【作者】 侯亚彬；

【导师】 姜德义；

【作者基本信息】 重庆大学 ， 安全技术及工程， 2008， 硕士

【摘要】 随着我国基础建设事业的高速发展和西部大开发的进一步推进,我国的公路工程、铁路工程和地下工程迅猛发展,大量长大、深埋隧道工程纷纷上马,同时新奥法信息化动态设计与施工技术越来越受到重视。本文以在建共和隧道为项目依托,分析隧道深埋段岩体性质及隧道初期整体式衬砌的可靠度,结合隧道现场监控量测主要做了以下几个方面的研究工作:①对隧道监控量测、反分析方法及可靠度的研究现状进行分析和总结,给出本文的主要研究内容和技术路线;②总结人工神经网络的基本原理和BP网络学习规则及网络的优缺点,并针对其缺点给出了相应的改进方法;同时也介绍了位移反分析的基本理论,并确定基于神经网络位移反分析的主要步骤;③简要介绍隧道新奥法施工,结合隧道监控量测技术分析得到共和隧道试验段在隧道开挖期间围岩周边收敛情况、锚杆轴力、接触应力及工字钢应力变化规律;④以共和隧道深埋段K43+170断面现场试验数据为基础,运用ANSYS有限元程序进行数值模拟,建立BP神经网络学习和训练样本;利用Matlab及其自带的神经网络工具箱编写BP神经网络程序,对隧道围岩参数进行反分析;⑤应用结构可靠性分析的基本原理和方法,建立隧道初期支护可靠度分析的极限状态方程,选用蒙特卡罗计算方法,运用ANSYS程序提供的PDS技术(Probabilistic Design System)进行8000次抽样计算,对隧道整体初期支护的可靠度进行分析和评价。更多还原

【Abstract】 With the rapid improvement of national infrastructure and further development of China’s Western Development Program, highway, railway and underground projects are all developing rapidly. A large number of long and deep tunnel projects are launched one after another. At the same time, New Austrian Tunneling Method of information dynamic design and construction technology are getting more and more attention. This thesis is based on Gonghe tunnel project which is under construction, rock behaviour and primary support’s reliability of deep tunnel are analysed. Combined with monitoring measurement in the tunnel, following areas are researched in this thesis:①Monitoring measurement in tunnel, back analysis method and reliability’s research status are analysed and summarized, main research contents and technical routes are given.②The basic principles of Artificial Neural Networks , learning rule and advantages and disadvantages of BP network’s are summarized, and appropriate improvement methods are given. Basic theories on reliability are listed, and the main steps of displacement back analysis based on neural network are confirmed.③The New Austrian Tunneling Method is briefly introduced, and the laws of convergence of the surrounding rock displacement, axial force of anchor bolt, contact stress and I-stress during the excavation of the tunnel are carried out combined with monitoring measurement.④Based on K43+170 section’s field-test data in deep buried of Gonghe tunnel, using ANSYS finite element program to simulate, the learning and training samples of BP neural network are established. Using BP neural network toolbox in Matlab compile program, Parameters of Surrounding Rock are calculated by back analysis.Using basic principle and methods of structural reliability analysis, the limit state equation of tunnel initial support’s reliability analysis is established. Using Probabilistic Design System in ANSYS, Monte Carlo Method was selected to make sample calculation for 8000 times. The reliability of integral primary support in tunnel is analysed and evaluated.更多还原