Applied Geophysics最新文献

Well logging curves serve as indicators of strata attribute changes and are frequently utilized for stratigraphic analysis and comparison. Deep learning, known for its robust feature extraction capabilities, has seen continuous adoption by scholars in the realm of well logging stratigraphic correlation tasks. Nonetheless, current deep learning algorithms often struggle to accurately capture feature changes occurring at layer boundaries within the curves. Moreover, when faced with data imbalance issues, neural networks encounter challenges in accurately modeling the one-hot encoded curve stratification positions, resulting in significant deviations between predicted and actual stratification positions. Addressing these challenges, this study proposes a novel well logging curve stratigraphic comparison algorithm based on uniformly distributed soft labels. In the training phase, a label smoothing loss function is introduced to comprehensively account for the substantial loss stemming from data imbalance and to consider the similarity between different layer data. Concurrently, spatial attention and channel attention mechanisms are incorporated into the shallow and deep encoder stages of U²-Net, respectively, to better focus on changes in stratification positions. During the prediction phase, an optimized confidence threshold algorithm is proposed to constrain stratification results and solve the problem of reduced prediction accuracy because of occasional layer repetition. The proposed method is applied to real-world well logging data in oil fields. Quantitative evaluation results demonstrate that within error ranges of 1, 2, and 3 m, the accuracy of well logging curve stratigraphic division reaches 87.27%, 92.68%, and 95.08%, respectively, thus validating the effectiveness of the algorithm presented in this paper.

In sharp contrast to prefabricated cracks, the damage to rock masses resulting from external disturbances such as excavation disturbances and tectonic movement varies substantially as to the incidence, density, and magnitude of defects. The growth ratio of the energy dissipation density proportion D(R_ω (α)) of the damaged rock under impact loading is closely related to the static damage factor D(α) and is theoretically explored based on the Weibull distribution in this paper. Sandstones with varied damage levels after distinct static precompression, as described by CT imaging, are used to evaluate the impact load of different driving pressures. In addition, a high-speed camera and geometric fractal are used to exhibit the ejection and fragmentation characteristics of the pulverized sandstones after impact loading. The experimental outcomes confirm the theoretical study where the function of D(R_ω (α)) involving D(α) obeys the Weibull distribution, and the D(R_ω (α)) slowly rises with the expansion of the damage factor. With the increase of either the damage level or driving pressure of the sandstone, the number of pulverized rocks, the fragmentation degree, and the D(R_ω (α)) all increase. These results further advance rock dynamic theory and corroborate the energy evolution, ejection, and fragmentation characteristics of damaged sandstone under impact loading. These results can also serve as references for rock dynamic risk mitigation under dynamic catastrophes..

Seismic imaging of complicated underground structures with severe surface undulation (i.e., double complex areas) is challenging owing to the difficulty of collecting the very weak reflected signal. Enhancing the weak signal is difficult even with state-of-the-art multi-domain and multidimensional prestack denoising techniques. This paper presents a time–space dip analysis of offset vector tile (OVT) domain data based on the τ-p transform. The proposed N-th root slant stack method enhances the signal in a three-dimensional τ-p domain by establishing a zero-offset time-dip seismic attribute trace and calculating the coherence values of a given data sub-volume (i.e., inline, crossline, time), which are then used to recalculate the data. After sorting, the new data provide a solid foundation for obtaining the optimal N value of the N-th root slant stack, which is used to enhance a weak signal. The proposed method was applied to denoising low signal-to-noise ratio (SNR) data from Western China. The optimal N value was determined for improving the SNR in deep strata, and the weak seismic signal was enhanced. The results showed that the proposed method effectively suppressed noise in low-SNR data.

Rapid acquisition of rock structure information during tunnel construction is crucial for optimizing subsequent construction strategies and avoiding engineering rock disasters. In this regard, this study proposes the best time to obtain rock structure information on the tunnel face during the construction period. By summarizing relevant studies on rock information acquisition locally and abroad and combining them with the actual situation during the construction of the Lushan Tunnel, this study analyzed the factors affecting the quality of rock information acquisition during the construction period and the approximate range of the optimal timing of acquisition. We also conducted experiments on the concentration of respiratory dust and the concentration of total dust on each section of the Lushan Tunnel construction site and explored the optimal timing of acquiring rock information on the tunnel face by conducting several dust experiments at the construction site. The results showed that the best time to obtain information was before the erection of the steel arch, which was also the best time for the engineers to conduct mechanical characterization of the tunnel face and the lining inspection of the tunnel. The optimal acquisition timing identified in this study is crucial for improving the accuracy of rock information acquisition and guiding subsequent construction programs.

The Chepaizi Exploration Area, Junggar Basin (NW China) holds substantial importance for seismic exploration endeavors, yet it poses notable challenges due to the intricate nature of its subsurface and near-surface conditions. To address these challenges, we introduce a novel and comprehensive workflow tailored to evaluate and optimize seismic acquisition geometries while considering the impacts of near-surface viscosity. By integrating geological knowledge, historical seismic data, and subsurface modeling, we conduct simulations employing the visco-acoustic wave equation and reverse-time migration to produce detailed subsurface images. The quality of these images is quantitatively evaluated using a local similarity metric, a pivotal tool for evaluating the accuracy of seismic imaging. The culmination of this workflow results in an automated optimization strategy for acquisition geometries that enhances subsurface exploration. Our proposed methodology underscores the importance of incorporating near-surface viscosity effects in seismic imaging, offering a robust framework for improving the accuracy of subsurface imaging. Herein, we aim to contribute to the advancement of seismic imaging methodologies by providing valuable insights for achieving high-quality seismic exploration outcomes in regions characterized by complex subsurface and near-surface conditions.

Effective attenuation of seismic multiples is a crucial step in the seismic data processing workflow. Despite the existence of various methods for multiple attenuation, challenges persist, such as incomplete attenuation and high computational requirements, particularly in complex geological conditions. Conventional multiple attenuation methods rely on prior geological information and involve extensive computations. Using deep neural networks for multiple attenuation can effectively reduce manual labor costs while improving the efficiency of multiple suppression. This study proposes an improved U-net-based method for multiple attenuation. The conventional U-net serves as the primary network, incorporating an attentional local contrast module to effectively process detailed information in seismic data. Emphasis is placed on distinguishing between seismic multiples and primaries. The improved network is trained using seismic data containing both multiples and primaries as input and seismic data containing only primaries as output. The effectiveness and stability of the proposed method in multiple attenuation are validated using two horizontal layered velocity models and the Sigsbee2B velocity model. Transfer learning is employed to endow the trained model with the capability to suppress multiples across seismic exploration areas, effectively improving multiple attenuation efficiency.

The time-frequency electromagnetic method (TFEM) is a controllable source electromagnetic method applied to oil and gas exploration. In traditional TFEM, the amplitude and phase of the electric field are used to correct the data in the frequency domain. It also contains information about the parameters of the observation system; therefore, it is challenging to directly correct the data, which affects the efficiency and precision of data processing. This study introduces the generalized apparent resistivity defined by the horizontal current field source and obtains the apparent resistivity in the frequency domain of a TFEM by combining the characteristics of the TFEM emitter and an observation system. The feasibility of the proposed method was verified by analyzing the characteristics of the electric field amplitude and apparent resistivity of the primary observation parameters, such as the offset and launch distance. The calculation of apparent resistivity in the frequency domain of TFEM and research based on the model provide scientific references for the design of TFEM exploration and data processing.

As geological exploration conditions become increasingly complex, meeting the requirements of precise geological exploration necessitates the development of a controlled-source audio magnetotelluric (CSAMT) inversion method that considers anisotropy to improve the effectiveness of inversion accuracy and interpretation accuracy of data. This study is based on the 3D finite-difference forward modeling of axis anisotropy using the reciprocity theorem to calculate the Jacobian matrix by applying the search method to automatically search for the Lagrange operator. The aim is to establish inversion iteration equations to achieve the axis anisotropic Occam’s 3D inversion of tensor CSAMT in data space. Further, we obtain an underground axis anisotropic 3D geoelectric model by inverting the impedance data of tensor CSAMT. Two synthetic data examples show that using the isotropic tensor CSAMT algorithm to directly invert data in anisotropic media can generate false anomalies, leading to incorrect geological interpretations. Meanwhile, the proposed anisotropic inversion algorithm can effectively improve the accuracy of data inversion in anisotropic media. Further, the inversion examples verify the effectiveness and stability of the algorithm.

Enhancing the mining speed of a working face has become the primary approach to achieve high production and efficiency in coal mines, thereby further improving the production capacity. However, the problem of rock bursts resulting from this approach has become increasingly serious. Therefore, to implement coal mine safety and efficient extraction, the impact of deformation pressure caused by different mining speeds should be considered, and a reasonable mining speed of the working face should be determined. The influence of mining speed on overlying rock breaking in the stope is analyzed by establishing a key layer block rotation and subsidence model. Results show that with the increasing mining speed, the compression amount of gangue in the goaf decreases, and the rotation and subsidence amount of rock block B above goaf decreases, forcing the rotation and subsidence amount of rock block A above roadway to increase. Consequently, the contact mode between rock block A and rock block B changes from line contact to point contact, and the horizontal thrust and shear force between blocks increase. The increase in rotation and subsidence of rock block A intensifies the compression degree of coal and rock mass below the key layer, thereby increasing the stress concentration degree of coal and rock mass as well as the total energy accumulation. In addition, due to the insufficient compression of gangue in the goaf, the bending and subsidence space of the far-field key layer are limited, the length of the suspended roof increases, and the influence range of mining stress and the energy accumulation range expand. Numerical test results and underground microseismic monitoring results verify the correlation between mining speed and stope energy, and high-energy events generally appear 1–2 d after the change in mining speed. On this basis, the statistical principle confirms that the maximum mining speed of the working face at 6 m/d is reasonable.

A systematic terahertz spectroscopy study of the mineral phase transformation process of natural pyrite samples heated in a nitrogen atmosphere is conducted. In addition, the pyrolysis process of pyrite in the 400 °C–800 °C temperature range is analyzed and discussed. This study is based on X-ray diffraction (XRD) and thermogravimetric-derivative thermogravimetric (TG-DTG) analysis of the corresponding thermal transformation sequences of pyrite, magnetopyrite, and sulfurous pyrite as the desulfurization process proceeds. Terahertz time-domain spectroscopy is employed to characterize the optical properties of the pyrolysis products. The results show that pyrite, magnetopyrite and sulfurous pyrite exhibit different absorption coefficients and refractive indices in the terahertz frequency band. The different optical properties of these products provide useful information for the investigation of the pyrolysis process of pyrite and the magnetic properties of environmental sediments.