隧道位移反分析研究.pdf

■试验研究 20 11 年 隧道位移反分析研究 李虎军，李德武 （兰州交通大学，甘肃 兰州730070） 摘 要 以兰渝铁路线天池坪隧道为工程背景，以现场监控量测资料为依据，基于1STOPT 中的通用全局 优化算法克服了当今世界上在优化计算领域中使用迭代法必须给出合适初始值的难题，而由1STOPT 随机给 出，找出最优解来预测围岩最终位移值，再分别基于Matlab 中BP 神经网络和径向基神经网络进行位移反分析 法，为保证系统训练样本符合工程实际，利用FLAC3D 有限差分软件对实际隧道开挖过程进行数值模拟，再用实 际监控量测的位移值分析反演出符合实际工程的围岩参数。 关键词 隧道；位移反分析；1STOPT ；MATLAB ；工具箱；BP 神经网络；径向基神经网络 The Displacement Back Analysis of Tunnel LI Hu- jun，LI De- wu 库 （Lanzhou Jiaotong University,Lanzhou,730070） Abstract: Tianchiping Tunnel in Lan- Yu Railway was taked for example and the field test data was deemed as the 例 references,nowadays,1STOPT based on the general global optimization algorithm is overcome in the field of optimization calculation using iterative method must be given initial value problem in the world,right by random given,find out the optimal solution to predict eventually displacement value of surrounding rock, based on Matlab, however, BP neural network 案 and RBF neural network to carry on the displacement back analysis method, in order to ensure that the systematic training E samples comply with the actual project, using finite difference of actual software FLAC3D numerical simulation of tunnel A excavation process，in the practical monitoring displacement value analysis of the actual project performance of the C surrounding rock parameters. Keywords: tunnel; displacement back analysis; 1STOPT;MATLAB toolbox; BP neural network ; RBF neural network 模 0 引言