作者：朱春柏¹, 刘志贺¹, 甘晓露^2,3, 李洛宾^2,3, 俞建霖^2,3, 龚晓南^2,3, 刘念武⁴

作者单位：1. 江苏中车城市发展有限公司, 江苏 无锡 214105;
2. 浙江大学 滨海和城市岩土工程研究中心, 浙江 杭州 310058;
3. 浙江大学 浙江省城市地下空间开发工程技术研究中心, 浙江 杭州 310058;
4. 浙江理工大学建筑工程学院, 浙江 杭州 310018

关键词：盾构隧道;循环神经网络;地表位移;全局敏感性

摘要：

依托杭州地区淤泥质黏土地层中某盾构隧道工程，利用循环神经网络建立盾构隧道施工引发地表最大位移的预测模型，基于该预测模型和Morris方法分析地表最大位移对各个施工参数变化的全局敏感程度，以探究影响盾构掘进引发地表最大位移的主导因素。研究结果表明，建立的循环神经网络预测模型对地表最大位移的预测效果良好；盾构隧道掘进过程中的同步注浆压力是影响地表最大位移的主导因素，因此施加合适的注浆压力是控制地表最大沉降和隆起量的关键。研究成果可为实际盾构隧道工程中的地表变形控制提供有益参考。

Global sensitivity analysis for construction parameters of shield tunneling using recurrent neural network
ZHU Chunbai¹, LIU Zhihe¹, GAN Xiaolu^2,3, LI Luobin^2,3, YU Jianlin^2,3, GONG Xiaonan^2,3, LIU Nianwu⁴
1. Jiangsu CRRC Urban Development Co., Ltd., Wuxi 214105, China;
2. Research Center of Coastal and Urban Geotechnical Engineering, Zhejiang University, Hangzhou 310058, China;
3. Engineering Research Center of Urban Underground Development, Hangzhou 310058, China;
4. School of Civil Engineering and Architecture, Zhejiang Sci-Tech University, Hangzhou 310018, China
Abstract: Based on a shield tunnel project in the muddy clay stratum of Hangzhou, a prediction model for the maximum ground displacement caused by shield tunneling is established using the recurrent neural network. The proposed prediction model and the Morris method are combined to obtain the global sensitivity of the maximum ground displacement to various construction parameters and determine the key influence parameters for the ground displacement due to shield excavation. The results show that the proposed model based on the recurrent neural network can predict the tunneling-induced maximum ground displacement fairly well. The synchronous grouting pressure during shield tunneling is the dominant factor affecting the maximum ground displacement, so the application of appropriate grouting pressure is important for the control the tunneling-induced ground settlement or heave in muddy clay. The research results can provide a useful reference for the control of the ground displacement in practical shield tunnel engineering.
Keywords: shield tunnel;recurrent neural network;ground displacement;global sensitivity analysis
2023, 49(5):158-163  收稿日期: 2021-07-28;收到修改稿日期: 2021-12-18
基金项目: 国家自然科学基金资助项目(5177858575)；浙江省重点研发计划项目(2019C03103)
作者简介: 朱春柏(1978-),男,河北唐山市人,高级工程师,主要从事岩土工程、结构工程等相关领域的工作
参考文献
[1] PINTO F, WHITTLE A J. Ground movements due to shallow tunnels in soft ground. I: Analytical solutions[J]. Journal of Geotechnical and Geoenvironmental Engineering, 2014, 140(4): 04013040
[2] LAVASAN A A, ZHAO C, BARCIAGA T, et al. Numerical investigation of tunneling in saturated soil: the role of construction and operation periods[J]. Acta Geotechnica, 2017, 13(3): 671-691
[3] 侯永茂, 郑宜枫, 杨国祥, 等. 超大直径土压平衡盾构施工对环境影响的现场监测研究[J]. 岩土力学, 2013, 34(1): 235-242
[4] 丁智, 王凡勇, 魏新江. 软土双线盾构施工地表变形实测分析与预测[J]. 浙江大学学报(工学版), 2019, 53(1): 61-73
[5] SUWANSAWAT S, EINSTEIN H H. Artificial neural networks for predicting the maximum surface settlement caused by EPB shield tunneling[J]. Tunnelling and Underground Space Technology incorporating Trenchless Technology Research, 2006, 21(2): 133-150
[6] CHEN R P, ZHANG P, KANG X, et al. Prediction of maximum surface settlement caused by earth pressure balance (EPB) shield tunneling with ANN methods[J]. Soils and Foundations, 2019, 59(2): 284-295
[7] ZHANG P, CHEN R P, WU H N. Real-time analysis and regulation of EPB shield steering using Random Forest[J]. Automation in construction, 2019, 106: 1-13
[8] 李洛宾, 龚晓南, 甘晓露, 等. 基于循环神经网络的盾构隧道引发地面最大沉降预测[J]. 土木工程学报, 2020, 54(S1): 13-19
[9] CHO K, MERRIENBOER B V, GULCEHRE C, et al. Learning phrase representations using RNN encoder-decoder for statistical machine translation[J]. Computer Science, 2014.
[10] 严勤, 邓高峰, 胡涛, 等. 基于深度循环神经网络的异常用电检测方法[J]. 中国测试, 2020, 47(7): 99-104
[11] 余晓, 吕飞鹏, 张国星, 等. 循环神经网络输电线路双端故障测距方法[J]. 中国测试, 2021, 47(9): 119-125
[12] MIRO S, HARTMANN D, SCHANZ T. Global sensitivity analysis for subsoil parameter estimation in mechanized tunneling[J]. Computers & Geotechnics, 2014, 56: 80-88
[13] MORRIS M D. Factorial sampling plans for preliminary computational experiments[J]. Technometrics, 1991, 33(2): 161-174
[14] CAMPOLONGO F, CARIBONI J, SALTELLI A. An effective screening design for sensitivity analysis of large models[J]. Environmental Modelling & Software, 2007, 22(10): 1509-1518