基于加速度传递率函数的结构损伤识别研究

发布时间：2018-06-30 09:46

本文选题：损伤识别 + 传递率函数　；参考：《青岛理工大学》2013年硕士论文

【摘要】：大型土木工程结构长时间服役将受到各类载荷的作用以及各种不确定外在因素的影响而面临结构损伤的问题，一旦发生破坏将导致灾难性的后果。结构损伤识别可以快速而准确的判别结构损伤与否、损伤的位置和损伤的程度，从而能够及时地提醒人们采取有效的措施进行补救，避免重大工程事故的发生。结构损伤识别方法通常需要已知激励信息，而激励信息一般难以准确测得，基于加速度传递率函数的结构损伤识别方法不需要已知激励信息，不需要进行模态分析。本文利用加速度传递率函数对结构进行损伤识别研究，主要内容如下：（1）介绍了结构损伤识别的研究背景及意义，综述了基于振动的结构损伤识别发展与研究现状。（2）介绍了传递率函数的概念与主成分分析的基本理论。利用加速度传递率函数和主成分分析进行结构损伤预警。首先，对加速度响应进行傅里叶变换构造传递率函数，并形成传递率函数原始数据矩阵。然后，利用主成分分析法对传递率函数矩阵进行降维，通过判断结构健康状态和未知状态下前两阶的主成分二维PCA空间散点图来判别结构是否发生损伤。同时，文中利用主成分置信度对结构进行了损伤预警。海洋平台数值模拟和振动台实验证明了这两种方法是有效的。（3）利用传递率函数的特性定义了新的结构定位损伤指标，求出结构中相邻连点间的损伤指标，通过比较相邻两点间的损伤指标的大小来判断海洋平台中支撑的损伤位置。海洋平台数值模拟和振动台实验证明了该方法是有效的。（4）介绍了马氏距离的概念。利用结构在健康和损伤情况下的实测传递率函数数据分别建立原始数据矩阵。对原始数据矩阵进行主成分分析，，分别求出结构在健康和损伤情况下的主成分。计算健康状态和损伤状态下传递率函数主成分数据间的马氏距离，根据计算出的马氏距离的大小对结构的损伤程度进行评估。海洋平台数值模拟和振动台实验证明了该方法是可行的。
[Abstract]:Long time service for large civil engineering structures will be faced with structural damage due to the effect of various loads and the influence of various uncertain external factors. Once the damage occurs, it will result in disastrous consequences. It is timely to remind people to take effective measures to remediate and avoid major engineering accidents. The method of structural damage identification usually requires known excitation information, and the incentive information is generally difficult to accurately measure. The structural damage identification method based on the acceleration transfer rate function does not need the known incentive information and does not need modal analysis. In this paper, the acceleration transmissibility function is used to study the damage identification of structures.
(1) introduces the research background and significance of structural damage identification, and summarizes the development and research status of structural damage identification based on vibration.
(2) the concept of transfer rate function and the basic theory of principal component analysis are introduced. The structure damage warning is made by using the acceleration transfer rate function and principal component analysis. First, the Fourier transform is used to construct the transfer rate function to the acceleration response, and the original data matrix of the transfer rate function is formed. Then, the principal component analysis method is used for the transfer rate. The function matrix is reduced to determine the damage of the structure by judging the structure health state and the first two order PCA space scatter plot of the principal component in the unknown state. At the same time, the damage early-warning is made by using the principal component confidence. The two methods are proved to be effective by the numerical simulation of the offshore platform and the shaking table experiment.
(3) a new damage index of structural location is defined by the characteristics of the transfer rate function, and the damage index between adjacent links in the structure is calculated. The damage position of the support in the offshore platform is judged by comparing the damage index between two adjacent two points. The numerical simulation of the offshore platform and the shaking table experiment prove that the method is effective.
(4) the concept of Mahalanobis distance is introduced. The original data matrix is established using the measured transfer rate function data of the structure under the condition of health and damage. The principal component analysis is carried out on the original data matrix, and the principal components of the structure under the condition of health and damage are calculated respectively. The main fraction of the transfer rate function in the health state and the damage state is calculated. According to the martensitic distance, the damage degree of the structure is evaluated according to the calculated martensitic distance. The numerical simulation of the offshore platform and the shaking table experiment prove that the method is feasible.
【学位授予单位】：青岛理工大学
【学位级别】：硕士
【学位授予年份】：2013
【分类号】：TU317

【引证文献】