低信噪比微地震监测方法与技术研究

发布时间：2018-07-05 01:15

本文选题：微地震 + 低信噪比　；参考：《长江大学》2015年硕士论文

【摘要】：随着油气勘探开发的进展,传统油气田中后期开发、新兴非常规油气田开采中大量使用水力压裂及微地震监测技术。油气开发中注水、注气、热驱或压裂时,地下岩层产生裂缝或断裂,产生地震波。该项技术通过在邻井中的检波器来监测压裂井在压裂过程中诱发的微地震波来描述压裂过程中裂缝生长的几何形状和空间展布。它能实时提供压裂施工产生裂隙的高度、长度和方位角,利用这些信息可以优化压裂设计、优化井网或其他油田开发措施,从而提高采收率。与天然地震和常规地震勘探相比,微地震具有地震强度低、频率高、持续时间短等特点。本文首先阐述微地震形成机理及其特征,并对微地震监测采集、处理、解释等技术进行描述。微地震监测技术中,压裂和采集方案设计、信号处理、波场分离、极化分析、初至拾取、正反演分析等都是重要步骤。在微地震监测中,受仪器设备制约以及强噪音的影响,接收到的微地震信号经常是低信噪比信号,所以微地震信号的处理、微地震事件的识别和初至拾取就成为整个流程中最关键的环节。微地震监测的资料处理和解释方法来源于天然地震和常规地震勘探。本文针对地震记录中噪音产生的原因、类型、特点进行了分析,介绍了常用的地震信号去噪方法。本文对K-L变换处理微地震资料的方法进行了探讨,并从傅里叶变换出发,对传统的时频分析方法进行综述,分别从从短时傅里叶变换和小波变换推导S变换,总结其特性：与傅里叶变换直接联系,其过程无损可逆；介于小波变换和短时傅里叶变换之间,是线性时频表示方法,没有交叉项的干扰；分辨率与信号频率直接相关；基本小波不用考虑容许性条件限制。本文对这些时频分析方法进行综合比较,S变换的时频分辨率在线性变换中有明显的优势,相对于二次型变换,没有交叉项干扰。结合微地震信号的特性,本文选择S变换进行微地震识别和信号重构。随机噪声的不可预测性以及与有效波之间的频谱重叠是微地震信号处理中的一个难点,低信噪比微地震信号的去噪处理中,容易出现信号处理的失真。在信号处理的过程中,信噪比的提高和信号的保真在算法的实现过程中是一个重要的平衡。传统的地震相拾取方法包括能量比法、AIC法、分形分维法、神经网络等。本文详细论述能量比法、AIC法等传统微地震信号初至拾取方法的算法,总结借鉴众多算法的优势和缺陷：能量比法拾取初至简单、有效,但是受时窗长度影响较大；AIC法能精确拾取初至,但是不能对信号中是否包含微地震事件进行有效识别。基于S变换对微地震信号重构获得较高信噪比信号的基础上, 本文设计先用能量比法识别微地震事件、再用AIC法进行精确拾取的两步法对微地震事件进行初至拾取。此方法不受时窗变化影响,能够实现重构后信号的有效和精确拾取。本文针对低信噪比微地震信号,利用S变换进行时频分析重构,能量比-AIC两步法进行拾取,获得更准确的初至信息,对压裂效果进行反演和解释,进而实现对压裂方案的优化。
[Abstract]:With the development of oil and gas exploration and development, in the middle and late development of traditional oil and gas fields, hydraulic fracturing and microseismic monitoring are widely used in the exploitation of new unconventional oil and gas fields. In the course of oil and gas development, water injection, gas injection, thermal flooding or fracturing, the underground rock produces crack or fracture and produces seismic waves. This technology is monitored by the geophone in adjacent wells. Microseismic waves induced by fracturing during fracturing describe the geometry and space distribution of fracture growth during fracturing. It can provide the height, length and azimuth of fractured construction in real time, using these information to optimize the fracturing design, optimize the well network or other oil field development measures, so as to improve the recovery and natural recovery. Compared with conventional seismic exploration, microseismic has the characteristics of low seismic intensity, high frequency and short duration. This paper first describes the mechanism and characteristics of microseismic formation, and describes the techniques of microseismic monitoring collection, processing and interpretation. In microseismic monitoring, the design of fracturing and acquisition schemes, signal processing, wave field separation, and extreme seismic detection are used in microseismic monitoring. In microseismic monitoring, the microseismic signals are often low signal to noise ratio signals, so the processing of microseismic signals, the recognition of microseismic events and the initial pick-up are the most critical links in the whole process. The methods of data processing and interpretation of seismic monitoring are derived from natural earthquakes and conventional seismic exploration. This paper analyzes the causes, types and characteristics of noise produced in seismic records, introduces common seismic signal denoising methods. This paper discusses the method of K-L transformation for processing microseismic data, and starts from Fu Liye transform, The traditional time-frequency analysis methods are summarized. The S transform is derived from the short time Fu Liye transform and the wavelet transform, and their characteristics are summarized. The process is directly connected with Fu Liye transform. The process is nondestructive and reversible; between the wavelet transform and the short-time Fu Liye transform, it is a linear time frequency representation method, without the interference of the cross term; resolution and letter. The frequency of the basic wavelet is directly related; the basic wavelet does not consider the admissibility conditions. This paper makes a comprehensive comparison of these time frequency analysis methods. The time frequency resolution of the S transform has obvious advantages in the linear transformation. Compared with the two type transformation, there is no cross term interference. In combination with the characteristics of the microseismic signal, this paper chooses the S transform to carry out the microseismic recognition. The unpredictability of the signal and the signal reconstruction. The unpredictability of random noise and the overlapping of the spectrum with the effective wave are a difficult point in the processing of the microseismic signal. In the denoising processing of the low signal to noise ratio microseismic signal, it is easy to appear the distortion of signal processing. In the process of signal processing, the enhancement of signal to noise ratio and the fidelity of signal in the process of signal processing are the realization of the algorithm. The traditional method of picking up seismic phase includes energy ratio method, AIC method, fractal fractal dimension method and neural network. This paper discusses the algorithm of the first arrival method of traditional microseismic signal, such as energy ratio method and AIC method, and summarizes the advantages and the deficiency of many algorithms: the energy ratio method is simple and effective, but it is effective, but subject to the energy ratio method. The length of the time window has great influence; the AIC method can pick up the first arrival accurately, but can not effectively identify the micro seismic events in the signal. Based on the S transform to obtain the signal to the high signal to noise ratio for the reconstruction of the microseismic signal, this paper first uses the energy ratio method to identify the microseismic events and then the two step method of picking up accurately with the method of AIC. This method is not affected by the change of the time window and can realize the effective and accurate pick-up of the reconstructed signal. In this paper, the S transform is used to reconstruct the time frequency analysis of the low signal to noise ratio microseismic signal, and the energy is picked up by the -AIC two step method, and the accurate initial information is obtained, and the fracturing effect is retrieved and the results are retrieved. In addition, the optimization of the fracturing scheme is realized.
【学位授予单位】：长江大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：P631.4

【参考文献】