盾构下穿既有隧道位移控制施工参数多目标优化  

Multi-objective Optimization of Displacement of Shield Tunneling Underpass Existing Tunnel Based on GA-LSSVM combining NSGA-Ⅱ

作　　者：曾铁梅 王金峰 吴贤国[2] 田金科 陶文涛 杨赛 陈彬 ZENG Tiemei;WANG Jinfeng;WU Xianguo;TIAN Jinke;TAO Wentao;YANG Sai;CHEN Bin(Wuhan Metro Group Co.,Ltd.,Wuhan 430070,China;School of Civil Engineering and Mechanics,Huazhong University of Science and Technology,Wuhan 430074,China)

机构地区：[1]武汉地铁集团有限公司,武汉430070 [2]华中科技大学土木工程与力学学院,武汉430074

出　　处：《铁道标准设计》2022年第2期98-104,共7页Railway Standard Design

基　　金：国家重点研发计划项目(2016YFC0800208);国家自然科学基金项目(51378235,71571078,51308240)。

摘　　要：盾构施工下穿既有隧道工程建设规模日益扩大,为研究盾构隧道下穿对既有隧道的安全影响和变形控制,建立GA-LSSVM与NSGA-Ⅱ算法相结合的多目标优化模型,以主要盾构施工参数为研究对象,对下穿施工引起的既有隧道底部水平和沉降变形为控制目标,进行施工参数优化控制分析。首先,基于收集的土仓压力、泡沫量、同步注浆量等6个盾构施工参数及对应的隧道底部位移样本数据;然后,利用GA改进LSSVM模型实现既有隧道底部水平和沉降位移高精度预测,并将GA-LSSVM预测得到的盾构施工参数诱发下穿既有隧道变形的非线性映射关系,作为NSGA-II适应度函数;最后,建立NSGA-II调用GA-LSSVM非线性预测函数的一套完整智能算法框架,结合数据范围进行施工参数多目标优化。以某地铁项目为例,研究结果表明,盾构下穿既有隧道底部水平和沉降位移预测模型泛化能力强、预测效果快速精确,通过多目标优化得到的盾构参数组合符合工程实际要求,既有隧道水平位移和沉降位移分别减小了约34%和10%,隧道变形得到有效控制。In order to study the safety impact and deformation control of the existing tunnel,a multiobjective optimization model combining GA-LSSVM and NSGA-Ⅱalgorithm is established.Taking the main shield construction parameters as the research object,the optimal control analysis of horizontal displacement and settlement displacement of the existing tunnel arch bottom caused by adjacent construction is carried out.First,six shield construction parameters and corresponding displacement data of the tunnel arch base,such as earth pressure,foam volume and synchronous grouting volume are collected.And then,GA improved LSSVM model is used to realize high precision prediction of the horizontal and settlement displacements of the existing arch base,and the two GA-LSSVM regression prediction functions are taken as the objective functions.Finally,the NSGA-Ⅱalgorithm is used to fulfill the construction parameters multiple objectives in combination with the data range.Based on a subway project,research results show that the prediction model of the horizontal and settlement displacements of the existing tunnel under the shield has strong generalization ability and fast and accurate prediction.The combination of shield parameters obtained by multi-objective optimization meets the actual requirements of the project,and the horizontal and settlement displacements of the existing tunnel are reduced by 34%and 30%respectively.The tunnel deformation is effectively controlled.

关 键 词：盾构隧道 下穿施工 变形 盾构施工参数 优化控制 GA-LSSVM NSGA-Ⅱ 

分 类 号：U231.3[交通运输工程—道路与铁道工程] U455.43[建筑科学—桥梁与隧道工程]