基于坍落度的砾砂地层盾构渣土改良试验及其预测研究  

Slump-based soil conditioning of EPB shield in gravelly sand and its prediction study

作　　者：李辉 佟方硕 李宾 裴柏铮 Li Hui;Tong Fangshuo;Li Bin;Pei Bozheng(The 4th Engineering Co.,Ltd.,China Railway No.9 Bureau Group,Shenyang 110819,Liaoning,China)

机构地区：[1]中铁九局集团第四工程有限公司,辽宁沈阳110819

出　　处：《岩土工程技术》2025年第2期290-296,共7页Geotechnical Engineering Technique

摘　　要：在砾砂地层中掘进时,渣土改良效果是影响盾构掘进效率的关键因素。通过坍落度试验研究了泡沫、膨润土泥浆和高分子聚合物对改良土体流塑性的影响。以试验结果作为数据样本集,采用SVR,KNR,RFR和BPNN等常用机器学习方法构建了土体坍落度的预测模型,并将预测值与实际值进行了对比分析。研究结果表明:(1)泡沫对砾砂渣土流塑性的改良效果较好;(2)对于高含水率的砾砂地层,应使用高黏度的膨润土泥浆或PAM溶液进行改良,以起到保水增黏、防止喷涌的目的;(3)对比SVR,KNR和BPNN模型,RFR模型在预测时的性能表现最佳,能够更准确地预测改良渣土的坍落度,并且对模型进行了可解释性分析。The impact of soil conditioning is a critical factor influencing shield tunneling efficiency in strata of gravelly sand.Through a slump test,the impacts of foam,bentonite slurry,and polymer on the enhanced soil’s flow plasticity were examined.A pre-diction model of soil slump was provided using machine learning techniques like SVR,KNR,RFR,and BPNN,utilizing the test res-ults as the data sample set.The predicted and real values were then compared and examined.The study indicates that:(1)Foam has a greater impact on enhancing the gravelly sandy soil’s flow flexibility.(2)High-viscosity bentonite slurry or PAM solution should be applied over gravelly sandy stratum with high water content to retain water,improve viscosity,and prevent blowout.(3)The RFR mod-el outperforms the SVR,KNR,and BPNN models regarding prediction accuracy.It can also forecast the slump of the improved waste soil with greater precision.The model’s interpretability was examined as well.

关 键 词：土压平衡盾构 砾砂地层 渣土改良 坍落度试验 机器学习 

分 类 号：U455[建筑科学—桥梁与隧道工程]