张骞,聂瑶奇,杜立杰

• 1:石家庄铁道大学安全工程与应急管理学院
• 2:石家庄铁道大学河北省大型结构健康诊断与控制重点实验室

摘要(Abstract)：

针对隧道掘进机(TBM)围岩级别识别中样本分布不均、模型精度有限以及可解释性不足等问题，提出一种融合SMOTE过采样、贝叶斯优化和加权集成学习的围岩级别智能识别方法。首先，利用SMOTE算法对少数类样本进行扩充，缓解数据不平衡对分类性能的不利影响，提升模型对不同围岩级别样本的识别能力。其次，采用贝叶斯优化分别对XGBoost和LightGBM模型的关键超参数进行自适应寻优，在提高模型泛化能力的基础上，构建BO-XGBoost-LightGBM加权融合模型，实现多模型优势互补。最后，引入SHAP方法对模型预测结果进行可解释性分析，揭示关键掘进参数对围岩级别判别的影响规律，并验证模型输出与工程实际之间的一致性。以实际工程中TBM施工数据为例进行验证，结果表明：所提方法在准确率、精确率、召回率和F1分数等指标上均优于Naive Bayes、Decision Tree、Random Forest、XGBoost和LightGBM等对比模型，表现出更高的识别精度、稳定性和鲁棒性。研究结果表明，所提出的方法不仅能够实现复杂地质条件下TBM围岩级别的高效识别，而且模型判别机制透明可溯，与现场经验一致，可为TBM智能掘进、施工参数优化及地质风险预警提供有效支撑。

关键词(KeyWords)： TBM;围岩级别;贝叶斯优化;加权策略;XGBoost;LightGBM;SHAP

Abstract:

Keywords:

基金项目(Foundation): 河北省自然科学基金(E2025210083);; 国家自然科学基金(52578482)

作者(Author): 张骞,聂瑶奇,杜立杰

参考文献(References)：

• [1] 洪开荣.超长深埋高地应力TBM隧道修建关键技术[J].铁道学报，2022,44(3):1-23.HONG Kairong.Key technology for construction of ultra-long and deep-buried TBM tunnels with high geostress[J].Journal of the China Railway Society,2022,44(3):1-23.
• [2] Nie Y Q,Zhang Q,Du Y L,et al.TBM jamming risk prediction method based on fuzzy theory and Bi-LSTM[J].Tunnelling and Underground Space Technology,2025,157:106329.
• [3] Zhang Q,Nie Y Q,Zhao W G,et al.Research on TBM parameter optimization based on failure probability[J].Engineering Failure Analysis,2025,167:109036.
• [4] Gong Q M,Lu J W,Xu H Y,et al.A modified rock mass classification system for TBM tunnels and tunneling based on the HC method of China[J].International Journal of Rock Mechanics and Mining Sciences,2021,137:104551.
• [5] Huang J T,Ji A K,Zhang L M,et al.Data-driven optimization for enhanced excavation efficiency in tunnel construction:a case study[J].Engineering Applications of Artificial Intelligence,2025,142:109868.
• [6] Lin P H,Zhang L M,Tiong R L K,et al.Physics-guided deep reinforcement learning for improved operation of TBM considering kinetic inertia stability[J].Advanced Engineering Informatics,2025,63:102943.
• [7] Sun R X,Mo J L,Zhang M Q,et al.Cutting performance and contact behavior of partial-wear TBM disc cutters:a laboratory scale investigation[J].Engineering Failure Analysis,2022,137:106253.
• [8] Sun R X,Mo J L,Zhang M Q,et al.Interaction between partial-worn TBM cutters and rocks:experimental and numerical investigation[J].Bulletin of Engineering Geology and the Environment,2023,82(4):111.
• [9] Lu J W,Gong Q M,Wei J Z,et al.Determination of rockburst reduction coefficient βA in the modified rock mass classification system for TBM tunnels and tunneling[J].International Journal of Rock Mechanics and Mining Sciences,2024,174:105657.
• [10] 丁自伟，高成登，张玲，等.基于数据驱动的TBM掘进地层岩性识别预测方法[J].采矿与安全工程学报，2025,42(1):147-160.Ding Ziwei,Gao Chengdeng,Zhang Ling,et al.Prediction method for data-driven lithology identification of TBM tunneling strata[J].Journal of Mining and Safety Engineering,2025,42(1):147-160.
• [11] Fu K,Qiu D H,Xue Y G,et al.TBM tunneling strata automatic identification and working conditions decision support[J].Automation in Construction,2024,163:105425.
• [12] Chen L,Liu Z T,Su H Y,et al.Automated rock mass condition assessment during TBM tunnel excavation using deep learning[J].Scientific Reports,2022,12:1722.
• [13] Gong Q M,Xu H Y,Lu J W,et al.Rock mass characteristics model for TBM penetration rate prediction-an updated version[J].International Journal of Rock Mechanics and Mining Sciences,2022,149:104993.
• [14] Zhou J,Qiu Y G,Zhu S L,et al.Optimization of support vector machine through the use of metaheuristic algorithms in forecasting TBM advance rate[J].Engineering Applications of Artificial Intelligence,2021,97:104015.
• [15] Shi C,Wang Y.Data-driven sequential development of geological cross-sections along tunnel trajectory[J].Acta Geotechnica,2023,18(4):1739-1754.
• [16] Guan Z,Wang Y.CPT-based probabilistic liquefaction assessment considering soil spatial variability,interpolation uncertainty and model uncertainty [J].Computers and Geotechnics,2022,141:104504.
• [17] Yang H Q,Wang H,Zhou X P.Analysis on the rock-cutter interaction mechanism during the TBM tunneling process[J].Rock Mechanics and Rock Engineering,2016,49(3):1073-1090.
• [18] Zhang Z X,Wang S F,Huang X,et al.TBM-block interaction during TBM tunneling in rock masses:block classification and identification[J].International Journal of Geomechanics,2017,17(5):E4016001.
• [19] 赵思光，王明年，童建军，等.基于模型分级概率的围岩亚级智能分级方法研究[J].铁道学报，2025,47(11):222-231.Zhao Siguang,Wang Mingnian,Tong Jianjun,et al.Rock based on model classification probabilities research on method for intelligent sub-grade classification of surrounding[J].Journal of the China Railway Society,2025,47(11):222-231.
• [20] Salimi A,Rostami J,Moormann C.Evaluating the suitability of existing rock mass classification systems for TBM performance prediction by using a regression tree[J].Procedia Engineering,2017,191:299-309.
• [21] Hou S K,Liu Y R,Yang Q.Real-time prediction of rock mass classification based on TBM operation big data and stacking technique of ensemble learning [J].Journal of Rock Mechanics and Geotechnical Engineering,2022,14(1):123-143.
• [22] Liu Q S,Wang X Y,Huang X,et al.Prediction model of rock mass class using classification and regression tree integrated AdaBoost algorithm based on TBM driving data [J].Tunnelling and Underground Space Technology,2020,106:103595.
• [23] 朱梦琦，朱合华，王昕，等.基于集成CART算法的TBM掘进参数与围岩等级预测[J].岩石力学与工程学报，2020,39(9):1860-1871.Zhu Mengqi,Zhu Hehua,Wang Xin,et al.Study on CART-based ensemble learning algorithms for predicting TBM tunneling parameters and classing surrounding rockmasses[J].Chinese Journal of Rock Mechanics and Engineering,2020,39(9):1860-1871.
• [24] Chen C,Seo H.Prediction of rock mass class ahead of TBM excavation face by ML and DL algorithms with Bayesian TPE optimization and SHAP feature analysis[J].Acta Geotechnica,2023,18(7):3825-3848.
• [25] 陈祖煜，范立涛，张云旆，等.知识驱动和数据驱动在TBM智能施工机器学习中的应用[J].土木工程学报，2024,57(6):1-12.Chen Zuyu,Fan Litao,Zhang Yunpei,et al.Knowledge-based and data-based machine learning in intelligent TBM construction[J].China Civil Engineering Journal,2024,57(6):1-12.
• [26] 李建斌，郑赢豪，荆留杰，等.基于岩体聚类分级的TBM掘进参数预测方法[J].岩石力学与工程学报，2020,39(S2):3326-3337.Li Jianbin,Zheng Yinghao,Jing Liujie,et al.Prediction method of TBM driving parameters based on rock mass clustering and classification[J].Chinese Journal of Rock Mechanics and Engineering,2020,39(S2):3326-3337.
• [27] 李宗林，谭忠盛，周振梁，等.基于TBM振动监测的围岩参数感知技术研究[J].铁道科学与工程学报，2025,22(4):1853-1869.Li Zonglin,Tan Zhongsheng,Zhou Zhenliang,et al.Research on surrounding rock parameter perception technology based on TBM vibration monitoring[J].Journal of Railway Science and Engineering,2025,22(4):1853-1869.
• [28] 吴志军，方立群，翁磊，等.基于TBM掘进性能的岩体分级及可掘性等级感知识别方法[J].岩石力学与工程学报，2022,41(S1):2684-2699.Wu Zhijun,Fang Liqun,Weng Lei,et al.Perception and identification method of rock mass classification and drivability grade based on TBM tunneling performance[J].Chinese Journal of Rock Mechanics and Engineering,2022,41(S1):2684-2699.
• [29] 曹晋镨，刘芳，申志福.基于长短期记忆网络的TBM掘进预测模型及围岩等级对预测精度的影响[J].土木工程学报，2022,55(S2):92-102.Cao Jinpu,Liu Fang,Shen Zhifu.A LSTM-based model for TBM performance prediction and the effect of rock mass grade on prediction accuracy[J].China Civil Engineering Journal,2022,55(S2):92-102.
• [30] 满轲，曹子祥，刘晓丽，等.基于GRU-RF模型的TBM掘进参数预测研究[J].应用基础与工程科学学报，2023,31(6):1519-1539.Man Ke,Cao Zixiang,Liu Xiaoli,et al.Prediction of TBM tunneling parameters based on a GRU-RF model[J].Journal of Basic Science and Engineering,2023,31(6):1519-1539.
• [31] Chen T Q,Guestrin C.XGBoost:a scalable tree boosting system[C]//Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining.New York:ACM,2016:785-794.
• [32] Ke G,Meng Q,Finley T,et al.LightGBM:a highly efficient gradient boosting decision tree[C]//Advances in Neural Information Processing Systems 30.Long Beach:Curran Associates,2017:3146-3154.
• [33] Nie Y Q,Zhang Q,Hou L L,et al.TBM rock mass classification using XGBoost and interpretable machine learning[J].Advanced Engineering Informatics,2025,66:103459.
• [34] Chawla N V,Bowyer K W,Hall L O,et al.SMOTE:synthetic minority over-sampling technique[J].Journal of Artificial Intelligence Research,2002,16:321-357.
• [35] Rodriguez J D,Perez A,Lozano J A.Sensitivity analysis of k-fold cross validation in prediction error estimation[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,2010,32(3):569-575.
• [36] 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.New York:ACM,2017:4768-4777.