Under-ventilated Conditions Research Articles

Radiative flamelet/progress variable modelling of the UMD turbulent line fires under varying oxygen concentrations

Advanced combustion models for fire simulations remain challenging due to the complex interaction of turbulence, combustion and radiation. Among these models, the flamelet-based model is a promising candidate to incorporate detailed chemistry at low computational cost. To this end, flamelet large-eddy simulations (LES) are performed to investigate extinction characteristics of a turbulent line fire, which is a target case in the Measurement and Computation of Fire Phenomena (MaCFP) workshop and has been experimentally studied at the University of Maryland. Two radiative flamelet formulations are investigated: (a) the radiative laminar flamelet model (RLFM-u) and (b) the radiative flamelet/progress variable (RFPV-u) model. In the normal air case where XO2 = 21%, RLFM-u and RFPV-u perform similarly, both yield reasonable agreement with experimentally measured temperature. However, in the under-ventilated case (XO2 = 13%), RFPV-u outperforms RLFM-u as it manages to capture the local flame extinction. Two dimensionless indices are introduced to quantify the role of radiation heat loss and local strain in flame weakening. It is revealed that radiative heat loss dominates the flame weakening in the normal air case while strain-induced local extinction contributes significantly to flame weakening in under-ventilated conditions, which is only captured by RFPV-u, highlighting its advantages in predicting fires with strong extinction.

Experimental investigation and numerical modelling of CO and HCN release during the combustion of flexible polyurethane foam

The combustion of flexible polyurethane foam (FPUF) in fires releases relatively large amounts of carbon monoxide (CO) and hydrogen cyanide (HCN), threatening the lives of those exposed thereto. It is therefore necessary to use computational fluid dynamics method to explore the distribution of CO and HCN in a fire involving the combustion of FPUF. However, the release of CO and HCN varies significantly with the combustion conditions, which makes the numerical modelling challenging. In the present research, the yields of CO and HCN during the combustion of a non-flame-retardant FPUF were determined using a steady-state tube furnace (SSTF) facility, considering the effect of the ventilation condition. Meanwhile, a pyrolysis and combustion model of FPUF was constructed based on the fire dynamics simulator (FDS), and a four-step chemistry model was proposed to deal with the reactions in the gas phase. The experimental results showed that the yields of CO and HCN both increased with the limitation of ventilation, which was more pronounced in under-ventilated conditions. Meanwhile, the numerical simulation applying the four-step chemistry model predicted the variations in the yields of CO and HCN under ventilated conditions. In addition, when the appropriate kinetic parameters were applied to the reactions governing the oxidation of CO and HCN, the comparative results indicated that the global average deviations between the predicted and experimental yields could be reduced to 13.61 % (CO) and 8.19 % (HCN), respectively.

A comparison of carbon monoxide yields and particle formation at various global equivalence ratios in vitiated and under-ventilated conditions

There have been previous studies comparing experimental methods for the purpose of capturing gaseous yields at a range of global equivalence ratios. However, no work has investigated the capability of the open controlled atmosphere cone calorimeter for collecting such data where its two modes of operation are directly compared. The aim of this study is to compare carbon monoxide yields collected using vitiated and under-ventilated modes of atmospheric control in order to identify the preferable method of replicating carbon monoxide yields reported from larger scale enclosure fire experiments. Cone irradiances of 30, 50 and 65 kW/m2 were applied to PMMA and plywood samples. Vitiated tests were conducted using a mixed air/diluent gas, with an inflow rate of either 100, 150 or 180 L/min, resulting in a reduced oxygen concentration of 17.5 vol. %. Under-ventilated tests were conducted using flow rates of 5, 10 and 20 L/min in an air atmosphere. Particle formations and emissions were also measured using a particle analyser and have been reported herein. Results indicate that the under-ventilated mode of equivalence ratio control offers a more promising method of capturing species yields with favourable comparisons to other bench scale methods.

Experimental investigation of the smoke explosion phenomenon

This study investigates the burning history of medium-density fibre (MDF) cribs in an underventilated enclosure with two vertical openings in the front wall. Such compartment fires have been explored before by other researchers and demonstrated the conditions that would lead to smoke explosions (Sutherland, 1999; Chen, 2012) [1,2]. A total of 19 experiments were completed, 5 of which led to smoke explosions. A gas conditioning system is utilised for the experiments which included a hydrocarbon analyser and an enhanced Phi-meter. Heat release rate, mass loss rate, temperatures, pressures, and O2, CO2 and CO gas concentrations within the compartment are measured for each experiment. Compartment fires with MDF crib fuel in underventilated conditions can lead to smoke explosions for both porosity-controlled and surface-controlled cribs. The smoke explosion is found to be occurring when the changes of pyrolysate gases and oxygen concentrations in the compartment form a flammable mixture. The equivalence ratio of the compartment was found between 1.5 and 2.0 before the occurrence of smoke explosions. Transformation of flaming combustion to smouldering was generally observed as a precursor for the accumulation of unburned fuel and the occurrence of smoke explosion.

Enhancing the phi-meter by incorporating carbon dioxide measurement

This paper describes the development, calibration and experimental application of an enhanced equivalence ratio measurement apparatus, or ‘phi-meter’. The original phi-meter developed by Babrauskas et al. used only an oxygen analyser and chemically removed the CO2. In this study, an enhanced phi-meter was developed to measure the equivalence ratio from experimental enclosure fires by O2 and CO2 measurements. As with all combustion calorimetry, the accuracy of the results increases with the number of species being directly measured and removing the need for chemical removal of the CO2 is an advantage. This is of practical importance as scrubbing agent required to remove CO2 from the combustion products requires careful handling and increases test costs. Moreover, in severely under-ventilated conditions, the production of CO significantly increases, and its measurement by phi-meter provides real-time validation of the occurrence of complete combustion within the phi furnace. The original equations derived for the phi-meter incorporated the depletion of oxygen, whereas the phi-meter equations developed in this paper incorporate O2 and CO2 measurements. The equivalence ratios measured in compartment fires for heptane, LPG and timber crib fuels ranging from 0.2 for oxygen-rich to 3.0 for fuel-rich conditions are found to agree with theoretical values.

Experimental investigation of underventilated fires in enclosures with two front vertical openings and occurrence of smoke explosions

In addition to classical compartment fire development stages, smoke explosions can occur for underventilated conditions while the ventilation openings remain the same. Smoke explosions are different from backdrafts, involving changes in ventilation by additional openings. In this study, the burning history of medium-density fibre (MDF) cribs located at three different elevations has been investigated for the first time in an underventilated enclosure having two vertical openings at the front wall. The effects are examined of different fuel locations, the crib geometry and the ventilation on the mass loss and heat release rates. The fuel mass loss and heat release rates being nearly independent of the crib location and type depend on a new ventilation parameter determined from the distance between the two openings in a simplified way including effects of mass pyrolysis rate and gas temperatures. The mass air inflow is also proportional to this ventilation parameter and independent of the gas temperatures. In some experiments, however, the fires produced single or multiple smoke explosions after a dip in the heat release rate ejecting flames outside. The occurrence of a smoke explosion in relation to the porosity of crib, appearance of smouldering and the vitiated environment is also discussed.

Effects of opening aspect ratio on facade gas temperatures with and without sidewalls for underventilated conditions

This study focuses on the gas temperature distributions in the emitted flames and the hot flow on façade walls above openings without and with sidewalls at the edges of the opening. A wide range of aspect ratios of the openings was used for the first time having width over height W/H = 0.5 to 10. These situations are important when considering fire spread on facades for high-rise buildings especially combined with measurements and correlations of heat fluxes performed in previous publications of the senior author. We have performed experiments using a small compartment with the aim of developing a predictive model for maximum gas temperatures in the vicinity of the walls above the openings. Specifically, experiments were performed using gaseous burners of varying heat release rates in the compartment having openings without and with sidewalls on the side edges, of sufficient depth to prevent entrainment from the lateral sides of the opening. The heat release rate for each opening was chosen to be equal to 1800 AH1/2 (kW) where A (m2) is the area and H (m) is the height of the opening, corresponding to actual wood crib burning at under ventilated conditions. The results are explained and well correlated based on a new theory and correlations for facade fire flows using two length scales characterize these flows. One length scale represents the lateral horizontal extent of the flow parallel to the wall and the other length scale represents the extension of the flow at the opening normal to the wall before the flow becomes vertical owing to buoyancy. For high aspect ratios (W/H≥3) no significant difference is observed between the temperature distributions in fires without or with sidewalls. By contrast, for relatively tall openings (W/H < 3) the temperature distributions are different because entrainment is not allowed from the sides when sidewalls are present.

Experimental investigation and numerical simulation of the combustion of flexible polyurethane foam with larger geometries

Numerical modelling of the combustion of flexible polyurethane foam (FPUF) is challenging, as the structural collapse and shrinkage of the foam complicate its pyrolysis, and two fuel items are involved in the combustion. Two-layer pyrolysis models were established based on the bench-scale tests over the past decade, but the accuracy was limited when simulating the combustion of FPUF with larger geometries. To improve the accuracy of the numerical simulation, small-scale experiments were conducted to investigate the combustion of FPUF with a larger geometry. Firstly, numerical simulations using a two-layer pyrolysis model proposed in the most recent research were performed to simulate the combustion of FPUF in the small-scale experiments. It was found that the heat release rate (HRR) was over-predicted in the initial combustion stage. Subsequently, based on the analysis of the visual and measured data obtained from the small-scale experiments, a three-layer model was proposed to describe the pyrolysis of FPUF in flaming combustion. The three-layer model was validated with the comparison of the predicted and experimental data. The results indicate that the numerical simulation using the three-layer model has a better performance in replicating the combustion of FPUF under well-ventilated conditions. While, the capability of the three-layer model was limited when it was used to simulate the combustion of FPUF in under-ventilated conditions, as it is found that ventilation influences the HRR of FPUF to a significant extent.

Experimental and numerical studies of smoke dynamics in a compartment fire

Fires in confined spaces are of major concern in fire safety engineering. Indeed, fires with ensuing fatalities, generally occur in apartment or room fires. The decision and response time of rescue teams mainly results from empirical understanding. However, since compartment fires are multi-physical and multi-scale problems, a clear fundamental approach is needed. One of the main issues concerns the transition from localized to generalized fire. The most important vector leading to generalized fires is smoke. Indeed, smoke temperature is very high and many species i.e. burned and unburned gases are already or might mix. In this study, an experimental facility composed of a maritime container is set-up. It allows enclosure fires up to 1MW of power. This power is representative of moderate room fires. The global behavior of the smoke is investigated through the analysis of mean experimental temperature fields, smoke dynamics by large scale PIV as well as numerical simulations. For numerical simulations, the Fire Dynamics Simulator (FDS) software is used. The experimental measurements are used to evaluate the validity of FDS in under-ventilated conditions. It is observed that FDS code is able to reproduce both temperature and velocity data of enclosure fires with accuracy depending on the power to volume ratio. A criterion capable to discriminate the ventilation status of confined fire is observed and discussed.

Validation of the lumped parameter code COCOSYS against large‐scale OECD PRISME 2 fire experiments

SummaryFire safety analysis is a major issue for nuclear power plants (NPPs) in the context of deterministic safety assessments as well as of probabilistic safety analyses. Oil reservoirs and cables represent major fire loads. Therefore, simulations of oil and cable fires are of interest for quantifying the risk of such internal hazards in NPPs. To investigate the applicability of lumped parameter (LP) modelling, validations against fire experiments are required. In this way, results obtained with the LP code COCOSYS for simulations of oil and cable fire experiments conducted in the OECD PRISME 2 Project are presented. The PRISME 2 VSP (vertical smoke propagation) tests involving oil fires in a confined and mechanically ventilated facility were used to assess the ability of the LP code to simulate smoke propagation through a horizontal opening from the fire compartment to a compartment on top of it. As it was already identified in the “International Collaborative Fire Modelling Project (ICFMP),” this type of opening might cause problems in fire simulations, particularly for zone or LP fire models. In these simulations, attention has been paid to the coupling between the fire and the surrounding environment due to the decrease of oxygen concentration. Furthermore, different cable materials have been tested in the PRISME 2 CORE (completing and repeating) test campaign. By simulating the CFS‐3 (cable fire spreading) test with confined underventilated conditions, the applicability of the COCOSYS cable fire model with input parameters deduced from open atmosphere fire tests (CORE‐2) was analysed. Results show that the applicability of a LP fire model to predict the pyrolysis rate is partly limited for both oil and cable fires, in confined environment. However, simulations with prescribed pyrolysis rates show encouraging results in good agreement with the experimental data and underline the capability of the LP code COCOSYS to simulate the interaction between the thermal hydraulics inside compartments and the fire source.

Experimental study of transitional behavior of fully developed under-ventilated compartment fire and associated facade flame height evolution

For compartment fires with an opening, two distinct states are widely recognized: well-ventilated (oxygen is sufficient for combustion inside the compartment) and under-ventilated (oxygen is completely used inside the compartment). And as a consequence, the flame ejects out through the opening with excess fuel burning outside when reaching under-ventilated condition. However, in this work, a further transitional behavior of under-ventilated compartment fires with increasing fuel supply is revealed and quantified. Experiments were carried out using an under-ventilated fire compartment (0.4 m cubic) with a fixed opening width (0.25 m) and various opening heights (0.0125 m to 0.15 m) (corresponding ventilation factors, AH, 3.49 × 10−4 to 1.45 × 10−2 m2.5). The temperature inside the compartment and the facade flame ejected through the opening were recorded with increasing fuel supply (or namely total heat release rate) for under-ventilated conditions. The results showed that when the under-ventilated compartment fire reached a transitional state, the temperature inside the compartment experienced a sudden drop associated with a sudden increase of facade flame height outside the opening, posing a more severe impact on the building facade. The critical fuel supply rate for reaching the transitional state could be divided into two different mechanisms by a critical value of the opening ventilation factor ([0.5AH]=0.0033kg/s), based on the mass balance and flame stoichiometric extinction analysis inside the compartment. The formula for describing the critical fuel supply rate by considering the evolution of the critical air-fuel equivalence ratio (χ) for both mechanisms was proposed. The classic model on the facade flame height for common under-ventilated compartment fires was found to be not applicable when the transitional state of under-ventilated condition had been reached. A new model was proposed to correlate the facade flame height for the transitional state of under-ventilated condition, which was shown to have a 2/3 power dependence on the total heat release rate taking the opening width as the characteristic length.

Experimental Study about the Influence of Heat Tightness of an Enclosure Fire on Ignition Risk of Unburnt Gases in a Connected Exhaust System

This experimental study was performed from a reduced scale enclosure with a length/height and width of 2 m. In this work, we investigate whether heat tightness of enclosure can change the compartment fire dynamics and pose ignition risk in exhaust system. An external ventilation system provides an air supply rate ranging from 24 to 40 m3/h, corresponding respectively to 3 and 5 Air Change Per Hour (ACPH). A circular dodecane pan with a diameter of 40 cm is placed in the middle of the enclosure. Two situations with and without insulation of the enclosure are compared for several ACPH. The results show that heat tightness of enclosure leads to faster fire growth implying more important peak in heat release rate, and thus more dangerous fire with regarding the ignition risk. Heat tightness of enclosure makes the depression level of the compartment decrease, and consequently the air inlet supply rate decrease too, but the mass loss rate of liquid fuel increase. With heat tightness of enclosure, the fire becomes very-under-ventilated, and the vaporized fuel does not completely contribute to the heat release due to a significant formation of the unburnt gases as hydrocarbons, CO and H2. It is found that ignition of unburnt volatiles near the exhaust system occurs more easily when the compartment is more heat-tight. Oscillating flames in under-ventilated conditions were observed experimentally due to the ignition/extinction of the liquid fuel pan in a vitiated air enclosure. After the fire extinction, the remaining flames as ghosting flame in the vicinity of the enclosure ceiling were not observed under lowered oxygen vitiation with a mass fraction of around 7%. With a time delay of about 15 min in the current situation, the energy released per mass of oxygen consumed allows to raise the temperature of fuel-air mixture to about 300°C high enough for ignition of the unburnt gases. When the molar fraction of unburnt fuels is above the low flammability limit, any return of air from dilution duct after the fire extinction in a vitiated air enclosure could lead to spontaneous ignition near the extraction duct without needing a pilot flame. Note that heat-tightness of enclosure contributes to a reduction in the ignition delay time at entrance of the connected extraction duct due to a quick accumulation of unburnt fuel gases there. The experimental results from such reduced scale enclosure can be extended to a full scale one by imposing a preservation of the ratio (/L5) between heat release rate () and characteristic length (L). This relationship is derived from dimensionless variable of the energy equation.

Experimental and Numerical investigations of the flow characteristics in confined fires

Fire compartments are of major concern since they are responsible for fatal accidents. The decision and response times of rescue teams are mainly based on human decisions which result from empirical approaches. Since compartment fires are multi-physical and multi-scale problems, a perfect knowledge of the situation is impossible. A fundamental approach of fire compartments which involves fluid mechanics is of great interest. One of the main issues concerns the transition from located to generalized fire. The most important vehicle leading to generalized fires results in smoke transport. Indeed, temperatures are very high and many species are mixed. In this study, an experimental facility composed of a maritime container is set-up. In parallel, numerical simulations-Fire Dynamics Simulator (FDS) software-based on a LES approach, are used. This dual approach makes it possible to validate the FDS code in ventilated and under-ventilated conditions, considering both the temperature fields and the flow dynamics. The compartment fire allows power up to 1MW. Temperature fields are measured with thermocouples racks, and velocity fields are obtained by large scale Particle Image Velocimetry (PIV). The objectives of the paper are: I) compare and analyze temperature fields from both experimental and numerical data, II) introduce preliminary results on large scale PIV measurements located at the smoke exit. Data are also compared with simulations and III) check the numerical assumptions.

Treatment of subtalar coalition with excision and lateral column lengthening

Polymer pellets have been burned in 12 large scale tests in under-ventilated conditions in a 2.5 m3 fire enclosure with attached corridor. The polymers investigated were low density polyethylene (LDPE), polymethyl methacrylate (PMMA), polystyrene (PS), and polyamide 6.6 (PA6.6). The enclosure, based on a scaled-down ISO 9705 room attached to a horizontally divided corridor, has been designed to replicate under-ventilated burning in enclosure fires, while allowing the equivalence ratio to be determined. This provides a more robust, alternative verification of equivalence ratio, based on measured air flow, than the phi meter used in other reported work. Comparison between the yields of carbon monoxide (CO), carbon dioxide (CO2) and hydrogen cyanide (HCN), with values from the same materials from a bench-scale test together with generically similar materials, shows very good agreement, demonstrating the ability of bench-scale apparatuses, such as the steady state tube furnace (ISO 19700) and the fire propagation apparatus (ISO 12136) to replicate the high toxic yields of under-ventilated combustion. Depending on the material, CO yields between 0.1 and 0.25 g/g were obtained for under-ventilated flaming of the four polymers, alongside HCN yields of 0.01–0.02 g/g for PA6.6.

Risk analysis of unburnt gas ignition in an exhaust system connected to a confined and mechanically ventilated enclosure fire

The main purpose is to focus on the assessment of an ignition risk due to a large amount of unburnt fuel gases accumulated in the extraction duct connected to a confined and mechanically ventilated enclosure fire combining numeric and experiment. The current numerical study includes the initial well ventilated fire, spreading of flame in the enclosure, subsequent decay during under-ventilated conditions and exhaust of unburnt gas ignition in an extraction duct. Globally, Large Eddy Simulation (LES) combined with an Eddy Dissipation Concept (EDC) combustion model shows the feasibility for simulations of the air vitiation effect on transient combustion events occurring in a closed environment. A particular effort is undertaken to properly predict the pressure level inside a confined facility, and consequently, air inflow supply rate by using a HVAC system. Overall, the numerical results are in fair agreement with the experimental data for the minor species production (CO, H2), and good agreement for pressure pulse, temperature peak, the major species and heat release rate. In spite of results for minor species that could be improved, the current work confirms the feasibility of a numerical treatment of under-ventilated fire phenomena. The possibility of simulating an ignition risk in an extraction duct connected to a very under-ventilated enclosure fire, has been demonstrated with success in medium-scale facility.

Asphyxiant yields from common polymers in under-ventilated fires in the large instrumented fire enclosure (LIFE)

Polymer pellets have been burned in 12 large scale tests in under-ventilated conditions in a 2.5m3 fire enclosure with attached corridor. The polymers investigated were low density polyethylene (LDPE), polymethyl methacrylate (PMMA), polystyrene (PS), and polyamide 6.6 (PA6.6). The enclosure, based on a scaled-down ISO 9705 room attached to a horizontally divided corridor, has been designed to replicate under-ventilated burning in enclosure fires, while allowing the equivalence ratio to be determined. This provides a more robust, alternative verification of equivalence ratio, based on measured air flow, than the phi meter used in other reported work. Comparison between the yields of carbon monoxide (CO), carbon dioxide (CO2) and hydrogen cyanide (HCN), with values from the same materials from a bench-scale test together with generically similar materials, shows very good agreement, demonstrating the ability of bench-scale apparatuses, such as the steady state tube furnace (ISO 19700) and the fire propagation apparatus (ISO 12136) to replicate the high toxic yields of under-ventilated combustion. Depending on the material, CO yields between 0.1 and 0.25g/g were obtained for under-ventilated flaming of the four polymers, alongside HCN yields of 0.01–0.02g/g for PA6.6.

Hydrogen jet fires in a passively ventilated enclosure

This paper describes a combined experimental, analytical and numerical modelling investigation into hydrogen jet fires in a passively ventilated enclosure. The work was funded by the EU Fuel Cells and Hydrogen Joint Undertaking project Hyindoor. It is relevant to situations where hydrogen is stored or used indoors. In such situations passive ventilation can be used to prevent the formation of a flammable atmosphere following a release of hydrogen. Whilst a significant amount of work has been reported on unignited releases in passively ventilated enclosures and on outdoor hydrogen jet fires, very little is known about the behaviour of hydrogen jet fires in passively ventilated enclosures. This paper considers the effects of passive ventilation openings on the behaviour of hydrogen jet fires. A series of hydrogen jet fire experiments were carried out using a 31 m3 passively ventilated enclosure. The test programme included subsonic and chocked flow releases with varying hydrogen release rates and vent configurations. In most of the tests the hydrogen release rate was sufficiently low and the vent area sufficiently large to lead to a well-ventilated jet fire. In a limited number of tests the vent area was reduced, allowing under-ventilated conditions to be investigated. The behaviour of a jet fire in a passively ventilated enclosure depends on the hydrogen release rate, the vent area and the thermal properties of the enclosure. An analytical model was used to quantify the relative importance of the hydrogen release rate and vent area, whilst the influence of the thermal properties of the enclosure were investigated using a CFD model. Overall, the results indicate that passive ventilation openings that are sufficiently large to safely ventilate an unignited release will tend to be large enough to prevent a jet fire from becoming under-ventilated.

Experimental and numerical study of low frequency oscillatory behaviour of a large-scale hydrocarbon pool fire in a mechanically ventilated compartment

The contribution presents an experimental and numerical study of the oscillatory behaviour of a pool fire in a confined and mechanically ventilated enclosure. The fire safety issue concerns the loss of dynamic confinement of the enclosure due to large pressure variations and therefore the possible release of toxic products outside the compartment. The experimental analysis is based on large fire tests showing a periodic low frequency oscillatory behaviour of the burning rate. The frequency is 0.005–0.007Hz (period of 150–200s) with amplitude of about twice the mean level of the burning rate. A parametric analysis is performed to identify the most influential parameters. This oscillatory phenomenon is explained as a coupling process between the burning rate, the room pressure, the ventilation flowrate, the oxygen concentration and then a feedback effect on the burning rate. The phenomenon occurs for under-ventilated conditions for which the burning region moves within the room leading to the displacement of the flame. Numerical simulations with the fire field model ISIS is performed to check the ability of a standard CFD modelling to reproduce the flame oscillatory behaviour and to give perspective issues for numerical developments. The average values of fuel mass loss rate, compartment pressure, ventilation flow rate, oxygen concentrations and gas temperatures are well predicted. The oscillatory behaviour of the fuel mass loss rate is also obtained with a dominant low frequency although the amplitude of the fluctuations is underestimated due to a poor simulation of the flame displacement inside the compartment. The simulations points out the key effect of the pyrolysis model, the combustion model, the treatment of local extinction and the effect the ventilation flow rate.

CFD studies of the propagation and extinction of flame in an under-ventilated and over-ventilated enclosure

This piece of work is a contribution to the model-building and numerical simulation of fire phenomenon concerning fire outbreak, in confined enclosures. The numerical tool that has been used is the industrial Computational Fluid Dynamics (CFD) code called Fire Dynamics Simulator (FDS), in its 5.5.3 version. The chosen configuration of Foote and Chen et al. has allowed us to evidence the ventilation phenomena. The results of calculations that we carried out under over-ventilation conditions were satisfactory. A gap is nevertheless observed between calculations and measurements when the measurement point is located close to the air inlet duct. In under-ventilated conditions, the results are equally satisfactory, but appear to require heat flow sensitivity analysis to be excellent.

An experimental investigation and statistical characterization of intermittent flame ejecting behavior of enclosure fires with an opening

Enclosure fires with an opening are commonly distinguished by two distinct phases namely, internal combustion with no flames ejected out (over-ventilated conditions), and internal combustion extending to external combustion having flames continuously ejected outside (under-ventilated conditions). A transitional phase not previously reported is investigated in this paper with flames being ejected out intermittently with a certain probability. Experiments were carried out in a 0.8 m cubic enclosure for seven different opening geometries. A Liquefied Petroleum Gas (LPG) porous gas burner was used as a fire source with fuel supply rate controlled by a flow meter. The intermittent ejection of flames was recorded by a digital CCD camera. The fraction of the time that the flames are outside the opening is defined as a probability, P. It is found that this probability P increases with the increase in fuel supply rate above a lower critical value (at P = 0) until an upper critical value (at P = 1). These critical lower and upper fuel supply rates or equivalently the corresponding total heat release rates (HRRs), are well correlated using the so called ventilation factor ( A H ) of the enclosure, where A is the area of the opening and H is height with length scales in meter. The ventilation factor ( A H ) is proportional to the rate of air mass inflow for under-ventilated conditions. The values of the lower and upper critical HRR in excess of 1500 A H (kW) (corresponding to heat released inside at under-ventilated conditions) are normalized by a characteristic HRR value applicable for the specific opening geometry. It is proposed that the intermittent ejection of flames is due to the probability of ignition and buoyant turbulent fluctuation of gases ejected from the opening which are affected by both the excess fuel amount and the temperature inside the enclosure which is equal to the temperature of the gases exiting the enclosure. Consequently, a simple mathematical model quantifies the ejection probability P as a function of normalized excess HRR Q ˙ ex ∗ and opening factor parameter A H / A T , where A T is the total exposed surface area of the enclosure. Indeed, the first parameter Q ˙ ex ∗ is related to the excess fuel amount that is ejected outside the enclosure as the conditions change from over-ventilated to under-ventilated, whereas the second parameter A H / A T is related to the heat losses to the wall which together with the heat released inside the enclosure determine the temperature inside the enclosure. It is shown that the lower critical normalized excess HRR depends on the parameter A H / A T . It is also found that the normalized lower-to-upper critical HRR difference is constant. The results are significant for combustion and fire safety science, because the conditions for which the external flame can be continuously sustained outside are clarified based on the characterization of the intermittent transition regime. For example, it is expected that the intermittent presence of flames outside will result in increased production of carbon monoxide when no combustion occurs of the excess fuel, although the radiation impact of the flames outside the enclosure will be reduced.