Fire

The annual number of forest fires at the provincial scale in China was predicted to support regional resource allocation and preventive mitigation planning. Drawing on climate, forest resource, and socioeconomic factors together with forest fire incident records from 2003 to 2023, we construct a deep learning model that integrates a Convolutional Neural Network—Gated Recurrent Unit (CNN-GRU) with a Multi-Head Attention (MHA) mechanism and a Kepler Optimization Algorithm (KOA). The CNN-GRU captures spatiotemporal features, MHA enhances the recognition of intrinsic data relationships, and KOA automatically optimizes the hyperparameters of the hybrid model. Our KOA-CNN-GRU-MHA model outperformed all traditional machine learning baselines. Compared with the native CNN-GRU model, the proposed model reduced the mean absolute error (MAE), mean absolute percentage error (MAPE), and root mean square error (RMSE) by 26.75%, 64.35%, and 16.47%, respectively, for the total annual Number of Forest Fires (NFF) and by 39.07%, 76.03%, and 32.82% for the total annual Number of Small Forest Fires (NSFF). Ranking the influence of predictor variables further supports the identification of key factors associated with provincial fire occurrence trends, which can assist region-level planning. Full article

This study investigates fire dynamics and passenger evacuation for a lithium-battery-powered sightseeing vessel using Pyrosim 2024 and Pathfinder 2019. Fire scenarios with varying ignition locations (lower battery compartment, first-deck cabin, second-deck cabin), heat release rates (4000 kW and 500 kW), and exit conditions were simulated to evaluate temperature, CO concentration, visibility, and smoke propagation. Evacuation simulations examined the effects of occupant load (100 vs. 185 persons), spatial distribution, and age composition on total evacuation time. Results demonstrate that ignition location governs fire development: lower-deck battery fires pose the highest risk, with smoke engulfing the entire vessel within 200 s and all compartments exceeding tenability limits within 180 s—55% faster than upper-deck fires. Reducing fire power from 4000 kW to 500 kW delays critical threshold attainment by only 22% (sensitivity coefficient = 0.032), confirming that fire location dominates over fire intensity in enclosed ship compartments. Exit blockage accelerates environmental deterioration in the fire-origin cabin by 56% but shows limited effect on remote compartments. Evacuation time increases by 56% when passenger load rises from 100 to 185, and uneven passenger distribution causes delays equivalent to adding 60–70 additional occupants. In contrast, age composition exerts only a marginal effect. The forward deck consistently remains the safest area across all scenarios. These quantitative findings provide prioritized engineering guidelines: lower-deck battery compartment fire protection, occupant density control, and uniform passenger distribution are the most critical factors for enhancing fire safety in battery-powered passenger vessels. Full article

Wildfires are an escalating natural hazard in Australia and a growing public health concern. Exposure to wildfire smoke is associated with acute and chronic adverse health effects, placing wildland firefighters at elevated risk due to recurrent occupational exposure. Although protective behaviours and personal protective equipment (PPE) are recommended, uptake remains inconsistent due to operational and behavioural barriers. This descriptive qualitative study applied the Health Belief Model to explore firefighters’ risk awareness, protective practices, motivators, and barriers. Semi-structured interviews were conducted with eleven wildland firefighters in Southeast Queensland, Australia. Participants demonstrated strong awareness of short-term health effects, informed by personal experience, but limited understanding of long-term health risks. Compliance with mandatory PPE (clothing, gloves, boots, helmets) was high and motivated by prevention of immediate injury. However, respiratory protective equipment use was inconsistent due to discomfort, poor fit, communication difficulties, and operational constraints. Additional barriers to health-protective behaviours included limited decontamination access and inequities in PPE design, particularly for women. Participants identified improved health education, practical PPE options, and further research as priorities for strengthening occupational health protection. Addressing these barriers through improved knowledge translation, organisational support, and approaches such as cultural burning may reduce exposure and improve workforce health. Full article

Perfluorohexanone (C₆F₁₂O) has been widely utilized as an eco-friendly gaseous fire extinguishing agent. However, it is highly susceptible to thermal decomposition under high-temperature conditions during the fire protection process, generating substantial amounts of corrosive hydrogen fluoride (HF). To overcome the limitations of single-component suppression media, this study systematically investigates the effects of E-CF₃-CH=CClH addition on the thermal pyrolysis of the C₆F₁₂O agent, focusing on the evolution behaviors of fluorine-containing species and HF yield variation under various addition ratios and temperatures via a tube furnace. Additionally, ReaxFF-MD simulations were deployed to elucidate the microscopic pyrolysis fragment evolution and intermolecular interaction mechanisms within the blended system. The results indicate that the pyrolysis conversion of the blended suppressants across all mixing ratios increases monotonically with rising temperature. Once the C₆F₁₂O volume fraction surpasses 20% vol, chlorinated organic species completely disappear from the pyrolysis yield, accompanied by the detection of trace amounts of hydrofluoroalkanes. Additionally, maintaining the C₆F₁₂O content at 40% vol and 50% vol enables effective control over the net HF yield. Modeling reveals that blending predominantly enhances the C₆F₁₂O pyrolysis conversion during the early reaction phase. Mechanistically, HF mitigation within the binary network is governed by a multi-pathway synergistic mechanism. The insights gained from this work provide crucial theoretical foundations and empirical evidence for the formulation optimization of novel environmentally benign blended suppressants and the mitigation of secondary hazards in fire scenes. Full article

With the rapid increase in electric vehicle ownership, fire risks associated with electric vehicles in tunnels have attracted increasing attention. In this study, a numerical model of an electric vehicle fire in a two-lane tunnel with an arched ceiling was established using the Fire Dynamics Simulator. The effects of transverse and longitudinal fire source locations on smoke layer thickness and longitudinal temperature distribution were investigated. Smoke layer thickness was identified using the visibility discontinuity method, and theoretical models incorporating the geometrical characteristics of the arched ceiling were established. The results indicate that fire source location exerts a marked influence on smoke layer development and longitudinal temperature attenuation. The maximum smoke-layer-thickness ratio exceeded unity in all cases and was generally greater downstream than upstream of the fire source. During the quasi-steady stage, the ceiling-temperature rise exhibited exponential decay along the tunnel. Relative to the centerline fire condition, displacement of the fire source to the center of the right lane increased the magnitude of the downstream composite attenuation coefficient by 5.6–11.2%; a further displacement toward the sidewall increased it by 20.6–34.1%. Most predictions from the proposed longitudinal temperature model agreed with the comparison data within ±20%. These findings underscore the need to jointly account for transverse and longitudinal fire-source locations in smoke-control design and evacuation-safety assessment for arched road tunnels. Full article

Self-contained breathing apparatus (SCBA) is essential for firefighter safety but contributes substantially to equipment weight and can restrict mobility and task performance. In order to improve operational performance in firefighting tasks, our research consortium developed a prototype SCBA incorporating weight reduction and ergonomic redesign. For field applicability assessment, this study aimed to quantify the impact of the prototype SCBA on firefighter mobility relative to a widely deployed conventional SCBA. Twelve firefighters conducted range of motion (ROM) tests and standardized scenario-based firefighting performance tests while wearing both SCBA configurations. Objective measures included ROM and completion time for scenario-based firefighting tasks. Subjective ratings of satisfaction for each SCBA were collected after each objective test using Likert-scale questionnaires. The results showed that use of the prototype SCBA increased overall ROM by approximately 5.6%. The average completion time for the scenario-based firefighting performance assessment decreased from 221.82 s with the conventional SCBA to 182.91 s with the prototype SCBA, corresponding to a reduction of approximately 17.5%. User satisfaction ratings consistently favored the prototype SCBA, particularly in perceived weight, fit, perceived pressure, and mobility. The consistency between objective results and user perceptions suggests that the lightweight and ergonomic improvements in SCBA can produce meaningful functional benefits during firefighting tasks. These findings provide evidence to guide the development of lighter and better-fitting SCBA. Full article

Wood dust poses fire and explosion hazards in wood-processing and biomass-handling facilities, but its behavior varies with material properties and operating conditions. This review develops a framework linking particle-scale conversion processes with industrial accident pathways and protection strategies. It distinguishes stand-alone smoldering and flaming fires, direct dust-cloud ignition, flash fires, confined deflagrations, and secondary explosions. Lignocellulosic pyrolysis, volatile combustion, and char oxidation are examined in relation to ignition delay, flame propagation, and pressure development. Ignition and explosion parameters, including minimum ignition energy, minimum ignition temperature, minimum explosible concentration, limiting oxygen concentration, maximum explosion pressure, and deflagration index, are evaluated alongside the effects of wood composition, particle-size distribution, moisture, turbulence, and confinement. For example, increasing the tested Pinus massoniana dust particle-size fraction from 0–63 to 250–500 μm reduced the maximum explosion pressure (P_max) from 7.56 to 6.80 bar and the deflagration index (K_st) from 129 to 59.2 bar·m·s⁻¹, illustrating the greater particle-size sensitivity of K_st in that study. Such results remain material- and test-dependent and do not establish universal wood-dust design values. Prevention and mitigation measures are assessed through risk-based selection of source control, inerting, venting, isolation, and suppression. Computational fluid dynamics applications are critically examined in terms of model assumptions, validation, and engineering applicability. Throughout the review, wood-dust-specific findings are distinguished from woody-biomass and other analog evidence. Major gaps remain in quantitative links between composition and explosion behavior, transitions between fire and explosion pathways, and industrial-scale validation of suppression and numerical models. Addressing these gaps would strengthen scenario-specific protection of dust collectors, conveying ducts, dryers, and silos. Full article

Remote sensing is widely used to assess wildfire effects, with established approaches providing continuous estimates of fire severity from spectral information. This study developed a canopy-condition classification framework that maps post-fire forest canopies into three ecologically interpretable classes: green, scorched, and burned, using Landsat imagery, the Tasseled Cap Transformation, and a Random Forest classifier. Training sample points were obtained through visual interpretation of post-fire Google Earth imagery from eleven wildfire events in Portugal. Model performance was evaluated using stratified group K-fold cross-validation and an independent validation dataset comprising eight additional wildfire events. The classifier achieved an overall cross-validation accuracy of 0.906 and an independent validation accuracy of 0.93, demonstrating good generalization across independent wildfire events. Among the Tasseled Cap components, Greenness was the most important predictor for discriminating canopy condition. Rather than introducing another fire severity metric, the proposed framework provides a straightforward description of the dominant post-fire canopy condition that complements existing fire severity products. The resulting maps offer an additional perspective on canopy fire effects and can support post-fire assessment, monitoring, and management in Mediterranean forest ecosystems. Full article

The interactions between multiple-pool fires can lead to significantly higher burning rates and flame heights compared to single-pool fires, increasing the likelihood of a domino effect and greater environmental impact. This study investigates the burning characteristics of multiple-pool fires relative to isolated fires, focusing on the interactions between adjacent fires and the effects of height differences. Using isooctane as the fuel, experiments were conducted with pools of varying sizes and heights. The results demonstrated that burning rates and flame heights exhibit nonlinear variations, while flame temperatures exhibit linear changes. Flame merging was most pronounced at a pool height of H = 10 cm. Based on the behavior of elevated pool fires, an energy analysis was performed to assess their burning rate. A model was developed to express the burning rate as a function of pool size, height, and the burning rate and flame height of single-pool fires. The proposed equation was validated against experimental data and literature, demonstrating reasonable reliability and consistency. Full article

Accurate and timely recognition of flame and smoke is critical for fire warning systems to mitigate casualties and property damage. Most existing fire datasets are image-based and thus fail to capture the dynamic spatiotemporal information of fire. Furthermore, the lack of large-scale fire video datasets poses a significant challenge to the training of neural networks. To address these limitations, we developed the Flame-Smoke Video Recognition (FSVR) dataset, a large-scale collection of 30,000 video clips that significantly exceeds the scale of prior datasets in the same domain. Since flame and smoke exhibit multiscale characteristics, small-scale targets are often obscured by complex backgrounds. Existing methods lack robust multiscale feature learning capabilities, which limits their ability to achieve precise fire recognition under complex conditions. To address this issue, we proposed the Multi-scale Fire Feature Learning Network (MFFLNet), which integrates multiple Multi-scale Fire Feature Learning (MFFL) blocks into a Transformer backbone. Each MFFL block comprises two key components, i.e., the multiscale fire Conv3D layer and the fire spatiotemporal feature learning layer. The experimental results obtained from the FSVR and LFVR datasets demonstrate that MFFLNet surpasses the baseline model and other comparative methods. When the backbone network is initialized with pre-trained weights from the Kinetics-710 dataset, MFFLNet attained an accuracy of 79.54% and a macro-F1 score of 79.22% on the FSVR dataset, while achieving an accuracy of 95.64% and an F1 score of 95.46% on the LFVR dataset. Full article

Fast-growing poplar wood is an attractive resource for sustainable construction, but its high flammability and limited durability restrict its use in fire-rated applications. This study evaluates a bio-based flame-retardant treatment combining tannic acid (TA) and diammonium phosphate (DAP) impregnation with a subsequent gas-phase ammonia (NH₃) fuming step intended to fix the flame-retardant components within the wood matrix. The combined TA+DAP treatment increased the limiting oxygen index from 23% (untreated) to 77% and reduced peak heat release rate by roughly 4.2x and total heat release by roughly 2.5x, consistent with a stable char layer confirmed by thermogravimetric analysis. Adding tannic acid within this fully fumed system also markedly improved fixation relative to DAP alone, shifting leach retention from a net mass loss (–13.0%) to a positive retention of 28.9%, while modestly increasing hardness and flexural stiffness. Although this leach retention remains below that of covalently fixed systems, and the fire performance and mechanical properties of leached specimens were not directly retested in this study, the results indicate that the combined TA+DAP treatment with ammonia fuming offers a simple, scalable route to enhancing the fire performance and reducing the equilibrium moisture content of low-density plantation wood; both properties were evaluated on unleached specimens only, and whether the fire-performance benefit is retained after leaching remains unconfirmed. Full article

Wildfire smoke contains complex mixtures of inorganic and organic species, yet metals in experimental smoke preparations are rarely quantified in a manner directly useful for toxicology. Here, an inductively coupled plasma–mass spectrometry (ICP-MS) workflow for multi-element analysis of bushfire smoke preparations was explored for use in future cell-based studies. The novel contribution is the integration of matrix-matched blank correction, near-LOD handling, micromolar unit conversion and multivariate visualisation into a reproducible workflow for preparation-specific smoke exposure characterisation. This study is an analytical-methods contribution that provides composition-based exposure-characterisation inputs for future toxicological and risk-assessment studies, rather than a direct assessment of biological toxicity or health risk. Eight case-study preparations were analysed: two laboratory-generated bushfire smoke stock solutions (BF#1, BF#2); four laboratory bushfire smoke filter extracts with dominant vegetation types of Eucalyptus, Pine, Banksia, and Mallee; and paired blue gum smoke chamber T- and G-phase samples, each with matched blanks. The workflow combined blank mapping and subtraction, handling of values near the limit of detection, and conversion of concentrations to blank-corrected micromolar units in analytical extracts. Major ions (Na, K, Ca, Mg, P, S) dominated the metal burden, redox-active metals (Al, Mn, Fe, Cu, Zn) were present at sub- to low-micromolar levels, and toxic metals (Pb, Cd) were generally low, with Pb elevated in one filter extract. Silicon was negligible in stock and chamber samples but high in filter extracts, consistent with quartz filter contributions. Heatmaps, burden indices and principal component analysis provided illustrative views of between-sample differences. This workflow enables high-quality single-sample metal characterization and supports more transparent interpretation of wildfire smoke toxicology studies. Full article

To address the compatibility challenges in selecting gas fire suppression systems for three typical complex protected environments—data centers, museum artifact storage rooms, and aircraft cargo holds—this paper establishes a multi-level evaluation framework comprising 25 third-tier indicators across five dimensions: protected object characteristics, fire suppression effectiveness, safety performance, environmental impact, and economic cost. Instead of computing weighted scores and producing a numerical ranking, this study determines indicator importance through a rule-based qualitative procedure, classifying each indicator into four importance grades—critically important, important, generally important, and not important—according to standards, safety consequences, and engineering applicability. The importance of each indicator is differentially classified across the three scenarios, and the performance differences among four mainstream gas fire suppression systems (perfluoroketone-based, HFC-227ea, IG541, and CO₂) are systematically analyzed under these criteria. The results reveal three typical patterns of indicator importance differentiation: scenario-specific dominance, high cross-scenario consensus, and gradient differentiation. The sets of critically important indicators vary significantly among scenarios: data centers prioritize equipment safety, artifact storage rooms emphasize both personnel and cultural heritage safety, and aircraft cargo holds focus primarily on structural safety. This “qualitative” indicator system provides a structured evaluation framework for the differentiated selection of gas fire suppression systems across diverse application scenarios, rather than yielding a composite score or a quantitative ranking of candidate systems. Full article

Active management making forests more resistant to wildfire and more resilient afterwards is a widely held assumption that is rarely tested against comparable fires. In March 2023, two wildfires ignited three days and 54 km apart under the same synoptic weather in southern Korea, one in a naturally regenerating forest of Jirisan National Park and one in a pine-dominated production forest. Treating the pair as a natural experiment, we compared burn severity and three-year recovery using Sentinel-2 imagery, hourly ERA5-Land weather, the 1:5000 national stand map and a composition-weighted unburned reference. Both fires burned under light mean winds over nearly equal areas (117 and 122 ha), yet high severity covered 0.4% of the park fire and 11.4% of the managed-forest fire. Broadleaved stands burned alike at the two sites. Pine stands burned far more severely in the managed forest (mean differenced Normalised Burn Ratio 0.37) than in the park (0.25), and pine made up 57% of the managed-forest burn, whereas conifers, mostly pine, made up only 14% of the park burn. The park’s steeper and more sun-exposed terrain would, if anything, have favoured more severe fire. The 2025 stand map reclassified 98% of the managed forest’s pine but classified 19% of its broadleaved stands as unstocked, consistent with the post-fire clearing of pine. Three years later, the park had regained at least 80% of its pre-fire levels for all three spectral indices, whereas the managed forest remained below 80% for NBR and NDMI. Broadleaved and mixed stands showed similar spectral recovery at both sites; pine showed slower spectral recovery at both sites, and the managed forest’s lightly burned pine regained about half the share of its loss observed in the park’s pine, which remained classified as stocked forest, although the intervals of the two estimates overlapped. The managed regime was associated with greater burn severity and slower spectral recovery; the latter coincided with its pine-dominated composition and the post-fire reclassification of most pine stands as unstocked. Full article

Coal spontaneous combustion fire is a major hazard threatening mine safety and production. Accurate identification of hidden fire sources is the key to preventing and controlling coal spontaneous combustion fires. Given that traditional detection methods are unable to accurately identify high-temperature zones, this study develops an acoustic-signal-based approach for reconstructing the temperature field associated with coal spontaneous combustion. By constructing a two-dimensional acoustic thermometry experimental system for temperature field measurement, multipath acoustic signals were collected in both the air medium and a loose coal medium, respectively. The variation characteristics of the acoustic spectrum, energy attenuation, and travel time were analyzed under different heat source temperatures, media, and sound source positions. The results show that the dominant frequencies of the received signals under different heat source temperatures are concentrated in the range of 1000–1300 Hz. An increase in temperature does not significantly alter the dominant frequency range, but it reduces the peak amplitude of the received signals. Compared with propagation in air, acoustic waves experience greater energy loss in loose coal. Their travel time is also generally longer, indicating that the pore structure alters the propagation path and influences the acoustic propagation process. To account for pore-structure-induced deviations between the actual acoustic propagation path and the geometric path, a path correction coefficient is introduced. Using the compensated correction model, the temperature fields under working conditions of 100 °C, 150 °C, and 200 °C were reconstructed, with peak temperatures of 91.507 °C, 140.554 °C, and 188.665 °C, respectively. The gap between the reconstructed results and those obtained in the air medium is significantly narrowed. This suggests the model effectively corrects path deviations due to pore structure and enhances acoustic reconstruction accuracy for non-uniform temperature fields in loose coal. These findings offer a methodological basis for acoustic temperature field reconstruction and hidden fire source detection in loose coal, with practical value for enhancing mine fire prevention efficiency. Full article

External heat flux is a primary thermal input parameter in cone calorimeter tests, with the nominal value conventionally established at the center of the specimen surface. However, the radiant heat flux generated by the cone heater specified in ISO 5660-1 is known to be spatially non-uniform. This study quantified the spatial distribution of external heat flux at nominal levels of 20, 35, and 50 kW/m² and applied area-weighted spatial averaging to quantify the relationship between the nominal heat flux and the spatially averaged external heat flux over the specimen. The measured heat flux decreased radially toward the specimen boundaries, with local relative deviations reaching approximately −30% near the corners of a standard 0.100 × 0.100 m specimen. Despite this local non-uniformity, the normalized spatial distributions were similar across all three heat flux levels. For the entire specimen, the area-weighted relative deviations were −3.54%, −3.28%, and −3.41% at 20, 35, and 50 kW/m², respectively, with comparable deviations in total external radiant power. These findings demonstrate that the magnitude of local heat-flux deviation should be distinguished from the specimen-wide spatially averaged deviation. The proposed area-weighted spatial interpretation provides a quantitative basis for relating the nominal external heat flux established at the specimen center to the spatially averaged external heat flux over the specimen under the ISO 5660-1 cone calorimeter configuration. Full article

High-density commercial pedestrian streets are open urban spaces where commercial activities, public interactions, and everyday pedestrian movements are highly concentrated. However, they are also prone to crowd aggregation and local congestion during emergency evacuation. In mountainous cities, compact street spaces and terrain constraints may further amplify evacuation risks, while quantitative guidance for evacuation-oriented street-facility layout remains insufficient. To address this gap, this study investigates the effects of typical facility layouts on fire evacuation in high-density commercial pedestrian streets of mountainous cities. Using Chongqing, China, as a representative case, an AnyLogic agent-based evacuation model was developed for a 30 m by 90 m pedestrian-street segment with an initial density of 1.0 persons/m². Archived surveys of 15 commercial pedestrian streets were used to identify recurring facility categories, geometric characteristics, and layout patterns rather than to reconstruct the surveyed streets as separate simulation geometries. Routine pedestrian-flow observations from two counting sections in the Three Gorges Square commercial center were used only for a qualitative ordinary-flow consistency assessment; they did not provide independent validation of emergency behavior under fire conditions. The simulations compared four clear edge-to-edge separation-barrier spacings, two archived advertising-board configurations, and a series of tree-pit-bench obstruction envelopes within an idealized 30 m × 90 m level corridor under simplified fire-related speed-reduction assumptions. Across the four tested barrier spacings of 800, 1100, 1400, and 1800 mm, the 95th-percentile evacuation time generally decreased as spacing increased. However, because only four spacing values were examined and 1800 mm was the only tested value above 1400 mm, the results do not establish an optimum spacing range or a response plateau. Differences were also observed between the archived advertising-board configurations and among the tested tree-pit-bench scenarios. The advertising-board comparison cannot be interpreted as a validated physical orientation effect because the original plan-view geometry could not be fully reconstructed. Similarly, the comparison between square and circular tree-pit-bench envelopes confounded boundary shape with occupied area and therefore does not demonstrate a pure shape effect. Most scenarios contained six simulation runs, and smoke, heat, visibility, toxic exposure, and mountainous terrain were not dynamically simulated. Accordingly, the findings are exploratory, model-specific scenario-comparison results and should not be interpreted as validated fire-safety thresholds or directly transferable engineering design requirements. Full article

As global urbanization transitions toward intensive underground space utilization, utility-tunnels have emerged as indispensable “urban lifelines.” Among their various compartments, the power compartment represents the most critical yet vulnerable zone, characterized by high-density fire loads and extreme confined boundary constraints. This review summarizes and critically discusses current research on the state-of-the-art in utility-tunnel fire safety, establishing a holistic framework that bridges fundamental physical mechanisms with advanced engineering applications. The review focuses on the fire dynamics specific to power compartments, including cable aging-induced ignition, multidirectional flame spread, heat release, and smoke propagation. Subsequently, the study critically evaluates the limitations of current fire codes and identifies the methodological gaps in reduced-scale experiments and CFD simulations when addressing complex heterogeneous topologies. As for fire safety protection, highlight transformative frontiers in smart prevention, including nano-enhanced intrinsically safe materials, AI-driven digital-twin frameworks for real-time situational awareness, and synergistic “gas–liquid–solid” extinguishing systems (e.g., liquid-nitrogen and fine water mist) were summarized and analyzed. Finally, a strategic roadmap for future research is proposed, advocating for a shift from “straight-segment” analysis to “network-wide” resilience assessment. This synthesis aims to guide the development of next-generation intelligent fire-protection systems for resilient underground urban infrastructures. Full article