基于小样本学习的盾构穿越粉质黏土夹粉土地层掘进参数控制值研究  

Few-shot Learning-based Shield Structure Crossing Powdery Clay with Powdered Ground Layer Research on Control Values of Digging Parameters

作　　者：张孝兵 李国田 张岩 谢阶祥 谭智颜 许卓淋 ZHANG Xiaobing;LI Guotian;ZHANG Yan;XIE Jiexiang;TAN Zhiyan;XU Zhuolin(China Railway Fifth Bureau Group Second Engineering Co.,Ltd.Hengyang 421002,China;School of Civil Engineering,Central South University Changsha 410075,China)

机构地区：[1]中铁五局集团第二工程有限责任公司,湖南衡阳421002 [2]中南大学土木工程学院,长沙410075

出　　处：《广东土木与建筑》2025年第2期97-99,共3页Guangdong Architecture Civil Engineering

摘　　要：随着城市地铁建设的迅速发展,盾构施工技术在软弱地层的应用受到广泛关注。研究基于武汉地铁12号线盾构穿越粉质黏土夹粉土地层的实际工程背景,通过在试验段收集掘进参数样本,并运用小样本学习算法训练预测模型,对施工方案及参数进行优化。研究表明,所提出的盾构掘进参数预测与优化方法能够有效利用有限数据进行学习,为盾构施工的安全高效提供了有力支持,对类似工程具有重要的参考价值。With the rapid development of urban metro construction,the application of shield construction technology in soft strata has received wide attention.Based on the actual engineering background of Wuhan Metro Line 12 shield structure traversing powdery clay sandwiched with pulverised soil stratum,this study optimises the construction scheme and parameters by collecting samples of tunneling parameters in the test section and training the prediction model using the small-sample learning algorithm.The study shows that the proposed shield tunneling parameter prediction and optimisation method can effectively use the limited data for learning,which provides strong support for the safety and high efficiency of shield construction and has important reference value for similar projects.

关 键 词：盾构施工 盾构穿越软弱地层 小样本学习 掘进参数优化 

分 类 号：U455.43[建筑科学—桥梁与隧道工程]