Nuclear power plants (NPPs) under construction are required to meet the stringent safety standards of operational facilities while also facing the heightened risk characteristics of construction projects. The combination of dense worker populations and generally low safety awareness presents serious challenges for ensuring construction safety. To address this, the present study proposes a BN–Game–SD multi-algorithm fusion model that systematically examines the evolution of behavioral risks from both group and individual perspectives. First, a behavioral indicator system was constructed using Bayesian Networks (BNs) to identify key risk factors. Then, a dynamic payoff matrix game model was introduced to analyze the incentive mechanisms between individuals and groups. Finally, a BN–Game–SD model was developed to capture the dynamic evolution of worker behaviors in NPP construction. Simulation results reveal that, under fixed probabilities of safety strategy selection, clear thresholds exist in group resistance to individual behavioral deviation and vice versa. Applied to a real NPP construction site, the model helped achieve a 10.39% reduction in safety violations. This study provides a theoretical foundation for promoting self-organized safety behavior evolution in nuclear enterprises and presents an innovative methodological framework for safety management in nuclear engineering.
Residents of coal mining communities include both coal mine workers and local residents, making the guidance of public opinion essential for maintaining social stability in such communities. Therefore, this paper utilizes an improved word2Vec model to extract factors influencing public opinion from a large number of accident cases. It then develops a coal mining community public opinion dissemination model based on game theory, focusing on two groups: official media and opinion leaders. By integrating cellular automata (CA) and agent-based modeling, this study examines the dissemination patterns of public opinion in coal mining communities. The simulation results and practical applications indicate that public opinion in coal mining communities spreads rapidly and can be effectively shaped. Positive guidance from official media plays a crucial role in directing public opinion. However, as interactions evolve, public opinion dynamics may become less favorable. Strengthening the intensity of positive guidance from official media further enhances its ability to shape and influence public opinion in coal mining communities. This research provides a novel perspective and methodology for studying community safety management, offering significant theoretical and practical implications.
Prolonged violation behavior is characterized by sustained temporal presence, slow action changes, and similarity to normal behavior. Due to the complex construction environment, intelligent recognition algorithms face significant challenges. This paper proposes an improved YOLOv8-based model, DGEA-YOLOv8, to address these issues, using “playing with mobile phones” as a case study. The model integrates the DCNv3 module in the backbone to enhance behavior deformation adaptability and the GELAN module to improve lightweight performance and global perception in resource-limited environments. An ECA attention mechanism is added to enhance small target detection, while the ASPP module boosts multi-scale perception. ByteTrack is incorporated for continuous tracking of prolonged violation behavior in construction scenarios. Experimental results show that DGEA-YOLOv8 achieves 94.5% mAP50, a 2.95% improvement over the YOLOv8s baseline, with better data capture rates and lower ID change rates compared to algorithms like Deepsort and Strongsort. A construction-specific dataset of over 3000 images verifies the model’s effectiveness. From the perspective of data symmetry, the proposed model demonstrates strong capability in addressing asymmetric feature distributions and behavioral imbalance inherent in prolonged violations, restoring spatiotemporal consistency in detection. In conclusion, DGEA-YOLOv8 provides a precise, efficient, and adaptive solution for recognizing prolonged violation behaviors in construction sites.