Search

The use of fire models to support fire protection engineering decisions requires an understanding of model shortcomings and assurance in their predictive robustness. This note is a summary of the online 'compartment fire' workshop that was organized prior to the International Association of Fire Safety Science (IAFSS) symposium, hosted online by the University of Waterloo (Canada) in April 2021. The objectives of the Workshop were to identify, discuss and prioritize key compartment fire modeling challenges. It is recognized that the substantial changes in the built environment and the variety of subsequent fire dynamics problems necessitate significant advances in modelling, particularly with respect to (amongst other aspects) (i) under-ventilated fires (the prediction of soot and CO concentrations, extinction and reignition), (ii) heat transfer and the interaction with structural and non-structural elements, and (iii) the interaction with water-based fire suppression systems. High-quality and well-documented experimental tests play an essential role in fostering model development and validation; a good synergy between experimentalists and modellers is of utmost importance. • Key compartment fire modeling challenges are discussed. • Significant advances in modelling are necessary. • High-quality and well-documented experimental tests are of utmost importance. • A strong collaboration between modelers and experimentalists fosters progress. [ABSTRACT FROM AUTHOR]

The paper presents experimental measurements for the transient profiles of burning rates and liquid temperatures (at different depths below the surface) in methanol and heptane pool fires. Square burners are made of steel walls and have 4 different side lengths: 5.5, 8.0, 10.5 and 15 cm. The burner is 5 cm deep and the initial fuel height is about 4.6 cm. An experimental procedure has been developed to measure the enhancement in heat transfer within the liquid due to convective motion, by estimating the internal Nusselt number, Nu i. The results suggest that for heptane Nu i = 11.8 ± 3.6 and for methanol Nu i = 6.2 ± 2.9. Additionally, based on the experimental data, a two-zone approach for the thermal structure within the liquid has been proposed. This approach will be implemented in a numerical code. • In-depth liquid temperature measurements are carried out for pool fires. • The internal Nusselt number is 11.8 ± 3.6 for heptane and 6.2 ± 2.9 for methanol. • A two-zone approach is proposed for the thermal structure within the liquid. [ABSTRACT FROM AUTHOR]

Abstract Focusing on advancing predictive fire modeling, large eddy simulations using improved approaches related to thermophysical, turbulence, combustion and radiation modeling are presented. More specifically, the consideration of a non-unity Lewis number, the use of the Hirschfelder-Curtiss diffusion model, the inclusion of differential diffusion and Soret effects, the application of the dynamic Smagorinsky model with a variable turbulent Prandtl number, an eddy dissipation concept model for combustion and the weighted-sum-of-gray-gases model for radiation have been included in a modified version of fireFoam 2.2.x. A comprehensive comparison between the predictions of the modified code against the standard version of fireFoam and experimental data of a medium-scale 22.6 kW methanol pool fire is presented. Evaporation modeling is not yet included. [ABSTRACT FROM AUTHOR]

The work described in this paper is undertaken with the purpose of providing a detailed assessment of the current modelling capabilities of the effects of fire suppression systems (e.g., sprinklers) in fire-driven flows. Such assessment will allow identifying key modelling issues and, ultimately, improving the reliability of the numerical tools in fire safety design studies. More specifically, we studied herein the heating and evaporation of a single water droplet. This rather 'simple' configuration represents the first step in a tedious and rigorous verification and validation process, as advocated in the MaCFP (Measurement and Computation of Fire Phenomena) working group (see https://iafss.org/macfp/). Such a process starts ideally with single-physics 'unit tests' and then more elaborate benchmark cases and sub-systems, before addressing 'real-life' application tests. In this paper, we are considering the recently published comprehensive and well-documented experimental data of Volkov and Strizhak (Applied Thermal Engineering, 2017) where a single suspended water droplet of initial diameter between 2.6 and 3.4 mm is heated up by a convective hot air flow with a velocity between 3 and 4.5 m/s and a temperature between 100 and 800 °C. In the present numerical study, 36 experimental tests have been simulated with the Fire Dynamics Simulator (FDS 6.7.0) as well as with an in-house code. The results show that the droplet lifetime is overpredicted with an overall deviation between 26 and 31%. The deviation in the range 300–800 °C is even better, i.e., 5–8%, whilst the cases of 200 and, more so 100 °C, showed much stronger deviations. The measured droplet saturation temperatures did not exceed 70 °C, even for high air temperatures of around 800 °C, whereas the predicted values approached 100 °C. A detailed analysis shows that the standard Ranz & Marshall modelling of the non-dimensional Nusselt and Sherwood numbers may not be appropriate in order to obtain a simultaneous good agreement for both the droplet lifetime and temperature. More specifically, the heat-mass transfer analogy (i.e., Nu = Sh) appears to be not always valid. [ABSTRACT FROM AUTHOR]

Abstract The paper presents a detailed modelling approach of liquid pool evaporation that is suitable for pool fires and which relies on a combination of mass and heat transfer-based models. The mass transfer approach is based on the so called 'film theory' where mass diffusion is driving the molecular transport process across the film thickness. This driving mechanism is dominant during the early stages of the fire. However, as the liquid surface approaches the boiling point, heat transfer takes over and the evaporation rate becomes controlled by the heat fluxes at the liquid surface. In order to illustrate the potential of the proposed 'comprehensive approach', an in-house code has been developed and validated against a standardized experimental test for the vaporization rate of water in the ASTM E2058 fire propagation apparatus under an external heat flux of 50 kW/m2. The results show that, when the film thickness is calculated assuming a natural convection regime instead of a turbulent forced convection regime, the prediction of the early stage of the vaporization process is significantly improved in comparison to previously published results. However, the water evaporation case appeared to be solely driven by mass transfer (i.e., water vapor concentration gradients). Therefore, additional illustrative numerical tests with several liquid fuels and higher heat fluxes have been carried out. These tests point out the importance of combining the mass transfer-based approach with the heat transfer-based approach when the liquid surface reaches the boiling point. Highlights • A 'comprehensive approach' is proposed for liquid pool evaporation, combining heat and mass transfer-based modelling. • An in-house code is developed and validated using a standardized experimental water vaporization test. • Better results are obtained with natural convection correlations as opposed to forced turbulent convection. • The potential of the 'comprehensive approach' is demonstrated for several liquid fuels. [ABSTRACT FROM AUTHOR]

The paper presents a comprehensive set of numerical simulations performed to examine the current Computational Fluid Dynamics (CFD) capabilities in the prediction of the interaction of a water mist spray with a vertical upward jet of hot air within an Eulerian-Lagrangian framework. The experimental tests considered herein are described by Zhou [Proceedings of the Combustion Institute, 2015]. The spray is a 30° full cone water mist spray emerging from a nozzle that delivers a water flow rate of 0.084 lpm at a pressure of 750 kPa. The vertical jet of hot air at 205 ∘ C is issued from a 72 mm-diameter nozzle placed at 560 mm below the water spray nozzle. Three exit velocities of 3.3, 4.2 and 5.3 m/s were examined. Gas phase simulations (described in the companion paper, Part I) have allowed to determine a set of parameters (e.g., cell size of 4 mm and modified Deardorff model for the turbulent viscosity) that are suitable for the water mist spray simulations. Moreover, it is shown here that a prescribed complex spray pattern with a full discharge angle of 60° is required in order to match water spray profiles in the nozzle near-field. The three regimes of spray-jet interaction (i.e., water spray dominated, vertical jet dominated or equal influence of the spray and the vertical jet) are qualitatively well captured by the numerical simulations. However, the location of the interaction boundary is underestimated by up to 26%. This could be partially attributed to modelling aspects related to, for example, turbulent dispersion or turbulence inflow conditions of the droplets. Uncertainties in the experimental measurements must also be considered. [ABSTRACT FROM AUTHOR]

The paper reports a comprehensive set of large-eddy simulations (LES) of a turbulent hot air jet impinging onto a ceiling. The hot air source is a 72-mm diameter circular nozzle with an exit temperature maintained at 205 °C. Three exit velocities have been tested: 3.3, 4.2 and 5.3 m/s, corresponding to Reynolds numbers of respectively 6800, 8600 and 10900 and Froude numbers of respectively 3.9, 5.0 and 6.3. The horizontal aluminium ceiling plate of 1.22 m × 1.22 m has been placed at a distance of 590 mm above the hot air nozzle. This configuration has been examined experimentally by Zhou [Proceedings of the Combustion Institute, 2015] to characterize gas phase conditions prior to experiments which aim at studying the interaction between hot air jets and water sprays. This paper constitutes the first part of a numerical study that aims at assessing the current modelling capabilities of the two-phase flow configuration examined by Zhou [Proceedings of the Combustion Institute, 2015]. The results show that the centerline mean vertical velocity profiles of the vertical jet are predicted with maximum deviations of less than 6% from the experimental data at the condition of an appropriate set-up of the inflow conditions (i.e., geometry of the inlet and turbulence inflow boundary conditions). Furthermore, the best results were obtained with the dynamic Smagorinsky model for the turbulent viscosity. The modified Deardorff results are nevertheless very good given the substantial decrease in computational time (in comparison to the dynamic Smagorinsky model). A good prediction of the vertical jet allowed relatively good predictions of the ceiling jet maximum velocity, boundary layer thickness and Gaussian momentum width with maximum deviations of respectively 20%, 1 mm and 18%. The numerical modelling of the gas phase described in this paper can thus be relied upon in the two-phase simulations described in the companion paper [Part II: Two-phase flow simulations]. [ABSTRACT FROM AUTHOR]

Large eddy simulations (LES) using advanced modelling approaches related to thermophysical, turbulence and combustion modelling are presented and their advantages when compared to some of the standard models used in the fire community are analyzed. More specifically, the consideration of a non-unity Lewis number and the Hirschfelder–Curtiss diffusion model, the inclusion of differential diffusion and Soret effects, the application of a dynamic turbulence model with a variable turbulent Prandtl number formulation, along with the EDC combustion model, have been included in a modified version of FireFOAM 2.2.x. A comparison between the predictions of the new and the standard models available in the code against experimental data of a medium-scale 24.6 kW methanol pool fire is presented. The predictions with the advanced modelling approaches are qualitatively and quantitatively better when compared to the standard models in the code, while having only a 20% increased computational cost. [ABSTRACT FROM AUTHOR]

The paper presents a detailed sensitivity analysis on the volume flux probability density function (PDF) to represent water spray patterns with computational fluid dynamics (CFD). The effects of the turbulent viscosity model and the cell size are also investigated. The test case considered herein is a 30 ° full cone water mist spray emerging from a nozzle that operates at a pressure of 750 kPa and delivers a water flow rate of 0.084 lpm. The errors solely induced by the limited number of computational droplets per second, N p , are proportional to 1 / N p and could reach up to 35%. The computational time generally increases linearly with N p . The paper illustrates also the better numerical performance of the lognormal-Rosin-Rammler droplet size distribution over the Rosin-Rammler distribution, especially in terms of reaching a converged volume-median diameter with increased N p . Furthermore, a uniform angular distribution is shown to provide results in better agreement with experimental data than a Gaussian-type distribution for the case at hand. For a sufficiently fine grid, the dynamic Smagorinsky and the modified Deardorff models converge to similar radial profiles of the water volume flux at 300 mm from the nozzle, with a deviation of less than 6% from the experiments. The deviations for the volume-median diameter are about 50% in the core region of the spray. [ABSTRACT FROM AUTHOR]

Numerical simulations of fire-induced smoke movement in the stairwell of a high-rise building are conducted using FDS, version 6.0.1, with default settings. Twelve scenarios are considered. The required fineness of the grid has been determined in earlier work by considering both the fire source and the vent flow, and by assessing the velocity profile at the bottom opening and the vertical distribution of temperature in the stairwell. In the present study, the results including the airflow velocity at the bottom opening, vertical distribution of temperature, the temperature at the middle opening, pressure distribution, and neutral plane height in the stairwell, are compared to experimental data. For the average velocity through the bottom opening, a maximum deviation of 16.23% is obtained. Good agreement is achieved for the vertical temperature inside the stairwell (maximum relative deviation of 12.3%). By analyzing the temperature at the middle opening, it is found that the smoke moves faster than in the experiment. The influence of the staircase on the pressure distribution is demonstrated by comparing two cases: one with and one without staircase. The difference between the pressure inside the stairwell and the pressure outside increases with height, due to fire-induced buoyancy. However, the pressure difference evolution is non-monotonic when there are staircases inside the stairwell. The neutral plane height value, as obtained by post-processing the simulation results, is too high in the simulations, compared to experimental data and the corresponding analytical expression. Finally, the influence of the turbulence model is shown to be negligible. [ABSTRACT FROM AUTHOR]

A global analysis is presented of a series of four multi-compartment full-scale fire tests carried out in an apartment located in a high-rise building (called the ‘Rabot’ tower in the city of Ghent, Belgium) in 2012. The data set, available at http://multimedialab.elis.ugent.be/rabot2012/ , is called ‘Rabot2012’. These tests are believed to provide a valuable set of experimental data for fire modelling (CFD and two-zone) and monitoring purposes, for a number of reasons. First of all, the fuel package of intermediate complexity ( i.e. two to three furniture items) and the ignition location were designed to focus on the initial flame spread and then the possible occurrence of secondary (and in one test tertiary) ignition. Also, very good repeatability is illustrated during the first minutes of the tests. Regarding the sensing aspect, in addition to the conventional tools such as thermocouples and velocity probes, a Video Fire Analysis (VFA) has been performed by applying smoke and flame detection algorithms to the footages obtained from a multi-view video network. Moreover, laboratory tests have been performed, putting the fuel packages as applied in the full-scale tests under a hood in free-burning conditions. The repeatability during the first minutes is confirmed and subsequent differences in fire development (free burning versus compartment set-up) are explained. This presentation of the ‘Rabot2012’ fire tests comprises: (1) a detailed description of the experimental set-up ( i.e. geometry, fuel package and instrumentation), (2) an overview of the fire scenarios obtained in each of the four tests, and (3) a detailed account of the data processing carried out. Finally, elements of exploitation of the collected set of data are discussed. [ABSTRACT FROM AUTHOR]

This paper presents a comprehensive set of numerical simulations for the characterization of a water mist spray emerging from a nozzle positioned at 2.2 m from floor level and operating at a pressure of 1.0 MPa. The droplet volume-median diameter is about 90 μm and the spray half-angle is around 42 °. The spray shape is visualized using a laser sheet and the water flux density distribution on the ground is measured with a 'bucket' test. An initial comprehensive numerical study using the Fire Dynamics Simulator (FDS) has been carried out by varying several numerical and physical models and parameters (e.g., cell size and turbulence modelling). The simulated sprays were very narrow (in comparison to the actual spray), yielding overestimations of the peak water flux density at floor level by about 430% (on average). Subsequently, it was found that there is a significant impact of the drag force modelling because the spray at hand is dense (near the nozzle). An ad-hoc reduction of the drag coefficient to a constant value leads to better results. The current study calls upon new developments for droplet aerodynamic modelling in dense sprays. [ABSTRACT FROM AUTHOR]

The work described in this paper presents a set of experimental tests carried out in a Controlled Atmosphere Cone Calorimeter (CACC) in order to study the evaporation rate of a 250 ml-'batch' of n-heptane and methanol poured into an insulated 0.10 × 0.10 × 0.04 m3 steel pan and exposed to two 'nominal' irradiation levels of 25 and 50 kW/m2. A low oxygen concentration is imposed during the CACC tests. The intent is to provide experimental data for the assessment and validation of the numerical modelling of liquid heat-up and evaporation in pool fires, without the need to model combustion and the subsequent associated complex phenomena. The analysis of the mass loss rate profiles shows that the evaporation rate of n-heptane increases with time due to the preheating of the liquid, whereas for methanol, a quasi-steady-state is reached. This is due to differences in specific heat, boiling point and latent heat of vaporization. Numerical simulations of the liquid phase using the Fire Dynamics Simulator (version 6.7.0) point out the importance to account for increased heat transfer due to the convective motion of the liquid by increasing the thermal conductivity of the liquid by 10–20 times. Other limitations and uncertainties in the modeling are also highlighted. [ABSTRACT FROM AUTHOR]

The present paper studies the fire-induced pressure evolution in an air-tight compartment. Two cases with different sizes (i.e., 0.3 m × 0.3 m and 0.5 m × 0.5 m) of square methanol pool fire are considered. Repeated tests were performed for each case, leading to a total of four fire experiments. The leakage of the compartment is measured under different pressure conditions in order to obtain the relationship between the leakage and the pressure increase. The pressure inside the compartment reaches peak values of 133 Pa and 697 Pa for the small and the large pool fire, respectively. Simulation results using the Fire Dynamics Simulator (FDS) 6.5.3 match well with experiments. The simulated leakage flow rates, which reach values up to 65 m3/h and 200 m3/h, respectively, depend on the pressure variation and leakage settings. The pressure rises strongly in the growth phase of the fire, in line with a rapid increase in fire heat release rate. The gradually increasing heat losses subsequently lead to a decrease in pressure. [ABSTRACT FROM AUTHOR]

Computational Fluid Dynamics (CFD) simulations of fire-induced pressure and ventilation flow in a mechanically-ventilated air-tight compartment, representative of a passive house, are presented. Experiments conducted by University of Mons are used to assess the results. The simulation heat release rate is prescribed based on experimental measurements. The experimental duct flow rate is about 80 m3/h. Two methods are used to meet the aforementioned value in simulations: 1) by modeling the actual fan curve and dampers, and 2) by modifying the fan curve to avoid simulations of dampers. The combination of real fan curve and dampers provides better results. The simulations reproduce well the pressure profile (maximum 433 Pa) and the duct flows. The simulated reverse inlet flow and increased outlet flow rates reach values of 157 m3/h and 165 m3/h, respectively. Besides, the fire-induced pressure is high enough to hinder evacuation and fire rescue operations due to the impossibility of opening inward-open doors. Moreover, the adjacent room pressure also reaches a high level. Reducing the gap area between rooms significantly reduces the adjacent room pressure, but leads to an increase of the fire room pressure. Temperature deviations are observed and are improved by modeling a more realistic fire source. [ABSTRACT FROM AUTHOR]

The paper presents a numerical study on an 18 cm-diameter liquid pool fire in a reduced-scale and mechanically-ventilated enclosure. The enclosure walls are made of a calcium silicate layer and a steel layer on five sides, the remaining sidewall is made of heat-tempered glass. The ventilation system is extracted mechanically with free air inlet. Three air renewal rates (ARR) and open atmosphere conditions are considered in the numerical study. Computational Fluid Dynamics (CFD) simulations are carried out with Fire Dynamics Simulator (FDS 6.7.5). The default evaporation model in FDS (i.e., forced convection approach) is compared to a natural convection approach implemented in the source code. The results show that the oscillatory combustion, occurring at ARR = 15 h−1, is best captured by the natural convection approach. It is also found that prescribing the fuel Auto-Ignition Temperature (AIT) is important to capture the fire oscillations and (local) flame quenching. At ARR = 8 h−1, extinction occurs. It is found that prescribing the fuel AIT allows to prevent spurious oscillatory combustion. At ARR = 22.5 h−1, there is a stable steady-state burning regime for which (as well as for open atmosphere conditions) prescribing the fuel AIT is not crucial. • Fire burning regimes are well predicted in a wide range of ventilation conditions. • Oscillatory combustion is captured by natural convection for evaporation modeling. • Prescribing Auto-Ignition Temperature is important in capturing fire oscillations. • Prescribing AIT prevents spurious fuel re-ignition in under-ventilated conditions. [ABSTRACT FROM AUTHOR]

Abstract Small-scale experiments have been conducted to study the sealing effect of an air curtain for fire-induced smoke confinement in a tunnel configuration. The processed data confirmed the results obtained earlier from blind Computational Fluid Dynamics (CFD) simulations [1] using the Fire Dynamics Simulator (FDS) 6.5.3 [2,3]. Furthermore, the CFD simulations provided complementary information on the detailed flow and temperature fields which are difficult to obtain in experiments with the available techniques. A parametric study is performed, covering a range of air curtain velocities, slot widths, injection angles and total fire heat release rates (HRRs). The momentum ratio R, defined as the ratio of the vertically downward air curtain momentum to the horizontal smoke layer momentum at the position of the air curtain, is confirmed to be a key parameter for the air curtain performance. A ratio R ≈ 10 is recommended for the optimum sealing effect in terms of smoke confinement. In addition, two other important parameters that determine the performance of air curtains for smoke confinement are presented. The first parameter is the dimensionless shape factor AR (AR = Width/Length) that characterizes the dilution effect of the air curtain jet. The second parameter is the injection angle θ that characterizes the horizontal force of the air curtain. The air curtain sealing effectiveness increases with both the increase of slot width (shape factor AR) and injection angle (θ). The air curtain width has a limited influence on the performance of the air curtain whilst the injection angle has a more significant influence on the sealing effectiveness of the air curtain for the scenarios considered in this study. An optimal injection angle of 30° inclined to the fire source is recommended in the engineering design of the air curtain for smoke confinement for situations where the fire location can be pre-determined to be only at one side of an air curtain. [ABSTRACT FROM AUTHOR]

Abstract Computational Fluid Dynamics (CFD) simulation results, obtained with FDS 6.0.1 (McGrattan et al., 2013), are presented of reduced-scale tunnel fire tests. In (Sun et al., 2016), an extensive data set has been discussed in terms of temperature measurements in a reduced-scale tunnel, involving longitudinal ventilation and a variety of water supply through nozzles, in the context of potential smoke blockage. In (Sun et al., 2016), 10 different scenarios have been discussed, for different numbers of nozzles and different nozzle configurations. Given the limitation of the experimental instrumentation, as a series of thermocouple trees, a full interpretation of the flow field was impossible. Nevertheless, a detailed characterization and interpretation of this turbulent flow field under different circumstances is essential in the discussion of potential smoke blockage. To that purpose, CFD can be a very useful tool. In this paper, as a first step, results are presented for 2 cases, without mechanical longitudinal ventilation, in order to illustrate the validity and potential of the CFD simulations, with and without the water system (4 nozzles) activated. The validity of the CFD results, using the default FDS settings for turbulence and combustion modeling, is illustrated first through comparison of the temperature profiles with the experimental data. A comprehensive sensitivity study on the computational mesh and model settings for the water sprays is included. Subsequently, the mean flow and temperature fields are analyzed, providing significant additional insight into the impact of the water system. The entrainment, induced by the water sprays, is illustrated. This causes downward motion of the smoke in the sprays. By bumping onto each other, the impinging flows onto the floor in their turn create an upward flow in between the water spray envelopes that impinges onto the ceiling in the absence of longitudinal ventilation. The global effect is smoke blockage by the water system. [ABSTRACT FROM AUTHOR]

Corner fires are known to spread more intensely in comparison with single wall fires. In view of the challenges associated with prediction of such fire behavior, the fire growth in a corner configuration of Medium Density Fiberboard (MDF) panels is investigated to provide a set of experimental data, performing Single Burning Item (SBI) tests. First, though, test results with inert calcium silicate panels are discussed for three values of HRR (10, 30 and 55 kW), allowing to address the main physics involved. The experimental data for 30 kW, the default SBI HRR, is used for detailed discussion of the observations. The SBI testing methodology, materials, and set-up are described. The results of total Heat Release Rates (HRR) and Smoke Production Rates (SPR), as well as the panel temperatures and total heat fluxes at several characteristic locations are analyzed. Moreover, the puffing frequency of the corner fire is characterized thanks to Video Fire Analysis (VFA) of the experimental footage. Additionally, flame heights are discussed, including the concept of mirroring. A new correlation for mean flame height is introduced, using the hypotenuse of the triangle as characteristic length for entrainment of air into the fire plume, and expressing that the flame height increases proportional to the square root of the fire heat release rate. The 30 kW propane burner of the standard SBI test is shown to feature a mean flame height of nearly 0.9 m and a puffing frequency of 2  ±  0.3 Hz, and an average total heat flux exceeding 44 kW/m 2 near the burner early on in the test. The completeness of the dataset is expected to be useful for testing and development of CFD codes for corner fire scenarios. [ABSTRACT FROM AUTHOR]

Having explained the characteristics of a corner fire in the configuration of Single Burning Item (SBI) test in Part I [1], the results of three flame spread experiments conducted with Medium Density Fiberboard (MDF) panels are discussed. The fire growth, in terms of flame heights and spread, is examined from two different angles visually and through Video Fire Analysis (VFA) with a flame detection algorithm. Total Heat Release Rates (HRR) and Smoke Production Rates (SPR), as well as total heat fluxes at several characteristic locations, are presented. Moreover, temperature evolutions are discussed for multiple locations and through the thickness of the panels. Also the backside temperatures, important as boundary condition for numerical simulations, are reported. The corner fire tests with MDF panels yield an average peak HRR of 151 kW and an average total heat flux exceeding 60 kW/m 2 close to the burner, with average flame heights surpassing 1.5 m in about 60 s. [ABSTRACT FROM AUTHOR]

Abstract: In this paper a novel multi-modal flame and smoke detector is proposed for the detection of fire in large open spaces such as car parks. The flame detector is based on the visual and amplitude image of a time-of-flight camera. Using this multi-modal information, flames can be detected very accurately by visual flame feature analysis and amplitude disorder detection. In order to detect the low-cost flame related features, moving objects in visual images are analyzed over time. If an object possesses high probability for each of the flame characteristics, it is labeled as candidate flame region. Simultaneously, the amplitude disorder is also investigated. Also labeled as candidate flame regions are regions with high accumulative amplitude differences and high values in all detail images of the amplitude image''s discrete wavelet transform. Finally, when there is overlap of at least one of the visual and amplitude candidate flame regions, fire alarm is raised. The smoke detector, on the other hand, focuses on global changes in the depth images of the time-of-flight camera, which do not have significant impact on the amplitude images. It was found that this behavior is unique for smoke. Experiments show that the proposed detectors improve the accuracy of fire detection in car parks. The flame detector has an average flame detection rate of 93%, with hardly any false positive detection, and the smoke detection rate of the TOF based smoke detector is 88%. [Copyright &y& Elsevier]

The paper presents experimental results indicating the influence of the fire heat release rate (HRR) evolution on the fire-induced pressure variation in air-tight compartments, which increasingly appear in modern buildings. The fire evolution is stipulated according to a power law with variations in fire growth/decay rate coefficient, maximum steady state HRR, fire growth/decay exponent and fire duration. It is illustrated that the higher the fire growth/decay rate coefficient, the higher the over-pressure/under-pressure peaks become. The peak values approach a maximum level when the fire growth/decay rate coefficient exceeds a certain level (0.0926 kW/s2 in the present cases). Besides, the over-pressure and under-pressure peak values increase monotonously with the increase of the maximum steady state HRR. For cases with fire growth/decay exponent below 1, the pressure reaches peak values before the HRR reaches a steady state (maximum value or zero) because the net heat gained per unit time in the gas phase already reaches its peak values then. In cases with fire growth/decay exponent above 1, the pressure reaches peak values at the moment where the HRR reaches the steady state. Finally, in one test low-frequency oscillatory behavior is illustrated with gaseous fuel during the well-ventilated phase. • Comprehensive studies are conducted by parameterizing the heat release rate. • Increase in fire growth/decay rate causes the pressure to rise to a plateau level. • Higher maximum heat release rate leads to higher over-pressure and under-pressure. • The fire growth/decay exponent affects pressure variation and under-pressure peak. • Oscillatory behavior is reported with gaseous fuel in the well-ventilated phase. [ABSTRACT FROM AUTHOR]

The paper presents a series of 48 enclosure fire experiments carried out in a 1. 40 m cubic compartment with a door-like opening of 0. 80 m width and a variable height from 0.25 to 1. 40 m. The fire source was a propane-fueled square burner of 0. 17 m side length positioned in the center of the compartment and delivering a steady mass flux that has been varied in the tests from 0.5 to 5. 0 g/s. In addition to the well-known well-ventilated and under-ventilated regimes, an oscillatory combustion regime has been uncovered for an opening height of 0.25 or 0. 35 m. This regime is characterized by an alternation between the well-ventilated regime (combustion exclusively inside the compartment) and the under-ventilated regime (combustion inside and outside the enclosure). It is believed that the time for the onset of the oscillations is linked to the smoke filling time. Furthermore, depending on the specific conditions, the oscillations are either steady (until shut-off of the fuel supply) or transitory (leading to a sustained under-ventilated regime). The frequencies ranged from about 9 to 19 mHz. The oscillatory regime has previously been mainly reported for liquid fuels and thus associated to radiative heat feedback effects. The new data for a gaseous fuel calls upon a better understanding and a further analysis of the underlying physics. • Naturally-ventilated compartment fire experiments using propane were carried out. • An oscillatory combustion regime has been uncovered. • There are several oscillatory modes: sustained and steady or transitory. [ABSTRACT FROM AUTHOR]

A novel two-zone model is implemented, within an in-house code, for the heating of an evaporating droplet. While it is not yet possible to fully assess and validate the novel model due to lack of available detailed experimental data, the concept is described. Differences compared to the 'common' isothermal model are explained. Firstly, a quantitative assessment is performed, using experimental datasets where a suspended water droplet is exposed to a hot air flow. Secondly, the behavior of the novel model is examined over a wider range of conditions in terms of air velocity, ambient temperature and initial droplet diameter. The novel model predicts droplet lifetimes and saturation temperatures that are similar to the isothermal model. Nevertheless, the results also show that, during the early stages, the cooling of the surrounding gas due to the evaporation of the droplet is more pronounced with the two-zone model. During a second stage, the evaporation term is slightly lower than with the isothermal model. The available data do not allow to fully assess which result is closer to reality, but the important observation is that there exist differences. These differences increase with higher ambient temperature, relative velocity or initial diameter. • A novel two-zone model for the heating of an evaporative droplet is implemented. • The two-zone model is compared to the isothermal model. • Similar predictions for the lifetimes and saturation temperatures are found. • Evaporative cooling is more pronounced at the early stages with the two-zone model. [ABSTRACT FROM AUTHOR]

• Little knowledge is openly available on marine engine fuel sprays. • A transient injection pressure profile improve spray penetration predictions. • Spray tip penetration is a three-stage function of time after injection. • Spray-induced turbulence is affected by ambient gas density and engine load. The understanding of diesel sprays is very important to enable a better and cleaner marine engine design, but unfortunately little knowledge is openly available on marine engine fuel sprays. In this paper, evaporating sprays for medium speed marine engines were studied in a constant volume combustion chamber by performing optical measurements through Schlieren and Mie diagnostic techniques. The effects of ambient gas temperature and ambient gas density on vapor and liquid penetration were investigated by changing the target condition in the combustion chamber. A comparative study of two injectors with different nozzle diameters (0.38 mm and 0.44 mm) was also carried out at ambient density of 22.5 kg / m 3 . Some empirical correlations of spray penetration have been modified to fit the spray measurement data. Due to the transient characteristics of the pump-line-nozzle injection system, a time-dependent injection pressure profile is suggested for calculation of spray penetration. The spray tip penetration at large distance under low density (7.6 and 15.2 kg / m 3 ) conditions is expected to be proportional to t 2 / 3 , which is supported by the model considering spray-induced gas turbulence effect. The t 1 / 2 law, where turbulence is not taken into account, is still valid under high density (22.5 kg / m 3 ) conditions with higher engine load. The comparison of two models demonstrates that the effect of gas turbulence is influenced by the ambient gas density and engine load. [ABSTRACT FROM AUTHOR]

Air cavities and gaps between material layers are common in construction systems, e.g. ventilated façades. Air cavity may provide a pathway for smoke and flame spread in case of fire. Performing physical testing to investigate different systems and fire scenarios is resource demanding. Fire Dynamics Simulator (FDS version 6.7.0) was used to simulate fire driven flow between two parallel vertical walls. Flame heights, thermal impact to the interior wall surface and upward flow velocities were predicted with FDS and compared with experimental results. The fire source was a propane burner with 8 × 391 mm2 gas outlet area. Heat release rates were 6.6 kW and 12.4 kW and the distance between the parallel walls was 40 mm. Two different convective heat transfer coefficient sub grid scale models available in FDS were investigated. In this study the cavity width to mesh cell size ratio was equal or above 10, resulting in good predictions of flame heights, upward flow velocities and wall temperatures. 2 mm grid resulted in 25% lower HRR in locations near the burner gas inlet, compared to 4 mm grid, indicating the importance of well resolved gas outlet boundary. • Cavity width divided by mesh cell size should be greater than 10 in FDS simulations. • Predicted flame heights and flow velocities corresponded well with experiments. • Refining the grid in FDS resulted in lower cumulative HRR in the flame region. • FDS models for heat transfer at surfaces must be revised for transitional flows. [ABSTRACT FROM AUTHOR]

Abstract The distribution of mass density through the thickness of Medium Density Fiberboard (MDF) panels is known to be non-uniform. A few studies have previously investigated the influence of this non-uniform through-thickness density distribution on the thermal behavior of MDF panels in small-scale tests. This study assesses the significance of this material property on flame spread simulations in a medium-scale set-up, namely that of Single Burning Item (SBI) corner fire tests. Simulations are performed using FireFOAM 2.2.x, considering both uniform and non-uniform MDF material density profiles, using model-effective material properties determined from bench-scale pyrolysis tests conducted in a Fire Propagation Apparatus (FPA). The heat transfer from the gas phase is modeled by means of empirical expressions with adjusted parameters. The simulations are assessed against the results of several SBI experiments with MDF panels and a test with Calcium Silicate (CS) panels. When the non-uniform nature of the through-thickness density is taken into account, the fire growth prediction in terms of the total Heat Release Rate (HRR) is considerably different (20% higher peak HRR), mainly due to the characteristic high peak mass loss rate at the initiation of pyrolysis of MDF material, resulting from the higher mass density near the surface of the panels. Furthermore, total heat fluxes on the panels, lateral flame spread, surface pyrolysis and through-thickness char formation visibly depend on the non-uniform distribution of mass density, particularly in regions further away from the corner where the influence of thermal attack from the burner is less dominant. A new diagnostic is proposed for determining the pyrolysis front location and spread on the surface of the charring panels. [ABSTRACT FROM AUTHOR]

A water system, consisting of several water mist nozzles, has been installed in a reduced-scale tunnel. Its effectiveness in blocking fire-induced smoke and heat is tested, with and without longitudinal ventilation. A total of 14 fire tests have been carried out, with 250 ml methanol in an iron tray (25 cm × 20 cm) as fuel. Temperatures have been measured by 30 thermocouples, located upstream and downstream of the fire location. The aim is to assess the effectiveness of the water system in preventing smoke spread and in reducing the temperature in the tunnel. Interaction of the water with the fire is avoided. The impact of water pressure, ventilation velocity and nozzle arrangement on the effectiveness in smoke blocking and temperature reduction is discussed. The result confirms that the water system effectively reduces the temperatures and prevents smoke spreading in the absence of longitudinal ventilation. However, strong longitudinal ventilation (0.8 m/s ventilation velocity in the reduced-scale tunnel, corresponding to critical velocity in full-scale (1:10) tunnel) reduces the effectiveness in blocking the smoke spreading by the water system, although the temperature reduction downstream the water system remains in place. Higher water pressure makes the cooling effect stronger, because more and smaller water droplets are injected into the tunnel. For a given level of water pressure level, the impact of the nozzle row configuration is small in the tests. [ABSTRACT FROM AUTHOR]

Numerical simulations are performed to investigate the classical compartment fire problem involving a single door opening. Three different configurations were considered, namely a full-scale (ISO 9705), an intermediate and a small-scale enclosure with various opening heights and widths. A large number of numerical simulations was carried out using the CFD code Fire Dynamics Simulator (FDS)(version 6.7.0). Based on the variation of the average temperature inside the compartment, three combustion regimes were identified namely well-ventilated, transitional and under-ventilated regime. This variation also allowed us to identify the boundaries between regimes. Furthermore, by adopting a non-dimensional representation of the fire heat release rate inside the compartment as a function of the Global Equivalence Ratio (GER), a clear demarcation between these combustion regimes was obtained. A linear correlation has been established between the maximum heat release rate inside the compartment and the ventilation factor. The latter is expressed as Q ̇ i n max = 850 A H. A linear relation between the maximum air flow rate m ̇ i n and the ventilation factor A H was found, i.e. m ̇ i n max = 0.46 A H where A is the door surface area and H its height. [ABSTRACT FROM AUTHOR]

Abstract: In this study, a global soot formation model based on the LSP (Laminar Smoke Point) concept in combination with the soot oxidation model developed by Leung et al. is validated in three laminar flames: a non-smoking and a smoking ethylene flame, as well as a non-smoking propane flame, and then applied in two turbulent pool fires fueled by ethylene and methane, respectively. In this global soot model, the sooting propensities of different fuels are accounted for by a pre-exponential factor determined from the LSP height, providing a general and practical solution for soot modeling in multi-fueled fires. The flame fields are solved by FLUENT with UDFs to provide material properties and add additional soot governing equations. A-CSE (Alternative Conditional Source-term Estimation) approach is adopted to handle the interaction between soot chemistry and turbulent flow in the turbulent fires. The model parameters such as the pre-exponential factor, soot inception limits and soot particulate surface area are determined and calibrated against the experimental data. Mixture fraction and temperature are first verified to provide a good premise for soot modeling. Good agreements between the predicted and measured soot volume fraction, as well as the reproduction of transition from non-smoking to smoking flames, demonstrate the capability of current global soot model in accurately predicting soot for both laminar flames and turbulent fires. Using the A-CSE soot modeling approach together with the global soot model, this study presents a general effective yet computationally efficient global soot modeling framework for fires. [Copyright &y& Elsevier]

The paper presents a comprehensive set of experiments on the effect of mechanical ventilation conditions and fire dynamics on temporal pressure evolution in a reduced-scale, air-tight and mechanically-ventilated enclosure. A square propane burner with flow controller imposes a quadratic fire growth followed by a steady-state (0.1 or 0.2 g/s) and then a quadratic decay phase. Eight tests are discussed with different ventilation conditions in terms of flow resistances and initial ventilation flow rates ranging from 12 to 40 m3/h, corresponding to air renewal rates of 6.4–21.3 h−1. The pressure evolution is characterized by an over-pressure peak (up to 900 Pa) followed by a quasi-steady state and then, an under-pressure peak (up to −760 Pa). The pressure variation is due to the mechanical effect (i.e., ventilation configurations), while also influenced by thermal effects. The pressure amplitudes increase with ventilation resistances. Both the total network resistance and individual resistances in admission and extraction ducts are important for the pressure variation. The enhancement and reduction of ventilation flow rates depend on both the fire-induced pressure and ventilation resistances. Experimental results show that the mechanical effect does not strongly affect gas temperatures. [ABSTRACT FROM AUTHOR]

In order to better interpret the fire-induced pressure variation in air-tight compartment fires, the present paper discusses the energy balance equation starting from a commonly used form, taking into account the mutual influence between compartment pressure and ventilation behavior. The obtained pressure formulation, under certain assumptions, can explain the observed linear relationship between pressure variation and fire evolution. Based on the mathematical development, the difference between fire-induced pressure and initial compartment pressure is expected to depend on the net heat gained in the gas phase. This is subsequently validated by experimental results. Differences in heat losses through walls, acquired by varying the fire duration, are illustrated to influence the pressure variation during the fire decay phase, while they hardly affect the pressure during the fire growth phase. Besides, it is found from the theoretical reasoning that the mass flow rate difference between admission and extraction ducts depends on the net heat gained in the gas phase, and is not strongly related to the ventilation resistance. This is also confirmed by the experiments. However, a higher ventilation resistance results in higher compartment pressure variations and longer development time of the pressure. [ABSTRACT FROM AUTHOR]

The individual and combined effect of a water mist system (WMS) and longitudinal ventilation (LV) on the fire HRR, smoke temperature and back-layering (BL) degree were explored in a 3 m (width) × 2.2 m(height) × 30 m (length) tunnel model. A series of 39 tests has been conducted in 6 configurations, varying the velocity, water volume flow rate and nozzles arrangement. When the WMS or LV is imposed individually, the reduction effect on the HRR gets stronger with larger water volume flow rates or higher velocities. Yet, under the combined effect, the effect of the LV velocity on the HRR is non-monotonic: the HRR first reduces more, and then rises again with higher LV velocities for given WMS settings. The reduction in temperature monotonically increases with higher LV velocities, in contrast to the HRR behavior. The above phenomenon is explained to be due to the relatively 'low' flame height in the present paper, as a consequence of which the interaction between the flame and the water droplets mainly occurs in the lower part of the tunnel, while the interaction between the hot smoke and the water droplets mainly happens in the upper part. The fire HRR behavior mainly relates to the interaction in the lower part, while the temperature depends on both regions. Besides, for a certain BL distance, the required ventilation velocities reduce by 13%–55% with activated WMS, compared to the same conditions without WMS. The effect of covering the fire with a shield is also discussed. For relatively small velocities, the fire HRR, smoke temperature and BL distance are larger with the shield in place. [ABSTRACT FROM AUTHOR]

The corner fire characteristics with two inert walls are compared against those with one or two combustible corner walls through a set of standard Single Burning Item (SBI) tests. The study examines how the corner fire behavior is affected by flame spread over one or two walls. In the case of flame spread over one corner wall, the inert side is a Calcium Silicate (CS) panel. As for the combustible wall, two different materials have been tested, namely, plywood and Medium-Density Fiberboard (MDF) panels. The results indicate that the simultaneous burning of two panels in the corner geometry increases the Heat Release Rate (HRR) contribution of each panel by 45–50%. This is because the burning of one corner wall increases the temperatures on the adjacent wall significantly, occasionally by more than 200 °C, and increases the heat fluxes on both corner walls at the same time, sometimes by nearly 20%. The results of HRR, Smoke Production Rate (SPR), flame heights, pyrolysis front propagation, panel temperatures, and total heat fluxes at several characteristic locations are presented. [ABSTRACT FROM AUTHOR]

Corrigendum to "Development and evaluation of two new droplet evaporation schemes for fire dynamics simulations" [Fire Saf. [Extracted from the article]

This paper presents a detailed experimental and numerical study on the interaction, at a reduced scale, between a turbulent buoyant propane flame of about 15.5 kW and a water mist spray with a flow rate of 0.43 L/min from a nozzle positioned at 1 m above the burner. The water spray has been characterized experimentally without a fire, by measuring the water mass flux and droplet size distributions at the level of the burner surface. The whole assembly was installed under a hood and the following three parameters were measured: (1) the chemical heat release rate (HRR) (using the oxygen consumption method), (2) the rise in gas temperature at the top, and (3) the radiative heat flux at 0.72 m from the axis of the burner and at a height of 0.05 m. Reductions of about 40% and 30% were recorded for, respectively, the gas temperature (in the hood) and the radiative heat flux, while the chemical HRR did not change. Computational Fluid Dynamics (CFD) simulations with the Fire Dynamics Simulator (FDS, version 6.7.0) predicted relatively well the gas temperature, without a reduction in the HRR, but, in contrast to the experiments, the radiative heat flux did not change. • A gas analysis system is used to accurately characterize the heat balance. • The cooling power of the spray is estimated using a single temperature measurement. • The water spray is well characterized experimentally and numerically. • The cooling power of the spray is underpredicted by about 34%. • The reduction in the radiative heat flux to the surroundings has not been predicted. [ABSTRACT FROM AUTHOR]

• The impact of the longitudinal ventilation and water spray on smoke was studied. • The impact of longtidinal ventilation flow on smoke blocking effect is presented. • A wavy pattern is obtained in the smoke layer thoughtout the tunnel. • The water spray system affects the flow and temperature fields up to the fire. Computational Fluid Dynamics (CFD) simulation results, obtained with FDS 6.0.1, are presented of reduced-scale tunnel fire tests with longitudinal mechanical ventilation and an activated water spray system. This is a sub-set of a more extensive data set, in the context of blocking fire-induced smoke tunnel tests conducted previously. As discussed in Sun et al. (2018), CFD can complement the experimental data, providing thermocouple tree temperature measurements, yielding insight into the turbulent flow field generated by the fire and the water spray. The CFD settings, decided upon on the basis of an extensive mesh sensitivity study in Sun et al. (2018), are applied here to illustrate the impact of longitudinal mechanical longitudinal ventilation in cases with and without the water spray system activated. Analysis of the mean flow and temperature fields provides significant additional insight into the impact of the combined longitudinal ventilation and the water spray system on the fire-induced smoke flow. The upward flow in between the water sprays in the absence of longitudinal ventilation is still observed, but it does not impinge onto the ceiling anymore: the horizontal momentum from the mechanical ventilation makes the 'pushing' force of the spray-induced downward flows impinging onto the floor less strong and moves them further downstream. Consequently, the smoke crosses the water spray system region. The smoke stratification is lost downstream of the spray region. Upstream of the water sprays, and throughout the tunnel in the absence of the water spray system, a wavy pattern is obtained in the smoke layer. This is related to the Froude or Richardson number for the flows with the settings at hand. The presence of the water spray system downstream of the fire affects the flow and temperature fields up to the fire source. [ABSTRACT FROM AUTHOR]

A 1/3 scaled tunnel model was built to explore the effect of fire source elevation on fire and smoke behavior, and especially the critical velocity, in a tunnel with longitudinal ventilation. An experimental campaign has been carried out with a series of tests in 9 configurations, varying the fire size, ventilation velocity and fire elevation. Results in terms of combustion rate, ceiling temperature and critical velocity are discussed. If the flame reaches the ceiling when the fire source is elevated, the fuel burning rate increases by 15% due to increased heat feedback created by the part of the flame below the ceiling. The location of the maximum ceiling temperature can move towards the downstream region in such conditions. A variable named as the height of the effective tunnel, H ' , was proposed to substitute the tunnel height, H , in the Li et al. model [1] to consider the effect of fire elevation in predicting critical velocity. The accuracy was illustrated by means of the results of the present tests and those of other research studies. Based on a dimensional analysis, three conditions are identified on the basis of the ratio of the flame height to the height difference between fire surface and the tunnel ceiling, L f l H ' . This allows to demonstrate the variation of the critical velocity with changing fire elevations. The accuracy was illustrated by means of the results of the present tests and those of other research studies. The increase of fire elevation leads to the increase of the dimensionless variables Q ∗ (W H ' ) − 1 4 , L f l H ' and V ∗ (although this remains sometimes unchanged), rather than only a rise or diminution of critical velocity. [ABSTRACT FROM AUTHOR]