【作者】 覃天；

【导师】 王彦春；

【作者基本信息】 中国地质大学（北京） ， 地球物理工程， 2007， 博士

【摘要】 获得零偏移距剖面是地震数据处理中很重要的一个中间环节。常规获得零偏移距剖面的方法有共中心点(CMP)叠加、倾角时差校正(DMO)叠加、叠前偏移等,这些方法要求知道宏观的速度模型,并且其成像结果好坏依赖于速度模型的精确程度。上世纪八十年代末九十年代初,一些地球物理学家着手建立与速度无关的旅行时公式,希望用更多参数来描述界面与走时的关系。代表性的是德国Karlsurhe大学Hubral教授九十年代末提出的共反射面叠加(CRS)方法。共反射面叠加的理论基础是射线理论。它利用共反射点道集一个邻域内(菲涅尔带)道之间的相关性,并将相干区域内道集的能量相加来增强地震信号能量,并借助于相邻CMP道集数据形成CRS超道集,利用超道集的高覆盖次数来压制噪声最终得到高信噪比的零偏移距叠加剖面。该方法可大幅度地提高地震资料信噪比和分辨率,被视为今后复杂地区地震资料处理方法的重要发展途径。CRS叠加在提供高质量的叠加剖面的同时,还提供了与之相关的运动波场参数。针对这些重要的波场参数,本文主要做了以下几方面的工作:1)研究基于二维CRS叠加及其属性的速度反演。在二维情况下,CRS叠加的三个波场属性参数都是特征波传播矩阵的相关参数,与速度参数必然有内在联系。通过研究,建立了与之对应的关系式,从而反演得到地下介质的速度模型。2)研究基于CRS叠加的剩余静校正方法。该方法具体实现途径为:首先通过CRS获得零偏移距剖面,再将零偏移距剖面的每一道作为互相关的模型道,对CRS超道集进行动校正。然后用超道集动校后的每道与零偏移距剖面的相应道之间进行互相关,再将所有属于相同炮或相同检波点的相关结果叠加,并用叠加能量最大值对应的时移量确定剩余静校正值。3)提出了CRS速度反演与叠前偏移联合建立速度模型的方法,以改善叠前偏移速度的分析的精度,缩短资料的处理周期。同时,将得到的速度模型,结合叠后偏移方法,得到基于CRS叠加的叠后偏移结果。理论模型及实际资料处理表明,基于CRS叠加的速度反演能得到较为精确的速度场,为后续的偏移处理、AVO分析提供良好的基础资料;基于CRS叠加的剩余静校正方法对于低信噪比资料处理也非常有益,能够改善常规速度谱的质量,提高速度分析的精度,增强反射波同相轴的连续性。更多还原

【Abstract】 Zero-offset (ZO) section is an important intermediate result in seismic reflection imaging. Common-midpoint (CMP) stack, normal-moveout correction/dip-moveout correction/stack(NMO/DMO/stack), as well as pre-stack migration are the main methods to obtain zero-offset section in conventional seismic processing. These conventional methods require macro-velocity model of subsurface and, the image quality are depended on the accuracy of the velocity model. In the eighties of last century, new stacking techniques have been established which yield better stacking results than the conventional methods mentioned above. They used several kinematic wavefield attributes instead of only one in NMO/DMO/stack. On behalf of the new type of multi-coverage method, the common reflection surface (CRS), which developed by Prof. Hubral of Karlsurhe university in Germany, has been widely accepted by geophysicists.Common reflection surface stack is a macro-model independent seismic imaging method and takes also the local curvature of the reflector at the reflection point into account. Based on the similarity of common reflection point(CRP) trace gathers in one coherent zone (Fresnel zone), CRS stack effectively improves S/N ratio by using more CMP trace gathers to stack. Compared with conventional CMP stack and DMO stack , CRS stack can focus more energy in the vicinity of the reflector. It not only improves the quality of simulated zero-offset section and S/N ratio in deeper layers but also provides important seismic three-parameters section which can be used for inversion of a macro-velocity model and for depth image. It is regarded as one important method of seismic data processing.In this paper, we have studied several forward and inverse problems by means of CRS stack and its kinematic wavefield attributes.Firstly, we developed several methods to reconstruct macro-velocity model in 2D seismic datasets. The CRS stack makes full use of the multi-coverage seismic reflection data and provides additional traveltime parameters. These parameters are very useful for the extraction of further attributes of the seismic medium or for an inversion of a meaningful subsurface velocity model. We can obtain three kinematic wavefield attributes through 2D CRS stacking. These attributes, expressed in terms of wavefront curvatures and emergence angle, can be combined to estimate the RMS velocities and/or interval velocities within the illuminated part of the subsurface model.Secondly, a new residual static correction approach by means of CRS attributes is presented. We considered to make use of the CRS stack which provides additional information about the subsurface by means of kinematic wavefield attributes. The new approach uses the CRS attributes for the moveout correction and the CRS stacked ZO traces as pilot traces. The result of the cross correlation of each pre-stack trace with the pilot trace is assigned to the corresponding source and receiver locations. This new approach is based on the stack power maximization method. In general, the maximum of this cross correlation stack corresponds to the surfaceconsistent residual static time shift of the respective source or receiver. Thus, one of the advantages of this method is that it makes use of more information of the seismic multi-coverage reflection data.Finally, we combined velocity inversion by means of CRS attributes with prestack migration to establish initial macro-velocity model, replacing the conventional velocity analysis and a Dix inversion in a seismic processing routine. The obtained velocity model can be used in a subsequent prestack migration or post-stack migration to determine a depth/time domain image of the subsurface. Otherwise, the projected Fresnel zone based on the kinematic wavefield attributes can be used to to determine an optimal migration aperture in Kirchhoff migration. As is shown, not only the prestack depth migration but also the poststack migration benefits from this approach.Application of the comprehensive processing methods by means of CRS on synthetic examples and field seismic records, the results of these methods show an excellent performance of the algorithm both in accuracy and efficiency. They can provide high accuracy of the velocity model for AVO, and improve the quality of velocity spectra, enhance the continuity of reflection events and the signal-to-noise (S/N) ratio after stacking.更多还原