时间序列在隧道位移监测中综合应用-安全技术及工程专业论文.docx

摘要监控量测是新奥法隧道施工的重要组成部分，为评价设计方案、对围岩变化 情况及确定支护结构形式提供依据，并为施工及时提供围岩稳定程度和支护结构 可靠性的安全信息，预见事故和险情。隧道位移监测是隧道监控量测的主要内容 之一，对隧道施工、优化设计具有举足轻重的作用。本文基于时间序列分析方法， 对隧道位移监测数据进行了建模分析，主要研究成果如下：① 建立了适合隧道位移监测数据的混合模型，该模型既体现了研究对象的物 理意义，又包含了系统特性。② 鉴于模型参数估计是时间序列建模最关键的部分，估计精度和结果的好坏 直接决定了模型的选取和预测的精度。针对自回归滑动平均模型的非线性参数估 计特征，提出了将参数估计分―两步走‖，一是参数初估计，二是参数精估计。首先 分别应用自编 MATLAB 参数初估计程序和 EViews 参数估计程序，选择 RSS 较小 者的参数估计值作为最新的初估计，然后再应用自编 MATLAB 参数精估计程序求 解参数精估计值，结果表明这种方法较大地改善了参数的估计精度和预测效果。③ 模型的选取包括两部分：模型定阶和诊断检验。通过应用多种定阶准则和 统计检验求出最佳模型。在实例分析时，应用不同的时间序列建模方法进行建模， 并对结果作分析和对比，尽量选择参数较少的低阶模型。④ 用本文方法对隧道位移监测数据进行时间序列建模分析并进行短期预测、 数据回复和模型比较，最终得到了合适的时间序列模型，该模型的逼近效果、拟 合优度和精确性均较优，能够为指导施工实践和修改支护系统设计提供依据。关键词：隧道位移，时间序列分析，参数估计，建模方法，混合模型ABSTRACTMonitoring control and survey is an important component in NATM tunnel construction, which can provide basis for estimating design scheme, investigating changes of surrounding rock and confirming supporting structural forms, also in good time propose secure information about stability degree of surrounding rock and supporting structural reliability to foresee accident and danger. Displacement monitor of tunnel is a part of main contents in monitoring control and survey, so it plays an important role in tunnel construction and optimum design. Based the method of time series analysis, monitoring data of tunnel displacement is modeled and analyzed, from which the main studying productions are listed as follows:① Mixed model that is suitable for monitoring data of tunnel displacement isestablished. It not only incarnates physical meaning of studying object, but also contains system’s characteristic.② In view of the fact that estimate of model’s parameters is the most critical partin time series, model estimate precision and good or bad results determine straightly choice of model and prediction accuracy. Aim to characteristics of nonlinear parameter estimation in autoregressive moving average model, it is brought forward that parameter estimation is divided into two steps: the first step is pa