基于偏微分方程的模型光滑处理在偏移成像中的应用

本文针对射线类偏移成像当中的速度模型光滑处理问题,借鉴数字图像处理当中的偏微分方程法,基于能量泛函,应用变分方法导出基于速度模型的偏微分方程实现射线类偏移成像当中的速度模型的光滑处理.由于偏微分方程法具有线性叠加特性、模型解的唯一性和局部特征保持性,因此,应用该算法可以实现基于原始速度模型空间结构的模型光滑处理.通过在原始速度模型以及光滑处理后的速度模型上计算速度的空间分布以及地震波走时、射线路径可以得出,偏微分方程法对速度模型的光滑处理能够很好地保持原始模型的空间结构,偏移成像结果也证明了该方法的实用性.     

Interpretation of MHD-sounding data from the Kola Peninsula by the electromagnetic migration method

The electromagnetic migration method, based on principles analogous to those of seismic migration, is developed.The concept of migrated fields is based on the Stratton-Chu type integrals, written in reverse time for the observed electromagnetic fields. Four types of migrated fields, which form a system of migration transformations of the transient electromagnetic field, are introduced.Study of the properties of the migrated fields by means of theoretical and model examples makes it possible to determine the optimal parameters of the migration procedure in which the anomalous field sources are localized by means of migration transformation.Information about the experiment using the MHD-generator (experiment ‘Chibini’) carried out on the Kola Peninsula to study the geoelectrical structure of the Earth's crust and upper mantle is included. The electromagnetic migration method for the interpretation of MHD-sounding data recorded on the Kola Peninsula, along a profile crossing the mineral-rich region of the Imandra-Varzuga structure, permits us to determine the location of a conducting zone at the depth of 10 km in the Earth's crust.     

Effects of seismic wave propagation upon buried pipelines

The behaviour of long straight buried pipelines subjected to seismic wave propagation is investigated. Well-known relationships for determining upper bounds for the axial strain and curvature in the pipeline as well as relationships for relative displacement and rotation at the pipeline joints are discussed. The assumption that the seismic excitation can be modelled as a travelling wave having a shape which remains unchanged as it traverses the pipeline is examined in detail. It is shown that this assumption is unconservative when the effective propagation velocity of the seismic waves with respect to the pipeline is such that the actual time lag (separation distance between points divided by effective propagation velocity) is less than a ‘cross-over’ time lag. Cross-over time lags for 22 pairs of ground displacements recorded during the 1971 San Fernando Earthquake are presented in this paper. Finally, methods for estimating the propagation speed of the seismic waves along or with respect to the pipeline are discussed.     

DEPTH IMAGING OF OFFSET VERTICAL SEISMIC PROFILE DATA1

Depth migration consists of two different steps: wavefield extrapolation and imaging. The wave propagation is firmly founded on a mathematical frame-work, and is simulated by solving different types of wave equations, dependent on the physical model under investigation. In contrast, the imaging part of migration is usually based on ad hoc‘principles’, rather than on a physical model with an associated mathematical expression. The imaging is usually performed using the U/D concept of Claerbout (1971), which states that reflectors exist at points in the subsurface where the first arrival of the downgoing wave is time-coincident with the upgoing wave. Inversion can, as with migration, be divided into the two steps of wavefield extrapolation and imaging. In contrast to the imaging principle in migration, imaging in inversion follows from the mathematical formulation of the problem. The image with respect to the bulk modulus (or velocity) perturbations is proportional to the correlation between the time derivatives of a forward-propagated field and a backward-propagated residual field (Lailly 1984; Tarantola 1984). We assume a physical model in which the wave propagation is governed by the 2D acoustic wave equation. The wave equation is solved numerically using an efficient finite-difference scheme, making simulations in realistically sized models feasible. The two imaging concepts of migration and inversion are tested and compared in depth imaging from a synthetic offset vertical seismic profile section. In order to test the velocity sensitivity of the algorithms, two erroneous input velocity models are tested. We find that the algorithm founded on inverse theory is less sensitive to velocity errors than depth migration using the more ad hoc U/D imaging principle.     

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.     

火山岩地震屏蔽层的转换波叠前时间偏移成像

在反射地震转换波资料处理中,准确求取共转换点一直是一个难题,采用叠前时间偏移技术能避免共转换点道集的抽取,而且能够使转换波归位到真正的反射点上,实现准确成像.本文针对火山岩地震屏蔽层的转换波成像问题,通过对转换波共近似转换点道集进行速度分析,建立了转换波叠前时间偏移的初始速度场,通过速度扫描和纵、横波速度比值扫描确定最佳的偏移速度场和纵、横波速度比值,实现了在火山岩高速层覆盖区域的转换波偏移成像.实际资料的成像结果表明,本文采用的近似转换点计算以及转换波叠前时间偏移方法是有效的.     

Velocity imaging by tau–p transformation of refracted seismic traveltimes

We present a new method for producing a ‘brute’ velocity image rapidly and automatically from traveltimes picked from densely sampled refraction data. The procedure involves imaging by data transformation from the time–offset domain into the tau–p (intercept–slope) domain, and does not include conventional modelling steps. Differences in apparent velocity and tau along reciprocal paths in the up- and downdip directions allow the estimation of the true velocity and geometrical position of the ray turning points. The tau–velocity–turningpoint (τ–ν–x) map distributes phases automatically on the basis of geometry and velocity to give a two-dimensional representation of subsurface structure. This map may be converted simply to depth and two-way-time images. Such images have potential for direct geological interpretation, for use as a starting model for seismic inversion, for superimposition on to conventional reflection images, or for input into prestack depth migration and other processing routines.     

Direct interpretation of 2D potential fields for deep structures by means of the quasi-singular points method

The founder of the Russian school of direct interpretation of potential fields (with minimal prior geological‐geophysical information) was V.M. Berezkin, who introduced the operator of total normalized gradient for the 2D interpretation of profile gravity data sets. This operator was successfully applied in searches of hydrocarbon reservoirs. The further development of this approach (the so‐called quasi‐singular points method) has allowed solution also to various structural problems, using mathematical criteria for the transition from extremes of total normalized gradient fields to coordinates of anomalous sources. The main numerical evaluation strategy is based on stabilized downward continuation of field derivatives and specific use of the filtration properties of Fourier series approximation. The characteristic properties of the quasi‐singular points method are: 1) presentation of a more general total normalized gradient function through additional parameters (derivative order m, form of smoothing function Q, number of Fourier coefficients N* with maximal N), optimum values being chosen during a peak‐spectrum analysis of the interpreted function; 2) calculation of the set of total normalized gradient fields for various values of N*/N, representing coordinate systems {x,N*/N} as an ‘axes tree’ of extrema, where each 2D total normalized gradient field is representationally compressed in a 1D line, permitting a) immediate overview of the positions of the axes in all variants of the calculated fields and b) reduction of the retained information, as required in subsequent interpretation; 3) development of two criteria for transition from extrema of total normalized gradient fields to the coordinates of anomaly sources. The quasi‐singular points method is intended for tracing limiting gently‐sloping boundaries, if their micro‐relief features are sources of the interpreted anomaly but sub‐vertical contacts may also be traced. The method has been tested in delineating various geological structures. One of the most challenging, successfully achieved, was tracing of the Moho discontinuity and study of the upper mantle, using only Bouguer anomaly data along interpretation profiles. This is attested in an example of two regional profiles intersecting the European part of Russia. The central part of one of them coincides with the results from a deep seismic profile.     

Event-oriented velocity estimation based on prestack data in time or depth domain1

A 2D reflection tomographic method is described, for the purpose of estimating an improved macrovelocity field for prestack depth migration. An event-oriented local approach of the ‘layer-stripping’ type has been developed, where each input event is defined by its traveltime and a traveltime derivative, taken with respect to one of four coordinates in the source/receiver and midpoint half-offset systems. Recent work has shown that the results of reflection tomography may be improved by performing event picking in a prestack depth domain. We adopt this approach and allow events to be picked either before or after prestack depth migration. Hence, if events have been picked in a depth domain, such as the common-shot depth domain or the common-offset depth domain, then a depth-time transformation is required before velocity estimation. The event transformation may, for example, be done by conventional kinematic ray tracingr and with respect to the original depth-migration velocity field. By this means, we expect the input events for velocity updating to become less sensitive to migration velocity errors. For the purpose of velocity estimation, events are subdivided into two categories; reference horizon events and individual events. The reference horizon events correspond to a fixed offset in order to provide basic information about reflector geometry, whereas individual events, corresponding to any offset, are supposed to provide the additional information needed for velocity estimation. An iterative updating approach is used, based on calculation of derivatives of event reflection points with respect to velocity. The event reflection points are obtained by ray-theoretical depth conversion, and reflection-point derivatives are calculated accurately and efficiently from information pertaining to single rays. A number of reference horizon events and a single individual event constitute the minimum information required to update the velocity locally, and the iterations proceed until the individual event reflection point is consistent with those of the reference horizon events. Normally, three to four iterations are sufficient to attain convergence. As a by-product of the process, we obtain so-called uncertainty amplification factors, which relate a picking error to the corresponding error in the estimated velocity or depth horizon position. The vector formulation of the updating relationship makes it applicable to smooth horizons having arbitrary dips and by applying velocity updating in combination with a flexible model-builder, very general macro-model structures can be obtained. As a first step in the evaluation of the new method, error-free traveltime events were generated by applying forward ray tracing within given macrovelocity models. When using such ‘perfect’ observations, the velocity estimation algorithm gave consistent reconstructions of macro-models containing interfaces with differential dip and curvature, a low-velocity layer and a layer with a laterally varying velocity function.     

Determination of optimum kernel bandwidth for northern Marmara region,Turkey

As a relatively recent development, spatial smoothing methods have been introduced to identify seismic patterns. Among the methods developed to model the spatial variation, methods based on utilization of 3-D Gaussian isotropic kernels have a wide reception. The most important question remaining to be answered in the application of these methods is the determination of the optimum kernel bandwidth. At the present, researchers’ efforts to clarify the subject have still not yet finalized, this study aims to provide insightful knowledge for future efforts. In this study, for the region bounded by 27°–33° longitudes and 39°–41° latitudes, where the western section of the famous Northern Anatolian fault lies, smoothing techniques are implemented to determine the optimum smoothing kernel bandwidth for point sources. The influence of the modeling of seismicity through the computation of the optimum smoothing kernel bandwidth is examined. In addition, the sensitivity of each smoothing technique to the seismic patterns, whether densely clustered or scarcely populated, is investigated. In the end, the smoothing method based on optimum neighbor number is identified as highly sensitive to the density of seismicity and seismic clusters, and better in modeling high seismicity compared to the model based on single optimum smoothing distance used for the entire region of interest.     

Estimation of Thomsen＇s anisotropic parameters from walkaway VSP and applications

Estimation of Thomsen＇s anisotropic parameters is very important for accuratetime-to-depth conversion and depth migration data processing. Compared with othermethods, it is much easier and more reliable to estimate anisotropic parameters that arerequired for surface seismic depth imaging from vertical seismic profile （VSP） data, becausethe first arrivals of VSP data can be picked with much higher accuracy. In this study, wedeveloped a method for estimating Thomsen＇s P-wave anisotropic parameters in VTImedia using the first arrivals from walkaway VSP data. Model first-arrival travel times arecalculated on the basis of the near-offset normal moveout correction velocity in VTI mediaand ray tracing using Thomsen＇s P-wave velocity approximation. Then, the anisotropicparameters 0 and e are determined by minimizing the difference between the calculatedand observed travel times for the near and far offsets. Numerical forward modeling, usingthe proposed method indicates that errors between the estimated and measured anisotropicparameters are small. Using field data from an eight-azimuth walkaway VSP in TarimBasin, we estimated the parameters 0 and e and built an anisotropic depth-velocity modelfor prestack depth migration processing of surface 3D seismic data. The results showimprovement in imaging the carbonate reservoirs and minimizing the depth errors of thegeological targets.     

Regularized inversion of amplitude-versus-incidence angle (AVA) based on a piecewise-smooth model

Different from the stacked seismic data, pre-stack data includes abundant information about shear wave and density. Through inversing the shear wave and density information from the pre-stack data, we can determine oil-bearing properties from different incident angles. The state-of-the-art inversion methods obtain either low vertical resolution or lateral discontinuities. However, the practical reservoir generally has sharp discontinuities between different layers in vertically direction and is horizontally smooth. Towards obtaining the practical model, we present an inversion method based on the regularized amplitude-versus-incidence angle (AVA) data to estimate the piecewise-smooth model from pre-stack seismic data. This method considers subsurface stratum as a combination of two parts: a piecewise smooth part and a constant part. To fix the ill-posedness in the inversion, we adopt four terms to define the AVA inversion misfit function: the data misfit itself, a total variation regularization term acting as a sparsing operator for the piecewise constant part, a Tikhonov regularization term acting as a smoothing operator for the smooth part, and the last term to smoothly incorporate a priori information for constraining the magnitude of the estimated model. The proposed method not only can incorporate structure information and a priori model constraint, but also is able to derive into a convex objective function that can be easily minimized using iterative approach. Compared with inversion results of TV and Tikhonov regularization methods, the inverted P-wave velocity, S-wave velocity and density of the proposed method can better delineate the piecewise-smooth characteristic of strata.     

再论地震数据偏移成像

利用地震波正向传播方程对属于波形线性反演问题近似求解方法的地震数据偏移成像进行重新推导,得到了适合散射地震数据的散射偏移成像方法和适合反射地震数据的反射偏移成像方法.以地震波传播的散射理论为出发点,首先根据描述一次散射波正向传播的线性方程研究建立散射地震数据的偏移成像方法理论;利用高频近似对产生散射波场的地下速度扰动函数的空间变化进行近似,推导出地下反射率函数,再由散射波传播方程推导出基于反射率函数的反射波传播方程,然后根据描述一次反射波正向传播的线性方程研究建立反射地震数据的偏移成像方法理论.本文指出和修正了Claerbout偏移成像方法中的不足,提出的地震数据偏移成像方法是对当前偏移成像方法理论的完善,使反射地震数据偏移成像具有了更坚实的数学物理理论基础,得到的偏移成像结果相位正确、位置准确、分辨率提高.     

Refraction traveltime and amplitude corrections for very near- surface inhomogeneities

The performance of refraction inversion methods that employ the principle of refraction migration, whereby traveltimes are laterally migrated by the offset distance (which is the horizontal separation between the point of refraction and the point of detection on the surface), can be adversely affected by very near‐surface inhomogeneities. Even inhomogeneities at single receivers can limit the lateral resolution of detailed seismic velocities in the refractor. The generalized reciprocal method ‘statics’ smoothing method (GRM SSM) is a smoothing rather than a deterministic method for correcting very near‐surface inhomogeneities of limited lateral extent. It is based on the observation that there are only relatively minor differences in the time‐depths to the target refractor computed for a range of XY distances, which is the separation between the reverse and forward traveltimes used to compute the time‐depth. However, any traveltime anomalies, which originate in the near‐surface, migrate laterally with increasing XY distance. Therefore, an average of the time‐depths over a range of XY values preserves the architecture of the refractor, but significantly minimizes the traveltime anomalies originating in the near‐surface. The GRM statics smoothing corrections are obtained by subtracting the average time‐depth values from those computed with a zero XY value. In turn, the corrections are subtracted from the traveltimes, and the GRM algorithms are then re‐applied to the corrected data. Although a single application is generally adequate for most sets of field data, model studies have indicated that several applications of the GRM SSM can be required with severe topographic features, such as escarpments. In addition, very near‐surface inhomogeneities produce anomalous head‐wave amplitudes. An analogous process, using geometric means, can largely correct amplitude anomalies. Furthermore, the coincidence of traveltime and amplitude anomalies indicates that variations in the near‐surface geology, rather than variations in the coupling of the receivers, are a more likely source of the anomalies. The application of the GRM SSM, together with the averaging of the refractor velocity analysis function over a range of XY values, significantly minimizes the generation of artefacts, and facilitates the computation of detailed seismic velocities in the refractor at each receiver. These detailed seismic velocities, together with the GRM SSM‐corrected amplitude products, can facilitate the computation of the ratio of the density in the bedrock to that in the weathered layer. The accuracy of the computed density ratio improves where lateral variations in the seismic velocities in the weathered layer are known.     

INTEGRATED INTERPRETATION, 3D MAP MIGRATION AND VSP MODELLING PROJECT,NORTHERN U.K. SOUTHERN GAS BASIN1

Depth conversion in the northern part of the U.K. Southern Gas Basin is complicated by the presence of Zechstein (Permian) salt swells and diapirs. In addition, the post-Zechstein (post-Permian) section displays large lateral velocity variations. The primary agents which control the velocity of this stratigraphic section are: (1) depth of burial, (2) lithological variation within individual formations, and (3) the effects of subsequent tectonic inversion. An integrated approach which combines well velocity, seismic velocity and seismic interpretation is required for accurate depth estimation. In 1988 Mobil and partners drilled an exploratory well in the northern part of the U.K. Southern Gas Basin. This well was located near the crest of a Zechstein salt diapir. Over 2000 m of Zechstein was encountered in the well. The Permian Rotliegendes objective was penetrated at a depth of over 3700 m. The initial delineation of the objective structure was based on the results of 3D map migration of the seismic time interpretation. Spatially-variant interval velocity functions were used to depth convert through five of the six mapped horizons. Both well and model-based seismic interval velocity analysis information was used to construct these functions. A moving-source well seismic survey was conducted. The survey was run in two critical directions. In conjunction with presurvey modelling, it was possible to confirm immediately the structural configuration as mapped to a distance of 7 km from the well. Post-survey 3D map migration and modelling was employed to further refine the structural interpretation. Although the question of stratigraphic anisotropy was considered in the evaluation of the long offset modelling, no evidence was found in the field data to support a significant effect. Finally, comparisons were made of: curved-ray versus straight-ray migration/modelling, midpoint-depth velocity versus (depth-dependant instantaneous velocity functions, and Hubral- versus Fermat-based map depth migration algorithms. Significant differences in the results were observed for structural dips exceeding 15^o and/or offsets exceeding 6 km. Map depth migration algorithms which employed both curved rays and spatially-variant instantaneous velocity functions were found to best approximate the ‘true’ geological velocity field in the study area.     

地震波干涉偏移及预条件正则化最小二乘偏移成像方法对比

地震波干涉偏移和偏移反演成像是近年来十分活跃的两个研究领域.干涉偏移提供了一个新的地震波数据成像工具,而偏移反演则提供了高逼近度地震成像.二者的共同目的是改善传统直接偏移方法的成像效果,展宽成像区域并提高成像的分辨率.本文研究干涉偏移方法和偏移反演方法对于地震成像效果的影响,探讨二者在提高成像分辨率上的异同.对于偏移反演,通过建立正则化模型,研究了预条件共轭梯度迭代正则化方法及改进措施,并通过绕射点模型数值模拟验证了该方法比直接偏移能够提高振幅的保真度和成像的分辨率.对于干涉偏移和偏移反演这两种方法,对层速度地震模型进行了数值模拟.结果表明干涉偏移和偏移反演成像方法比传统的偏移方法在成像效果上是更加有效的,因而对于实际的地震成像问题很有应用前景.     

Dimensionality‐reduced estimation of primaries by sparse inversion

Wave‐equation based methods, such as the estimation of primaries by sparse inversion, have been successful in the mitigation of the adverse effects of surface‐related multiples on seismic imaging and migration‐velocity analysis. However, the reliance of these methods on multidimensional convolutions with fully sampled data exposes the ‘curse of dimensionality’, which leads to disproportional growth in computational and storage demands when moving to realistic 3D field data. To remove this fundamental impediment, we propose a dimensionality‐reduction technique where the ‘data matrix’ is approximated adaptively by a randomized low‐rank factorization. Compared to conventional methods, which need for each iteration passage through all data possibly requiring on‐the‐fly interpolation, our randomized approach has the advantage that the total number of passes is reduced to only one to three. In addition, the low‐rank matrix factorization leads to considerable reductions in storage and computational costs of the matrix multiplies required by the sparse inversion. Application of the proposed method to two‐dimensional synthetic and real data shows that significant performance improvements in speed and memory use are achievable at a low computational up‐front cost required by the low‐rank factorization.     

MULTI-OFFSET ACOUSTIC INVERSION OF A LATERALLY INVARIANT MEDIUM: APPLICATION TO REAL DATA1

The aim of seismic inversion methods is to obtain quantitative information on the subsurface properties from seismic measurements. However, the potential accuracy of such methods depends strongly on the physical correctness of the mathematical equations used to model the propagation of the seismic waves. In general, the most accurate models involve the full non-linear acoustic or elastic wave equations. Inversion algorithms based on these equations are very CPU intensive. The application of such an algorithm on a real marine CMP gather is demonstrated. The earth model is assumed to be laterally invariant and only acoustic wave phenomena are modelled. A complete acoustic earth model (P-wave velocity and reflectivity as functions of vertical traveltime) is estimated. The inversion algorithm assumes that the seismic waves propagate in 2D. Therefore, an exact method for transforming the real data from 3D to 2D is derived and applied to the data. The time function of the source is estimated from a vertical far-field signature and its applicability is demonstrated by comparing synthetic and real water-bottom reflections. The source scaling factor is chosen such that the false reflection coefficient due to the first water-bottom multiple disappears from the inversion result. In order to speed up the convergence of the algorithm, the following inversion strategy is adopted: an initial smooth velocity model (macromodel) is obtained by applying Dix's equation to the result of a classical velocity analysis, followed by a smoothing operation. The initial reflectivity model is then computed using Gardner's empirical relationship between densities and velocities. In a first inversion step, reflectivity is estimated from small-offset data, keeping the velocity model fixed. In a second step, the initial smooth velocity model, and possibly the reflectivity model, is refined by using larger-offset data. This strategy is very efficient. In the first step, only ten iterations with a quasi-Newton algorithm are necessary in order to obtain an excellent convergence. The data window was 0–2.8 s, the maximum offset was 250 m, and the residual energy after the first inversion step was only 5% of the energy of the observed data. When the earth model estimated in the first inversion step is used to model data at moderate offsets (900 m, time window 0.0–1.1 s), the data fit is very good. In the second step, only a small improvement in the data fit could be obtained, and the convergence was slow. This is probably due to the strong non-linearity of the inversion problem with respect to the velocity model. Nevertheless, the final residual energy for the moderate offsets was only 11%. The estimated model was compared to sonic and density logs obtained from a nearby well. The comparison indicated that the present algorithm can be used to estimate normal incidence reflectivity from real data with good accuracy, provided that absorption phenomena play a minor role in the depth interval considered. If details in the velocity model are required, large offsets and an elastic inversion algorithm should be used.