摘要
针对现有预测地表沉降方法的不足,结合土压平衡式盾构一个掘进周期中不同阶段引起地表沉降的状况不同,提出了应用最近邻抽样回归模型(NNBR模型)对纵向地表沉降进行动态预测的方法。文章初步总结出土舱压力和平均掘进速度、注浆压力和平均掘进速度、监测时间可分别作为开挖面前方变形、盾尾空隙沉降、固结沉降三个阶段的主要预测因子,并列出了相应阶段的相似计算公式。利用该方法对某一工程实例的固结沉降阶段进行了预测,在样本数较少的情况下,预测结果最大相对误差只有14.04%,从而论证了基于NNBR模型的动态预测方法的准确性和可行性。
Because the surface settlements are different in different stages of an excavation cycle by an EPB shield, and some defects exist in the current prediction methods, this paper discusses the dynamic prediction method for lon- gitudinal surface settlement using the nearest neighbor bootstrap re average driving speed, grouting pressure and average driving spee gr d, essive model （NNBR model）. Soil pressure and monitoring time are considered to be the main predictors of surface settlement in front of the excavation face, gap settlement at the shield tail, and consolida- tion settlement. Similarly, calculation formulas for corresponding stages are presented. The consolidation settlement in one case was predicted by the above method, with the maximum relative error of 14.04%, which demonstrates the ac- curacy and feasibility of the dynamic prediction method based on NNBR model
出处
《现代隧道技术》
EI
北大核心
2012年第6期82-87,共6页
Modern Tunnelling Technology
基金
超大规模地下空间绿色施工技术(11dz1201503)
关键词
最近邻抽样回归模型
盾构法
地表沉降
动态预测
Nearest-neighbor bootstrap regressive model
Shield construction
Surface settlement
Dynamic predic- tion
作者简介
王红卫（1974-），男，博士研究生，工程师，主要从事岩土工程风险研究等，E-mail：wang_hongwei2011@163．com．
