【作者】 毕有益；

【导师】 孙渊；

【作者基本信息】 长安大学 ， 地球探测与信息技术， 2010， 硕士

【摘要】 井间地震是油气田勘探开发的一项关键技术,能够实现两井之间构造和储层等地质目标的高精度成像。井间地震波场十分丰富且十分复杂,而丰富、复杂的波场信息必然会加大波场分离的难度,干扰了人们对有效波的识别,同时严重影响了地震资料的处理质量。因此,如何从复杂的井间地震波场中,识别和分离出可以用于油气藏精细研究的高分辨率井间反射波场,是地球物理工作者首先要解决的问题。井间地震反射波场分离可以看成是VSP波场分离的一种延伸。与VSP不同,井间地震波场较为复杂,分离的难度较大。本论文首先从运动学和动力学两方面详细分析了井间地震的波场特征；通过借鉴VSP波场分离技术与常规地面地震处理技术,系统概述了实际井间地震反射波场的几种分离技术方法原理；重点研究了SVD滤波、中值滤波和τ-p滤波这三种常用的波场分离技术。在此基础上,分析探讨了该三种方法在算法上存在的问题,给出了相应的解决办法,并编程实现了算法。通过理论模型和实际数据验算,结果表明了该三种方法在井间地震波场分离应用中的可行性。其中针对SVD主要用豪斯赫尔德(Householder)变换及变形QR算法对一般实矩阵进行奇异值分解,再通过对奇异值曲线设计带通滤波器来实现滤波；中值滤波实现了三种滤波方式：空间方向中值滤波、时间方向中值滤波及径向中值滤波,针对中值滤波边界处理的局限性,对边界做了拓边充零处理；τ-p滤波主要使用线性radon变换算法,在τ_p域内实现了滤波区域的鼠标选择,并对区域边界做了e指数衰减处理,同时针对线性radon变换假频及端点效应使用了相关函数、速度滤波法进行压制,最后对滤波后数据的振幅及能量进行了恢复校正。经过理论模型数据及实际资料数据的验算表明：SVD滤波对线性相关性高的波场的分离有很好的效果,它基本上不改变分离前后波场分量的能量及振幅相对关系,是保幅的滤波方法；中值滤波对不同方向线性相关性高的波场的分离有很好效果,当参数选取合适时,它能有效的保护信号的细节成分,是一种较为保幅的滤波方法；τ-p滤波是不保幅的滤波方法,它改变了波场能量及振幅相对关系,需要对滤后数据进行振幅校正,经过对振幅处理后,保真度得到很大改善,适合于分离具不同视速度线性或双曲线性波场。更多还原

【Abstract】 Cross-well seismic which can achieve the high-precision imaging of the structure and reservoir between two wells is a key technology in exploration and oil development. Cross-well seismic wave field is very rich and complex, which will increase the difficulty of the wave field separation and thus interferes with the recognition of significant wave, at the same time it also seriously affects the quality of seismic data processing. Therefore, how to distinguish and isolate the high-resolution reflected wave field from the complex cross-well seismic wave field which can be used to study the fine structure of the reservoir is the first question for geophysicists to solve.The separation of the cross-well seismic wave field can be regarded as an extension of the VSP wave field separation. It’s distinct from VSP wave field separation that the cross-well seismic wave field is relatively complex and difficult to separate. This paper firstly in detail analyzed the characteristics of the cross-well seismic wave field from two aspects of kinematics and dynamics and then summarized the principle of some cross-well seismic reflection wave field separation methods which focus on SVD filtering, median filtering andτ-p filtering using VSP wave field separation and conventional seismic processing technology. On this basis, we also discussed some existing problems about the algorithm of the three methods and presented the corresponding solution. We finally implemented the algorithm by programming. Through the theoretical model and actual data calculation, the result shows that the three kinds of methods are feasible in cross-hole seismic separation application.For SVD filtering, we carried out the singular value decomposition for general real matrix mainly using Householder transforming and deformation QR method and then after that realized the filter through designing band-pass filter for singular value curve. The median filter achieved three ways for filtering:space direction median filtering, time direction median filtering and radial median filtering. To the limitation of the boundary treatment in median filter, we took zeroized edge extension processing. Linear radon transforming method is mainly used forτ-p filtering. In theτ-p domain, mouse can be used to select the filtering area whose boundary was simultaneously processed by e-exponent; furthermore, we used correlation function and velocity filtering to suppress the alias and endpoint effect in linear radon transforming and finally recovered the amplitude and energy of the post filtered data.The checking calculation of theoretical model data and practical data shows that SVD filter belong to amplitude-retaining filtering method, and the former has good effect in the separation of the wave field with higher correlation in horizontal orientation while the latter in different orientations, it basically don’t change the relative relations of the separation of the wave field energy and amplitude; the Median filter has good effect in the separation of the wave field with higher correlation in different orientations,it can effectively protect the details of the signal, it’s a amplitude-retaining filtering method. Theτ-p filter is not an amplitude-retaining filtering method, it’s suitable for separating the linear and hyperbolic wave field with different apparent velocities, it changed the relative relations of the separation of the wave field energy and amplitude, so it needs correction of amplitude after filtering data, and then the amplitude fidelity are improved greatly.更多还原