摘要
为了准确而迅速地拾取大量地震事件的P波初至,将深度学习方法引入微地震P波初至到时拾取研究中,对卷积神经网络的结构进行改造,以便适应地震波形数据的特点和P波初至拾取的要求。该算法只需要输入10s窗口的三分量地震波形数据,就可以自动地判定P波初至时刻,无需扫描连续波形,运算时间远远小于长短窗、模板匹配等传统方法。使用该算法训练汶川地震主震后2008年7—8月7467条人工拾取的余震P波初至到时,将得到的模型对测试集中1867条数据的计算结果与人工拾取结果对比,误差小于0.5 s者占比达到98.9%。在低信噪比条件下,该方法仍能保持较好的拾取能力。
In order to accurately and quickly pick up P-wave first-arrival of a large number of seismic events,deep learning method is introduced into the micro seismic P-wave first-arrival picking problem. The structure of convolution neural network is adjusted to apply to the characteristics of the seismic waveform data and first-arrival picking problem. The algorithm takes a 10s-window three-component seismic waveform data as input instead of scanning the continuous waveform. So the running time is far less than traditional methods such as STA/LTA and template matching. The algorithm is applied to aftershocks of 2008 Wenchuan earthquake in July and August,using 7467 manual picked first-arrival data as training dataset. Among the 1867 testing data, 98.9% of the P arrival times picked using this algorithm have an error less than 0.5s compare to the results picked manually. This method can still maintain good pick-up capability under the condition of low signal-to-noise ratio.
引文
[1]Satriano C,Wu Y M,Zollo A,et al.Earthquake early warning:concepts,methods and physical grounds.Soil Dynamics and Earthquake Engineering,2011,31(2):106?118
[2]Stevenson P R.Microearthquakes at Flathead Lake,Montana:a study using automatic earthquake processing.Bulletin of the Seismological Society of America,1976,66(1):61?80
[3]Earle P S,Shearer P M.Characterization of global seismograms using an automatic-picking algorithm.Bulletin of the Seismological Society of America,1994,84(2):366?376
[4]Liu H,Zhang J.STA/LTA algorithm analysis and improvement of microseismic signal automatic detection.Progress in Geophysics,2014,29(4):1708?1714
[5]Gibbons S J,Ringdal F.The detection of low magnitude seismic events using array-based wave-form correlation.Geophysical Journal International,2006,165(1):149?166
[6]Zhang J,Tang Y,Li H.STA/LTA fractal dimension algorithm of detecting the P-wave arrival.Bulletin of the Seismological Society of America,2017,108(1):230?237
[7]张范民,李清河.利用人工神经网络理论对地震信号及地震震相进行识别.地震工程学报,1998,20(4):43?49
[8]郑秀芬,欧阳飚,张东宁,等.“国家数字测震台网数据备份中心”技术系统建设及其对汶川大地震研究的数据支撑.地球物理学报,2009,52(5):1412?1417
[9]Ripley B D.Pattern recognition and neural networks.Technometrics,1999,39(2):233?234
[10]McCulloch W S,Pitts W.A logical calculus of the ideas immanent in nervous activity.Bulletin of Mathematical Biology,1990,52(1/2):99?115
[11]LeCun Y,Bengio Y.The handbook of brain theory and neural networks.Cambridge:MIT Press,1995,255?258
[12]Perol T,Gharbi M,Denolle M.Convolutional neural network for earthquake detection and location.Science Advances,2018,4(2):e1700578
[13]LeCun Y,Bottou L,Bengio Y,et al.Gradient-based learning applied to document recognition.Proceedings of the IEEE,1998,86(11):2278?2324
[14]Springenberg J T,Dosovitskiy A,Brox T,et al.Striving for simplicity:the all convolutional net[EB/OL].(2014?12?21)[2018?04?13].https://arxiv.org/pdf/1412.6806.pdf
[15]Glorot X,Bordes A,Bengio Y.Deep sparse rectifier neural networks//Proceedings of the Fourteenth International Conference on Artificial Intelligence and Statistics.Crete,2011:315?323
[16]刘希强,周蕙兰.用于三分向记录震相识别的小波变换方法.地震学报,2000,22(2):125?131
[17]Chakraborty A,Okaya D.Frequency-time decomposition of seismic data using wavelet-based methods.Geophisics,1995,60(6):1906?1916
[18]Farge M.Wavelet transforms and their applications to turbulence.Annual Review of Fluid Mechanics,1992,24(1):395?458
[19]Grinsted A,Moore J C,Jevrejeva S.Application of the cross wavelet transform and wavelet coherence to geophysical time series.Nonlinear Processes in Geophysics,2004,11(5/6):561?566
[20]Torrence C,Compo G P.A practical guide to wavelet analysis.Bulletin of the American Meteorological Society,1998,79(1):61?78