构建煤矿灾害智能化精准管控体系有助于提高事故预测预警准确度，实现智能化风险研判。现有研究大多聚焦单一灾害进行监测预警技术或算法优化，缺乏多灾害监测预警协同机制和事故应急联动下的避灾路径规划，且数据传输时延大，管控效率较低。提出了一种煤矿灾害“云边端”一体化智能精准管控技术体系，介绍了该体系架构和监控预警数据流转与交互机制，从数据精准感知、边缘计算、云平台3个角度重点分析了该体系的关键技术，包括：在端侧，研发瓦斯、火灾、粉尘、顶板等多灾害智能传感器，构建基于IPv6的5G+4G+WiFi6的高速低时延通信网络，优化感知设备及联动控制装备部署方案；在边侧，建立基于深度学习AdaTT模型的煤矿重大灾害数据融合分析模型，研发矿用AI视频分析设备以实现安全隐患图像识别，开发基于边缘计算驱动的煤矿工作面协同管控技术；在云侧，采用数字孪生技术实现可视化推演，基于Delphi理论和深度学习模型实现煤矿重大灾害安全态势分析，设计灾害环境下的路径时变网络路径规划算法。基于该技术体系开发了煤矿灾害融合监控预警与管控数字化决策平台，并在平顶山天安煤业股份有限公司十二矿成功应用，显著提高了多灾种风险分析决策效率与智能管控水平。

Constructing an intelligent and precise disaster management system for coal mines helps improve the accuracy of accident prediction and early warning, enabling intelligent risk assessment. Existing research mainly focuses on monitoring and early warning technologies or algorithm optimization for individual disasters, lacking a coordinated mechanism for multi-disaster monitoring and early warning, as well as disaster avoidance path planning under emergency response. Additionally, data transmission latency is high, and management efficiency remains low. An integrated "cloud-edge-end" intelligent and precise management and control technology system for coal mine disasters was proposed in this study. The system architecture and the data flow and interaction mechanism for monitoring and early warning were introduced. Key technologies were analyzed from three perspectives: precise data perception, edge computing, and cloud platform. On the end side, intelligent sensors for multiple disasters, including gas, fire, dust, and roof hazards, were developed. A high-speed, low-latency communication network based on IPv6 and a 5G+4G+WiFi6 framework was established, and the deployment of sensing devices and linked control equipment was optimized. On the edge side, a coal mine major disaster data fusion analysis model based on the deep learning AdaTT model was developed. AI-powered video analysis devices for mining applications were designed to enable image-based hazard identification. Additionally, a coalface collaborative management and control technology driven by edge computing was developed. On the cloud side, digital twin technology was applied for visual simulation, while coal mine major disaster safety situation analysis was conducted using Delphi theory and deep learning models. Furthermore, a time-varying network path planning algorithm was designed for disaster environments. Based on the technical system, a coal mine disaster fusion monitoring and intelligent decision-making platform was developed and successfully applied at the No.12 Mine, Pingdingshan Tian'an Coal Mining Co., Ltd. The platform significantly improves the efficiency of multi-disaster risk analysis decision-making and the level of intelligent management and control.