基于SVM动能指标的桥梁时域损伤识别

发布时间：2018-02-09 07:48

本文关键词： 桥梁时域损伤识别 SVM 动能指标　出处：《武汉理工大学》2014年硕士论文　论文类型：学位论文

【摘要】：本文通过对桥梁结构时域损伤识别方法的总结分析发现，对于桥梁服役期间车辆荷载对桥梁损伤的影响的研究较少。有鉴于此，本文将直接利用结构动力响应的速度信号，在时域内构建基于SVM动能能量比的损伤指标，进行损伤定位，依据动能能量比指标的变化判断桥梁损伤的程度。并应用该方法到有限元数值模拟试验中，主要研究工作包括：（1）通过阅读大量的国内外文献，对桥梁结构损伤识别的背景和意义以及其发展历史和研究的现状进行总结。重点介绍了主要的优化识别方法和模式识别方法在桥梁时域损伤识别中应用的情况。并对这些方法在桥梁损伤识别应用中存在的问题进行分析。（2）对基于SVM动能能量指标的时域损伤识别方法进行阐述，包括支持向量机基本理论、基本概念以及优势，最优分类超平面，核函数和分类边界，SVM算法，，动能比损伤识别理论。通过移动荷载作用下梁的动力方程，从而导出桥梁-车辆耦合系统中梁在强迫振动下某时刻的动力响应，最后利用得到的动力响应数据构建动能能量比指标，进行结构损伤程度和位置的识别。（3）通过数值模拟实验运用上述方法。首先介绍实验方法和过程，给出混凝土损伤模型以及汽车车道荷载实现的技术途径。通过建立简支梁桥有限元模型探讨于SVM动能能量指标的时域损伤识别方法的实现过程，然后为了研究该方法的适用性，拓展该方法应用到一座3跨预应力混凝土连续梁桥按一定比例缩放的有限元模型中，同样按照这种方法进行桥梁损伤位置和损伤程度的识别。（4）徐葛大桥时域损伤识别。以大跨度斜拉桥——徐葛大桥为研究对象，通过一座实际的徐葛大桥实测的传感器优化布置数据模型为基础，验证了利用动能能量指标的SVM方法在时域损伤识别的可行性，取得较好的效果，对于主塔的损伤识别给出了几种探讨，同时又由于这一方法的研究目前尚处于理论研究中，要使其能真正应用到桥梁结构健康监测，需要进一步研究。
[Abstract]:Based on the summary and analysis of time-domain damage identification methods for bridge structures, it is found that there are few studies on the effects of vehicle loads on bridge damage during bridge service. In view of this, the velocity signals of dynamic responses of structures will be used directly in this paper. The damage index based on SVM kinetic energy ratio is constructed in time domain, and the damage location is carried out. The damage degree of bridge is judged by the change of kinetic energy ratio index. The main research work includes:. By reading a great deal of literature at home and abroad, In this paper, the background and significance of bridge structure damage identification, its development history and research status are summarized, and the application of main optimization and pattern recognition methods in bridge damage identification in time domain is introduced. The problems existing in the application of these methods in bridge damage identification are analyzed. (2) the damage identification method in time domain based on SVM kinetic energy index is expounded, including the basic theory, basic concepts and advantages of support vector machine, the optimal classification hyperplane, kernel function and classification boundary SVM algorithm. The kinetic energy ratio damage identification theory. Through the dynamic equation of the beam under moving load, the dynamic response of the beam under forced vibration at a certain time in the bridge-vehicle coupling system is derived. Finally, the kinetic energy ratio index is constructed by using the dynamic response data to identify the damage degree and location of the structure. 3) the above methods are used in numerical simulation experiments. First, the experimental methods and processes are introduced. The concrete damage model and the technical way to realize the vehicle lane load are given. By establishing the finite element model of the simply supported beam bridge, the realization process of the time domain damage identification method based on the SVM kinetic energy index is discussed, and then the applicability of the method is studied. The method is extended to the finite element model of a 3-span prestressed concrete continuous beam bridge which is scaled to a certain scale. The damage location and damage degree of the bridge are also identified according to this method. (4) time domain damage identification of Xuge Bridge. Taking the long-span cable-stayed bridge-Xuge Bridge as the research object, the data model of optimal sensor arrangement of a practical Xuge Bridge is used as the basis. The feasibility of the SVM method using kinetic energy index in time domain damage identification is verified, and good results are obtained. Several discussions on damage identification of the main tower are given. At the same time, the research of this method is still in the theoretical research at present. Further research is needed to make it applicable to bridge structure health monitoring.
【学位授予单位】：武汉理工大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：U446

【参考文献】