A SIMULATED ANNEALING APPROACH TO SEISMIC MODEL OPTIMIZATION WITH SPARSE PRIOR INFORMATION1

It is well known that seismic inversion based on local model optimization methods, such as iterative use of linear optimization, may fail when prior information is sparse. Where the seismic events corresponding to reflectors of interest remain to be identified, a global optimization technique is required. We investigate the use of a global, stochastic optimization method, that of simulated annealing, to solve the seismic trace inversion problem, in which the two-way traveltimes and reflection coefficients are to be determined. The simulated annealing method is based on an analogy between the model-algorithm system and a statistical mechanical system. We exploit this analogy to produce improved annealing schedules. It is shown that even in cases of virtually no prior information about two-way traveltimes and reflection coefficients, the method is capable of producing reliable results.     

L1范数约束被动源数据稀疏反演一次波估计

对于被动源地震数据,运用常规的互相关算法得到的虚拟炮记录中,不仅含有一次波反射信息,还包括了表面相关多次波.然而,通过传统的被动源数据稀疏反演一次波估计(EPSI)方法,可以求得只含有一次波,不含表面相关多次波的虚拟炮记录.本文改进了传统的被动源数据稀疏反演一次波估计问题的求解方法,将被动源稀疏反演一次波估计求解问题转化为双凸L1范数约束的最优化求解问题,避免了在传统的稀疏反演一次波估计过程中用时窗防止反演陷入局部最优化的情况.在L1范数约束最优化的求解过程中,又结合了2DCurvelet变换和小波变换,在2DCurvelet-wavelet域中,数据变得更加稀疏,从而使求得的结果更加准确,成像质量得到了改善.通过简单模型和复杂模型,验证了本文提出方法的有效性.     

Assessing uncertainty in refraction seismic traveltime inversion using a global inversion strategy

To analyse and invert refraction seismic travel time data, different approaches and techniques have been proposed. One common approach is to invert first‐break travel times employing local optimization approaches. However, these approaches result in a single velocity model, and it is difficult to assess the quality and to quantify uncertainties and non‐uniqueness of the found solution. To address these problems, we propose an inversion strategy relying on a global optimization approach known as particle swarm optimization. With this approach we generate an ensemble of acceptable velocity models, i.e., models explaining our data equally well. We test and evaluate our approach using synthetic seismic travel times and field data collected across a creeping hillslope in the Austrian Alps. Our synthetic study mimics a layered near‐surface environment, including a sharp velocity increase with depth and complex refractor topography. Analysing the generated ensemble of acceptable solutions using different statistical measures demonstrates that our inversion strategy is able to reconstruct the input velocity model, including reasonable, quantitative estimates of uncertainty. Our field data set is inverted, employing the same strategy, and we further compare our results with the velocity model obtained by a standard local optimization approach and the information from a nearby borehole. This comparison shows that both inversion strategies result in geologically reasonable models (in agreement with the borehole information). However, analysing the model variability of the ensemble generated using our global approach indicates that the result of the local optimization approach is part of this model ensemble. Our results show the benefit of employing a global inversion strategy to generate near‐surface velocity models from refraction seismic data sets, especially in cases where no detailed a priori information regarding subsurface structures and velocity variations is available.     

Semiautomatic and Automatic Cooperative Inversion of Seismic and Magnetotelluric Data

Natural source electromagnetic methods have the potential to recover rock property distributions from the surface to great depths. Unfortunately, results in complex 3D geo-electrical settings can be disappointing, especially where significant near-surface conductivity variations exist. In such settings, unconstrained inversion of magnetotelluric data is inexorably non-unique. We believe that: (1) correctly introduced information from seismic reflection can substantially improve MT inversion, (2) a cooperative inversion approach can be automated, and (3) massively parallel computing can make such a process viable. Nine inversion strategies including baseline unconstrained inversion and new automated/semiautomated cooperative inversion approaches are applied to industry-scale co-located 3D seismic and magnetotelluric data sets. These data sets were acquired in one of the Carlin gold deposit districts in north-central Nevada, USA. In our approach, seismic information feeds directly into the creation of sets of prior conductivity model and covariance coefficient distributions. We demonstrate how statistical analysis of the distribution of selected seismic attributes can be used to automatically extract subvolumes that form the framework for prior model 3D conductivity distribution. Our cooperative inversion strategies result in detailed subsurface conductivity distributions that are consistent with seismic, electrical logs and geochemical analysis of cores. Such 3D conductivity distributions would be expected to provide clues to 3D velocity structures that could feed back into full seismic inversion for an iterative practical and truly cooperative inversion process. We anticipate that, with the aid of parallel computing, cooperative inversion of seismic and magnetotelluric data can be fully automated, and we hold confidence that significant and practical advances in this direction have been accomplished.     

Closed-loop SRME based on 3D L1-norm sparse inversion

In many situations, the quality of seismic imaging is largely determined by a proper multiple attenuation as preprocessing step. Despite the widespread application of surface-related multiple elimination (SRME) and estimation of primaries by sparse inversion (EPSI) for the removal of multiples, there still exist some limitations in the process of prediction and subtraction (SRME) or inversion (EPSI), which make the efficiency of multiple attenuation less satisfactory. To solve these problems, a new fully data-driven method called closed-loop SRME was proposed, which combines the robustness of SRME and the multi-dimensional inversion strategy of EPSI. Due to the selection of inversion approach and constraint, primary estimation by closed-loop SRME may fall into a local optimum during the solving process, which lowers the accuracy of deep information and weakens the continuity of seismic events. To avoid these shortcomings, we first modified the solving method for closed-loop SRME to an L1 norm-based bi-convex optimization method, which stabilizes the solution. Meanwhile, in the L1 norm constraint-based optimization process, the 3D sparsifying transform, being a 2D Curvelet-1D wavelet transform, is brought in as a 3D sparse constraint. In the 3D sparsifying domain, the data become sparser, thus making the result of optimization more accurate, the information of seismic events more continuous and the resolution higher. Examples on both synthetic and field data demonstrate that the method proposed in this paper, compared with the traditional SRME and closed-loop SRME, have an excellent effect on primary estimation and suppress multiples effectively.     

Seismic-impedance inversion with fuzzy clustering constraints: an example from the Carlin Gold District,Nevada, USA

The seismic reflection method provides high-resolution data that are especially useful for discovering mineral deposits under deep cover. A hindrance to the wider adoption of the seismic reflection method in mineral exploration is that the data are often interpreted differently and independently of other geophysical data unless common earth models are used to link the methods during geological interpretation. Model-based inversion of post-stack seismic data allows rock units with common petrophysical properties to be identified and permits increased bandwidth to enhance the spatial resolution of the acoustic-impedance model. However, as seismic reflection data are naturally bandlimited, any inversion scheme depends upon an initial model, and must deal with non-unique solutions for the inversion. Both issues can be largely overcome by using constraints and integrating prior information. We exploit the abilities of fuzzy c-means clustering to constrain and to include prior information in the inversion. The use of a clustering constraint for petrophysical values pushes the inversion process to select models that are primarily composed of several discrete rock units and the fuzzy c-means algorithm allows some properties to overlap by varying degrees. Imposing the fuzzy clustering techniques in the inversion process allows solutions that are similar to the natural geologic patterns that often have a few rock units represented by distinct combinations of petrophysical characteristics. Our tests on synthetic models, with clear and distinct boundaries, show that our methodology effectively recovers the true model. Accurate model recovery can be obtained even when the data are highly contaminated by random noise, where the initial model is homogeneous, or there is minimal prior petrophysical information available. We demonstrate the abilities of fuzzy c-means clustering to constrain and to include prior information in the acoustic-impedance inversion of a challenging magnetotelluric/seismic data set from the Carlin Gold District, USA. Using fuzzy c-means guided inversion of magnetotelluric data to create a starting model for acoustic-impedance proved important in obtaining the best result. Our inversion results correlate with borehole data and provided a better basis for geological interpretation than the seismic reflection images alone. Low values of the acoustic impedance in the basement rocks were shown to be prospective by geochemical analysis of rock cores, as would be predicted for later gold mineralization in weak, decalcified rocks.     

Model-based optimization of source locations for 3D acoustic seismic full-waveform inversion

Besides classical imaging techniques, full-waveform inversion is an increasingly popular method to derive elastic subsurface properties from seismic data. High-resolution velocity models can be obtained, and spatial sampling criteria are less strict than for imaging methods, because the entire information content of the seismic waveforms is used. As high operational costs arise from seismic surveys, the acquirable data volume is often limited by economic criteria. By selecting optimal locations for seismic sources, the information content of the data can be maximized, and the number of sources and thus the acquisition costs can be reduced compared with standard acquisition designs. The computation of such optimized designs for large-size 3D inverse problems at affordable computational cost is challenging. By using a sequential receiver-wise optimization strategy, we substantially reduce the computational requirements of the optimization process. We prove the applicability of this method by means of numerical 3D acoustic examples. Optimized source designs for different receiver patterns are computed for a realistic subsurface model, and the value of the designs is evaluated by comparing checkerboard inversion tests with different acquisition designs. Our examples show that inversion results with higher accuracy can be obtained with the optimized designs, regardless of the number of sources, the number of receivers, or the receiver distribution. Larger benefits of the optimized designs are visible when a sparse receiver geometry is used.     

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.     

数据域特征波反射走时反演

地震波的运动学信息(走时、斜率等)通常用于宏观速度建模.针对走时反演方法,一个基本问题是走时拾取或反射时差的估计.对于成像域反演方法,可以通过成像道集的剩余深度差近似计算反射波时差.在数据域中,反射地震观测数据是有限频带信号,如果不能准确地确定子波的起跳时间,难以精确地确定反射波的到达时间.另一方面,如果缺乏关于模型的先验信息,则很难精确测量自地下同一个反射界面的观测数据同相轴和模拟数据同相轴之间的时差.针对走时定义及时差测量问题,首先从叠前地震数据的稀疏表达出发,利用特征波场分解方法,提取反射子波并估计局部平面波的入射和出射射线参数.进一步,为了实现自动和稳定的走时拾取,用震相的包络极值对应的时间定义反射波的到达时,实现了立体数据中间的自动生成.理论上讲,利用包络极值定义的走时大于真实的反射波走时,除非观测信号具有无限带宽(即delta脉冲).然而,走时反演的目的是估计中-大尺度的背景速度结构,因此走时误差导致的速度误差仍然在可以接受的误差范围内.利用局部化传播算子及特征波聚焦成像条件将特征波数据直接投影到地下虚拟反射点,提出了一种新的反射时差估计方法.既避免了周期跳跃现象以及串层等可能性,又消除了振幅因素对时差测量的影响.最后,在上述工作基础之上,提出了一种基于特征波场分解的新型全自动反射走时反演方法(CWRTI).通过对泛函梯度的线性化近似,并用全变差正则化方法提取梯度的低波数部分,实现了背景速度迭代反演.在理论上,无需长偏移距观测数据或低频信息、对初始模型依赖性低且计算效率高,可以为后续的全波形反演提供可靠的初始速度模型.理论和实际资料的测试结果证明了本文方法的有效性.     

模拟退火叠前AVA同步反演方法

叠前反演技术已经成为岩性预测和流体识别的关键技术,并将在油气勘探中发挥重要作用,本文提出了模拟退火AVA同步反演方法,该方法以多个角道集地震数据体和先验纵波速度、横波速度、密度模型等为输入,综合利用地震、地质、测井等多种信息为约束,采用模拟退火全局优化方法,直接采用Knott-Zoeppritz方程求解反射系数,同步反演获得纵波阻抗、横波阻抗、纵横波速度、密度、Mu、Lambda等23种岩石弹性参数模型.其主要特点如下：１）采用Zoeppritz方程直接求解,精度高;2) 目标函数选择灵活;３）多角度提取子波;４）可获得深度域的反演结果.实际资料应用证明,该反演方法保留了地震反射振幅随偏移距不同或入射角不同而变化的特征,可有效提高储层预测和流体识别精度.实际应用表明,本方法预测结果可靠,具有较强的实用性.     

Depth and morphology of reflectors from the non-linear inversion of arrival-time and waveform semblance data. Part I: method and applications to synthetic data

We propose a two-dimensional, non-linear method for the inversion of reflected/converted traveltimes and waveform semblance designed to obtain the location and morphology of seismic reflectors in a lateral heterogeneous medium and in any source-to-receiver acquisition lay-out. This method uses a scheme of non-linear optimization for the determination of the interface parameters where the calculation of the traveltimes is carried out using a finite-difference solver of the Eikonal equation, assuming an a priori known background velocity model. For the search for the optimal interface model, we used a multiscale approach and the genetic algorithm global optimization technique. During the initial stages of inversion, we used the arrival times of the reflection phase to retrieve the interface model that is defined by a small number of parameters. In the successive steps, the inversion is based on the optimization of the semblance value determined along the calculated traveltime curves. Errors in the final model parameters and the criteria for the choice of the best-fit model are also estimated from the shape of the semblance function in the model parameter space. The method is tested and validated on a synthetic dataset that simulates the acquisition of reflection data in a complex volcanic structure. This study shows that the proposed inversion approach is a valid tool for geophysical investigations in complex geological environments, in order to obtain the morphology and positions of embedded discontinuities.     

基于稀疏反演三维表面多次波压制方法

三维表面多次波压制是海洋地震资料预处理中的重要研究课题,基于波动理论的三维表面多次波压制方法(3DSRME)是数据驱动的方法,理论上来说,可有效压制复杂构造地震数据表面多次波.但该方法因对原始地震数据采集要求高而很难在实际资料处理中广泛应用.本文基于贡献道集的概念,将稀疏反演方法引入到表面多次波压制中,应用稀疏反演代替横测线积分求和,无需对横测线进行大规模重建,进而完成三维表面多次波预测,这样可有效解决实际三维地震数据横测线方向稀疏的问题.基于纵测线多次波积分道集为抛物线的假设,为保证预测后三维表面多次波和全三维数据预测的多次波在运动学和动力学特征上基本一致,文中对预测数据实施基于稳相原理的相位校正.理论模型和实际数据的测试结果表明,本文基于稀疏反演三维表面多次波压制方法可在横测线稀疏的情况下,有效压制三维复杂介质地震资料中的表面多次波,从而更好地提高海洋地震资料的信噪比,为高分辨率地震成像提供可靠的预处理数据保障.     

Enforcing smoothness and assessing uncertainty in non-linear one-dimensional prestack seismic inversion

Estimation of elastic properties of rock formations from surface seismic amplitude measurements remains a subject of interest for the exploration and development of hydrocarbon reservoirs. This paper develops a global inversion technique to estimate and appraise 1D distributions of compressional‐wave velocity, shear‐wave velocity and bulk density, from normal‐moveout‐corrected PP prestack surface seismic amplitude measurements. Specific objectives are: (a) to evaluate the efficiency of the minimization algorithm (b) to appraise the impact of various data misfit functions, and (c) to assess the effect of the degree and type of smoothness criterion enforced by the inversion. Numerical experiments show that very fast simulated annealing is the most efficient minimization technique among alternative approaches considered for global inversion. It is also found that an adequate choice of data misfit function is necessary for a reliable and efficient match of noisy and sparse seismic amplitude measurements. Several procedures are considered to enforce smoothness of the estimated 1D distributions of elastic parameters, including predefined quadratic measures of length, flatness and roughness. Based on the general analysis of global inversion techniques, we introduce a new stochastic inversion algorithm that initializes the search for the minimum with constrained random distributions of elastic parameters and enforces predefined autocorrelation functions (semivariograms). This strategy readily lends itself to the assessment of model uncertainty. The new global inversion algorithm is successfully tested on noisy synthetic amplitude data. Moreover, we present a feasibility analysis of the resolution and uncertainty of prestack seismic amplitude data to infer 1D distributions of elastic parameters measured with wireline logs in the deepwater Gulf of Mexico. The new global inversion algorithm is computationally more efficient than the alternative global inversion procedures considered here.     

基于改进差分进化算法的大地电磁测深和重力宽范围物性约束联合反演

宽范围物性约束技术容易实现、具有一定容错性,目前已在大地电磁测深（MT）和地震、MT和重力联合反演中实现,但该技术是结合模拟退火算法实现的.差分进化算法（DE）是一种全局优化算法,但该算法在地球物理联合反演领域应用较少.基于此,本文以双种群设置方案为框架改进了DE算法,并提出了基于改进DE算法的宽范围物性约束技术.MT和重力联合反演的模型试验表明：与传统的DE算法相比,改进的DE算法收敛速度更快,寻优能力更强;基于改进DE算法的宽范围物性约束技术可以促进不同岩石物性参数在一定"范围"内实现耦合,既可以利用岩石物性关联的导向作用,又可以发挥优化算法的寻优能力,进而降低地球物理联合反演对先验信息的要求;此外,该技术的实现也验证了宽范围物性约束思想在联合反演领域中的适用性,具有进一步推广至其他优化算法中的潜质.     

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.     

地震数据互相关驱动的多道反演方法

地震反演是储层定量描述和地震油气识别的关键技术,反演结果在复杂构造区域的横向连续性和保真性是影响地震资料定量解释精度的重要因素.基于此,本文发展了地震数据互相关驱动的多道反演方法.考虑地层反射系数与地震数据在结构上具有相似性的特点,基于地震数据互相关描述地层反射系数的结构特征,并将其作为多道地震反演的横向约束条件;此外,为改善地震数据本身横向连续性差对反演结果的影响,在目标泛函的惩罚项中引入局部优化算子,构建了一个易于求解的多道地震反演目标泛函.与常规多道地震反演方法相比,本文方法能够设计更合理、更符合实际情况的横向约束算子,提高反演结果的横向连续性,并且能有效降低地震资料质量对反演结果的影响.模型测试和实际应用验证了本方法的可靠性和稳定性.     

2D多尺度混合优化地球物理反演方法及其应用(英文)

局部优化和全局优化方法广泛应用到地球物理反演,但是两者各有其优缺点。将两类方法结合起来可以取长补短。将退火遗传算法（SAGA）和单纯形算法相结合,得到了一种高效、健全的2D非线性混合地震走时反演方法。首先,利用SAGA进行大范围的全局搜索,然后由单纯形方法进行快速局部搜索。为了降低层析成像的多解性,我们采用了多尺度逐次逼近的技巧。把速度场划分为不同的空间尺度,定义网格节点上的速度作为待反演参数,采用双三次样条函数模型参数化,正问题采用有限差分走时计算方法,反问题采用多尺度混合反演方法。一个低速度异常体的数值模拟试验和抗走时扰动试验表明该方法是有效和健全的。我们将该方法应用到青藏高原东北缘阿尼玛卿rlet,Meyer,Marr,缝合带东段上部地壳速度结构研究中。数字模型试验和实际资料的应用表明了方法的有效性和健全性。     

Look-ahead vertical seismic profiling inversion approach for density and compressional wave velocity in Bayesian framework

To reduce drilling uncertainties, zero-offset vertical seismic profiles can be inverted to quantify acoustic properties ahead of the bit. In this work, we propose an approach to invert vertical seismic profile corridor stacks in Bayesian framework for look-ahead prediction. The implemented approach helps to successfully predict density and compressional wave velocity using prior knowledge from drilled interval. Hence, this information can be used to monitor reservoir depth as well as quantifying high-pressure zones, which enables taking the correct decision during drilling. The inversion algorithm uses Gauss–Newton as an optimization tool, which requires the calculation of the sensitivity matrix of trace samples with respect to model parameters. Gauss–Newton has quadratic rate of convergence, which can speed up the inversion process. Moreover, geo-statistical analysis has been used to efficiently utilize prior information supplied to the inversion process. The algorithm has been tested on synthetic and field cases. For the field case, a zero-offset vertical seismic profile data taken from an offshore well were used as input to the inversion algorithm. Well logs acquired after drilling the prediction section was used to validate the inversion results. The results from the synthetic case applications were encouraging to accurately predict compressional wave velocity and density from just a constant prior model. The field case application shows the strength of our proposed approach in inverting vertical seismic profile data to obtain density and compressional wave velocity ahead of a bit with reasonable accuracy. Unlike the commonly used vertical seismic profile inversion approach for acoustic impedance using simple error to represent the prior covariance matrix, this work shows the importance of inverting for both density and compressional wave velocity using geo-statistical knowledge of density and compressional wave velocity from the drilled section to quantify the prior covariance matrix required during Bayesian inversion.     

带粒子滤波约束的PP-PS联合反演的稀疏解算法

随着地震勘探目标从构造型油气藏向岩性油气藏的转变,地震勘探难度日益增大,这就要求从地震数据中获得更多可靠且具有明确地质含义的属性信息,并充分利用这些属性信息来对储层的岩性、岩相进行分析.AVO三参数反演能够从振幅随炮检距的变化信息中直接提取纵波速度、横波速度以及密度来估计岩石和流体的性质,进而对储层进行预测.然而,AVO反演本身是一个不适定的问题,加上地震纵波反射系数对横波速度和密度的不敏感,会造成单纯利用纵波地震数据进行反演的结果误差大.随着地震接收和数据处理技术的发展,越来越多的学者对PP-PS联合反演方法进行了研究并在实际资料中得以运用.融合转换横波地震数据的联合反演在一定程度上提高了反演的精度,降低了解的不稳定性.但是在信噪比较低的情况下,联合反演的效果受到了限制.本文从优化理论出发,提出了基于粒子滤波提供先验知识的l₁范数约束极小化问题的稀疏解算法.并将上述方法运用到了不同的模型中,通过比较分析,证实了该方法在不同信噪比资料中的有效性和在信噪比较低情况下的优势.     

Adaptive reconstruction method of impedance model with absolute and relative constraints

Inversion for seismic impedance is an ill-posed and band-limited problem so that inversion results are non-unique and unstable and low and high frequency components of inversion results are missed. Combining regularization with constraints of sonic log data and geological structure information can help to alleviate these problems. To achieve this, we developed an inversion method by constructing a new objective function which includes edge-preserving regularization and soft constraint based on Markov random field (MRF). The method directly introduces absolute constraints with prior impedance and sonic log data in the objective function and indirectly achieves relative constraints with geologic structures of layer interfaces and faults by adjusting the regularization parameter which is the scaling parameter δ. Moreover, we improved the inversion result using anisotropic diffusion smoothing method. Optimization approach utilized in inversion is a fast simulated annealing (FSA). We test the method on both synthetic and field data examples. Tests on 2-D synthetic data indicate that aspects of the discontinuity in the inversion results are significantly improved by adding δ values in faults and layer interfaces. We obtained better results by combining the first-order neighborhood and the third-order neighborhood of MRF. The inversion results of the field data provide more detailed information of the layers. The results of nearby faults were improved by introducing the geological structure constraints.