Applied Geophysics最新文献

Landslides are a type of natural disaster that can cause substantial harm to humanity. Monitoring and predicting the initiation of potential landslides is critical to avoiding losses due to disasters and economic activities. The impact of the controlled-source audio-frequency magnetotelluric method on investigating landslide surfaces is assessed through numerical simulations with a finite element approach. A Dirichlet boundary condition is selected to match the truncated boundary, resulting in a remarkable improvement in simulation efficiency. Rederivation of the formulas for a layered medium adept to the controlled-source audio-frequency magnetotelluric method is necessary to determine the electromagnetic field at any location along the truncated boundary. After the reliability evaluation of the new codes, a landslide model with a slide surface is designed, and the characteristics of its electromagnetic field and the apparent resistivity are studied. Instead of the total electromagnetic field, which is strongly influenced by topography variation, the apparent resistivity should be used for sliding surface detection. The normalized pure anomalous electromagnetic field may also be employed to quickly assess the detectability of the sliding surface. Overall, this study demonstrates that the controlled-source audio-frequency magnetotelluric method can be employed for investigating landslides, and recommends survey parameters, including configuration, frequency range, and length of survey line in landslide exploration.

To detect bull’s-eye anomalies in low-frequency seismic inversion models, the study proposed an advanced method using an optimized you only look once version 7 (YOLOv7) model. This model is enhanced by integrating advanced modules, including the bidirectional feature pyramid network (BiFPN), weighted intersection-over-union (wise-IoU), efficient channel attention (ECA), and atrous spatial pyramid pooling (ASPP). BiFPN facilitates robust feature extraction by enabling bidirectional information flow across network scales, which enhances the ability of the model to capture complex patterns in seismic inversion models. Wise-IoU improves the precision and fineness of reservoir feature localization through its weighted approach to IoU. Meanwhile, ECA optimizes interactions between channels, which promotes effective information exchange and enhances the overall response of the model to subtle inversion details. Lastly, the ASPP module strategically addresses spatial dependencies at multiple scales, which further enhances the ability of the model to identify complex reservoir structures. By synergistically integrating these advanced modules, the proposed model not only demonstrates superior performance in detecting bull’s-eye anomalies but also marks a pioneering step in utilizing cutting-edge deep learning technologies to enhance the accuracy and reliability of seismic reservoir prediction in oil and gas exploration. The results meet scientific literature standards and provide new perspectives on methodology, which makes significant contributions to ongoing efforts to refine accurate and efficient prediction models for oil and gas exploration.

In this study, we use the Bohai Sea area as an example to investigate the characteristics of secondary microseisms and their impact on seismic noise based on the temporal frequency spectral analysis of observation data from 33 broadband seismic stations during strong gust periods, and new perspectives are proposed on the generation mechanisms of secondary microseisms. The results show that short-period double-frequency (SPDF) and long-period double-frequency (LPDF) microseisms exhibit significant alternating trends of strengthening and weakening in the northwest area of the Bohai Sea. SPDF microseisms are generated by irregular wind waves during strong offshore wind periods, with a broad frequency band distributed in the range of 0.2–1 Hz; LPDF microseisms are generated by regular swells during periods of sea wind weakening, with a narrow frequency band concentrated between 0.15 and 0.3 Hz. In terms of temporal dimensions, as the sea wind weakens, the energy of SPDF microseisms weakens, and the dominant frequencies increase, whereas the energy of LPDF microseisms strengthens and the dominant frequencies decrease, which is consistent with the process of the decay of wind waves and the growth of swells. In terms of spatial dimensions, as the microseisms propagate inland areas, the advantageous frequency band and energy of SPDF microseisms are reduced and significantly attenuated, respectively, whereas LPDF microseisms show no significant changes. And during the propagation process in high-elevation areas, LPDF microseisms exhibit a certain site amplification effect when the energy is strong. The results provide important supplements to the basic theory of secondary microseisms, preliminarily reveal the relationship between the atmosphere, ocean, and seismic noise, and provide important theoretical references for conducting geological and oceanographic research based on the characteristics of secondary microseisms.

Waveforms of artificially induced explosions and collapse events recorded by the seismic network share similarities with natural earthquakes. Failure to identify and screen them in a timely manner can introduce confusion into the earthquake catalog established using these recordings, thereby impacting future seismological research. Therefore, the identification and separation of natural earthquakes from continuous seismic signals contribute to the monitoring and early warning of destructive tectonic earthquakes. A 1D convolutional neural network (CNN) is proposed for seismic event classification using an efficient channel attention mechanism and an improved light inception block. A total of 9937 seismic sample records are obtained after waveform interception, filtering, and normalization. The proposed model can obtain better classification performance than other major existing methods, exhibiting 96.79% overall classification accuracy and 96.73%, 94.85%, and 96.35% classification accuracy for natural seismic events, collapse events, and blasting events, respectively. Meanwhile, the proposed model is lighter than the 2D convolutional and common inception networks. We also apply the proposed model to the seismic data recorded at the University of Utah seismograph stations and compare its performance with that of the CNN-waveform model.

Full waveform inversion methods evaluate the properties of subsurface media by minimizing the misfit between synthetic and observed data. However, these methods omit measurement errors and physical assumptions in modeling, resulting in several problems in practical applications. In particular, full waveform inversion methods are very sensitive to erroneous observations (outliers) that violate the Gauss–Markov theorem. Herein, we propose a method for addressing spurious observations or outliers. Specifically, we remove outliers by inverting the synthetic data using the local convexity of the Gaussian distribution. To achieve this, we apply a waveform-like noise model based on a specific covariance matrix definition. Finally, we build an inversion problem based on the updated data, which is consistent with the wavefield reconstruction inversion method. Overall, we report an alternative optimization inversion problem for data containing outliers. The proposed method is robust because it uses uncertainties. This method enables accurate inversion, even when based on noisy models or a wrong wavelet.

Rotary steering systems (RSSs) have been increasingly used to develop horizontal wells. A static push-the-bit RSS uses three hydraulic modules with varying degrees of expansion and contraction to achieve changes in the pushing force acting on the wellbore in different sizes and directions within a circular range, ultimately allowing the wellbore trajectory to be drilled in a predetermined direction. By analyzing its mathematical principles and the actual characteristics of the instrument, a vector force closed-loop control method, including steering and holding modes, was designed. The adjustment criteria for the three hydraulic modules are determined to achieve rapid adjustment of the vector force. The theoretical feasibility of the developed method was verified by comparing its results with the on-site application data of an imported rotary guidance system.

This study combines ground penetrating radar (GPR) and convolutional neural networks for the intelligent detection of underground road targets. The target location was realized using a gradient-class activation map (Grad-CAM). First, GPR technology was used to detect roads and obtain radar images. This study constructs a radar image dataset containing 3000 underground road radar targets, such as underground pipelines and holes. Based on the dataset, a ResNet50 network was used to classify and train different underground targets. During training, the accuracy of the training set gradually increases and finally fluctuates approximately 85%. The loss function gradually decreases and falls between 0.2 and 0.3. Finally, targets were located using Grad-CAM. The positioning results of single and multiple targets are consistent with the actual position, indicating that the method can effectively realize the intelligent detection of underground targets in GPR.

Stress field movements result directly from earthquakes; therefore, observing the stress field is significant. Experiments on the relationships among wave velocity, stress factors, and faults show that the wave velocity of rock media under stable stress fields corresponds one-to-one with stress factors. Therefore, the wave velocity gradient can indicate the direction of stress loss, and the gradient divergence can indicate the strength of the stress field. To verify the results, considering the limitations of wave velocity measurement in solid crustal media, two quantities, namely the apparent wave velocity and Poisson ratios relating to wave velocity, were used to reflect the stress field state. The seismic data of the Tangshan and Luzhou regions were studied separately. The calculated apparent wave velocity and Poisson ratios were interpolated to achieve regional data gridding. The gradients and the gradient divergences of the apparent wave velocity and Poisson ratio fields in the two regions were analyzed, and it was found that their spatial distribution in the same region was the same. They are believed to reflect the vertical projection of the stress direction loss and strength on the surface in the stress field, consistent with the experimental results. Whether it can effectively reflect the stress field requires further analysis of the specific situation of the local medium and the movement mode of the stress field.

The verification of nuclear test ban necessitates the classification and identification of infrasound events. The accurate and effective classification of seismic and chemical explosion infrasounds can promote the classification and identification of infrasound events. However, overfitting of the signals of seismic and chemical explosion infrasounds easily occurs during training due to the limited amount of data. Thus, to solve this problem, this paper proposes a classification method based on the mixed virtual infrasound data augmentation (MVIDA) algorithm and multiscale squeeze-and-excitation ResNet (MS-SE-ResNet). In this study, the effectiveness of the proposed method is verified through simulation and comparison experiments. The simulation results reveal that the MS-SE-ResNet network can effectively determine the separability of chemical explosion and seismic infrasounds in the frequency domain, and the average classification accuracy on the dataset enhanced by the MVIDA algorithm reaches 81.12%. This value is higher than those of the other four types of comparative classification methods. This work also demonstrates the effectiveness and stability of the augmentation algorithm and classification network in the classification of few-shot infrasound events.

The propagation of seismic waves in viscous media, such as the loess plateau and shallow gas regions, alters their amplitude, frequency, and phase due to absorption attenuation, resulting in reductions in the resolution and fidelity of seismic profiles and the inaccurate identification of subtle structure and lithology. Q modeling and Q migration techniques proposed in this paper are used to compensate for the energy and frequency attenuation of seismic waves, obtain high-quality depth imaging results, and further enhance structural imaging to address the aforementioned problem. First, various prior information is utilized to construct an initial Q model. Q tomography techniques are employed to further optimize the precision of the initial Q model and build a high-precision Q model. Subsequently, Q prestack depth migration technology is employed to compensate for absorption and attenuation in the three-dimensional space along the seismic wave propagation path and correct the travel times, realizing the purposes of amplitude compensation, frequency recovery, and phase correction, which can help improve the wave group characteristics while enhancing the resolution. Model data and practical application results demonstrate that high-precision Q modeling and Q migration techniques can substantially improve the imaging quality of underground structures and formations in the loess plateau region with extremely complex surface and near-surface conditions. The resolution and fidelity of seismic data, as well as the capability to identify reservoirs, can be improved using these techniques.