New generation information technology-enabled innovative model of dual prevention mechanism for mine safety

The dual prevention mechanism provides a systematic solution to address the difficulty of achieving accurate identification and dynamic control of risks and hazards in the traditional mine safety management model, which overly relies on manual inspection and post-incident handling. This study presents the current research and application status of the dual prevention mechanism for mine safety and points out that the current mechanism faces problems such as static and mechanized risk assessment and control, fragmented and isolated hazard investigation and data management, lagging and shallow dynamic response and informatization application, and structural imbalance between personnel capabilities and resource allocation. The applicability of new generation information technologies, including digital twin, large Artificial Intelligence (AI) models, and mine Internet of Things, in the dual prevention mechanism for mine safety is analyzed. A fusion framework integrating digital twin and large AI models based on the mine Internet of Things is developed, and an innovative model of the dual prevention mechanism for mine safety is proposed based on this framework. In this model, digital twin is used to achieve virtual mapping and real-time simulation of risk scenarios. An intelligent control platform is relied upon to complete multi-source data integration and strategy linkage. Large AI models are leveraged to enable knowledge-driven intelligent analysis, forming a dual prevention system with self-learning and self-optimization capabilities. The analysis indicates that the application of the innovative model needs to address data security and privacy protection, model reliability and drift issues. The future development of the dual prevention mechanism for mine safety is expected to achieve three major transformations: from passive response to active defense, from single-point control to system-wide coordination, and from experience-driven to data- and mechanism-driven approaches.