【作者】 景海涛；

【导师】 章杨松； 程知言；

【作者基本信息】 南京理工大学 ， 建筑与土木工程， 2012， 硕士

【摘要】 隧道等地下工程建设期间的安全性一直是隧道建设关注的热点,开展隧道建设期风险分析研究对提高隧道建设的安全性与风险预测具有重要的理论意义和应用价值。为了对隧道建设期的风险进行分析,本文基于贝叶斯网络理论,应用数理统计与有限元模拟结果,建立了隧道围岩分级与围岩初期支护稳定性的贝叶斯网络风险分析模型,并基于网络技术建立了隧道建设期风险预测与风险减缓快速决断的专家系统,实现隧道建设期风险因素的快速识别与决断。本文针对影响隧道建设期风险发生规律以及影响隧道稳定性的因素作了较为深入的研究,主要内容包含以下几方面：(1)对隧道建设面临的主要风险问题进行分析,确定了影响围岩分级、围岩与初期支护稳定性的影响因素,阐述了贝叶斯网络的基本原理及其在隧道风险分析中应用的可行性,确定了贝叶斯网络风险分析各个结点以及它们之间关系。(2)对已建隧道围岩的分级情况进行了统计,得到影响Q分级和[BQ]分级各因素的先验概率,基于统计的数据建立了隧道围岩分级的贝叶斯网络模型,模型能够对围岩的级别做出快速的判断,并能由已知证据变量来预测未知变量的取值范围。(3)利用有限元软件对隧道开挖支护过程进行了模拟,得到岩体的强度参数与初期支护应力之间的表达式,在此基础上建立了围岩初期支护稳定性的贝叶斯网络模型。通过对模型的分析与调整,得到了围岩初期支护失稳的概率。利用所建网络,在改变网络中任何一个结点的状态变量时,得出其它变量状态的变化,通过这种变化为找到各个因素之间规律成为可能。(4)建立隧道施工方法专家系统,将搜集已建隧道的数据输入到专家系统,基于统计数据与专家经验建立隧道施工方法查询的系统。更多还原

【Abstract】 During construction of the tunnel, safety has been focus on the construction of tunnel,it’s value to carry out the risk analysis to improve the safety of the tunnel and to predicate the risk.In order to analysis the tunnel’s risk in the construction,this paper based on the theory of bayesian network, the mathematical statistics and the results of the finite element simulation was used.the bayesian network risk analysis models for the classify of the rock and stability of the initial rock-support has been built, and it set up an expert system for forecasting and reducing risk and determinating quickly during the contruction of the tunnel base on the network.lt achieves the fast identification and determination for the risk of the tunnel.This paper researched the rules of the risks and the stability of the tunnel which occurred during period of the construction tunnel, the main contents include the following aspects:First, the main risk problems were analyzed for the tunnel construction and determined the factors which affected classification of rock and the stability of initial rock-support.This paper expounded the basic principle of bayesian networks and the feasibility of risk analysis, and set up a bayesian network for risk analysis and their relationships between them.Second,classify of rock has been statisticed from the built-tunnels.And we got the prior probability which influence Q classification and [BQ],and based on the statistical data to set up the bayesian network models about the classify of rock,by use of the models, bayesian network models can make a rapid judgment for the level of rock.And we can be predict the unknown variables by the known variables of the tunnels.Third, we usethe finite element software to simulate the excavation of the tunnel and to support process.we got the formulas which can express the relation between the strength parameters of the rock and stress of supporting, and set up the bayesian network models.By the analysis of the model and adjustment, we obtained probability about instability of supporting,and if we changes the state of the variables in the network, other variables will change, through this change we can find various factors rules between them.Forth, an expert system has been set up which use to choose the right construction methods, we collected a lot of datas which came from the built-tunnel, and inputed these data to the system.We set up the tunnel construction method of the system based on the statistical datas and expert experiences.更多还原