地下水位变化被认为是地震前兆的重要潜在信号,研究其与地震活动之间的关系对地震预测具有重要意义。为识别地下水位的异常变化特征,提出了一种基于TCN-GRU模型的异常检测方法,并结合EWMA控制图精确定位异常发生时间。实验结果表明,TCN-GRU模型对异常波动最为敏感,具有显著的鲁棒性和实时检测能力,能够适应不同井条件下的复杂变化。研究揭示了地下水位异常与地震活动的密切关联,为地震前兆信号的早期识别提供了科学依据,对地震预测与减灾防灾具有重要应用价值。
Abstract
The change in groundwater level is considered an important potential signal for earthquake precursors, and studying its relationship with seismic activity is of great significance for earthquake prediction. An anomaly detection method based on TCN-GRU model is proposed to identify the abnormal changes in groundwater level, and combined with EWMA control chart to accurately locate the time of anomaly occurrence. The experimental results show that the TCN-GRU model is most sensitive to abnormal fluctuations, has significant robustness and real-time detection ability, and can adapt to complex changes under different well conditions. The study reveals the close relationship between abnormal groundwater level and seismic activity, providing scientific basis for early identification of earthquake precursor signals and having important application value for earthquake prediction and disaster reduction.
关键词
TCN-GRU /
深度学习 /
地震前兆 /
异常检测 /
EWMA控制图
Key words
TCN-GRU /
deep learning /
earthquake precursors /
anomaly detection /
EWMA control chart
{{custom_sec.title}}
{{custom_sec.title}}
{{custom_sec.content}}
参考文献
[1] 车用太,刘成龙,鱼金子,等.汶川M_S 8.0地震的地下流体与宏观异常及地震预测问题的思考[J].地震地质,2008,30(4):828-838.
[2] 余维,王博,陈真林.MODFLOW在井灌区地下水数值模拟中的应用[J].中国农村水利水电,2006(11):17~21.
[3] 杨玮. 基于GM(1,1)与ARIMA的长江干线航道水位预测研究[D].武汉:武汉理工大学,2018.
[4] Yu P S,Chen S T,Chang I F.Support vector regression for real-time flood stage forecasting[J].Journal of Hydrology,2006,328(3-4):704-716.
[5] He Y,Zhao J.Temporal Convolutional Networks for Anomaly Detection in Time Series[J].Journal of Physics:Conference Series,2019,1213(4):042050.
[6] 杨丽,吴雨茜,王俊丽,等.循环神经网络研究综述[J].计算机应用,2018,38(S2):1-6,26.
[7] 薛红盼,张乐,陆丽娜,等.夏垫断裂带地震地下流体的同震响应研究[J].地震,2020,40(1):184-202.
[8] 尹宏伟,梁丽环,韩文英,等.河北省数字水位短临异常特征研究——以2016年唐山4.2级地震为例[J].地震地磁观测与研究,2017,38(4):119-125.
[9] Biagi P,Ermini A,Cozzi E,Khatkevich Y,Gordeev E.Hydrogeochemical Precursors in Kamchatka (Russia) Related to the Strongest Earthquakes in 1988-1997[J].Natural Hazards,2000,21(2):263-276.
[10] 赵晗清,陈新房,杨丽佳,等.基于LSTM的地下水水位预测及地震前兆异常分析[J].电脑与电信,2024(6):68-72.
[11] 郑麟露. 基于Lp深度的EWMA控制图研究与应用[D].成都:四川师范大学,2024.
基金
廊坊市科学技术研究与发展计划自筹经费项目,项目编号：2024011018; 河北省教育厅科学研究项目,项目编号：ZC2023108
