The connectivity of natural fractures is conducive to reservoir reconstruction and can be used as a connecting channel for different oil and gas reservoir units, which is the research focus of exploration and development. The existing fracture connectivity calculation is mainly applied to the connectivity analysis of regional fault system, which relies on the quality of seismic data and imaging logging. The data requirements are high, and the transformation of micro-fractures to reservoirs is neglected. In order to solve the above problems, a slice connectivity calculation method based on semantic segmentation deep learning was proposed. Taking the thin section image of the tight sandstone reservoir of Huangliu Formation in Qiongdongnan Basin as an example, the connectivity of thin section fractures was calculated. Compared with the traditional manual identification, this method has higher efficiency. In addition to the inaccurate identification of some fractures, the overall calculation error was less than 10%, which has a good application prospect. The fracture connectivity of tight sandstone reservoirs in different fault sections of Huangliu Formation in Qiongdongnan Basin was calculated, and the connectivity of Songtao section > Lingshui section > Yaxi-Yanan section was obtained, which was in good agreement with the degree of fault development. This method combines artificial intelligence technology and geological information to effectively calculate the connectivity of micro-fractures at the slice scale, which is helpful to clarify the transformation effect of micro-fractures on reservoirs. Moreover, this method has low data requirements and is suitable for the evaluation of reservoir properties in the early stage of exploration and development.
SHI Ning, LIU Jing-shou, ZHANG Guan-jie, LEI Xin, LUO Yang, WANG Xiang.
Fracture Connectivity Calculation of Tight Sandstone Reservoir Thin Sections Using Semantic Segmentation Deep Learning[J].
Science Technology and Engineering, 2026, 26(7): 2834-2843 DOI:10.12404/j.issn.1671-1815.2501614
本次裂缝语义分割识别采用PSPNet(pyramid scene parsing network)网络模型进行识别模型的训练,PSPNet是一种基于卷积神经网络的语义分割模型[图3(a)],首先由扩张网络获得原图像1/8大小的特征图,然后采用金字塔池化模块将特征图通过4个并行池化层得到不同大小的输出,将4个输出进行上采样,然后还原到特征图的大小并与特征图连接,最后经过卷积层得到预测图像。
ZhaoXing, WangGuiwen, LiYafeng, et al. Development characteristics and logging identification of natural fractures in shale oil reservoirs of the Yingxiongling Area, Qaidam Basin[J]. Natural Gas Geoscience, 2025, 36(4): 713-733.
LiuJiangyu, RenDengfeng, QinShiyong, et al. 3D acid fracturing technique for ultra-deep fractured-vuggy carbonate reservoirs in Fuman Oilfield of Tarim Basin[J]. Petroleum Geology & Oilfield Development in Daqing, 2025, 44(3): 77-84.
ZhangGuanjie, WuKongyou, FanCaiwei, et al. Characteristics and controlling factors of fractures in low-permeability sandstone reservoirs: a case study of the Miocene Huangliu Formation in L Gasfield, Yinggehai Basin[J]. Marine Origin Petroleum Geology, 2022, 27(4): 383-395.
LiXiaomei, ZhangJigang, ChenChao, et al. Fracture development characteristics and quantitative prediction of tight reservoirs in Well Block Ma 2, Northwestern Margin, Junggar Basin[J]. Special Oil & Gas Reservoirs, 2022, 29(3): 50-56.
HuYanzhi, LiXiao, ZhangZhaobin, et al. A review on hydro-mechanical coupling simulations of hydraulic fracture network in shale reservoirs[J]. Journal of Engineering Geology, 2024, 32(4): 1381-1396.
HanPengyuan, DingWenlong, YangDebin, et al. Characteristics and main controlling factors of fracture development in the Ordovician Carbonate Reservoir, Tahe Oilfield[J]. Earth Science Frontiers, 2024, 31(5): 209-226.
ZengLianbo, GongLei, SuXiaocen, et al. Natural fractures in deep to ultra-deep tight reservoirs: distribution and development[J]. Oil & Gas Geology, 2024, 45(1): 1-14.
[15]
SandersonD J, NixonC W. The use of topology in fracture network characterization[J]. Journal of Structural Geology, 2015, 72: 55-66.
LiangXiaobo, JuWei. Fault connectivity evaluation based on topological structure analysis: a case study of multi-stage faults of deep shale gas reservoirs in central Luzhou Block, Southern Sichuan[J]. Petroleum Reservoir Evaluation and Development, 2024, 14(3): 446-457.
LiYong, HeJianhua, DengHucheng, et al. Analysis of connectivity characterization and mechanical effectiveness of natural fracture in deep shale reservoirs: a case study of the Wufeng-Longmaxi Formations in the Dingshan-Dongxi Area, Southeastern Margin of Sichuan Basin[J]. Natural Gas Geoscience, 2024, 35(2): 230-244.
HuHuiting, MaYong, ZhangSuian, et al. Anisotropy and influencing factors of shale fracture and matrix permeability[J]. Science Technology and Engineering, 2023, 23(24): 10252-10263.
LiShunyong, LiQinghui, XingYuman. Intelligent rock recognition based on lightweight network and transfer learning[J]. Science Technology and Engineering, 2025, 25(5): 1774-1782.
WanChengzhou, JiXiaohui, YangMei, et al. Mineral image recognition based on progressive deep learning across different granularity levels[J]. Earth Science Frontiers, 2024, 31(4): 112-118.
ChengGuojian, YangJing, HuangQuanzhou, et al. Rock image classification recognition based on probabilistic neural networks[J]. Science Technology and Engineering, 2013, 13(31): 9231-9235.
LiuZiyu, JiaWanli, LiJianping, et al. Sedimentary composition of turbidite fan and reservoir difference of main units: a case study of Meishan Formation in Qiongdongnan Basin[J]. Chinese Journal of Geology(Scientia Geologica Sinica), 2024, 59(5): 1268-1279.
HuLin, XiongFuhao, ChenAnqing, et al. Granitic magmatism and tectonic evolution in Qiongdongnan Basin and their constraints on the properties of buried hill reservoirs[J]. Geological Bulletin of China, 2024, 43(7): 1191-1206.
LiJianping, JiaWanli, LiHongyi, et al. Sedimentary reservoir characteristics and exploration significance of submarine fan under the dual control of provenance and geomorphology: taking the Meishan Formation in Qiongdongnan Basin as an example[J]. Chinese Journal of Geology(Scientia Geologica Sinica), 2024, 59(1): 148-165.
TangWu, XieXiaojun, XiongLianqiao, et al. Characteristics of deep source-to-sink systemsand sand prediction in Qiongdongnan Basin: a case study[J]. Marine Geology & Quaternary Geology, 2024, 44(4): 123-133.
LiuShuo, FangLiyu, LuoDing, et al. Hydrocarbon accumulation conditions and models in buried hill of Yacheng 13-A structure in Qiongdongnan Basin[J]. Sino-Global Energy, 2024, 29(9): 39-47.
ZhouJilin, LiLixia, ZhuZhenyu, et al. Thedistribution characteristics and the potential target of gas hydrate and shallow gas in the Qiongdongnan Basin[J]. Acta Geologica Sinica, 2024, 98(9): 2630-2640.
YouLi, JiangRufeng, GongYu, et al. Main controlling factors of accumulation and exploration breakthrough of deep-water submarine fan lithologic trap in the Qiongdongnan Basin[J]. Earth Science, 2024, 49(2): 749-758.
TianLanxi, ZhangGuanjie, LeiXin, et al. Evaluation of fault sealing in the No.2 Fault Zone of the Qiongdongnan Basin[J]. Geoscience, 2024, 38(5): 1235-1247.
SongRuiyou, HeXiaohu, WanYang, et al. Development characteristics and main control factors of reservoir formation of buried hills in Qiongdongnan Basin[J]. World Petroleum Industry, 2023, 30(6): 30-42.
PeiYangwen, WangJing, FanCaiwei, et al. Characterization of micro-structural deformation and its petroleum significance in Yacheng area of Qiongdongnan Basin[J]. Journal of China University of Petroleum(Edition of Natural Science), 2023, 47(3): 15-26.
[52]
CaoM, SharmaM. Effect of fracture geometry, topology and connectivity on energy recovery from enhanced geothermal systems[J]. Energy, 2023, 282: 212-315.
[53]
LauraL A, NascimentoD J L G, FernandesI G, et al. Impact of fracture topology on the fluid flow behavior of naturally fractured reservoirs[J]. Energies, 2021, 14(17): 5488-5488.
WangMinshui, WangMingchang, WangJingyu, et al. Remote sensing image classification based on fusion of attentionmechanism and weight balance algorithm[J]. Journal of Jilin University(Earth Science Edition), 2025, 55 (2): 697-704.
LiuYu, YuMing, ZhuYe. Bi-aggregation and self-merging network for few-shot imagesemantic segmentation[J]. Chinese Journal of Liquid Crystals and Displays, 2024, 39(10): 1421-1430.
JiaLili, ZhuShibo, WangJianrong, et al. Prediction of prospecting targets based on semantic segmentation with deep learning: a case study of the Yangjiang-Maoming Area in Guangdong Province[J]. Geology and Exploration, 2023, 59(5): 1093-1102.
RacolteG, MarquesA, MenezesE, et al. Stochastic gradient descent optimization to estimate the power-law fractal index in fracture networks[J]. Computers and Geosciences, 2024, 191: 105677.
[62]
HongH, TsangaratosP, IliaI, et al. Introducing a novel multi-layer perceptron network based on stochastic gradient descent optimized by a meta-heuristic algorithm for landslide susceptibility mapping[J]. Science of the Total Environment, 2020, 74: 2140549.
[63]
ChoiW P, LamK M, SiuW C. Extraction of the euclidean skeleton based on a connectivity criterion[J]. Pattern Recognition, 2003, 36(3): 721-729.
[64]
ZhuS C, YuilleA L. FORMS: a flexible object recognition and modeling system[J]. International Journal of Computer Vision, 1996, 20(3): 187-212.
[65]
XieW J, ThompsonR P, PerucchioR. A topology-preserving parallel3Dthinning algorithm for extracting the curve skeleton[J]. Pattern Recognition, 2003, 36(7): 1529-1544.
ZhongJia, DaiLong, ZhuPeiyuan, et al. Reservoir characteristics and its main controlling factors of the central canyon-channel of Huangliu Formation in Qiongdongnan Basin[J]. Marine Origin Petroleum Geology, 2023, 28(2): 169-177.