Rock fracture type recognition based on deep feature learning of microseismic signals
-
摘要:
岩石破裂类型识别是实现煤矿冲击地压灾害预测和预警的重要前提。微震是岩石破裂监测的有效手段之一，但常规的微震信号机器学习方法存在特征提取能力有限，以及受噪声影响导致的准确率不高且泛化性较差等问题。针对上述问题，提出了一种基于微震信号深度特征学习的岩石破裂类型识别方法。首先，通过巴西圆盘劈裂试验和直剪试验分别获取张拉型破裂微震信号和剪切型破裂微震信号，并将其时频谱图、Log−Mel频谱图和梅尔频率倒谱系数合并构造微震信号聚合（MSA）声谱图；然后，通过加入多特征并行密集连接块（MP−DenseBlock）和压缩与激发过渡层（SE−TransLayer）的改进DenseNet（SE−MPDenseNet）对MSA声谱图进行深度特征提取；最后，将提取的特征向量输入至添加Hinge Loss损失函数的改进LightGBM（HBL−LightGBM）进行分类，识别岩石破裂类型。通过真三轴加载试验模拟接近地下工程实际环境中的冲击地压灾害，结果表明，所提方法对于岩石破裂类型识别的准确率达92.12%，且具有较强的特征提取能力和泛化能力。
Abstract:Accurate identification of rock fracture types is crucial for the prediction and early warning of coal mine rockburst hazards. Microseismic monitoring has been widely used for detecting rock fractures. However, conventional machine learning methods for microseismic signal analysis exhibited limited feature extraction capabilities and were highly susceptible to noise, leading to reduced classification accuracy and poor generalization performance. To address these limitations, this study proposed a novel rock fracture type recognition method based on deep feature learning of microseismic signals. In this study, microseismic signals corresponding to tensile and shear fractures were collected through Brazilian disc splitting and direct shear tests, respectively. These signals were then processed to construct a microseismic signal aggregation (MSA) spectrogram, which integrated time-frequency spectrograms, log-Mel spectrograms, and Mel-frequency cepstral coefficients. To enhance feature extraction efficiency, an improved DenseNet model (SE-MPDenseNet) was developed by incorporating multi-feature parallel dense blocks (MP-DenseBlock) and squeeze-and-excitation transition layers (SE-TransLayer). The extracted deep feature vectors were subsequently fed into an optimized LightGBM classifier (HBL-LightGBM), which was modified with a Hinge Loss function to improve classification performance. To evaluate the effectiveness of the proposed method, a true triaxial loading test was conducted to simulate rockburst hazards under realistic underground engineering conditions. Experimental results demonstrated that the proposed approach achieved a rock fracture type recognition accuracy of 92.12%, significantly outperforming conventional methods in both feature extraction capability and generalization ability. The findings indicate that the proposed method provides a robust and effective framework for microseismic-based rock fracture classification. It offers valuable insights for rockburst hazard monitoring and mitigation in mining and geotechnical engineering.
-
-
![]()
图 1 基于微震信号深度特征学习的岩石破裂类型识别方法架构
Figure 1. Schematic of rock fracture type recognition method based on deep feature learning of microseismic signals
![]()
图 4 张拉型及剪切型破裂微震信号时频谱
Figure 4. Time-frequency spectrograms of microseismic signals from tensile and shear fractures
![]()
图 9 真三轴加载试验波形及阶段划分
Figure 9. Waveform and stage division of true-triaxial loading test
![]()
图 10 真三轴加载试验拉剪破裂识别结果
Figure 10. Recognition results for tensile and shear fractures of true-triaxial loading test
![]()
图 11 真三轴加载试验后岩样破坏情况
Figure 11. Post-test failure morphology of rock samples from true-triaxial loading test
表 1 岩石破裂类型识别结果
Table 1 Recognition results for different rock fracture types
% 类别 准确率 召回率 F1分数 总体准确率 剪切 90.34 87.82 88.87 92.12 张拉 93.90 92.51 91.06 
下载: 导出CSV
表 2 不同改进策略对识别准确率的影响
Table 2 Effect of different model enhancement strategies on recognition accuracy
特征提取器 分类器 准确率/% DenseNet HBL−LightGBM 90.64 DenseNet+MP−DenseBlock HBL−LightGBM 91.78 DenseNet+SE−TransLayer HBL−LightGBM 90.97 SE−MPDenseNet LightGBM 91.20 SE−MPDenseNet HBL−LightGBM 92.12
下载: 导出CSV
-
[1] FENG Xiating,CHEN Bingrui,LI Shaojun,et al. Studies on the evolution process of rockbursts in deep tunnels[J]. Journal of Rock Mechanics and Geotechnical Engineering,2012,4(4):289-295. DOI: 10.3724/SP.J.1235.2012.00289
[2] SU Guoshao,JIANG Jianqing,ZHAI Shaobin,et al. Influence of tunnel axis stress on strainburst:an experimental study[J]. Rock Mechanics and Rock Engineering,2017,50(6):1551-1567. DOI: 10.1007/s00603-017-1181-7
[3] SU Guoshao,SHI Yanjiong,FENG Xiating,et al. True-triaxial experimental study of the evolutionary features of the acoustic emissions and sounds of rockburst processes[J]. Rock Mechanics and Rock Engineering,2018,51(2):375-389. DOI: 10.1007/s00603-017-1344-6
[4] 赵国富,苏国韶,胡诗红,等. 一种基于微震试验的硬岩破裂模式识别方法[J]. 实验力学,2022,37(1):107-117. ZHAO Guofu,SU Guoshao,HU Shihong,et al. A recognition method of cracking types for hard rock based on microseisms tests[J]. Journal of Experimental Mechanics,2022,37(1):107-117.
[5] 李德春,葛宝堂,舒继森. 岩体破坏过程中的电阻率变化试验[J]. 中国矿业大学学报,1999,28(5):80-82. LI Dechun,GE Baotang,SHU Jisen. Experiment of resistivity variation of rocks in failure process[J]. Journal of China University of Mining & Technology,1999,28(5):80-82.
[6] 李术才,许新骥,刘征宇,等. 单轴压缩条件下砂岩破坏全过程电阻率与声发射响应特征及损伤演化[J]. 岩石力学与工程学报,2014,33(1):14-23. LI Shucai,XU Xinji,LIU Zhengyu,et al. Electrical resistivity and acoustic emission response characteristics and damage evolution of sandstone during whole process of uniaxial compression[J]. Chinese Journal of Rock Mechanics and Engineering,2014,33(1):14-23.
[7] 李忠辉,王恩元,何学秋,等. 含水量对煤岩电磁辐射特征的影响[J]. 中国矿业大学学报,2006,35(3):362-366. LI Zhonghui,WANG Enyuan,HE Xueqiu,et al. Effect of moisture content on electromagnetic radiation characteristic of coal or rock[J]. Journal of China University of Mining & Technology,2006,35(3):362-366.
[8] 李夕兵,万国香,周子龙. 岩石破裂电磁辐射频率与岩石属性参数的关系[J]. 地球物理学报,2009,52(1):253-259. LI Xibing,WAN Guoxiang,ZHOU Zilong. The relation between the frequency of electromagnetic radiation (EMR) induced by rock fracture and attribute parameters of rock masses[J]. Chinese Journal of Geophysics,2009,52(1):253-259.
[9] 高煜,胡宾鑫,朱峰,等. 微震初至波到时自动拾取研究[J]. 工矿自动化,2020,46(12):106-110. GAO Yu,HU Binxin,ZHU Feng,et al. Research on automatic picking of microseismic first arrival[J]. Industry and Mine Automation,2020,46(12):106-110.
[10] 何风贞,李桂臣,阚甲广,等. 岩石多尺度损伤研究进展[J]. 煤炭科学技术,2024,52(10):33-53. HE Fengzhen,LI Guichen,KAN Jiaguang,et al. Research progress on multi-scale damage of rock[J]. Coal Science and Technology,2024,52(10):33-53.
[11] 张凯,张东晓,赵勇强,等. 损伤岩石声发射演化特征及响应机制试验研究[J]. 煤田地质与勘探,2024,52(3):96-106. ZHANG Kai,ZHANG Dongxiao,ZHAO Yongqiang,et al. Experimental study on acoustic emission evolution characteristics and response mechanism of damaged rocks[J]. Coal Geology & Exploration,2024,52(3):96-106.
[12] 张泽坤,宋战平,程昀,等. 加载速率影响下类硬岩声发射及破裂响应特征[J]. 煤田地质与勘探,2022,50(2):115-124. ZHANG Zekun,SONG Zhanping,CHENG Yun,et al. Acoustic emission characteristics and fracture response behavior of hard rock-like material under influence of loading rate[J]. Coal Geology & Exploration,2022,50(2):115-124.
[13] 朱权洁,张尔辉,李青松,等. 岩石破坏失稳的声发射响应与损伤定量表征研究[J]. 中国安全生产科学技术,2020,16(1):92-98. ZHU Quanjie,ZHANG Erhui,LI Qingsong,et al. Study on acoustic emission response and damage quantitative characterization of rock destruction and instability[J]. Journal of Safety Science and Technology,2020,16(1):92-98.
[14] 李炜强,许沁舒,成功,等. 单轴压缩下砂岩微破裂演化力学行为研究[J]. 煤炭科学技术,2020,48(11):60-67. LI Weiqiang,XU Qinshu,CHENG Gong,et al. Study on mechanical behavior of sandstone micro-fracture evolution under uniaxial compression test[J]. Coal Science and Technology,2020,48(11):60-67.
[15] 刘健,王晓军,徐莎莎,等. 基于声发射RA−AF值识别不同岩爆倾向性灰岩破裂特征[J]. 金属矿山,2022(10):16-23. LIU Jian,WANG Xiaojun,XU Shasha,et al. Identification of limestone fracture characteristics with different rockburst propensities based on acoustic emission RA-AF values[J]. Metal Mine,2022(10):16-23.
[16] 纪洪广,张春瑞,张月征,等. 岩石材料破裂过程中声发射信号的应力状态及能量演化研究[J]. 中国矿业大学学报,2024,53(2):211-223. JI Hongguang,ZHANG Chunrui,ZHANG Yuezheng,et al. Research on stress state and energy evolution of acoustic emission signal during rock materials fracture process[J]. Journal of China University of Mining & Technology,2024,53(2):211-223.
[17] 陈炳瑞,冯夏庭,符启卿,等. 综合集成高精度智能微震监测技术及其在深部岩石工程中的应用[J]. 岩土力学,2020,41(7):2422-2431. CHEN Bingrui,FENG Xiating,FU Qiqing,et al. Integration and high precision intelligence microseismic monitoring technology and its application in deep rock engineering[J]. Rock and Soil Mechanics,2020,41(7):2422-2431.
[18] 黄杰,苏国韶,王拯扶,等. 基于声谱图特征的硬岩破裂类型识别方法[J]. 人民长江,2021,52(8):198-203. HUANG Jie,SU Guoshao,WANG Zhengfu,et al. Recognition method of hard rock cracking types based on spectrograms characteristics[J]. Yangtze River,2021,52(8):198-203.
[19] CHANG Zhenghao,HE R,YU Yongsheng,et al. A two-stream convolution architecture for ESC based on audio feature distanglement[C]. Asian Conference on Machine Learning,New York,2023:153-168.
[20] HUANG Gao,LIU Zhuang,VAN DER MAATEN L,et al. Densely connected convolutional networks[C]. IEEE Conference on Computer Vision and Pattern Recognition,Honolulu,2017:2261-2269.
[21] HU Jie,SHEN Li,SUN Gang. Squeeze-and-excitation networks[C]. IEEE/CVF Conference on Computer Vision and Pattern Recognition,Salt Lake City,2018:7132-7141.
[22] YANG Huazhong,CHEN Zhongju,YANG Huajian,et al. Predicting coronary heart disease using an improved LightGBM model:performance analysis and comparison[J]. IEEE Access,2023,11:23366-23380. DOI: 10.1109/ACCESS.2023.3253885
[23] WANG Huajun,LI Genghui,WANG Zhenkun. Fast SVM classifier for large-scale classification problems[J]. Information Sciences,2023,642. DOI: 10.1016/j.ins.2023.119136.
[24] WU Di,FAN Zheyi,YI Shuhan. Crowd counting based on multi-level multi-scale feature[J]. Applied Intelligence,2023,53(19):21891-21901. DOI: 10.1007/s10489-023-04641-1
计量
- 文章访问数: 22
- HTML全文浏览量: 3
- PDF下载量: 4
