温度敏感性结构基于小波包变换的损伤识别研究与应用

发布时间：2018-03-29 15:26

本文选题：损伤识别　切入点：小波包变换　出处：《北京交通大学》2017年硕士论文

【摘要】：在结构的长期健康监测过程中,环境因素的变化会引起结构自身动力参数的改变,这些改变会使结构的损伤识别发生误判。为了避免由于温度变化而使损伤识别发生误判,本文利用结构在不同温度下的加速度响应,通过小波包变换提取小波包能量谱,并以此作为样本,利用多元统计分析中的主成分分析方法剔除或者减少温度对损伤特征参数的影响来进行损伤识别。本文分别以简支梁模型、ASCE的Benchmark数值模型和试验数据,以及古建监测数据为对象,运用小波包变换和主成分分析法提取特征参数并去除温度影响,对其进行损伤识别分析,主要工作包括以下几个方面:1)利用小波包变换构建损伤指标进行损伤识别研究。针对一跨长为5m的简支梁数值模型,对其动力响应进行小波包变换,得到小波包能量谱,并构建损伤指标,对不同位置损伤的进行损伤定位,并研究不同程度损伤和噪声对损伤指标的影响。2)以5m跨长的简支梁模型为对象,加入温度的影响,对得到动力加速度响应进行小波包变换和主成分分析,建立损伤指标和识别率因子,对不同温度下结构的损伤进行识别,并研究损伤位置、损伤程度及噪声对损伤指标的影响。同时对ASCE提出的第一阶段的Benchmark有限元模型加入温度的变化,得到不同温度下的加速度响应数据,用上述方法进行损伤识别,验证方法的有效性。3)对ASCE第二阶段的Benchmark模型试验得到的环境激励数据进行分析,通过自然激励技术和小波包变换得到的小波包能量谱,并对其进行归一化处理,构建损伤指标进行结构的损伤识别分析。4)对实测加速度数据样本利用本文方法进行分析。构建两组时间间隔为一年的样本,对样本进行数据的预处理,通过随机减量技术得到结构自由振动响应,然后利用小波包变换和主成分分析计算损伤指标,对结构进行损伤的判别,并分析其原因。
[Abstract]:In the process of long-term health monitoring of structure, the change of environmental factors will cause the change of dynamic parameters of the structure itself, and these changes will make the damage identification of the structure misjudge, in order to avoid the misjudgment of the damage identification due to the change of temperature. In this paper, the energy spectrum of wavelet packet is extracted by wavelet packet transform using the acceleration response of structure at different temperatures, and it is used as a sample. The method of principal component analysis (PCA) in multivariate statistical analysis is used to eliminate or reduce the effect of temperature on damage characteristic parameters to identify damage. In this paper, the Benchmark numerical model and experimental data of simply supported beam model ASCE are used, respectively. Using wavelet packet transform and principal component analysis method to extract the characteristic parameters and remove the influence of temperature, the damage identification analysis is carried out. The main work includes the following aspects: (1) damage identification is studied by using wavelet packet transform to construct damage index. For a simple supported beam model with a span of 5 m, wavelet packet transform is applied to the dynamic response of the model, and the wavelet packet energy spectrum is obtained. The damage index is constructed to locate the damage in different positions, and the effects of damage and noise on the damage index are studied. 2) the model of simply supported beam with 5m span is taken as the object, and the effect of temperature is added. The wavelet packet transform and principal component analysis are used to analyze the dynamic acceleration response. The damage index and recognition rate factor are established to identify the damage of structures at different temperatures, and the damage location is studied. The influence of damage degree and noise on the damage index. At the same time, the change of temperature is added to the Benchmark finite element model of the first stage proposed by ASCE, and the acceleration response data at different temperatures are obtained, and the above methods are used to identify the damage. Verify the validity of the method. 3) analyze the environmental excitation data obtained from the Benchmark model test in the second stage of ASCE. Through the natural excitation technique and wavelet packet transform, the wavelet packet energy spectrum is analyzed, and the wavelet packet energy spectrum is normalized. Construction of damage indicators for structural damage identification analysis .4) the actual acceleration data samples are analyzed by the method in this paper. Two groups of samples with a time interval of one year are constructed, and the samples are preprocessed. The free vibration response of the structure was obtained by random decrement technique. Then the damage index was calculated by wavelet packet transform and principal component analysis (PCA). The damage of the structure was judged and the causes were analyzed.
【学位授予单位】：北京交通大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TU317

【参考文献】