|
|
|
| High-fidelity mesoscopic heterogeneity modeling and mechanical properties investigation of rock based on mask R-CNN and DDA |
| GUO Longxiao1, WANG Yan1, SHAO Wenhui1, CHEN Guangqi1,2, MA Guowei1 |
(1. School of Civil and Transportation Engineering, Hebei University of Technology, Tianjin 300401, China;
2. Department of Civil and Structural Engineering, Kyushu University, Fukuoka, 819–0395, Japan) |
|
|
|
|
Abstract The macroscopic mechanical behavior of rock is fundamentally governed by its mesostructure. However, most existing mesoscopic DDA models rely on ideal geometric units and structural simplifications, making it challenging to accurately represent the actual mesoscopic characteristics of rock. To address this limitation, this study proposes a mesoscopic numerical modeling approach for rock that integrates Mask R-CNN and DDA, enabling the direct construction of high-fidelity numerical models from real mesoscopic images. The proposed method first enhances mineral boundary features in rock images through wavelet transform, and then utilizes Mask R-CNN for accurate identification and segmentation of minerals and pores. Based on this, a high-precision DDA model is established to systematically analyze the influence of mineral composition and pore characteristics on the mechanical properties of rock. The results indicate that quartz can inhibit the initiation and propagation of microcracks due to its high elastic modulus and strong interface contact strength, while the low contact strength of feldspar exacerbates rock stress concentration, leading to a linear positive correlation between the quartz-to-feldspar ratio (QFR) and uniaxial compressive strength (UCS). The geometric heterogeneity of pores and minerals induces local stress concentration in the rock, promoting an increase in microcracks in high-porosity areas and resulting in a nonlinear negative correlation between porosity and UCS. This method overcomes the limitations of traditional modeling associated with oversimplification, achieving a realistic reproduction of rock mesostructures and highlighting the impact of mesoscopic heterogeneity on macroscopic mechanical behavior. It provides a novel tool and insights for revealing the mesoscopic mechanisms underlying the macroscopic mechanical behavior of rock.
|
|
|
|
|
|
[1] TANG C A，HUDSON J A. Rock failure mechanisms：explained and illustrated[M]. Boca Raton：CRC，2010：256–263.
[2] LIU G，CAI M，HUANG M. Mechanical properties of brittle rock governed by micro-geometric heterogeneity[J]. Computers and Geotechnics，2018，104(12)：358–372.
[3] 李明耀，彭 磊，左建平，等. 基于真实细观结构的FFT数值方法对岩石材料非线性力学行为的研究[J]. 矿业科学学报，2022，7(4)：456–466.(LI Mingyao，PENG Lei，ZUO Jianping，et al. Research on nonlinear mechanical behavior of rock materials using FFT numerical method based on real mesostructure[J]. Journal of Mining Science and Technology，2022，7(4)：456–466.(in Chinese))
[4] ZHAO Z，SHOU Y，ZHOU X. Microscopic cracking behaviors of rocks under uniaxial compression with microscopic multiphase heterogeneity by deep learning[J]. International Journal of Mining Science and Technology，2023，33(4)：411–422.
[5] LI F，YANG Y，FAN X，et al. Numerical analysis of the hydrofracturing behaviours and mechanisms of heterogeneous reservoir rock using the continuum-based discrete element method considering pre-existing fractures[J]. Geomechanics and Geophysics for Geo-Energy and Geo-Resources，2018，4(9)：383–401.
[6] YANG Y，WU W，XU D，et al. A nodal-based 3D discontinuous deformation analysis method with contact potential for discrete rock block system[J]. Rock Mechanics and Rock Engineering，2023，56(6)：4 043–4 059.
[7] ZHENG F，ZHUANG X，ZHENG H，et al. Discontinuous deformation analysis with distributed bond for the modelling of rock deformation and failure[J]. Computers and Geotechnics，2021，139(11)：104413.
[8] ZHANG K，LIU F，XIA K，et al. On the calibration and verification of Voronoi-based discontinuous deformation analysis for modeling rock fracture[J]. Journal of Rock Mechanics and Geotechnical Engineering，2023，15(8)：2 025–2 038.
[9] 高善铧，张开雨，杨 玫，等. 基于GB-DDA方法的岩石细观非均质性对其压缩力学特性影响研究[J]. 岩石力学与工程学报，2025，44(6)：1 612–1623.(GAO Shanhua，ZHANG Kaiyu，YANG Mei，et al. Research on the influence of rock microscopic heterogeneity on its compressive mechanical properties based on the GB-DDA method[J]. Chinese Journal of Rock Mechanics and Engineering，2025，44(6)：1 612–1 623.(in Chinese))
[10] GUO L，MA G，CEN G. A micro frost heave model for porous rock considering pore characteristics and water saturation[J]. Computers and Geotechnics，2024，166(10)：106029.
[11] 孙 闯，敖云鹤，张家鸣，等. 花岗岩细观破裂特征及宏观尺度效应的颗粒流研究[J]. 岩土工程学报，2020，42(9)：1 687–1 695.(SUN Chuang，AO Yunhe，ZHANG Jiaming，et al. Particle flow study on meso-fracture characteristics and macro-scale effects of granite[J]. Chinese Journal of Geotechnical Engineering，2020，42(9)：1 687–1 695.(in Chinese))
[12] 丁自伟，李小菲，唐青豹，等. 砂岩颗粒孔隙分布分形特征与强度相关性研究[J]. 岩石力学与工程学报，2020，39(9)：1 787–1 796. (DING Ziwei，LI Xiaofei，TANG Qingbao，et al. Study on correlation between fractal characteristics of sandstone grain-pore distribution and strength[J]. Chinese Journal of Rock Mechanics and Engineering，2020，39(9)：1 787–1 796.(in Chinese))
[13] SHENG G，ZHAO H，SU Y，et al. An analytical model to couple gas storage and transport capacity in organic matter with noncircular pores[J]. Fuel，2020，268(5)：117288.
[14] 王志宽，李 昕. 基于语义一致性的精确目标移除[J]. 计算机与数字工程，2025，53(1)：234–239.(WANG Zhikuan，LI Xin. Semantic consistency-based precise object removal[J]. Computer and Digital Engineering，2025，53(1)：234–239.(in Chinese))
[15] KHAN S S，KHAN M，ALHARBI Y. Multi focus image fusion using image enhancement techniques with wavelet transformation[J]. International Journal of Advanced Computer Science and Applications，2020，11(5)：414–420.
[16] AHMAD Q A，EHSAN M I，KHAN N，et al. Numerical simulation and modeling of a poroelastic media for detection and discrimination of geo-fluids using finite difference method[J]. Alexandria Engineering Journal，2022，61(5)：3 447–3 462.
[17] ZHANG Y，MA Y，LI Y，et al. A deep learning approach of RQD analysis for rock core images via cascade mask R-CNN-based model[J]. Rock Mechanics and Rock Engineering，2024，57(8)：11 381–11 398.
[18] 王桂林，王润秋，孙 帆. 块体离散元颗粒模型细观参数标定方法及花岗岩细观演化模拟[J]. 长江科学院院报，2022，39(1)：86–93.(WANG Guilin，WANG Runqiu，SUN Fan. A discrete element GBM simulation method for meso-parameter calibration and granite meso-evolution simulation[J]. Journal of Yangtze River Scientific Research Institute，2022，39(1)：86–93.(in Chinese))
[19] 杨圣奇，孙博文，田文岭. 不同层理页岩常规三轴压缩力学特性离散元模拟[J]. 工程科学学报，2022，44(3)：430–439.(YANG Shengqi，SUN Bowen，TIAN Wenling. Discrete element simulation of conventional triaxial compression mechanical properties of shale with different bedding planes[J]. Chinese Journal of Engineering，2022，44(3)：430–439.(in Chinese))
[20] 岳中琦. 基于数字图像的真实细观非均质岩土力学分析数值方法[J]. 地球物理学报，2022，65(1)：108–118.(YUE Zhongqi. Digital image-based numerical method for geomechanical analysis of real meso-heterogeneous soils and rocks [J]. Chinese Journal of Geophysics，2022，65(1)：108–118.(in Chinese))
[21] SHI G H. Discontinuous deformation analysis-a new numerical model for the statics and dynamics of block system[Ph. D. Thesis][D]. Berkeley：University of California，1988.
[22] 梁绍敏，冯云田，赵婷婷，等. 颗粒材料破碎行为数值分析方法研究综述[J]. 力学学报，2024，56(1)：1–22.(LIANG Shaomin，FENG Yuntian，ZHAO Tingting，et al. A review of numerical methods for modelling particle breakage[J]. Chinese Journal of Theoretical and Applied Mechanics，2024，56(1)：1–22.(in Chinese))
[23] HE K，GKIOXARI G，DOLLAR P，et al. Mask r-CNN[C]// Proceedings of the IEEE International Conference on Computer Vision. [S. l.]：[s. n.]，2017：2 961–2 969.
[24] WANG J，CHEN G，JABOYEDOFF M，et al. Loess landslides detection via a partially supervised learning and improved Mask-RCNN with multi-source remote sensing data[J]. Catena，2023，231(10)：107371.
[25] HU Z，MEI H，YU L. An intelligent prediction method for rock core integrity based on deep learning[J]. Scientific Reports，2025，15(1)：6 456.
[26] LI B，LIMA D. Facial expression recognition via ResNet-50[J]. International Journal of Cognitive Computing in Engineering，2021，2(6)：57–64.
[27] ZHU L，LEE F，CAI J，et al. An improved feature pyramid network for object detection[J]. Neurocomputing，2022，262(5)：1 202–1 208.
[28] VU T，JANG H，PHAM T X，et al. Cascade RPN：Delving into high-quality region proposal network with adaptive convolution[J]. Advances in neural information processing systems，2019，32(5)：7 537–7 546.
[29] JIANG B，LUO R，MAO J，et al. Acquisition of localization confidence for accurate object detection[C]//Proceedings of the European Conference On Computer Vision(ECCV). [S. l.]：[s. n.]，2018：784–799.
[30] GUARIGLIA E，GUIDO R C. Chebyshev wavelet analysis[J]. Journal of Function Spaces，2022，2022(1)：5542054.
[31] OTHMAN G，ZEEBAREE D Q. The applications of discrete wavelet transform in image processing：a review[J]. Journal of Soft Computing and Data Mining，2020，1(2)：31–43.
[32] 赖 文，蒋璟鑫，邱检生，等. 南京大学岩石教学薄片显微图像数据集[J]. 中国科学数据：中英文网络版，2020，5(3)：26–38.(LAI Wen，JIANG Jingxin，QIU Jiansheng，et al. Rock teaching thin section microscopic image dataset of nanjing university[J]. Science Data of China：Chinese and English Online，2020，5(3)：26–38.(in Chinese))
[33] WANG L，LU J，LUO Y，et al. An automated quantitative methodology for computing gravel parameters in imaging logging leveraging deep learning：a case analysis of the Baikouquan Formation within the Mahu Sag[J]. Processes，2024，12(7)：1 337.
[34] ROSTIANINGSIH S，SETIAWAN A，HALIM C I. COCO (creating common object in context) dataset for chemistry apparatus[J]. Procedia Computer Science，2020，171(4)：2 445–2 452.
[35] SOFIIUK K，PETROV I A，KONUSHIN A. Reviving iterative training with mask guidance for interactive segmentation[C]// 2022 IEEE International Conference on Image Processing (ICIP). [S. l.]：[s. n.]，2022：3 141–3 145.
[36] 汤翔中，高丙朋. 融合注意力空洞卷积和Transformer的矿石图像分割[J]. 科学技术与工程，2023，23(16)：6 974–6 982.(TANG Xiangzhong，GAO Bingpeng. Ore image segmentation based on attention hole convolution and transformer[J]. Science Technology and Engineering，2023，23(16)：6 974–6 982.(in Chinese))
[37] IFTIKHAR C M A，KHAN A S，NAMBORI V. The effect of temperature on the mechanical behavior of Berea sandstone under confining pressure：experiments[J]. International Journal of Geo-Engineering，2023，14(11)：14–18.
[38] GUO P，ZHANG P，BU M，et al. Impact of cooling rate on mechanical properties and failure mechanism of sandstone under thermal–mechanical coupling effect[J]. International Journal of Coal Science and Technology，2023，10(1)：26.
[39] 冯佳歆，李 皋，谢 强. 致密砂岩细观非均质数值模型构建及其应用[C]// 2022油气田勘探与开发国际会议论文集Ⅱ. 西安：[s. n.]，2022：653–660.(FENG Jiaxin，LI Gao，XIE Qiang. Construction and application of meso-heterogeneity numerical model of tight sandstone[C]// Proceedings of the 2022 International Field Exploration and Development Conference II. Xi?an：[s. n.]，2022：653–660.(in Chinese))
[40] 林铁军，曾 馨，孙 鑫，等. 基于等效模量法的数字岩芯三轴压缩行为研究[J]. 西南石油大学学报：自然科学版，2025，47(2)：84–94.(LIN Tiejun，ZENG Xin，SUN Xin，et al. A Study on triaxial compression behavior of digital core based on equivalent modulus method[J]. Journal of Southwest Petroleum University：Science and Technology，2025，47(2)：84–94.(in Chinese))
[41] TAROKH A，FAKHIMI A. Discrete element simulation of the effect of particle size on the size of fracture process zone in quasi-brittle materials[J]. Computers and Geotechnics，2014，62(10)：51–60.
[42] MISHRA D A，RAM B K. A review on evaluation of microstructural parameters to estimate the strength of virtually isotropic rock materials[J]. Geotechnical and Geological Engineering，2024，42(6)：4 627–4 649.
[43] PENG S, MARONE F, DULTZ S. Resolution effect in X-ray microcomputed tomography imaging and small pore’s contribution to permeability for a Berea sandstone[J]. Journal of Hydrology，2014，510(3)：403–411.
|
|
|
|
2026, Vol. 45 
