Visco‐acoustic prestack reverse‐time migration based on the time‐space domain adaptive high‐order finite‐difference method

It is important to include the viscous effect in seismic numerical modelling and seismic migration due to the ubiquitous viscosity in an actual subsurface medium. Prestack reverse‐time migration (RTM) is currently one of the most accurate methods for seismic imaging. One of the key steps of RTM is wavefield forward and backward extrapolation and how to solve the wave equation fast and accurately is the essence of this process. In this paper, we apply the time‐space domain dispersion‐relation‐based finite‐difference (FD) method for visco‐acoustic wave numerical modelling. Dispersion analysis and numerical modelling results demonstrate that the time‐space domain FD method has great accuracy and can effectively suppress numerical dispersion. Also, we use the time‐space domain FD method to solve the visco‐acoustic wave equation in wavefield extrapolation of RTM and apply the source‐normalized cross‐correlation imaging condition in migration. Improved imaging has been obtained in both synthetic and real data tests. The migration result of the visco‐acoustic wave RTM is clearer and more accurate than that of acoustic wave RTM. In addition, in the process of wavefield forward and backward extrapolation, we adopt adaptive variable‐length spatial operators to compute spatial derivatives to significantly decrease computing costs without reducing the accuracy of the numerical solution.     

深度均匀采样梯形网格有限差分地震波场模拟方法

由于重力引起的岩石压实效应,一般来说,地震波传播速度由浅入深整体逐渐增大.梯形坐标系设计可耦合速度由浅入深逐渐增大的变化,该坐标系中均匀网格采样所对应的物理直角坐标系网格由浅入深逐渐增大,也即浅部低速区对应细网格,深部高速区对应粗网格.在梯形坐标系表征波动方程后利用有限差分求解,本文实现一种深度均匀采样、横向采样间隔随深度增加逐渐线性增大的有限差分地震波模拟方法.梯形坐标系波动方程离散后,仍采用常规均匀网格有限差分算法对其求解.由于横向网格大小由浅入深线性增加,本方法可避免不同大小网格区域过渡所产生的虚假反射.梯形坐标系波场模拟浅层精度高,深层横向响应范围广,可有效减少有限差分网格数量.本文提出的方法是在更广义的坐标系下利用有限差分求解波动方程,正交坐标系仅为该梯形坐标系之特例.本文旨在为大速度动态范围深地高效高精度地震波场模拟提供一种思路.     

基于时空域自适应高阶有限差分的声波叠前逆时偏移

叠前逆时偏移在理论上是现行偏移方法中最为精确的一种成像方法,其实现过程中的核心步骤之一是波动方程的波场延拓,而波场延拓的本质是求解波动方程,所以精确、快速地求解波动方程对逆时偏移至关重要.本文采用一种基于时空域频散关系的有限差分方法来求解声波方程,分析其频散和稳定性,实现波场数值模拟,并将分析和模拟结果与传统有限差分法进行对比.分析结果和模型数值模拟结果都表明时空域有限差分法模拟精度更高、稳定性更好.将时空域高阶有限差分法应用到叠前逆时偏移波场延拓的方程求解中,然后再利用归一化互相关成像条件成像,理论模型数据偏移处理获得了精度更高的成像.同时,在逆时偏移波场延拓的实现中,采用自适应变长度的空间差分算子求解空间导数的有限差分策略,在不影响数值模拟和成像精度的前提下,有效地提高了计算效率.     

An efficient wave extrapolation method for anisotropic media with tilt

Wavefield extrapolation operators for elliptically anisotropic media offer significant cost reduction compared with that for the transversely isotropic case, particularly when the axis of symmetry exhibits tilt (from the vertical). However, elliptical anisotropy does not provide accurate wavefield representation or imaging for transversely isotropic media. Therefore, we propose effective elliptically anisotropic models that correctly capture the kinematic behaviour of wavefields for transversely isotropic media. Specifically, we compute source‐dependent effective velocities for the elliptic medium using kinematic high‐frequency representation of the transversely isotropic wavefield. The effective model allows us to use cheaper elliptic wave extrapolation operators. Despite the fact that the effective models are obtained by matching kinematics using high‐frequency asymptotic, the resulting wavefield contains most of the critical wavefield components, including frequency dependency and caustics, if present, with reasonable accuracy. The methodology developed here offers a much better cost versus accuracy trade‐off for wavefield computations in transversely isotropic media, particularly for media of low to moderate complexity. In addition, the wavefield solution is free from shear‐wave artefacts as opposed to the conventional finite‐difference‐based transversely isotropic wave extrapolation scheme. We demonstrate these assertions through numerical tests on synthetic tilted transversely isotropic models.     

Using the pseudospectral method on curved grids for 2D elastic forward modelling1

When applying the conventional Fourier pseudospectral method (FSM) on a Cartesian grid that has a sufficient size to propagate a pulse, spurious diffractions from the staircase representation of the curved interfaces appear in the wavefield. It is demonstrated that these non-physical diffractions can be eliminated by using curved grids that conform to all the interfaces of the subsurface. Methods for solving the 2D acoustic wave equation using such curved grids have been published previously by the authors. Here the extensions to the full 2D elastic wave equations are presented. The curved grids are generated by using the so-called multiblock strategy which is a well-known concept in computational fluid dynamics. In principle the sub-surface is divided into a number of contiguous subdomains. A separate grid is generated for each subdomain patching the grid lines across domain boundaries to obtain a globally continuous grid. Using this approach, even configurations with pinch outs can be handled. The curved grid is taken to constitute a generalized curvilinear coordinate system. Thus, the elastic equations have to be written in a curvilinear frame before applying the numerical scheme. The method implies that twice the number of spatial derivatives have to be evaluated compared to the conventional FSM on a Cartesian grid. However, it is demonstrated that the extra terms are more than compensated for by the fewer grid points needed in the curved approach.     

Wave-equation migration velocity analysis. II. Subsalt imaging examples

Subsalt imaging is strongly dependent on the quality of the velocity model. However, rugose salt bodies complicate wavefield propagation and lead to subsalt multipathing, illumination gaps and shadow zones, which cannot be handled correctly by conventional traveltime‐based migration velocity analysis (MVA). We overcome these limitations by the wave‐equation MVA technique, introduced in a companion paper, and demonstrate the methodology on a realistic synthetic data set simulating a salt‐dome environment and a Gulf of Mexico data set. We model subsalt propagation using wave paths created by one‐way wavefield extrapolation. Those wave paths are much more accurate and robust than broadband rays, since they inherit the frequency dependence and multipathing of the underlying wavefield. We formulate an objective function for optimization in the image space by relating an image perturbation to a perturbation of the velocity model. The image perturbations are defined using linearized prestack residual migration, thus ensuring stability, relative to the first‐order Born approximation assumptions. Synthetic and real data examples demonstrate that wave‐equation MVA is an effective tool for subsalt velocity analysis, even when shadows and illumination gaps are present.     

基于自适应优化有限差分方法的全波VSP逆时偏移

与地面地震资料相比,VSP资料具有分辨率高、环境噪声小及能更好地反映井旁信息等优点.常规VSP偏移主要对上行反射波进行成像,存在照明度低、成像范围受限等问题.为了增加照明度、拓宽成像范围、提高成像精度,本文采用直达波除外的所有声波波场数据(全波),包括一次反射波、多次反射波等进行叠前逆时偏移成像.针对逆时偏移中的四个关键问题,即波场延拓、吸收边界条件、成像条件及低频噪声的压制,本文分别采用自适应变空间差分算子长度的优化有限差分方法(自适应优化有限差分方法)求解二维声波波动方程以实现高精度、高效率的波场延拓,采用混合吸收边界条件压制因计算区域有限所引起的人工边界反射,采用震源归一化零延迟互相关成像条件进行成像,采用拉普拉斯滤波方法压制逆时偏移中产生的低频噪声.本文对VSP模型数据的逆时偏移成像进行了分析,结果表明:自适应优化有限差分方法比传统有限差分方法具有更高的模拟精度与计算效率,适用于VSP逆时偏移成像;全波场VSP逆时偏移成像比上行波VSP逆时偏移的成像范围大、成像效果好;相对于反褶积成像条件,震源归一化零延迟互相关成像条件具有稳定性好、计算效率高等优点.将本文方法应用于某实际VSP资料的逆时偏移成像,进一步验证了本文方法的正确性和有效性.     

TOWARDS OPTIMUM ONE-WAY WAVE PROPAGATION1

Numerical wavefield extrapolation represents the backbone of any algorithm for depth migration pre- or post-stack. For such depth imaging techniques to yield reliable and interpretable results, the underlying wavefield extrapolation algorithm must propagate the waves through inhomogeneous media with a minimum of numerically induced distortion, over a range of frequencies and angles of propagation. A review of finite-difference (FD) approximations to the acoustic one-way wave equation in the space-frequency domain is presented. A straightforward generalization of the conventional FD formulation leads to an algorithm where the wavefield is continued downwards with space-variant symmetric convolutional operators. The operators can be precomputed and made accessible in tables such that the ratio between the temporal frequency and the local velocity is used to determine the correct operator at each grid point during the downward continuation. Convolutional operators are designed to fit the desired dispersion relation over a range of frequencies and angles of propagation such that the resulting numerical distortion is minimized. The optimization is constrained to ensure that evanescent energy and waves propagating at angles higher than the maximum design angle are attenuated in each extrapolation step. The resulting operators may be viewed as optimally truncated and bandlimited spatial versions of the familiar phase shift operator. They are unconditionally stable and can be applied explicitly. This results in a simple wave propagation algorithm, eminently suited for implementation on pipelined computers and on large parallel computing systems.     

Wavefield interpolation in 3D large‐step Fourier wavefield extrapolation

Extrapolating wavefields and imaging at each depth during three‐dimensional recursive wave‐equation migration is a time‐consuming endeavor. For efficiency, most commercial techniques extrapolate wavefields through thick slabs followed by wavefield interpolation within each thick slab. In this article, we develop this strategy by associating more efficient interpolators with a Fourier‐transform‐related wavefield extrapolation method. First, we formulate a three‐dimensional first‐order separation‐of‐variables screen propagator for large‐step wavefield extrapolation, which allows for wide‐angle propagations in highly contrasting media. This propagator significantly improves the performance of the split‐step Fourier method in dealing with significant lateral heterogeneities at the cost of only one more fast Fourier transform in each thick slab. We then extend the two‐dimensional Kirchhoff and Born–Kirchhoff local wavefield interpolators to three‐dimensional cases for each slab. The three‐dimensional Kirchhoff interpolator is based on the traditional Kirchhoff formula and applies to moderate lateral velocity variations, whereas the three‐dimensional Born–Kirchhoff interpolator is derived from the Lippmann–Schwinger integral equation under the Born approximation and is adapted to highly laterally varying media. Numerical examples on the three‐dimensional salt model of the Society of Exploration Geophysicists/European Association of Geoscientists demonstrate that three‐dimensional first‐order separation‐of‐variables screen propagator Born–Kirchhoff depth migration using thick‐slab wavefield extrapolation plus thin‐slab interpolation tolerates a considerable depth‐step size of up to 72 ms, eventually resulting in an efficiency improvement of nearly 80% without obvious loss of imaging accuracy. Although the proposed three‐dimensional interpolators are presented with one‐way Fourier extrapolation methods, they can be extended for applications to general migration methods.     

基于解析时间波场外推与波场分解的逆时偏移方法研究

双程波方程逆时深度偏移是复杂介质高精度成像的有效技术,但其结果中通常包含成像方法引起的噪音和假象,一般的滤波方法会破坏成像剖面上的振幅,其中的假象也会给后续地质解释带来困扰.将波场进行方向分解然后实现入射波与反射波的相关成像能够有效地消除这类成像噪音,并提高逆时偏移成像质量.波传播方向的分解通常在频率波数域实现,它会占用大量的存储和计算资源,不便于在沿时间外推的逆时深度偏移中应用.本文提出解析时间波场外推方法,可以在时间外推的每个时间片上实现波传播方向的显式分解,逆时深度偏移中利用分解后的炮检波场进行对应的相关运算,实现成像噪音和成像信号的分离.在模型和实际数据上的测试表明,相比于常规互相关逆时偏移成像结果,本文方法能够有效地消除低频成像噪音和特殊地质构造导致的成像假象.     

Effective ellipsoidal models for wavefield extrapolation in tilted orthorhombic media

Wavefield computations using the ellipsoidally anisotropic extrapolation operator offer significant cost reduction compared to that for the orthorhombic case, especially when the symmetry planes are tilted and/or rotated. However, ellipsoidal anisotropy does not provide accurate wavefield representation or imaging for media of orthorhombic symmetry. Therefore, we propose the use of ‘effective ellipsoidally anisotropic’ models that correctly capture the kinematic behaviour of wavefields for tilted orthorhombic (TOR) media. We compute effective velocities for the ellipsoidally anisotropic medium using kinematic high-frequency representation of the TOR wavefield, obtained by solving the TOR eikonal equation. The effective model allows us to use the cheaper ellipsoidally anisotropic wave extrapolation operators. Although the effective models are obtained by kinematic matching using high-frequency asymptotic, the resulting wavefield contains most of the critical wavefield components, including frequency dependency and caustics, if present, with reasonable accuracy. The proposed methodology offers a much better cost versus accuracy trade-off for wavefield computations in TOR media, particularly for media of low to moderate anisotropic strength. Furthermore, the computed wavefield solution is free from shear-wave artefacts as opposed to the conventional finite-difference based TOR wave extrapolation scheme. We demonstrate applicability and usefulness of our formulation through numerical tests on synthetic TOR models.     

螺旋坐标下的因子分解合成炮叠前深度偏移

在合成炮叠前深度偏移的基础上，提出了一种高效的合成炮叠前深度偏移方法，即在螺旋坐标下用因子分解进行波场外推的混合法. 用因子分解进行波场外推分为因果过程和反因果过程两个显式求解过程. 这种螺旋坐标下的显式求解过程,提高了波场外推的效率. 根据相位编码原理,对多个射线参数的合成波场进行了编码叠加,基于射线参数实现了MPI并行计算,进一步提高了计算效率. 在推导了有关公式并进行定量分析之后,对Marmousi复杂模型进行了计算,并作了比较,结果表明本文方法具有精度高和速度快的特点,可用于实际资料的计算.     

基于波场垂向和水平方向外推的高陡构造偏移

常规的单程波波动方程偏移成像方法对大角度的高陡构造偏移成像存在内在的限制.根据波动方程在各个空间方向的数学特性和高陡构造反射地震波的传播特征,通过把地震波分解为垂向的上下行波、水平方向的前后行波和左右行波,提出基于波场垂向外推和水平方向外推相结合的单程波波动方程高陡构造偏移成像方法,即用波场垂向外推的单程波波动方程偏移成像方法解决中低角度平缓构造的偏移成像,用波场水平方向外推的单程波波动方程偏移成像方法解决中高角度陡倾构造的偏移成像.这种基于波场垂向和水平方向外推相结合的高陡构造偏移成像方法是常规单程波波动方程叠前深度偏移成像方法的补充和改进,它相对基于全波方程的逆时偏移具有计算效率上的优势.     

波动方程深度偏移的频率相关变步长延拓方法

发展了波动方程深度延拓的频率相关变步长深度延拓方法和表驱动的单点波场插值技术.前者通过减少深度延拓的次数减少了波动方程深度偏移的计算量,而后者用很少的计算量实现了等间距、理想采样的深度成像.就同一偏移方法,采用频率相关变步长深度延拓加单点插值,其计算量大约是常规的等间距采样延拓方法的三分之一,但两者的成像效果基本相同.文中以最优分裂Fourier方法为例,用二维理论数据（Marmousi模型）和三维实际地震资料验证了这一方法,但这一方法可适用于各类频率域波动方程深度偏移方法.     

Wavefield dependency on virtual shifts in the source location

The wavefield dependence on a virtual shift in the source location can provide information helpful in velocity estimation and interpolation. However, the second‐order partial differential equation (PDE) that relates changes in the wavefield form (or shape) to lateral perturbations in the source location depends explicitly on lateral derivatives of the velocity field. For velocity models that include lateral velocity discontinuities this is problematic as such derivatives in their classical definition do not exist. As a result, I derive perturbation partial differential wave equations that are independent of direct velocity derivatives and thus, provide possibilities for wavefield shape extrapolation in complex media. These PDEs have the same structure as the wave equation with a source function that depends on the background (original source) wavefield. The solutions of the perturbation equations provide the coefficients of a Taylor's series type expansion for the wavefield. The new formulas introduce changes to the background wavefield only in the presence of lateral velocity variation or in general terms velocity variations in the perturbation direction. The accuracy of the representation, as demonstrated on the Marmousi model, is generally good.     

Elastic wave mode separation for tilted transverse isotropy media

Seismic waves propagate through the earth as a superposition of different wave modes. Seismic imaging in areas characterized by complex geology requires techniques based on accurate reconstruction of the seismic wavefields. A crucial component of the methods in this category, collectively known as wave‐equation migration, is the imaging condition that extracts information about the discontinuities of physical properties from the reconstructed wavefields at every location in space. Conventional acoustic migration techniques image a scalar wavefield representing the P‐wave mode, in contrast to elastic migration techniques, which image a vector wavefield representing both the P‐ and S‐waves. For elastic imaging, it is desirable that the reconstructed vector fields are decomposed into pure wave modes, such that the imaging condition produces interpretable images, characterizing, for example, PP or PS reflectivity. In anisotropic media, wave mode separation can be achieved by projection of the reconstructed vector fields on the polarization vectors characterizing various wave modes. For heterogeneous media, because polarization directions change with position, wave mode separation needs to be implemented using space‐domain filters. For transversely isotropic media with a tilted symmetry axis, the polarization vectors depend on the elastic material parameters, including the tilt angles. Using these parameters, we separate the wave modes by constructing nine filters corresponding to the nine Cartesian components of the three polarization directions at every grid point. Since the S polarization vectors in transverse isotropic media are not defined in the singular directions, e.g., along the symmetry axes, we construct these vectors by exploiting the orthogonality between the SV and SH polarization vectors, as well as their orthogonality with the P polarization vector. This procedure allows one to separate all three modes, with better preserved P‐wave amplitudes than S‐wave amplitudes. Realistic synthetic examples show that this wave mode separation is effective for both 2D and 3D models with strong heterogeneity and anisotropy.     

Seismic wavefield modeling based on time-domain symplectic and Fourier finite-difference method

Seismic wavefield modeling is important for improving seismic data processing and interpretation. Calculations of wavefield propagation are sometimes not stable when forward modeling of seismic wave uses large time steps for long times. Based on the Hamiltonian expression of the acoustic wave equation, we propose a structure-preserving method for seismic wavefield modeling by applying the symplectic finite-difference method on time grids and the Fourier finite-difference method on space grids to solve the acoustic wave equation. The proposed method is called the symplectic Fourier finite-difference (symplectic FFD) method, and offers high computational accuracy and improves the computational stability. Using acoustic approximation, we extend the method to anisotropic media. We discuss the calculations in the symplectic FFD method for seismic wavefield modeling of isotropic and anisotropic media, and use the BP salt model and BP TTI model to test the proposed method. The numerical examples suggest that the proposed method can be used in seismic modeling of strongly variable velocities, offering high computational accuracy and low numerical dispersion. The symplectic FFD method overcomes the residual qSV wave of seismic modeling in anisotropic media and maintains the stability of the wavefield propagation for large time steps.     

利用矩阵分解理论的双程波方程叠前深度偏移方法

叠前深度偏移理论及方法一直是地震数据成像中研究的热点问题.业界对单程波叠前深度偏移方法和逆时深度偏移开展了深入的研究,但对双程波方程波场深度延拓理论及成像方法的研究还鲜有报道.本文以地表记录的波场值为基础,利用单程波传播算子估计波场对深度的偏导数,为在深度域求解双程波方程提供充分的边界条件,并提出利用矩阵分解理论实现双程波方程的波场深度外推.通过对强速度变化介质中传播波场的计算,与传统的单程波偏移方法相比,本文提出的偏移方法计算的波场与常规有限差分技术计算的波场相一致,证明了本方法计算的准确性.通过对SEAM模型的成像,在相同的成像参数下,与传统的单程波偏移算法和逆时深度偏移算法方法相比,本文提出的偏移方法能够提供更少的虚假成像和更清晰的成像结果.本文所提偏移算法具有深度偏移和双程波偏移的双重特色,推动和发展了双程波叠前深度偏移的理论和实践.     

求解二阶解耦弹性波方程的低秩分解法和低秩有限差分法

时间域的波场延拓方法在本质上都可以归结为对一个空间-波数域算子的近似.本文基于一阶波数-空间混合域象征,提出一种新的方法求解解耦的二阶位移弹性波方程.该方法采用交错网格,连续使用两次一阶前向和后向拟微分算子,推导得到了解耦的二阶位移弹性波方程的波场延拓算子.由于该混合域象征在伪谱算子的基础上增加了一个依赖于速度模型的补偿项,可以补偿由于采用二阶中心差分计算时间微分项带来的误差,有效地减少模拟结果的数值频散,提高模拟精度.然而,在非均匀介质中,直接计算该二阶的波场延拓算子,每一个时间步上需要做N次快速傅里叶逆变换,其中N是总的网格点数.为了减少计算量,提出了交错网格低秩分解方法;针对常规有限差分数值频散问题,本文将交错网格低秩方法与有限差分法结合,提出了交错网格低秩有限差分法.数值结果表明,交错网格低秩方法和交错网格低秩有限差分法具有较高的精度,对于复杂介质的地震波数值模拟和偏移成像具有重要的价值.     

Directional‐oriented wavefield imaging: a new wave‐based subsurface illumination imaging condition for reverse time migration

The key objective of an imaging algorithm is to produce accurate and high‐resolution images of the subsurface geology. However, significant wavefield distortions occur due to wave propagation through complex structures and irregular acquisition geometries causing uneven wavefield illumination at the target. Therefore, conventional imaging conditions are unable to correctly compensate for variable illumination effects. We propose a generalised wave‐based imaging condition, which incorporates a weighting function based on energy illumination at each subsurface reflection and azimuth angles. Our proposed imaging kernel, named as the directional‐oriented wavefield imaging, compensates for illumination effects produced by possible surface obstructions during acquisition, sparse geometries employed in the field, and complex velocity models. An integral part of the directional‐oriented wavefield imaging condition is a methodology for applying down‐going/up‐going wavefield decomposition to both source and receiver extrapolated wavefields. This type of wavefield decomposition eliminates low‐frequency artefacts and scattering noise caused by the two‐way wave equation and can facilitate the robust estimation for energy fluxes of wavefields required for the seismic illumination analysis. Then, based on the estimation of the respective wavefield propagation vectors and associated directions, we evaluate the illumination energy for each subsurface location as a function of image depth point and subsurface azimuth and reflection angles. Thus, the final directional‐oriented wavefield imaging kernel is a cross‐correlation of the decomposed source and receiver wavefields weighted by the illuminated energy estimated at each depth location. The application of the directional‐oriented wavefield imaging condition can be employed during the generation of both depth‐stacked images and azimuth–reflection angle‐domain common image gathers. Numerical examples using synthetic and real data demonstrate that the new imaging condition can properly image complex wave paths and produce high‐fidelity depth sections.