Physical and/or economic constraints cause acquired seismic data to be incomplete; however, complete are required for many subsequent processing procedures. Data reconstruction is a crucial and long-standing topic in the exploration seismology field. We extended our previous works on deep learning (DL)-based irregularly regularly missing 2-D 3-D data. A key motivation that convolutional neural network (CNN) can take full advantage of nature data, additional dimension allows more information...

Inspired by image-to-image translation, we applied deep learning (DL) to regularly missing data reconstruction, aimed at translating incomplete into their corresponding complete data. With this purpose in mind, first construct a network architecture based on an end-to-end U-Net convolutional network, which is generic DL solution for various tasks. We then meticulously prepare the training with both synthetic and field seismic This article implemented Python Keras (a high-level library)....

Abstract Carbonates are characterized by a complex system of pores, caves, vugs and fractures that significantly influence fluid flow the physical behaviors rocks. Six rock samples taken from carbonate reservoir in China's Sichuan Basin to perform computed tomography (CT), X‐ray diffraction thin section analyses. The classified into fractured, fractured‐vuggy pore‐cavity types based on their microstructural properties. Ultrasonic low frequency tests performed with different pressures fluids...

Shear wave velocity plays an important role in both reservoir prediction and pre-stack inversion. However, the current deep learning-based shear methods have certain limitations, including lack of training dataset, poor model generalization, physical interpretability. In this study, theoretical rock physics models are introduced into construction labeled dataset for learning algorithms, a forward simulation is utilized to supplement that incorporates geological geophysical knowledge. This...

Recently, intelligent data processing and interpretation based on deep learning (DL) have received considerable attention. Training are vital for DL-based approaches. In geosciences, researchers been facing a significant obstacle, i.e., the absence of authoritative representative open training testing artificial neural networks (ANNs). Although open-source works in geosciences increasing, quantity is currently limited. With aid Society Exploration Geophysicists (SEG) 2020 Machine Learning...

Sparse-spike deconvolution (SSD) is an important method for seismic resolution enhancement. With the wavelet given, many trace-by-trace SSD methods have been proposed extracting estimate of reflection-coefficient series from stacked traces. The main drawbacks are that they neither use information adjacent seismograms nor do take full advantage inherent spatial continuity data. Although several multitrace consequently proposed, these generally rely on different assumptions and theories...

SUMMARY Deep learning (DL) has achieved remarkable progress in geophysics. The most commonly used supervised (SL) framework requires massive labelled representative data to train artificial neural networks (ANNs) for good generalization. However, the labels are limited or unavailable field seismic applications. In addition, SL generally cannot take advantage of well-known physical laws and thus fails generate physically consistent results. weaknesses standard non-negligible. Therefore, we...

Abstract We present a detailed 3‐D P ‐wave velocity model obtained by first‐arrival travel‐time tomography with seismic refraction data in the segment boundary of Sumatra subduction zone across Simeulue Island, and an image top subducted oceanic crust extracted from depth‐migrated multi‐channel reflection profiles. have picked first arrivals air‐gun source recorded local networks ocean‐bottom seismometers, inverted travel‐times for zone. This shows anomalous intermediate velocities between...

Seismic waves propagating in the subsurface suffer from attenuation, which can be represented by quality factor Q. Knowledge of Q plays a vital role hydrocarbon exploration. Many methods to measure have been proposed, among central frequency shift (CFS) and peak (PFS) are commonly used. However, both under assumption particular shape for amplitude spectra, will cause systematic error estimation. Recently new method estimate has proposed overcome this disadvantage using weighted exponential...

SUMMARY In fully fluid-saturated rocks, two common phenomena are documented both experimentally and theoretically for frequency-dependent elastic moduli attenuation, that is, the drained/undrained transition relaxed/unrelaxed transition. When investigating these transitions with forced oscillation method in laboratory, it is crucial to consider boundary differences between laboratory underground. A 1-D poroelastic numerical model was previously established describe their effects; however,...

Poroelastic moduli of rocks are difficult to measure in the laboratory setting. Digital rock technology has potential support interpretation results and provide insights into pore-scale deformation patterns. We apply a digital workflow images obtained for sand pack estimate Biot coefficient drained pore modulus by exploiting its relation bulk solid phase modulus. Comparison with measured values sandstone similar porosities indicate that simulated value is plausible but appears too high. The...

Acoustic impedance (AI) inversion is of great interest because it extracts information regarding rock properties from seismic data and has successful applications in reservoir characterization. During wave propagation, anelastic attenuation dispersion always occur the subsurface not perfectly elastic, thereby diminishing resolution. AI based on convolutional model requires that input be free effects; otherwise, low-resolution results are inevitable. The intrinsic instability occurs while...

Forward seismic modeling is a useful tool for simulating the response of low velocity zones near ground surface and delicate structures at deeper depth in northwestern China. The main objective forward to solve wave equation accurately. finite difference method (FDM) commonly used numerical solving equation. It can obtain very high accuracy with high‐order scheme if grid configuration appropriate. However it shows dispersion its stability criteria also strict. In order deal these problems,...

Seismic data are nonstationary due to subsurface anelastic attenuation and dispersion effects. These effects, also referred as the earth’s [Formula: see text]-filtering can diminish seismic resolution. We previously developed a method of sparse reflectivity inversion (NSRI) for resolution enhancement, which avoids intrinsic instability associated with inverse text] filtering generates superior compensation results. Applying NSRI sets that contain multiples (addressing surface-related only)...

Coda waves are usually regarded as noise in the conventional seismic exploration fields. Our work is to use energy of coda estimate stochastic parameters random media, which necessary characterize subsurface reservoir and assess oil or gas total volume heterogeneous reservoir. In this paper, we briefly present Monte Carlo radiative transfer (MCRT) theory acoustic often used model envelopes approximated media seismology. Then, fluctuation strength correlation length 2D based on MCRT...

In recent years, several experimental methods have been introduced to measure the elastic parameters of rocks in relatively low-frequency range, such as differential acoustic resonance spectroscopy (DARS) and stress–strain measurement. It is necessary verify validity feasibility applied measurement method quantify sources levels error. Relying solely on laboratory measurements, however, we cannot evaluate complete wavefield variation apparatus. Numerical simulations wave propagation, other...

ABSTRACT The subsurface media are not perfectly elastic, thus anelastic absorption, attenuation and dispersion (aka Q filtering) effects occur during wave propagation, diminishing seismic resolution. Compensating for is imperative resolution enhancement. values required most of conventional ‐compensation methods, the source wavelet additionally some them. Based on previous work non‐stationary sparse reflectivity inversion, we evaluate a series methods with/without knowing wavelet. We...

Squirt flow is a wave-induced fluid mechanism to account for the velocity dispersion and attenuation of fluid-saturated porous media. Theoretical models squirt cannot deal with complex microcrack-pore networks thus give accurate curves respect frequency. We adopt numerical oscillatory compressibility test method, based on quasi-static poroelastic equation Biot, model calculate corresponding P-wave modulus attenuation. also designed nine medium containing different crack-pore configurations...

Summary S-wave velocity plays an important role in both reservoir prediction and pre-stack inversion. However, the current deep learning-based methods have certain limitations. In order to solve problems of insufficient training samples real field areas poor generalization learning model, we combine DNN theoretical rock physics model predict limestone reservoir. Firstly, four parameters (P-wave velocity, crack-density, density, porosity) affecting are selected according synthetic data...

Summary Lithology identification is one of the important tasks in reservoir evaluation and basis for solving parameters. However, traditional lithology methods have problems such as large workload, poor generalization difficulty obtaining labels, resulting low accuracy identification. In this abstract, multivariate properties including elastic parameters, physical parameters fluid are incorporated into by theoretical rock physics models, we use label data generated models to train a deep...