煤矿职工群体安全行为研究进展综述

Review on the Research Progress of Group Safety Behavior of Coal Mine Workers

摘要: 煤矿安全生产直接关系到矿工生命安全和国家能源供应，而煤矿职工的不安全行为是导致事故发生的重要因素之一。近年来，煤矿职工群体安全行为的研究逐步从个体行为分析向系统化、智能化、群体行为模拟方向发展。本文系统综述了煤矿职工群体安全行为的国内外研究进展，重点探讨了安全行为的理论基础、群体行为模拟方法及安全行为预警机制。在安全行为理论方面，介绍了事故因果连锁理论、瑞士奶酪模型、HFACS等经典理论，以及系统安全管理的最新进展；在群体行为模拟方面，分析了系统动力学、多智能体建模、QSIM等方法在煤矿职工群体行为预测中的应用；在安全行为预警方面，探讨了基于人工智能、贝叶斯网络、模糊综合评判等技术的智能预警方法，并分析了当前预警系统的不足。研究表明，煤矿职工的安全行为受个体、群体和组织管理等多重因素影响，现有研究在群体行为影响因素整合、行为模拟精准性、安全行为预警实时性等方面仍存在不足。未来研究应进一步结合人工智能、深度学习、数据融合和智能监测等技术，提高煤矿职工群体安全行为的监测、模拟和预测能力，构建更加精准、智能的安全预警体系，以提升煤矿安全管理的科学性和有效性。

Abstract: Coal mine safety is directly related to miners’ life security and national energy supply, while unsafe behaviors of coal mine workers are one of the key factors leading to accidents. In recent years, research on coal mine workers’ group safety behavior has gradually evolved from individual behavior analysis to systematic, intelligent, and group behavior simulation approaches. This paper systematically reviews the research progress on coal mine workers’ group safety behavior both domestically and internationally, focusing on the theoretical foundations of safety behavior, group behavior simulation methods, and safety behavior early warning mechanisms. In terms of safety behavior theories, classic models such as the Accident Causation Chain Theory, Swiss Cheese Model, and HFACS are introduced, along with the latest advancements in systematic safety management. Regarding group behavior simulation, methods including System Dynamics, Multi-Agent Modeling, and QSIM are analyzed for their applications in predicting coal mine workers’ group behavior. For safety behavior early warning, this paper discusses intelligent early warning methods based on Artificial Intelligence, Bayesian Networks, and Fuzzy Comprehensive Evaluation, while also analyzing the current limitations of early warning systems. Research findings indicate that coal mine workers’ safety behavior is influenced by multiple factors, including individual psychology, group interactions, and organizational management. However, current studies still have limitations in integrating group behavior impact factors, improving simulation accuracy, and enhancing the real-time capability of safety behavior early warning systems. Future research should further integrate Artificial Intelligence, Deep Learning, Data Fusion, and Intelligent Monitoring to enhance the monitoring, simulation, and prediction of group safety behavior in coal mines. The goal is to develop a more precise and intelligent safety early warning system, thereby improving the scientific and practical effectiveness of coal mine safety management.