地质不确定性驱动的 TBM 智能掘进技术体系构建
薛亚东 , 魏飞 , 池建军 , 杨晋文 , 郑志强
有色金属(矿山部分) ›› 2026, Vol. 78 ›› Issue (6) : 25 -35.
地质不确定性驱动的 TBM 智能掘进技术体系构建
Study on TBM intelligent tunneling technology system driven by geological uncertainty
针对深部金属矿山地质条件不确定性强、传统围岩分级方法难以满足TBM 实时掘进决策需求的工程难题,构建了“破岩机理认知- 围岩实时感知- 围岩智能预测- 掘进参数优化”的四层智能掘进技术体系,系统阐述了各层级的核心方法与内在逻辑。在破岩机理层面,基于滑移线场理论揭示滚刀破岩三区演化规律与模式转变机制,引入安装半径因子与协同效应因子改进Colorado School of Mines(CSM)模型,建立了破岩比能与最优贯入度的确定方法;在围岩感知层面,提出可掘进性与适应性二维综合分级方法,构建了自适应分段常数近似(APCA)时序分割与堆叠稀疏自编码器- 模糊C 均值(SSAE-FCM)深度聚类相结合的无监督实时围岩分类模型;在围岩预测层面,建立知识与数据双驱动特征工程框架,开发基于树结构Parzen 估计器-LightGBM(TPE-LightGBM)的投票预测模型,并通过联合训练提升跨工程泛化能力;在参数优化层面,提出CSM 简化模型与破岩模式分区方法,形成多模式掘进参数优化策略。工程验证表明,该体系可有效应对地质不确定性,实现TBM 掘进效率与安全性的协同提升,可为深部金属矿山TBM 智能施工提供技术参考。
To address the engineering challenges of strong geological uncertainty in deep metal mines and the inadequacy of traditional surrounding rock classification methods for real-time TBM excavation decision-making,a four-layer intelligent excavation technology system comprising "rock fragmentation mechanism cognition-real-time surrounding rock perception-intelligent surrounding rock prediction-excavation parameter optimization" is constructed,with the core methods and intrinsic logic of each layer systematically elaborated. At the rock fragmentation mechanism level,the slip-line field theory is employed to reveal the three-zone evolution law and mode transition mechanism of disc cutter rock fragmentation;the Colorado School of Mines (CSM) model is improved by introducing an installation radius factor and a synergistic effect factor;and a method for determining the specific energy of rock fragmentation and the optimal penetration depth is established. At the surrounding rock perception level,a two-dimensional comprehensive classification method based on boreability and adaptability is proposed,and an unsupervised real-time surrounding rock classification model integrating Adaptive Piecewise Constant Approximation (APCA) time series segmentation with Stacked Sparse Autoencoder-Fuzzy C-Means (SSAE-FCM) deep clustering is constructed. At the surrounding rock prediction level,a knowledge-and data-driven dual feature engineering framework is established,a voting prediction model based on Tree-structured Parzen Estimator-LightGBM is developed,and the cross-project generalization capability is enhanced through joint training. At the parameter optimization level,a simplified CSM model and a rock fragmentation mode zoning method are proposed,forming a multi-mode excavation parameter optimization strategy. Engineering verification demonstrates that the proposed system can effectively cope with geological uncertainty,achieve synergistic improvement of TBM excavation efficiency and safety,and provide technical reference for intelligent TBM construction in deep metal mines.
| [1] |
刘泉声,潘玉丛,余宏淦, |
| [2] |
|
| [3] |
|
| [4] |
|
| [5] |
|
| [6] |
|
| [7] |
|
| [8] |
|
| [9] |
|
| [10] |
|
| [11] |
Armaghani D J,Mohamad E T,Narayanasamy M S, |
| [12] |
|
| [13] |
|
| [14] |
|
| [15] |
|
| [16] |
|
| [17] |
|
| [18] |
|
| [19] |
Barton N. TBM tunnelling in difficult ground[J]. Tunnelling and Underground Space Technology,2000,15(4):391-408. |
| [20] |
霍军周,赵军,谭青, |
| [21] |
Huo Junzhou,Zhao Jun,Tan Qing,et al. Research on design theory and method of TBM disc cutter spacing[J]. Journal of Mechanical Engineering,2015,51(9):162—168. |
| [22] |
Teale R. The concept of specific energy in rock drilling[J]. International Journal of Rock Mechanics and Mining Sciences & Geomechanics Abstracts,1965,2(1):57-73. |
| [23] |
|
| [24] |
|
| [25] |
薛亚东,李兴,刁振兴, |
| [26] |
|
| [27] |
|
| [28] |
|
| [29] |
|
国家自然科学基金资助项目(52478412)
新疆水利发展投资集团有限公司科研项目资助(ABHB-20551030)
/
| 〈 |
|
〉 |