小波降噪在结构损伤识别中的应用

发布时间：2019-04-11 18:56

【摘要】：结构健康监测已逐渐发展成为土木工程结构领域研究的热点,损伤识别是结构健康监测的重要环节,在实际工程中,一般使用动态响应实测数据对结构进行损伤识别。基于实际情况,在进行损伤识别的研究过程中,都会将噪声作为考虑因素。本文以此为出发点,以小波分析为工具,对含噪结构响应信号进行降噪处理,使得结构本身信息在待处理信号中更加凸显,避免噪声对于损伤识别效果的干扰。首先,探究了影响结构本身响应信号的因素,对常见土木工程结构在常见荷载作用下的响应信号进行了分析。了解分析结构自身响应信号特点,并将其和噪声信号相对比,此工作有利于对结构自身有用信号和噪声信号进行区分,使得降噪工作具有针对性,以获得更加良好的降噪效果。然后,利用小波降噪的三种方法对含噪信号进行降噪处理,以降噪前后信号信噪比的变化和均方根误差为评价指标,验证了小波降噪方法的有效性,并探究了小波分解层数和小波函数的选择对于降噪效果的影响。通过时域和频域内降噪前后信号的对比,分析降噪效果。对降噪方案中小波函数选择,分解层数选择,阈值选取规则和阈值量化函数等参数选择给出建议。最后,利用简支梁模型算例,在各种工况下,分别对未加噪信号、加噪信号、降噪后信号进行损伤识别。通过对比损伤识别结果,探究小波降噪方案在损伤程度、损伤位置、噪音强度等影响因素下的适用范围,对小波阈值降噪方法适用工况给出建议。本文以小波分析为载体,研究土木工程领域中降噪方法,对小波降噪方案中各种参数给予建议,为实际工程结构信号分析提供依据,具有重要的科学意义和广阔的应用前景。
[Abstract]:Structural health monitoring has gradually become a hot spot in the field of civil engineering structure research. Damage identification is an important part of structural health monitoring. In practical engineering, dynamic response data are generally used to identify structural damage. Based on the actual situation, noise is considered as a factor in the research of damage identification. This paper takes this as a starting point and uses wavelet analysis as a tool to de-noising the noise-containing structural response signal, which makes the structural information more prominent in the signal to be processed, and avoids the noise interference to the damage identification effect. Firstly, the factors that influence the response signals of civil engineering structures are discussed, and the response signals of common civil engineering structures under common loads are analyzed. By understanding and analyzing the characteristics of the structure's own response signal and comparing it with the noise signal, this work is helpful to distinguish the useful signal from the noise signal of the structure itself, and make the noise reduction work targeted, so as to obtain a better noise reduction effect. Then, three methods of wavelet de-noising are used to de-noising the noisy signal. Taking the change of signal-to-noise ratio and root mean square error before and after de-noising as the evaluation indexes, the effectiveness of the wavelet denoising method is verified. The influence of wavelet decomposition layer number and wavelet function on noise reduction is also discussed. The de-noising effect is analyzed by comparing the signal before and after de-noising in time domain and frequency domain. The selection of wavelet function, decomposition layer, threshold selection rule and threshold quantization function are suggested. Finally, an example of simple-supported beam model is used to identify the damage of the non-noised signal and the de-noised signal under various working conditions. By comparing the damage identification results, this paper explores the application range of wavelet de-noising scheme under the influence factors such as damage degree, damage location and noise intensity, and gives some suggestions on the application of wavelet threshold de-noising method under the conditions of damage degree, damage location and noise intensity. In this paper, the wavelet analysis is used as the carrier to study the de-noising method in the field of civil engineering, and the various parameters in the wavelet de-noising scheme are suggested, which provides the basis for the signal analysis of the practical engineering structure, and has important scientific significance and broad application prospect.
【学位授予单位】：华中科技大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：TU317

【相似文献】