【作者】 王艺霖；

【导师】 刘西拉；

【作者基本信息】 上海交通大学 ， 结构工程， 2011， 博士

【摘要】 对桥梁结构进行有效地损伤识别是目前的研究热点之一。本论文针对梁式桥,参照处理反问题的通用途径,以“尽量有效地增加有用信息量以减小损伤识别的空间尺度”为指导思想展开了研究。论文通过分析桥梁上可能出现的各种损伤类型及其特征明确了理想损伤识别指标的标准,从局部性指标(静力应变)、半局部性指标(静力差分曲率)到整体性指标进行分步寻求,经过对静力应变指标和差分曲率指标的研究后发现,“整体性指标加荷载移动”的模式是构建理想损伤识别指标的有效途径。进而,论文基于影响线二次差值的思想,对支座反力和局部点位移两个整体性指标做进一步的挖掘,分别建立起基于竖向支座反力测量的“反力影响线差值指标DIILSR(Difference Index of Influence Lines for Support Reaction)”基于跨中点位移测量的“跨中位移影响线差值指标DIILMSD(Difference Index of Influence Lines for Mid-span Displacement )”和基于两对称点位移测量的“对称点位移影响线差值指标DIILDSP(Difference Index of Influence Lines for Displacement at Symmetrical Points)”。这些新方法利用了“虚拟分割”的概念来将结构划分成多个小区段,假定损伤区域的长度等于这些小区段的长度,这样能在计算上减少待求未知量的数目,同时能做到较好的损伤定位和定量。新方法通过“让荷载移动起来”的方式,建立了指标与结构局部损伤位置之间的关系,不再需要利用测点位置来定位损伤,消除了对测点数量的依赖性。论文提出的新方法计算过程稳定,应用效果可以通过增加加载值和虚拟分割长度的取值来进行调节:加载值越大、虚拟分割的长度值越小,则方法的效果越好。在实际应用时,新方法所需的测量简单、加载容易、成本低。因此待测桥梁的尺寸越大、所能配置的测量装置越少,采用本方法的优势就越明显。更重要的是,新方法立足于“差值”,对实际桥梁的材料性质和截面尺寸参数不均匀、边界条件的不明确不敏感。此外,围绕损伤识别的需要,作者还针对埋入式应变传感器的耐用性、高精度位移数据的获取等问题设计了实用的新型装置并获得国家专利授权,有关细节也在论文中做了介绍。为了进一步说明新方法的应用价值,本文最后建议了一套完整的针对梁式桥的损伤识别方案,可具有成本低、效果好、通用性强等优势。更多还原

【Abstract】 At present, the damage detection of bridges is one of the research focuses in structural engineering. Focusing on beam bridges and considering the general approach in inverse problems,‘Make every effort to increase the useful information effectively and reduce the dimensions of damage detection’is treated as guidance in the present study. At first, the types and characteristics of damages which may occur on bridges are analyzed. The criterions of ideal damage detection index are then proposed. Research is carried out step by step from local index (static strain), semi-local index (static difference curvature) to global index. The static strain and difference curvature indices are studied in detail and then it is found that‘global index plus mobile load’can be treated as an effective pattern for establishing more proper damage detection index.Further, based on two global indices (support reaction and local point displacement), the idea of‘influence line second order difference’is introduced. Three novel indices termed as‘Difference Index of Influence Lines for Support Reaction, DIILSR’,‘Difference Index of Influence Lines for Mid-span Displacement, DIILMSD’and‘Difference Index of Influence Lines for Displacement at Symmetrical Points, DIILDSP’are then developed, respectively. The philosophy of these new methods is different from the existing damage detection methods. They use the concept of“virtual mashing”to divide the whole bridge into a number of segments with the same length. Therefore, not determining the exact damage location but detecting which segments contain damage is the work needed to be done. The number of unknown parameters which have to be solved is decreased. With the help of this concept, the localization and quantitative analysis of damage can be achieved more effectively. Besides, mobile load is employed to relate the local location of bridge with index values. Therefore, the number of measurement points will not affect the effect of these methods.The algorithm of these new methods is stable and their effect can be adjusted freely by changing the load value and the segment length. Better effect can be achieved with the increase of load value and the decrease of segment length. During the application process, simple measurement, convenient loading and low cost are advantages of new methods. These advantages are further highlighted when the bridge size increases and the number of offered sensors decreases. Moreover, based on the‘difference value’concept, these new methods can be applied to the actual bridges with non-homogeneous material/section parameters and unclearly defined support conditions.In addition, in order to satisfy the requirements of damage detection, innovative devices have also been designed to improve the durability of embedded optical fiber strain sensors and the acquisition of displacement data with high precision etc., which are all introduced in detail in the present thesis. These devices have been patented by National Intellectual Property Bureau.Finally, an integrated bridge damage detection scheme which has low cost and well effect for beam bridges is proposed to show the application value of this study.更多还原