戴风华,史经峰,王帅帅,高轩,汪久洋,于礼杰

• 1:中交第二公路工程局有限公司
• 2:山东大学岩土与地下工程研究院

摘要(Abstract)：

智能凿岩台车随钻参数中的钻孔速度是当前反映隧道掘进掌子面前方围岩性质最为科学且数据易于获取的指标。为定量描述钻孔速度、爆破参数与围岩超欠挖之间的关系，依托某高原隧道智能凿岩台车施工工程，系统采集84个爆破循环、共3732条周边孔的完整MWD数据。基于对数据组成与爆破影响因素的梳理，构建了包含领钻阶段识别、正常钻进阶段钻孔速度聚类、装药结构向量化与降维、几何约束及最小抵抗线计算等环节的处理流程，提取周边孔外插角、孔距、最小抵抗线、装药结构与钻孔速度等特征，建立了基于MWD的Light GBM超欠挖预测模型。模型以设计原因导致的超欠挖为输出，结合早停与贝叶斯超参数优化保证收敛与泛化。结果表明:模型训练收敛性良好，验证/测试阶段整体误差处于工程可用水平，能够较好地刻画超欠挖随最小抵抗线、装药结构及钻孔速度等因素的变化趋势;在此基础上开展现场应用验证，平均线性超挖由24.8 cm降至13.2 cm，半孔率由53%提升至79%，轮廓控制与爆破质量得到改善。

关键词(KeyWords)： 钻爆法;超欠挖;随钻参数;钻孔速度;机器学习;LightGBM

Abstract:

Keywords:

基金项目(Foundation): 山东大学高端工程机械智能制造全国重点实验室开放基金项目（ACMKF2024-07）~~

作者(Author): 戴风华,史经峰,王帅帅,高轩,汪久洋,于礼杰

参考文献(References)：

• [1]龚伟毅，姚颖康，杜宇翔．中等断面隧道长进尺直孔掏槽爆破开挖与超欠挖控制现场试验[J]．爆破，2024,41(2):32-39.[1]GONG Wei-yi,YAO Ying-kang,DU Yu-xiang. Field test of long footage burn cut blasting excavation and control of overcut and undercut in medium section tunnel[J]. Blasting,2024,41(2):32-39.(in Chinese)
• [2]ZHANG Wan-mao,LIU Dun-wen,TANG Yu,et al. Multifractal characteristics of smooth blasting overbreak in extra-long hard rock tunnel[J]. Fractal Fract,2023,7:842.https:∥doi. org/10. 3390/fractalfract7120842.
• [3]WANG Hao-teng,HE Ming-ming. Determining method of tensile strength of rock based on friction characteristics in the drilling process[J]. Rock Mech Rock Eng,2023,56:4211-4227. https:∥doi. org/10. 1007/s00603-023-03276-5.
• [4]WANG Shao-feng,WU Yu-meng,CAI Xin,et al. Strength prediction and drillability identification for rock based on measurement while drilling parameters[J]. J Cent South Univ,2023,30:4036-4051. https:∥doi. org/10. 1007/s11771-023-5492-4.
• [5]LAKSHMINARAYANA C R,TRIPATHI A K,PAL S K.Experimental investigation on potential use of drilling parameters to quantify rock strength[J]. Geo-Engineering,2021,12:23. https:∥doi. org/10. 1186/s40703-021-00152-5.
• [6]LIU Can-can,ZHENG Xi-gui,SHAHANI N M,et al. An experimental investigation into the borehole drilling and strata characteristics[J]. PLo S ONE,2021,16(7):e0253663.
• [7]张树才，仇文革，张齐芳，等．基于凿岩台车钻进参数的岩石强度预测模型研究[J]．隧道建设(中英文)，2023,43(12):2007-2017.[7]ZHANG Shu-cai,QIU Wen-ge,ZHANG Qi-fang,et al. Research on rock strength prediction model based on drilling parameters of drilling jumbo[J]. Tunnel construction(English and Chinese),2023,43(2):2007-2017.(in Chinese)
• [8]易文豪，王明年，童建军，等．基于支持向量机的大断面岩质隧道掌子面围岩非均一性判识方法[J]．中国铁道科学，2021,42(5):112-122.[8]YI Wen-hao,WANG Ming-nian,TONG Jian-jun,et al.Method for identifying the heterogeneity of surrounding rock of large-section rock tunnel face based on support vector machine[J]. China Railway Science,2021,42(5):112-122.(in Chinese)
• [9]王明年，赵思光，童建军，等．基于炮孔钻进参数的隧道掌子面围岩三维精细化分级方法[J]．铁道学报，2024,46(10):163-173.[9]WANG Ming-nian,ZHAO Si-guang,TONG Jian-jun,et al. A three-dimensional refined classification method for surrounding rock of tunnel face based on borehole drilling parameters[J]. Journal of the China Railway Society,2024,46(10),163-173.(in Chinese)
• [10]HANSEN T F,LIU Zhong-qiang,TORRESEN J. Predicting rock type from MWD tunnel data using a reproducible ML-modelling process[J]. Tunnelling and Underground Space Technology,2024,152:105843. https:∥doi. org/10. 1016/j. tust. 2024. 105843.
• [11]薛翊国，孔凡猛，杨为民，等．川藏铁路沿线主要不良地质条件与工程地质问题[J]．岩石力学与工程学报，2020,39(3):445-468.[11]XUE Yi-guo,KONG Fan-meng,YANG Wei-min,et al.Major adverse geological conditions and engineering geological problems along the Sichuan-Tibet Railway[J].Journal of Rock Mechanics and Engineering,2020,39(3):445-468.(in Chinese)
• [12]RYBAKOVA E O,LIMONOVA E E,NIKOLAEV D P.Fast gaussian filter approximations comparison on SIMD computing platforms[J]. Applied Sciences,2024,14(11):4664. https:∥doi. org/10. 3390/app14114664.
• [13]BELHAOUARI S B. Unsupervised outlier detection in multidimensional data[J]. J Big Data,2021,8:80. https:∥doi. org/10. 1186/s40537-021-00469-z.
• [14]KAREN K,SPIEGELMAN C H. An alternative to ordinary q-q plots:Conditional q-q plots[J]. Computational Statistics&Data Analysis,1986,4(3):167-184. https:∥doi. org/10. 1016/0167-9473(86)90032-0.
• [15]FRALEY C,RAFTERY A E. Model-based clustering,discriminant analysis,and density estimation[J]. Journal of the American Statistical Association,2002,97(458):611-631. https:∥doi. org/10. 1198/016214502760047 131.
• [16]ANDREI A T,GRIGORE O. Low-cost optimized U-Net model with GMM automatic labeling used in forest semantic segmentation[J]. Sensors,2023,23:8991. https:∥doi. org/10. 3390/s23218991.
• [17]MURTAGH F,CONTRERAS P. Algorithms for hierarchical clustering:an overview[J]. WIREs Data Mining Knowl Discov,2012,2:86-97. https:∥doi. org/10.1002/widm. 53.
• [18]赵茉溪，杨玉民，周传波，等．基于MD-PCA-BP模型的露天矿山爆破振动速度预测[J]．爆破，2024,41(2):203-211.[18]ZHAO Mo-xi,YANG Yu-min,ZHOU Chuan-bo,et al.Prediction of blasting vibration velocity in open-pit mine based on MD-PCA-BP model[J]. Blasting,2024,41(2):203-211.(in Chinese)
• [19]HE Biao,ARMAGHANI D J,LAI Sai-hin. Assessment of tunnel blasting-induced overbreak:A novel metaheuristic-based random forest approach[J]. Tunnelling and Underground Space Technology,2023,133:104979. https:∥doi. org/10. 1016/j. tust. 2022. 104979.
• [20]LIU Yao-sheng,LI Ang,DAI Feng,et al. An AI-powered approach to improving tunnel blast performance considering geological conditions[J]. Tunnelling and Underground Space Technology,2024,144:105508. https:∥doi. org/10. 1016/j. tust. 2023. 105508.