Machine learning is the main technical means for lithofacies logging identification. As the main target of shale oil spatial distribution prediction, mud shale petrography is subjected to the constraints of stratigraphic inhomogeneity and logging information redundancy. Therefore, choosing the most applicable machine learning method for different geological characteristics and data situations is one of the key aspects of high-precision lithofacies identification. However, only a few studies have been conducted on the applicability of machine learning methods for mud shale petrography. This paper aims to identify lithofacies using commonly used machine learning methods. The study employs five supervised learning algorithms, namely Random Forest Algorithm (RF), BP Neural Network Algorithm (BPANN), Gradient Boosting Decision Tree Method (GBDT), Nearest Neighbor Method (KNN), and Vector Machine Method (SVM), as well as four unsupervised learning algorithms, namely K-means, DBSCAN, SOM, and MRGC. The results are evaluated using the confusion matrix, which provides the accuracy of each algorithm. The GBDT algorithm has better accuracy in supervised learning, while the K-means and DBSCAN algorithms have higher accuracy in unsupervised learning. Based on the comparison of different algorithms, it can be concluded that shale lithofacies identification poses challenges due to limited sample data and high overlapping degree of type distribution areas. Therefore, selecting the appropriate algorithm is crucial. Although supervised machine learning algorithms are generally accurate, they are limited by the data volume of lithofacies samples. Future research should focus on how to make the most of limited samples for supervised learning and combine unsupervised learning algorithms to explore lithofacies types of non-coring wells.

The 8# coal seam in the Benxi Formation of the southeastern margin of the Ordos Basin is a deep coal seam with abundant coalbed methane resources. Calcite veins are commonly developed within the 8# coal seam, and their formation processes and mechanisms have significant implications for the enrichment of deep coalbed methane. Genesis of the calcite veins was analyzed to reveal the impact of the calcite veins formation on coalbed methane accumulation, with an integrated application of petrographic study by thin section, cathodoluminescence analysis, carbon-oxygen isotope analysis, and homogeneous temperature measurements of fluid inclusions. The research findings indicate that the calcite veins in the 8# coal seam can be classified into three stages: C1, C2, and C3. The diagenetic fluids of C1 primarily originated from contemporaneous seawater. The fluids responsible for the formation of C2 primarily consist of organic fluids enriched in biogenic gas, whereas the fluids contributing to the formation of C3 are primarily associated with liquid hydrocarbons originated form decarboxylation of organic matter. Furthermore, the development of both C2 and C3 is influenced by deep hydrothermal fluids resulting from tectonic heating events during the Early Cretaceous. By combining analysis of the hydrocarbon accumulation history and burial history in the study area, it has been established that C2 formation occurred during the Late Triassic to Early Jurassic, while C3 formation took place during the Late Jurassic to Early Cretaceous. The exploration and production practices in the study area have firmly established the crucial significance of the formation and evolution of calcite veins within the 8# coal seam for the migration and accumulation of coalbed methane. The research outcomes provide valuable insights for the exploration of deep coalbed methane enrichment areas.