Search

The dynamic behaviour of wildfires is mainly influenced by weather, fuel, and topography. Based on fundamental conservation laws involving numerous physical processes and large scales, atmospheric models require substantial computational resources. Therefore, coupling wildfire and atmospheric models is impractical for high resolutions. Instead, a static atmospheric wind field is typically input into the wildfire model, either as boundary conditions on the control surface or distributed over the control volume. Wildfire propagation models may be (i) data-driven ; (i i) theoretical ; or (i i i) mechanistic surrogates. Data-driven models are beyond the scope of this paper. Theoretical models are based on conservation laws (species, energy, mass, momentum) and are, therefore, computationally intensive; e.g. the Fire Dynamics Simulator (FDS). Mechanistic surrogate models do not closely follow fire dynamics laws, but related laws observed to make predictions more efficiently with sufficient accuracy; e.g. FARSITE, and FDS with the Level Set model (FDS-LS). Whether theoretical or mechanistic surrogate, these wildfire models may be coupled with or decoupled from wind models (e.g. Navier-Stokes equations). Only coupled models account for the effect of the fire on the wind field. In this paper, a series of simulations of wildfire propagation on grassland are performed using FDS-LS to study the impact of the fire-induced wind on the fire propagation dynamics. Results show that coupling leads to higher Rates of Spread (RoS), closer to those reported from field experiments, with increasing wind speeds and higher terrain slopes strengthening this trend. Aiming to capture the fire–wind interaction without the hefty cost of solving Navier-Stokes equations, a conceptual framework is proposed: 1) a swarm of unmanned aerial vehicles measure wind velocities at flight height; 2) the wind field is constructed with the acquired data; 3) the high-altitude wind field is mapped to near-surface, and 4) the near-surface wind field is fed into the wildfire model periodically. A series of simulations are performed using an in-house decoupled physics-based reduced-order fire propagation model (FireProM-F) enhanced by wind field "measurements". In this proof of concept, wind velocities are not measured but extracted from physics-based Large Eddy Simulations taken as ground truth. Unsurprisingly, higher measurement frequencies lead to more accurate and realistic predictions of the propagating fire front. An initial attempt is made to study the effect of wind measurement uncertainty on the model predictions by adding Gaussian noise. Preliminary results show that higher noise leads to the fire front displaying more irregular shapes and slower propagation. • Higher wildfire model prediction accuracy normally achieved by coupling it with wind model such as Navier-Stokes equations. • Coupled fire-wind models often too computationally expensive for real-time use in operational contexts. • Prediction accuracy of wildfire model uncoupled from wind model enhanced by periodic inputs of updated wind field. • Wind field to be constructed from UAV swarm measurements, surrogated here to velocities obtained from full CFD simulations. • Proposed method to enhance wildfire model predictions robust to Gaussian noise added to represent measurement uncertainty. [ABSTRACT FROM AUTHOR]

Chemical warehouses are one of the important sources in process units. Locations are one of the issues that should be investigated in closed environments due to the many types of chemicals and storage conditions. In this study, fire dynamics simulation (FDS), was used to simulation of fire in the presence and non-presence of the ventilation system. Then the spread of fire in the warehouse was evaluated based on domino methods and finally the Bayesian network (BN) was used to update the rate of fire spread. The simulation results in this approach showed that the highest amount of heat flux related to iso-octane fire at L/D= 0.3 when the ventilation system is off and the doors are closed. The updated results of the fire spread modeling in BN showed that if a fire occurs in isooctane barrels, it has the highest rate of spread with a probability of 22% and the lowest rate of spread in Diglycolamine(DGA) barrels with a probability of 0.4%. From the results of this study, it is possible to understand the importance of arranging materials according to the principles of compatibility of chemical materials, and based on this, appropriate control measures can be provided. [Display omitted] [ABSTRACT FROM AUTHOR]

Recent experiments have shown that exposed cross-laminated timber (CLT) can have a significant effect on the fire dynamics of large compartments. A simulation with the Fire Dynamic Simulator has been conducted to better understand the fire behaviour of open-plan compartments with exposed CLT. The simulation was set up to replicate a large-scale experiment, FRIC-02, with exposed CLT on the back wall and ceiling. The compartment was 95 m2 (18.8 m × 5.0 m × 2.5 m), with one long wall open (opening factor 0.18 m1/2). A continuous wood crib was used as the variable fuel load. The characteristic results of FRIC-02 with a rapid fire development and non-symmetrical external flames were successfully reproduced. With the wind coming diagonally from behind, as in FRIC-02, the external flames emerged mainly out of one window. The flames covered the entire window height, which effectively inhibited the inflow of air through that window. The imbalance in air supply also created large temperature differences throughout the compartment. With no implementation of wind, external flames and temperatures were more symmetrical. Despite a good match to FRIC-02, the method still has several limitations, including the adaption of the burning rate to the feedback from surroundings. • An FDS model for fire spread on a compartment with exposed CLT is developed. • Global matching of fast fire spread with experiment. • Non-symmetrical flames caused by external wind are reproduced. • CLT elements on the back wall cause larger external flames. [ABSTRACT FROM AUTHOR]

External walls are part of the construction system where new solutions appear every day. These systems are usually built with new materials that offer greater thermal resistance but with poor fire behaviour. This fire behaviour can be evaluated in large-scale tests, such as the ones prescribed at BS 8414-1, LEPIR II or ISO-13785-1 standards, that consider a typical scenario where the fire outbreaks inside the building and spreads to outside through a window or venting. However, there are situations in which these systems can be affected by external fires, such as forest fires, goods burning next to the external wall, or even sabotage, that are of interest to be evaluated. This paper presents the results of experimental tests for evaluating the behaviour of building external walls made of sandwich panels, of four different insulation materials in the core (mineral wool, polyisocyanurate - PIR, organic polymer and polyurethane - PUR). The tests were carried out according to BS 8414-2 standard, considering the common internal and a new external fire scenario. The external fire scenario was intended to simulate the case of goods or equipment stored next to the wall, outside the building, that catches fire. The walls' performance was checked according to BRE 135 Annex B criteria. In these tests were registered the temperatures in the exposed and unexposed face of the walls and the incident radiant heat flux in the wall surface. The external fire scenario showed that can be more aggressive to the external wall cladding than the internal fire, demonstrating that this condition must be carefully investigated to propose this situation in standards for the future. [ABSTRACT FROM AUTHOR]

Data-driven techniques are increasingly being applied to complement physics-based models in fire science. However, the lack of sufficiently large datasets continues to hinder the application of certain machine learning techniques. In this paper, we use simulated data to investigate the ability of neural networks to parameterize dynamics in fire science. In particular, we investigate neural networks that map five key parameters in fire spread to the first arrival time, and the corresponding inverse problem. By using simulated data, we are able to characterize the error, the required dataset size, and the convergence properties of these neural networks. For the inverse problem, we quantify the network's sensitivity in estimating each of the key parameters. The findings demonstrate the potential of machine learning in fire science, highlight the challenges associated with limited dataset sizes, and quantify the sensitivity of neural networks to estimate key parameters governing fire spread dynamics. [Display omitted] • Neural networks parameterize first arrival times and the associated inverse problem. • Various neural network architectures are considered. • The error and sensitivity of the networks are analyzed and compared. [ABSTRACT FROM AUTHOR]

One crucial mechanism in the spread of wildfires is the so-called fire-spotting: a random phenomenon that occurs when embers are transported over large distances. Fire-spotting speeds up the rate of spread and starts new ignitions that can jeopardise firefighting operations. Unfortunately, operational fire-spread simulators may not account for spotting events, thus overlooking the harmful consequences associated with this phenomenon. In this work, three fire spotting parametrisations are integrated in the operational wildfire simulator PROPAGATOR based on Cellular Automata (CA). RandomFront, a physics-based parametrisation of fire-spotting, is tested for the first time in the context of CA simulators. RandomFront is compared with other two parametrisations already adopted in CA based simulators, those by Alexandridis and co-authors and by Perryman and collaborators. A wildfire occurred in the summer of 2021 in the municipality of Campomarino (Molise, Italy), and where spotting effects were clearly reported, is used as a case study. This case study, featuring evident airborne transport of firebrands, paves the way for a framework for comparing parameterised spotting models used in operational scenarios. RandomFront produced a more complex burning probability pattern than the other parametrisations and it predicted a higher probability of burning in the zone mainly affected by the fire-spotting. • Comparison study among fire-spotting parametrisations. • Implementation of RandomFront fire-spotting routine in PROPAGATOR. • A real wildfire event is reproduced in operational time. • Results are analysed through probabilistic scores. [ABSTRACT FROM AUTHOR]

This paper presents the mathematical formulation, numerical solution, calibration and testing of a physics-based model of wildfire propagation aimed at faster-than-real-time simulations. Despite a number of simplifying assumptions, the model is comprehensive enough to capture the major phenomena that govern the behaviour of a real fire — namely the pyrolysation of wood; the combustion of a mono-phase medium composed of premixed gas of fuel and air; and the heat transferred by conduction, convection, radiation, mass diffusion and transport due to atmospheric wind. The model consists of a system of coupled partial differential equations, one ensuring the balance of enthalpy, and a set of equations representing the mass formation of each chemical species involved in the combustion. Dimensionality reduction is sought by modelling these three-dimensional phenomena in a two-dimensional space, which has been achieved by means of heat-sources and heat-sinks to account for the third dimension in the energy balance equation. Once calibrated with a widely used non-physics-based commercial wildfire simulator, the proposed Fire Pro pagation M odel for F ast simulations (FireProM-F) is tested, returning similar predictions in terms of the size and shape of the burnt area although similarity deteriorates for windy conditions. FireProM-F has the added benefit of being both physics-based and computationally inexpensive so that its interaction with fire suppressants may also be modelled in the future and simulated in real time. [ABSTRACT FROM AUTHOR]

The 3D structure of a fire front propagating through a homogeneous porous solid-fuel layer was studied numerically at laboratory and field scales. At laboratory scale, wind-tunnel fires propagating through laser-cut cardboard fuel were numerically reproduced, while at field scale, simulations of grassland fires with quasi-infinite fire front were carried out for different wind speeds. These simulations were performed using FIRESTAR3D, based on a multiphase formulation that includes the main physical phenomena governing fire behavior. An unsteady RANS approach and a Large Eddy Simulation (LES) approach were used to simulate the reactive turbulent flow, whereas turbulent combustion was modeled using Eddy Dissipation Concept (EDC). Unlike other 3D wildfire tools available in the community, such as FIRETEC and WFDS, the model is based on an implicit, low-Mach number resolution of the governing equations, and makes no empirical assumptions in the resolution of the radiative transfer equation. The comparison with the experimental data concerned mainly the Rate of Spread (ROS) of fire, the fireline intensity, the flame-zone depth, and the wavelength characterizing the crest-and-trough structure of the fire front along the transverse direction. Particular attention was drawn to the similitude in the fire front dynamics between small and large scales. In order to highlight the physical mechanisms responsible for this dynamics, a dimensional analysis was carried out by introducing Byram's convection number N C based on the fireline intensity and Froude's numbers Fr based on the characteristic wavelength of the fire-front structure. The analysis shows that all the results (wind-tunnel fires and grassland fires, experimental and numerical) collapsed on a single scaling law in the form F r = N C − 2 / 3. [ABSTRACT FROM AUTHOR]

A fact often overlooked is that large-scale wildfires, although occurring infrequently, are responsible for the overwhelming majority of fire-related suppression costs, economic losses, and natural resources damages. Fortunately, the increasingly severe problems of large-scale wildfires worldwide have been receiving ever-growing academic attention. The high-intensity burning behaviors in wildfires stem from the significant interaction of combustion with heat transfer and atmospheric flow under complicated fuel, meteorology, and topography conditions. Therefore, mitigating measures against large-scale wildfire disasters have grown into a challenging research focus for combustion scientists. Research over the past century has resulted in incrementally enhanced insights into the mechanisms of combustion dynamics underlying the various erratic behaviors in large-scale wildfires, with theories and models of fire accelerations developed and validated. These advances are expected to improve the efficacy of large-scale wildfire predictions significantly. Nevertheless, the physical interpretation of the acceleration of large-scale wildfires is far from adequate and complete. This paper intends not to make a comprehensive review of the entire wildfire research field, but to depict an overall pattern of the essential factors that lead an initial small-scale spreading flame to a large-scale wildfire beyond control. It is outlined that the complicated transformation of fuel preheating mechanisms determines the growth of surface fire spread, while varied large-size flame fronts and unique spread modes induced in specific fire environments play an essential role in fire spread acceleration. Additionally, multiple fires burning and merging often act as crucial steps for accelerating surface fire spread, generating large-size flames, and triggering unique spread modes. These major potential factors strike the energy balance of a low-intensity wildfire and push it to a high-intensity state. Several issues regarding intensely burning behaviors in large-scale wildfires are selected for in-depth discussions, for which an overview of the progress and challenges in research is presented. It is concluded that the fundamental exploration targeted at developing application tools capable of dealing with large-scale wildfires remains at its early stages. Opportunities for innovation are abundant, yet systematic and long-term research programs are required. [ABSTRACT FROM AUTHOR]

The external heating test is widely used to evaluate the hazards of battery thermal runaway, but the efficiency and effect of the heating source are rarely quantified. This work performs thermal runaway propagation tests in a 3-layer cylindrical battery pile with a uniform state of charge (SOC) ranging from 30 % to 75 %. A cylindrical heater is in contact with two cells in the first layer and has a power varying from 50 W to 300 W to trigger thermal runaway. Results indicate that for the current system, the heating efficiency to a single cell is around 15 %, and the effective heating power is insensitive to the SOC. The intensity of thermal runaway increases with the external heating power and the cell SOC. The influence of external heating on the propagation of thermal runaway is reflected in the intensity of thermal runaway in the first-layer cells and the preheating effect of subsequent layers. A simplified heat-transfer model is established to quantify the thermal impact on both thermal runaway intensity and preheating depth. Finally, a new approach for selecting the appropriate heating power is proposed to help optimize battery thermal-runaway tests and improve the safety regulations for battery modules. [ABSTRACT FROM AUTHOR]

The aim of this work is to improve the understanding of the fire behaviour of building materials used at the wildland urban interface using laboratory scale experiments. To this end, wood cladding was exposed to a propagating flame front across a decking slab. This slab element was ignited by a radiant panel. Three slab configurations were studied: thermoplastic slabs and wooden slabs with slats oriented parallel or perpendicular to the cladding. The vertical distance between the slabs and the cladding varied from 0 to 20 cm and two types of batten arrangement were used to separate the cladding from the supporting cellular concrete. The experiments first focused on the measurement of the heat fluxes received by a non-combustible wall in place of the cladding, depending on the type of slab used. Radiation was the dominant heat flux measured on the wall and it increased up to 35 kW m−2 as the flame front approached to the cladding. The spreading of the fire from the slab to the cladding was then investigated. Ignition of the cladding occurred in all configurations tested. The slab configuration (material or orientation) has little effect on the combustion of the cladding once it has been ignited. In addition, the distance between the deck slab and the cladding did not have a significant impact on the fire behaviour of the cladding, even at the 20 cm distance recommended by the standards. [ABSTRACT FROM AUTHOR]

Highlights • A series of experiments are conducted by considering fire spreading between vehicles in a longitudinally ventilated tunnel. • The vehicular obstacle ignition time delay is analyzed with different ventilation velocity and vehicle distance. • The temperature profile near ignited vehicle and along tunnel centerline is investigated. • A mathematical model for maximum temperature is proposed by considering ignited vehicle and ventilation. Abstract The vehicle being ignited in tunnel fire could disturb the temperature distribution and result in more serious hazards. However, the temperature profile about the tunnel fire with an ignited vehicle has not been clarified. This paper conducted a series of 1/6 small scale experiments with different heat release rates to investigate the vehicle ignition time delay and temperature profile. The longitudinal ventilation velocity varied from 0.5 m/s to 2.0 m/s, and the blockage-fire distance varied from 0.35 m to 1.6 m. The temperature and the ignition time delay were measured. Results show that the ignition time delay is shorter with shorter blockage-fire distance. The ignited vehicle results in the temperature hump and larger high temperature region at downstream. A modified model to predict the maximum temperature beneath the tunnel ceiling by considering the combined effect of longitudinal ventilation and blockage-fire distance is proposed. Finally, the predictions obtained by proposed model are compared with the experimental data and previous experiments. [ABSTRACT FROM AUTHOR]

The University of Edinburgh and its alumni have made significant contributions to knowledge in the field of tunnel fire safety engineering. This paper summarises the situation of tunnel fire safety in the early 1970s, when the department of fire engineering was founded and briefly discusses all the contributions to knowledge in the field, made by Edinburgh and its alumni in the past four decades. Research carried out at Edinburgh has changed the way the tunnel safety industry estimates heat release rates in tunnels, has influenced way design fires are specified and has challenged industry opinion about the use of water sprays in tunnels. This paper is part of a celebration of four decades of fire research at Edinburgh. • 40 years of tunnel fire research at Edinburgh are reviewed. • 40 years of tunnel fire incidents are discussed. • Edinburgh researchers and alumni have changed the way tunnel fires are perceived. [ABSTRACT FROM AUTHOR]

Internationally, informal settlements (also known as slums, favelas, shantytowns, ghettos, etc.) are expanding rapidly which have been associated with numerous large fires (where more than 10,000 people can be left homeless in a single disaster [ 1 ]), and an increase in fire related injuries and deaths. As a result of an increase in fire related incidences, various proposed interventions have been funded from governmental or private sources. However, many interventions have been unsuccessful because of technical, political, economic, social or practical issues. It is with the above mentioned in mind that this paper seeks to provide an understanding of fire behavior between multiple informal settlement dwellings (ISDs) as a baseline for future research and experiments. The paper starts by defining informal settlements in terms of construction and continues to look at the potential causes and hazards associated with informal settlement fires. The following two full-scale informal settlement dwelling burn experiments were conducted; (a) three ISDs clad with steel sheeting and (b) three ISDs clad with timber. The heat fluxes and temperatures of both experiments are shown and compared to understand the effect of cladding materials (i.e. timber cladding versus steel cladding, which are mostly used in informal settlements) on fire spread. Some general behavioral observations made during the full-scale experiments are given and discussed (i.e. how fire spreads, collapse mechanisms, etc.). Fire spread times are reported for both experiments, and range between 4 and 9 min (i.e. from the start of flashover in the first dwelling to the end of flashover in the last dwelling). These experiments are one of the bigger ISD experiments conducted to date and are the first experiments that measured heat fluxes for ISDs, which is crucial to understand fire spread. An important finding in this paper is that there is a critical separation distance of 3.8 m needed to prevent fire spread between dwellings. [ABSTRACT FROM AUTHOR]

The aim of this work is to improve the understanding of the fire behaviour of building materials used in the wildland urban interface at a large scale. To this end, wood cladding was exposed to the heat flux from the burning of a cistus hedge ignited by a spreading fire front in a fuel bed of excelsior litter. Six experiments were performed with different configurations: hedge/cladding distance, cistus bulk density, hedge length. During the experiments, the assessment of the cladding ignition was observed by camera. Wind direction and velocity were measured continuously since the tests were conducted outdoors and wind is a significant driver of flame spread. Heat fluxes and temperatures were also measured in order to compare the intensity of the thermal aggression on the cladding coming from the burning hedge. When the distance between the hedge and the cladding was 3 m, the wood did not ignite even when the wind was in the direction of propagation towards the cladding. In these cases, the maximum heat flux was 18 kW m−2. This value did not allow the ignition of the wood cladding. When the distance was reduced, the ignition of the cladding depended on the wind direction. When the cladding was ignited, the heat fluxes from the burning hedge could exceed 100 kW m−2. • Ignitability of wood cladding at the Wildland-Urban Interface. • Fire spread from a vegetation hedge to a building material. • Comparison of the critical heat flux for ignition at material and large scale. [ABSTRACT FROM AUTHOR]

Cross-laminated timber (CLT) is becoming increasingly popular due to its many advantages. However, it has been shown that exposed CLT can have a significant effect on fire dynamics and spread rates. Further studies are therefore needed to better understand the impact of CLT to fire safety. Two large-scale CLT compartment fire experiments (95 m2) representing a modern office building have been performed, #FRIC-01 and #FRIC-02. This paper presents the second experiment, #FRIC-02, with exposed CLT on the back wall and the ceiling. The fire developed fast and spread across the room in less than 3.5 min from ignition of the wood crib on the floor and in 1.5 min after the ignition of the ceiling. Large external flames were observed, despite the compartment being well-ventilated. The 5-layer CLT, which comprised a 40 mm thick exposed outer layer and was face-bonded using a common European polyurethane adhesive, exhibited glue-line integrity failure and led to a second flashover after a significant period of decay. Subsequent layers of 20 mm also delaminated before the fire was manually extinguished after 3 h. Compared to #FRIC-01, the fire spread rate was faster, and temperatures, charring rates, heat release rates and external flames were higher. [Display omitted] • Flashover after 3 min and 13 s. • Large and asymmetrical external flaming despite compartment being well-ventilated. • Self-extinguishment of flames at CLT while the wood crib was burning. • 2nd flashover after a long decay phase, followed by continuous fluctuations of fire intensity. • Strongly non-uniform charring and delamination. [ABSTRACT FROM AUTHOR]

Surface fire is the most prevalent type of wildland fire and dominates the cost of wildland firefighting. The high-intensity burning behaviors in surface fires stem from the significant interaction of combustion with heat transfer and fluid flow under complicated fuel, meteorology, and topography conditions. Over the past several decades, research has resulted in incrementally enhanced insights into the mechanism transformations and behavior transitions related to surface fires, with theories and models developed and validated. These advances have improved the fundamental understanding of surface fire evolution significantly. However, the current knowledge is still inadequate for dealing with practical surface fire problems. In this paper, we specifically address two complexities associated with fire growth. First, the fuel preheating mechanisms involving various forms of flame heat transfer may change under different conditions. Second, a surface fire may transition to other intensely burning behaviors by interacting with environmental conditions. For the first complexity, we initially discuss the fuel preheating mechanism, flame propagation across discrete fuels, and flame geometry transformation in wind-driven fire spread. Then, we examine the firelines interaction and its induced junction fire. Followed are the fireline transformations under different topography conditions, focusing on the flame attachment in upslope terrain and the fireline behaviors in canyon terrain. For the second complexity, we highlight the transitions between surface fire and other behaviors with distinct mechanisms, including smoldering, crown fire, fire whirl, and spot fire. Despite significant advances, more understanding of the mechanisms underlying fuel preheating processes and fireline behaviors is needed. Furthermore, we need more exploration into the scaling law of combustion dynamics of surface fires and the non-steady transitions to other intense burning behaviors. All these challenges have proven an elusive goal for future work to enhance surface fire predictive capabilities. • Fuel preheating and flame geometry transformation under the wind are reviewed. • Mechanisms of burning and merging of multiple firelines are discussed. • Fireline transformations under different topography conditions are investigated. • Transitions between surface fires and other behaviors are examined. • Challenges for enhancing surface fire predictive capabilities are proposed. [ABSTRACT FROM AUTHOR]

• Fire simulators compute wind speed using fixed Wind Adjustment Factors (WAF). • Canopy cover is the variable with the greatest influence on WAF. • Canopy base height could improve the performance of the WAF model. • Wind at 2-m plays an essential role in low intensity fires and prescribed burns. • Our tree decision approach can be used to simulate canopy management alternatives. Wind speed, which is significantly affected by terrain and vegetation, is one of the most crucial factors in terms of fire spread. Fire simulators compute wind speed as open 10-m wind speed multiplied by the Wind Adjustment Factor (WAF). In forested sites, the sub-canopy wind speed plays a fundamental role in low intensity fire or prescribed burn, and therefore, in prescribed burn plan. This research aims to estimate a WAF model based on in-stand wind speed at 2-m above ground and forest characteristics. We use Chi-square Automatic Interaction Detection to classify WAF according to the most influential stand variables. Fifty-seven sampling sites were established for WAF training and testing. WAF ranged from 0.03 to 0.84, showing significant differences due to stand characteristics. Our findings showed that canopy cover is the variable with the greatest influence on WAF. On the one hand, the non-linear WAF model reached a coefficient of determination (R2) of 90 %. On the other hand, a decision tree performed four decision nodes based on canopy cover, stand height, and stand density. These approaches propose a novel method to the identification of WAF for fuel treatment and prescribed burning implementation, avoiding the errors that could be generated using fixed WAF for each fuel model. The proposed model can be used to simulate the effect of different canopy management alternatives, both fuel treatments and timber harvesting, in wind speed at 2-m height. [ABSTRACT FROM AUTHOR]

Electrical cabinet-related fire events are among the most recorded fire events in nuclear power plants. As part of the OECD PRISME-3 project, four fire tests were carried out to highlight the fire spread paths from an open-door electrical cabinet (CC) to two adjacent closed-door electrical cabinets (ACs). The cabinets, separated by a double wall with an air gap, were connected by overhead and subfloor cable trays. These last ones were filled with either a "low" or "well-qualified" (LQ or WQ, respectively) cable type. Each cable type was used for two tests. In addition, tests were conducted either in open atmosphere conditions (two) or in confined and mechanically ventilated conditions (two). These tests highlighted three fire spread paths from the CC to the ACs. Regarding the first one through the double wall with an air gap, several ignitions of electrical components (used as targets) occurred in the ACs when the wall temperature near the ignited target exceeded 600 °C (or the heat flux incident to the target was higher than 26 kW/m2) whatever the test conditions. Furthermore, the two other fire spread paths through the overhead and subfloor cable trays were highlighted only if the LQ cable type was used. • Three fire spread paths were shown between adjacent cabinets separated by a double wall and connected by cable trays. • The fire spread paths were highlighted for the two studied test conditions. • Determination of the temperature and heat flux criteria leading to the fire spread via the double wall. • Impact of the cable type on the fire spread paths via the cable trays. [ABSTRACT FROM AUTHOR]

Exposing cross-laminated timber (CLT) structures in buildings is increasingly popular in modern buildings. However, large timber surfaces, window facades, and different geometries can change the fire dynamics in a compartment. The effect of those parameters, therefore, needs to be studied. Two large-scale CLT compartment fire experiments (95 m2) have consequently been performed. The experiments were designed to represent a modern office building with an open-plan space and large window openings. In this experiment, #FRIC-01, the ceiling was exposed. The wood crib fire developed slowly and travelled approximately 1.5 m before the ceiling ignited at 32.5 min. Thereafter the fire spread rapidly across the ceiling and wood crib before it shortly after retracted. Three such cycles of rapid spread followed by a retraction occurred within 13 min, whereby the wood crib fire grew larger for each cycle. After the flames extended through the compartment for the fourth time, the fire remained fully developed. After a short period of intense burning, the CLT self-extinguished while the wood crib fire was still burning. The compartment withstood full burnout, and no reignition occurred despite some delamination and using an adhesive that lacks a demonstrated resistance against glue-line integrity failure. [Display omitted] • Late ignition of ceiling due to flames not impinging the ceiling. • Fast fire spread after ignition of the ceiling. • Fire spread in several flashing waves. • Self-extinction of the CLT ceiling while wood crib was still burning. • No 2nd flashover despite presence of areas with delamination. [ABSTRACT FROM AUTHOR]

A novel Large Eddy Simulation (LES) based fire field model that incorporates pyrolysis modelling has been developed. This model is specifically designed for flame propagation of wildland fire scenarios. It uniquely embraces the radiation heat feedback from the flame, gaseous combustion and soot products towards the fuel bed surface. It also considers the detailed chemical kinetics for combustion, primary soot incipient and oxidant for soot formation, turbulent microscopic fuel–air mixing which are fully coupled, interactive and non-linear. Numerical simulation has been performed to study the effect of slope angle on the flame propagation characteristic of pine needle fuel beds. The fire spread rate and temperature predictions are within 12% accuracy in comparison to experimental data. Owing to the unbalanced air entrainment drew by the flame combustion for the inclined slope angle cases, it can be observed from the visualised flame that it was tilted to the unburned portion of the board. This behaviour strongly promotes the radiative heat transfer from the flame onto the fuel bed, which led to a rapid increase in pyrolysis rate thus accelerated the overall flame spread on the board surface. This physical phenomenon was successfully captured by the pyrolysis model and was found to be significantly more accurate in predicting the fire spread rate for slope angles higher than 20° (error of 11.12%) when compared to empirical flame tracking methods (error of 87.19%). [ABSTRACT FROM AUTHOR]

Modelling the dynamics of wildfires is very computationally challenging. Although three-dimensional computational fluid dynamics (CFD) models have been successfully applied to wildfires, the computational time required makes them currently impractical for operational usage. In this study, we develop a two-dimensional propagation model coupled to a ‘pyrogenic’ potential flow formulation representing the inflow of air generated by the fire. This model can accurately replicate features of fires previously unable to be simulated using current two-dimensional models, including development of a fire line into a parabolic shape, attraction between nearby fires and the observed closing behaviour of ‘V’ shaped fires. The model is compared to experimental results with good agreement. The pyrogenic potential model is orders of magnitude faster than a full CFD model, and could be used for improved operational wildfire prediction. [ABSTRACT FROM AUTHOR]

Local point measurements of fire dynamics in field-scale experiments of wildland fires are highly useful. This is true both for understanding the mechanisms driving fire spread that result in the observed macroscopic behaviors, but also in terms of providing comparison points for numerical tools, such as detailed physics-based fire behavior models. This work describes measurements of temperature, velocity, and radiative heat flux that were made in a field-scale fire experiment in a pine forest, with the aim of providing both of the above benefits. Regions of both surface fire and crown fire were captured and are compared. The crown fire exhibited tall upright flames, compared to the shorter tilted flames of the surface fire. Crown fire resulted in a significant increase in integrated radiative preheating, by a factor of ∼1.75, as well as greater flow sheltering in the downstream region of the fire front. Further, a corrective factor is introduced for oblique sensor placement relative to the fire front, in order to improve the value of these and other measurements, particularly for model comparison. The presented methodology, while able to be improved, is shown to successfully characterize local differences in fire behavior. [ABSTRACT FROM AUTHOR]

The complex interactions between the inclined terrain and the flow generated by the fire make the slope one of the most influencing factors on fire spread. In order to gain a deeper understanding of the mechanisms involved in wildfires spreading upslope, the investigation of flow dynamics and heat transfers is fundamental. This paper reports a series of fire spread experiments conducted across a porous bed of excelsior in a large-scale facility, under both no-slope and 30° up-slope conditions. The coupling of particle image velocimetry and video imaging allowed characterizing the flow pattern with respect to the fire front. Simultaneous heat flux measurements with high scan rate were also performed at the edge of the fuel bed. From the collected data, the increase of the rate of spread with increasing slope is attributed to a major change in fluid dynamics surrounding the flame. For horizontal fire spread, flame fronts exhibit quasi-vertical plume resulting from the buoyancy forces generated by the fire. These buoyancy effects induce an inward flow of ambient air that is entrained laterally into the fire from both sides. Flame radiation is the dominant fuel preheating mechanism. Under upslope conditions, the fire plume is tilted toward the unburnt vegetation, increasing radiation levels. The air entrainment at the burnt side of the fire strongly influences the downstream flow, which becomes attached to the surface over a characteristic length scale. Ahead of the flame front, the induced wind blows away from the fire rather than toward it, enhancing convective heating. Periodical forward bursts of flame combined with distant fuel ignitions were also observed. The heat flux measurements confirmed the existence of such convective mechanisms. [ABSTRACT FROM AUTHOR]

In this paper, we report on our investigations into the behavior of successive fire spread between multiple houses in an urban area. More specifically, to acquire data necessary for the validation of our computational models, we conducted a fire spread experiment with nineteen 1/3-scale model houses aligned in a former quarry at Iwafune Mt. in Tochigi Prefecture, Japan. Most previous experiments used wood cribs or gas burners as fire sources to investigate the behavior of a fully-developed urban fire, but gave less attention to its formation due to successive occurrence of fire spread between houses. Successive fire spread between houses cannot be adequately simulated as discrete fuel elements in urban areas without replicating the exterior geometry, including openings allocated in each wall; however, this factor, which affects the intensity of heat transfer between houses, has not yet been considered in previous experiments except those that used disused houses. The experiment comprised two tests. In test 1, fire spread between model houses did not occur as was originally expected because of the short duration of the fully developed fire phase at the fire origin model house. The average mass loss rate in the fully developed fire phase was 1.33 kg/s with the maximum flame height of 7.8 m. In test 2, ceilings and roofs of model houses that were left unburnt in test 1 were reinforced so as to enhance their compartmentation, which in turn extends the duration of fire at each model house and the heating period of the adjacent model houses. As a result, fifteen out of eighteen model houses were burnt due to successive fire spread that lasted for about 50 min. The estimated maximum heat release rate of the group fire was about 100 MW with the maximum flame height of 11.0 m. The flame height after roof burn-through was modeled by introducing a conversion factor into one of the standard flame height model considering fresh air supply through openings due to the pressure difference between inside and outside of a model house. Although the quantitative agreement was not satisfactory, the overall change in flame height over time was adequately captured with the present model. [ABSTRACT FROM AUTHOR]

• Key element index (KI) indicates contribution of columns to overall strength of a building. • Collapse scale becomes drastically larger as the integrated value of KI exceeds a threshold. • The collapse threshold depends on the outbreak locations of fire. • Collapse scale can be estimated during fire spread using the integrated value of KI. In this study, the collapse behavior of a steel-framed building induced by the spread of a large-scale fire was analyzed using the ASI-Gauss code, and the collapse scales of buildings were estimated using a key element index. The results of the fire-induced collapse analyses indicated that the collapse scales were highly dependent on the fire range and location. The collapse scales tended to increase when the fire range widened on the lower floors. Moreover, there was a greater increase in scale when fire occurred in the peripheral area of the building, compared to when the fire occurred in the inner area. Based on the numerical results, the collapse scales of a building were examined using the relationships between the integrated values of the key element index (KI), a parameter that indicates the contribution of columns to the overall strength of a building, and the sum of the heights of the remains. It was confirmed that the collapse scales were larger when the integrated value of KI of the columns in the spreading range of the fire reached a specific threshold, indicating that the collapse scales of buildings can be estimated by accumulating the KI of damaged columns. [ABSTRACT FROM AUTHOR]

The computer simulation of fire spread based on modelling of complex processes related to fire by Computer Fluid Dynamics CFD theory should be a significant contribution to increase of safety of road tunnels. This paper is about instant project aimed at creating simulation models of formation and spread of fires in the road tunnel. The simulations will be compared with real in-situ measurements of fire scenarios suggested by customer organization in two road tunnels. Within the project, Tunnel Traffic and Operation Simulator, as unique facility at University of Žilina will be extended by visualizations of smoke stratification in the virtual tunnel tube. [ABSTRACT FROM AUTHOR]

Spread rate is an overall property of flame propagation that characterizes the condition of a flame better than any other property. As a result, prediction and measurement of spread rate is central to flame spread studies over solid fuels. Significant amount of data have been collected over last four decades of research on flame spread over various fuels under different conditions. In most of these studies, however, only average spread rate is reported which is adequate for steady phenomena. Given that a flame may not face the same conditions during the spread, it is possible for the spread rate to change during the duration of the spread continually. In this work a methodology for image analysis is presented with the goal of evaluating instantaneous spread rate to study time-dependent phenomena. The parameters that control the error and time resolution of the flame spread history are identified, and a sensitivity study is carried out to validate the results of a scale analysis. A MATLAB-based Flame Image Analyzer (FIA) package is developed and applied to flame spread videos recorded in several experiments in different regimes of opposed-flow flame spread. An expression for the error in spread rate for a given time resolution is expressed in terms of the imaging parameters. The two parameters that are found most important are the pixel resolution and the frame rate. A non-dimensional imaging parameter is identified that is shown to govern the quality of imaging for spread rate measurement. Theoretical prediction from the error analysis is confirmed by doing various case studies using the Analyzer. [ABSTRACT FROM AUTHOR]

The general objective of this research is to develop a prototype data-driven wildland fire spread simulator, called FIREFLY, using an ensemble-based data assimilation approach with the objective to forecast the location and speed of the fire. The specific focus of the present study is on evaluating the new features of FIREFLY at field scale in a controlled grassland fire experiment known as FireFlux I. FIREFLY features the following components: an Eulerian front-tracking solver that treats the fire as a propagating front and uses Rothermel's model for the rate of spread (ROS); a series of observations of the fire front position (based here on high-resolution fireline data previously generated by validated numerical simulations); and a data assimilation algorithm based on an ensemble Kalman filter configured in a parameter estimation mode to address model bias and uncertainties in the input data to the ROS model. In this work, FIREFLY is modified to allow for an estimation of spatially-distributed surface wind speed and direction. To generate a reliable ensemble and ensure an accurate correction, the ensemble Kalman filter requires sampling truncated probability density functions as well as localizing, i.e ., dynamically selecting the areas where the wind parameters are corrected. Results show that the spatialized parameter estimation approach allows for a successful reconstruction of observed fireline position and shape as well as a substantial improvement in the forecast performance compared to the standalone fire spread model. Results also show that the inferred wind parameters may not be accurate and should be viewed as effective values that incorporate multiple sources of uncertainties. Developing a better representation of fire-wind interactions is thus viewed as a key aspect to improve the FIREFLY forecast capability. [ABSTRACT FROM AUTHOR]

Insulation panels made of organic, combustible materials are frequently used in the exterior thermal insulation systems (ETIS) for buildings. Such combustible insulation panels have been involved in several catastrophic building fires in recent years in China. One potential strategy to mitigate this fire hazard is to limit fire spread over the ETIS. The present work evaluates the effectiveness of vertical fire barriers in inhibiting fire spread over exposed insulation walls made of expanded polystyrene (EPS) panels. Reduced-scale experiments were carried out indoors using EPS panels with or without two vertical barriers made of non-combustible mineral wool, the fire started at the bottom center of the middle panel. The interval and width of the barriers were varied systematically, while the temperature distribution on the wall, the radiation heat flux from the fire, and the infra-red (IR) images were recorded. To demonstrate the validity of the concept, an outdoor, full-scale experiment was carried out using a 7-floor building. Our reduced-scale experiments showed that the installation of two vertical fire barriers successfully stopped the lateral flame spread, decreasing the peak temperatures of the two side panels by about 300 °C for all barrier configurations tested. When barrier width was fixed at 5 cm, an increase of the barrier interval from 30 to 90 cm led to increases in the peak temperatures, radiation heat flux, and the maximum rate of upward flame spread. By contrast, when barrier interval was fixed at 90 cm, an increase of the barrier width from 2 to 5 cm had little influence on the combustion dynamics of the middle panel but the peak temperature on the side panels dropped, consistent with the smaller heat transferred with wider fire barriers. In the regions of the side panels next to the barriers, pyrolysis and deformation could be observed with barrier widths of 2 and 3 cm, but not 5 cm. Finally, our outdoor, full-scale experiment demonstrated that a 30 cm wide vertical barrier made of air-filled cement successfully stopped the lateral flame spread over exposed EPS wall. The study highlights the effectiveness of vertical fire barriers in preventing the lateral flame spread over the exposed EPS insulation wall and provides another option for enhancing the fire safety of the combustible insulation systems. [ABSTRACT FROM AUTHOR]

This work is devoted to the development of a small-world network model to predict real-time fire spread onboard naval vessels. This model takes into account short-range and long-range connections between neighboring and remote network compartments. Fire ignition and flashover, as well as fire transmissions through the walls and ventilation ducts are simulated using time-dependent normal probability density functions. Mean durations of fire transmission through the walls and ducts are determined by a three-zone model and a one-dimensional CFD code, respectively. Specific experiments are conducted in a steel room, representative of a naval vessel compartment, in order to validate the zone model. Then a proof of concept is developed by applying the network model to a full-scale vessel mockup composed of 113 compartments on 7 decks. A statistical study is conducted to produce fire risk maps, classifying the vessel compartments according to their propensity to burn. [ABSTRACT FROM AUTHOR]

Several studies in the literature explore the connection between rate of spread (ROS) and wind in wildland fires. These studies show very different positions about the role of radiation and convection as heat transfer mechanisms. In the case when the fuel bed is well-ordered and vertically-oriented, there seems to be a consensus leading to suggest that convective heating is the dominant heat transfer mode in that case. The purpose of this work is to propose a convective semi-physical model for the behaviour of the rate of spread in wind, when the fuel bed is vertically-oriented. Due to a specific fuel bed arrangement, flame radiation –i.e. radiation from the part of the flame above the vegetal stratum– is neglected. Only horizontal radiation from the fuel burning particles area and convective heating are taken into account. Convective heat transfer is assumed to be the primary heat transfer mechanism. The proposed model is confronted to 172 laboratory fires with a wide range of fuel characteristics. The predicted results are also compared with two simplified models from the literature. Statistical tools are used to check the agreement between the predicted ROS and the observed one where a strong agreement is generally observed, irrespective of fuel bed characteristics. [ABSTRACT FROM AUTHOR]

Fire behaviour and spread predictions guides suppression strategies and public warnings during wildfires. The scientific understanding of fire behaviour forms the core of these predictions, but is incomplete, and expert judgement and experience are required to augment the evidence based knowledge. Amicus is a new decision support system that implements contemporary, published and operationalised bushfire behaviour models (e.g. rate of spread, flame height, fireline intensity, spotting distance) in the Australian bushfire context. It enables the inclusion of expert judgement and local knowledge, allows users to analyse temporal trends and uncertainty in inputs, and facilitates reliable and practical predictions. This paper provides a comprehensive overview of Amicus, including its operation and functionality, identifies the boundaries of the current understanding of fire science, discusses the major limitations in existing knowledge, and provides a framework for allowing deterministic and anecdotal/local knowledge to be incorporated into formal fire behaviour predictions. [ABSTRACT FROM AUTHOR]

Parametric uncertainty quantification of the Rothermel's fire spread model is presented using the Polynomial Chaos expansion method under a Non-Intrusive Spectral Projection (NISP) approach. Several Rothermel's model input parameters have been considered random with an associated prescribed probability density function. Two different vegetation fire scenarios are considered and NISP method results and performance are compared with four other stochastic methodologies: Sensitivity Derivative Enhance Sampling; two Monte Carlo techniques; and Global Sensitivity Analysis. The stochastic analysis includes a sensitivity analysis study to quantify the direct influence of each random parameter on the solution. The NISP approach achieved performance three orders of magnitude faster than the traditional Monte Carlo method. The NISP capability to perform uncertainty quantification associated with fast convergence makes it well suited to be applied for stochastic prediction of fire spread. [ABSTRACT FROM AUTHOR]

Wildland-Urban Interface (WUI) fires are becoming more common globally. However, most research has been focused on developed regions such as North America and Australia, exposing a lack of geographical diversity in the field. This paper presents an analysis of South Africa's largest WUI fire disaster, in terms of the number of structures lost and the economic losses, the 2017 Knysna fires. The data used in this report is based on aerial reconnaissance conducted immediately after the event, a geo-located database identifying homes that were destroyed, satellite imagery and site visits. The analysis includes the fire spread mechanisms, firebreak efficiency, the influence of roof and wall types, vegetation, and weather conditions. The Canadian Fire Weather Index (FWI) values on the day of ignition were unprecedented in the 1979–2021 period, the culmination of a clear multi-decadal trend towards greater fire weather risk. Due to the combination of severe medium-term drought and extreme fire weather conditions homes up to 1 km inside the WUI were destroyed. The high wind speeds allowed the fire to cross natural and artificial barriers such as rivers and highways, allowing the ignition of several secondary fires due to ember attacks. Construction materials and fire exposure, which is influenced by the presence of vegetation, affected home survivability. Combustible materials (thatch and timber) typically made homes more susceptible to fire damage. It was found that in the areas where more structures were affected the vegetation were dense, highly flammable stands of alien invasive vegetation. [ABSTRACT FROM AUTHOR]

• Fire spread in tunnels using longitudinal ventilation was studied. • A simplified temperature model for tunnel with multiple fire sources was proposed. • Criterion for fire spread to two nearby vehicles in tunnel was investigated. • Critical fire spread distance in a tunnel with multiple fire sources was discussed. The characteristics of fire spread among multiple vehicles in tunnels using longitudinal ventilation were investigated by analyzing the experimental data from a series of fire tests in a 1:15 scale tunnel. Further, a simple theoretical model for gas temperature in a tunnel with multiple fire sources was proposed and used in analysis of the experimental data. The results showed that, for objects (wood piles) placed at a same separating distance downstream of the fire, the fire spread occurred faster and faster along the tunnel. Validation of the simplified temperature model for multiple fire sources was made against both model and full-scale tunnel fire tests. The model was further used to predict the critical conditions for fire spread to the second and third objects. Comparisons with the test data showed that average excess temperature of 465 K (or an equivalent incident heat flux of 18.7 kW/m2) could be used as the criterion for fire spread, and this was verified further by other model-scale tests and full-scale tests. The results showed that the critical fire spread distance monotonously increases with the heat release rate, and decreases with the tunnel perimeter. For multiple fire sources with equivalent heat release rates, as the separation distance between the first two fire sources increases, the critical fire spread distance from the second fire source to the third fire source decreases, but the total fire spread distance from the first fire source to the third one increases. If the total heat release rate at the site of a downstream fire source is greater than that at the former fire source, the critical fire spread distance becomes longer. [ABSTRACT FROM AUTHOR]

As wildland fire activity continues to surge across the western US, it is increasingly important that we understand and quantify the environmental drivers of fire and how they vary across ecosystems. At daily to annual timescales, weather, fuels, and topography are known to influence characteristics such as area burned and fire severity. An understudied facet, however, concerns how these factors inhibit fire spread and thereby contribute to the formation of fire boundaries. We evaluated how weather, fuels, and topography impeded fire spread in four large study areas in the western US, three in the Northern Rockies and one in the Southwest. Weather and fuels were the most important factors in the Northern Rockies, whereas fuels and topography were dominant in the Southwest. Within the categories of weather, fuels, and topography, we also evaluated which specific variables were most influential in impeding fire spread. We explicitly accounted for the presence and age of previous burns within the fuels category. We found that: (1) temperature was the most influential weather variable in the Northern Rockies; (2) previous burns (particularly those that were ⩽5 years old) were moderately to highly influential in all study areas; and (3) valley bottoms and ridgetops were moderately to highly associated with fire boundaries in all study areas. Our results elucidate the regionally varying roles of weather, fuels, and topography in impeding fire spread, emphasizing each ecosystem’s unique biophysical setting and fire regime. [ABSTRACT FROM AUTHOR]

CUrisk as a comprehensive fire risk analysis model have been under active development over the past decade. Recently a Barrier Failure submodel and a Fire Spread submodel have been integrated into the CUrisk system structure. CUrisk can provide data of barrier failure followed by results of fire spread through the Fire Spread submodel. This paper presented a fire risk analysis case study of a six-storey apartment building of light-frame construction, aimed at showing the performance of CUrisk and particularly the effect of wall barriers on building fire risk. The apartment building has 18 different designs of wall barriers, all of which are light frame assemblies but with different configurations. These wall barriers have different failure times under the unsuppressed fires in the apartment of fire origin, and consequently resulted in different times of fire spread to the adjacent compartments. The maps of fire spread in the unsuppressed fires at 60 minutes demonstrated good patterns and the important role of fire barriers in containing a fire. The impact of wall barrier designs on occupant risk is also discussed. Results showed that the building designed with better fire-resistive barriers could lead to lower occupant deaths, which are presented quantitatively. [ABSTRACT FROM AUTHOR]

• Effects of longitudinal ventilation on fire properties in a tunnel were studied. • Higher ventilation speed can reduce the high-temperature area near the fire source. • Fire spread can only occur by flame impingement if the fire is not very large. • The formula of flame tilt angle was developed. Fire spread between vehicles is of a great concern in a tunnel fire. Longitudinal ventilation, which is commonly used for the smoke control in tunnel fire, would inevitably affect fire properties and fire spread in tunnel at the same time. Such effects were studied using model-scale experiments in this paper. The results show that fire spread mostly occur due to the flame touching arising from flame tilting under longitudinal flow. Safety spacing should be kept for preventing fire spread between vehicles. The effect of longitudinal ventilation on the temperature distribution in the downstream area of the fire source is found mainly limited to the vicinity of the fire source. Maximum ceiling temperature, flame tilt angle and downwind flame temperature under longitudinal ventilation are found to be heavily dependent on Fr (Froude number) and dimensionless Q* (Dimensionless heat release rate). The formulas of flame tilt angle and mean downwind flame temperature were developed and predicted results were compared with those in modelling experiments. [ABSTRACT FROM AUTHOR]

• Develop an AI model to forecast dynamic tunnel fire scenarios with changing fire location and size. • Establish a numerical fire database of 300 transient tunnel fire scenarios to train the model. • Dual-agent deep learning model (TCNN + LSTM) to forest tunnel fire scene with 30 s lead time. • Demonstrate the capacity of AI model in forecast the rapid-changing tunnel fire for smart firefighting. Disastrous fire accidents occurred in the tunnel is fatal and destructive, which may pose great threats to the trapped person and firefighters. However, fire in a confined tunnel can develop rapidly and spread between vehicles, so it is difficult to forecast the fire development and possible catastrophic incidents. This work develops an intelligent model to predict the fire information, temperature distribution and critical events in real-time based on artificial intelligence algorithms. The numerical model is first validated by the full-scale tunnel fire test, and then a numerical database of 300 transient tunnel-fire scenarios is established under various initial fire locations, fire sizes, fire growth and spread rates, and ventilation conditions. The proposed dual-agent deep-learning model combining the Long Short-term Memory (LSTM) model and Transpose Convolution Neural Network (TCNN) is trained with the database. With the input data of on-site temperature sensors, the dual-agent model can forecast transient fire scenarios with changing location and size 30 s in advance. This study demonstrates the feasibility of the AI model in identifying and forecasting the rapid-changing fire scenarios inside a tunnel in smart firefighting practices. [ABSTRACT FROM AUTHOR]

Modelling oxygen-limited and self-sustained smoldering propagation is of significance for prevention of fire hazards and optimization of applied systems. However, two issues remain unsolved in the conventional models: (a) adjustment of TG-scale kinetic parameters applied to bed-scale propagation, and (b) decoupling oxidative reaction and oxygen transport in multi-scale porous media. In this work, an analytic expression of oxidative reaction rates limited by oxygen transport is derived from the conservation equations of oxygen species transport in gas and solid. Then, both oxidative reaction rates controlled by the kinetic reaction (Arrhenius equation) and oxygen transport (analytic expression) are integrated into the conservation equations of mass, energy, and oxygen species transport. In this model, five-step reaction scheme is considered and their kinetic parameters obtained from TG experiments are employed without any adjustment. Along with the Darcy air flow driven by thermal buoyancy, this model is applied to predict oxygen-limited and self-sustained smoldering propagation of underground coal fires. The proposed model is compared with laboratory experiments and the conventional model. Results show that the proposed model well predicts the oxygen-limited and self-sustained smoldering propagations of underground bituminous and anthracite coal fires. The predictability of the proposed model is better than the conventional model in spite of great effort to modify kinetic parameters best fitting with experimental data. It is validated that the proposed model addresses the two puzzled issues in the conventional model with respect to buoyancy-driven, oxygen-limited, and self-sustained smoldering propagation of underground coal fires. This work may help to improve models of self-sustained propagation of other smoldering fires and applied smoldering systems. [ABSTRACT FROM AUTHOR]

• An engineering CFD model for fire spread on wood cribs is demonstrated. • Fire spread characterised at stick-by-stick level with at least 2 × 2 cells. • Global matching of heat release rate, fire spread, burn-away and flame temperatures. • Approximate linear regression between different fire modes and incident radiant heat fluxes. • Heat release rate more sensitive to fire spread than stick burning rate. The temperature heterogeneity due to fire in large open-plan office compartments is closely associated with fire spread behaviour and has been historically limited to experimental investigations using timber cribs. This study explores the ability of Computational Fluid Dynamics (CFD) models, specifically the Fire Dynamics Simulator (FDS), to reproduce the results of full-scale tests involving fire spread over timber cribs for continuous fuel-beds. Mesh schemes are studied, with a fine mesh over the crib and 2 × 2 cells in the wood stick cross-section by default, this being relaxed in the surrounding regions to enhance computational efficiency. The simple pyrolysis model considers the charring phase and moisture. In application to the TRAFIR-Liège LB7 test, this calibrated "stick-by-stick" representation shows a good agreement for interrelated parameters of heat release rate, fire spread, gas phase temperature, and burn-away, a set of agreements which has not been demonstrated in previous studies. Fire spread shows relatively high sensitivities to: heat of combustion, ignition temperature, thermal inertia, radiation fraction, heat release rate per unit area, and the fuel load density. An approximately linear regression was found between the different fire modes and the thermal exposures, with "travelling" (and decaying) fires characterised by heat fluxes associated with the fire plume, while the growing fires were associated with proportionally higher heat fluxes on the horizontal surfaces of the sticks, in conditions where these receive more pre-heating. The trends in the overall HRR are more dependent on the fire spread rates than variations in the stick burning rates. [ABSTRACT FROM AUTHOR]

The determination of firefighting water requirements in large, tall and complex buildings in the UK has largely been based on data retrieved from studies into full-scale test fires, combined with some real fire research undertaken by the Fire Research Station (FRS) from 1955–1970. This scientific research effectively formed the basis for the design and configuration of rising fire mains (stand-pipes) in tall buildings, as well as water storage provisions for town centre and infrastructure planning. The resulting guidance is also applied to individually isolated buildings that are distanced from the nearest available water supply. However, it is suggested the UK has since fallen behind many international standards and codes, where firefighting water provisions are more reflective of modern building design as movable fire loads, compartment dimensions and window (ventilation) sizes have increased over the years. There have been recent calls in both the UK and USA for the development of a performance based method of calculating firefighting water requirements for design purposes, based on the quantity of water actually being used effectively by firefighters. The recent publication of BS PD 7974:5:2014 (7974) meets this need in calculating ‘adequate’ and effective firefighting water (s.8.5). The research by Glasgow Caledonian University (GCU) provides a framework upon which the ‘7974’ water strategy evolved and is based on an analysis of 5401 building fires that occurred in the UK from 2009 to 2012, where active firefighting was undertaken across a broad range of occupancies. When used in design, any recommended deviations from the prescriptive codes may achieve some cost/benefit advantages, whilst at the same time providing an improved firefighting water flow density (L/min/m 2 ). It is also worth noting that UK prescriptive building codes do not differentiate between the volume of firefighting water storage required for residential or commercial buildings. Furthermore, reductions in compartment size and fire load, or enhancements to passive/active fire protection, may mean even less storage water is required when using the 7974 strategy compared to current code compliance. The GCU research clearly established that the severity of building fires is dependent on the size of compartment in which the fire originates; the containment of fire spread through the provisions of adequate passive or active fire protection; the fire load density and the potential for a fire to reach a fuel controlled burning regime at peak heat release for the compartment at Q max . It is further demonstrated that the quantity of firefighting water that may be deemed as ‘adequate’ can be presented on a gradient ( Fig. 13 ), ranging from the lesser amount (L/min) required in residential buildings (low-flow), upwards through offices or commercial mixed use buildings (mid-flow), to industrial and storage facilities (high-flow). This finding is not reflected in current codes. In using the ‘7974’ methodology to calculate adequate firefighting water, simple equations are utilised in this paper for both sprinkler protected and non-sprinklered occupancies. The method can be also be used where either dry or wet rising fire mains (standpipes) are required. In referring to past fire case histories it is then demonstrated that fires can travel across large floor plates, with great speed. A spread rate of >20 m 2 /min may see a fire develop with such velocity and power that any fire service intervention can become compromised from the outset, unless adequate firefighting water provisions are available and effective fire protection measures are in place. [ABSTRACT FROM AUTHOR]

Five large-scale fire tests, including one pool fire test and four HGV mock-up fire tests, were carried out in the Runehamar tunnel in Norway in year 2003. New data and new analyzes are presented in this paper, together with a short summary of previous work on these tests. Heat release rate (HRR), radiation, fire spread, gas production, backside wall temperature, visibility, backlayering, fire growth rate, gas temperature, flame length, ventilation and pulsation are investigated. Simple theoretical models are developed to estimate and predict these parameters. The correlations developed can be used by engineers working on fire safety in tunnels. [ABSTRACT FROM AUTHOR]

Four different methodologies for calculating the ignition of different components on a mining vehicle in a mine drift were analysed. The results were compared with two full-scale fire experiments on mining vehicles. The four different methods are based on physical relations for fire spread between combustible components of the mining vehicles. The first two methods use a critical heat flux as ignition criterion while the other two methods use an ignition temperature. A sensitivity analysis was performed and the most influencing parameter of the methods was further analyzed. The calculated results were compared with the measured results from the experiments. The two methodologies applying an ignition temperature criterion were ruled out at as the surface temperatures of all fuel components never achieved the corresponding ignition temperatures. For the two methods applying a critical heat flux criterion it was found that the expression not including a flame radiation term was not suitable as it was found that the flame radiation played an important part with respect to spread mechanisms. The expression containing a flame radiation term was found to come very close to the observed ignition times, except in the case of the left, rear tyre of the drilling rig where it predicted a much higher ignition time than the one observed. The difference is unclear and would have to be investigated further. It was also found that the surface heat losses had none effect on the output results and could therefore be neglected in the calculations. In the case of the wheel loader the calculated heat release rate curves did not match the measured curve as well as in the case of the drilling rig. The difficulty in this case consists of accurately predicting the mechanical failure of a component – in this case a suction hose – that would initiate the very significant hydraulic oil pool fire. [ABSTRACT FROM AUTHOR]

Connecting the results from laboratory wildland fire experiments with in-situ field experiments under natural conditions remains an important challenge in wildland fire science, particularly because of differences in flow and boundary-layer conditions in the local atmosphere. Thus, as part of a broader investigation involving laboratory and field experimentation, a large-scale, portable wind tunnel capable of use in both the laboratory and the field was designed and implemented to bridge the two experimental domains. Two axial fans generate flow speeds of up to 8 m/s within the test section. Thorough flow-field characterization revealed good uniformity and minimal spatial variability within the flow field. Uniquely, the wind tunnel is portable; due to its modular design, the entire wind tunnel can be broken down to fit onto two shipping pallets for transportation. This portability allows for future experiments to be conducted in the field under the same boundary conditions as in the laboratory. These same boundary conditions created a unique research environment to iteratively test the effects of boundary conditions to validate computational models in different domains. The results from some of the wildland fire spread experiments conducted in the wind tunnel are presented to demonstrate and validate its laboratory capabilities. • Development of a portable wind tunnel for wildland fire research. • Providing well-controlled boundary conditions for field wildfire research. • Bridging scales in wildfires by multi-scale experiments. • Apparatus for generating validation data for computational research. [ABSTRACT FROM AUTHOR]

A modified particle swarm optimization algorithm is proposed in this paper to investigate the dynamic of pedestrian evacuation from a fire in a public building—a supermarket with multiple exits and configurations of counters. Two distinctive evacuation behaviours featured by the shortest-path strategy and the following-up strategy are simulated in the model, accounting for different categories of age and sex of the pedestrians along with the impact of the fire, including gases, heat and smoke. To examine the relationship among the progress of the overall evacuation and the layout and configuration of the site, a series of simulations are conducted in various settings: without a fire and with a fire at different locations. Those experiments reveal a general pattern of two-phase evacuation, i.e., a steep section and a flat section, in addition to the impact of the presence of multiple exits on the evacuation along with the geographic locations of the exits. For the study site, our simulations indicated the deficiency of the configuration and the current layout of this site in the process of evacuation and verified the availability of proposed solutions to resolve the deficiency. More specifically, for improvement of the effectiveness of the evacuation from the site, adding an exit between Exit 6 and Exit 7 and expanding the corridor at the right side of Exit 7 would significantly reduce the evacuation time. [ABSTRACT FROM AUTHOR]

Fire risk models support decision making for engineering problems under the consistent consideration of the associated uncertainties. Empirical approaches can be used for cost-benefit studies when enough data about the decision problem are available. But often the empirical approaches are not detailed enough. Engineering risk models, on the other hand, may be detailed but typically involve assumptions that may result in a biased risk assessment and make a cost-benefit study problematic. In two related papers it is shown how engineering and data-driven modeling can be combined by developing a generic risk model that is calibrated to observed fire loss data. Generic risk models assess the risk of buildings based on specific risk indicators and support risk assessment at a portfolio level. After an introduction to the principles of generic risk assessment, the focus of the present paper is on the development of a generic fire risk model for single family houses as an example. The risk model considers the building characteristics of a single family house as well as the uncertainties associated with the fire spread in a building and the intervention of the fire brigade. [ABSTRACT FROM AUTHOR]

Highlights: [•] Smoke removal in tunnel fire is significant with multiple ceiling vents. [•] Significant change in tunnel flow characteristics by varying heat source and ceiling vent locations. [•] As distance from fire source and ceiling vent increases smoke venting efficiency decreases. [•] Tunnel temperature distribution increases with decrease in vent size. [ABSTRACT FROM AUTHOR]

Abstract: The barrier effect of balcony is significant on vertical spread of window spill plume along building exterior facade. A series of reduced- scale fire simulation experiments were carried out in order to investigate the reduction of vertical fire spread due to the present of a balcony in different depth and width. Finally conclude:(1)the depth and width of balcony play different degrees of barrier effect on vertical spread of window spill plume along building exterior facade, (2)When the balcony width is longer enough, the balcony depth playing dominant barrier effect, (3)when the balcony depth is long enough, with th% balcon9 width increase, the wall temperature decrease slightly.These conclusions provide fire prevention of the building exterior faca$e with a reference. [Copyright &y& Elsevier]