【作者】 秦世强；

【导师】 蒲黔辉；

【作者基本信息】 西南交通大学 ， 桥梁与隧道工程， 2013， 博士

【摘要】 桥梁作为交通土建的重要组成部分,其安全运营具有重要意义。随着科学技术的进步,桥梁朝着“大跨度”、“新材料”及“新体系”的方向发展,这给桥梁的设计、施工及后期运营管理都带来了新的挑战。而已经发生的桥梁事故时刻警醒着我们保证桥梁运营安全的重要性；在这样的背景下,基于动力测试的桥梁健康监测系统成为近年来的研究热点。随着工业制造和仪器仪表业的发展,传感器的精度已经得到很大的提升,硬件设备不再是健康监测的羁绊；而结构损伤识别在桥梁健康监测中的问题更多的是从理论到实际的转化。论文围绕桥梁健康监测与工作模态分析的理论、应用及系统实现展开了以下几个方面的研究：1、在现有文献的基础上总结了桥梁健康监测的研究现状。研究了桥梁健康监测信号分析处理方法；分别介绍了桥梁动力信号前处理、基于FFT的稳态频谱分析及时频分析方法。针对实测动力信号的非平稳的特点,重点介绍了时频分析工具Hilbert-Huang变换(HHT),通过仿真信号和实测信号的分析验证了HHT在分析非平稳、非线性信号方面的优势,同时指出了经验模态分解中存在的一些问题；2、介绍了环境激励模态参数识别常用方法；按照“结构随机状态模型”、“系统矩阵识别”、“模态参数识别”及“结果不确定度分析”系统地介绍了时域随机子空间识别(SSI)理论；针对SSI中存在的系统阶次难以确定的问题,提出了一种基于奇异值分解的系统阶次加权判定法；针对目前对SSI识别结果的精确性无法衡量的问题,引入了基于敏感性分析的模态参数不确定度量化方法,并结合多个测试组识别结果的方差分析,形成了一套从整体到局部的模态参数不确定度量化方法,构建了识别结果的置信区间；通过一个两自由度振动系统的数值模拟验证了SSI理论和所提出的方法；3、首次考虑了结构初始状态对工作模态分析的影响,解决了工作模态分析存在的两大主要问题之一：因环境荷载频带窄而导致高阶振动和扭转模态难以识别。首先从结构输出协方差出发,分析了初始条件对其精度的影响；其次,基于蒙特卡洛方法,利用一个单自由度振动系统分析了采样时间、初始条件、阻尼比等各种参数对工作模态分析精度的影响；最后利用一高速铁路高架桥的工作模态分析,验证了在考虑结构初始状态时,部分高阶模态能被识别,且识别精度满足要求；4、研究了经验模态分解(EMD)在模态参数识别中的应用,首先介绍了一种基于EMD和随机减量法的模态参数识别方法；然后针对EMD存在的模态混叠问题和SSI存在的虚假模态问题,提出了基于限制带宽的EMD和SSI的模态参数识别方法；限制带宽的EMD有效地抑制了模态混叠,而利用其分解出的本征模态函数作为SSI的输入,稳定图的虚假模态也得到抑制；通过松头江大桥的试验模态分析验证了这种方法；5、基于Visual Studio2010平台开发了一套桥梁健康监测软件(HBHM1.0)；介绍了软件的总体设计及各部分功能模块,并介绍了数据结构和数据库的开发；提出了一种数据采集模拟系统,便于软件的初期开发及调试；论文重点介绍了软件的信号处理及模态参数识别模块,同时为了完整性,介绍了软件中的损伤识别及多层次性能评估模块；由于软件考虑多种硬件设备接口,因此可服务于不同工程；并通过宜昌长江大桥的实测数据验证了软件在实际工程中的运行情况。最后,给出了论文研究的主要结论,并展望了后续的研究内容。更多还原

【Abstract】 As an important part of transportation civil engineering, the operational safety of bridge structures is of great meaning. With the development of science and technology, the bridges are heading torward the directions of "long span","new material" and "new structure system", which also brings new chanllenges to the design, construction and operational management of bridges. The already happened bridge failures always remind us the importance of operational safety of civil structures. Under this background, the vibration-based structural health monitoring (SHM) becomes a continuing research topic of recent years. The quick development of industrial manufactory and instrumentation greatly improved the precision of sensors, making them not be a barrier of SHM anymore, while the biggest problem of damage detection in SHM is the transformation from theoretical to application. This dissertation mainly focuses on the signal processing and operational modal analyis in bridge SHM and the related theory, application and implementation are investigated in following aspects:1. The state-of-the-art of bridge SHM is summarized based on the existing literatures. The signal processing methods for bridge SHM are studied. The preprocessing methods, stationary spectrum analysis based on Fast Fourier Transform (FFT) and time-frequency analysis are introduced respectively. The Hilbert-Huang Transform (HHT) is especially introduced for the analysis of non-stationary bridge dynamic test signals. Through mathematical simulation and test signal analysis, it is proved that HHT has advantages in dealing with non-stationary and nonlinear signals. Also, the existing problems of HHT are pointed out.2. The common used methods for modal parameter identification from ambient vibration test are studied. The stochastic subspace identification (SSI) is systematically investigated following the sequence of "structure stochastic state-space model","system matrices identification","modal parameter identification", and "uncertainty analysis". A weighting judge method is proposed for determining model order in SSI. For the problem of evaluating the accuracy of identified modal parameters, an uncertainty calculation method based on sensitive analysis is introduced. A system modal parameter uncertainty quantification procedure is established. The confidence intervals of modal parameters are constructed. Through a simulation example of a two degree of freedom system, the effectiveness and robustness of SSI and the proposed methods are illustrated.3. The effects of initial conditions in operational modal analysis (OMA) are considered for the first time. One biggest problem of OMA, which is the high bending and torsion modes are difficult to identify from output-only data because the frequency content of ambient excitation is usually narrow-banded, is solved. Firstly, the accuracy of output correlation sequence by considering large-amplitude initial conditions is studied. Secondly, the influences of sampling time, initial conditions and damping ratios and many other parameters on the accuracy of OMA are analyzed through a simulation example of a single degree of freedom system based on Monte Carlo Analysis. Finally, a full scale application is presented where the modal parameters of a high speed railway bridge are determined from output-only data. It is found that some additional high bending and torsion modes can be identified with good accuracy when considering initial conditions.4. The application of empirical mode decomposition (EMD) in modal parameter identification is studied. Firstly, a modal analysis method based on EMD and random decrement technique (RDT) is introduced. Secondly, in order to deal with the problem of mode mixing of EMD and spurious modes of SSI, a bandwidth restricted EMD based SSI method is proposed for OMA. The test dynamic signals are first decomposed into a series of intrinsic mode functions (IMFs), each of which represents only one frequency component, and then SSI is performed on IMFs to extract modal parameters. The spurious modes in stabilization diagram are greatly restricted. Through the OMA of Songtoujiang railway bridge, the proposed method is proved to be very effective.5. The bridge SHM software, HBHM1.0is developed based on Visual studio2010. The overall design and each function mode of HBHM are introduced. The data construction and database are also introduced. A simulated data acquisition system is proposed to the early stage development of the software and also for the debugging. This paper mainly focuses on the implementation of the function modes of signal processing and modal parameter identification. For the completeness of the theory, the function mode of the damage detection and multi-level structure condition assessment are also introduced. As HBHM considered different data interface, it can serve for different bridge SHM projects. At last, the running of the HBHM is fulfilled by the dynamic test of Yichang Yangtze river bridge.Finally, the main research contents of the thesis and the basic conclusions are summarized. The outlook for future research is also made.更多还原