安全科学与韧性(英文)最新文献

{"title":"A lightweight four-channel multi-modal model to improve computational performance of automated fire detection","authors":"Jingshuo Yu ,&nbsp;Qian Chen","doi":"10.1016/j.jnlssr.2025.100254","DOIUrl":"10.1016/j.jnlssr.2025.100254","url":null,"abstract":"<div><div>The urgent need for advanced fire detection methods stems from the increased intensity of fire incidents, which cause massive property loss and irreversible damage. To overcome the limitations of traditional fire detection methods, such as those of smoke detectors, fire detection based on computer vision (CV) algorithms has been adopted to improve detection accuracy. Compared to single-modal fire detection, multi-modal fire detection has gained attention because it leverages the richer information present in both RGB and thermal images. However, prevalent multi-modal fire detection methods significantly increase model complexity by requiring two separate streams in the backbone to process RGB and thermal images independently. To address this issue, this paper proposes a four-channel single-stream fire detection method based on YOLOv5, which concatenates RGB and thermal images to form the required four-channel input. Comparison experiments with dual-stream YOLOv5 models using add fusion and transformer fusion demonstrate that the four-channel single-stream model reduces model complexity while improving detection accuracy. To further enhance detection accuracy and reduce model complexity, this study redesigned YOLOv5’s C3 module by integrating the convolutional block attention module (CBAM) to form the C3CBAM module and introduced the SCYLLA-Intersection over Union (SIoU) loss function. By comparing its performance with that of state-of-the-art (SOTA) models in multi-modal object detection, such as the YOLOv5-based dual-stream model, this study shows that the proposed approach improves detection in the diverse conditions presented in the selected dataset.</div></div>","PeriodicalId":62710,"journal":{"name":"安全科学与韧性(英文)","volume":"7 2","pages":"Article 100254"},"PeriodicalIF":3.4,"publicationDate":"2026-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145736496","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Fusing emergency data and seasonal effects for urban rail transit safety assessment: A severity probability distribution model","authors":"Bosong Fan ,&nbsp;Dongyu Luo ,&nbsp;Chunfu Shao","doi":"10.1016/j.jnlssr.2025.100264","DOIUrl":"10.1016/j.jnlssr.2025.100264","url":null,"abstract":"<div><div>An analysis of Beijing urban rail transit emergency data from 2016 to 2018 is conducted to identify the factors influencing emergency severity. A severity probability distribution model is proposed, incorporating seasonal effects, multi-dimensional features (passenger, vehicle, line, environment), random parameters, interaction terms, and weather variables (temperature, humidity, wind speed) to capture severity probability changes driven by seasonally influenced factors. High performance is achieved by the model, with 92.10 % accuracy, 95.17 % sensitivity, and 87.58 % specificity. Five key factors—bad weather (correlation = -0.0611, <em>p</em> &lt; 0.001), track failure (correlation = -0.0503, <em>p</em> &lt; 0.001), signal failure (correlation = -0.0483, <em>p</em> &lt; 0.001), urban area (correlation = -0.0441, <em>p</em> &lt; 0.001), and vehicle failure (correlation = -0.0435, <em>p</em> &lt; 0.001)—are identified, with their interactions amplifying risks, particularly the coupling effect between bad weather and train system failures. Notable seasonal effects are observed: In summer, high temperature and humidity increase vehicle and equipment failures; in autumn, temperature fluctuations raise signaling anomaly risks; in winter, low temperatures elevate track and equipment failure risks; and in spring and winter, windy weather exacerbates risks for suburban elevated lines. This study highlights the need for refined risk identification and response mechanisms at the seasonal and regional levels, such as targeted inspections and weather-triggered controls, to enhance operational safety in urban rail transit under complex weather conditions.</div></div>","PeriodicalId":62710,"journal":{"name":"安全科学与韧性(英文)","volume":"7 2","pages":"Article 100264"},"PeriodicalIF":3.4,"publicationDate":"2026-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145839661","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Leveraging digital twin for COVID-19 risk assessment in IoT-enhanced hospital pharmacies","authors":"Divya Saraswathy Jayaprakash,&nbsp;Aryadevi Remanidevi Devidas,&nbsp;Maneesha Vinodini Ramesh","doi":"10.1016/j.jnlssr.2025.100246","DOIUrl":"10.1016/j.jnlssr.2025.100246","url":null,"abstract":"<div><div>Over the past three years, the Coronavirus Disease 2019 (COVID-19) pandemic has caused widespread human casualties, economic disruptions, and public health challenges. It is vital to assess the risk of contagious airborne diseases like COVID-19 to safeguard public health. Hospitals, as potential hotspots due to patient presence, require meticulous risk evaluation, particularly in critical areas such as pharmacies, which witness significant foot traffic for medicine collection. This paper presents a Digital Twin (DT) enabled (Internet of Things) IoT-based architecture comprising different enablers and their data interactions for COVID-19 Digital Twin-Risk Assessment in hospital pharmacy (DT-RAHP) to capture dynamic risk variations. The Decision Intelligence Module in the proposed architecture employs a novel two-phase risk assessment approach, DT-RAHP, which includes: Risk Assessment in Hospital Pharmacy-Phase I (RAHP-I) assesses the risk based on the primary risk factors, whereas the second phase, Risk Assessment in Hospital Pharmacy-Phase II (RAHP-II), extrapolates the risk assessment based on the secondary risk factors. Using DT, the system gathers data, assesses risk, and provides alerts. Studies in two hospital pharmacies showcased efficient sub-region classification based on the risk assessment. Additionally, our study employs hierarchical clustering techniques to categorize risk dynamically, ensuring a scalable and efficient assessment model. The DT-RAHP mobile application further enhances risk communication by disseminating warnings to users and officials when the risk is high. This proactive risk management framework offers a robust methodology applicable to various airborne contagious diseases. Moreover, the flexibility of the methodology allows for incorporating additional risk factors, enhancing result precision and comprehensiveness.</div></div>","PeriodicalId":62710,"journal":{"name":"安全科学与韧性(英文)","volume":"7 2","pages":"Article 100246"},"PeriodicalIF":3.4,"publicationDate":"2026-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145924082","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Exploring cascading failures in supply chain risk management: A systematic review, 2013-2024","authors":"Xiaoxin Zhu ,&nbsp;Jiahui Zhu ,&nbsp;David Regan ,&nbsp;Zhimin Wen","doi":"10.1016/j.jnlssr.2025.100234","DOIUrl":"10.1016/j.jnlssr.2025.100234","url":null,"abstract":"<div><div>Supply chain risk management has become increasingly crucial due to the complexity and interconnectivity of modern supply chains. Among various risk factors, cascading failures have emerged as a significant concern for risk management. This study systematically reviews the literature published between 2013 and 2024, focusing on cascading failures in supply chain networks. A total of 92 articles were selected from the Web of Science (WoS) database for in-depth analysis. The findings indicate that future research needs to achieve breakthroughs in three dimensions: data acquisition, theoretical innovation, and practical application. Specifically, integrating multi-source data is crucial for enhancing the accuracy, comprehensiveness, and timeliness of data in supply chain networks. Higher-order network modeling and digital twin technology will aid in more accurately simulating and predicting cascading failures. From a theoretical standpoint, exploring hybrid failure mechanisms and multi-level propagation patterns will deepen the understanding of failure propagation in complex networks. From a practical perspective, developing resilience design standards, intelligent early warning systems, and differentiated policy tools will strengthen the risk resistance capabilities of supply chains. Overcoming these challenges will require a holistic approach that combines data-driven insights, theoretical advancements, and practical solutions to build a safer and more resilient global supply chain system.</div></div>","PeriodicalId":62710,"journal":{"name":"安全科学与韧性(英文)","volume":"7 2","pages":"Article 100234"},"PeriodicalIF":3.4,"publicationDate":"2026-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145520268","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Human detection with smoke occlusion based on AI-generated images","authors":"Zhen Xu,&nbsp;Xiyan Tang,&nbsp;Wenting Li,&nbsp;Yujie Zhao,&nbsp;Donglian Gu,&nbsp;Yuan Tian","doi":"10.1016/j.jnlssr.2025.100235","DOIUrl":"10.1016/j.jnlssr.2025.100235","url":null,"abstract":"<div><div>Detecting occupants trapped in buildings is challenging during fire accidents because of the dense, rapidly spreading smoke. To this end, this paper proposes a human detection method that accounts for smoke occlusion, based on artificial intelligence (AI)-generated images and an improved real-time detection transformer (RT-DETR) model. First, a procedure is designed to construct a dataset of human images with smoke occlusion using generative AI, addressing the current dataset shortage. Second, an RT-DETR model is optimized using a transformer-based global-local interaction network to increase the sensitivity to human characteristics with smoke occlusion. Third, corresponding validations were conducted. The results indicated that the proposed method could achieve remarkable average precision (up to 93.8 %) for smoke occlusion scenarios. Finally, the proposed method was employed in real-world fire cases. The proposed method can accurately detect humans with smoke occlusion in building fires, effectively aiding fire emergency evacuation and rescue efforts.</div></div>","PeriodicalId":62710,"journal":{"name":"安全科学与韧性(英文)","volume":"7 2","pages":"Article 100235"},"PeriodicalIF":3.4,"publicationDate":"2026-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145467589","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Influencing factors and mechanisms of super high-rise buildings safety risks: A Fuzzy-DEMATEL-AISM analysis","authors":"Ziyu Jia ,&nbsp;Jiakai Xiu ,&nbsp;Ke Liu ,&nbsp;Lihua Zhao ,&nbsp;Wei Zhang ,&nbsp;Ziqi Song ,&nbsp;Na Xiao ,&nbsp;Tiantian Du","doi":"10.1016/j.jnlssr.2025.100232","DOIUrl":"10.1016/j.jnlssr.2025.100232","url":null,"abstract":"<div><div>As global urbanization accelerates, super high-rise buildings increasingly face critical safety challenges due to mid-life vulnerabilities that can cause casualties and economic losses. This study examines two typical risk systems in super high-rise buildings—primary structure damage and building envelope detachment—to address urban risks and emerging safety management needs.</div><div>Through a mixed-methods approach, we developed a theoretical framework based on \"environmental pressure - building physical condition - management response\" relationships. Using literature analysis and the Delphi method, we established two comprehensive risk indicator systems and analyzed their interaction mechanisms using fuzzy DEMATEL-AISM. Our findings reveal that root-level environmental pressures (e.g., seismic activity, wind/snow loads) trigger structural disasters, while intermediate-level building physical conditions mediate these pressures and management responses, forming core disaster pathways. Management factors influence safety outcomes throughout a building’s lifecycle by operating across both root and result levels, with significant feedback loops in both risk systems.</div><div>Based on these findings, we propose four integrated management strategies: root-level environmental pressure monitoring for early prevention; efficient identification of physical defects from surface damage; full-lifecycle risk management with specialized operations and maintenance; and collaborative diagnostics driven by loop factor assessment. This research bridges engineering perspectives and safety management, offering stakeholders—including building owners of super high-rise buildings, government authorities, and design professionals—practical approaches to enhance safety for millions of inhabitants while contributing a comprehensive theoretical framework for risk analysis.</div></div>","PeriodicalId":62710,"journal":{"name":"安全科学与韧性(英文)","volume":"7 2","pages":"Article 100232"},"PeriodicalIF":3.4,"publicationDate":"2026-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145365211","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Maritime man-overboard search using a lightweight and efficient end-to-end detection transformer","authors":"Guokang Xu ,&nbsp;Jianchuan Yin ,&nbsp;Nini Wang ,&nbsp;Zeguo Zhang","doi":"10.1016/j.jnlssr.2025.100267","DOIUrl":"10.1016/j.jnlssr.2025.100267","url":null,"abstract":"<div><div>Maritime transportation plays a crucial role in global economic trade. However, maritime accidents occur frequently, posing significant threats to the safety of seafarers. In search and rescue scenarios for man-overboard, unmanned aerial vehicles (UAVs) are gradually replacing manned aircraft and helicopters. Besides, detecting man-overboard is challenging because of their small pixel size, weak signals, and indistinct features on the ocean surface. Furthermore, existing detectors struggle to strike a balance between lightweight design for UAVs and detection accuracy. To address this issue, the novel Man-overboard Detection Transformer (MOB-DETR) is proposed. On the one hand, the Token Enhancement layer is introduced, which conducts fine-grained filtering of spatial and channel dimensions, reducing redundant encoding caused by background queries. On the other hand, the Effusion Fusion Module, based on the RepViT Block, is proposed, effectively eliminating computational redundancy by decoupling the interaction mechanisms between spatial and channel dimensions. Additionally, to fill the existing gap in benchmark datasets for detecting man-overboard, the ManOverboard benchmark dataset has been established. In the experimental validation phase, MOB-DETR is conducted on ManOverboard and SeaDronesSeev2. Ablation experiments show that MOB-DETR achieves 11.7 % better lightweight performance and 14.4 % higher <span><math><mrow><mi>A</mi><msub><mi>P</mi><mi>small</mi></msub></mrow></math></span> than baselines. Comparison experiments on ManOverboard and SeaDronesSeev2 validate its effectiveness, offering an efficient solution for man-overboard detection. Overall, this research not only advances man-overboard detection but also significantly enhances the resilience of maritime transportation, ultimately protecting seafarers' lives and ensuring the reliability of the world's essential trade routes.</div></div>","PeriodicalId":62710,"journal":{"name":"安全科学与韧性(英文)","volume":"7 2","pages":"Article 100267"},"PeriodicalIF":3.4,"publicationDate":"2026-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145839662","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Intelligent fire modeling in wildland-urban interface: A comprehensive review of current progress, challenges, and future perspectives","authors":"Ao Zheng ,&nbsp;Rui Ba ,&nbsp;Wenyu Jiang ,&nbsp;Zijun Chen ,&nbsp;Menghao He ,&nbsp;Yuansheng Hua ,&nbsp;Song Zhu ,&nbsp;Jiasong Zhu ,&nbsp;Guochao Liu ,&nbsp;Zhuojie Zhu ,&nbsp;Xinyi Han","doi":"10.1016/j.jnlssr.2025.100252","DOIUrl":"10.1016/j.jnlssr.2025.100252","url":null,"abstract":"<div><div>Wildfire prevention and control, especially within complex Wildland-Urban Interface (WUI), face escalating challenges due to the synergistic impacts of climate change and expanding urban frontiers. While advanced wildfire spread prediction models are essential for developing disaster-resilient emergency systems, the increasing complexities of WUI wildfire scenarios highlight critical limitations in current modeling approaches. These complexities include the dynamic interactions between vegetation and built environments, the demands of multi-scale spatiotemporal forecasting, and the challenges of cross-platform integration. To better understand and address these challenges, this paper establishes a novel tripartite analytical framework through a systematic review of peer-reviewed studies from the Scopus and Web of Science databases: (1) fire combustion characterization, (2) fire dynamics mechanism, and (3) fire management system. Our critical analysis identifies three persistent research challenges, including modeling fire behaviors in heterogeneous WUI scenarios; balancing fire dynamics with computational speed, accuracy, and resolution; and managing and applying models in highly integrated systems. This study concludes with actionable priorities for subsequent research, providing methodological guidelines for model developers and evidence-based integration pathways for emergency management systems, particularly in addressing critical infrastructure protection in rapidly urbanizing, fire-prone regions.</div></div>","PeriodicalId":62710,"journal":{"name":"安全科学与韧性(英文)","volume":"7 2","pages":"Article 100252"},"PeriodicalIF":3.4,"publicationDate":"2026-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145789662","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Predicting models for critical velocity at the stair entrance under a two-story building fire","authors":"Zeng Long ,&nbsp;Cong Su ,&nbsp;Chang Liu ,&nbsp;Fazheng Chong ,&nbsp;Fan Tong ,&nbsp;Fengju Shang ,&nbsp;Jiaqing Zhang ,&nbsp;Jiansong Wu","doi":"10.1016/j.jnlssr.2025.100266","DOIUrl":"10.1016/j.jnlssr.2025.100266","url":null,"abstract":"<div><div>The cross-floor spread of fire smoke in a two-story building hinders personal evacuation and leads to casualties. To address this issue, sufficient wind pressure needs to be applied at the stair to prevent smoke from entering the upper space, and the critical velocity is an important parameter worthy of attention. By conducting a series of model experiments and numerical simulations in this study, the critical velocity below the ceiling screen at the stair is investigated, considering the factors of heat release rate (HRR), fire source location, ceiling screen depth and number of stairs. The results show that the critical velocity is proportional to one-third of the HRR, and the proportional coefficient decreases as the fire source is farther from the stair and the ceiling screen depth increases. The prediction models for the critical velocity under fire sources located in front and back of the stairs are proposed, respectively, and the proposed model is validated under three stairs. Additionally, the proposed prediction model is compared with the velocity value required by the relevant standard, and some dangerous situations are identified that require further strengthened ventilation. This study can provide theoretical guidance for ventilation design and emergency response in actual engineering with similar structures.</div></div>","PeriodicalId":62710,"journal":{"name":"安全科学与韧性(英文)","volume":"7 2","pages":"Article 100266"},"PeriodicalIF":3.4,"publicationDate":"2026-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145326926","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Machine learning hybrid dynamic best model selection algorithm for real-time fire prediction using IoT-enabled multi-sensor data in buildings","authors":"Mujeeb Ali Khan ,&nbsp;Weiguo Song ,&nbsp;Abbas Khan ,&nbsp;Mazhar Ali ,&nbsp;Rehmat Karim ,&nbsp;Jun Zhang","doi":"10.1016/j.jnlssr.2025.100236","DOIUrl":"10.1016/j.jnlssr.2025.100236","url":null,"abstract":"<div><div>Fire disasters in urban areas, including homes, offices, and industrial facilities, have increased significantly over the past decade, causing extensive damage and loss of life. Integrating intelligent fire detection systems with machine learning (ML) is crucial for providing early warnings and facilitating effective response coordination. In this research, a novel hybrid dynamic best model selection (HDBMS) ML-based algorithm is proposed for internet of things (IoT)-enabled fire detection in buildings, which outperforms traditional static models by providing higher accuracy and adaptability across diverse fire scenarios. The proposed algorithm employs feature selection pre-processing techniques, followed by the synergistic integration of five classifiers: support vector classifier (SVC), logistic regression, random forest, Gaussian Naïve Bayes (Gaussian NB), and decision tree, aiming to enhance prediction accuracy and robustness across various fire scenarios. This system dynamically selects the optimal classifier based on real-time performance metrics such as precision, accuracy, F1-score, and recall. This approach was rigorously validated using our developed dataset and real-time sensor data to monitor smoke, temperature, and humidity under various fire scenarios. Following algorithm validation, a laboratory-constructed multi-sensor fire detection node prototype wirelessly feeds sensor data to the ThingSpeak cloud platform for real-time data analysis and communication with ML algorithms in the back-end. The system's improved precision, accuracy, and root mean squared error (RMSE) confirm its effectiveness. The results demonstrate that the proposed approach achieves superior classification accuracy compared with existing methods in the literature. Furthermore, this study's novelty lies in the dynamic selection of the most effective model in real-time, a characteristic that is currently lacking in fire detection systems, thereby enhancing the system's flexibility and efficiency in various fire scenarios.</div></div>","PeriodicalId":62710,"journal":{"name":"安全科学与韧性(英文)","volume":"7 2","pages":"Article 100236"},"PeriodicalIF":3.4,"publicationDate":"2026-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145467616","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}