A SHAP-based interpretable machine learning approach for acoustic sensing and identification of damage in hydro-mechanically coupled sandstone
ZHANG Shuai1, 2, 3, LAI Xingping1, 2*, SUN Yao1, CAO Jiantao1, 2, CUI Feng1, 2, SHAN Pengfei1, 2, YAN Baoxu1, 2, DANG Yongchao4
(1. College of Energy and Mining Engineering, Xi?an University of Science and Technology, Xi?an, Shaanxi 710054, China; 2. Key Laboratory of Western Mines and Hazard Prevention, Ministry of Education, Xi?an University of Science and Technology, Xi?an, Shaanxi 710054, China; 3. Shaanxi Yanqiong Lai Intelligent Technology Co., Ltd., Xi?an, Shaanxi 710600, China;
4. Shaanxi Ruineng Coal Industry Co., Ltd., Yan?an, Shaanxi 727307, China)
Abstract：The synergistic evolution of the “stress field-fracture field-seepage field” in the surrounding rock during deep mining, combined with the significant alteration of the rock mass?s physical and mechanical properties by groundwater, makes it difficult to reconcile the scientific validity and practical applicability of traditional acoustic monitoring and early-warning criteria, thereby significantly increasing the difficulty of accurately predicting dynamic disasters such as rockbursts. To elucidate the damage evolution and acoustic response characteristics of hydraulically weakened sandstone and establish reliable detection methods, this study proposes a multidimensional discrimination model based on acoustic features. Four groups of sandstone specimens with different water contents—dry, naturally moist, weakly saturated, and fully saturated—were prepared. Synchronous mechanical-acoustic tests were conducted throughout uniaxial compression using an MTS electro-hydraulic servo system and a P-wave velocity monitoring platform. Ultrasonic P-wave velocity, waveform skewness, and kurtosis were extracted as key acoustic parameters to characterize their qualitative relationships with damage evolution. A random-forest-based classification model incorporating SHAP(SHapley Additive exPlanations)-weighted feature matching was then developed. The results indicate that water significantly modulates the evolution of P-wave velocity, waveform skewness, and kurtosis across different stress stages. Under fully saturated conditions, P-wave velocity does not decrease substantially during the yield stage; instead, it remains high or even increases slightly, producing a misleading acoustic signature. This finding highlights the limitations of relying solely on velocity reduction to assess rock instability. In contrast, waveform skewness and kurtosis exhibit complex but distinct variations associated with microcrack initiation and propagation. SHAP-based interpretability analysis clarifies the respective contributions of acoustic features, including P-wave velocity, skewness, and kurtosis, to damage-state identification under different saturation conditions. An interpretable random forest discrimination model integrating multidimensional acoustic features was established, achieving an overall accuracy of 71.6% in identifying the water-saturation state and stress stage of the rock. This study elucidates the acoustic characterization mechanisms of nonlinear rock damage under hydraulic weakening and proposes an integrated identification approach that combines machine-learning interpretability with rock-physics mechanisms. The findings provide an experimental foundation and theoretical support for stability assessment and intelligent early warning of instability in deeply buried, water-saturated rock masses.
[1] 谢和平. 深部高应力下的资源开采——现状、基础科学问题与展望[C]// 科学前沿与未来(第六集). 北京：中国环境科学出版社，2002：179–191.(XIE Heping. Resources development under high ground stress：present state，base science problems and perspective[C]// Science Foreland and Future(Volume VI). Beijing：China Environment Science Press，2002：179–191.(in Chinese))
[2] 谭云亮，郭伟耀，辛恒奇，等. 煤矿深部开采冲击地压监测解危关键技术研究[J]. 煤炭学报，2019，44(1)：160–172.(TAN Yunliang，GUO Weiyao，XIN Hengqi，et al. Key technology of rock burst monitoring and control in deep coal mining[J]. Journal of China Coal Society，2019，44(1)：160–172.(in Chinese))
[3] 彭苏萍. 我国煤矿安全高效开采地质保障系统研究现状及展望[J]. 煤炭学报，2020，45(7)：2 331–2 345.(PENG Suping. Current status and prospects of research on geological assurance system for coal mine safe and high efficient mining[J]. Journal of China Coal Society，2020，45(7)：2 331–2 345.(in Chinese))
[4] 王国法. 煤矿智能化最新技术进展与问题探讨[J]. 煤炭科学技术，2022，50(1)：1–27.(WANG Guofa. Latest Technological Advances and Issues in Intelligent Coal Mining[J]. Coal Science and Technology，2022，50(1)：1–27.(in Chinese))
[5] 窦林名，牟宗龙，陆菜平，等. 采矿地球物理理论与技术[M]. 北京：科学出版社，2014：224–226.(DOU Linming，MOU Zonglong，LU Caiping，et al. Mining geophysics theory and technology[M]. Beijing：Science Press，2014：224–226.(in Chinese))
[6] 汤连生，张鹏程，王思敬. 水–岩化学作用的岩石宏观力学效应的试验研究[J]. 岩石力学与工程学报，2002，21(4)：526–531.(TANG Liansheng，ZHANG Pengcheng，WANG Sijing. Experimental study on macroscopic mechanical effect of water-rock chemical action[J]. Chinese Journal of Rock Mechanics and Engineering，2002，21(4)：526–531.(in Chinese))
[7] 刘飞禹，梁崇旭，王 军，等. 不同含水率下花岗岩残积土循环剪切特性研究[J]. 岩石力学与工程学报，2023，42(8)：2 048–2 057. (LIU Feiyu，LIANG Chongxu，WANG Jun，et al. Study on cyclic shear characteristics of granite residual soil under different water contents[J]. Chinese Journal of Rock Mechanics and Engineering，2023，42(8)：2 048–2 057.(in Chinese))
[8] 刘跃东，康红普. 不同应力路径下含水泥岩崩解破坏跨尺度分形规律研究[J]. 岩石力学与工程学报，2023，42(增2)：4 162–4 173.(LIU Yuedong，KANG Hongpu. Study on cross-scale fractal law of disintegration and failure of water-bearing mudstone under different stress paths[J]. Chinese Journal of Rock Mechanics and Engineering，2023，42(Supp.2)：4 162–4 173.(in Chinese))
[9] 李波波，王忠晖，任崇鸿，等. 水–力耦合下煤岩力学特性及损伤本构模型研究[J]. 岩土力学，2021，42(2)：315–323.(LI Bobo，WANG Zhonghui，REN Chonghong，et al. Mechanical properties and damage constitutive model of coal rock under the coupled hydro-mechanical effect[J]. Rock and Soil Mechanics，2021，42(2)：315–323.(in Chinese))
[10] 周子龙，熊 成，蔡 鑫，等. 单轴载荷下不同含水率砂岩力学和红外辐射特征[J]. 中南大学学报：自然科学版，2018，49(5)：1 189–1 196.(ZHOU Zilong，XIONG Cheng，CAI Xin，et al. Mechanical and infrared radiation properties of sandstone with different water contents under uniaxial compression[J]. Journal of Central South University：Science and Technology，2018，49(5)：1 189–1 196.(in Chinese))
[11] 王 凯，蒋一峰，徐 超. 不同含水率煤体单轴压缩力学特性及损伤统计模型研究[J]. 岩石力学与工程学报，2018，37(5)：1 070–1 079. (WANG Kai，JIANG Yifeng，XU Chao. Mechanical properties and statistical damage model of coal with different moisture contents under uniaxial compression[J]. Chinese Journal of Rock Mechanics and Engineering，2018，37(5)：1 070–1 079.(in Chinese))
[12] 王思敬，唐大荣，杨志法，等. 声波技术在工程岩体测试中的初步应用[J]. 地质科学，1974，(3)：269–282.(WANG Sijing，TANG Darong，YANG Zhifa，et al. Preliminary application of acoustic technique in engineering rock mass testing[J]. Scientia Geologica Sinica，1974，(3)：269–282.(in Chinese))
[13] 张培森，赵成业，李腾辉，等. 红砂岩三轴加载过程中波速变化及能量演化规律试验研究[J]. 岩石力学与工程学报，2021，40(7)：1 369–1 382.(ZHANG Peisen，ZHAO Chengye，LI Tenghui，et al. Experimental study on wave velocity variation and energy evolution of red sandstone during triaxial loading process[J]. Chinese Journal of Rock Mechanics and Engineering，2021，40(7)：1 369–1 382.(in Chinese))
[14] 张旭龙，张 盛，王龙飞，等. 基于砂岩三轴加载过程声发射与波速变化的峰前特征应力识别研究[J]. 岩石力学与工程学报，2023，42(11)：2 711–2 726.(ZHANG Xulong，ZHANG Sheng，WANG Longfei，et al. Identification of pre-peak characteristic stress of sandstone under triaxial loading based on acoustic emission and wave velocity variation[J]. Chinese Journal of Rock Mechanics and Engineering，2023，42(11)：2 711–2 726.(in Chinese))
[15] HENG Z，XU T，KONIETZKY H，et al. New insights into the continuous-discontinuous failure characteristics of granite under Brazilian splitting test conditions using acousto-optic-mechanical (AOM) method[J]. International Journal of Rock Mechanics and Mining Sciences，2025，187：106041.
[16] 王蒙恩，王小琼，马德龙，等. 基于弹性波速的砂砾岩非均质性评价方法[J]. 岩石力学与工程学报，2026，45(2)：525–536.(WANG Meng?en，WANG Xiaoqiong，MA Delong，et al. An evaluation method for glutenite heterogeneity based on elastic wave velocity[J]. Chinese Journal of Rock Mechanics and Engineering，2026，45(2)：525–536.(in Chinese))
[17] 张俊楷，乔江美，倪祖甲，等. 基于精确矿物晶体模型的玄武岩宏观力学性质及波速分析[J]. 岩石力学与工程学报，2025，44(2)：442–458.(ZHANG Junkai，QIAO Jiangmei，NI Zujia，et al. Investigate the macroscale mechanical property and wave velocity of basalt based on accurate grain-based model[J]. Chinese Journal of Rock Mechanics and Engineering，2025，44(2)：442–458.(in Chinese))
[18] WANG D，TANG Y，SANG Y，et al. Investigation of microcrack activation in CO2 geological storage during shear slip via real-time ultrasonic monitoring[J]. Geomechanics and Geophysics for Geo-Energy and Geo-Resources，2025，https://doi.org/10.1007/s40948–025–00971–1
[19] 贾 蓬，祝鹏程，李 博，等. 单轴压缩过程中岩石的实时超声波特性[J]. 中南大学学报：自然科学版，2022，53(10)：3 967–3 977. (JIA Peng，ZHU Pengcheng，LI Bo，et al. Characteristics of real-time ultrasonic wave during uniaxial compression of rock [J]. Journal of Central South University：Science and Technology，2022，53(10)：3 967–3 977.(in Chinese))
[20] ZOU Y，LI J，ZHAO J. A novel experimental method to investigate the seismic response of rock joints under obliquely incident wave[J]. Rock Mechanics and Rock Engineering，2019，52：3 459–3 466.
[21] 李浩然，杨春和，刘玉刚，等. 单轴荷载作用下盐岩声波与声发射特征试验研究[J]. 岩石力学与工程学报，2014，33(10)：2 107–2 116. (LI Haoran，YANG Chunhe，LIU Yugang，et al. Experimental study of ultrasonic velocity and acoustic emission properties of salt rock under uniaxial compression load [J]. Chinese Journal of Rock Mechanics and Engineering，2014，33(10)：2 107–2 116.(in Chinese))
[22] 来兴平，张 帅，代晶晶，等. 水力耦合作用下煤岩多尺度损伤演化特征[J]. 岩石力学与工程学报，2020，39(增2)：3 217–3 228.(LAI Xingping，ZHANG Shuai，DAI Jingjing，et al. Multi-scale damage evolution characteristics of coal rock under hydraulic mechanical coupling[J]. Chinese Journal of Rock Mechanics and Engineering，2020，39(Supp.2)：3 217–3 228.(in Chinese))
[23] 来兴平，张 帅，崔 峰，等.含水承载煤岩损伤演化过程能量释放规律及关键孕灾声发射信号拾取[J]. 岩石力学与工程学报，2020，39(3)：433–444.(LAI Xingping，ZHANG Shuai，CUI Feng，et al. Energy release law during the damage evolution of water-bearing coal and rock and pick-up of AE signals of key pregnancy disasters[J]. Chinese Journal of Rock Mechanics and Engineering，2020，39(3)：433–444.(in Chinese))
[24] 董陇军，张义涵，孙道元，等. 花岗岩破裂的声发射阶段特征及裂纹不稳定扩展状态识别[J]. 岩石力学与工程学报，2022，41(1)：120–131.(DONG Longjun，ZHANG Yihan，SUN Daoyuan，et al. Stage characteristics of acoustic emission and identification of unstable crack state for granite fractures[J]. Chinese Journal of Rock Mechanics and Engineering，2022，41(1)：120–131.(in Chinese))
[25] 陈进帆，尚德磊，赵志宏，等. 基于机器学习的含充填裂隙岩样断裂行为预测方法[J]. 岩石力学与工程学报，2023，42(增1)：3 458–3 472.(CHEN Jinfan，SHANG Delei，ZHAO Zhihong，et al. Prediction method of fracture behavior of rock samples with an infilled fractures based on machine learning[J]. Chinese Journal of Rock Mechanics and Engineering，2023，42(Supp.1)：3 458–3 472.(in Chinese))
[26] 刘德军，戴庆庆，左建平，等. 基于Stacking集成算法的岩爆等级预测研究[J].岩石力学与工程学报，2022，41(增1)：2 915–2 926. (LIU Dejun，DAI Qingqing，ZUO Jianping，et al. Research on rockburst grade prediction based on stacking integrated algorithm[J]. Chinese Journal of Rock Mechanics and Engineering，2022，41(Supp.1)：2 915–2 926.(in Chinese))
[27] 李晓照，罗秋林，张卓祥，等. 基于机器学习的岩石动态断裂韧度预测方法[J]. 应用力学学报，2026，43(3)：534–544.(LI Xiaozhao，LUO Qiulin，ZHANG Zhuoxiang，et al. A machine learning-based prediction method for dynamic fracture toughness of rocks[J]. Chinese Journal of Applied Mechanics，2026，43(3)：534–544.(in Chinese))
[28] 李典泽，许华杰，张 勃. 基于微震信号深度特征学习的岩石破裂类型识别[J]. 工矿自动化，2025，51(3)：156–164.(LI Dienze，XU Huajie，ZHANG Bo. Rock fracture type recognition based on deep feature learning of microseismic signals[J]. Journal of Mine Automation，2025，51(3)：156–164.(in Chinese))
[29] LUNDBERG S M，LEE S I. A unified approach to interpreting model predictions[C]// Proceedings of the 31st International Conference on Neural Information Processing Systems(NIPS?17). Red Hook：Curran Associates Inc.，2017：4 768–4 777
2026, Vol. 45 
