基于Shearlet相关性的S变换在微地震初至拾取中的应用

发布时间：2018-01-17 19:17

本文关键词：基于Shearlet相关性的S变换在微地震初至拾取中的应用　出处：《吉林大学》2017年硕士论文　论文类型：学位论文

【摘要】：伴随着油气资源的大量消耗,非常规油气藏开采成为新的研究热点,微地震监测作为其主要的技术手段,对低渗透油田开发具有重要的指导作用。初至拾取是微地震监测技术的重要环节,是井中微地震数据处理的基本步骤,自动准确的拾取信号初至是实现震源精确定位的重要前提。由于微地震有效信号能量较弱、数据信噪比较低,传统方法的自动拾取结果不理想,经常发生错拾或漏拾初至的情况,严重影响了后续震源定位的精度,因此,建立一种适用于低信噪比微地震信号初至的到时拾取方法具有重要意义。本文对微震初至拾取展开研究,主要目的在于提高低信噪比下的拾取准确度。对于低信噪比微地震数据,有效信号与噪声间的时域特征差异不明显,难以在时域完成有效的初至拾取。为此,本文将Shearlet变换引入到初至拾取中,通过Shearlet变换将时域微地震数据映射到Shearlet域,基于信号与噪声在Shearlet域高频尺度层的系数差异,实现有效信号点的初步识别,然后利用信号系数在相邻尺度层间的尺度相关性对信号点进行校正,进一步提高信号点识别准确度,最后将校正后的第一个信号点视为初至点。模拟和实际微地震记录的处理结果表明,与传统的长短时窗平均能量比法(STA/LTA)和Akaike信息准则(AIC)算法相比,Shearlet变换法有效改善了低信噪比下的初至识别情况。Shearlet变换法的实现表明了在Shearlet域区分信号与噪声的有效性和可行性。但是该方法中的阈值及校正值等参数会对最终的拾取结果产生一定影响,使得其在实际应用中具有一定限制。为了进一步提高低信噪比下的拾取性能,本文在Shearlet变换基础上结合S变换,提出了基于Shearlet相关性的S变换(SC-ST)初至拾取算法。该算法首先利用相邻高频尺度层间信号Shearlet系数相关性,增强有效信号系数,使得信号与噪声间具有更明显的特征差异,然后对处理后的系数进行S变换,有效聚集信号能量,最后基于信号与噪声间的能量差异完成信号的初至拾取。由于S变换具有良好的时频聚集性,在S变换时频表示图中可以对有效信号进行更直观、更准确的识别,从而使得该算法具有更优的初至拾取性能。为了验证SC-ST算法的有效性,本文对加入高斯白噪声和实际噪声的合成微地震数据进行处理,并且与STA/LTA、AIC和Shearlet变换法进行对比,实验结果表明SC-ST算法在对噪声的适应性和初至拾取准确度上均优于其他方法。此外,本文通过大量实验分析了SC-ST算法在不同信噪比下的初至拾取性能。实际三分量微地震数据的处理结果验证了SC-ST算法在初至拾取上的优越性,它是一种有效的、具有高准确度的初至拾取方法,能够在低信噪比环境下提供准确可靠的初至结果。
[Abstract]:With the large consumption of oil and gas resources, unconventional oil and gas reservoir production has become a new research hotspot, micro-seismic monitoring as its main technical means. It plays an important role in guiding the development of low permeability oil fields. First arrival pickup is an important part of micro-seismic monitoring technology and a basic step of micro-seismic data processing in wells. Automatic and accurate signal pickup is an important prerequisite for the accurate location of the source. Due to the weak energy of the effective signal and the low signal to noise ratio of the data, the traditional automatic pickup method is not ideal. Frequent mispicking or missing initial arrival seriously affects the accuracy of subsequent focal location, therefore. It is of great significance to establish a method for detecting the first arrival of microseismic signals with low SNR. In this paper, the detection of the first arrival of microearthquakes is studied. The main purpose is to improve the picking accuracy at low SNR. For the low-SNR micro-seismic data, there is no obvious difference between the time domain characteristics of the effective signal and noise, so it is difficult to complete the effective first arrival pickup in the time domain. In this paper, the Shearlet transform is introduced into the first arrival pickup, and the time-domain microseismic data are mapped to the Shearlet domain by the Shearlet transform. Based on the difference of the coefficients between the signal and the noise in the high-frequency scale layer in Shearlet domain, the initial recognition of the effective signal points is realized. Then the signal points are corrected by the scale correlation of the signal coefficients between adjacent scale layers to further improve the accuracy of signal point recognition. Finally, the first signal point after correction is regarded as the first arrival point. The processing results of simulation and actual microseismic records show that. Compared with the traditional time-window average energy ratio method (STA-LTA) and Akaike information criterion (Akaike) algorithm. The implementation of the Shearlet transform method to improve the first arrival recognition in low signal-to-noise ratio shows that it is effective and feasible to distinguish signal from noise in Shearlet domain. However, the threshold value and correction value of the method will have a certain impact on the final pick-up results. In order to further improve the pickup performance of low signal-to-noise ratio, this paper combines S transform based on Shearlet transform. A S-transform SC-ST-based first-break pickup algorithm based on Shearlet correlation is proposed. The algorithm firstly uses the correlation of Shearlet coefficients of adjacent high-frequency scale interlayer signals. Enhance the effective signal coefficient, make the signal and noise have more obvious characteristic difference, then the processed coefficient of S transform, effectively gather the signal energy. Finally, based on the energy difference between the signal and the noise, the first arrival of the signal is picked up. Due to the good time-frequency aggregation of S-transform, the effective signal can be visualized in the S-transform time-frequency representation diagram. In order to verify the effectiveness of SC-ST algorithm, the synthetic microseismic data with Gao Si white noise and actual noise are processed in this paper. And compared with the method of STAP / LTA AIC and Shearlet transform. The experimental results show that the SC-ST algorithm is superior to other methods in the adaptability to noise and the accuracy of first-break pickup. This paper analyzes the performance of SC-ST algorithm in different signal-to-noise ratio through a large number of experiments. The processing results of the actual three-component micro-seismic data verify the superiority of SC-ST algorithm in the first-arrival pickup. It is an effective method with high accuracy and can provide accurate and reliable first arrival results in low signal-to-noise ratio (SNR) environment.
【学位授予单位】：吉林大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TE348;P631.4

【相似文献】