针对智能矿山火灾防控子系统建设过程中数据采集及执行层不全面、边缘计算功能未体现、 综合防控平台建设标准不明确等问题，对照国家及地方煤矿智能化建设标准中火灾防控相关条款， 提出了基于“云-边-端”协同的架构体系。增设光纤测温、MEMS 温度传感器、巡检机器人等火灾数 据感知装置及灌浆、注氮、自动洒水、巷道快速封闭装置等火灾执行控制装置，形成数据感知与执 行控制层，负责感知火灾数据信息以及执行云计算中心或边缘侧下发的控制命令；增加边缘计算装 置，形成边缘计算层，实现对现场防灭火装置的联动控制；以“智能矿山基础信息平台”为中台底座， 建立智能矿山火灾防控业务应用，形成平台防控层，融合专家知识库，对火灾防控关键指标进行专 业分析。形成云平台层智能分析、边缘层就近分析控制、感知执行层数据采集及执行的综合防控体 系，且辩证分析了“云−边−端”火灾防控体系三层之间的对立统一关系。探讨了无线低功耗火灾传感 器、巡检机器人、火灾危险区域监测点部署优化、边缘侧配套物理装置、边缘侧火灾预警算法、决 策边缘计算算法、边缘节点部署策略、综合防控平台、火灾数据治理、火灾预警及决策模型等关键 装置及技术。 

In view of the problems such as incomplete data collection and implementation layer, non embodiment of edge computing function, and unclear construction standards of integrated prevention and control platform during the construction of intelligent mine fire prevention and control subsystem, an architecture model based on “cloud-edge-end” collaboration was proposed in comparison with relevant provisions of national and local coal mine intelligent construction standards for fire prevention and control. Fire data sensing devices such as optical fiber temperature measurement, MEMS temperature sensors and inspection robot, as well as fire execution control devices such as grouting, nitrogen injection, automatic sprinkler and roadway rapid closure device were added to form the data sensing and execution control layer, which is responsible for sensing fire data information and executing control commands issued by cloud computing center or edge side; Edge computing devices were added to form the edge computing layer to realize the linkage control of on-site fire prevention and extinguishing devices; with the “Intelligent Mine Basic Information Platform” as the middle platform base, intelligent mine fire prevention and control business application was established to form a platform prevention and control layer, expert knowledge base was integrated, and professional analysis on key indicators of fire prevention and control were conducted. A comprehensive prevention and control system for intelligent analysis at the cloud platform layer, nearby analysis and control at the edge layer, and data collection and execution at the perception and execution layer were formed. The opposition and unity relationship among the three layers of “cloud-edge-end” fire prevention and control system were dialectically analyzed. Key devices and technologies such as wireless low-power fire sensor, patrol robot, monitoring point deployment optimization in fire hazard area, edge side supporting physical devices, edge side fire early warning algorithm, decision edge calculation algorithm, edge node deployment strategy, comprehensive prevention and control platform, fire data governance, fire early warning and decision model were discussed.