油气地质与采收率2026，Vol.33Issue(2)：36-47,12.DOI:10.13673/j.pgre.202511022
基于深度学习的纳米级页岩三维数字岩心重构及导电机理研究
Research on 3D digital rock reconstruction and conductive mechanism of nanoscale shale based on deep learning
摘要
Abstract
Rock resistivity is a critical petrophysical parameter for calculating reservoir water saturation.To address the challenges of complex electrical conduction mechanisms and difficult rock electrical experiments in shale oil reservoirs,a process for reconstructing 3D digital shale rocks at the nanoscale and conducting numerical simulations based on deep learning was proposed to reveal the electrical conduction mechanisms and calculate the electrical parameters of shale.First,by considering the anisotropic characteristics of shale pore structures,deep learning was employed to construct 3D digital rocks with nanoscale precision;subsequently,mathematical morphology methods were utilized to simulate the microscopic oil and water distribution states in reservoirs under different saturation conditions;the Waxman-Smits model was incorporated to improve the finite element electrical simulation method,enabling accurate calculation of shale rock electrical parameters.Numerical simulation results demonstrate that shale conductivity exhibits significant anisotropy.The horizontal direction exhibits superior electrical conductivity compared to the vertical direction,yet the additional conductivity effect of clay minerals is more pronounced in the vertical direction,resulting in a greater formation resistivity factor and cementation exponent in the vertical direction than in the horizontal direction,whereas the saturation exponent is smaller in the vertical direction than in the horizontal direction.The resistivity index and its corresponding saturation exponent are jointly controlled by pore structure and additional conductivity of clay minerals.A greater influence of pore structure indicates faster variation of the resistivity index curve,while stronger additional conductivity of clay minerals leads to a more gradual curve variation.Through this 3D digital rock numerical simulation workflow,the complex electrical conduction mechanisms of shale oil reservoirs can be effectively revealed,and key rock electrical parameters are obtained,providing an essential physical basis and technical support for hydrocarbon saturation calculation and reserve evaluation.关键词
页岩油/数字岩心/纳米级孔隙/电阻率/数值模拟/深度学习Key words
shale oil/digital rock/nanoscale pore/resistivity/numerical simulation/deep learning分类
能源科技引用本文复制引用
李忠新,曹小朋,吕琦,蒋龙,程紫燕,迟蓬,孙福璟,林承焰..基于深度学习的纳米级页岩三维数字岩心重构及导电机理研究[J].油气地质与采收率,2026,33(2):36-47,12.基金项目
国家科技重大专项"渤海湾盆地济阳坳陷古近系陆相页岩油勘探开发技术与集成示范"(2024ZD1405100),中国石化股份有限公司科技项目"页岩油地质精细评价技术研究"(P24031). （2024ZD1405100）
