摘要
针对冲击地压的超前预测问题,提出了一种基于深度学习的微震能量信号时间序列预测方法。该方法构建了多种深度神经网络模型,通过对大量微震信号数据进行分析,优化模型参数,实现了对微震信号的高效预测。实验结果表明,预测信号与后续现场采集的真实信号相关性达95%以上,验证了该方法的有效性。研究为冲击地压的定量监测与预警提供了新的数据支持,具有重要的实用价值和创新意义。
A time series prediction method for microseismic energy signals based on deep learning was proposed to address the issue of advanced prediction of rock burst.This method constructed various deep neural network models,the model parameters were optimized through the analysis of a large volume of microseismic signal data,achieving the efficient prediction of microseismic signals.Experimental results indicate that the predicted signals have a correlation of over 95%with subsequent field-collected real signals,validating the effectiveness of this method.This research provides new data support for quantitative monitoring and early warning of rock burst,holding significant practical value and innovative significance.
作者
唐晓明
邵华
宿国瑞
高文忠
狄洋阳
付恩三
贾慧霖
Tang Xiaoming;Shao Hua;Su Guorui;Gao Wenzhong;Di Yangyang;Fu Ensan;Jia Huilin(Xuzhuang Coal Mine,Shanghai Datun Energy Co.,Ltd.,Xuzhou 221600,China;Information Research Institute of the Ministry of Emergency Management,Beijing 100029,China;School of Materials Engineering,Changshu Institute of Technology,Suzhou 215506,China;School of Safety Engineering,China University of Mining and Technology,Xuzhou 221116,China)
出处
《能源与环保》
2025年第3期85-93,共9页
CHINA ENERGY AND ENVIRONMENTAL PROTECTION
基金
国家重点研发计划资助项目(2022YFC3004702)。
关键词
冲击地压
微震
超前预测
深度学习
rock burst
microseismic
advanced prediction
deep learning
