基于自适应轮动归类的微震震相识别方法  

Microseismic phase identification method based on adaptive wheel-motion classification method

作　　者：贾宝新[1] 李峰 JIA BaoXin;LI Feng(School of Civil Engineering,Liaoning Technical University,Fuxin Liaoning 123000,China)

机构地区：[1]辽宁工程技术大学土木工程学院,辽宁阜新123000

出　　处：《地球物理学报》2023年第2期731-745,共15页Chinese Journal of Geophysics

基　　金：国家自然科学基金项目(51774173);辽宁工程技术大学学科创新团队资助项目(LNTU20TD08);辽宁省“兴辽英才计划”项目(XLYC2007163);辽宁“百千万人才工程”培养经费资助。

摘　　要：震相识别为震源定位提供了可获取必需初始数据的微震信号片段.基于自适应高通滤波、背景噪声幅值上下界计算、轮动圆半径计算及轮动迭代、超限点归类分组4个信号处理流程提出了自适应轮动归类法,并分析了4个初始参数的影响程度及其取值依据.通过计算背景噪声与微震响应的特征值提出了用于判定微震信号有效性的微震信号信噪比(SNRmss),将SNRmss与信号长度分别同震相识别方法的计算速度与识别偏差进行了相关性分析.对比分析了模型试验与实际工程下该方法与改进长短时窗法(STA/LTA)在识别准确率、识别稳定性、计算速度、计算稳定性等方面的优劣.结果表明:相较改进STA/LTA方法,自适应轮动归类法对长度不同信号的识别容错率与稳定性更高.在模型试验下,自适应轮动归类法的识别准确率提高了25.0%,识别偏差标准差为前者的17.4%,计算时间平均值、标准差分别为前者的44.2%、67.4%;实际工程中,自适应轮动归类法的识别准确率提高了4.8%,识别偏差标准差为前者的34.2%,计算时间平均值、标准差分别为前者的40.9%、55.9%.Microseismic phase identification provides microseismic signal segments that can obtain necessary initial data for source location. Based on the four signal processing processes of adaptive high-pass filtering, calculation of upper and lower bounds of background noise amplitude, calculation of wheel-motion circle radius and wheel-motion iteration and classification and grouping of overrun points, the adaptive wheel-motion classification(AWC) method is proposed. On the basis of this method, the influence degree and value basis of the 4 initial parameters are analyzed. By calculating the eigenvalues of background noise and microseismic response, the signal-to-noise ratio of microseismic signal(SNRmss) for judging the effectiveness of microseismic signal is proposed. The correlations between the calculation speed, identification deviation of microseismic phase identification methods and SNRmss, signal length are analyzed respectively. The advantages and disadvantages of this method and the improved STA/LTA method in identification accuracy, identification stability, calculation speed and calculation stability under the model test and practical engineering are compared and analyzed. The results show that compared with the improved STA/LTA method, the AWC method has higher fault tolerance and stability for the signals with different lengths. Under the model test, the identification accuracy of the AWC method is improved by 25.0%, the standard deviation of identification deviation is 17.4% of the former, and the average of calculation time and standard deviation are 44.2% and 67.4% of the former respectively;Under the practical engineering, the identification accuracy of the AWC method is improved by 4.8%, the standard deviation of identification deviation is 34.2% of the former, and the average of calculation time and standard deviation are 40.9% and 55.9% of the former respectively.

关 键 词：微震监测 震相识别 模型试验 自适应轮动归类法 改进STA/LTA方法 

分 类 号：P631[天文地球—地质矿产勘探]