Missing traces interpolation is a basic step in the seismic data processing workflow. Recently, many seismic data interpolation methods based on different neural networks have been proposed. The existing research shows that when the seismic data are consecutively missing, the larger gap for missing traces, the more difficult task of interpolation, due to convolution operation in the neural network can only capture local relations. In this letter, we incorporate the coordinate attention block to the Unet for 2-D successive missing traces interpolation. The hybrid loss function combined with structural similarity (SSIM) and $\text {L}_{ {1}}$ norm is used as the loss function to further improve the interpolation performance of the designed network. Comparison experiments on 2-D synthetic and field seismic data show that the interpolation results obtained by the proposed method are more accurate and reasonable compared with Unet and Unets equipped state-of-the-art similar modules.

Deep learning has been widely used in seismic inversion. Since the label data obtained in production is actually a small amount of one-dimensional (1D) well-log data, most deep learning based seismic inversion is 1D trace-to-trace method based on poststack seismic data. However, the 1D algorithm does not take seismic data spatial correlation into consideration and the results lack of stability with good lateral continuity. Meanwhile, it is still a challenge to obtain high-precision multi-parameter inversion results simultaneously for prestack inversion. To mitigate the above issues, we propose a 2D Multi-task Full Attention U-Net for prestack three-parameter inversion. The proposed network can capture geological structural features in shared information and allow each task to automatically learn task-specific attention weights in a separate branch. The full attention mechanism fuses the shared features and the internal features of each task in stages, and pays attention in both channels and feature maps. Moreover, a joint hybrid loss function based on mean squared error and structural similarity is used to further enhance the continuity and resolution of inversion results. To obtain sufficient training samples, we generate a large number of small sample patches by sliding window both laterally and vertically with equidistant around the wells. In order to obtain stable inversion results, we incorporate the initial models and feed them into the network together with seismic data as input. Experiments on synthetic and field data show that our proposed method can simultaneously obtain three-parameter inversion results with high vertical resolution and transverse continuity.

In downhole microseismic monitoring, the velocity model plays a vital role in accurate mapping of the hydraulic fracturing image. For velocity model uncertainties in the number of layers or interface depths, the conventional velocity calibration method has been shown to effectively locate the perforation shots; however, it introduces nonnegligible location errors for microseismic events, especially for complex geologic formations with inclinations. To improve the event location accuracy, we have exploited the advantages of the reversible jump Markov chain Monte Carlo approach in generating different dimensions of velocity models and adopted a transdimensional Bayesian simultaneous inversion framework for obtaining the effective velocity structure and event locations simultaneously. The transdimensional inversion changes the number of layers during the inversion process and selects the optimal interface depths and velocity values to improve the event location accuracy. The confidence intervals of the simultaneous inversion event locations estimated by Bayesian inference enable us to evaluate the location uncertainties in the horizontal and vertical directions. Two synthetic examples and a field test are presented to illustrate the performance of our methodology, and the event location accuracy is shown to be higher than that obtained using the conventional methods. With less dependence on prior information, our transdimensional simultaneous inversion method can be used to obtain an effective velocity structure for facilitating highly accurate hydraulic fracturing mapping.

The velocity model plays a critical role in accurately determining microseismic event locations during hydraulic fracturing. Conventional velocity optimization methods, such as updating the velocity model using perforation shots, have the drawback of weak ray coverage during single-well monitoring. The microseismic events distributed in different areas can broaden the ray coverage and help improve the accuracy of the velocity model, whereas only a few perforation shots are available for a given stage. We have developed an incremental pseudomaster (IPM) method to reduce the velocity errors. The master event location technique is used to determine the positions of representative events in each iteration of the IPM method. These events then act as new pseudomasters to further optimize the velocity model using Bayesian inference. With progressive iterations, the number of pseudomaster events increases and their distribution range expands. The accuracy of the velocity model continuously improves, which means that the final IPM-based velocity model, generated based on the pseudomasters distributed throughout the study area, more accurately locates all microseismic events than the perforation shot alone method. One synthetic example and one field test have been conducted, and the results demonstrate that the IPM method could overcome the drawbacks of weak ray coverage and improve the accuracy of the inverted velocity model and event locations.

Due to various reasons, seismic data are often inevitably affected by noise. Therefore, random noise suppression of seismic data is a key step for seismic data processing workflow. Recently, deep learning method has performed well in seismic data denoising. In this study, we propose a self-supervised deep learning seismic data noise attenuation method. We introduced an effective Blind2Unblind (B2U) denoising framework, which can complete denoising using only a single noisy seismic data. Use a mask mapper with global awareness, which can sample all pixels at the blind spots on noisy data and map them to a same channel. At the same time, a re-visible loss function is used to train the network, which can optimize all blind spots, mitigating the information loss and retaining more details of geological structure. The denoising experiments on synthetic and field data show that our method has achieved superior results compared with previous work.