基于波动方程表面多次波预测与 自适应相减方法研究

多次波预测与自适应相减是基于波动方程表面多次波压制的两个重要环节.文中利用具有并行计算优势的GPU加速表面多次波预测,使得预测效率大为提高.在自适应相减算法中,文中将预测的多次波道、预测多次波道的Hilbert变换道、预测多次波道的高频重建道、以及它们的平移道用作自适应相减中的多次波模型道.Hilbert变换道用以补偿预测多次波的相位信息,高频重建道用以改善预测多次波的高频信息,补偿频带能量差异.文中在预测和相减过程中均采用迭代算法,迭代预测,可较好地获得多次波的运动学特性,迭代相减,可较好地获得多次波的动力学特性,迭代预测与相减使预测的多次波与地震数据中实际的多次波更好地匹配.将该方法应用于理论模拟的SMAART模型和实际海洋数据中,测试结果表明,该方法预测多次波效率较高,在保持有效波振幅条件下可有效地压制地震数据中的表面多次波.     

Migration methods for imaging different-order multiples

Multiples contain valuable information about the subsurface, and if properly migrated can provide a wider illumination of the subsurface compared to imaging with VSP primary reflections. In this paper we review three different methods for migrating multiples. The first method is model-based, and it is more sensitive to velocity errors than primary migration; the second method uses a semi-natural Green's function for migrating multiples, where part of the traveltimes are computed from the velocity model, and part of the traveltimes (i.e., natural traveltimes) are picked from the data to construct the imaging condition for multiples; the third method uses cross-correlation of traces. The last two methods are preferred in the sense that they are significantly less sensitive to velocity errors and statics because they use “natural data” to construct part of the migration imaging conditions. Compared with the interferometric (i.e., crosscorrelation) imaging method the semi-natural Green's function method is more computationally efficient and is sometimes less prone to migration artifacts. Numerical tests with 2-D and 3-D VSP data show that a wider subsurface coverage, higher-fold and more balanced illumination of the subsurface can be achieved with multiple migration compared with migration of primary reflections only. However, there can be strong interference from multiples with different orders or primaries when multiples of high order are migrated. One possible solution is to filter primaries and different orders of multiples before migration, and another possible solution is least squares migration of all events. A limitation of multiple migration is encountered for subsalt imaging. Here, the multiples must pass through the salt body more than twice, which amplifies the distortion of the image.     

地震多次波成像

地震资料含有各种类型多次波,而传统成像方法仅利用地震一次反射波成像,在地震成像前需将多次波去除.然而,多次波携带了丰富的地下结构信息,多次波偏移能够提供除反射波外的额外地下照明.修改传统逆时偏移方法,用包含一次反射波和多次波的原始记录代替震源子波,将SRME方法预测的表面多次波代替一次反射波作为输入数据,可将表面多次波成像.多次波成像的挑战和困难在于大量串扰噪声的产生,针对表面多次波成像中的成像噪声问题,将最小二乘逆时偏移方法与多次波分阶思想结合起来,发展可控阶数的表面多次波反演成像方法,有望初步实现高精度的表面多次波成像.在消除原始记录中的表面多次波后,通过逆散射级数方法预测得到层间多次波,将层间多次波作为逆时偏移方法的输入数据可将其准确归位到地下反射位置.数值实验表明,多次波成像能够有效地为地下提供额外照明,而可控阶表面多次波最小二乘逆时偏移成像方法几乎完全避免成像噪声.     

Estimating primaries by sparse inversion of the 3D curvelet transform and the L1-norm constraint

In this paper, we built upon the estimating primaries by sparse inversion (EPSI) method. We use the 3D curvelet transform and modify the EPSI method to the sparse inversion of the biconvex optimization and L1-norm regularization, and use alternating optimization to directly estimate the primary reflection coefficients and source wavelet. The 3D curvelet transform is used as a sparseness constraint when inverting the primary reflection coefficients, which results in avoiding the prediction subtraction process in the surface-related multiples elimination (SRME) method. The proposed method not only reduces the damage to the effective waves but also improves the elimination of multiples. It is also a wave equationbased method for elimination of surface multiple reflections, which effectively removes surface multiples under complex submarine conditions.     

衰减干涉假象的海洋一次波与多次波同时成像方法

海水与空气间的强波阻抗差使得海洋地震资料普遍发育自由表面相关多次波,如何利用好多次波所携带的有效信息已成为提高海洋地震资料成像品质的新突破点.基于面炮偏移的一次波与多次波同时成像方法能够避免多次波预测精度的影响,但是,正向传播的震源子波与反向延拓的自由表面相关多次波所产生的干涉假象严重制约了该技术的应用,本文提出了一种基于单程波偏移算子,可在成像域压制干涉假象的一次波与多次波同时成像方法.其中包含了三个步骤:第一,传统单程波偏移成像方法中的震源子波替换为一次波、多次波与震源子波,初始上行延拓波场为一次波与多次波,基于单程波算子的波场延拓与互相关成像条件的应用得到包含干涉假象的一次波与多次波同时成像;第二,以子波为震源,自由表面相关多次波为记录,按照传统单程波偏移成像方法得到干涉假象;第三,基于最小二乘匹配滤波算法,将第一步的成像结果与第二步的干涉假象进行匹配相减,得到干涉假象衰减后的一次波与多次波同时成像,避开了由于实际资料子波无法准确提取而造成一次波与多次波对成像能量级的不一致性.Sigsbee2B模型测试验证了本方法的有效性,并在我国某探区深海实际资料处理中得到了成功应用,深层基底得到了清晰刻画,并且照明均衡度明显改善.     

改进的多道预测算子压制浅水多次波方法

在浅水情况下,由于观测数据中缺少近偏移距信息,水层多次波的压制面临挑战.利用多道预测算子压制水层多次波是浅水环境下压制多次波的重要方法之一,这种方法先从输入数据中估计出多道预测算子,再将预测算子和输入数据做褶积预测出水层相关多次波.然而,估计的多道预测算子很容易受噪声污染,从而影响多次波模型的精度.所以,我们提出了改进的多道预测算子压制浅水多次波方法.该方法先从数据中估计出多道预测算子,并利用估计的算子构建出精确的水层模型;然后,通过计算算子的走时信息、估计振幅信息、合成新算子三个步骤来修正原始的多道预测算子.修正的算子不仅不受噪声影响,还含有精确的走时信息、可靠的振幅信息;最后,该方法用修正的算子来预测多次波,并结合自适应相减,将预测的多次波从输入数据中去除.通过合成数据和实际资料的验证表明,相比于原始的多道预测算子压制多次波方法,改进的方法能够取得更好的压制效果.     

3D targeted multiple attenuation

Short-period multiple reflections pose a particular problem in the North Sea where predictive deconvolution is often only partially successful. The targeted multiple attenuation (TMA) algorithm comprises computation of the covariance matrix of preflattened prestack or post-stack seismic data, the determination of the dominating eigenvectors of the covariance matrix, and subtraction of the related eigenimages followed by reverse flattening. The main assumption made is that the flattened multiple reflections may be represented by the first eigenimage(s) which implies that the spatial amplitude variations of primaries and associated multiples are similar. This assumption usually limits the method to short-period multiple reflections. TMA is applicable post-stack or prestack to common-offset gathers. It is computationally fast, robust towards random noise, irregular geometry and spatial aliasing, and it preserves the amplitudes of primaries provided they are not parallel to the targeted multiples. Application of TMA to 3D wavefields is preferable because this allows a better discrimination between primaries and multiples. Real data examples show that the danger of partially removing primary energy can be reduced by improving the raw multiple model that is based on eigenimages, for example by prediction filtering.     

改进的1范数匹配滤波法及在南海深水盆地的应用

地震勘探中多次波的存在会影响速度模型的建立、地震成像效果、反演和解释过程中的介质属性提取等,因此多次波的压制成为至关重要的研究课题.本文采用改进的1范数匹配滤波法,基于1范数对大值条件的开放性,利用预测出的多次波模型Hilbert变换道及求导道,从理论上克服2范数的大值条件,并改善常规1范数下的正交性条件,在一次波和多次波叠合之处及能量差异大的情况下有效压制多次波.模型及南海深水盆地实际海洋地震数据的处理结果显示了改进的1范数匹配方法在压制多次波方面相对于常规2范数、1范数的优越性.     

基于自适应变步长波场延拓的可控层分阶层间多次波模拟

地下低速夹层的存在导致地震数据中包含较强能量的层间多次波,有效识别和预测深部储层上覆地层产生的层间多次波是提高深部储层解释精度的重要环节,而准确模拟层间多次波是辅助识别地震数据中层间多次波的一种非常有效的方法.本文提出了一种基于自适应变步长波场延拓的可控地层分阶层间多次波模拟方法,该方法基于自适应变步长波场延拓,以递归循环的方式实现分阶层间多次波的模拟.通过对模型添加双重层位约束,可以模拟指定地层产生的各阶层间多次波.利用二维反周期延拓方法压制波场延拓的边界反射优于传统方法,例如吸收边界法.提出自适应变步长波场延拓技术,大大提升了波场模拟的效率.理论和数值例子表明,本文方法模拟的一次波和各阶层间多次波与常用的有限差分方法模拟结果具有很好的一致性,且克服了有限差分方法无法分阶模拟波场的不足,显著提升了层间多次波识别的效率.     

ACOUSTIC MODELING OF THE SEAFLOOR*

Data from routine seismic surveys contain considerable information about the geo-acoustic properties of the seafloor. Waves are reflected at a wide range of angles of incidence from near-vertical reflections (higher multiples) to supercritical reflections (primary and lower multiples). The reflection coefficient is approximately constant for small angles of incidence (< 10°) but varies greatly for larger angles of incidence. Near-vertical reflections are used to determine the seafloor density. The P-velocity in the seafloor is determined in advance from the critical distance using the amplitude variation of the primary as well as the multiples. The V_p/V_S ratio is determined by modeling the amplitude variation with the angle of incidence. The primary reflection from the seafloor and the first three multiples are included in the modeling. Seismic data obtained with both conventional and superlong airgun arrays have been modeled. Data collected from the Barents Sea show that even if the P-velocity is the same at different sites, the V_p/V_s ratio, density and Poisson's ratio vary significantly. The most extreme example shows that for a P-velocity of 2.80 km/s the V_p/V_s ratio varies between 1.9 and 6.0. The corresponding densities vary from 2.36 g/cm³ to 1.80 g/cm³ and the Poisson's ratio varies from 0.31 to 0.49. The acoustic modeling offers a method of assessing the mean geotechnical or mechanical properties of larger volumes of marine sediments in terms of incompressibility, shear modulus and Poisson's ratio.     

SLOPE DATA WATER-BOTTOM MULTIPLE ATTENUATION1

Methods for predicting and attenuating water-bottom multiples by wavefield extrapolation have been discussed by several investigators. Because these prediction methods operate on shot records, boundary conditions must be specified for every shot record. The approach presented operates in the common-offset plane; a model of expected water-bottom multiples is generated from the observed surface wavefield using a finite-difference wave-equation migration algorithm with an offset term. An accurate water-depth profile is required, but there is no restriction on the shape of the water bottom other than a dip limit of approximately 18–20°. In generating a multiple model, the water-bottom primary and each water-bottom multiple reflection of the observed surface wavefield are extrapolated to a higher order. Thus, the extrapolated water-bottom primary of the model is lined up with a water-bottom multiple in the data and each multiple in the model is lined up with a higher-order (or later) multiple in the data. Prestack multiple attenuation is achieved, for one offset at a time, by first adapting the model of expected multiples to the observed data and then subtracting the predicted multiple energy. An error-constrained adaptation algorithm is proposed in order to control instabilities. No assumptions are made about primary reflections and no subwater-bottom velocities are required. Computational efficiency of modelling and adaptation can be improved by applying this method only to near and intermediate offsets as the stacking process usually provides sufficient multiple attenuation at far offsets. A field data example demonstrates the potential of the proposed method for improving the primary-to-multiple ratio in prestack and post-stack data.     

DECOMPOSITION OF MULTICOMPONENT SEISMIC DATA INTO PRIMARY P- AND S-WAVE RESPONSES1

Inversion of multicomponent seismic data can be subdivided in three main processes: (1) Surface-related preprocessing (decomposition of the multicomponent data into ‘primary’ P-and S-wave responses). (2) Prestack migration of the primary P- and S-wave responses, yielding the (angle-dependent) P-P, P-S, S-P and S-S reflectivity of the subsurface. (3) Target-related post-processing (transformation of the reflectivity into the rock and pore parameters in the target). This paper deals with the theoretical aspects of surface-related preprocessing. In a multicomponent seismic data set the P- and S-wave responses of the subsurface are distorted by two main causes: (1) The seismic vibrators always radiate a mixture of P- and S-waves into the subsurface. Similarly, the geophones always measure a mixture of P- and S-waves. (2) The free surface reflects any upgoing wave fully back into the subsurface. This gives rise to strong multiple reflections, including conversions. Therefore, surface-related preprocessing consists of two steps: (1)Decomposition of the multicomponent data (pseudo P- and S-wave responses) into true P- and S-wave responses. In practice this procedure involves (a) decomposition per common shot record of the particle velocity vector into scalar upgoing P- and S-waves, followed by (b) decomposition per common receiver record of the traction vector into scalar downgoing P- and S-waves. (2) Elimination of the surface-related multiple reflections and conversions. In this procedure the free surface is replaced by a reflection-free surface. The effect is that we obtain ‘primary’ P-and S-wave responses, that contain internal multiples only. An interesting aspect of the procedure is that no knowledge of the subsurface is required. In fact, the subsurface may have any degree of complexity. Both the decomposition step and the multiple elimination step are fully determined by the medium parameters at the free surface only. After surface-related preprocessing, the scalar P- and S-wave responses can be further processed independently by existing scalar algorithms.     

Extending illumination using all multiples: application to 3D acquisition geometry analysis

Recent advances in survey design have led to conventional common‐midpoint‐based analysis being replaced by subsurface‐based seismic acquisition analysis, with emphasis on advanced techniques of illumination analysis. Among them is the so‐called focal beam method, which is a wave‐equation‐based seismic illumination analysis method. The objective of the focal beam method is to provide a quantitative insight into the combined influence of acquisition geometry, overburden structure, and migration operators on the resolution and angle‐dependent amplitude fidelity of the image. The method distinguishes between illumination and sensing capability of a particular acquisition geometry by computing the focal source beam and the focal detector beam, respectively. Sensing is related to the detection properties of a detector configuration, whereas illumination is related to the emission properties of a source configuration. The focal source beam analyses the incident wavefield at a specific subsurface grid point from all available sources, whereas the focal detector beam analyses the sensing wavefield reaching at the detector locations from the same subsurface grid point. In the past, this method could only address illumination by primary reflections. In this paper, we will extend the concept of the focal beam method to incorporate the illumination due to the surface and internal multiples. This in fact complies with the trend of including multiples in the imaging process. Multiple reflections can illuminate a target location from other angles compared with primary reflections, resulting in a higher resolution and an improved illumination. We demonstrate how an acquisition‐related footprint can be corrected using both the surface and the internal multiples.     

不同阶次自由表面相关多次波预测与成像方法

海水与空气间的强波阻抗界面使得海洋地震数据普遍发育自由表面相关多次波,多次波信息的利用是提高海洋地震资料成像品质的新突破点.近年来发展了一系列多次波成像方法,干涉假象是制约其应用推广的关键问题之一.为了避免假象影响,本文提出了不同阶次自由表面相关多次波预测与成像方法,首先,修改了传统SRME(表面相关多次波衰减)方法中的边界条件,通过多次波升阶次与匹配相减的方法预测出不同阶次自由表面相关多次波;其次,基于单程波偏移算子和"面炮"偏移策略,以一次反射波或第(N-1)阶自由表面相关多次波为下行波场正向延拓,以第1阶多次或第N阶多次波为上行波场逆向延拓,并在每一层互相关成像得到第1阶或N阶多次波单独成像.本方法避免了低阶多次波和高阶多次波产生的相关假象,且相对于全波算子的偏移方法具有较高的计算效率,增强了多次波成像方法的实用性.单层模型和三层模型测试验证了本方法的正确性,并在我国某深海探区实际资料处理中得到了成功应用.相对于传统一次波成像,分阶次多次波成像具有更高的照明均衡度、垂向分辨率和信噪比.本研究表明,海洋多次波成像是一次波成像的有力补充,对于稳定海底沉积的深海地区,具有一定的应用前景.     

数据自相关多次波偏移成像

在常规偏移方法中一般都需要压制地震数据中的多次波,仅利用一次波信息成像,把自由表面反射的多次波视为噪声,但是在多次波中也包含着地下结构信息,应该将其充分利用到成像中来.事实上,已经有不少成像方法试图利用多次波信息,但是大部分方法都需要对多次波进行预测.本文提出了基于傅里叶有限差分偏移算子的数据自相关偏移方法.在这种偏移方法中,对含有一次波和多次波的地震数据,分别进行下行和上行延拓,然后直接利用常规的互相关成像条件成像.由于波场延拓采用了傅里叶有限差分算子,其计算效率高,能够很好地对复杂介质中的地震数据进行延拓.在数值试验中,使用了一个含散射点的三层模型和Marmousi模型.合成数据测试结果表明,这种方法可以对更大范围的地下构造成像,比常规的只利用一次波的傅里叶有限差分法照明度更好,并且在浅层可以提供更高的分辨率.我们提出的数据自相关策略易于实现且避免了繁杂的多次波预测,这对于复杂地下构造成像可能有着重大意义.     

Using surface multiples to estimate primaries by sparse inversion from blended data

For data acquired with conventional acquisition techniques, surface multiples are usually considered as noise events that obscure the primaries. However, in this paper we demonstrate that for the situation of blended acquisition, meaning that different sources are shooting in a time‐overlapping fashion, multiples can be used to ‘deblend’ the seismic measurements. We utilize the recently introduced estimation of primaries by sparse inversion (EPSI) methodology, in which the primary impulse responses are considered to be the unknowns in a large‐scale inversion process. With some modifications the estimation of primaries by sparse inversion method can be used for blended seismic data. As output this process gives unblended primary impulse responses with point sources and receivers at the surface, which can be used directly in traditional imaging schemes. It turns out that extra information is needed to improve on the deblending of events that do not have much associated multiple energy in the data, such as steep events at large offsets. We demonstrate that this information can be brought in during acquisition and during processing. The methodology is illustrated on 2D synthetic data.     

IS IT POSSIBLE TO INCREASE THE RESOLUTION IN SEISMIC EXPLORATION FOR COAL BY USING HIGH FREQUENCY SIGNALS?*

In seismic exploration for coal data resolution is a fundamental problem. Modeling helps to understand those details of the geology that can be interpreted from the seismic image. For single seam exploration, the vertical resolution of a seismic section is defined by the bandwidth of the signals. If there are several seams, each seam acts as a high-pass filter for reflections and as a low-pass filter for transmitted waves. Synthetic seismograms show that reflections from deep seams have a low frequency content. Within a layered sequence of coal seams, many multiples are generated which disturb later primary reflections. The ratio of primaries to multiples depends on the frequency content of the seismic data and on the number of overlying seams. The multiple problem is more severe with high frequencies. Primary reflections from deep coal seams within a sequence can be detected only if low-frequency signals are used. However, the use of low-frequency signals reduces the resolution of the deeper data.     

稀疏反演多次波去除策略与效果分析

多次波去除是海上勘探中的一项关键技术,在一定程度上决定着最终成像结果的质量.表面相关多次波去除方法(Surface-Related Multiple Elimination,SRME)是一种基于反馈迭代模型的数据驱动类方法,该方法利用地震数据本身预测多次波,进一步对其进行匹配相减获得一次波结果,这往往导致与多次波重叠的一次波被错误地去除.稀疏反演一次波估计方法(Estimation of Primaries by Sparsity Inversion,EPSI)在原理上同样基于反馈迭代模型,通过求解一次波和表面相关多次波总残差最小化问题直接获得一次波.该方法无多次波的匹配相减过程,可减少与多次波发生重叠的一次波产生损失.另外,EPSI可以很好地解决SRME无法解决的近道缺失问题.理论数据和实际资料处理对比表明SRME对数据完整性以及一次波和多次波的分布情况有较强的依赖性,而EPSI几乎不受这两个因素的影响.EPSI相比SRME方法计算成本高,尚未在工业界广泛应用,但随着高性能计算的普及,该方法的实用性有望得到突破.     

Filtering vibroseis data in the precorrelation domain

Vibroseis data recorded at short source–receiver offsets can be swamped by direct waves from the source. The signal-to-noise ratio, where primary reflections are the signal and correlation side lobes are the noise, decreases with time and late reflection events are overwhelmed. This leads to low seismic resolution on the vibroseis correlogram. A new precorrelation filtering approach is proposed to suppress correlation noise. It is the ‘squeeze-filter-unsqueeze’ (SFU) process, a combination of ‘squeeze’ and ‘unsqueeze’ (S and U) transformations, together with the application of either an optimum least-squares filter or a linear recursive notch filter. SFU processing provides excellent direct wave removal if the onset time of the direct wave is known precisely, but when the correlation recognition method used to search for the first arrival fails, the SFU filtering will also fail. If the tapers of the source sweeps are badly distorted, a harmonic distortion will be introduced into the SFU-filtered trace. SFU appears to be more suitable for low-noise vibroseis data, and more effective when we know the sweep tapers exactly. SFU requires uncorrelated data, and is thus cpu intensive, but since it is automatic, it is not labour intensive. With non-linear sweeps, there are two approaches to the S,U transformations in SFU. The first requires the non-linear analytical sweep formula, and the second is to search and pick the zero nodes on the recorded pilot trace and then carry out the S,U transformations directly without requiring the algorithm or formula by which the sweep was generated. The latter method is also valid for vibroseis data with a linear sweep. SFU may be applied to the removal of any undesired signal, as long as the exact onset time of the unwanted signal in the precorrelation domain is known or determinable.