基于波动方程有限差分算法的接收函数正演与偏移

针对接收函数正演与偏移, 本文采用波动方程有限差分算法. 借鉴成熟的勘探地震学方法, 引入等效速度概念, 建立接收函数转换波与地震勘探反射波的等效走时方程, 实现了基于波动方程有限差分算法的接收函数正演与偏移. 数值计算表明, 波动方程有限差分叠后偏移方法可以对点绕射和穹隆构造模型实现高精度成像. 本文利用数值计算讨论了波动方程有限差分叠后偏移与Kirchhoff叠后偏移对于接收函数偏移的适用性, 还对偏移过程中速度模型的误差进行了分析.     

Estimation of geological dip and reflector curvature from zero-offset seismic reflections in heterogeneous anisotropic media

Depth conversion of selected seismic reflections is a valuable procedure to position key reflectors in depth in a process of constructing or refining a depth-velocity model. The most widespread example of such procedure is the so-called map migration, in which normal-incidence, zero-offset (stacked) seismic data are employed. Since the late seventies and early eighties, under the assumption of an isotropic velocity model, map migration algorithms have been devised to convert traveltime and its first and second derivatives into reflector position, dip and curvatures in depth. In this work we revisit map migration to improve the existing algorithms in the following accounts: (a) We allow for fully anisotropic media; (b) In contrast to simple planar measurement surface, arbitrary topography is allowed, thus enlarging the algorithms applicability and (c) Derivations and results are much simplified upon the use of the methodology of surface-to-surface paraxial matrices.     

REFLECTION AMPLITUDES AND MIGRATION AMPLITUDES (ZERO-OFFSET SITUATION)1

The effect of wave-equation migration on amplitudes is determined. This effect is derived for zero-offset traces and for second-order approximations of the traveltimes. Three steps are followed: firstly, the amplitudes of zero-offset traces are established; secondly minus half the traveltimes are used as input for downward continuation in migration (forward in space and time); thirdly, the amplitudes of the migrated events are determined by downward continuation (at zero-traveltimes). Layered models (piles of homogeneous layers) with smooth interfaces are used. The determinants of the 2 × 2 matrices B ₀ obtained for these models are responsible for the main effect on migration. The migration result primarily depends on the overburden as the inverse of det ( B ₀). Drastic effects can occur over small distances. For weakly reflecting media, it is confirmed that wave-equation migration gives “correct” results (but the input data must be multiplied by V₀T₀), i.e. amplitudes proportional to the reflection coefficient. For any velocity changes, the inverse of det ( B ₀) will, in general, give inaccurate migration amplitudes and inaccurate lithological interpretations. In a simple step, true amplitude migration, or exact migration, is derived from our results. It is assumed that no focus phenomena are present. The effect of buried foci is discussed briefly.     

共炮检距高斯波束偏移的一种快速实现算法

与共炮高斯波束偏移相比,共炮检距高斯波束偏移具有直接抽取炮检距域共成像点道集的优势.过去,共炮检距高斯波束偏移以损失成像精度的代价采用最速下降法来降低积分的维数,从而提高计算效率.但经过最速下降近似简化的偏移公式仍是频率域的,需要在每个频点进行计算.为此,本文提出一种快速实现算法来避免采用最速下降法.本文通过分析一个水平层状速度模型的偏移过程和Marmousi速度模型的成像结果来检验不同插值方法对快速实现算法的成像精度和计算效率的影响,并建议采用二维三次卷积插值方法.同时本文在Marmousi速度模型下验证了快速实现算法相对于最速下降法在成像精度和计算效率上的优势.此外,本文将采用二维三次卷积插值的快速实现算法应用于Sigsbee2A模型并获得了清晰的盐下图像.     

MODEL-BASED STACK: A METHOD FOR CONSTRUCTING AN ACCURATE ZERO-OFFSET SECTION FOR COMPLEX OVERBURDENS1

The interpretation of stacked time sections can produce a correct geological image of the earth in cases when the stack represents a true zero-offset section. This assumption is not valid in the presence of conflicting dips or strong lateral velocity variations. We present a method for constructing a relatively accurate zero-offset section. We refer to this method as model-based stack (MBS), and it is based on the idea of stacking traces within CMP gathers along actual traveltime curves, and not along hyperbolic trajectories as it is done in a conventional stacking process. These theoretical curves are calculated for each CMP gather by tracing rays through a velocity-depth model. The last can be obtained using one of the methods for macromodel estimation. In this study we use the coherence inversion method for the estimation of the macromodel since it has the advantage of not requiring prestack traveltime picking. The MBS represents an accurate zero-offset section in cases where the estimated macromodel is correct. Using the velocity–depth macromodel, the structural inversion can be completed by post-stack depth migration of the MBS.     

Multiwave velocity analysis based on Gaussian beam prestack depth migration

Prestack depth migration of multicomponent seismic data improves the imaging accuracy of subsurface complex geological structures. An accurate velocity field is critical to accurate imaging. Gaussian beam migration was used to perform multicomponent migration velocity analysis of PP- and PS-waves. First, PP- and PS-wave Gaussian beam prestack depth migration algorithms that operate on common-offset gathers are presented to extract offset-domain common-image gathers of PP- and PS-waves. Second, based on the residual moveout equation, the migration velocity fields of P- and S-waves are updated. Depth matching is used to ensure that the depth of the target layers in the PP- and PS-wave migration profiles are consistent, and high-precision P- and S-wave velocities are obtained. Finally, synthetic and field seismic data suggest that the method can be used effectively in multiwave migration velocity analysis.     

MIGRATION TO ZERO OFFSET (DMO) FOR A CONSTANT VELOCITY GRADIENT: AN ANALYTICAL FORMULATION1

Migration to zero offset (MZO) is a prestack partial migration process that transforms finite-offset seismic data into a close approximation to zero-offset data, regardless of the reflector dips that are present in the data. MZO is an important step in the standard processing sequence of seismic data, but is usually restricted to constant velocity media. Thus, most MZO algorithms are unable to correct for the reflection point dispersal caused by ray bending in inhomogeneous media. We present an analytical formulation of the MZO operator for the simple possible variation of velocity within the earth, i.e. a constant gradient in the vertical direction. The derivation of the MZO operator is carried out in two steps. We first derive the equation of the constant traveltime surface for linear V(z) velocity functions and show that the isochron can be represented by a fourth-degree polynomial in x, y and z. This surface reduces to the well-known ellipsoid in the constant-velocity case, and to the spherical wavefront obtained by Slotnick in the coincident source-receiver case. We then derive the kinematic and dynamic zero-offset corrections in parametric form by using the equation of the isochron. The weighting factors are obtained in the high-frequency limit by means of a simple geometric spreading correction. Our analytical results show that the MZO operator is a multivalued, saddle-shaped operator with marked dip moveout effects in the cross-line direction. However, the amplitude analysis and the distribution of dips along the MZO impulse response show that the most important contributions of the MZO operator are concentrated in a narrow zone along the in-line direction. In practice, MZO processing requires approximately the same trace spacing in the in-line and cross-line directions to avoid spatial aliasing effects.     

OFFSET CONTINUATION OF SEISMIC SECTIONS*

Geometrical acoustic and wave theory lead to a second-order partial differential equation that links seismic sections with different offsets. In this equation a time-shift term appears that corresponds to normal moveout; a second term, dependent on offset and time only, corrects the moveout of dipping events. The zero-offset stacked section can thus be obtained by continuing the section with maximum offset towards zero, and stacking along the way the other common-offset sections. Without the correction for dip moveout, the spatial resolution of the section is noticeably impaired, thus limiting the advantages that could be obtained with expensive migration procedures. Trade-offs exist between multiplicity of coverage, spatial resolution, and signal-to-noise; in some cases the spatial resolution on the surface can be doubled and the aliasing noise averaged out. Velocity analyses carried out on data continued to zero offset show a better resolution and improved discrimination against multiples. For instance, sea-floor multiples always appear at water velocity, so that their removal is simplified. This offset continuation can be carried out either in the time-space domain or in the time-wave number domain. The methods are applied both to synthetic and real data.     

基于全波形反演的探地雷达数据逆时偏移成像

逆时偏移成像(RTM)常用来处理复杂速度模型,包括陡倾角及横向速度变化剧烈的模型.与常规偏移成像方法(如Kirchhoff偏移)相比,逆时偏移成像能提供更好的偏移成像结果,近些年逆时偏移成像越来越广泛地应用到勘探地震中,它逐渐成为石油地震勘探中的一种行业标准.电磁波和弹性波在动力学和运动学上存在相似性,故本文开发了基于麦克斯韦方程组的电磁波逆时偏移成像算法,并将其应用到探地雷达数据处理中.时间域有限差分(FDTD)用于模拟电磁波正向和逆向传播过程,互相关成像条件用于获得最终偏移结果.逆时偏移成像算法中,偏移成像结果受初始模型影响较大,而其中决定电磁波传播速度的介电常数的影响尤为重要.本文基于时间域全波形反演(FWI)算法反演获得了更为精确的地下介电常数模型,并将其反演结果作为逆时偏移成像的初始介电常数模型.为了验证此算法的有效性,首先构建了一个复杂地质结构模型,合成了共偏移距及共炮点探地雷达数据,分别应用常规Kirchhoff偏移算法及逆时偏移成像算法进行偏移处理,成像结果显示由逆时偏移成像算法得到的偏移结果与实际模型具有较高的一致性;此外本文在室内沙槽中进行了相关的物理模拟实验,采集了共偏移距及共炮点探地雷达数据,分别应用Kirchhoff和叠前逆时偏移成像算法进行处理,结果表明叠前逆时偏移成像在实际应用中能获得更好的成像效果.     

波动方程有限差分法叠前深度偏移

从地震叠前反射椭圆方程出发,本文导出了基于波动理论的共偏移距地震剖面叠前偏移方程,然后对此方程进行参考速度场中的浮动坐标变换,获得了叠前深度偏移方程．为了解决叠前衍射方程中含有对深度的二阶导数引起波场延拓成像的不适定问题,文中采用低阶偏微分方程组近似描述全上行波的办法,得到了衍射方程的高阶近似方程,并给出了计算衍射方程和折射方程稳定的差分格式,最后用此方法编制的程序对某一碳酸岩地区的地震资料进行了试处理,效果良好．     

适于大规模数据的三维Kirchhoff 积分法体偏移实现方案

为适应实际生产中对大规模三维工区数据处理的效果及效率的要求,提出了按三维成像体输出成像结果的3D Kirchhoff积分法偏移实现方案.将地震数据按共偏移距道集形式排放,每个共偏移距数据的偏移类似于一个3D叠后Kirchhoff积分偏移,极大地降低了对计算机内存和局部盘及I/O通讯率的要求.每个地震道的成像(输出等时面)在由炮检点连线定义的旋转坐标系中进行,更好地考虑了偏移孔径计算及反假频处理.同时兼顾了超大规模地震数据PSTM成像处理中内存需求量、I/O通讯问题、并行处理方案及效率优化的细节问题.并行计算用偏移距号和每个共偏移距数据体中的线号作为一级和二级索引进行任务分解,更适应当前计算机集群中计算节点比较多的情况.最后考虑了在基本不影响效率的前提下的断点保护处理方案.理论及实际数据测试结果说明了该方案的可行性,与商业软件的对比验证了该方案的优越性.在此较完善的实现方案基础上,可以容易地把更优越的积分类偏移方法迅速推向实用化.     

Numerical tests of 3D true-amplitude zero-offset migration1

An amplitude-preserving migration aims at imaging compressional primary (zero-or) non-zero-offset reflections into 3D time or depth-migrated reflections so that the migrated wavefield amplitudes are a measure of angle-dependent reflection coeffcients. The principal objective is the removal of the geometrical-spreading factor of the primary reflections. Various migration/inversion algorithms involving weighted diffraction stacks proposed recently are based on Born or Kirchhoff approximations. Here, a 3D Kirchhoff-type zero-offset migration approach, also known as a diffraction-stack migration, is implemented in the form of a time migration. The primary reflections of the wavefield to be imaged are described a priori by the zero-order ray approximation. The aim of removing the geometrical- spreading loss can, in the zero-offset case, be achieved by not applying weights to the data before stacking them. This case alone has been implemented in this work. Application of the method to 3D synthetic zero-offset data proves that an amplitude-preserving migration can be performed in this way. Various numerical aspects of the true-amplitude zero-offset migration are discussed.     

Establishing an effective offset-dependent velocity model by ray tracing and its application in Kirchhoff time migration

Conventional Kirchhoff prestack time migration based on the hyperbolic moveout can cause ambiguity in laterally inhomogeneous media, because the root mean square velocity corresponds to a one-dimensional model under the horizontal layer assumption; it does not include the lateral variations. The shot/receiver configuration with different offsets and azimuths should adopt different migration velocities as they contribute to a single image point. Therefore, we propose to use an offset-vector to describe the lateral variations through an offset-dependent velocity corresponding to the difference in offset from surface points to the image point. The offset-vector is decomposed into orthogonal directions along the in-line and cross-line directions so that the single velocity can be expressed as a series of actual velocities. We use a simple Snell's law-based ray tracing to calculate the travel time recorded at the image point and convert the travel time to an equivalent velocity corresponding to a pseudo-straight ray. The double-square-root equation using such an equivalent velocity in the offset-vector domain is non-hyperbolic and asymmetrical, which improves the accuracy of the migration. Numerical examples using the Marmousi model and a wide azimuth field data show that the proposed method can achieve reasonable accuracy and significantly enhances the imaging of complex structures.     

关于共反射面元叠加方法在实际应用中的一些思考

共反射面元（Common Reflection Surface=CRS）叠加是一种特殊的零偏移距成像方法,实践中它具有独立于宏观速度模型和完全数据驱动实现的鲜明特色,CRS叠加理论认为在得到高质量的零偏移距剖面的同时,还可以得到三个有用的波场属性参数剖面反演宏观速度模型,CRS叠加剖面之后的叠后深度偏移质量将超过叠前深度偏移.虽然CRS叠加倡导的成像方式和承诺的上述理想境界带来了全新的启示,但是实践中这些特色同样带来了令人困扰的问题,为此我们提出了倾角分解CRS叠加方法解决这些问题.本文即是作者通过上述实践之后对CRS叠加方法形成的一些思考和总结.     

基于Hilbert变换的全波场分离逆时偏移成像

逆时偏移方法利用双程波算子模拟波场的正向和反向传播,通常采用互相关成像条件获得偏移剖面,是一种高精度的成像方法.但是传统的互相关成像条件会在偏移结果中产生低频噪声;此外,如果偏移速度中存在剧烈速度变化还可能进一步产生偏移假象.为了提高逆时偏移的成像质量,可在成像过程中先对震源波场和检波点波场分别进行波场分离,然后选择合适的波场成分进行互相关成像.本文基于Hilbert变换,推导了可在偏移过程中进行上下行和左右行波场分离的高效波场分离公式以及相应的成像条件,结合Sigsbee 2B合成数据,给出了不同波场成分的互相关成像结果.数值算例结果表明,采用本文提出的高效波场分离算法以及合理的波场成分互相关成像条件可以获得高信噪比的成像结果.     

共偏移距道集平面波叠前时间偏移与反偏移

在Dubrulle提出的共偏移距道集频率波数域叠前时间偏移的基础上,提出了共偏移距道集频率波数域叠前时间偏移与反偏移一对共轭算子.讨论了该对算子的变孔径实现过程.并把该对共轭算子串连起来实现了叠前地震数据的规则化处理.指出最小二乘意义下的叠前地震数据规则化会得到更好的效果.v(z)介质模型和Marmousi模型的数值试验结果表明,方法理论正确、有效.     

Separation of PP- and PS-wave reflected seismic data using two-dimensional finite offset common-reflection-surface traveltime approximation

Recently, the interest in PS-converted waves has increased for several applications, such as sub-basalt layer imaging, impedance estimates and amplitude-versus-offset analysis. In this study, we consider the problem of separation of PP- and PS-waves from pre-stacked multicomponent seismic data in two-dimensional isotropic medium. We aim to demonstrate that the finite-offset common-reflection-surface traveltime approximation is a good alternative for separating PP- and PS-converted waves in common-offset and common shot configurations by considering a two-dimensional isotropic medium. The five parameters of the finite-offset common-reflection-surface are firstly estimated through the inversion methodology called very fast simulated annealing, which estimates all parameters simultaneously. Next, the emergence angle, one of the inverted parameters, is used to build an analytical separation function of PP and PS reflection separation based on the wave polarization equations. Once the PP- and PS-converted waves were separated, the sections are stacked to increase the signal-to-noise ratio using the special curves derived from finite-offset common-reflection-surface approximation. We applied this methodology to a synthetic dataset from simple-layered to complex-structured media. The numerical results showed that the inverted parameters of the finite offset common-reflection-surface and the separation function yield good results for separating PP- and PS-converted waves in noisy common-offset and common shot gathers.     

A SIMPLE EFFICIENT METHOD OF DIP-MOVEOUT CORRECTION1

Much of the success of modern seismic data processing derives from the use of the stacking process. Unfortunately, as is well known, conventional normal moveout correction (NMO) introduces mispositioning of data, and hence mis-stacking, when dip is present. Dip moveout correction (DMO) is a technique that converts non-zero-offset seismic data after NMO to true zero-offset locations and reflection times, irrespective of dip. The combination of NMO and DMO followed by post-stack time migration is equivalent to, but can be implemented much more efficiently than, full time migration before stack. In this paper we consider the frequency-wavenumber DMO algorithm developed by Hale. Our analysis centres on the result that, for a given dip, the combination of NMO at migration velocity and DMO is equivalent to NMO at the appropriate, dip-dependent, stacking velocity. This perspective on DMO leads to computationally efficient methods for applying Hale DMO and also provides interesting insights on the nature of both DMO and conventional stacking.     

Common-reflection-point time migration

Since the early days of seismic processing, time migration has proven to be a valuable tool for a number of imaging purposes. Main motivations for its widespread use include robustness with respect to velocity errors, as well as fast turnaround and low computation costs. In areas of complex geology, in which it has well-known limitations, time migration can still be of value by providing first images and also attributes, which can be of much help in further, more comprehensive depth migration. Time migration is a very close process to common-midpoint (CMP) stacking and, more recently, to zero-offset commonreflection- surface (CRS) stacking. In fact, Kirchhoff time migration operators can be readily formulated in terms of CRS parameters. In the nineties, several studies have shown advantages in the use of common-reflection-point (CRP) traveltimes to replace conventional CMP traveltimes for a number of stacking and migration purposes. In this paper, we follow that trend and introduce a Kirchhoff-type prestack time migration and velocity analysis algorithm, referred to as CRP time migration. The algorithm is based on a CRP operator together with optimal apertures, both computed with the help of CRS parameters. A field-data example indicates the potential of the proposed technique.     

Fourier finite-difference migration for steeply dipping reflectors with complex overburden1

The improvement in accuracy and efficiency of wave-equation migration techniques is an ongoing topic of research. The main problem is the correct imaging of steeply dipping reflectors in media with strong lateral velocity variations. We propose an improved migration method which is based on cascading phase-shift and finite-difference operators for downward continuation. Due to these cascaded operators we call this method‘Fourier finite-difference migration’(FFD migration). In our approach we try to generalize and improve the split-step Fourier migration method for strong lateral velocity variations using an additional finite-difference correction term. Like most of the current migration methods in use today, our method is based on the one-way wave equation. It is solved by first applying the square-root operator but using a constant velocity at each depth step which has to be the minimum velocity. In a second step, the approximate difference between the correct square-root operator and this constant-velocity squareroot operator (the error made in the first step) is implemented as an implicit FD migration scheme, part of which is the split-step Fourier correction term. Some practical aspects of the new FFD method are discussed. Its performance is compared with that of split-step and standard FD migration schemes. First applications to synthetic and real data sets are presented. They show that the superiority of FFD migration becomes evident by migrating steeply dipping reflectors with complex overburden having strong lateral velocity variations. If velocity is laterally constant, FFD migration has the accuracy of the phase-shift method. The maximum migration angle is velocity adaptive, in contrast to conventional FD migration schemes. It varies laterally depending on the local level of velocity variation. FFD migration is more efficient than higher-order implicit FD schemes. These schemes use two cascaded downward-continuation steps in order to attain comparable migration performance.