低信噪比微地震事件辨识与震相初至自动拾取方法

发布时间：2018-07-05 08:24

本文选题：微地震事件辨识 + 震相初至　；参考：《山东科技大学》2017年硕士论文

【摘要】：基于声发射学、地震学发展而来的微地震监测技术是一种针对人类开采活动中发生的渺小地震事件进行采集、观测,进而分析出人类活动对地下岩层状态产生的影响以及影响范围的地球物理勘探技术。微地震监测技术已经广泛应用于冲击地压、矿井突水等矿井动力灾害监测预警,并已取得许多研究成果。微地震有效事件辨识及震相到时拾取是微地震监测技术的核心问题之一,把微地震事件从海量震动监测数据中识别出来,并进行震相初至时刻的准确拾取是数据处理的首要步骤,也是震源定位研究、震源参数反演研究和震源机制解释研究的基础与条件。微地震监测系统采集的微震数据具有非平稳性、多样性等特点,由于受到采集现场机械震动、电磁噪声、岩石破裂、爆破振动等多种外界因素的影响,采集的数据信噪比较低。常规的微地震事件辨识和震相初至拾取方法不能有效处理低信噪比信号。研究表明微地震信号在时频域较为稀疏,基于此,本文提出了一种新的微地震事件辨识和震相到时自动拾取方法,简称为S-AIC方法。该方法首先对采集的信号应用小波变换进行时频分析,再利用Renyi计算函数计量时频分析结果,根据熵值判定信号中是否存在微地震有效事件;其次,对含有微地震事件的监测数据应用基于小波变换的分层阈值法降噪;再次,基于时频分析法计算出震相到时的大致位置,并为赤池信息准则(AIC)算法选择合适的时窗;最后,在选择的时窗长度内利用AIC算法计算出精确的震相到时。论文中使用程序对微地震数据进行了实验,通过对比发现:以人工辨识结果为参考,误差在0-20ms范围内判定为准确辨识,本文方法对高信噪比信号的辨识准确率较高,为100%,对低信噪比信号的辨识准确率达到了 92%,优于常规辨识方法。应用AIC方法和S-AIC方法对抽选的高信噪比和低信噪比两组数据共100个微地震信号进行了震相初至拾取实验,以人工拾取结果为参考,误差在0-10ms范围内判定为准确拾取,S-AIC方法对高信噪比信号拾取准确率为100%,耗时0.812s;对低信噪比信号的拾取准确率为91%,平均耗时1.325s。
[Abstract]:The microseismic monitoring technology based on acoustic emission and seismology is a kind of acquisition and observation of small earthquake events occurring in human mining activities. Furthermore, the influence of human activities on the state of underground strata and the geophysical exploration technique of the influence range are analyzed. Micro-seismic monitoring technology has been widely used in monitoring and warning of mine dynamic disasters, such as impact ground pressure, mine water inrush and so on, and many research results have been obtained. Effective microseismic event identification and phase arrival detection is one of the core problems of microseismic monitoring technology. The micro-seismic events are identified from massive seismic monitoring data. It is the first step of data processing and the foundation and condition of source location, source parameter inversion and focal mechanism interpretation. The microseismic data collected by the micro-seismic monitoring system have the characteristics of non-stationarity and diversity. Due to the influence of many external factors such as mechanical vibration, electromagnetic noise, rock rupture and blasting vibration, the SNR of the collected data is low. The conventional methods of microseismic event identification and phase first arrival pickup can not effectively deal with low signal-to-noise ratio (SNR) signals. It is shown that the microseismic signal is sparse in time-frequency domain. Based on this, a new method of microseismic event identification and phase arrival automatic pick-up, called S-AIC method, is proposed in this paper. In this method, wavelet transform is applied to time-frequency analysis of collected signals, and then Renyi function is used to calculate the time-frequency analysis results to determine whether there are micro-seismic effective events in the signals according to entropy. For the monitoring data containing micro-seismic events, the hierarchical threshold method based on wavelet transform is used to reduce the noise. Thirdly, the approximate position of phase arrival is calculated based on time-frequency analysis method, and the appropriate time window is selected for the red pool information criterion (AIC) algorithm. The exact phase arrival time is calculated by using the AIC algorithm in the selected window length. In this paper, the microseismic data are tested by the program. It is found that the error is identified accurately in the range of 0-20ms with the reference of manual identification results, and the accuracy of identification of high signal-to-noise ratio (SNR) signals by this method is high. The accuracy of low signal-to-noise ratio (SNR) signal identification is 92, which is superior to the conventional identification method. Using AIC method and S-AIC method, a total of 100 microseismic signals with high SNR and low signal-to-noise ratio (SNR) were tested. The results of manual pickup were used as a reference. The accuracy rate of the S-AIC method is 0.812 s for high SNR signals and 91s for low signal-to-noise ratio signals, with an average time of 1.325 s.
【学位授予单位】：山东科技大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：P631.4

【参考文献】