Search Results (3,488)

Autonomous navigation technologies have been widely adopted in the automotive and aviation sectors, significantly reducing human-error-induced accidents and operational costs. However, their application to maritime systems remains limited due to the complexity of conventional propulsion systems. Electric propulsion ships, with well-defined system boundaries and accessible operational data, offer a promising platform for autonomous navigation. In this study, we propose an FMEA-guided selective multi-fidelity digital twin framework for fault detection, where model fidelity is adaptively selected between low- and high-fidelity models based on risk priority numbers derived from failure mode and effects analysis. This approach enables selective execution of computationally expensive models only under high-risk conditions, thereby improving computational efficiency. In addition, a sliding window-based algebraic aggregation method is employed to achieve lightweight and real-time fault diagnosis. The proposed framework is validated using operational sensor data from a 100 kW electric propulsion ship under multiple fault scenarios, including power supply faults and signal anomalies. Experimental results show that the proposed method reduces computational cost while maintaining stable real-time performance, compared to conventional data-driven AI-based approaches. These results demonstrate that the proposed framework provides an effective and efficient solution for enhancing the reliability and safety of autonomous ship systems. Full article

Compared with other construction operations, high-altitude operations are more dangerous. Falling from a height is the main type of accident in construction. It is important to study the human risk of falling from height to reduce falling accidents. Based on the Human Factors Analysis and Classification System (HFACS) model, a preliminary evaluation index system for fall risk in building construction was established. Through the Delphi method and sensitivity analysis, the initial indicators were screened, the index factors that did not meet the requirements were removed, and the final human risk index evaluation system was determined. The system includes five first-level indicators and 17 s-level indicators of organizational influence, unsafe supervision, preconditions for unsafe behavior, and unsafe behavior. Subsequently, the analytic network process–entropy weight method (ANP-EWM) is used to subjectively and objectively weight the evaluation indicators, and the combined weight is obtained through game theory. The matter–element extension model is constructed to evaluate the human risk of falling from height in construction. Finally, an empirical analysis is carried out with the Y project as a case study. The novelty of this study lies in integrating human-factor analysis with the matter–element extension model for fall risk assessment in construction, while combining ANP, the entropy weight method, and game theory to balance subjective and objective weighting. The proposed model provides a practical tool for evaluating and controlling human risk in high-altitude construction operations. The results show that the correlation degree calculated according to the matter–element extension model is K₄ = 3.5, and the human risk of falling from height in the construction of Y project has generally reached an excellent level. However, the evaluation level of some evaluation indexes is still low, which is consistent with the actual situation of construction enterprises in Y project. This model provides a direction for the study of human risk assessment of falling from different construction heights. Full article

With the rapid development of society, the number of road vehicles has increased significantly, leading to a growing severity of traffic accident issues. Timely and accurate detection of road traffic anomalies or accidents is crucial for reducing fatalities and alleviating traffic congestion. Consequently, the detection of road traffic anomalies has become a focal point of research in recent years. With the assistance of computer technologies such as deep learning, researchers have developed more accurate and effective methods for detecting road traffic anomalies. However, the small proportion of anomaly-prone areas in surveillance video frames, combined with the complex and difficult-to-capture patterns of accidents, presents new challenges for the application of deep models to traffic anomaly detection from a surveillance perspective. In light of this, this paper annotates the TADS dataset we previously proposed, a popular text-assisted video representation learning method, to develop a more efficient detection method. Utilizing the well-known video-text model CLIP, we have constructed a detection model that leverages unique text and eye-gaze annotation data from the TADS dataset to learn anomaly representations more effectively, thereby improving the detection of road traffic anomalies from a surveillance perspective. Experimental results demonstrate the superiority of our model for detecting traffic anomalies from a surveillance perspective, as well as the utility of the text and eye-gaze data included in the dataset. Full article

Traffic accident detection is a critical component of intelligent transportation systems (ITS), enabling timely incident response and traffic management. While most existing approaches rely on temporal information from video sequences, such methods are not always applicable in resource-constrained surveillance environments. This study investigates the feasibility of detecting traffic accidents from single static images by formulating the task as a binary classification problem. Representative architectures, including Vision Transformer (ViT), Swin Transformer, and ResNet-50, are systematically evaluated on the Car Crash Dataset (CCD) under multiple training configurations. To assess generalization capability, cross-domain evaluation is conducted using an external crash video dataset (ECVD) constructed to approximate real-world deployment conditions. Experimental results show that all models achieve strong performance under in-domain evaluation. However, cross-domain testing reveals substantial performance degradation, particularly in recall, indicating limited generalization capability under domain shift. Qualitative analysis further shows that missed detections are associated with weak visual cues, occlusion, and complex traffic environments, while false positives are caused by visually ambiguous patterns resembling accident scenarios. Unlike prior studies that primarily report performance improvements, this work provides empirical evidence that model behavior in static-image-based accident detection is governed by dataset composition rather than architectural design. Therefore, static-image-based accident detection should be interpreted as a coarse-level screening tool rather than a fully reliable decision-making system. This study highlights the importance of data-centric design and cross-domain evaluation for improving real-world applicability. Full article

In the complex and high-risk underground environment of coal mining, ensuring occupational health and safety is a fundamental pillar of social sustainability. Traditional safety compliance is insufficient to prevent unpredictable accidents and sustain long-term enterprise resilience. Thus, fostering proactive safety citizenship behavior is essential for enhancing organizational resilience. Drawing on the cognitive appraisal theory of stress, this study constructs a double-edged sword model of felt accountability on miners’ safety citizenship behavior. A three-wave time-lagged survey was conducted among 375 frontline coal miners in China, with data analyzed using SPSS 26.0 and AMOS 24.0. The findings show that felt accountability can increase work engagement and promote employee safety citizenship behavior, while also enhancing psychological strain and inhibiting employee safety citizenship behavior. In addition, safety-specific transformational leadership amplifies the positive impact of felt accountability on work engagement and mitigates its effects on psychological strain. These findings enrich our understanding of the impact of felt accountability, and provide practical insights for coal enterprise managers to improve sustainable safety performance and foster a socially sustainable work environment. Full article

Aiming to address the problems of low accuracy in coal–rock identification during coal mining, which lead to energy waste and safety hazards, a high-precision coal–rock medium identification method combining terahertz time-domain spectroscopy technology and multiple machine learning algorithms is proposed. By preparing coal–rock samples with a gradient change in coal content, terahertz time-domain spectroscopy data of coal–rock mixed media are collected, and optical parameters such as the refractive index and absorption coefficient are extracted. Principal component analysis is used to reduce the dimensionality of the terahertz data, and machine learning algorithms such as support vector machine, least squares support vector machine, artificial neural networks, and random forests are adopted for classification and identification. The study found that terahertz waves are more sensitive to coal–rock media in the 0.7–1.3 THz frequency band, and that the refractive index and absorption coefficient of coal–rock mixed media are significantly positively correlated with coal content within the range of 0–30%. After feature extraction and K-fold cross-validation, the random forest model achieved a coal–rock classification accuracy of over 96% on the test set, significantly outperforming other comparison algorithms. The research verifies the efficiency and practicality of terahertz technology combined with multiple machine learning algorithms in coal–rock identification, providing a new method for fields such as mineral separation. This method has, to a certain extent, broken through the accuracy bottleneck of traditional coal–rock identification technologies within its applicable range, providing a new solution for real-time detection of coal–rock interfaces and is expected to further reduce the risks of ineffective mining and roof accidents in the future. Full article

Road accidents related to long-distance driving remain a major safety concern, primarily driven by fatigue and musculoskeletal discomfort. This study investigates the acceptance of driver assistance technologies, specifically seat adjustment and fatigue warning applications using the Technology Acceptance Model (TAM). Data were collected from 1600 drivers in Thailand. Participants were categorized into low-discomfort (n = 785) and high-discomfort (n = 815) groups based on the Cornell Musculoskeletal Discomfort Questionnaire (CMDQ). Multi-group Structural Equation Modeling (SEM) demonstrated that the TAM framework adequately explains technology acceptance in both groups. However, significant differences in the underlying mechanisms were observed. For drivers with lower discomfort levels, attitude toward technology played a more prominent role in shaping usage intention. In contrast, perceived usefulness emerged as the dominant factor for drivers experiencing higher musculoskeletal discomfort. These findings indicate that physical discomfort functions as a critical moderator in technology acceptance. Practically, the results highlight the need for driver assistance systems that prioritize ergonomic effectiveness to enhance adoption among physically strained users. Full article

Occupational safety analytics is increasingly moving toward data-driven methodologies; however, existing models often struggle to capture the multidimensional nature of accident causation. This study presents a multimodal Hybrid Transformer-LSTM framework for classifying occupational fatalities by jointly modeling unstructured narratives, cyclical temporal features, and regional spatial indicators. Utilizing a large-scale dataset of 14,914 OSHA fatality records, the proposed architecture leverages BERT-based embeddings for semantic extraction and Bidirectional LSTMs as non-linear pattern encoders for spatiotemporal context. Conceptually grounded in the Swiss Cheese Model, the framework treats different data modalities as proxies for distinct layers of system risk, ranging from proximal unsafe acts to environmental preconditions. Experimental results show that the multimodal architecture achieves an accuracy of 84.56%, representing a 5.33% gain over unimodal BERT baselines. To address the inherent “black-box” nature of deep learning, a SHAP-based explainability framework is incorporated to quantify the contributions of both textual tokens and environmental features to the model’s decision-making process. The results indicate that integrating narrative semantics with temporal and spatial context enhances discriminative performance and enables context-aware classification within a weakly supervised setting. By providing a scalable and interpretable classification framework, this study offers a data-driven decision-support approach for safety professionals and regulatory bodies seeking to implement evidence-based risk management strategies in high-risk industrial sectors. Full article

Driver emotion recognition is a crucial method for reducing traffic accidents. Most existing research focuses on experienced drivers as the primary research subjects, overlooking novice drivers, who are inexperienced in driving. However, novice drivers can easily lose control of their emotions due to the high mental load during driving, which can lead to serious traffic accidents. Therefore, to recognize the emotions of novice drivers for timely warnings, we propose an emotion recognition model based on multimodal information. The model consists of a facial feature extraction module, an eye movement feature extraction module and a classifier. The facial feature extraction module uses the ViT-B/16 to extract the facial features of novice drivers. The eye movement feature extraction module is a hybrid network containing Bi-LSTM and Transformer. It extracts eye movement features of novice drivers. Facial features and eye movement features are fused and fed to the classifier. The classifier can output the five major emotion categories of surprise, anger, calm, happy, and other for novice drivers. The experimental results demonstrate that our model accurately recognizes the emotions of novice drivers with an accuracy of 98.72%, surpassing that of other models. Full article

Ship collision accidents remain a critical concern in maritime safety because of their potential to cause operational disruption as well as environmental and economic damage in areas with dense shipping activity. Complex traffic interactions, differences in vessel characteristics, and dynamic environmental conditions make collision risk increasingly difficult to manage using traditional navigation measures alone. This paper presents a structured review of ship collision research, focusing on collision impacts, collision avoidance strategies, risk assessment methodologies, and safety index development. The review synthesizes reported collision cases and their environmental consequences, examines commonly used analytical frameworks including probabilistic, data-driven, and multicriteria approaches, and discusses recent developments in AIS-based analysis, sensor-based monitoring, and intelligent prediction techniques. The analysis identifies several methodological gaps in existing studies. Collision avoidance methods and risk assessment models are often developed independently, while their integration with safety index frameworks remains limited. In addition, safety index formulations differ considerably in terms of indicator selection and modeling approaches, which reduces comparability between studies conducted in different waterways. The findings highlight how different analytical approaches contribute to maritime safety evaluation at strategic, operational, and real-time levels and provide insights for developing more integrated safety assessment frameworks to support navigation risk monitoring in high-traffic maritime environments. Full article

Currently, the lack of analysis and applicable circuit models for the evolution of cable joint faults is responsible for explosions or fire accidents in the distribution network system. In this paper, the modeling of high-resistance to low-resistance grounding faults for 10 kV cable joints is investigated. Firstly, the physical evolution from high-resistance to low-resistance grounding faults in 10 kV cable joints is analyzed. Secondly, the common discharge characteristics under different evolution stages are extracted by simulation experiments and fault-recording data. Thirdly, an interface breakdown circuit model and a radial breakdown circuit model are established to quantitatively describe the high-resistance to low-resistance grounding faults of cable joints. Fourthly, the corresponding arc resistance models are proposed, and the controlled parameter values of the models under different evolution stages are given. Finally, the fault identification control model is implemented for relay protection. This paper provides theoretical and modeling support for the fault identification of 10 kV cable joints, filling the knowledge gap of this critical fault type in relay protection. Full article

Poor visibility contributes to nighttime accidents at highway intersections, especially in developing countries where vehicles mix with vulnerable road users (VRUs) such as pedestrians and cyclists. Unlike downtown intersections with traffic signals and ambient lighting, rural intersections have no signals and minimal ambient light, forcing drivers to rely on roadway lighting for hazard recognition. Improving illumination arrangements can significantly reduce the likelihood of crashes. However, there are significant differences in the effects of illumination on drivers’ visual search ability at different vehicle speeds. Therefore, the collaborative matching of illumination and speed limits can effectively improve traffic efficiency and reduce the probability of nighttime accidents. In this paper, we establish a collaborative optimization model of illumination and speed limits at rural highway intersections that considers drivers’ visual recognition of VRUs. We then design an experiment with illuminance, vehicle speed, and VRU type/location as control variables to collect recognition distances, and finally analyze their effects to calculate speed limits under different illuminances. Results indicate that pedestrians and cyclists appearing from the left side are recognized 24.73% and 15.79% earlier than those from the right, suggesting that VRUs from the right side are more vulnerable. Additionally, the safety benefit of improving illumination on increasing speed limits gradually diminishes as illuminance rises. Therefore, determining the most suitable illumination and speed limit configuration requires a comprehensive evaluation of the cost–benefit relationship between lighting investments and the gains resulting from higher speed limits. Full article

The prestressed concrete containment structure constitutes the core protective structure of a nuclear power plant. This paper utilises the prestressed concrete containment vessel (PCCV) of the Hualong Pressurised Reactor 1000 (HPR-1000)—a third-generation pressurised water reactor (PWR)—as the primary research prototype. Utilising ANSYS, a finite element model was established, with key points selected at critical locations such as the dome, cylinder, and base slab for stress analysis calculations. Reinforcement quantification derived from the design methodologies and analytical formulations prescribed in ACI 349 and ACI 359 were compared under various loading conditions. This investigation identified the core discrepancies and influencing factors between the two codes in reinforcement design, alongside a sensitivity analysis to identify key parameters affecting reinforcement design in different structural zones. The results indicate that discrepancies in reinforcement requirements stem primarily from the divergent design philosophies and strength assessment formulations, with this influence outweighing variations in load combinations. Furthermore, significant spatial differences exist in the sensitivity of reinforcement designs for key components to parameters such as the height-to-diameter ratio, shutdown seismic actions, accident pressure, and temperature effects. The conclusions of this study establish theoretical foundations and furnish empirical data to enhance the computational efficiency of prestressed concrete containment design for pressurised water reactor (PWR) facilities, while supporting the alignment of national and international regulatory standards. Furthermore, they serve as a technical reference for advancing nuclear power structural design practices. Full article

In rural areas affected by Chernobyl, accelerated erosion has become a major source of particulate ¹³⁷Cs in sediment load. The long-term dynamics of the activity concentration in eroded soil material transported from individual slope catchments can be better understood by exploring the ¹³⁷Cs depth distribution in sediments deposited near cultivated fields. This study focuses on three cultivated slope catchments located in the Chernobyl-affected area of Central Russia. A depth incremental campaign was conducted within zones of sediment accumulation in 2022–2025. The behavior of radiocaesium associated with particles after the Chernobyl accident was controlled by the prompt implementation of remediation measures. Shortly after the accident, the values decreased by more than two times. The radionuclide flux then began to depend on soil erosion processes. Gradually, the thickness of the upper soil that had been eroded became large enough to allow soil material from deeper layers to be involved during ordinary plowing and led to a subsequent decrease in the ¹³⁷Cs activity concentration. Given the decreasing snowmelt runoff and lack of increase in high-intensity rainfall in the 21st century, the activity concentration of ¹³⁷Cs in slope runoff has remained quite stable. This phenomenon requires consideration of whether a physically based model for the transport of particulate radionuclides should be developed. Full article

Same-level falls are the most frequent occupational accidents, yet traditional manual analysis of accident reports is labor-intensive and limits large-scale prevention strategies. In this pilot study, we aimed to evaluate the accuracy of using large language models (LLMs) to automate the classification of occupational accident text data without task-specific pretraining. We analyzed data from 2619 same-level-fall-related injury cases, using expert manual classification as the reference standard. Four models—GPT-4o mini, GPT-4.1 mini, GPT-4.1, and o4-mini—were compared using accuracy and Cohen’s kappa. The o4-mini model demonstrated the highest performance, showing statistical superiority in the complex “causal agent” category with 72.8% accuracy. For other classification tasks, the top models achieved accuracies of 82–92%, with Cohen’s kappa coefficients > 0.7, indicating substantial agreement with expert judgments. These findings suggest that LLMs can classify occupational accident text with substantial agreement with the expert-derived reference standard in this dataset. This automated approach enables efficient, high-frequency analysis of large datasets, offering a promising tool for large-scale occupational accident surveillance and screening. Full article