When oil-based mud invades into a formation,the resistivity distribution and response characteristics of the invaded zone are complex and cannot be represented by a simple radial two-layer step model.Considering the actual structure and dimensions of the array induction logging tool,this study derives the expressions for the induced electromotive force and effective tool constant of the three-coil measurement structure with a built-in copper tube.The outer diameter of the copper tube is adjusted depending on the derived expression of the effec-tive tool constant and the effective tool constant of the actual tool is simulated to approximate the radial detection characteristics of the actual tool.Subsequently,a forward model for the radial layered medium model of array in-duction logging is constructed.Based on the analysis of invasion response characteristics under oil-based mud conditions in the fractured tight sandstone formation of the Cretaceous Bashijiqike Formation in the Dabei and Keshen areas of the Kuqa Depression in the Tarim Basin,the radial step invasion model is selected for forward modeling of the complex invasion response characteristics of array induction logging in oil-based mud.The study shows that the radial three-layer step model can reasonably represent the complex response characteristics of for-mation invaded by oil-based mud,laying a foundation for the subsequent inversion application of array induction logging data.
The actual underground medium is very complex. There are inevitable errors in the numerical forward modeling results by employing computer software, and it is difficult to simulate the complex underground medium. The dynamic photoelasticity method can obtain the propagation characteristics of wave fields at the seismic scale in the laboratory, and can observe the wave field characteristics in real time. The wave field characteristics in solid media are obtained by dynamic photoelastic physical simulation device experiment, and the propagation law of ultrasonic pulse waves in media is observed. By carrying out the sample model pixel calibration of the instantaneous wave field image and longitudinal pixel sampling of the wave field image, the simulated single shot seismic record in solid media can be obtained. Comparison of the simulated single shot seismic records of the non-hole medium model and containing-hole medium model shows that there are obvious differences in waveform characteristics between the two models. By comparing the simulated single shot seismic record with the actual field single shot seismic record, it is found that the two have sound correspondence, which shows that the laboratory high-frequency small-scale and field low-frequency large-scale can also reflect the same internal physical essence without meeting the fixed ratio observation theory.
In order to study the propagation law of seismic waves and the generation mechanism of noise in the loess plateau,this paper establishes 3D realistic models(velocity and quality factor Q)of the loess plateau in Or-dos basin according to actual surface elevation and velocity structure of the Qingcheng North 3D work area.Then under the condition of undulating land surface,3D numerical modelling using the coordinate transforma-tion method is accomplished based on a viscoacoustic wave equation with explicitly expressed quality factor.This method can simulate the strong energy noise near the offset caused by complex near surface and strong at-tenuation effects,and obtain simulation results similar to the real seismic data.Through wave field analysis of the numerical modelling results of the models with different complexity,this paper gradually clarifies the genera-tion mechanism of near-offset strong energy noise,multiple reflection and multiple refraction.Finally,this paper analyzes the differences in wave field characteristics at different locations(plateau,ridge,hillside,ditch)and at different source depth in the loess plateau area based on the numerical modelling and actual data.Analysis shows that when excited in the ditch,the near-offset strong energy noise is weaker,and the data quality is rela-tively good;As the source depth increases,the high-frequency component of the signal gradually increases;For real seismic data acquiring in the loess plateau,it is necessary to avoid excitation in the dry loess layer and try to excite in the clay layer or deeper layer.The conclusion of wave field analysis has positive guiding significance for real data acquiring and processing in the loess plateau.
Full waveform inversion(FWI)can establish high-precision velocity models of subsurface media.Due to geophone limitations,towed-streamer data often lacks reliable low-frequency components.Conse-quently,FWI based on towed streamer data often encounters the cycle skipping problem,resulting in the objec-tive function converging to a local extremum.Ocean bottom node(OBN)seismic data contains more low-fre-quency components compared to towed-streamer data.Combining OBN data with towed-streamer data in wave-form inversion can compensate for the low-frequency deficiency in towed-streamer data.However,the source wavelet of towed-streamer data is different from that of OBN data,and the joint waveform inversion is more se-verely affected by the source wavelet.This paper proposes a joint waveform inversion method for towed-streamer and OBN data that does not depend on source wavelets.When constructing the objective function,the simulated and observed data are respectively convolved with the average traces of the other,solving the problem of wavelet error in joint waveform inversion and achieving the effect of not relying on the source wavelet.The feasibility and effectiveness of the proposed method are verified through model testing using the Marmousi model synthesized towed-streamer data and OBN data.
The high-resolution and small trace marine seismic exploration method has the characteristics of high dominant frequency and wide frequency band, which is an important means for the study of near-seabed stratigraphic sequence and fine structure, gas hydrate exploration and so on. The particularity of acquisition and the inherent problems of equipment cause unequal floating effect of the cable, resulting in unevenness, even bending and shaking of the in-phase axis on the gathers after NMO correction，which affects the effect of in-phase stacking and reduces the resolution and imaging accuracy of seismic data processing. The conventional residual moveout correction methods such as: interactive pick-up seabed reflection and ghost reflection traveltime fitting method, gather residual moveout correction method, etc., due to the large amount of pick-up, high quality requirements of gathers, and the residual moveout is a comprehensive reflection of the cable unequal floating, inaccurate speed and anisotropy, it is difficult to accurately identify whether residual moveout is caused by the unequal floating of the cable, which will cause errors in the correction results. This paper proposes a cable floating correction method based on ghost reflection traveltime controlled by coherent analysis. Firstly, the ghost reflection traveltime of the seabed reflection and receiver points are interactively picked up on the shot gathers in intervals, the information of all shots is obtained through fitting interpolation, and the real-time depth of the cable is obtained based on the moveout, and the initial moveout correction amount is calculated; Then the reference moveout correction amount is obtained through a two-dimensional coherent algorithm based on model trace constraints. Finally, based on the cross-correlation coefficient, a weighted selection is made to obtain an accurate residual moveout of unequal cable floating, and the moveout is applied after smoothing, the effect of in-phase stacking is achieved, thereby improving the imaging quality of seismic data. The test analysis combining synthetic data and actual processing shows that, compared with the conventional correction methods, the method proposed in this paper is more accurate and effective.
对于物性参数反演这类存在多解性的问题,常见的方法往往难以高效、高精度地获得最优解.因此,根据贝叶斯统计学,采用高斯先验模型引入测井数据约束,并发展储层物性参数贝叶斯反演方法,以克服反演的多解性.其次,在BISQ理论和等效流体理论的基础上建立了岩石物理正演模型,该模型反映了孔隙介质中固体和流体的多种物理作用,可精细地刻画储层中地震波的传播过程,进而提升反演结果的可靠性.此外,由于确定性优化算法通常只有局部收敛能力,无法收敛到反演问题的全局最优解,所以引入具有全局收敛能力的杂交遗传算法,以提高反演方法的精度和效率.最终,利用所提反演方法预测中国东部A油田目标区域的孔隙度和含水饱和度,获得了较好效果.
分析了卷积神经网络(CNN)、去噪卷积神经网络(DnCNN)、U-Net深度神经网络、前反馈(BP)神经网络、空洞卷积神经网络(DCNN)、残差网络(ResNet)、迁移学习等为代表的深度学习方法的概念、发展现状、方法原理、去噪效果以及优缺点等;对比了传统去噪方法、字典学习及深度学习方法的去噪效果;展望了深度学习技术在地震去噪领域的发展前景.获得以下认识:①深度学习方法的实际去噪效果优于传统方法和字典学习方法,不需要设定结构模型,泛化性更强,且计算时间短、精度更高.②深度学习方法存在诸多不足:实际数据的去噪效果往往差于合成数据;普适性不强;神经网络的"黑匣子"特性使其物理可解释性大大降低;网络性能与训练数据的泛化性密切相关;用于训练网络的数据集因人而异,难以评价网络性能.③期待深度学习在以下方面取得进展和突破:搭建适用于不同噪声的去噪神经网络结构,并将更优的网络结构引入地震随机噪声压制;将地震信号转换到变换域构造网络的损失函数;改进学习策略的同时制作更具代表性的数据集,尽可能地使训练数据覆盖所有类型的解,提高网络泛化性;自动化的参数调优;结合模型驱动与数据驱动的方法.
各向异性介质拟声波方程正演模拟及逆时偏移存在伪横波干扰及数值不稳定现象,成像质量不佳;同时计算效率也是各向异性逆时偏移的重要影响因素.为此,提出一种兼具成像质量和计算效率的纯qP波逆时偏移成像方法.首先,从波动方程伪解析解和各向异性精确频散关系出发,分别构建了VTI和TTI介质纯qP波伪解析解波场延拓公式,避免了推导显式波动方程,无需对频散关系做平方处理,由此消除了波场模拟存在的伪横波干扰;然后,通过设计三维球型或二维圆形差分模板,并借助低秩有限差分法求解与模型相适应的差分系数,推导了基于低秩有限差分的纯qP波波场延拓公式,在时间方向延拓时,无需进行多次Fourier变换,降低了计算复杂度;最后,在此基础上,利用GPU并行计算正、反向地震波场,提高了逆时偏移的成像效率.典型二维和三维模型试算结果表明:该方法能够消除伪横波干扰,保证计算过程稳定,在二维和三维各向异性介质中都具有较强的适应性和较高的计算效率.
在老油田开发中后期,精准描述剩余油分布成为亟待解决的难题.叠前地震资料包含不同入射角的道集信息,通过分析振幅随入射角的变化特征可预测储层岩性和流体分布规律.为此,以大庆长垣BSX区块为例,在含流体岩石物理特征分析基础上,开展相控叠前地质统计学反演预测剩余油分布特征,以寻找潜力部位,实施精准挖潜.首先,明确随注水开发、含水饱合度逐渐增大引起的纵、横波响应特征,建立纵横波速度比、纵波阻抗的弹性流体岩石物理定量解释图板;其次,以纵横波速度比、纵波阻抗与含水(油)饱和度之间的关系为纽带,采用相控蒙特卡洛反演算法,生成高分辨率的岩石弹性参数体、含油饱和度体.应用结果表明,剩余油的分布预测结果与取心井岩心实测数据符合率达76.5％.该方法指导了BSX区5种剩余油的挖潜对策,可为其他陆相油田的高效开发提供借鉴.
为了深入分析激电法差分装置的响应特征,提出一种高精度、高效率的三维数值模拟方法.该数值模拟方法通过二维傅里叶变换将三维激电数值模拟问题转化为波数域一维数值模拟问题,使用一维有限单元法求解,同时引入算子进行迭代,最终获得高精度数值解.该算法可避免进行大型稀疏线性方程组的直接求解,大幅度降低了计算量和存储需求,提高了计算效率.以提出的三维数值模拟方法为工具,分析激电法差分装置响应特征.基于高阻/低阻的高极化异常体模型测试结果表明,单孔井—井测量下差分装置对井旁盲矿的分辨效果优于其他装置,且低阻体较高阻体对电流具有更强的敏感度;基于不同背景介质下的低阻高极化异常体模型的测试结果表明,井—地测量下采用差分装置得到的视电阻率和极化率比采用二极装置得到的视电阻率和极化率能更准确地反映背景介质和异常体在垂向上的空间分布特征.因此,激电法差分装置勘探效果优于其他方法装置,研究结论可为野外勘探测量装置的选择提供参考.
目前基于激光雷达(LiDAR)数据分析陆地重力近区、中区地形改正方法及其影响因素的文献较少.以往有关地形改正方法的精度评价都是基于某一个数学模型的假设条件,只能讨论高程测量误差对地形改正的影响,不能分析由扇形锥、扇形柱等数学模型产生的误差.为此,采用美国地调局三维高程计划(3DEP)的1m×1m间距LiDAR高程数据,选择12个数据区作为研究对象,以直棱柱模型重力解析式计算结果为地形改正参考值,从统计角度对比以往近区和中区地形改正数学模型与计算方法的精度——主要创新点;在此基础上,分析不同计算方法、不同地形网格间距的差异,并且对比不同方法的计算效率.结果表明:①近区地形改正范围与计算模型是影响近区地形改正计算的主要因素.混合模型方案可有效降低计算误差,当近区地形改正半径为20m时,采用方案V可以保证K区以外的11个区的平均相对误差小于20％;当近区地形改正半径为50m时,采用方案Ⅷ可以将全区的平均相对误差控制在13％以内,是计算精度最高的方案.②地形起伏较平缓的丘陵(K区)的差异最大值与中山(J区)相当;在复杂地表的露天铜矿(L区),7种计算方案的差异最大值处于相同量级.③补角地形改正的修正公式不适用于丘陵,区域重力调查规范中的内接口补角地形改正的简化公式的适用性更强.④影响中区地形改正计算结果的因素包括高程数据网格间距、中区地形改正范围和计算方法,其中网格间距是主要因素.⑤综合计算精度和计算时间两个因素,质量线模型法为中区地形改正的首选计算方法.
叠前地震反演技术可以获取多种弹性参数的空间展布特征,是识别致密砂岩甜点的一种重要手段.致密砂岩的甜点均位于内部,砂体与甜点往往具有不同的敏感弹性参数,因而需要进行反演参数之间的转换,这往往降低了预测结果的精度和效率.为此,利用叠前地震数据,提出了一种基于贝叶斯概率反演的致密砂岩甜点直接预测方法.首先,推导岩性与甜点敏感弹性参数的Zeoppritz近似方程,结合褶积模型构建敏感参数与地震记录之间的关系;然后,在贝叶斯理论的框架下,建立敏感参数的后验概率分布表达式,利用马尔科夫链—蒙特卡洛算法对后验概率分布进行抽样,并在抽样过程中充分考虑反演参数之间的相关性;最后,利用改进的贝叶斯概率反演方法实现致密砂岩甜点的直接预测.该方法引入基于条件概率分布的抽样策略以约束反演参数的采样空间,有效提高了预测结果的精度和效率.模型数据和实际数据的测试均验证了该方法的可行性.
最小结构模型约束正则化二维地震初至旅行时反演中存在模型边界刻画不清的问题,尤其是地质体内射线分布稀疏的情况下,反演效果不理想.为此,引入模糊C均值(FCM)聚类模型约束函数,旨在提高反演结果对模型边界的成像精度.该约束项将先验信息作为参考聚类中心,在迭代过程通过反复修改聚类中心及每个网格单元对聚类中心的隶属度,实现对速度的自动分类.在此基础上,采用以模型灵敏度信息为依据的多重网格反演策略,以提高反演的稳定性及效果;应用简单模型讨论了FCM聚类模型约束权重、先验信息引导项权重等参数选取方案;对比无监督学习与先验信息监督学习的反演效果,后者改善了反演速度模型边界刻画模糊现象,有效提高了反演结果的分辨率;最后,通过实测数据反演,验证该方法在实际应用中的实用性和有效性.
用常规交错网格有限差分法(C-SFD)进行速度—应力声波方程数值模拟,虽然空间差分算子能达到2M阶差分精度(M表示空间差分算子中与差分中心点等距的坐标轴网格点的组数),但差分离散声波方程仅具有2阶差分精度,所以其模拟精度低、稳定性差.为此,联合利用坐标轴网格点和非坐标轴网格点构建空间差分算子近似一阶空间偏导数,建立了一种适用于速度—应力声波方程的混合交错网格有限差分法(M-SFD),并基于时空域频散关系和泰勒级数展开建立差分系数求解方程组,导出了差分系数通解.M-SFD给出的差分离散声波方程可达到4、6、8阶、甚至任意偶数阶差分精度.频散分析表明:在特定的Courant条件数取值(如Courant条件数r=0.3)条件下,C-SFD通过调整M的取值,无法将数值频散误差控制在1％以内;M-SFD通过调整M和N的取值(N表示空间差分算子中与差分中心点等距的非坐标轴网格点的组数),基本可以将频散误差控制在1‰以内,甚至可以控制在0.1‰以内.稳定性分析显示:M取值相同时,M-SFD的稳定性强于C-SFD.数值模拟实例表明:计算效率基本相同时,M-SFD比C-SFD能更有效地压制数值频散,模拟精度更高;M-SFD还能采用比C-SFD更大的时间采样间隔以获得更高的计算效率,且模拟精度更高.进一步将M-SFD推广应用于逆时偏移,M-SFD作为逆时偏移中的波场传播算子,能够有效消除由于数值频散造成的成像假象,从而提高深层的构造成像精度和分辨率.
地下介质的吸收衰减效应严重降低了地震信号的探测深度,限制了高分辨率地震勘探的发展.反Q滤波技术是补偿地层吸收衰减、提高地震分辨率的重要手段.然而,传统的反Q滤波方法本身并不具备信号识别能力,在补偿地震信号的同时会放大噪声干扰.为此,文中提出一种信号自适应识别的多道吸收补偿方法,其核心思想是利用地震信号的空间构造特征提升吸收补偿算法的稳定性和精度.该方法主要实施流程包括四步:首先,利用空间预测算子对反射信号的空间构造特征进行数学表达;再根据信号预测算子构造空间反射结构约束项;然后,将反射结构约束项加入多道吸收补偿目标函数,得到吸收补偿反演系统;最后,求解多道反演目标函数,得到吸收补偿结果.模型数据测试和实际资料应用结果表明:所提方法不仅能提高地震资料的分辨率,还可保护补偿结果的空间构造特征和信噪比,具有较强的稳定性和一定的实用性.
对重磁资料进行综合处理与解释有助于降低多解性、提高解释的可靠性.重磁对应分析法是重磁联合定性解释的一种常用方法,但该方法会受剩磁影响,易产生虚假信息,因而实用性不强.为此,从重力异常二阶垂向导数与归一化磁源强度表达式出发,推导出基于重力二阶垂向导数和归一化磁源强度的重磁泊松公式,提出了一种新的重磁相关性分析方法,并采用噪声扰动提高重磁相关度的有效性.该方法物理意义明确,易于实现,无需化极处理,且受重磁背景场影响小.模型试验表明,常规重磁对应分析法仅在噪声扰动下重磁完全同源时具有良好效果,在重磁非完全同源时会出现虚假信息;新方法能够较好地反映重磁是否同源及同源位置,在复杂情况下也能获得良好的效果.将新方法应用于江西相山铀矿田重磁资料处理和解释,准确地反映了碎斑熔岩、流纹英安岩及花岗斑岩这三类重磁同源性岩石的位置及规模.另外,依据大面积分布的强负相关与强正相关特征推测了古火山喷发的主要岩石类型,并预测了一个碎斑熔岩火山口.同时,相山地区铀矿点主要分布在强正相关与强负相关的过渡带上,进一步揭示了铀矿分布与岩性接触面存在着密切关系.这些研究成果可为相山地区深部铀矿勘查提供数据基础.
水力压裂是改善非常规储层物性、实现非常规油气资源有效开发的核心技术.与微地震监测不同,电磁监测方法以压裂液与围岩的复电阻率差异为基础,具有更好的物性基础.对于致密页岩等非常规储层,水力压裂形成的裂隙储层结构复杂,多种激发极化机理复合作用,给实际地层的激电建模带来一定的困难,限制了压裂电磁监测的储层评估能力.为此,开展了裂隙介质双重孔隙介质理论建模,结合五峰组致密储层页岩的压裂前、后激发极化特征实验,分析了储层样品在不同压力、压裂前后不同饱和度、填充不同支撑剂等状态下的激电参数变化及各向异性特征.研究结果表明:压裂后裂隙储层的激电物理特征受裂隙面影响较大,储层孔隙结构连通性增强是引起电阻率下降的主要原因;支撑剂的定向分布有利于增强裂隙储层的各向异性.该研究成果为储层激电建模提供了理论与实验基础.
传统偏移成像方法是建立在波场一次反射的假设条件之上,事实上,完整的地震成像来自于地下全波场(一次波和多次波).为了能够利用全波场信息以提高成像质量,提出了基于单程波算子的全波场最小二乘偏移(FW-LSM)方法.首先,引入地层上、下界面反射系数和背景速度,推导了基于单程波算子闭循环延拓的全波场正演算子,模拟了平滑速度模型情况下的全波场信息;其次,在二范数意义下求解FW-LSM的误差泛函和梯度项表达式,构建基于反演框架的全波场最小二乘偏移方法;最后,针对透镜体与加入水层的Marmousi模型进行测试分析,验证了方法的有效性.研究表明:多轮次反演迭代压制了复杂波场产生的相干假象,多次反射波信息的利用显著改善了成像品质;该方法拓宽了地震成像的手段,尤其在多次波发育的海域地震资料处理中有着重要的作用和潜在的推广价值.
高斯束偏移方法具有较高的计算效率与成像精度,以其灵活性和高效性被广泛应用.但当地震数据的信噪比较低时,采用常规弹性波高斯束方法易受噪声干扰,导致最终成像结果不理想.针对此问题,文中提出一种基于阈值控制的弹性波高斯束偏移的优化方法.通过设置阈值函数,并将其应用于Tau-p变换,得到基于阈值控制的倾斜叠加公式;然后利用互相关成像条件,得到基于阈值控制的弹性波高斯束偏移成像公式.模型试算的结果表明,采用基于阈值控制的弹性波偏移高斯束成像方法,可有效压制成像噪声,提高成像信噪比,显著改善偏移成像质量.
