Geophysical Prospecting最新文献

Digital rock techniques are increasingly important in petroleum exploration and petrophysics. Digital rocks are typically acquired via scanning or imaging techniques, but the resulting images may lack clear, detailed information due to resolution limitations. Super-resolution reconstruction using deep learning offers new possibilities for digital rock technology development. In current research on super-resolution reconstruction of digital rock images, most networks employ attentional mechanisms in a single dimension, ignoring more comprehensive interactions from both spatial and channel dimensions.

To address the above problems, we propose a bi-dimensional large kernel attention network for super-resolution reconstruction of digital rock images. The network consists of three components: a bi-dimensional large kernel building block, a contrast channel attention block and an enhanced spatial attention block. In addition, the traditional method of stacking network modules to build the network leads to an increase in computation and network size, so we adopt Transformer's MetaFormer architecture, which integrates multivariate feature extraction to improve the efficiency of the network. In the process of feature information circulation, we effectively prevent shallow feature loss by two efficient attention modules working at different network depth positions. Extensive experiments on Sandstone2D and Carbonate2D rock datasets show that our proposed model significantly outperforms existing image super-resolution networks.

Fault interpretation is crucial for subsurface resource extraction. Recent research has demonstrated that deep learning techniques can successfully detect faults. However, the network's prediction results still suffer from discontinuity and low accuracy problems due to insufficient exploitation of the spatial and global distribution characteristics of faults. This paper presents a novel approach for seismic fault detection using a dual-attention mechanism and multi-scale feature fusion. The proposed network uses ResNeSt residual blocks as encoders to extract multi-scale features of faults. During multi-scale feature fusion, a global context and a spatial dual-attention module are introduced to suppress interference from non-fault features. This improves the ability to detect faults. Five adjacent seismic slices were used as inputs to obtain the spatial distribution characteristics of faults. Data augmentation methods were used to enrich the fault morphology of synthetic seismic data. The Tversky loss function was used in the proposed model to alleviate the effect of data imbalance on fault identification tasks. Transfer learning methods were also used to evaluate the model's performance on field data from the F3 block in the Dutch North Sea and field data from the New Zealand Great South Basin. The model's performance was compared with some state-of-the-art methods, including DeepLabV3+, Pyramid Scene Parsing Network, Feature Pyramid Network and U-Net. The results show that the proposed fault detection method has excellent accuracy and fault continuity.

Fractures are omnipresent features in the shallower regions of the Earth's crust. In the context of rock physics, fracture characterization techniques rely largely on the determination of normal fracture compliances. Despite being thoroughly investigated through wave propagation experiments, this parameter is seldom estimated locally. In this work, we measure and compare local displacement-jump- and transmission-related fracture compliances using forced oscillations and ultrasonic propagation techniques, respectively. The experiments are carried out on an aluminium standard and on four different sandstone samples that contain a single planar fracture, considering a range of axial stresses. The results show that, for most rocks, both transmission-related and locally measured dry normal compliances are of the same order and also present similar tendencies with axial loads. However, transmission methods predict larger dry normal fracture compliances than those retrieved from local strain estimations. The results of this study may help to assess the validity of linear slip theory, which is widely used in fracture characterization efforts in the specific literature.

Fractures represent a critical structural feature in unconventional reservoirs, as they create essential pathways for the migration and accumulation of oil and gas. Therefore, fracture characterization is a fundamental task in the exploration of unconventional hydrocarbon resources. Conventional fracture characterization methods typically do not account for the inherent anisotropy of the formation, which arises from the sedimentary environment and fluid distribution, often leading to inaccurate fracture predictions. To address this challenge, we propose a petrophysical model that incorporates inherent anisotropy, employing rock physics modelling to accurately characterize fracture distribution. Furthermore, to reduce the substantial workload involved in manually calibrating the petrophysical model, we introduce a one-dimensional convolutional neural network combined with an attention mechanism. By leveraging the advanced nonlinear learning capabilities of the convolutional neural network, we aim to fit the petrophysical model and extend its application across all exploration wells and the entire field. The effectiveness and feasibility of the proposed method are demonstrated through experiments using actual borehole data from a fracture-dominated reservoir.

Deep-strata high-pressure reservoirs are a key research area in subsurface resource exploration. The complex mix of in situ pressure, anisotropy and fluid saturation in rocks leads to unclear seismic responses and uncertainties in wave propagation. Using acoustoelasticity theory and assuming weak anisotropy, we derived equations for the elastic parameters of stressed orthotropic media. These equations use anisotropic parameters to describe the unstressed elastic properties of orthotropic media. Then, using the Gassmann equation and low-frequency poro-elasticity, we found elastic parameters for single fluid-saturated orthotropic media. Non-welded interfaces serve as a reasonable approximation for tiny fractures and are ubiquitous in subsurface formations, and the viscous fluid present within these interfaces contributes to the observable attenuation of seismic waves. Using elastic parameters of stressed, fluid-saturated orthotropic media, we formulated reflection and transmission coefficient equations for these interfaces based on linear-slip theory. Using these equations, we analysed how stress, fluid saturation and interface changes affect seismic response and wave propagation. We then analysed how frequency, porosity, viscosity, fracture weakness and other physical properties affect seismic behaviour within and at the medium's interface. By constructing exact equations, we have achieved a more realistic simulation of subsurface seismic response. This enhancement in simulation accuracy facilitates a deeper understanding of the seismic response patterns observed in deep and complex subsurface reservoirs. Furthermore, it provides a solid theoretical foundation for fluid identification and reservoir prediction in actual subsurface reservoir scenarios.

The transformation of elastic impedance (EI) from partial-angle-stacked seismic data is a crucial technique in the domains of reservoir modelling. Conventionally, EI inversion is performed on a per-angle basis, leading to significant discrepancies in EI values across different angles, which may not accurately represent actual conditions. When the signal-to-noise ratio (SNR) of seismic data is low, the inverted EI tends to be unstable, resulting in poor-quality inversion outcomes. This research proposes a novel method that allows for enabling the derivation of EI for various angles simultaneously inverted from multiple partial angle-stack seismic datasets in one process. The aim of simultaneous inversion is to potentially ensure consistent EI results. To obtain this aim, we utilize an advanced regularization method called the Gramian constraint. Consequently, the objective function for the simultaneous inversion of multiple EIs is developed. Results from both synthetic and field data demonstrate improved stability in EI inversion, especially for the case of low SNR.

The transient electromagnetic method (TEM) is a prominent geophysical technique, and the TEM inversion for resistivity models is a crucial aspect of physical exploration. However, TEM inversion faces challenges such as nonlinearity, multiple solutions and ill-conditioning, which can lead to inaccurate results. In response to these challenges, metaheuristic algorithms have been extensively studied for their innovative approaches to solving inverse problems. Despite this, many existing metaheuristic inversion algorithms exhibit limitations, including premature convergence, slow convergence speed and inadequate computational accuracy. To address these issues, an improved bat algorithm (IBA) that incorporates logistic chaotic mapping and a spiral flight strategy (Logistic Chaotic Mapping and Spiral Flight Strategy-Based Bat Algorithm, LSBA) has been proposed. The logistic chaotic mapping strategy is utilized to initialize the population of the bat algorithm to enhance the initial convergence rate. Moreover, the spiral flight strategy facilitates the bats’ escape from local optima, thereby improving the algorithm's local exploration capabilities and solution accuracy. Numerical simulations, synthetic models and field experiments have demonstrated that the LSBA significantly enhances solution precision (the degree of closeness between the algorithm's inverted parameters and the true values), convergence speed and anti-noise performance. The LSBA effectively retrieves the stratigraphic parameters of the true model and accurately represents the geological information of actual mining areas, thereby validating the efficacy and feasibility of the proposed approach in TEM inversion.