Xuepeng Cui

Acoustic impedance inversion is crucial for reservoir prediction and lithologic boundary characterization. However, conventional sparsity-constraint inversion methods often encounter issues such as excessive smoothing or staircase effects, leading to potential loss of detail. This study proposes a novel approach that integrates high-order Lp quasi-norm (HOLp) and overlapping group sparse (OGS) regularization to improve detail fidelity in inversion results. HOLp sharpens edges, while OGS regularization smooths global staircase artifacts, providing complementary benefits. To stabilize the inversion process and address multiconstraint conditions, we employ the alternating direction method of multipliers, utilizing the majorization–minimization algorithm to solve the HOLp and a reweighted L1 algorithm to handle the OGS term. Numerical experiments and field data applications indicate that our method improves the accuracy and resolution of acoustic impedance results, particularly in identifying thin reservoirs and complex stratigraphic features. Hence, the proposed method effectively mitigates common inversion artifacts, leading to more precise reservoir characterization.

The accurate prediction of coal structure is important to guide the exploration and development of coal reservoirs. Most prediction models are interpreted for a single sensitive coal seam, and the selection of sensitive parameters is correlated with the coal structure, but they ignore the interactions between different attributes. Part of it introduces the concept of the geological strength index (GSI) of coal rocks in order to achieve a multi-element macroscopic description and quantitative characterization of coal structure; however, the determination of coal structure involves some uncertainties among the properties of coal, such as lithology, gas content and tectonic fracture, due to their complex nature. Fuzzy inference systems provide a knowledge discovery process to handle uncertainty. The study shows that a type-2 fuzzy inference system (T2-FIS) with multi-attribute fusion is used to effectively fuse pre-stack and post-stack seismic inversion reservoir parameters and azimuthal seismic attribute parameters in order to produce more accurate prediction results for the Hengling block in the Shanxi area. The fuzzy set rules generated in this paper can provide a more reliable prediction of coal structure in the GSI system. The proposed system has been tested on various datasets and the results show that it is capable of providing reliable and high-quality coal structure predictions.

Following the advancement of machine learning-based seismic feature classification techniques for complex reservoirs, the acquisition and analysis of reliable seismic samples involved in seismic facies analysis and network-based inversion have emerged as a current research hotspot in the field of intelligent seismic processing. Many investigations focus on the improvement of model classification algorithms and neural networks. However, creating and collecting labels for massive seismic data are highly time-consuming and laborious, and suffer from sample unreliability and category imbalance in the case of small-sample labels. To address such problems, a multiscale and multilabel consistent principal component analysis-linear discriminant analysis (PCA-LDA) algorithm to learn a robust feature discriminative dictionary for classification is presented. In addition to the automatic use of multilabels from well logs and core analysis, we have associated multiscale with well trajectory locations to enrich sample information and enhance the reliability of the samples during 3D sample acquisition. More specifically, we begin by proposing an approach for the automatic collection of multiscale multilabel 3D poststack seismic samples along the well track. Next, the multilabel sequence in the scan window is fed into the Boyer-Moore majority vote algorithm for sample segmentation, which constructs multilabel hierarchies for each sample. Then to enhance the model training bias due to small-sample label imbalance, we develop a novel label-shuffling balanced strategy, which obtains a complete database by filling random unduplicated augmented training samples (spatial and frequency-domain augmentation operations). Finally, the linear robustness decision-making space of PCA-LDA is obtained using the feature mapping space of PCA, as well as its visual representation. Experimental results on synthetic and field seismic data demonstrate that robust feature extraction with a trustworthy and complete multiscale and multilabel sample database increases classification accuracy.