基于Hilbert-Huang理论斜拉索损伤识别方法研究
发布时间：2018-11-12 17:23
【摘要】：由于斜拉桥所具有的优势,在现代桥梁的建设中很受工程师和人们的青睐,已成为大跨度桥梁的主要桥型。本文充分利用前人在信号分析领域研究的成果和实际服役的斜拉桥,提出了解决Hilbert-Huang变换中的难点问题,并与小波分析进行对比,通过建立有限元模型,系统的研究了基于Hilbert-Huang变换的斜拉桥拉索损伤识别问题。本文主要研究内容包括:首先,通过分析Hilbert-Huang变换理论,给出了EMD结束准则,并分析了其理论中不成熟部分,通过与小波分析方法对比,给出了其在损伤识别中的优点,并引用算例证明了HHT变换较小波分析方法简便、有效。其次,基于原始的能量变化曲线来判别噪声与信号的分界点缺乏稳定性缺点,并根据信号与噪声的不同的统计特性,提出了新的EMD去噪方法-基于自相关的函数特性EMD去噪方法,通过实验用该去噪方法与傅里叶变换滤波去噪方法及小波变换去噪方法进行比较分析,验证了所提方法的有效性。再次,对端点效应的机制以及现有处理方法进行总结,在其基础上提出了一种自适应的端点相位正弦延拓方法(只对EMD进行端点效应处理),仿真分析表明,该方法不仅比目前现有的延拓方法,筛选次数更少,提高信号的EMD分解精度,而且有较好的解决信号端点效应的作用,有效减小Hilbert谱的端点效应。最后,将改进的Hilbert-Huang变换方法结合有限元分析模型,对在役斜拉桥拉索进行损伤识别分析讨论,通过实验分析,各损伤工况下均能有效的识别出拉索的损伤位置与程度。综上,本文基于Hilbert-Huang变换的信号分析处理方法,在对给出EMD结束准则,运用提出的基于自相关的函数特性EMD去噪方法和自适应的端点相位正弦延拓方法,结合有限元模型,实现了对某一在役斜拉桥拉索的准确损伤定位与程度识别。
[Abstract]:Because of the advantages of cable-stayed bridges, they are favored by engineers and people in the construction of modern bridges, and have become the main bridge type of long-span bridges. This paper makes full use of the achievements of previous researches in the field of signal analysis and the actual cable-stayed bridge in service, and puts forward a solution to the difficult problem in Hilbert-Huang transform, and compares it with wavelet analysis, and establishes the finite element model. The problem of cable damage identification of cable-stayed bridge based on Hilbert-Huang transform is studied systematically. The main research contents of this paper are as follows: firstly, by analyzing the Hilbert-Huang transform theory, the end criterion of EMD is given, and the immature part of the theory is analyzed. Compared with the wavelet analysis method, the advantages of EMD in damage identification are given. An example is given to prove that the HHT transform is simpler and more effective than the wavelet analysis method. Secondly, judging the boundary point of noise and signal based on the original energy change curve is lack of stability, and according to the different statistical characteristics of signal and noise, In this paper, a new EMD denoising method, EMD denoising method based on autocorrelation function characteristic, is proposed. The proposed method is compared with Fourier transform filtering method and wavelet transform method to verify the effectiveness of the proposed method. Thirdly, the mechanism of endpoint effect and the existing processing methods are summarized, and an adaptive end-point phase sinusoidal continuation method (only the endpoint effect is processed for EMD) is proposed. The simulation results show that, This method not only has fewer screening times than the existing continuation methods, and improves the EMD decomposition accuracy of signals, but also has a better effect on solving the endpoint effects of signals and effectively reducing the endpoint effects of Hilbert spectra. Finally, the improved Hilbert-Huang transform method is combined with the finite element analysis model to analyze and discuss the damage identification of the cable in service cable-stayed bridge. Through the experimental analysis, the location and extent of the cable damage can be effectively identified under each damage condition. In summary, based on the signal analysis and processing method of Hilbert-Huang transform, the EMD end criterion is given, and the proposed EMD denoising method based on autocorrelation and the adaptive end-point phase sinusoidal continuation method are used to combine the finite element model. The accurate damage location and degree identification of a cable in service cable-stayed bridge are realized.
【学位授予单位】：东华理工大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：U446
本文编号：2327725
[Abstract]:Because of the advantages of cable-stayed bridges, they are favored by engineers and people in the construction of modern bridges, and have become the main bridge type of long-span bridges. This paper makes full use of the achievements of previous researches in the field of signal analysis and the actual cable-stayed bridge in service, and puts forward a solution to the difficult problem in Hilbert-Huang transform, and compares it with wavelet analysis, and establishes the finite element model. The problem of cable damage identification of cable-stayed bridge based on Hilbert-Huang transform is studied systematically. The main research contents of this paper are as follows: firstly, by analyzing the Hilbert-Huang transform theory, the end criterion of EMD is given, and the immature part of the theory is analyzed. Compared with the wavelet analysis method, the advantages of EMD in damage identification are given. An example is given to prove that the HHT transform is simpler and more effective than the wavelet analysis method. Secondly, judging the boundary point of noise and signal based on the original energy change curve is lack of stability, and according to the different statistical characteristics of signal and noise, In this paper, a new EMD denoising method, EMD denoising method based on autocorrelation function characteristic, is proposed. The proposed method is compared with Fourier transform filtering method and wavelet transform method to verify the effectiveness of the proposed method. Thirdly, the mechanism of endpoint effect and the existing processing methods are summarized, and an adaptive end-point phase sinusoidal continuation method (only the endpoint effect is processed for EMD) is proposed. The simulation results show that, This method not only has fewer screening times than the existing continuation methods, and improves the EMD decomposition accuracy of signals, but also has a better effect on solving the endpoint effects of signals and effectively reducing the endpoint effects of Hilbert spectra. Finally, the improved Hilbert-Huang transform method is combined with the finite element analysis model to analyze and discuss the damage identification of the cable in service cable-stayed bridge. Through the experimental analysis, the location and extent of the cable damage can be effectively identified under each damage condition. In summary, based on the signal analysis and processing method of Hilbert-Huang transform, the EMD end criterion is given, and the proposed EMD denoising method based on autocorrelation and the adaptive end-point phase sinusoidal continuation method are used to combine the finite element model. The accurate damage location and degree identification of a cable in service cable-stayed bridge are realized.
【学位授予单位】：东华理工大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：U446
【参考文献】
相关期刊论文 前1条
1 李忠献,杨晓明,丁阳;应用人工神经网络技术的大型斜拉桥子结构损伤识别研究[J];地震工程与工程振动;2003年03期
，本文编号：2327725
本文链接：https://www.wllwen.com/kejilunwen/jiaotonggongchenglunwen/2327725.html
