Compressional and shear wave velocities (Vp and Vs, respectively) are important elastic parameters to predict reservoir parameters, such as lithology and hydrocarbons. Due to acquisition technologies and economy, the shear wave velocity is generally lacking. Over the last few years, some researchers proposed deep learning algorithms to predict the shear wave velocity using conventional logging data. However, these algorithms focus either on spatial feature extraction for different physical properties of rocks or on sequential feature extraction in the depth direction of rocks. Only focusing on feature extraction in a direction of rocks might lead to a decrease in prediction accuracy. Therefore, we propose a hybrid network of a two-dimensional convolutional neural network and the gated recurrent unit (2DCNN-GRU), which can establish more complex nonlinear relationships between the input and output data based on the spatial features extracted by 2DCNN and the sequential features extracted by GRU. In this study, two cases are used to validate the reliability and prediction accuracy of the proposed network. Comparing the prediction results of 2DCNN, GRU, and the proposed network, the proposed network shows better performance. Meanwhile, for improving the prediction accuracy of the proposed network, the relationship is analyzed between the prediction accuracy of the proposed network and the length of the input sample.
Compression-wave velocity and shear-wave velocity are important elastic parameters describing deeply tight sandstone. Limited by cost and technical reasons, the conventional logging data generally lack shear-wave velocity. In addition, the existing rock physics theory is difficult to accurately establish the rock physics models due to the complex pore structure of tight sandstone reservoir. With the rapid development of the artificial intelligence, the attention mechanism that can increase the sensitivity of the network to important characteristics has been widely used in machine translation, image processing, and other fields, but it is rarely used to predict shear-wave velocity. Based on the correlation between the shear-wave velocity and the conventional logging data in the spatiotemporal direction, a gate recurrent unit (GRU) fusion network based on the spatiotemporal attention mechanism (STAGRU) is proposed. Compared with the convolutional neural network (CNN) and gate recurrent unit (GRU), the network proposed can improve the sensitivity of the network to important spatiotemporal characteristics using the spatiotemporal attention mechanism. It is analyzed that the relationship between the spatiotemporal characteristics of the conventional logging data and the attention weights of the network proposed to verify the rationality of adding the spatiotemporal attention mechanism. Finally, the training and testing results of the STAGRU, CNN, and GRU networks show that the prediction accuracy and generalization of the network proposed are better than those of the other two networks.
Microseismic monitoring is an important technology used to evaluate hydraulic fracturing, and denoising is a crucial processing step. Analyses of the characteristics of acquired three-component microseismic data have indicated that the vertical component has a higher signal-to-noise ratio (SNR) than the two horizontal components. Therefore, we propose a new denoising method for three-component microseismic data using re-constrain variational mode decomposition (VMD). In this method, it is assumed that there is a linear relationship between the modes with the same center frequency among the VMD results of the three-component data. Then, the decomposition result of the vertical component is used as a constraint to the whole denoising effect of the three-component data. On the basis of VMD, we add a constraint condition to form the re-constrain VMD, and deduce the corresponding solution process. According to the synthesis data analysis, the proposed method can not only improve the SNR level of three-component records, it also improves the accuracy of polarization analysis. The proposed method also achieved a satisfactory effect for field data.
Shear wave velocity prediction is critical for applications in petrophysics, reservoir characterization, and unconventional energy resource development. While empirical formulas and theoretical rock physics models offer solutions, they are often limited by geological complexity, high cost, and computational inefficiency. After the emergence of deep learning methods, a series of new approaches have been provided to tackle these problems. In this study, a novel Inception–attention–BiLSTM hybrid network is proposed to enhance shear wave prediction accuracy and stability by integrating the strengths of three components: Inception for multi-scale feature extraction, attention mechanisms for dynamically highlighting key temporal features, and BiLSTM for capturing long-term dependencies in logging data. The test dataset of this network comes from the Jurassic Badaowan Formation in the Junggar Basin, achieving superior performance compared to standalone Inception and BiLSTM networks. The proposed hybrid network demonstrated MAE and R2 values of 0.211 and 0.994, respectively, outperforming Inception (MAE 0.671, R2 0.981) and BiLSTM (MAE 0.215, R2 0.991). These results underscore its robustness in handling complex logging data, providing a more accurate and generalizable framework for Vs prediction while addressing limitations of traditional methods. This work highlights the potential of hybrid deep learning architectures in advancing logging data analysis and reservoir characterization.
The modeling method of hydraulic fracturing fracture network combined with microseismic monitoring is currently a hot research topic. In the process of hydraulic fracturing production, previous studies have established a fracture model after fracturing based on the recorded source coordinates, earthquake time, and moment magnitude throughout the entire fracturing cycle, in order to evaluate the fracturing results. However, in the process of reservoir fracturing transformation, real-time evaluation of fracturing effect and provision of fracturing guidance are required. The method of establishing fracture models through long-term data collection is obviously not applicable. Nevertheless, existing research lacks efficient and direct 3D fracture network modeling methods based on microseismic events. To address these challenges, this study introduces a novel three-dimensional fracture network model reconstruction process, DEFME (Dimensionality Enhancement and Fracture Modeling of Microseismic Event Location Information), specifically tailored for datasets containing only the source coordinate information of microseismic events. This innovative process initially incorporates the improved DB-RANSAC (Density-Based Random Sample Consensus) method, introducing new regional microseismic event normal vector constraint information. It transforms the conventional DBSCAN(Density-Based Spatial Clustering of Applications with Noise) clustering algorithm into the NC-DBSCAN (Normal Vector Constraint-DBSCAN) method, combining normal vector constraints with density constraints for microseismic event point clustering. Finally, the Alpha-shape method is employed to extract contour information for each fracture, constructing a three-dimensional plane geometric model of the fracture. To validate the efficacy of the proposed processing workflow, synthetic data and public dataset are utilized. The corresponding results indicate that DEFME has successfully overcome noise interference and efficiently and accurately established a three-dimensional fracture network model based on microseismic events.
The arrival time of a microseismic event is an important piece of information for microseismic monitoring. The accuracy and efficiency of arrival time identification is affected by many factors, such as the low signal-to-noise ratio (SNR) of the records, the vast amount of real-time monitoring records, and the abnormal situations of monitoring equipment. In order to eliminate the interference of these factors, we propose a method based on phase-only correlation (POC) to estimate the relative arrival times of microseismic events. The proposed method includes three main steps: (1) The SNR of the records is improved via time-frequency transform, which is used to obtain the time-frequency representation of each trace of a microseismic event. (2) The POC functions of all pairs of time-frequency representations are calculated. The peak value of the POC function indicates the similarity of the traces, and the peak position in the time lag axis indicates the relative arrival times between the traces. (3) Using the peak values as weighting coefficients of the linear equations, consistency processing is used to exclude any abnormal situations and obtain the optimal relative arrival times. We used synthetic data and field data to validate the proposed method. Comparing with Akaike information criterion (AIC) and cross-correlation, the proposed method is more robust at estimating the relative arrival time and excluding the influence of abnormal situations.