基于GRU神经网络的矿井工作面瓦斯突出预警模型研究  

Research on a GRU Neural Network-based Gas Emission Warning Model for Mine Working Faces

作　　者：龚选平 杨鹏 GONG Xuanping;YANG Peng(China Coal Energy Research Institute Co.,Ltd.,Xi’an,Shaanxi 710054)

机构地区：[1]中煤能源研究院有限责任公司,陕西西安710054

出　　处：《中国煤炭地质》2025年第2期67-73,40,共8页Coal Geology of China

基　　金：国家重点研发计划“储层地质参数高精度测量关键技术研究及应用”(2023YFF0615400)。

摘　　要：瓦斯突出是矿井工作中的一大安全隐患,严重威胁着井下员工的人身安全和生产工作的顺利开展。建立一个可以及时防范瓦斯突出的模型对降低事故风险、保护矿工人身安全有着举足轻重的意义。造成瓦斯突出的原因有许多,这些原因造成瓦斯突出危险性的增加。在实际掘进活动中,把传感器采集到的多种信息加以综合,利用GRU神经网络强大的时间序列数据处理能力和记忆能力,构建出完整的瓦斯突出预警模型。实验结果表明,基于GRU神经网络的瓦斯突出预警模型在预测精度和分级预警效果上均表现出色,能够在关键时刻发出有效的预警信号,为煤矿安全生产提供了有力保障。Gas outburst poses a significant safety hazard in mining operations,gravely threatening the personal safety of underground workers and the smooth progress of production activities.Establishment of a model capable of promptly preventing gas outbursts is of paramount importance in reducing accident risks and safeguarding the lives of miners.Numerous factors contribute to gas outbursts,collectively increasing their risk.During actual excavation activities,by integrating various information gathered by sensors and leveraging the robust time-series data processing and memory capabilities of GRU(Gated Recurrent Unit)neural networks,a comprehensive gas outburst early warning model has been developed.Experimental results demonstrate that this GRU-based gas outburst early warning model excels in both prediction accuracy and graded warning effectiveness,enabling the issuance of effective warning signals at critical moments,thereby providing robust support for safe coal mine production.

关 键 词：GRU神经网络 瓦斯突出 预警模型 影响因素 

分 类 号：TP389.1[自动化与计算机技术—计算机系统结构]