| Fine-grained sedimentary rocks represent a widespread class of rocks that exhibit small grain size,complex composition and type,high heterogeneity,and susceptibility to the influence of special hydrodynamics and water-rock environments.These rocks possess considerable scientific research and industrial value,particularly with regard to unconventional oil and gas resource exploration and development,thereby elevating their significance as research hotspots and challenges in the fields of sedimentary geology and petroleum geology.This thesis focuses on the fine-grained sedimentary rocks of Member 7 of the Yanchang Formation in northern Shaanxi Province,China.Based on previous studies and field observations,as well as thin section analysis,X-ray diffraction analysis,and environmental scanning electron microscopy,the characteristics of the fine-grained sedimentary rocks in the study area were systematically studied.In addition,deep learning techniques were used to optimize the pre-trained model for accurately identifying the lithology of fine-grained sedimentary rocks.Based on this,the sedimentary environment and paleoclimate of Member 7of the fine-grained sedimentary rocks in the study area were explored using geochemical analysis.The following insights were obtained:The Chang7 fine-grained sedimentary rock samples form the Yanchang Formation exhibit a predominant mineral composition,xconsisting of quartz,feldspar,and clay minerals.Notably,feldspar minerals are the most abundant,surpassing clay minerals in content,while carbonate minerals are relatively scarce.Moreover,the presence of minerals such as albite,jarosite,goethite,and pyrite is also evident within the samples.The primary pore types identified in these rocks encompass intra-granular dissolution pores,residual intergranular pores,and intercrystalline pores.Regarding the sedimentary structures,they can be classified into two primary categories:horizontal bedding structures and laminated structures.Siltstone and mixed fine-grained rocks are the prevailing rock types,occasionally accompanied by claystone,while the occurrence of carbonate rocks is infrequent.By integrating the characteristics of sedimentary structures,color variations,and mineral composition,the fine-grained sedimentary rocks in the study area can be delineated into six distinct lithofacies: gray-black layered siltstone facies,gray-brown laminated siltstone facies,dark organic-rich shale facies,dark muddy siltstone facies,dark gray claystone facies,and gray-black mixed fine-grained rock facies.The application of deep learning in the field of microscopic image recognition was studied,and a large number of training samples were collected and enhanced.Res Net18 was chosen as the pre-trained model,After conducting comprehensive testing and comparison of three pretrained models,namely Goog Le Net,Alex Net,and Res Net.and its structure was optimized by modifying and replacing the last three layers and adjusting the input image size.Through continuous testing,the model was finally able to accurately identify the fine-grained sedimentary rocks in the study area.The accuracy of this model was higher than that of traditional methods,and it improved efficiency and precision.This study represents an interdisciplinary exploration of geology and computer artificial intelligence.The paleoredox conditions,paleosalinity characteristics,and paleoclimate characteristics of the Yanchang Formation in northern Shaanxi were analyzed using geochemical data.It was found that the study area was mainly in a strong reducing environment and a freshwater to slightly brackish water environment,and the paleotemperature was generally high.Based on these data analyses,it was concluded that the paleoclimate of the study area was warm and humid. |
