复杂地质隧洞施工突涌水的集成学习预测与稳定性优化

doi:10.3724/1000-6915.jrme.2026.0046

Abstract
Figure/Table
References (0)
Related Citation (15)

Download: PDF (1599 KB) HTML (1 KB)
Export: BibTeX | EndNote (RIS)

Abstract To address the low accuracy of tunnel water-inrush prediction under complex geological conditions and the difficulty in capturing the nonlinear relationships among disaster-inducing factors, an intelligent prediction framework driven by ensemble learning is proposed. Based on a review of 249 documented water-inrush disaster cases in China and abroad, 22 disaster-inducing factors were selected. The synthetic minority over-sampling technique (SMOTE) was employed to overcome the class imbalance of the minority-class data, and principal component analysis (PCA) combined with K-means clustering was used to establish a hazard assessment system. A stacked generalization (SG) strategy was then adopted to integrate a random forest (RF) model and a backpropagation neural network (BPNN) to construct the prediction model, and the analytic hierarchy process (AHP) was further introduced to identify the key disaster-inducing factors. The results show that: (1) By exploiting the complementary error characteristics between the base learners, the SG model achieves a coefficient of determination (R2) of 0.98, a mean absolute error (MAE) of 0.68 and a mean square error (MSE) of 0.87, and its overall prediction performance and generalization ability are superior to those of a single base learner. (2) For major water-inrush cases with water-inflow volumes exceeding 300 m3/(d•m), the relative prediction error of the SG model remains stably within 10%, which effectively reduces the prediction fluctuations of a single base learner in the high-flow range. (3) Based on the initial weights obtained from the RF model and the correction made by engineering judgment, the AHP identifies the hydraulic conductivity, adverse geological conditions, geological structure, negative landform and groundwater occurrence conditions as the five core disaster-inducing factors, with a cumulative weight of more than 48.56%. These findings provide a relatively reliable technical approach for water-inrush disaster prevention and early warning in underground engineering.

Key words： tunnel engineering machine learning water inrush disaster random forest (RF) BP neural network (BPNN) stacked generalization (SG) K-means clustering

	Service

	E-mail this article
	Add to my bookshelf
	Add to citation manager
	E-mail Alert
	RSS
	Articles by authors
	CHEN Zhou1
	2*
	WANG Kai1
	2
	CHEN Zihao1
	2
	WANG Jinguo1
	2
	ZONG Chenyu1
	2
	DOU Zhi1
	2
	ZHU Chun1
	2

Cite this article:

CHEN Zhou1,2*,WANG Kai1, et al. Ensemble learning-based prediction and stability enhancement of water inrush during tunnel construction under complex geological conditions[J]. , 2026, 45(11): 3403-3419.

URL:

https://rockmech.whrsm.ac.cn/EN/10.3724/1000-6915.jrme.2026.0046 OR https://rockmech.whrsm.ac.cn/EN/Y2026/V45/I11/3403

[1] 《中国公路学报》编辑部. 中国交通隧道工程学术研究综述·2022[J]. 中国公路学报，2022，35(4)：1–40.(Editorial Department of China Journal of Highway and Transport. Review on China¢s traffic tunnel engineering research：2022[J]. China Journal of Highway and Transport，2022，35(4)：1–40.(in Chinese))

[2] 张庆松，李术才，韩宏伟，等. 岩溶隧道施工风险评价与突水灾害防治技术研究[J]. 山东大学学报：工学版，2009，39(3)：106–110.(ZHANG Qingsong，LI Shucai，HAN Hongwei，et al. Study on risk evaluation and water inrush disaster preventing technology during construction of karst tunnels[J]. Journal of Shandong University：Engineering Science，2009，39(3)：106–110.(in Chinese))

[3] 李术才，许振浩，黄鑫，等. 隧道突水突泥致灾构造分类、地质判识、孕灾模式与典型案例分析[J]. 岩石力学与工程学报，2018，37(5)：1 041–1 069.(LI Shucai，XU Zhenhao，HUANG Xin，et al. Classification，geological identification，hazard mode and typical case studies of hazard-causing structures for water and mud inrush in tunnels[J]. Chinese Journal of Rock Mechanics and Engineering，2018，37(5)：1 041–1 069.(in Chinese))

[4] AALIANVARI A，KATIBEH H，SHARIFZADEH M. Application of fuzzy Delphi AHP method for the estimation and classification of Ghomrud tunnel from groundwater flow hazard[J]. Arabian Journal of Geosciences，2012，5(2)：275–284.

[5] 周宗青，李术才，李利平，等. 岩溶隧道突涌水危险性评价的属性识别模型及其工程应用[J]. 岩土力学，2013，34(3)：818–826.(ZHOU Zongqing，LI Shucai，LI Liping，et al. Attribute recognition model of fatalness assessment of water inrush in karst tunnels and its application[J]. Rock and Soil Mechanics，2013，34(3)：818–826.(in Chinese))

[6] MAHMOODZADEH A，MOHAMMADI M，NOORI K M G，et al. Presenting the best prediction model of water inflow into drill and blast tunnels among several machine learning techniques[J]. Automation in Construction，2021，127：103719.

[7] 刘再斌. 基于数据挖掘分类技术的煤层底板突水预测[硕士学位论文][D]. 北京：煤炭科学研究总院，2008.(LIU Zaibin. Prediction of water inrush from coal seam floor based on data mining classification technology[M. S. Thesis][D]. Beijing：China Coal Research Institute，2008.(in Chinese))

[8] 李付法. 锦屏水电站辅助洞突水、突泥机理及预测预报研究[硕士学位论文][D]. 成都：西南交通大学，2006.(LI Fufa. Study on mechanism and forecast of water outbrush and mud outbrush of Jin-ping tunnel[M. S. Thesis][D]. Chengdu：Southwest Jiaotong University，2006.(in Chinese))

[9] 李新平，瞿江文，唐结齐，等. 基于组合赋权法–TOPSIS法的北天山隧道突水风险评价研究[J]. 水利水电技术，2019，50(9)：114–119.(LI Xinping，QU Jiangwen，TANG Jieqi，et al. Study on combination weighting method-TOPSIS method-based risk assessment of water inrush in construction of North Tianshan Mountain Tunnel[J]. Water Resources and Hydropower Engineering，2019，50(9)：114–119.(in Chinese))

[10] 杨卓，马超. 基于BP神经网络方法的岩溶隧道突涌水风险预测[J]. 隧道建设，2016，36(11)：1 337–1 342.(YANG Zhuo，MA Chao. Risk prediction of water inrush of karst tunnels based on BP neural network[J]. Tunnel Construction，2016，36(11)：1 337–1 342.(in Chinese))

[11] 李志林，王星华，谢李钊. 基于模糊小波神经网络的岩溶隧道风险评估及综合超前地质预报技术[J]. 现代地质，2013，27(3)：719–726.(LI Zhilin，WANG Xinghua，XIE Lizhao. Risk evaluation and comprehensive geological prediction based on fuzzy wavelet neural network during tunneling in karst area[J]. Geoscience，2013，27(3)：719–726.(in Chinese))

[12] 李术才，王康，李利平，等. 岩溶隧道突水灾害形成机制及发展趋势[J]. 力学学报，2017，49(1)：22–30.(LI Shucai，WANG Kang，LI Liping，et al. Mechanical mechanism and development trend of water–inrush disasters in karst tunnels[J]. Chinese Journal of Theoretical and Applied Mechanics，2017，49(1)：22–30.(in Chinese))

[13] MAHMOODZADEH A，GHAFOURIAN H，MOHAMMED A H，et al. Predicting tunnel water inflow using a machine learning-based solution to improve tunnel construction safety[J]. Transportation Geotechnics，2023，40：100978.

[14] 朱珍，王旭春，袁永才，等. 基于加权平均法的岩溶隧道突涌水风险评估[J]. 公路工程，2015，40(6)：51–54.(ZHU Zhen，WANG Xuchun，YUAN Yongcai，et al. Risk assessment of water inrush in karst tunnels based on weighted average method[J]. Highway Engineering，2015，40(6)：51–54.(in Chinese))

[15] 黄小城，陈秋南，阳跃朋，等. 可拓理论对复杂条件下岩溶隧道的风险评估[J]. 地下空间与工程学报，2013，9(5)：1 179–1 185. (HUANG Xiaocheng，CHEN Qiunan，YANG Yuepeng，et al. Risk evaluation of karst tunnel under complex geological condition with extension theory[J]. Chinese Journal of Underground Space and Engineering，2013，9(5)：1 179–1 185.(in Chinese))

[16] 殷颖，田军，张永杰. 岩溶隧道灾害案例统计分析研究[J]. 公路工程，2018，43(4)：210–214.(YIN Ying，TIAN Jun，ZHANG Yongjie. Study on statistical analysis of karst tunnel disaster cases[J]. Highway Engineering，2018，43(4)：210–214.(in Chinese))

[17] 欧光照.基于机器学习的隧道施工期工程灾害风险评估研究[博士学位论文][D]. 武汉：中国地质大学，2022.(OU Guangzhao. Hazard risk assessment of tunnel during construction based on machine learning[Ph. D. Thesis][D]. Wuhan：China University of Geosciences，2022.(in Chinese))

[18] 王旺盛，陈长生，王家祥，等. 滇中引水工程香炉山深埋长隧洞主要工程地质问题[J]. 长江科学院院报，2020，37(9)：154–159.(WANG Wangsheng，CHEN Changsheng，WANG Jiaxiang，et al. Major engineering geological problems of Xianglushan deep-buried long tunnel in Central Yunnan Water Diversion Project[J]. Journal of Changjiang River Scientific Research Institute，2020，37(9)：154–159.(in Chinese))

[19] 李利平，李术才，陈军，等. 基于岩溶突涌水风险评价的隧道施工许可机制及其应用研究[J]. 岩石力学与工程学报，2011，30(7)：1 345–1 355.(LI Liping，LI Shucai，CHEN Jun，et al. Construction license mechanism and its application based on karst water inrush risk evaluation[J]. Chinese Journal of Rock Mechanics and Engineering，2011，30(7)：1 345–1 355.(in Chinese))

[20] 郑顺华，王迎超，陈帆，等. 地铁隧道涌水涌砂灾害特征及灾变模式分析[J]. 浙江大学学报：工学版，2025，59(1)：152–166.(ZHENG Shunhua，WANG Yingchao，CHEN Fan，et al. Characteristics of water and sand gushing disasters in subway tunnels and disaster modes analysis[J]. Journal of Zhejiang University：Engineering Science，2025，59(1)：152–166.(in Chinese))

[21] 吴志军，戴骞，储昭飞，等. 基于DNN和属性数学的岩溶隧道突涌水风险预测[J]. 中国公路学报，2024，37(7)：82–99.(WU Zhijun，DAI Qian，CHU Zhaofei，et al. Risk prediction of water inrush in karst tunnels based on DNN and attribute mathematics[J]. China Journal of Highway and Transport，2024，37(7)：82–99.(in Chinese))

[22] HE H，MA Y，eds. Imbalanced learning：Foundations，algorithms，and applications[M]. Hoboken，NJ：Wiley-IEEE Press，2013：13–41.

[23] 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.

[24] FERNÁNDEZ A，GARCÍA S，HERRERA F，et al. SMOTE for learning from imbalanced data：Progress and challenges，marking the 15-year anniversary[J]. Journal of Artificial Intelligence Research，2018，61：863–905.

[25] WU Z J，FAN L F，LIU Q S，et al. Micro-mechanical modeling of the macro-mechanical response and fracture behavior of rock using the numerical manifold method[J]. Engineering Geology，2017，225：49–60.

[26] BREIMAN L. Randomizing outputs to increase prediction accuracy[J]. Machine Learning，2000，40(3)：229–242.

[27] SONG M Y，YANG M H，YAO G Z，et al. Artificial intelligence driven tunneling-induced surface settlement prediction[J]. Automation in Construction，2024，168：105819.

[28] RUMELHART D E，HINTON G E，WILLIAMS R J. Learning internal representations by error propagation[C]// RUMELHART D E，MCCLELLAND J L，PDP RESEARCH GROUP，ed. Parallel Distributed Processing：Explorations in the Microstructure of Cognition. Vol. 1，Foundations. Cambridge，MA：MIT Press，1986：318–362.

[29] BRYSON A E，HO Y C. Applied optimal control：Optimization，estimation，and control[M]. Washington，D C：Hemisphere Publishing Corporation，1975：5–41.

[30] MCCULLOCH W S，PITTS W. A logical calculus of the ideas immanent in nervous activity[J]. The Bulletin of Mathematical Biophysics，1943，5(4)：115–133.

[31] LIU B，WANG R，ZHAO G，et al. Prediction of rock mass parameters in the TBM tunnel based on BP neural network integrated simulated annealing algorithm[J]. Tunnelling and Underground Space Technology，2020，95：103103.

[32] WOLPERT D H. Stacked generalization[J]. Neural Networks，1992，5(2)：241–259.

[33] 卢佳乐，张念，牛萌萌，等. 基于Stacking集成学习的隧道突水危险预测模型[J]. 中国安全科学学报，2025，35(4)：137–144.(LU Jiale，ZHANG Nian，NIU Mengmeng，et al. Hazard prediction model of tunnel water inrush based on Stacking ensemble learning[J]. China Safety Science Journal，2025，35(4)：137–144.(in Chinese))

[34] 许振浩，李术才，李利平，等. 基于层次分析法的岩溶隧道突水突泥风险评估[J]. 岩土力学，2011，32(6)：1 757–1 766.(XU Zhenhao，LI Shucai，LI Liping，et al. Risk assessment of water or mud inrush of karst tunnels based on analytic hierarchy process[J]. Rock and Soil Mechanics，2011，32(6)：1 757–1 766.(in Chinese))

[35] 许增光，王亚萍，肖瑜，等. 长深隧洞突涌水危险性等级指标及评价方法[J]. 中国公路学报，2018，31(10)：91–100.(XU Zengguang，WANG Yaping，XIAO Yu，et al. Risk rating index and evaluation method for water inrush in long-deep tunnels[J]. China Journal of Highway and Transport，2018，31(10)：91–100.(in Chinese))

[36] 许增光，线美婷，熊伟，等. 基于集对分析模型的岩溶区浅埋穿河隧道突涌水危险性评价[J]. 应用力学学报，2023，40(1)：135–145.(XU Zengguang，XIAN Meiting，XIONG Wei，et al. Risk assessment of water inrush in karst shallow tunnel under river based on SPA model[J]. Chinese Journal of Applied Mechanics，2023，40(1)：135–145.(in Chinese))