基于小波变换奇异性理论的框架结构损伤识别的方法研究

发布时间：2018-07-07 22:30

本文选题：结构健康诊断 + 结构损伤识别　；参考：《辽宁工程技术大学》2013年硕士论文

【摘要】：近二十年对建筑结构进行健康诊断的意义越来越明显,结构损伤识别与结构损伤识别方法都有了很大发展,但是现在大部分损伤识别方法都是利用结构损伤前后特性的改变,需要分析损伤前后的动力响应,进而比较得到的损伤指标,因此具有一定的局限性。而且,这些损伤识别方法对于分析信号只能在频域内或者时域内进行,没有对信号的时频信息进行充分利用。本文针对这些不足之处,基于小波时频分析特点,研究了基于小波变换理论的建筑结构方面的健康诊断。本文详细阐述了小波变换的原理及小波变换在建筑结构健康诊断中的应用,利用ASCE编制的Benchmark模型数据生成程序获取了结构的响应信号,通过对不同小波变化细节信号的比较,选取了Bior小波对建筑结构响应信号进行小波变换,利用变换后的细节信号对建筑物结构健康进行了诊断,判定了结构损伤的存在以及结构损伤发生的时刻和位置。论文研究发现,当建筑结构突然出现损伤时,经过小波变换后的细节信号会出现尖峰,并且当点距离损伤位置越近,这点加速度响应的细节信号经小波变换后的尖峰高度越大；距离损伤位置越远,细节信号尖峰高度越小。本文还研究并验证了采样频率对于结构损伤识别的影响,经过仿真研究发现,当采样频率大于1kHz时结构损伤识别可以得到较好的效果。
[Abstract]:In the past two decades, it is more and more important to make health diagnosis of building structures. The methods of structural damage identification and structural damage identification have been developed greatly. However, most of the damage identification methods are based on the changes of structural damage characteristics before and after damage. It is necessary to analyze the dynamic response before and after the damage, and then compare the damage index, so it has some limitations. Moreover, these damage identification methods can only be used in the frequency domain or in time domain, and the time-frequency information of the signal can not be fully utilized. In this paper, based on the characteristics of wavelet time-frequency analysis, the health diagnosis of building structure based on wavelet transform theory is studied. In this paper, the principle of wavelet transform and the application of wavelet transform in building structure health diagnosis are described in detail. The response signals of the structure are obtained by using benchmark model data generation program compiled by ASCE, and the details of different wavelet changes are compared. In this paper, the wavelet transform is selected to transform the response signal of the building structure, the health of the building structure is diagnosed by the detail signal after the transformation, and the existence of the structural damage and the time and position of the damage are determined. It is found that when the building structure is damaged suddenly, the detail signal after wavelet transform will appear spike, and when the point is close to the damage position, the peak height of the detail signal of the point acceleration response after wavelet transform will be higher. The farther away from the damage location, the smaller the peak height of the detail signal. The effect of sampling frequency on structural damage identification is also studied and verified. The simulation results show that when the sampling frequency is greater than 1kHz, the structure damage identification can get better results.
【学位授予单位】：辽宁工程技术大学
【学位级别】：硕士
【学位授予年份】：2013
【分类号】：TU317

【参考文献】