地震前兆钻孔应变数据的异常检测对于地震预测具有重要意义。以江苏省地震局提供的钻孔应变时间序列作为实验数据,首先使用趋势分离和残差去噪相结合的预处理方法,有效提取了数据中的长期趋势与关键波动信息,降低了噪声干扰。随后进行LSTM-Attention模型的设计,并将其与LSTM、CNN和AutoEncoder三个模型进行性能对比,对比结果表明LSTM-Attention模型在MSE、MAE、RMSE三项指标中均优于其他模型,最后使用该模型进行异常检测。通过实验验证,LSTM-Attention模型在地震前兆钻孔应变异常检测任务中能够较为准确地捕捉到数据中的异常事件。
Abstract
Anomaly detection in borehole strain data as earthquake precursors is of great significance for earthquake prediction. This paper takes the borehole strain time series provided by the Jiangsu Earthquake Agency as experimental data. Firstly, a preprocessing method combining trend separation and residual denoising is used to effectively extract the long-term trends and critical fluctuation information from the data, thereby reducing noise interference. Subsequently, an LSTM-Attention model is designed and its performance is compared with three other models: LSTM, CNN, and AutoEncoder. The comparison results indicate that the LSTM-Attention model outperforms the other models in terms of MSE, MAE, and RMSE. Finally, this model is used for anomaly detection. Experimental validation demonstrates that the LSTM-Attention model can accurately capture anomalies in the data for the task of anomaly detection in borehole strain precursors of earthquakes.
关键词
地震前兆异常 /
钻孔应变数据 /
时间序列 /
LSTM /
attention
Key words
earthquake precursor anomaly /
borehole strain data /
time series /
LSTM /
attention
{{custom_sec.title}}
{{custom_sec.title}}
{{custom_sec.content}}
参考文献
[1] 韩培锋,田述军,樊晓一,等.芦山地震触发次生灾害特点统计分析及规律预测[J].自然灾害学报,2018,27(1):120-126.
[2] Zhu K,Chi C,Yu Z,et al.Extracting borehole strain precursors associated with the Lushan earthquake through principal component analysis[J].Annals of Geophysics,2018,61:SE447.
[3] L,Telesca,T,et al.Relationship between seismicity and water level in the Enguri high dam area (Georgia) using the singular spectrum analysis[J].Natural Hazards and Earth System Science,2012,12(8):2479-2485.
[4] 邱泽华,唐磊,张宝红,等.用小波-超限率分析提取宁陕台汶川地震体应变异常[J].地球物理学报,2012,55(2):538-546.
[5] 池成全. 钻孔应变前兆观测数据分析与异常提取研究[D].长春:吉林大学,2020.
[6] Yu Z N.Pre-earthquake Anomaly Extraction and Evaluation of Borehole Strain Observations[D].Changchun:Jilin University,2021:1-2, 21-65(in Chinese).
[7] Chi C,Zhu K,Yu Z,et al.Detecting Earthquake-Related Borehole Strain Data Anomalies With Variational Mode Decomposition and Principal Component Analysis:A Case Study of the Wenchuan Earthquake[J].IEEE Access,2019,7:1-1.
[8] Yu Z,Zhu K,Hattori K,et al.Investigation of borehole strain observations based on a state-space model and ApNe analysis associated with the 2013 Lushan earthquake[J].IEEE Access,2021,9:12167-12179.
[9] 石贵飞,王俊,刘杰,等.基于LSTM深度神经网络的TGS地震数据去噪重构在滇中引水工程中的应用[J/OL].水电能源科学,1-6[2024-12-30].DOI:10.20040/j.cnki.1000-7709.2025.20242064.
[10] 何伦. 基于Attention和LSTM循环神经网络的文本情感分析[J].电脑编程技巧与维护,2024(4):31-34.
[11] 王志涛. 基于注意力机制的网络状态与流量预测[J]. 黑龙江科学,2024,15(24):77-79.
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：2024011006
