Fire

Wildfires are a critical global threat, emphasizing the need for efficient detection systems capable of identifying fires and distinguishing fire-related from non-fire events in their early stages. This study integrates the swintransformer into the Faster R-CNN backbone to overcome challenges in detecting small flames and smoke and distinguishing complex scenarios like fog/haze and chimney smoke. The proposed model was evaluated using a dataset comprising five classes: flames, smoke, clouds, fog/haze, and chimney smoke. Experimental results demonstrate that swintransformer-based models outperform ResNet-based Faster R-CNN models, achieving a maximum mAP50 of 0.841 with the swintransformer-based model. The model exhibited superior performance in detecting small and dynamic objects while reducing misclassification rates between similar classes, such as smoke and chimney smoke. Precision–recall analysis further validated the model’s robustness across diverse scenarios. However, slightly lower recall for specific classes and a lower FPS compared to ResNet models suggest a need for further optimization for real-time applications. This study highlights the swintransformer’s potential to enhance wildfire detection systems by addressing fire and non-fire events effectively. Future research will focus on optimizing its real-time performance and improving its recall for challenging scenarios, thereby contributing to the development of robust and reliable wildfire detection systems. Full article

Portugal’s increasing wildfire frequency has led to home destruction, large areas burned, ecological damage, and economic loss, emphasizing the need for effective fire exposure assessments. This study builds on a Canadian approach to wildfire exposure and evaluates wildfire exposure in the Portuguese municipality of Mafra, using artificial territories (AT) as a proxy for the wildland–urban interface (WUI) and integrates land use land cover (LULC) data with a neighborhood analysis to map exposure at the municipal scale. Fire exposure was assessed for three fire transmission distances: radiant heat (RH, <30 m), short-range spotting (SRS, <100 m), and longer-range spotting (LRS, 100–500 m) using fine resolution (5 m) LULC data. Results revealed that while AT generally exhibited lower exposure (<16% “very high” exposure), adjacent hazardous LULC subtypes significantly increase wildfire hazard, with up to 51% of LULC subtypes classified as “very high exposure”. Field validation confirmed the accuracy of exposure maps, supporting their use in wildfire risk reduction strategies. This cost-effective, scalable approach offers actionable insights for forest and land managers, civil protection agencies, and policymakers, aiding in fuel management prioritization, community preparedness, and the design of evacuation planning. The methodology is adaptable to other fire-prone regions, particularly mediterranean landscapes. Full article

Live fuel moisture content (LFMC) significantly influences fire activity and behavior over different spatial and temporal scales. The ability to estimate LFMC is important to improve our capability to predict when and where large wildfires may occur. Currently, there is a gap in providing reliable near-real-time LFMC estimates which can contribute to better operational decision-making. The objective of this work was to develop near-real-time LFMC estimates for operational purposes in Portugal. We modelled LFMC using Random Forests for Portugal using a large set of potential predictor variables. We validated the model and analyzed the relationships between estimated LFMC and both fire size and behavior. The model predicted LFMC with an R² of 0.78 and an RMSE of 12.82%, and combined six variables: drought code, day-of-year and satellite vegetation indices. The model predicted well the temporal LFMC variability across most of the sampling sites. A clear relationship between LFMC and fire size was observed: 98% of the wildfires larger than 500 ha occurred with LFMC lower than 100%. Further analysis showed that 90% of these wildfires occurred for dead and live fuel moisture content lower than 10% and 100%, respectively. Fast-spreading wildfires were coincident with lower LFMC conditions: 92% of fires with rate of spread larger than 1000 m/h occurred with LFMC lower than 100%. The availability of spatial and temporal LFMC information for Portugal will be relevant for better fire management decision-making and will allow a better understanding of the drivers of large wildfires. Full article

Urban fire incidents pose significant risks to public safety and infrastructure, necessitating precise spatiotemporal prediction to enhance fire prevention and emergency response strategies. However, predicting fire occurrences remains a complex task due to the intricate interplay between spatial and temporal factors, including dynamic environmental conditions, historical dependencies, and inter-regional correlations. Temporal variables, such as past fire incidents and external influences like meteorological conditions, significantly impact fire risk, while spatial attributes, including regional characteristics and cross-regional interactions, further complicate predictive modeling. This study introduces UFSTP, an innovative framework for Urban Fire Spatial–Temporal Prediction that integrates multi-source data for enhanced predictive accuracy. UFSTP employs a neural region state representation to capture both intrinsic and extrinsic temporal dependencies, alongside a spatiotemporal propagation mechanism to model inter-regional correlations. Key environmental and historical features are extracted from real-world datasets to construct a comprehensive fire risk representation, facilitating the precise forecasting of fire occurrence in both time and space. Extensive evaluations on real-world datasets from Anci and Guangyang Districts demonstrate that UFSTP achieves a 16.2% average reduction in Mean Absolute Error (MAE) for time prediction and a 3.3% average improvement in top-1 hit rate for regional prediction over state-of-the-art baselines. The proposed framework offers a robust and interpretable approach to urban fire risk assessment, providing critical insights to optimize fire prevention measures and emergency resource allocation. Full article

Fire is a combustion reaction where fuel reacts with oxygen in the presence of heat, releasing energy as light, heat, and flames. The main components of fire are fuel, oxygen, and heat. All three components must be present to cause a fire. Fire is a significant threat to residential and commercial buildings, often intensified by high fuel content in building materials such as wood and synthetics. This paper summarizes fire types and damages, loss of property and life, fuel content in building materials, and a method to reduce fire risk by minimizing the building material’s fuel content. This method uses minerals (coal combustion residual (CCR)), primarily inorganic oxides bonded with a small percentage of polyurethane binder, to manufacture a composite material moldable into multiple building products. The composite was tested as per the ASTM for mechanical, thermal, and fire safety performance. ASTM D635-based fire testing showed self-extinguishing behavior with significantly reduced burn rate and lengths (1–2 mm). A low calorific value of 6.6 MJ/kg was determined separately. The test results demonstrate that CCR-based mineral composites offer a fire-resistant, structurally sound, and eco-friendly alternative to wood products. This research supports recycling inorganic minerals into fire-resistant building products that enhance safety. Full article

This study addresses the problems of multi-source data redundancy, insufficient feature capture timing, and delayed risk warning in the prediction of gas concentration in fully mechanized coal-mining operations by constructing a three-pronged technical approach that integrates feature dimensionality reduction, hybrid modeling, and intelligent early warning. First, sparse kernel principal component analysis (SKPCA) is used to accomplish the feature decoupling of multi-source monitoring data, and its optimal dimensionality reduction performance is verified using long-term and short-term neural networks (LSTMs). Second, an innovative TCN–Transformer hybrid architecture is proposed. The transient fluctuation characteristics of gas concentration are captured using causal dilation convolution, while a multi-head self-attention mechanism is used to analyze the cross-scale correlation of geological mining parameters. A flood optimization algorithm (FLA) is used to establish a hyperparameter collaborative optimization framework. Compared to TCN-LSTM, CNN-GRU, and other hybrid models, the hybrid model proposed in this study exhibits superior point prediction performance, with a maximum R² of 0.980988. Finally, a dynamic confidence interval is established using the locally weighted kernel density estimation (LWD-KDE) method with an optimized bandwidth, and an unsupervised early warning mechanism for the risk of gas concentration fluctuations in coal mines is constructed. The results provide a comprehensive approach to preventing and controlling gas disasters in fully mechanized mining operations. This research effectively promotes the transformation and upgrading of coal-mine-safety-monitoring systems to an active defense paradigm. Full article

Obstacles “upstream” of the exit significantly impact evacuation efficiency and deserve attention. Based on the discrete cellular automaton model, this paper studies the impact of different obstacle settings on evacuation efficiency in different emergency levels under different exit loads. Through simulation, we found that at low emergency levels, the appearance of obstacles has little impact on evacuation efficiency, while at high emergency levels, the changes in evacuation efficiency vary greatly under different obstacle settings: when the exit is relatively wide (evacuation pressure is low) and has the “faster is faster” effect, obstacles upstream of the exit reduce the evacuation efficiency, and setting obstacles directly opposite of the safety exit has the most obvious impact on the evacuation efficiency; while when the exit is narrow (evacuation pressure is high) and has the “faster is slower” effect, appropriately setting obstacles can slightly improve the evacuation efficiency. Our findings help to understand the impact of obstacles on evacuation efficiency under different exit loads to set upstream obstacles reasonably. Full article

In the high-temperature cooking process of Chinese-style catering, the oil fume accumulates on the inner wall of the flue during the cooling process, forming grease stains, which can easily trigger flue fires and cause a kitchen fire. Statistics indicate flue fires are a primary cause of kitchen fires in China. The changes in the composition of grease stains are due to different freezing points, which will adhere to different parts of the flue and be repeatedly heated and cooled if not cleaned in time. This leads to changes in combustion performance, subsequently affecting the progression of flue fire propagation. This paper takes grease deposits with different deposition times and locations in the flue of commercial kitchens as the research object. The research selected a medium-sized commercial kitchen flue (kitchen chimney) in Langfang City, with deposition times of the parts of the inlet and outlet for 2 months and grease in the inlet for a deposition time of 7 days, 60 days, and more than 1 year. This paper analyzed the grease deposits at different deposition positions at the flue inlet and outlet using a thermogravimetric analyzer and a gas-mass spectrometer. It is found that the primary components of the grease at the outlet have low molecular weight, thermal decomposition starting temperature ignition temperature, and activation energy in the first stage and will catch fire first; the grease at the inlet has a high comprehensive combustion performance, and the combustion is violent with little effect from the oxygen supply. Then, the pyrolysis analysis of grease stains located at the entrance of the flue is performed at different deposition times under air and nitrogen atmosphere. The results showed that the pyrolysis process of grease stains with a more than 1 year deposition time consists of two stages. One stage is the first weightlessness stage, which has the lowest activation energy, the longest combustion process, and the greatest fire risk; the other is the pyrolysis combustion process of grease stains with a deposition time of 7 days. Its activation energy is the highest, and the fire risk is the smallest. The research results can be a reference for the setting of the fire dampers and the cleaning time for the flue. Full article

Forest fires are a critical ecological disturbance that significantly impact vegetation dynamics, biodiversity, and ecosystem services. This study investigates the impacts of forest fires in the Ghalahmah Mountains, Saudi Arabia, using remote sensing data and fire impact models to assess fire severity, environmental drivers, and post-fire vegetation recovery. The research integrates Landsat 8, Sentinel-2, and DEM data to analyze the spatial extent and severity of a 2020 fire event using the Relativized Burn Ratio (RBR). Results reveal that high-severity burns covered 49.9% of the affected area, with pre-fire vegetation density (NDVI) and moisture (NDWI) identified as key drivers of fire severity through correlation analysis and Random Forest regression. Post-fire vegetation recovery, assessed using NDVI trends from 2021 to 2024, demonstrated varying recovery rates across vegetation types. Medium NDVI areas (0.2–0.3) recovered fastest, with 134.46 hectares exceeding pre-fire conditions by 2024, while high NDVI areas (>0.3) exhibited slower recovery, with 26.55 hectares still recovering. These findings underscore the resilience of grasslands and shrubs compared to dense woody vegetation, which remains vulnerable to high-severity fires. The study advances fire ecology research by combining multi-source remote sensing data and machine learning techniques to provide a comprehensive understanding of fire impacts and recovery processes in semi-arid mountainous regions. The results suggest valuable insights for sustainable land management and conservation, emphasizing the need for targeted fuel management and protection of ecologically sensitive areas. This research contributes to the broader understanding of fire ecology and supports efforts to post-fire management. Full article

Firefighting foam is widely recognized for its excellent fire suppression performance. However, research on the effect of foam expansion ratio on the suppression efficiency of forest surface fires remains limited. In this study, the expansion ratio was adjusted by varying the air-to-liquid ratio in a compressed air foam system, and laboratory-scale foam suppression experiments were conducted. Key performance indicators, including extinguishing coverage time, internal cooling rate, and resistance to reignition, were systematically measured. The effects of expansion ratio on the diffusion and penetration behavior of foam on the fuel bed surface were then investigated to understand how these characteristics influence suppression performance. The results indicate that both excessively low and high expansion ratios can weaken fire suppression effectiveness. Low-expansion foam, characterized by low viscosity and high water content, exhibits strong local penetration and cooling capabilities. However, it struggles to rapidly cover the fuel bed surface and isolate oxygen, thereby reducing the overall suppression efficiency. In contrast, high-expansion foam has greater viscosity, allowing it to spread across the fuel bed surface under pressure gradient forces and form a stable coverage layer, effectively limiting the oxygen supply required for combustion. However, its limited depth penetration and lower water content reduce internal cooling efficiency, increasing the risk of reignition. The optimal expansion ratio was determined to be 15.1. Additionally, increasing the liquid supply flow rate significantly improved suppression performance; however, this improvement plateaued when the flow rate exceeded 10 L/min. Full article

Smoke detection is crucial for early fire prevention and the protection of lives and property. Unlike generic object detection, smoke detection faces unique challenges due to smoke’s semitransparent, fluid nature, which often leads to false positives in complex backgrounds and missed detections—particularly around smoke edges and small targets. Moreover, high computational overhead further restricts real-world deployment. To tackle these issues, we propose RT-DETR-Smoke, a specialized real-time transformer-based smoke-detection framework. First, we designed a high-efficiency hybrid encoder that combines convolutional and Transformer features, thus reducing computational cost while preserving crucial smoke details. We then incorporated an uncertainty-minimization strategy to dynamically select the most confident detection queries, further improving detection accuracy in challenging scenarios. Next, to alleviate the common issue of blurred or incomplete smoke boundaries, we introduced a coordinate attention mechanism, which enhances spatial-feature fusion and refines smoke-edge localization. Finally, we propose the WShapeIoU loss function to accelerate model convergence and boost the precision of the bounding-box regression for multiscale smoke targets under diverse environmental conditions. As evaluated on our custom smoke dataset, RT-DETR-Smoke achieves a remarkable 87.75% mAP@0.5 and processes images at 445.50 FPS, significantly outperforming existing methods in both accuracy and speed. These results underscore the potential of RT-DETR-Smoke for practical deployment in early fire-warning and smoke-monitoring systems. Full article

As an emerging development field, in recent years, emergency industrial parks in China have faced increasingly complex and high-risk challenges. This article proposes the establishment of a scientific safety risk assessment and grading model to help improve the safety management level of emergency industrial parks, in response to the problems of the multi-source heterogeneity of fire risks in emergency industrial parks and the difficulty of comprehensive assessment using traditional methods. This approach combines enterprise type classification with multi-level assessment for the first time, effectively identifying high-risk links such as fires and explosions and playing an effective role in preventing accidents such as fires in the park. Enterprises within the park are categorized into seven distinct groups based on their characteristics and associated safety risks: medical and healthcare, new energy storage, composite materials and new materials, intelligent manufacturing, mechanical manufacturing, consulting and technical services, and construction and installation. The following models are constructed: (1) a risk assessment model based on AHP-FCE, which can assess the safety risk levels of individual enterprises and the industrial park at a macro level; (2) a risk grading model based on the risk matrix method, which can inspect and control specific risk sources at a micro level. The integration of these two methods establishes a comprehensive model for safety risk assessment and grading in emergency industrial parks, significantly improving both the accuracy and the systematic nature of risk management processes. Full article

Both single-skin and double-skin glass facades are extensively employed in commercial high-rise buildings and are gaining increasing popularity. However, the capability to deploy firefighting agents in such ultra-high structures remains limited and has been minimally investigated. To provide guidance for single-layer exterior wall fire protection, this study examines the impact of vertical walls on window ejected plumes by simulating the upper portion of jet plumes using a square burner flame. Experimental and numerical simulations were conducted. The findings revealed that plumes from propane burners could attach to the wall even when L_{E. burner fire} > 0.7W, contradicting previous criteria. This discrepancy arises because prior studies underestimated the induced pressure difference in large fires. This pressure difference propels the plume toward the wall, behaving like a rigid body. Full article

Wildfire activity in the western United States has been on the rise since the mid-1980s, with longer, higher-risk fire seasons projected for the future. Prescribed burning mitigates the risk of extreme wildfire events, but such treatments are currently underutilized. Fire managers have cited lack of firefighter availability as a key barrier to prescribed burning. We use both principal component analysis (PCA) and logistic regression modeling methodologies to investigate whether or not (and if yes, under what conditions) personnel shortages on a given day are associated with lower odds of a prescribed burn occurring in the Okanogan–Wenatchee National Forest. We utilize the logit model to further assess how personnel availability compares to other potential barriers (e.g., meteorological conditions) in terms of association with odds of a prescribed burn occurring. Our analysis finds that fall and spring days in general have distinct constellations of characteristics. Unavailability of personnel is associated with lower odds of prescribed burning in the fall season, controlling for meteorological conditions. However, in the spring, only fuel moisture is observed to be associated with the odds of prescribed burning. Our findings suggest that if agencies aim to increase prescribed burning to mitigate wildfire risk, workforce decisions should prioritize firefighter availability in the fall. Full article

With increasing wildfire severity and duration driven by climate change, accurately predicting fire behavior over extended time frames is critical for effective management and mitigation of such wildfires. Fire propagation models play a pivotal role in these efforts, providing simulations that can be used to strategize and respond to active fires. This study examines the fire area simulator (FARSITE) model’s performance in simulating recent wildfire events that persisted over 24 h with limited firefighting intervention in mostly remote access areas across diverse ecosystems. Our findings reveal key insights into a prolonged wildfire scenarios potentially informing improvements in operational fire management and long-term predictive accuracy, as the area comparisons indexes showed reasonable results between the detected fires from the fire information for resource management systems (FIRMSs) in the first 24 h of the fire and the following days. A case study of a recent wildfire in Madeira Island highlights the integration of real-time weather predictions and post-event weather data analysis. This analysis underscores the potential of combining accurate forecasts with retrospective validation to improve predictive capabilities in dynamic fire environments, which guided the development of a software platform designed to analyse ongoing wildfire events in real-time, leveraging image satellite data and weather predictions. Full article

Fires pose significant threats to human safety, health, and property. Traditional methods, with their inefficient use of features, struggle to meet the demands of fire detection. You Only Look Once (YOLO), as an efficient deep learning object detection framework, can rapidly locate and identify fire and smoke objects in visual images. However, research utilizing the latest YOLO11 for fire and smoke detection remains sparse, and addressing the scale variability of fire and smoke objects as well as the practicality of detection models continues to be a research focus. This study first compares YOLO11 with classic models in the YOLO series to analyze its advantages in fire and smoke detection tasks. Then, to tackle the challenges of scale variability and model practicality, we propose a Multi-Scale Convolutional Attention (MSCA) mechanism, integrating it into YOLO11 to create YOLO11s-MSCA. Experimental results show that YOLO11 outperforms other YOLO models by balancing accuracy, speed, and practicality. The YOLO11s-MSCA model performs exceptionally well on the D-Fire dataset, improving overall detection accuracy by 2.6% and smoke recognition accuracy by 2.8%. The model demonstrates a stronger ability to identify small fire and smoke objects. Although challenges remain in handling occluded targets and complex backgrounds, the model exhibits strong robustness and generalization capabilities, maintaining efficient detection performance in complicated environments. Full article

In order to study the characteristics of the thermal runaway process of a full-size prefabricated cabin energy storage system, a full-scale prefabricated cabin energy storage physical fire test platform was designed using 100% SOC energy storage battery packs as the thermal runaway object, and full-scale prefabricated cabin energy storage system physical fire experiments were conducted. This experiment analyzes the early change rules of parameters such as temperature, voltage, CO, and VOC after the energy storage system enters thermal runaway and explores the technical methods to improve the fire protection of electrochemical energy storage systems. The results show that the time when the surface temperature of the runaway cell undergoes a sudden change is 37 s later than the time when the voltage undergoes a sudden change; the CO at the bottom and middle of the runaway cluster reaches the alarm threshold 25 s and 39 s earlier than that at the top of the cluster, respectively, and the peak concentration of CO at the bottom and middle of the cluster is more than three times that at the top of the cluster. The opening of the fan causes the CO concentration on the left side of the thermal runaway cluster to be higher than that of the runaway cluster; before the battery thermal runaway, the VOC concentration at the middle and top of the runaway cluster is generally higher than that at the bottom of the cluster. After thermal runaway occurs, the VOC concentration at the bottom of the thermal runaway cluster exceeds that at other positions of the runaway cluster and the adjacent cluster; the t_VOC at the top, middle, and bottom of the thermal runaway cluster is 2296 s, 1681 s, and 1464 s earlier than the t_CO, respectively, but the initial detection value of VOC fluctuates more than that of CO. Full article

Wildfires in the Mediterranean basin, particularly in Portugal, pose significant ecological challenges by altering landscapes and ecosystems. This study examines vegetation recovery in Serra do Açor seven years after the 2017 wildfires, using remote sensing and field data to analyze post-fire dynamics. The primary goal was to assess whether fire severity, measured via the dNBR index from Sentinel-2 imagery, impacts vegetation recovery or if site-specific factors and pre-fire floristic composition are more influential. Randomly assigned plots based on previous land use and fire severity were analyzed for floristic attributes. To quantify and classify cover changes, a supervised classification methodology based on the random forest algorithm was applied to Sentinel-2 data. The results showed no clear link between fire severity and recovery; instead, local factors like soil and topography, along with dominant pre-fire species, influenced recovery. Acacia and eucalyptus communities grew faster and increased the occupied area but exhibited lower diversity than native vegetation communities. Supervised classifications achieved high accuracy (Kappa > 0.90), showing increased shrubland areas and expansion of eucalyptus and acacia. The study highlights the methodology’s effectiveness and potential for broader applications in future research. Full article

Small-scale model experiments were conducted to examine the effectiveness of a point smoke exhaust system in an asymmetrical V-shaped tunnel. Using a design fire power of 20 MW as a reference, the study explored the optimal smoke exhaust rate for the point exhaust system in tunnels. Additionally, the impact of different slope combinations and various smoke vent opening configurations on the smoke control efficiency was analyzed when the fire source was positioned at the slope transition point of the V-shaped tunnel. The results indicate that an exhaust rate of approximately 140–160 m³/s is effective in controlling smoke for a 20 MW fire. In V-shaped tunnels, when three exhaust vents are symmetrically opened on both sides of the fire source, smoke diffusion on the large slope side remains uncontrolled. To address this issue, increasing the number of smoke exhaust vents on the large slope side can enhance smoke control. However, when the slope difference between the two sides of an asymmetrical V-shaped tunnel is excessively large—especially when the large side has a very high inclination—effective smoke control becomes significantly more challenging. Full article