Toward predictive simulations of pool fires in mechanically ventilated compartments

The objective of this work is to propose an effective modeling to perform predictive simulations of pool fires in mechanically ventilated compartments, representative of a nuclear installation. These predictive simulations have been conducted using original boundary conditions (BCs) for the fuel mass loss rate and the ventilation mass flow rate, which depend on the surrounding environment. To validate the proposed modeling, the specific BCs were implemented in the ISIS computational fluid dynamics (CFD) tool, developed at IRSN, and three fire tests of the PRISME-Door experimental campaign were simulated. They involved a hydrogenated tetrapropylene (HTP) pool fire in a confined room linked to another one by a doorway; the two rooms being connected to a mechanical ventilation system. The three fire scenarios offer different pool fire areas (0.4 and 1 m²) and air change rates (1.5 and 4.7 h⁻¹). For the one square meter pool fire test, the study presents, in detail, the effects of the boundary conditions modeling. The influence of the ventilation and fuel BCs is analyzed using either fixed value, or variable, function of the surrounding environment, determined by a Bernoulli formulation for the ventilation mass flow rate and by the Peatross and Beyler correlation for the fuel mass loss rate. The results indicate that a full coupling between these two BCs is crucial to correctly predict the main parameters of a fire scenario as fire duration, temperature and oxygen fields, over- and under-pressure peaks in the fire compartment. Variable BCs for ventilation and fuel rates were afterward both used to predictively simulate the fire tests with a pool surface area of 0.4 m². The predicted results are in good agreement with measurements signifying that the model allows to catch the main patterns characteristic of an under-ventilated fire.     

Firebrands generated from a full-scale structure burning under well-controlled laboratory conditions

Firebrand production from a real-scale structure under well-controlled laboratory conditions was investigated. The structure was fabricated using wood studs and oriented strand board (OSB). The entire structure was placed inside the Building Research Institute's (BRI) Fire Research Wind Tunnel Facility (FRWTF) in Japan to apply a wind field of 6 m/s onto the structure. As the structure burned, firebrands were collected using an array of water pans. The size and mass distributions of firebrands collected in this study were compared with sparsely available firebrand generation data from actual full-scale structure burns, individual building component tests, and historical structure fire firebrand generation studies. In this experiment, more than 90% of firebrands were less than 1 g and 56% were less than 0.1 g. It was found that size and mass of firebrands collected in this study were similar to the literature studies, yet differences existed as well. Different experimental conditions, as well as varied firebrand collection strategies, were believed to be responsible for the differences in firebrand size and mass measured in the present work, and those in the literature. The present study has provided much needed data on firebrand generation from structures.     

Auto-extinction of engineered timber: Application to compartment fires with exposed timber surfaces

A series of compartment fire tests with multiple exposed timber surfaces have been undertaken to explore the effect of exposed timber on the fire dynamics and the potential for auto-extinction. A test with exposed wall and ceiling achieved auto-extinction after approximately 21 min. Firepoint theory is applied using temperature data at the charline, is shown to predict a mass loss rate dropping below the critical value at 20–21 min, and thus is successful in predicting auto-extinction. Additional uncertainties caused by delamination are explored, and recommendations for the use of auto-extinction in design are given.     

The effect of fuel area size on behavior of fires in a reduced-scale single-track railway tunnel

A set of experiments was carried out in a 1/9 reduced-scale single-track railway tunnel to investigate the effect of fuel area size on the temperature distribution and behavior of fires in a tunnel with natural ventilation. Methanol pool fires with four different fuel areas 0.6 × 0.3 m² (1 pan), 1.2 × 0.3 m² (2 pans), 2.4 × 0.3 m² (4 pans) and 3.6 × 0.3 m² (6 pans), were used in these experiments. Data were collected on temperatures, radiative heat flux and mass loss rates. The temperature distribution and smoke layer in the tunnel, along with overflow dimensions and radiant heat at the tunnel entrance were analyzed. The results show that as the fuel area enlarges, the fire gradually becomes ventilation-controlled and the ceiling temperature over the center of fire source declines. Burning at the central region of fire source is depressed due to lack of oxygen. This makes the temperature distribution along the tunnel ceiling change from a typical inverted V-shape to an M-shape. As observed in the experiments, a jet flame appeared at tunnel entrances and both the size and temperature of the flame increased with the enlargement of fuel area leading to a great threat to firefighters and evacuees in actual tunnel fires.     

Interactive fire spread simulations with extinguishment support for Virtual Reality training tools

Virtual Reality training for fire fighters and managers has two main advantages. On one hand, it supports the simulation of complex scenarios like big cities, where a fire cannot be simulated in the real world. On the other hand, fire fighting VR simulators allow trainees to experience situations as similar as possible to real fire, reducing the probability of accidents when they are practising exercises with real fire.The success of the Virtual Reality training tools also depends on how close to reality the simulation process is. This work provides fire spread algorithms for forest and urban environments, which can be used at interactive rates. Due to the interactive nature of the algorithms, the users are able to fight the fire by throwing extinguishing agents.Although the algorithms assume many simplifications of the problem, their behaviour is satisfactory. This is due to the efficient management of the cell states in a 3 m×3 m cell grid. Also the variables that have more influence on fire propagation constitute the core of the algorithms. The overall system deals with user extinguishment actions, natural and artificial firebreaks, variable wind conditions (even at a cell level) and non-contiguous fire propagation (embers and spotting fires). The unified forest/urban model leads to an object oriented architecture which supports the fire propagation algorithms. This also allows the system to compute efficiently mixed forest–urban environments.     

Predicting the structural response of a compartment fire using full-field heat transfer measurements

Inverse heat transfer analysis (IHT) was used to measure the full-field heat fluxes on a small scale (0.9 m×0.9 m×0.9 m) stainless steel SS304 compartment exposed to a 100 kW diffusion flame. The measured heat fluxes were then used in a thermo-mechanical finite element model in Abaqus to predict the response of an aluminum 6061-T6 compartment to the same exposure. Coupled measurements of deflection and temperature using Thermographic Digital Image Correlation (TDIC) were obtained of an aluminum compartment tested until collapse. Two convective heat transfer coefficients, h =35 W/m²-K and h =10 W/m²-K were examined for the thermal model using the experimentally measured heat fluxes. Predictions of the thermal and structural response of the same compartment were generated by coupling Fire Dynamics Simulator (FDS) and Abaqus using the two values for h, h =35 W/m²-K and h from convection correlations. Predictions of deflection and temperature using heat fluxes from IHT and FDS with h=35 W/m²-K agreed with experimental measurements along the back wall. The temperature predictions from the IHT-Abaqus model were independent of h, whereas the temperature predictions from the FDS-Abaqus model were dependent on h.     

An experimental study on backdraught: The dependence on temperature

This paper presents the results of a series of reduced scale experiments to investigate the temperature conditions leading to backdraught in a fire compartment (0.8 m×0.4 m×0.4 m), using solid polypropylene pellets as the fuel. The factors of primary interest are the pre-burn time, before the fire becomes oxygen limited, the duration of door closure, and the temperature distribution in the compartment. It is shown that the temperature inside the compartment is crucial for the occurrence of backdraught. Above 350 °C, backdraught by auto-ignition is possible. If a pilot spark is present, backdraught may occur at temperatures down to 300 °C. It is shown that backdraught conditions can be achieved in the early stages of a fire as long as a suitable temperature is reached, at considerably lower temperatures than those generated during flashover. Further investigation on gas concentration is essential to understand the chemistry of backdraught combustion.     

Determination of smoldering time and thermal characteristics of firebrands under laboratory conditions

The laboratory experiment was conducted to simulate the transfer of smouldering particles produced in forest wildfires by a heated gas flow. The pine bark pieces with the linear dimensions L=(15; 20; 30) mm and a thickness of h=(4−5) mm were selected as model particles. The rate and temperature of the incident flow varied in the range of 1–3 m/s and 80–85 °C, respectively. The temperature of the samples was recorded using a thermal imager. To determine the minimum smouldering temperature of pine bark, the thermal analysis was conducted. The minimum smouldering temperature of pine bark was found to be 190 °C. This temperature will cause thermal decomposition of bark only at the first stage (oxidation of resinous components). In the study the smouldering time, the temperature and the weight of samples were obtained and analyzed under various experimental conditions. The data analysis shows that the increase in the particle size leads to the decrease in their mass loss, and the rate change of the incident flow does not practically influence the mass change. For particles with the linear dimensions of 10 mm and 20 mm, the mass varies from 6% to 25%. The maximum mass loss is observed for the flows with a rate of 1 and 2 m/s. The results have shown that the increase in the particle size leads to the increase in the smouldering time. The position of the particle plays an important role, the effect of which increases with increasing the particle size. The calculations showed that the smouldering time of bark samples is long enough for the particles to serve as new sources of spot fires. The particles were found to be transported to a distance of 218 m from the fire line which can certainly influence the propagation of the fire front.     

Investigation on combustion of fire retardant board under different N2–O2 mixture gas atmospheres by using thermogravimetric analysis

By thermogravimetric analysis (TGA) and kinetic study, the thermal decomposition and combustion processes of a fire retardant board were investigated under different N₂–O₂ mixture gas atmospheres, with 20%, 15%, 10% and 5% oxygen, respectively. The samples were heated from 373 K to 1223 K with a constant heating rate of 30 K/min. The trend of the thermogravimetry (TG) curves and derivative thermogravimetry (DTG) curves were interpreted regarding the conversion yields and the reactivity of samples. There are nine reaction stages which are discernible easily from DTG curves for samples under different gas atmospheres. Each sample’s top value among maximum mass loss rates of all reaction stages appears during reaction 4 stage. The reaction orders and apparent activation energies of all samples during all reaction stages vary between 1.4 and 4.0 and between 7.358 kJ/mol and 535.555 kJ/mol.     

Automatic fire detection in road traffic tunnels

Fire detection experiments in a road traffic tunnel were performed in the Runehamar test tunnel 5th–8th March 2007. The Runehamar test tunnel is a full profile road traffic tunnel, 1.65 km long, located outside Åndalsnes, Norway. The goal was to examinate smoke and heat detection systems to determinate what kind of principle best suited for detecting a fire in an early stage. The systems were tested during small Heptane pool fires, varying between 0.16 m² and 1 m², giving heat release rates from 0.2 MW to 2.4 MW accordingly, and one car fire of about 3–5 MW, and with wind conditions varying from 1.1 m s^?1 to 1.6 m s^?1. The size of the fires, were designed to be in the range from impossible to difficult to detect. The results were conclusive. Earliest detection of a car fire, fire starts inside, was by smoke detection given fixed limits (3000 μg m^?3). With open pool fires, or immediate flames, continues fibre optical heat detection systems was faster given the limits 3 °C/4 min.     

Fracture behavior of a four-point fixed glass curtain wall under fire conditions

The cracking and subsequent fallout of glazing could significantly affect compartment fire dynamics by creating a new opening for air to enter. Twenty-four 1200×1200×6 mm³ soda-lime glass panes in eight different fixing forms were heated by a 500×500 mm² N-heptane pool fire to investigate the influence of fixing conditions on glass breakage and fallout. The time of crack initiation, behavior of crack propagation, heat release rates, central gas temperatures, glass surface temperatures and loss of integrity of the glazing assembly were investigated. The relationship between fixing form and crack behavior is discussed, based on the experimental results. The results show that all the cracks initiated at the supporting point and intersected rapidly, causing glass fallout. Mechanical stress caused by supporting pins and thermal stress caused by glass temperature difference (ranging from 48 °C to 159 °C) are the causes of breaking for this kind of curtain wall. It is concluded that various fixing locations have a significant effect on glass breaking. Among the eight cases, the glass panes whose supporting points were located at 10 cm (Case 1) or 5 cm (Case 8) from the edges performed best: these support locations are recommended in practical engineering because of the good fire resistance and structural beauty of such panes.     

Deflagrations involving stratified heavier-than-air vapor/air mixtures

The distribution of explosion damage in a structure is a major indicator of the type of explosive material involved and its location. A solid-phase explosive material typically produces localized or “seated” damage, where a vapor/air explosive mixture typically produces generalized, omnidirectional damage. Investigators have been taught that the finding of more intense blast damage to lower portions of an enclosure indicates that the vapors were heavier than air, while explosion damage to upper portions indicates a lighter-than-air gaseous fuel. Most of the explosion pressure data in the literature deal with well-mixed mixtures that are uniform in concentration prior to ignition. This study explores the pressure distributions produced by the ignition of shallow (0.05–0.2 m deep) layers of hexane vapor created by the evaporation of liquid in a still, isothermal compartment. The floor-level vapor layers thus produced were ignited by an electric arc and the pressures at five different locations in the room were monitored. It was found that pressures increased in an exponential fashion over a period of 300–400 ms after ignition until the relief panel failed (at ∼5–6 kPa). The peak pressures observed at all five locations in the compartment coincided in time (to within ±5 ms) and intensity suggesting that the pressures produced within the 3.6 m×2.4 m×2.4 m chamber equilibrated very quickly. Any failure of the compartment, then, would be the result of failure of the weakest part of the confining structure, rather than the result of pre-ignition distribution of the fuel/air mixture. A small (∼−2 kPa), but reproducible negative pressure peak was observed some 60–70 ms after the maximum positive pressure. This finding shows that negative pressure peaks can be produced by deflagrating vapor/air mixtures that could exert physical effects on lightweight debris dislodged by the initial positive pulse.     

Smoke concentrations inside and outside of a corridor-like enclosure

This work presents smoke measurements and correlations inside and outside of a corridor-like enclosure fires in order to determine the effects of burning on smoke concentrations inside and outside the enclosure. Thirty eight experiments were performed in a three metre long corridor-like enclosure having a cross section 0.5 m×0.5 m, door like openings in the front panel and a gaseous burner located near the closed end. Smoke concentrations were measured at two locations inside the enclosure and also in the exhaust duct of a hood collecting the fire gases from the enclosure. It was found that smoke concentration in the exhaust duct decreased whereas smoke concentration inside the enclosure increased after the flames started moving towards the opening and external burning occurred. This increased smoke concentration inside the enclosure was caused by reversion of the flow pattern inside the enclosure after the flames moved past a point towards the opening. Namely, the flow pattern changed direction behind the flame front in the sense that hot gases in the upper layer were travelling backwards towards the closed end of the corridor thus contributing to smoke increase inside the enclosure. This change of flow pattern was confirmed in all experiments by bidirectional probe velocity measurements in the upper and lower layer as well as by oxygen concentrations and temperature measurements inside the enclosure. These results are useful for CFD validation and specifically applicable for assessing smoke hazards in corridor fires in buildings where smoke concentrations can be much larger than anticipated owing to leakage to adjacent rooms behind a moving flame front.     

An experimental study on the intermittent extension of flames in wind-driven fires

Experiments were conducted to study the intermittent extension of flames from wind-driven line fires using stationary burners. These fires are thought to share similar features with propagating wildland fires, where forward pulsations of flame have been observed to quickly ignite material far ahead of the mean flame front. However, stationary burners offer the ability to study the movement of the flame and its heating processes in greater detail than a spreading fire. In these stationary experiments, propane gas was used as a fuel with different burner sizes, 25–30 cm wide and 5–25 cm long in the direction of the flow. A specially-built wind tunnel was used to provide a well-characterized laminar flow for the experimental area. The free-stream flow velocity, measured by a hot-wire anemometer, ranged in the experiments from 0.2 to 2.7 m/s. The shape of the flame was measured using a high-speed video camera mounted perpendicular to the apparatus. A method was developed to track the extension of the flame close to the surface, simulating flame contact with unburnt fuel downstream of the fire. This extension length was then measured frame by frame and frequencies of flame presence/absence determined as a function of downstream distance. The location of maximum pulsation frequency, x_max, for each burner/wind configuration, was obtained using a level-crossing approach (essentially the variable-interval time-average (VITA) method). Further study indicates that x_max can be well estimated using mean flame properties. Probability distributions describing the location of the flame over time also showed that, the probability the flame extends far beyond the mean flame front is sensitive to increasing ambient winds and fire size.     

Gas temperatures in heavy goods vehicle fires in tunnels

Large-scale fire tests were carried out with heavy goods vehicle (HGV) cargos in the Runehamar tunnel in Norway. The tunnel is a decommissioned, two-way-asphalted road tunnel that is 1600 m long, 6 m high and 9 m wide, with a slope varying between 0.5% uphill and 1% downhill. In total four tests were performed with fire in an HGV set-up and a longitudinal ventilation flow of approximately 3 m/s. In three tests, mixtures of different cellulose and plastic materials were used; in the fourth test a commodity consisting of furniture and fixtures was used. In all tests the mass ratio was approximately 82% cellulose and 18% plastic. A polyester tarpaulin covered the cargo.One purpose of the large-scale tests was to obtain new relevant gas temperature-time data from large-scale HGV fires in tunnels. There is presently a lack of such information for road tunnels. The maximum heat release rates produced by the four different fire loads varied between 66 and 202 MW resulting in maximum gas temperatures at the ceiling ranging between 1281 and 1365 °C. A comparison with literature values shows that the gas temperatures obtained here are uniformly higher than those obtained in other similar large-scale test series conducted using solid materials. A mathematical correlation of a temperature–time curve is given and this is the best representation of the measured temperature and a combination of frequently used temperature curves for tunnels (the HC curve and the RWS curve).     

Extinguishing characteristics of a diffusion flame with water vapor produced from a water droplet impacting onto a heated plate

In this study, we have investigated an extinguishing method of a diffusion flame with water vapor produced from a water droplet impacting onto a headed plate, which is called as indirect fire attack. In order to clarify the extinguishing characteristics, the extinguishing experiments of a methane-air diffusion flame have been performed by using a pure water droplet with the diameter of 3.2 mm. The droplet dropped from the height of 400 mm. The wall superheat and the burner height were varied from 0 K to 330 K and from 32 mm to 102 mm, respectively. As a result, under certain wall-heat conditions, the water-vapor vortex ring is formed and visualized by the white water fog. At wall superheat of 150 K, the formation probability of the vortex ring is unity and the extinguishing probability always shows the peak values regardless of the burner height. As a result, it can be said in our study that the wall superheat of 150 K is the most effective value for the indirect extinguishing method.     

Simulation of fire scenarios due to different vehicle types with and without traffic in a bi-directional road tunnel

This paper presents findings obtained by CFD modelling for simulating the effects of fire due to different vehicle types in a bi-directional road tunnel. Four different burning vehicles placed in the centre of the driving lane at tunnel middle length were considered. Peaks of the heat release rate (HRR) of: 8, 30, 50, and 100 MW were simulated for the two cars, the bus, the heavy goods vehicle (HGV), and the petrol tanker, respectively. The fire effects on tunnel structure and on environmental conditions along people evacuation path were especially evaluated. The effects of the traffic jam, in contrast with the isolated vehicles, on temperatures, radiant heat flux, visibility distance, and toxic gases concentrations, were also investigated. The worst scenario was identified to be that pertaining to the petrol tanker and more critical conditions were also found when the tunnel was full of vehicles. The maximum gas temperatures reached in the presence of traffic at the side wall (and at the tunnel ceiling reported in brackets) were found to be: 360 °C (170 °C) for the two cars; 740 °C (465 °C) for the bus; 835 °C (735 °C) for the HGV and 1305 °C (1145 °C) for the petrol tanker, respectively. The presence of the traffic, in contrast with the isolated vehicle, involved an increase in the maximum temperatures equal to 16–17% for the two cars, and contained in the range 12–29% with percentages increasing starting from the tanker, to the HGV and to the bus. In other words when the maximum temperatures produced by the isolated vehicle are very high (e.g. for the tanker), the presence of the traffic had a minor effect. With reference to environmental conditions along the evacuation path, the results showed that in the case of petrol tanker fire the emergency ventilation ensures a tenable level of temperature, radiant heat flux, and toxic gases concentrations up to 5 min from the fire starting. This time increases up to 6.5 min for the HGV and 8 min for the bus. This means that the tunnel users in order to be safe in all scenarios should leave the tunnel within 5 min after the fire starting. Toxic gases concentrations, however, were found to be below the limit values in all cases and also in the presence of traffic. In the light of the aforementioned results, tunnel occupants should be promptly informed of the fire risk and guided to the exit portals. This might be done by equipping the tunnel with illuminated emergency signs located along the tunnel length and by installing traffic lights before the entrances so that the tunnel can be closed in case of emergency. By activating the traffic lights at the portals and the emergency signs (more especially those at the ceiling) at the same time as the emergency ventilation is activated, safer conditions for the people evacuation are expected.     

Improvement of residual compressive strength and spalling resistance of high-strength RC columns subjected to fire

This paper investigates the spalling properties of high-strength concrete in order to improve the residual compressive strength and spalling resistance in specimens subjected to 3 h of unloading fire conditions. This study consists of three series of experiments with eighteen different specimens varying in fiber type and content, finishing material and simultaneous fiber content and lateral confinement. They were fabricated to a 300 × 300 × 600 mm mock-up size. Results of the fire test showed that the control concrete was explosive, while the specimens that contained more than 0.1 vol% of polypropylene (PP) and polyvinylalcohol (PVA) fibers were prevented from spalling. One specimen, finished by a fire endurance spray, exhibited even more severe spalling than the control concrete. The specimen containing 0.1 vol% of PP fiber and using a confining metal fabric at the same time, showed the most effective spalling resistance; in particular, the residual compressive strength ratio was even higher than that of the control concrete before the fire test. It was demonstrated that adding fibers in concrete prevented the spalling occurrence and confining metal fabric around the main bars of concrete specimens can secure the strength of structures during the conditions of elevated temperature.     

Heat release characteristics of ethanol-water mixtures: Small-scale experiments

Small-scale pool fire tests of ethanol-water mixtures are described in this paper. The main goal of the tests was to obtain data on the heat release rate per unit area, the total heat release per unit area, and the effective heat of combustion of such mixtures with ethanol percentage of 20 vol% or more. The results can be utilised in the interpolation and extrapolation of these quantities for different ethanol percentages and irradiance levels. In addition, the results give information on the heat release of possible pool fires of alcoholic beverages in a retail store in the case of a fully developed fire. It was concluded that significant fire load is not formed until the ethanol percentage of the alcoholic beverage is ca. 20 vol% or more. The data reported can also be used as an input to simulations of ethanol-water mixture pool fires in the performance-based fire safety design.