基于围岩力学参数反演的隧道结构受力分析  

Mechanical analysis on tunnel structure based on back-analysis of surrounding rock mechanical parameters

作　　者：李讯[1] 何川[1] 张政[1] 汪波[1] 

机构地区：[1]西南交通大学地下工程系,四川成都610031

出　　处：《铁道建筑》2012年第8期43-46,共4页Railway Engineering

基　　金：中央高校基本科研业务费专项资金资助(SWJTU12CX066);交通部西部交通科技项目(200731800039)

摘　　要：本文以重庆摩天岭隧道为例,采用正交数值试验,获得了神经网络的训练样本,并结合现场监控数据反演了围岩物理力学参数,最后根据反演结果对隧道结构受力进行了分析。研究结果表明:在规范规定的同种围岩级别力学参数范围内取值得到的数值模拟结果相差甚大;采用正交试验、现场监控量测与BP神经网络结合的手段反演结果较好,得到的力学参数能较好地反映隧道的实际受力状况;隧道衬砌为受压控制,支护结构处于安全状态,原设计合理。Taking Motianling tunnel in Chongqing as an engineering background,this paper obtained the training samples of neural network by orthogonal experiment method, and analyzed the mechamcal characteristics of tunnel structure based on the inversion results by back analysis of surrounding rock parameters which used the field monitoring data. The study results indicated that the numerical simulation results vary considerably by adopting the values among the range of the rock mechanics parameters in the same level specified by the specification,inversion results are better by using the orthogonal experiment method and combining the field monitoring measurement with BP neural network, the mechanical parameters obtained by this way can reflect the actual stress status of the tunnel, tunnel lining is in pressure control, supporting structure is in safe state, and the original design is reasonable.

关 键 词：隧道 参数反演 数值模拟 正交试验 神经网络 

分 类 号：U451.2[建筑科学—桥梁与隧道工程]