Title

Picking the P-phase first arrival of microseismic data with strong noise

单位

1.西南科技大学核废物与环境安全国防重点学科实验室,四川绵阳621010;2.成都理工大学地质灾害防治与地质环境保护国家重点实验室,四川成都610059;3.四川轻化工大学,四川自贡643002

Organization

1.Fundamental Science on Nuclear Wastes and Environmental Safety Laboratory,Southwest University of Science and Technology,Mianyang 621010,China;2.State Key Laboratory of Geohazard and Prevention and Geoenvironment Protection,Chengdu University of Technology,Chengdu 610059,China;3.Sichuan University of Science & Engineering,Zigong 643002,China

摘要

针对水力压裂微地震监测信号信噪比普遍偏低,难以准确拾取初至的问题,提出了一种基于特征函数构建的峰度和小波多尺度分解的P震相初至精确拾取方法。首先利用小波多尺度分解法提取低信噪比微地震数据的主成分,进而构建针对主成分数据的特征函数,并计算该特征函数序列的峰度值,最终将峰度曲线的全局最大斜率定义为P震相的初至。与传统峰度法、小波分解和高阶统计量联合方法相比,该方法能够显著减小拾取误差。将该方法应用于不同信噪比的模拟微地震数据的P震相初至拾取,结果表明:其拾取误差为0.0302×10^-3～1.3002×10^-3 s,同时,相比于小波分解与高阶统计量联合方法,其计算效率稍有提高。将该方法应用于实测微地震数据的P震相初至拾取的结果表明,与人工拾取和传统峰度法相比,拾取结果更接近于人工拾取结果,具有更高的准确率。

Abstract

The arrival time of P waves in hydro-fracture microseismic monitoring is difficult to identify due to its low signal-to-noise ratio(SNR).A method which combines the wavelet multilevel analysis and an improved kurtosis was proposed.Firstly,the wavelet multilevel analysis was employed to extract the dominant signal of microseismic recordings with low SNR,and the characteristic function for the dominant effective signal was constructed.Then,the kurtosis of the sequence of the characteristic function was calculated.The maximum slope point of the kurtosis curve was defined as the first arrival of the P phase.Compared with the kurtosis picker and the automatic detection method of microseismic signals based on wavelet decomposition and high-order statistics,the proposed method can significantly reduce the picking error.Tests on synthetic microseismic recordings with different SNR indicated a picking error of the proposed method between 0.0302×10^-3 s and 1.3002×10^-3 s.Moreover,the proposed method has a higher computing efficiency compared to that of existing pickup methods.The method was also tested on actual data,and the results showed that the first arrival time identified by the proposed method was closer to the time identified by manual pickup,thereby resulting in higher accuracy with respect to other existing methods.

关键词：

水力压裂微地震监测; P震相初至拾取; 小波多尺度分解; 峰度法; 特征函数;

Keywords:

microseismic monitoring for hydro-fracturing;; P-phase first arrival time pickup;; wavelet multilevel analysis;; kurtosis;; characteristic function;

基金项目

国家自然科学基金项目(41774118,41604088,41604153)和四川省科技厅项目(2017JY0006,2019YFG0294)共同资助。

DOI

10.3969/j.issn.1000-1441.2020.03.004

一种强噪声微地震信号P震相初至拾取的新方法