FIRE PROTECTION AND LIFE SAFETY IN BUILDINGS AND TRANSPORTATION SYSTEMS

Advanced Research Workshop FIRE PROTECTION AND LIFE SAFETY IN BUILDINGS AND TRANSPORTATION SYSTEMS Workshop Abstracts UNIVERSIDAD DE CANTABRIA Dpto...
Author: Baldwin Allison
1 downloads 1 Views 402KB Size
Advanced Research Workshop

FIRE PROTECTION AND LIFE SAFETY IN BUILDINGS AND TRANSPORTATION SYSTEMS

Workshop Abstracts

UNIVERSIDAD DE CANTABRIA Dpto. de Transportes y Tecnología de Proyectos y Procesos GIDAI – Seguridad contra Incendios – Investigación y Tecnología Avda. Los Castros, s/n 39005 Santander. España

This Advanced Research Workshop was organized by: GIDAI – Fire Safety – Research and Technology UNIVERSIDAD DE CANTABRIA Dpto. de Transportes y Tecnología de Proyectos y Procesos Avda. Los Castros, s/n 39005 Santander. Spain Tf. + 34 942 201826. Fax. +34 942 201873; [email protected]; http://grupos.unican.es/GIDAI With the collaboration of: Society of Fire Protection Engineers SFPE Scientific Committee - Editorial Board: - Dr. Daniel Alvear

National Fire Protection Association NFPA

International Association for Fire Safety Science IAFSS

Universidad de Cantabria

España

- Dr. Jorge A. Capote

Universidad de Cantabria

España

- Dr. Richard Carvel

University of Edinburgh

UK

- Dr. Michael Delichatsios

University of Ulster

UK

- Dr. Carlos Fernández-Pello

University of California

USA

- Dr. Charles M. Fleischmann

University of Canterbury

N. Zealand

- Dr. George Hadjisophoucleus

University of Carleton

Canada

- Mr. Morgan J. Hurley

SFPE

USA

- Dr. Timo Korhonen

VTT Research Centre

Finland

- Dr. James A. Milke

University of Maryland

USA

- Dr. Frederick W. Mowrer

University of Maryland

USA

- Dr. Paulo Piloto

Polytechnic Institute of Bragança

Portugal

- Dr. David Purser

Hartford Environmental Research

UK

- Dr. James G. Quintiere

University of Maryland

USA

- Dr. Guillermo Rein

University of Edinburgh

UK

- Dr. José Torero

University of Edinburgh

UK

- Dr. Arnaud Trouve

University of Maryland

USA

Sponsored by: MINISTERIO DE CIENCIA E INNOVACIÓN

Convocatoria de Ayudas para la realización de Acciones Complementarias Ref.: BIA2009-05701-E/

I

Index

Página Index ..................................................................................................................................

i

The Regulatory use of Advanced Fire Engineering Techniques ........................................

1

Applied Psychosociology to Intervention Teams in Situations of Emergency...................

2

Experiments to Investigate Radiation Heat Fluxes on Adjacent Buildings........................

4

Analysis of eleven evacuation events in Finland.................................................................

5

Analysis Stability on fire conditions of the Roof Structure of Building Intermodal in the new Airport Terminal T1 Barcelona. .............................................................................

6

Numerical simulations of some possible fire scenarios in a closed car park with RANS and LES ...................................................................................................................

8

Modelling of pedestrian movement around 90° and 180° bends........................................

9

An Experimental Review of the Homogeneous Temperature Assumption in PostFlashover Compartment Fires. .............................................................................................

10

Experimental Research - Large-Scale Tunnel Fire Tests and the use of CFD modelling to predict Heat Flux and Thermal Behaviour. ...................................................

12

i

Landing Distance of Droplets from Water Mist Suppression Systems in Tunnels with Longitudinal Ventilation. .....................................................................................................

13

Comparison of STEPS and FDS+EVAC simulations to an evacuation of a 4 storey building. ................................................................................................................................

15

Effectiveness assessment of road tunnel fire-fighting strategies by ventilation and water mist systems................................................................................................................

16

The Smoke Layer Interface during a Fire in an Atrium: a New Method to Locate it using a Zone Computer Model.............................................................................................

18

Low and medium power full-scale atrium fire tests and numerical validation of FDS......

19

Can Active Fire Protection Systems in Tunnels Prevent Minor Fire Incidents From Becoming Disasters?. ...........................................................................................................

21

How do you Determine Length of the Staircase when Performing Simulex?....................

22

Façade Flames for Corridor (Tunnel) Like Enclosure Fires. ..............................................

24

Thermal Analysis in Fire-Resistance Furnace. ...................................................................

25

Influence of the BENCH SCALE TEST in the calculation of THE HEAT RELEASE RATE for Aircraft Materials: Cone Calorimeter vs. OSU Apparatus ........

27

Sthocastical Egress Model of Passengers Train. .................................................................

29

Application of human behaviour and toxic hazard analysis to the validation of CFD modelling for the Mont Blanc Tunnel fire incident.............................................................

30

An Immersive Simulation of Fire Evacuation Based on Virtual Reality ……..................

32

ii

THE REGULATORY USE OF ADVANCED FIRE ENGINEERING TECHNIQUES Dr. Jur Vincent Brannigan Prof. Emeritus.Department of Fire Protection Engineering University of Maryland Numerous disasters over the past 30 years have shown the inadequacy of traditional approaches to controlling fire safety. Engineers and fire scientists have worked steadily to improve our understanding of fire. However it is not clear that regulatory structures have kept up with the changes in engineering design philosophy and novel technological methods. In case after case, disasters occur despite regulatory compliance and often in ways that were entirely predictable with advanced engineering analysis. Movement of products in international trade and a dependence on possibly inadequate performance based standards can only make the situation worse. The use of advanced fire engineering techniques requires an appraisal of the capability of the regulatory system to adapt to the unique advantages and disadvantages of the techniques. This requires a flexible approach with a suitable balance between public and private controls, pre building approval and post building regulation and the ability to adapt to changes in understanding of the underlying hazard. Sophisticated definitions of fire load, human behavior and system redundancy wter into binding contracts ill be required. There is also potential for a contract approach in which builders and developers create a mutually enforceable private law agreement which tailors the social safety requirements more precisely to the technical capabilities. Such systems can also support third party enforcement. Fire safety has the engineering and technological base to move into the 21st century, but It needs a regulatory system and philosophy to support the use of the technology.

1

APPLIED PSYCHOSOCIOLOGY TO INTERVENTION TEAMS IN SITUATIONS OF EMERGENCIE de las Heras Fernández, Mariano 1; Beltrán Bengoechea, Luis 1; de las Heras Ramos, Aída Muriel 1; Ferrer Garcés, Rafael 2; Iglesias Rodríguez, Manuel 3 1

Dep. Construcciones Arquitectónicas y su Control. Grupo de Investigación GIEPPE para el Estudio de la Problemática del Patrimonio Edificado. Escuela de Arquitectura Técnica e Ingenieros de la Edificación Universidad Politécnica de Madrid. 2 Dep. Expresión Gráfica Arquitectónica. Grupo de Investigación GIEPPE para el Estudio de la Problemática del Patrimonio Edificado. Escuela de Arquitectura Técnica e Ingenieros de la Edificación Universidad Politécnica de Madrid. 3 Escuela de Bomberos y Protección Civil. Ayuntamiento de Madrid, España The members of emergency teams are involved in emergency situations in which they live close and dramatic situations with a high burden of human misery, taking responsibility for resolving claims for incidents that occur that are outside the normal experiences and those who are highly emotionally involved. In performing their normal work in direct contact with affected by disasters of various kinds of more or less serious, emotionally involved with the suffering of others, so they often are in a position to risk not only physical but also psychological. In addition, the member of the Emergency Team is sometimes so involved in saving the lives of others that they forget to pay attention to itself, which can produce physical and emotional harm. If you are not worried about their personal safety during operations or do not recognize their own stress, can cause your effectiveness and efficiency in the workplace suffer. For the effective performance of their work is essential to the team Emergency mastered their craft and learn to perfection the technique and method of intervention, but also need to know to handle emotions properly, both own and those of affected by the incident, so that the tension does not affect the effective performance of the task of intervention.

2

The objectives that we want cover are: - Identification of the emotions and expressions of both individual and group, more frequently experienced by those affected by calamity, to make use of communication techniques and appropriate psychological support to minimize the negative reactions and to handle properly the normal emotional reactions in people affected . - Identify the likely reactions of the members of emergency teams on the sinister and post, in order to establish methods to assume the normal emotions and reactions. - Identify the tools to handle stressful situations and support mechanisms for those affected: - Provide a comprehensive relief of physical and emotional suffering. - Early recognition of own reactions to stress and the mechanisms to be put in place to monitor and manage the stress effectively, avoiding the psychological and let cronifique affecting the future effectiveness claims.

3

EXPERIMENTS TO INVESTIGATE RADIATION HEAT FLUXES ON ADJACENT BUILDINGS Hao Cheng and George V. Hadjisophocleous Department of Civil and Environmental Engineering Carleton University. ABSTRACT This paper presents the results of full-scale experiments conducted to determine heat fluxes to adjacent buildings from flames issuing from a window. The experiments were conducted using a three-story facility at the National Research Council of Canada. The size of the fire room was 5.95 m x 4.4 m x 2.75 m high. The room had a window on one wall whose dimensions were varied to study the impact of the window size on radiation heat fluxes on a target wall. The façade of the exterior wall was covered by gypsum board. A 5 x 5-m wood-framed target wall was constructed with and placed across from the window at 4 different distances: 2.2 m, 3.0 m, 3.5 m and 4.0 m. The exterior façade of the target wall was covered with non-combustible cement boards. Sixteen thermocouples were installed on the external wall of the fire room above the window to measure the surface temperature and temperatures of the window plume. Thermocouples were also installed below the top of the window to measure the temperature of the gases leaving the fire room. Radiometers were installed on the target wall to measure heat fluxes at different heights. In addition, thermocouples were used to measure the surface temperatures of the target wall. Two infrared cameras were also used during the experiments to record temperature distribution on the target wall and the wall above the window. Twelve experiments have been conducted; nine using propane and three using woodcribs as the fuel. The parameters studied were the window area and aspect ratio, number of windows, and distance of target wall from window. Results of heat fluxes on the target wall and temperature of target wall and the window plume will be presented and discussed.

4

ANALYSIS OF ELEVEN EVACUATION EVENTS IN FINLAND T. Korhonen, T. Rinne, and K. Tillander VTT Technical Research Centre of Finland ABSTRACT In this work, eleven evacuation events in Finland were studied in order to provide validation data for some aspects of evacuation modelling. Three of these events were real evacuations, which occur every now and then due to false alarms or real fires. The rest (eight) were fire drills, which are normally carried out as part of the safety training of the staff in public buildings and offices. The studied buildings included college buildings, shopping malls, a hospital, a factory, office buildings, a church, and comprehensive schools. The main techniques used for the observation of evacuation events were digital video camera and surveillance camera recordings. A large amount of information was obtained and some problems in the application of the observation techniques were identified. The data analysis is still going on and this paper presents the first results of the quantitative data analysis. Some qualitative observations were also collected during the preliminary data analysis phase and these are also shortly presented. The data obtained from the fire drill of a college building is presented in detail, where the recordings of the digital video cameras and surveillance cameras were used to measure, e.g., the flow rates of people and their walking speeds. The first results are very promising and it is supposed that the quantitative analysis of the other observed evacuation events will produce much more useful data. An evacuation drill in a church is analysed further by making evacuation simulations using the FDS+Evac programme. The church has a quite good surveillance camera system but four additional digital video cameras were used in order to cover the interior of the church better. There were about 200 persons attending the service and the “alarm” was given by starting a cold smoke generator and the priest telling the situation and asking people to leave the building and pointing towards the different exits. Thus, the initial location of the persons are known quite well and the reaction time distribution was quite narrow, which makes the case easier to model, because the initial conditions for the movement time computation are quite good known.

5

ANNALISYS STABILITY ON FIRE CONDITIONS OF THE ROOF STRUCTURE OF BUILDING INTERMODAL IN THE NEW AIRPORT TERMINAL T1 BARCELONA Dr. Ing. Alfredo Arnedo Pena(*)(**) / Ing. Fco. Javier Salcedo López(*) / Ing. Arcadi Sanmartín Carrillo(*) (*) SENER. Ingeniería y Sistemas.(**) Profesor asociado UPC. On June 17, 2009 is scheduled to inaugurate the new Terminal T1 from Barcelona Airport, one of the largest air terminals in the world and possibly the most important transport infrastructure that will launch in coming months in Europe. The Intermodal Building is the link of several modes of transport converging in the Terminal: the airport itself, the high speed rail and road transport and metropolitan bus, taxi or private car. The roof of the building covers an area of 8,000 m2, has been entirely designed in structure-based on steel trusses and composite slab deck and is supported by steel tubular columns with concrete inside. The structure is formed by a system of primary and secondary lattices and purlins which are connected to a deck built of concrete. The general design criteria stipulated that in the terminal was not necessary considering the action of fire, because of the height, size, load and fire resistant expected margin for the structure. Once designed and built the building, the general report on fire safety has reconsidered the initial position. So it has been mandatory a calculation of the structure based on the permanent loads actually acting, also to increase the vent surface air and to perform modifications in some joints next door of other sectors of fire. This paper presents a summary of all calculations and analysis covering a wide range of issues related to the protection against fire. The calculations dealing with fire safety has been done according to the simplified method of calculation for steel and composite structures included in Eurocode 3 and 4, 12 that now has been partially incorporated in the core document of the SI Technical Building Code CTE , mandatory in Spain today.

6

For each element of the roof the so called critical temperature is obtained, at which the collapse occurs when the combination of actions is the accidental one. Taking into account this critical temperature the time of stability is determined and compared with the requirements considering the standard fire curve ISO 834. The fire resistance of each element is enough safe for the time stability required when is above the equivalent exposure time. At calculating this exposure time it plays a very important role the determination of the density of fire load, the active-fighting operation measures in the Terminal, and the increase of area of ventilation by adding of vents automatic devices.

7

NUMERICAL SIMULATIONS OF SOME POSSIBLE FIRE SCENARIOS IN A CLOSED CAR PARK WITH RANS AND LES Mehdi Jangi,*, Nele Tilley, Bart Merci Ghent University, Department of Flow, Heat and Combustion Mechanics Numerical simulations for the evaluation of the risk of several possible fire scenarios in a closed car park will be presented. This study is of fundamental interest for the development of universal design standards for closed car parks in urban areas. As in [1], we use CFD simulations as numerical experiments. A design fire is adopted from available experimental heat release rate curves for the case of cars on fire in a tunnel [2]. Fire is simulated by insertion of a thermal heat source in the computational domain. The energy conservation equation with this additional source term is solved. The time evolution of the thermal source is such that it satisfies the chosen design fire heat release rate curve. Detailed effects of sprinklers are not considered. Thus, smoke and heat exhaust ventilation is provided only by an inlet air mass flow. Various air mass flow rates are examined to determine the associated heat and smoke backlayering length of each condition (reference is made to the results reported in [3]). These calculations are conducted by OpenFoam [4] and FDS [5] for the purpose of comparison and validation of the numerical model and results. In addition, calculations with OpenFoam are performed using RANS and LES to investigate the effect of turbulence models on the results. Several fire scenarios are examined, considering various configurations of car and fires in the car park in order to assess their risk level. Flow field and temperature distributions at each condition will be discussed and the differences between the results of RANS and LES are addressed. Finally, time evolutions are shown for the temperature distribution at the ceiling, near the cars. These graphs envisage the effects of the fire on the construction of the car park.

8

MODELLING OF PEDESTRIAN MOVEMENT AROUND 90° AND 180° BENDS Bernhard Steffen, Armin Seyfried Jülich Supercomputing Centre For the planning of large pedestrian facilities, the movement of pedestrians in various situations has to be modelled. These include different densities and different usages, at least standard operation mode and evacuation. Many tools for pedestrian planning are based on cellular automata (CA), discrete in space and time. It is common experience that CA have problems with modelling sharp bends in wide corridors, and especially with 180° turns as they are frequent in staircases. They tend to move the pedestrians to the innermost lanes far too strongly, thereby reducing the capacity. In the extreme, only a small fraction of the width will be used after the bend. There have been some remedies proposed, but a systematic investigation is lacking, and there is no accord on the causes of the problem. In the paper, we present a comparison of various implementations of CA- for treating 90° and 180° bends with different width. We compare different models for the static floor field which provides the persons orientation. We also indicate how a density dependant static floor field can give a much increased capacity while still allowing single walkers to cut corners. We test at which position the walkers are most likely to get stuck and relate this to differences between CA models and real-world behaviour. The behaviour in the model is compared with observation of people walking through a 90° bend in a corridor and walking down a wide staircase. It is shown that with a proper floor field, at least the average behaviour of people can be simulated. Capturing the fluctuations present in reality is beyond the capacity of the simple models investigated.

9

AN EXPERIMENTAL REVIEW OF THE HOMOGENEOUS TEMPERATURE ASSUMPTION IN POST-FLASHOVER COMPARTMENT FIRES Jamie Stern-Gottfried1,2, Guillermo Rein1, and Jose L. Torero1 1

:BRE Centre for Fire Safety Engineering The University of Edinburgh, UK 2 :Arup Fire, UK

Abstract Post-flashover fires are of particular relevance to the analysis of structural fire performance. Traditional methods for quantifying compartment fires assume homogeneous temperature conditions, i.e. the gas phase temperature is uniform and does not have significant temperature gradients. For example, the methodologies for structural fire analysis, such as the standard temperature-time curve and the parametric temperature-time curves, assume this uniform temperature regardless of the compartment size and fire power. This assumption has been necessary to develop simple analytical solutions to the temperature evolution and further the understanding of post-flashover compartment fires and the subsequent structural response. However, the accuracy and range of validity of the homogeneous temperature assumption has not been thoroughly examined. This is generally due to the limited data from postflashover fire experiments available and especially the low spatial resolution of temperature measurements in the gas phase. This paper examines this assumption in previously conducted fire tests. A statistical analysis of the data provides insights into the temperature field distribution. Two types of fire test are included: Fire tests with non-uniform ventilation. These tests include one large, long enclosure in Cardington (UK, 1993) and a set of fourteen tests from a smaller compartment in Victoria (Australia, 1999). Fire tests with a high degree of spatial resolution in temperature measurements. These tests include one real high-rise apartment block in Dalmarnock (UK, 2006) and a series of eight tests in a 12m x 12m room at Cardington (UK, 1999-2000). 10

The fire tests with non-uniform ventilation conditions show non-uniform burning conditions. As the data from these tests have a low spatial resolution, no detailed statistical analysis can be made. However, the results of these tests clearly show that the non-uniform burning produce non-homogeneous temperature conditions. The temperature conditions in the tests with high spatial resolution are statistically examined in terms of deviation from the average. For example, the figure below provides a temperature-time curve for the Dalmarnock fire test, including the average temperature, ± one standard deviation from the average temperature, the minimum temperature, and the maximum temperature located in the test compartment.

Temperature-Time Curves for the Dalmarnock Fire Test

The review of the data concludes that fully uniform temperature conditions are not reached. Peak local temperatures range from 6% to 74% higher than the compartment average, with an average increase of 25%. These results are then applied to a typical steel beam to examine the impact of the assumption of homogeneous temperature conditions on its heating. The use of local peak temperatures results in reductions of the time to failure (quantified as time to reach 550°C) from those calculated from the average gas temperature from 7% to 41%, with an average of 24%. These results can be used to estimate the safety factors for structural analysis required to avoid underestimation of the fire severity to the structure.

11

EXPERIMENTAL RESEARCH - LARGE-SCALE TUNNEL FIRE TESTS AND THE USE OF CFD MODELLING TO PREDICT HEAT FLUX AND THERMAL BEHAVIOUR Gabriele Vigne, Jimmy Jönsson ArupFire(Spain) ABSTRACT Large-scale fire tests were performed in a 50m² cross-section tunnel situated in La Ribera del Folgoso (Ponferrada), Spain. The tunnel installation belongs to FSB, Fundacion Santa Barbara. The tests were conducted to investigate the temperature and heat flux behaviour in the surroundings of a fire, approximately 20m upstream and 20m downstream of the fire. The test-setup was made with the intent of getting valuable information to be used for CFD-software validation, with particular emphasis on the tests uncertainty. Therefore the same test was repeated several times; the fire size and the ventilation conditions were not changed in order to get valuables information about relative errors. The test-setup consisted of 15 thermocouples, 10 radiation plates, velocimeters and load cells to measure the mass loss and thus the Heat Release Rate. The fire source was made up of a Heptane pool with a peak Heat Release Rate of approximately 5 MW, burning constantly at this level for about 10 minutes. After 3 preliminary tests, 6 more tests were conducted, each one with the same fire size, ventilation and ambient conditions. The main aim of the study was to investigate the adequacy of using CFD-software to determine the thermal conditions in the surroundings of the fire source in a tunnel.

12

LANDING DISTANCE OF DROPLETS FROM WATER MIST SUPPRESSION SYSTEMS IN TUNNELS WITH LONGITUDINAL VENTILATION. Angelo Cavallo(*)(**), Richard Crosfield(*), Francesco Colella(*)(**), Ricky Carvel(*), Guillermo Rein(*) and José L. Torero(*) (*)BRE Centre for Fire Safety Engineering, University of Edinburgh, UK (**)Dipartimento di Energetica, Politecnico di Torino, Italy ABSTRACT This paper presents results from a simple analysis of water mist droplet trajectories under ventilation conditions that are commonplace in tunnels. An ODE model solving for the motion of a single droplet falling from the ceiling of a tunnel to the deck under a cold cross flow is used as an approximation to the problem. Unimodal turbulence is considered. The results for a realistic scenario of a modern tunnel show that most water mist droplets (i.e. those less than 170 μm) are carried between 60 and 200 m downstream before they reach the road deck with 5 m/s of average ventilation flow (see Figure). A significant proportion of the water mist may be carried very large distances downstream under similar conditions or larger ventilation velocities. The results from this simple model agree well with CFD calculations in a 3D domain using RANS turbulence modelling and with experimental measurements of falling droplets in a scaled tunnel 0.3 m diameter and 6 m. long.

13

1.5

pdf

1

0.5

0 27.5

28

28.5

29

29.5 30 30.5 landing distance (m)

31

31.5

32

Figure. Two sets of trajectories in a 6 m tall tunnel. Top) droplets of 170 μm under 1, 3, 5 and 10 m/s average ventilation flows. Bottom) Probability density function resulting from 100 simulations of 300 μm droplets under 3 m/s average ventilation flow.

According to these results, if the current ventilation strategies are to be employed in the event of a fire, the zone length should be considerably longer than the 50 m long stipulated by the World Road Association (PIARC) for flows up to 10 m/s. An alternative strategy, in order to reduce the lengths of nozzle zone, would be to reduce the ventilation flow during emergency response. The results suggest that more experimental research into the interaction of water mist, tunnel ventilation and fire plumes is required.

14

COMPARISON OF STEPS AND FDS+EVAC SIMULATIONS TO AN EVACUATION OF A 4 STOREY BUILDING Katrina Speirs (SAFE Consulting), Kristen Salzer-Frost (SAFE Consulting), Victoria Gray (Buro Happold (FEDRA)) In 2006 an evacuation of the Saltire Centre at Glasgow Caledonian University was video recorded from all the escape stairs to allow evacuation times to be determined. Simulations of the evacuation using both the well validated STEPS software and the newer FDS+EVAC package are to be compared to the data from the real evacuation. The FDS+EVAC package will also be used to investigate a fire scenario in order to highlight the differences between evacuation only simulations and one incorporating the interaction of the occupants with the smoke. This is a large complex building formed by the integration of a number of existing buildings and a modern complex. It is a large complex space therefore people may have way-finding difficulties and managing/modelling an evacuation could be challenging. The building is connected to the other two existing buildings of the university, George Moore building and Hamish Wood building. The ground floor of the building is not accessible from the outside to students; the main entrance for the building is at first floor level. The Saltire Centre is a library facility, with computer, study and meeting facilities. This is a 5-storey building with glass façade and a roof garden. The building is used for a variety of purposes; predominately it is a library building. The ground floor is a café, meeting area and computer work area; this is generally the meeting point for people and often full of students. The ground floor creates a void the full height of the building. This is a typical example of a building which is beyond the scope of prescriptive building regulations and for which simulations such as these are used as an integral part of the design process. As such, it makes an ideal case study for assessing the effectiveness of the simulation tools that are available and identifying areas that are in need of further refinement and those parts of the simulation that correlate well with actual scenarios.

15

EFFECTIVENESS ASSESSMENT OF ROAD TUNNEL FIREFIGHTING STRATEGIES BY VENTILATION AND WATER MIST SYSTEMS Giovanni Manzini, Luca Iannantuoni Politecnico di Milano - Department of Energy In order to verify the effectiveness of different fire-fighting strategies applied to an urban road tunnel, a numerical analysis was done, including three different numerical tools: SES (Subway Environment Simulation, developed by the Department of Transportation of United States of America), ECART (a proprietary code of ERSE - ENEA Ricerca sul Sistema Elettrico, developed and owned before by CESI RICERCA) and FDS (Fire Dynamics Simulator, developed at Building and Fire Research Laboratory of NIST). The shallow twin-bore tunnel, object of this study, is approximately 800 m long and equipped with jet fans at portals plus a transversal ventilation system, to guarantee the fresh air income during operation and, with properly sized fans, the smoke extraction in case of fire. Moreover, the impact of a water mist fire deluge system had to be analyzed in terms of smoke confinement, to protect the egress paths and prevent the fire and smoke spread along the tunnel. Once the sizes of design fires and their location was set, a one-dimensional, incompressible, turbulent, slug-flow model, called SES, was used to determine airflows and temperatures in all the tunnel sections and ventilation shafts. The SES usage was enhanced by an external tool, named proSES (developed by Politecnico di Milano Department of Energy - and owned by Metropolitana Milanese S.P.A.), which simplifies the model description and adds useful feature to do parametric studies and post process the output data. This results was then necessary to set the boundary conditions to apply ECART (a lumped parameter model, but capable to treat the water mist injection, transport and dynamics) and FDS (a CFD code). In fact, because of computational issues, primarily regarding the mesh size and the simulation time required, only a portion of the entire physical domain was described during modeling by these two codes.

16

With ECART and FDS, the analysis has also addressed the smoke destratification (due to water mist injection and other transport phenomena) which is relevant to determine the safety condition for people during the egress. The main criticality was about uncertainties in spray phenomena and characterization of water mist droplet size distribution and its behaviour in strong ventilated compartments, so there was not treated the interaction between combustion reactions and water droplets. In conclusion, it was shown that the ventilation strategies adopted when fire occurs at a given location, and the fan sizes was sufficient to prevent the spread of smoke along the tunnel in most cases, up to fire sizes around 10 MW. When the maximum heat release rate becomes higher, a water mist deluge system demonstrates its efficacy in terms of smoke confinement and temperature mitigation.

17

THE SMOKE LAYER INTERFACE DURING A FIRE IN AN ATRIUM: A NEW METHOD TO LOCATE IT USING A ZONE COMPUTER MODEL

Jorge A. Capote; Daniel Alvear; Orlando Abreu; Mariano Lázaro; Pablo Espina GIDAI Group – Fire Safety, Research and Technology. University of Cantabria. ABSTRACT The scale models, analytical models and fire computer simulations zone and field, are useful tools for scientific analysis of the smoke movement in atria. However, the results obtained by these methods are sometimes different due to the complexity of the phenomenon and the assumptions implicit in the equations that characterize the smoke movement in atria, enclosures where the surface height (A/H2) and the magnitude of the fire have a strong relevance. Among the different systems used for the smoke control in large spaces, the smoke filling strategy is particularly suitable. Based on its buoyancy, it is to ensure the safety evacuation of occupants. For this reason, the smoke layer interface position during the development of a fire is one of the most important aspects. This study proposes a general method, using the strategy outlined above, for determining the smoke layer interface location from the fire in atria. Based on the analysis of functions of the curves corresponding to the position of the smoke layer interface during the fire, from the development of numerous tests as Froude scale model, and their fitting to a type curves family - using the least squares method - so, making corrections on the model Computational Consolidated Fire Zone Model for Fire and Smoke Transport CFAST, can achieve a more accurate prediction of the smoke layer interface location using this computational model. The results provide a new approach to locate more precisely the smoke layer interface location through the computational zone model CFAST, suggesting a general method for the determination of this parameter, especially important for the fire safety.

18

LOW AND MEDIUM POWER FULL-SCALE ATRIUM FIRE TESTS AND NUMERICAL VALIDATION OF FDS Cándido Gutiérrez-Montes a, Enrique Sanmiguel-Rojas a, Antonio Viedma b, Guillermo Rein c a

Fluid Dynamics Division of the Department of Mining and Mechanical Engineering, University of Jaen, Jaen, Spain b Department of Thermal and Fluid Engineering, Technical University of Cartagena, Murcia, Spain c BRE Centre for Fire Safety Engineering, The University of Edinburgh, EH9 3JL, UK Abstract: The inclusion of atria within modern large buildings is relative recent. These structures are important architectonical features since the 60’s. Atria are a source of discussion within the fire science community. They introduce complex designs and non conventional architectonical elements that can lead to fire environments diverging from those in current codes. Because of this, the current trend in fire safety in atria is towards performance based design which relies on fire modelling. At this point, it is still necessary to improve and validate the existing numerical models. For this aim, some tests were carried out at the Murcia Fire Facility. These consist of 19 full-scale fire tests that provide with new experimental data of atrium fires. The fire size, the smoke extraction rate and make-up openings size and location were varied. At the present paper, results from three experiments in a 20 m cubic facility are reported. Later comparisons with the predicted results from Fire Dynamics Simulator (FDS) v.4 are also presented. In the far field, good agreement has been found at the upper regions of the facility (above 10 m from the ground) whereas poor agreement has been found in the near field, near the flame and at the smoke layer interface.

19

a)

b)

c)

Figure 1. Test facility in a; sensors’ location scheme in b; temperature measurements (solid line), and predictions (dash-dot line), at exhaust fans, in red, near the walls at 15 m high, in blue, and 5 m high, in grey of one of the tests in c.

20

CAN ACTIVE FIRE PROTECTION SYSTEMS IN TUNNELS PREVENT MINOR FIRE INCIDENTS FROM BECOMING DISASTERS? Ricky Carvel BRE Centre for Fire Safety Engineering, The University of Edinburgh Abstract The majority of road tunnels of significant length have a ventilation system, intended to be used for smoke control in the event of a fire incident. Many road tunnels in Japan and Australia have sprinkler systems intended to suppress fire incidents. An increasing number of tunnels in Europe are being equipped with water mist systems for fire suppression in the event of a fire incident. This paper will briefly review the (positive and negative) claims made about these systems, will discuss the experimentally verified capabilities of these systems with regard to fire control and will highlight the differences between what has been claimed and what has been demonstrated. The paper will then consider a few historical tunnel fire incidents – including the fires in the Channel Tunnel (1996), the Mont Blanc Tunnel (1999), the Fréjus Tunnel (2005) & US Interstate 5 (2007) – and will speculate what the outcome might have been had different active fire protection systems been installed in these tunnels.

21

HOW DO YOU DETERMINE LENGTH OF THE STAIRCASE WHEN PERFORMING SIMULEX? Hyun-seung. Hwang, Jun-ho. Choi* & Won-hwa. Hong School of Architecture, Kyungpook National University ABSTRACT SIMULEX, which has been developed in the UK, is a world-renowned exclusive evacuation computer program to analyze the evacuees' escaping situation on the assumption that multi-level buildings are exposed to fire. Within the process of simulation, SIMULEX user must need to link each floor to staircases in order to obtain the evacuation results of simulation, including the Personnel Evacuation Time (PET) and Cumulative Wait Time (CWT), and then input the parameter of staircase-width in accordance with size of the plan. However, we are usually in a delicate situation when inputting the parameter of staircase-length because almost evacuees tend to make a trace of semicircle while passing through the stair landings, by comparison, SIMULEX population can evacuate through the shape of rectilinear staircase only, thus the gaps of the final evacuation time data occur because of the difference between the valid distance of evacuees' escaping movement and the staircase-length in SIMULEX. Therefore, this study selected a subject of high-rise building for calculating the distance of evacuees' escaping movement while evacuating through the stair landing with 351 invited participants, executed the trial empirical CCTVs recorded experiment progressed to go downward from 30th floor to 1st floor, classified approximately 5,000 cases: 1. Classified the evacuees who moved on stair landings near the wall of staircases or the stair railing when a participant went through the stair landing. 2. Classified the evacuees when two participants went through the stair landing

22

And then measured the frequency distribution curve, calculated the precise average distance of evacuees' escaping movement in staircase considering the walk speed of each participant and developing the method of simulation for an application of the accurate staircase-length as well as simulated to classify two cases(1st case to be applied the staircase-length according to input the distance of the half of stair landings’ length which was liable to err and 2nd case to be applied the average distance of evacuees' escaping movement which was proposed by this paper) in SIMULEX for verifying the validity of the paper. As a result, the output of the 2nd case was more precise than the 1st case's. Therefore, the users can input more accurate staircase-length when performing SIMULEX from now on, furthermore controversial points of high-rise evacuation (esp. bottleneck and the increase of CWT) will be able to decrease if the average distance of evacuees' escaping movement might be applied when designing the plan of fire protection.

23

FAÇADE FLAMES FOR CORRIDOR (TUNNEL) LIKE ENCLOSURE FIRES T. Beji, S. Ukleja, J.Zhang,W. Yao and M.A. Delichactios* Institute for Fire Safety engineering Research and Technology (FireSERT),University of Ulster. Abstract Flame heights and heat fluxes in facade flames originating from enclosure fires are needed in fire safety engineering design for preventing flame spread from floor to floor and to adjacent buildings. By systematically performing a series of small scale experiments having various enclosure geometries, door-like openings and fire locations, the physics and new relations are underpinned for the emerging flames on inert facades in ventilation controlled (under-ventilated) fires at the floor of fire origin. To limit the variables and uncertainties, propane and ethylene gas burners create a controlled (theoretical) heat release rate at the source. Gas temperatures inside the enclosure and at the opening, heat fluxes on the façade wall, flame contours (by a Charge Coupled Device camera-CCD camera) and heat release rates (by oxygen calorimetry) inside and outside the enclosure have been measured. The present study extends previous work in regular enclosures ( up to aspect ratio 3:1, cross section 0.5 m by 0.5 m) by allowing to gain insight on the burning rates, unburned fuel, external flame heights and temperatures and heat fluxes involved in a corridor-like small scale enclosure ( aspect ratio 6:1, consisting of six boxes having cross section 0.5 m by 0.5 m) for which experimental data remain scarce. The opening sizes at the front of the corridor were 0.1 x 0.25 m , 0.2 x 0.2 m and 0.25 x 0.1m whereas location of the burner was varied from the back to the front box. The fuel supply to the burner was increased linearly with time. The severity of the fire both inside and outside on the facade of the enclosure was assessed from video observations, the gas temperature measurements, the heat release rates and the flame heights. It was established, unexpectedly, that the inflow of air for under ventilated conditions is not affected by the corridor like geometry or the location of the burner and that the vertical distribution of gas temperatures inside the enclosure is uniform with height. For the lowest height ventilation opening (0.25 x .0.1 m ) the fire was extinguished when the burner was in the back box owing to the low inflow arte of air and reduced temperatures inside the enclosure from heat losses to the walls. The experimental data recorded is also used for exploration of capabilities of the FDS code modified with a soot formation model based on the smoke point height of the fuel.

24

THERMAL ANALYSIS IN FIRE-RESISTANCE FURNACE Piloto, P.A.G. 1, Mesquita, L.M.R. 1; Alexandre Pereira 2 1

- Applied Mechanics Dep., Polytechnic Inst. of Bragança, Portugal. 2 - Fellow research, Polytechnic Inst. of Bragança, Portugal.

Fire resistance rating of building construction elements is defined under fire-resistance test furnace. The geometry and shape of fire-resistance furnaces are not defined by any prescriptive document, being necessary to comply thermally with specified nominal fire curves, such as ISO 834 or hydrocarbon [1,2]. This research work intends to measure temperatures inside volume furnace, using sixteen plate thermocouples and compare average plane temperatures, with reference thermocouple that is responsible for controlling furnace operation, see figure 1. Three tests were performed, the first two running with ISO 834, during 45 minutes and the last one with hydrocarbon curve, during 30 minutes.

a) Furnace model with four b) Geometric position for c) Furnace with opened burners Bi (i=1,4). plate thermocouples inside door after running test furnace volume. 3. Figure 1 - Temperature instrumentation inside volume furnace.

Experimental results reveal that relative temperature differences are smaller than 30 % in the initial test stage, being smaller than 5 %, after 500 [s] until the end of the tests.

25

Numerical simulation was also performed to validate experimental results, based on the following transport equations: mass, momentum, energy and chemical species, together with radiative heat transfer model and turbulence model, using Fluent 3D. The numerical results agree well with experimental results, being the relative difference smaller than 5% for each nominal test. Numerical simulation reveal the localized effect of each burner, see figure 2.

a) Finite volume mesh.

b) Temperature distribution over four reference planes. Figure 2 – Numerical model.

26

INFLUENCE OF THE BENCH SCALE TEST IN THE CALCULATION OF THE HEAT RELEASE RATE FOR AIRCRAFT MATERIALS: CONE CALORIMETER VS. OSU APPARATUS J. A. Capote1, D. Alvear1, O. Abreu1, M. Lázaro1, I. Sáez de Ocariz2, I. López2 1

GIDAI Group – Fire Safety – Science and Technology, University of Cantabria 2 CTA – Aeronautical Technologies Centre

ABSTRACT Nowadays there are several standards to calculate the heat release rate of a material what means, if it behaves properly in case of a fire. The aeronautical standard FAR 25.853 (d) use the OSU (Rate of Heat Release Apparatus) to calculate the heat flux release by a sample of a material which is exposed to a heat flux of 35 kW/m2. Initially developed at the Ohio State University (OSU), this test apparatus was later modified and standardized by the American Society for Testing and Materials and the Federal Aviation Administration. Cone Calorimeter is a more standard apparatus in the field of fire safety, and it allows to perform tests of several international standards (ISO 5660, ASTM E 1354, ASTM E 1474, ASTM E 1740, ASTM F 1550, ASTM D 6113, NFPA 264, CAN ULC 135 and BS 476 Part 15). As is known, both equipments have different fundaments due to the fact that the Cone Calorimeter uses the oxygen consumption principle while the OSU apparatus was designed as an insulated box for enthalpy flow sensing but allowed to measure the heat release rate of different materials. In order to compare both test method, three aircraft materials were selected. The first one was a honeycomb panel that conform the walls inside the aircraft, the second one was the carpet that cover the floor of the cabin and the last one was a sample with the different layers that are present in the seats (upholstery, fire blocking layer and polyurethane foam). The Cone Calorimeter samples had the normalized size for the test (10x10 cm) and six heat fluxes (25, 35, 40, 50, 60 and 75 kW/m2) were applied to them. In case of OSU tests, samples had a size of 15 x 15 cm and only a heat flux of 35 kW/m2 was used. Important differences were found in the results according to the critical heat flux of the sample, but good correlations can be obtained if the critical heat flux of material is lower than 25 kW/m2. The results of the seat showed a good correlation between both tests, not in case of 35 kW/m2 but in case of 50 kW/m2 using Cone and 35 kW/m2 using OSU.

27

This material has a critical heat flux lower than 25 kW/m2 (smallest heat flux tested with the Cone Calorimeter) and the different concept of the ignition source of the tests could be the cause of these differences. The correlations in case of panel results were not successful but this is a better material than the previous one and its ignition temperature is between 40 and 50 kW/m2 (largest than the OSU heat flux). The carpet results were close to the seat one, getting the best values in case of a heat flux of 40 kW/m2 using the Cone Calorimeter and a heat flux of 35 kW/m2 using OSU apparatus. In this case the values of the first peak of Cone Calorimeter curve were a little larger than the OSU one but the second peak results were very close for both apparatus.

28

STOCHASTICAL EGRESS MODEL OF PASSENGER TRAIN J. A. Capote, D. Alvear, O. Abreu, M. Lázaro, A. Cuesta, V. Alonso GIDAI Group – Fire Safety – Science and Technology, University of Cantabria. ABSTRACT The existing egress models describe the evacuation process primarily for buildings and, most of them have a deterministic approach. The particular environments of passenger trains require a special approach to model the evacuation process. Similarly, more and more authors are agree on the need to adopt a stochastic approach to egress models due to the many random factors that influence this process, as the uncertainties in the input data, usually obtained from drills and experiments. That is the reason we developed a specific egress model for passenger trains which includes the geometry particularities, the common behavioral characteristics and the intrinsically stochastic nature of evacuation process under different emergency situations that may arises passengers. The model has been considered primarily the geometries of the passenger coaches of the rail network of Spain, especially the coaches of 102 and 130 series, and Talgo VI, but the concept of Object Oriented Programming in which is based his computational realization and, facilities provided for the entry of the geometric characteristics of the coaches, make it easy to potentially extend to other types. There are included situations of evacuation between coaches during the train movement and to a station platform. A new set of quantitative behavioral variables of passengers was designed. They was considered random variables and because was proposed a set of statistical methods of they treatment for its approximation to know probability density functions or for its estimation. A series of virtual experiments was conducted to determinate the statistical characteristics of input random behavioral variables. These characteristics were compared with the samples resulting of processing the evacuation drills videos and direct measurements during travels in real trains. The compatibility of the samples and obtained join statistical parameters was demonstrated. Model itself was conceived like a fine grid improve cellular automata movement model. The computational model was deployment in Microsoft C# 2008 Express Edition and .NET 3.5 SP1 Framework. The verification and validation of the model was conducting through the comparison of model results with the results of evacuation drills and another egress models. The validation of model demonstrate it correctness. Furthermore, the broad conception of the model together with the employ of Object Oriented Paradigm of the computer implementation lets an open application to other kinds of transport and scenarios with narrows escape ways. 29

APPLICATION OF HUMAN BEHAVIOUR AND TOXIC HAZARD ANALYSIS TO THE VALIDATION OF CFD MODELLING FOR THE MONT BLANC TUNNEL FIRE INCIDENT. David Purser The Mont Blanc road tunnel is 11.6km long and runs through the mountains between France and Italy. At 10:46 on 24th March 1999 a lorry carrying margarine entered from France and started to emit smoke in the tunnel. The driver became aware and stopped at 10.53 near garage 21 (6.5 km) from the French portal near the summit of the tunnel and just across the border between France and Italy. The driver tried to get a hand extinguisher, but very rapid fire engulfed the lorry, with flames rising quickly to the tunnel ceiling. Under normal ventilation operating conditions fresh air is blown into the tunnel from upper and lower vents along its length. The result is that at any particular time there is a neutral plane at some point along the tunnel, from which air flow at increasing velocities in both directions (towards France and Italy). However the location of the neutral plane depends upon the meteorological conditions, and the ventilation settings in the French and Italian halves of the tunnel. Due to the ventilation conditions on the day of the fire, with the neutral plane close to the point of the incident, the smoke flowed towards the French portal, and subsequent changes to the mechanical ventilation increased this flow. The driver escaped on foot towards Italy. Four vehicles from France were able to pass the lorry (PL0) before the fire and smoke became too severe, then 34 vehicles stopped behind over a distance of 500 metres. All 38 occupants of these vehicles died. Vehicles from the Italian side of the fire were able to stop and back up or turn round and escape in clear air. The fire continued to burn for three days. A team including Dr. Joel Kruppa and scientists then working at BRE, UK (Geoff Cox, Suresh Kumar, Stuart Miles and David Purser) were appointed experts to the Tribunal De Grande Instance De Bonneville Procès de la catastrophe du tunnel du Mont-Blanc, 20042006, with a brief to determine the development of the fire conditions during the incident, the effects on the tunnel occupants and the extent to which the consequences might have differed if different actions had been taken by the tunnel operators on the day of the incident. Using information on the structure and ventilation system of the tunnel and the on the conditions immediately before and during the incident determined from real time measurements in the tunnel, and subsequent investigations of the incident, the BRE team performed a set CFD analyses using JASMINE of the fire development, temperatures and effluent spread through the tunnel during the incident. In order to do this it was 30

necessary to make certain assumptions about the boundary conditions (in particular the effects of meteorological conditions and mechanical ventilation on pressure gradients along the tunnel, and the size and growth curve for the initial fire), so a number of different possible scenarios were run using different values for several of these key parameters. David Purser sought to establish the conditions faced by the vehicle occupants during the fire, their behaviours during the incident and causes of death. In order to achieve this, two independent methods were used. For one method, reports of the Mont Blanc Tunnel incident and on human behaviour in fires were used to establish the time of arrival of each vehicle and its location, the locations of deceased occupants who left their vehicles, the times taken to leave the vehicles and walking speeds and the times at which conditions caused incapacitation at different distances from PL0 (the location of burning lorry). For the second method the results of the fire modelling (CFD) analysis of Cox Miles and Kumar was used to establish the fire conditions in the tunnel at different times and locations taken from the CFD charts in terms of smoke density, heat, and toxic gases. A physiological Fractional Effective Dose (FED) analysis was then carried out to predict time to incapacitation and death for tunnel occupants and their locations within the tunnel. From the first method of analysis it was possible to establish the times at which occupants left different vehicles at different locations and attempted to move back along the tunnel towards France, and the times at which they were overcome as conditions became untenable at different locations. Using the second met hod of analysis it was possible to determine if the conditions at different times and locations predicted from the CFD analysis were compatible with vehicle occupant being able to leave their vehicles and move along the tunnel. It was also possible using Fractional Effective Dose (FED) analysis to calculate the times and locations a which each individual would have been overcome, collapsed and died in the tunnel as the conditions became untenable. By comparing the results from the two methods of analysis is was therefore possible to determine the extent to which the conditions at different times and locations predicted by the CFD analysis, as the smoke and heat front moved along the tunnel towards the French portal, were in agreement with the findings in terms of the locations and conditions of the actual decedents as determined form the incident investigation and behavioural analysis. The results of this analysis showed a good agreement between the two methods for the main fire scenario used, while other variant scenarios using different boundary conditions were able to be excluded as leading to FED tenability findings less compatible with the findings from the actual incident.

31

AN IMMERSIVE SIMULATION OF FIRE EVACUATION BASED ON VIRTUAL REALITY TANG FANGQIN, REN AIZHU, XU ZHEN, LUO YUAN Department of Civil Engineering, Tsinghua University, Beijing 100084, China Abstract: The virtual reality technology makes it possible for users to get a better understanding of dangerous scenes, e.g., a building on fire, through computational simulations. This paper presents a virtual reality system, which can conduct immersive simulations for human evacuation at fire scenes. FDS (Fire Dynamics Simulator) was selected as the tool for fire numerical simulation. By introducing the simulation results to the system, the process of fire growth and smoke spread was dynamically represented in the virtual environment. A GIS (Geographic Information System)-based computational engine was implemented to analyze the geometry features, thereby realizing the simulation of spatial cognition and way finding. The human movements and behaviors were visualized according to the evacuation calculations. The textures were set for the building models to achieve high visual quality. The users of the system can get immersed in the virtual environment when observing the simulation with specific equipments such as stereo glasses. The case study shows that the system can be used as an effective tool to form a full-scale comprehensive view of the process of fire evacuation. Key words: evacuation simulation; fire; virtual reality; geographic information system (GIS)

32

Suggest Documents