs to flee/escape. To this end, several provisions usually have already been incorporated in the tunnel infrastructure/geometry (Ashe and Hall, 1999) such as: escape routes, cross passages, emergency lighting, and emergency signs4. This research does not focus on these aspects. Other provisions will be activated by the tunnel operator or the emergency responders such as the public address system, ventilation, or sound markers. Finally, it could be necessary for emergency responders (fire fighters) to enter the endangered tunnel tube to assist in the evacuation5. Victims might need to be cut out of their cars or need to be guided to the safe areas.

2 Also called recognition time or wake-up time: the time between the accident and the moment a victim realizes a dangerous situation develops. 3 Also called response time: time to prepare to evacuate, for example to grab one's belongings or to inform people. 4 Boer (2002) conducted large-scale fire evacuation tests in the 2nd Benelux tunnel in the Netherlands. His research made elements of the escape routes such as escape doors, cross passages and emergency signs need to be highly visible and uni-interpretable. 5 Nibra (2002) and Rosmuller and Van den Brand (2003) argue that entering the endangered tunnel tube involves serious risk for the firefighters themselves.

To summarize, we have three types of main actors involved in cross-border incident management:

• the individual in the tunnel6

• the tunnel operator at both sides of the border

• the emergency response organisations at both sides of the border

The figure below schematizes the main actors involved in the evacuation process, including the person(s) in the tunnel, the tunnel operators at both sides of the border, and the emergency responders at both sides of the border.7

border

e.g. country B e.g. country A = persons in tunnel

= tunnel operator

highway = emergency tunnel responders

= accident

Figure 3: Cross-border incident management

Several situations might develop as a result of an accident and depending the level of involvement in the accident, the accident/threat awareness and individual actions of car drivers:

• involvement in the accident

First, the car drivers in the tunnel might be involved in the accident or not. If the driver is involved, he might be stucked in his/her car, and hence (s)he needs help from emergency responders or other car drivers to evacuate. If the driver is not stucked, (s)he might run away from the accident spot or help other victims.

6 According to Boer (2002) we are aware of the fact that individuals might follow the behavior of others in the tunnel, and imitate this behavior. As a result, group behavior develops instead of all kind of individual actions. 7 The tunnel lay out is not relevant in this figure. This tunnel is schematized involving bidirectional traffic. For the same purpose, two unidirectional tunnel tubes could have been used.

• awareness of the accident

If the driver is not involved, (s)he might be aware of the accident, or not. If the driver is aware (s) he can run away to a safe area, stay in the car or help other victims. If the driver is not aware of the accident (s)he will stay in the car and probably wait for the moment (s) he is able to continue the trip

• individual action

Among others, depending the awareness of the accident, an individual has the opportunity to run away, help or stay in the car.

These possibilities a car driver has are relevant with regard to the actions of the tunnel operator and the emergency responders. When a car driver is stucked in the car, emergency response will be necessary to free this person. In case no one is stucked, in particular the tunnel operator has a role to assess the situation and if necessary, to direct the people in the tunnel to the safe areas. To this end, radio broadcasting and public address systems could be useful, as well as messages on route information signs. If car drivers stay in their cars or cannot find the emergency exits, emergency response activities in the tunnel might be necessary to evacuate the people in the tunnel.

Self-evacuation can start directly following the accident. Intervention by the tunnel operators might take a small period because the tunnel operator has to be aware of the accident and has to decide what will be the appropriate action. Intervention by emergency responders in the tunnel will take the longest period, caused by the transportation time from the emergency response turn out point to the tunnel, and in the tunnel to the right cross passage.

In particular the coordination and communication between these actors should be prepared/managed adequately. It is exactly the communication process between the person in the tunnel, the tunnel operators and emergency responders in cross-border situations that need to be prepared/managed adequately to facilitate the self-evacuation from the tunnel, or the organised evacuation.

1.6 Outline

In chapter 4, the research approach for collecting data is described, so the reader gets insights in the way the results of this research can be reproduced. In the same chapter, we delineate several of the key terms in this research, such as cross-border. In chapter 5, the cross-border aspects that have been found by literature study, case studies and analysis of cross-border tunnel incident management procedures are presented and analysed. Chapter 7 presents proposals for guidelines with regard to cross-border incident management in tunnels. Limitations of the research and recommendations are presented in chapter 8, respectively 9.

This report describes the aspects that need to be prepared in order to facilitate adequate incident management in cross-border tunnels. The focus is on road tunnels, however, many of these aspects might count for rail tunnels as well.

22. Data collection

In order to develop structural knowledge with regard to cross-border incident management for tunnel situations, we might be able draw lessons from cross-border emergency management in general. Hence, apart from tunnels and evacuation, cross-border emergency response activities might give clues for aspects that need to be considered for tunnel situations as well8. Hence, literature and the Internet (see appendix 2 for a more detailed insights in the performed search) will be searched for clues with regard to cross-border emergency response activities, such as for example cross-border wood fire or earth quakes. Subsequently, the following research questions need to be answered:

1) What cross-border lessons can be learned from cross-border tunnel accidents?

Several large-scale cross-border tunnel accidents are selected, being the Mont Blanc inferno (1999) and the Eurotunnel accident. The investigations reports and the recommendations will be studied, in particular searching for cross-border evacuation aspects.

2) What procedures have been prepared by tunnel operators and emergency response organisations involved in cross-border tunnel situations?

To generate this information, an overview of cross-border tunnels in Europe will be developed. Subsequently, we will select several tunnels for a closer view on the cross-border evacuation aspect. Selection criteria that will be used are the scale of the road (national roads/highway), (inter)national (involving multiple countries), and trans-boundary (involving multiple autonomous regions within a single country). For the selected tunnels, a study after the evacuation management procedures will be conducted.

3) What guidelines could be provided with regard to cross-border incident management?

Based upon the answering the above-formulated questions, several guidelines for cross-border incident management are proposed. These proposals have been discussed tunnel safety experts who participated in SafeT.

As a result, a three-way research approach is followed, that is visualized in the figure below.

literature cases procedures

guidelines for cross border evacuation management

Figure 2: Research approach.

8 Conform Ashe and Hall (1997) and Boer (2002) we recognize that evacuation from a tunnel is different from an 'ordinary' building, due to for example the facts that victims have to leave their car behind or have never been experienced to a tunnel evacuation.

23. Data analysis

Analogue the three defined research questions, in this chapter we will present the retrieved data by literature study, case study and procedures and conduct the analysis.

1.7 Literature

Cross-border evacuation management itself is not specific for tunnels. Other application fields were found, such as natural disaster management (e.g. bush fires, floods, earth quakes).

The issues for cross-border evacuation as presented here are derived from documents that in majority are not tunnel related. However, the issues presented here are applicable to tunnels as well. Cross- border issues that are not relevant for tunnels are left out.

In the spring and summer of 2004, we searched the journal of Tunnelling and underground space technology and the Internet. The former did not generate relevant information for this research. The Internet was searched in March 2004 (Google) based on a number of search strings as depicted in appendix 2..

The issues found could be presented by phase of the so-called safety chain: proaction, prevention, preparation, suppression (intervention), post-event care (BZK, 1993)9. The table below summarizes the aspects that have been identified for which attention has to be paid in cross-border evacuation tunnel situations. In the first row, we depicted the phase of the safety chain. Below each of these phases, we listed the relevant cross-border aspect. In this table, in the column preparation, attention should be paid to information and communication. A specific subject in this category is the language in which actors communicate.

Table 1: Cross-border incident management aspects. Preparation Suppression post event care administrative arrangements10 scenario prediction casualty information organizational structure information and accident investigations communication operational procedures on-scene actions juridical investigations Planning technical measures insurance procedures Resources securing possessions of victims Exercises information and communication technical measures (other than inf. and comm.)

9 However, because the focus is on reacting to accidents (evacuation), the phases proaction and prevention are left out here. Proaction and prevention deal with causes of an accident. Prevention could for example involve escape route or cross passages.

10 See for example the bi-national agreement between The Netherlands and Belgium in the Brabant border area with regard to cross-border emergency response protocol.

Appendix 2 presents in more detail the cross-border aspects. The identified aspects were formulated as either points of attention or practiced measures according to the documents. Because the legal, organizational and institutional context of emergency services and agencies concerned with a tunnel may vary considerably for different tunnels, the issues are formulated in a manner that is independent of individual tunnel characteristics.

1.8 Case studies: Eurotunnel and Mont Blanc Tunnel

In this paragraph, we will summarize the main cross-border evacuation lessons that have been learned from two major cross-border tunnel accidents: the Eurotunnel (1996) and the Mont Blanc Tunnel (1999).

Eurotunnel fire (November 18th, 1996)

In November 18th, 1996 a severe fire occurred in the 50-kilometre Eurorailtunnel between the united Kingdom and France. One of the trucks on the train caught fire before the train entered the tunnel. The train was one of the Eurotunnel's lorry shuttles with the truck drivers in a separate car next to the locomotive. There were no fatalities, 2 persons got injured. The damage to the tunnel was huge. Persons in the staff car and the truck drivers managed to evacuate through the neighboring door leading to the parallel service tunnel. After 29 hours, the fire was under full control (Desfray and Beech, 1997).

Cross border lessons from this incident are:

• As the two nations on both sides of the tunnel speak different languages, it is necessary to take care of good coordination and communication. It is of paramount importance/essential to have the fire fighters both from the first line and from the second line of response (FLOR and SLOR) train together so that they're familiar with each other's equipment and tactics.

• Now, exercises between the FLOR units and external units take place regularly.

• Members of both the French and the English fire brigades now patrol the service tunnel in order to reduce response time to an incident.

• Communication was one of the major problems. Emergency personnel now are provided with a special handset, which they can plug into the landline when they arrive on the scene in the service tunnel.

• Radio frequency and the channels to use for communication are standardized for the French and the English side.

• A strict accountability system has been introduced in order to cope with the former personnel accountability. Both in Calais and in Folkestone are control boards.

Mont Blanc tunnel fire (March, 24th, 1999) March 24th, 1999 a severe fire took place in the 11,6 km long Mont Blanc road tunnel between France and Italy. A lorry with flour and margarine got on fire. After several days, the fire was extinguished. In the end 39 fatalities were counted (including a fire fighter , tens of vehicles were destroyed and the tunnel was heavily damaged. It took about 55 hours to gain full control over the fire. The tunnel was reopened in 2002, march, 9th (Duffé, M. and P. Marec, 1999).

Cross border lessons learned from this incident:

• The installation of a unique control centre. Now, there is a main command post on the French side and a back-up command post at the Italian side, manned around the clock.

• First rescue services at both portals are coordinated.

• It is recommended to have one operating company in charge, even if there are more lessees, so the operating and investment policy is valid for the tunnel as a whole.

• Good communication inside the tunnel is indispensable during a crisis.

• Regular drills in which participate both nationalities have to be held.

1.9 Procedures in some TERN tunnels

Cross-border tunnels are not new. These tunnels already exist and it might be that tunnel operators and emergency responders have already prepared several cross-border aspects in order to facilitate evacuations from people that have become part of a tunnel accident11. Therefore, we searched several evacuation management procedures of existing cross-border tunnels.

To get the information of cross-border cooperation in general and with regard to cross-border incident management in particular, we followed three strategies:

• the SafeT-partners where asked (minutes, action A-37,) in the Vienna meeting (June 2004) and in August 2004 to submit tunnel specific guidelines or national regulations (SafeT, 2004)

• in December 2004, several tunnel experts in France, Slovenia, Sweden, Austria and the UK where asked by e-mail to submit tunnel specific guidelines or national regulations

• in December 2004, the helpdesk of the tunnels as mentioned in table where asked by e-mail to submit tunnel specific guidelines or national regulations

11 A brief survey of tunnels in the Netherlands showed that apart from the cross-border aspect, evacuation plans for tunnel have hardly been developed.

The table below lists the tunnels for which evacuation management information was received and for which subsequently emergency response procedures have been studied. For more information about these tunnels, see the lines in italic in appendix 1 per country for these tunnels.

Table 2: Tunnel of which evacuation procedures have been studied. Tunnel Between Fréjus France – Italy Öresund12 Denmark – Sweden Eurotunnel (rail) France – United Kingdom Karawanken Slovenia – Austria

Below, we summarize the most important findings per tunnel with regard to cross-border incident management

Fréjus tunne13l

The French (SFTRF) and Italian (Sitaf) tunnel operator have developed a bi-national emergency plan for security in de Fréjus tunnel. The cross-border aspects are extracted from this bi-national emergency plan (SITAF and SFTRF, 2004):

• bi-national development of the bi-national emergency management plan

• the bi-national emergency management plan is in two languages: French and Italian

• international cooperation

• cooperation agreement between SITAF and SFTRF

• the ventilation system can be operated both at the French and Italian operator room. The French control room is in charge.

• French and Italian emergency responders can use the same radio frequency for communications

• in case of an traffic accident, there is direct coordination between the French and the Italian operator

• a bi-national alarm scheme is developed in which the corresponding French and Italian hierarchical levels are indicated

12 The Öresund tunnel connect Denmark and Sweden, but is completely on Danish territory. 13 The research conducted after the Fréjus tunnel took place before the early June 2005 tunnel accident.

Öresund tunnel

The information with regard to cross-border emergency management is provided by the health and safety manager of the Öresund.

• all aspects of emergency planning are dealt with in common between the Danish and Swedish authorities

• there is only one single control room which is manned with professional from both Sweden and Denmark

• in case of emergencies, tunnels users are warned in Swedish, Danish and English by radio on 3 FM channels

• the bi-national emergency management plan is in two languages: Danish and Swedish

• in case of an accident, fire fighters from both countries turn out. the first fire engine at the accident scene is in charge, no matter in which country the accident occurred. however, when fire fighters from both countries are at the accident scene, the lead is taken by the fire fighters from the country where the accident took place. There is always only one single leader on the accident scene.

• there is one single emergency radio system for communication purposes for both the Swedish and Danish emergency responders

• Fire hydrants have outlets for both the Swedish and Danish system

• bi national training exercises are hold

• law enforcement is the one in which country the accident has happened

• there is one insurance covering the total link

• all accidents are evaluated including authorities from both countries

Eurotunnel Johnson has conducted research related to cross-border aspects in general and with regard to the Eurotunnel in particular. The Eurotunnel results are presented below (Johnson, 2000):

• unified word use and using the same terminology

• bi-national safety committee and emergency planning bi-national committee

• emergency plans in both English and French

• attendance of English and French team members at major accident exercises and planning meetings

• matching the hierarchical levels between the French and English emergency response organisations

• appointments on strict standards for operational communications between British and French agencies

• harmonization in operational procedures and work practices

• bi-national exercises

• the exchange of casualty details between casualty bureaux in more than one country

Karawanken tunnel The information with regard to cross-border emergency management is extracted from the agreement between the Austrian and Slovenian ministry for Traffic and the construction agreement between Austria and Slovenia. This agreement concerns in particular transport of hazardous materials (Austrian and Slovenian Ministries of Transport, 1997 and 2003).

• vehicles transporting hazardous materials are assisted/guided by vehicles form Austrian and Slovenian traffic ministry. These vehicles have permission to cross the border for these purposes.

• the exit of the tunnel is only closed after the entrance of the tunnel in the other country is closed

• users are informed in both German and Slovenian

• emergency responders are allowed to act on the territory of the other country

We conclude that several tunnels have already prepared cross-border aspects in their evacuation and intervention management. However, there is a variety in cross-borders aspects that is covered. The table below summarizes the results of the elaboration. In this table, the first row indicates the cross- border tunnel, the first column indicates the various cross-border aspects. The cells indicate (using ) per tunnel if the particular cross-border aspect is incorporated in the evacuation and intervention management.

Table 3: Cross-border aspects in several cross-border tunnels.

Fréjus Öresund Eurotunnel Karawanken bi-national emergency plan multi language emergency plan

International cooperation of tunnel providers one single commander in emergency management communication appointments

Organisational match multi language user warnings hybrid (bi- national) fire hydrants bi-national incident evaluation bi-national exercises emergency responders permission to cross-border

24. Practical examples

For the practical examples, we refer to paragraph 5.2, in which we studied the case studies of the Eurotunnel fire in 1996 and the Mont Blanc tunnel fire in 1999. These examples show some practical examples of cross-border aspects that went wrong during suppression activities. Furthermore, we refer tot paragraph 5.3 for available cross-border intervention procedures. The practical examples concern the Fréjus, Öresund, Eurotunnel and Karawanken tunnel.

25. Proposals for EU guidelines

Based upon the research activities and their results, we list recommendation for cross-border tunnel evacuation management below. Some of the guidelines are specific for cross-border tunnels between countries with different languages (these are indicated with *).