Interpolation and random noise removal is a pre-requisite for multichannel techniques because the irregularity in observed data can affect their performances. Projection Onto Convex Sets (POCS) method better handle seismic interpolation if data's signal-to-noise ratio (SNR) high, while it has difficulty noisy situations inserts each iteration. Weighted POCS weaken effects, performance affected by choice of weight factors still unsatisfactory. Thus, new weighted derived through Iterative Hard...

Facies identification is a powerful means to predict reservoirs. We achieve facies using relevance vector machine (RVM) and develop discriminant method based on multikernel RVM (MKRVM). An has the same functional form as support (SVM) that widely used in geophysics shows promising performance disposing of small-samples, nonlinear high-dimensional problems. The inherits these superiorities, its training implemented under Bayesian framework. Thus, it can provide probability information about...

SUMMARY Seismic inversion is one of the most commonly used methods in oil and gas industry for reservoir characterization from observed seismic data. Deep learning (DL) emerging as a data-driven approach that can effectively solve inverse problem. However, existing DL-based utilize only data input, which often leads to poor stability results. Besides, it has always been challenging train robust network since real survey limited labelled pairs. To partially overcome these issues, we develop...

Lithofacies identification is a crucial work in reservoir characterization and modeling. The vast inter-well area can be supplemented by facies of seismic data. However, the relationship between lithofacies information that affected many factors complicated. Machine learning has received extensive attention recent years, among which support vector machine (SVM) potential method for classification. classification involves identifying various types generally nonlinear problem, needs to solved...

Lithofacies classification is an indispensable procedure in well logging and seismic data interpretation. We propose a novel deep classified autoencoder learning approach to identify lithofacies for high-dimensional complex problems. Deep (DAE) unsupervised method via layerwise pretraining multiple autoencoders. It can learn features automatically reconstruct the original with small error. Introducing sparse constraint (i.e., autoencoder) potentiates ability of autoencoder. On this...

Due to the environment effects, economy restrictions, and acquisition equipment limitations, observed seismic data always have several traces missing contain some random noise, affecting performance of surface-related multiple elimination (SRME), wave-equation-based imaging, inversion. Projection onto convex sets (POCS) is an effective interpolation algorithm, while unsatisfactory in noisy situations. Weighted POCS (WPOCS) method can weaken noise effects extent, but still unsatisfactory....

Seismic data are usually contaminated by various noises. Noise suppression plays an important role in seismic processing. In this article, we propose a new denoising method based on the nonlocal weighted robust principal component analysis (RPCA). First, divided into many patches and grouped similarity. For each group, then, establish similar block matrix set up objective function of RPCA. Next, introduce iterative log-thresholding algorithm augmented Lagrangian to solve problem....

Unconventional reservoirs usually have strong anisotropy. Generally, they can be regarded as transversely isotropic media with a vertical symmetry axis [(VTI) media] in the absence of fractures. Therefore, it is great significance to develop high-accuracy inversion method for VTI media. The three elastic parameters and two Thomsen anisotropy are obtained by indirect calculation or methods based on approximate formulas. However, cumulative errors caused low accuracy formulas limit estimation...

Most existing amplitude variation with offset (AVO) inversion methods are based on the Zoeppritz’s equation or its approximations. These assume that of seismic data depends only reflection coefficients, which means wave-propagation effects, such as geometric spreading, attenuation, transmission loss, and multiples, have been fully corrected attenuated before inversion. However, these requirements very strict can hardly be satisfied. Under a 1D assumption, reflectivity-method-based inversions...

Abstract In VTI media, the conventional inversion methods based on existing approximation formulas are difficult to accurately estimate anisotropic parameters of reservoirs, even more so for unconventional reservoirs with strong seismic anisotropy. Theoretically, above problems can be solved by utilizing exact reflection coefficients equations. However, their complicated expression increases difficulty in calculating Jacobian matrix when applying them Bayesian deterministic inversion....

ABSTRACT Elastic parameters such as Young's modulus, Poisson's ratio, and density are very important characteristic that required to properly characterise shale gas reservoir rock brittleness, evaluate characteristics of reservoirs, directly interpret lithology oil‐bearing properties. Therefore, it is significant obtain accurate information these elastic parameters. Conventionally, they indirectly calculated by the physics method or estimated approximate formula inversion. The cumulative...

ABSTRACT Facies boundaries are critical for flow performance in a reservoir and significant lithofacies identification well interpretation prediction. based on supervised machine learning methods usually requires large amount of labelled data, which sometimes difficult to obtain. Here, we introduce the deep autoencoder learn hidden features conduct facies classification from elastic attributes. Both unlabelled data involved training process. Then, develop semi‐supervised by taking mean...

Seismic estimation of the fluid factor and shear modulus plays an important role in reservoir identification characterization. Various amplitude variation with offset inversion methods have been used to estimate these two parameters, which are generally based on approximate formulations Zoeppritz equations. However, accuracy is limited because forward modeling ability equations incorrect under conditions strong impedance contrast large incidence angles. Therefore, improve accuracy, we...

Anisotropy is widespread in the Earth’s crust, and VTI (vertical axis symmetry transverse isotropy) anisotropy common due to stratigraphic pressure. Disregarding leads inaccurate inversion results media. To estimate accurate elastic parameters, exact reflection coefficient equation of media should be used. This nonlinear more than commonly used linear equation. Although based on a complex problem, it still computable. Therefore, for media, we derive objective function combining Bayesian...

Prediction of lithology/fluid (LF) properties from seismic data can be very valuable in all phases oil and gas exploration production, but the resolution accuracy predicted results are reduced due to band-limited wavelet noise data. Deep learning review data, discover specific trends patterns that would not apparent humans, has been successfully used many applications, including geophysics. Also, time-frequency (T-F) analysis tools show how energy signal is distributed over 2-D T-F space,...

The accurate identification of lithofacies is indispensable for reservoir parameter prediction. In recent years, the application multivariate statistical methods has gained more and attention in petroleum geology. terms lithofacies, commonly used include discriminant analysis cluster analysis. Fisher Bayesian analyses are two different methods, which intrinsic advantages disadvantages. Given efficiency calculation cost, difficulty degree determining parameters, ability to analyze...

The numerical solution of the inverse problem is usually obtained by solving a set linear algebraic equations, while system equations may suffer from ill-posedness due to insufficient data. Regularisation technique for making estimation problems well posed adding indirect constraints on estimated model, but regularisation parameter selection difficult. In geophysics, without explicit calculation methods and quantitative evaluation criteria, it based experience inversion engineers try achieve...

Abstract Inverse Q filtering, which can compensate amplitude and correct phase, is an effective method to improve the resolution of seismic data. The conventional inverse usually based on wavefield continuation, unstable or under‐compensation. Based forward filtering equation in this work proposes a new attenuation compensation method, stable accurate, taking advantages theory regularization strategy obtain compensated data eventually. It also computationally efficient because only frequency...