低速行驶地铁轨道振动信号小波去噪参数研究  

Study on Wavelet De-noising Parameters of Track Vibration Signal for Metro with the Train Traveling on at Low Speed

作　　者：曾宇 邬玉斌[1] 宋瑞祥[1] 户文成[1] ZENG Yu;WU Yubin;SONG Ruixiang;HU Wencheng(Beijing Municipal Institute of Labour Protection, Beijing 100054, China)

机构地区：[1]北京市劳动保护科学研究所,北京100054

出　　处：《噪声与振动控制》2019年第1期151-155,200,共6页Noise and Vibration Control

基　　金：北京市财政资助项目(PXM2018_178304_000008);北京市科学技术研究院创新团队计划资助项目(IG201803N)

摘　　要：针对地铁低速行驶时的轨道振动信号,分析小波去噪参数的选取对去噪效果的影响。以北京某地铁站附近的隧道断面的轨道铅垂向振动加速度信号为例,对轨道振动信号进行小波分解、阈值去噪和重构。以信噪比作为去噪质量评价指标,采用Kruskal-Wallis非参数检验方法分析小波系、小波阶数、小波分解层数和阈值的选取对去噪效果的影响。基于信噪比最大化原则,确定地铁低速行驶时轨道振动信号的最优小波去噪参数。结果表明:显著性水平0.05下,小波系、小波阶数的选取对地铁轨道振动信号小波去噪效果没有显著影响,不同的小波分解层数、阈值对地铁轨道振动信号小波去噪效果的影响不全相同;以Symlets A小波系的4阶小波为去噪小波基,经过3层小波分解,基于无偏似然估计阈值和软阈值函数进行小波去噪后的信号获得最大的信噪比,去噪后的振动信号保留了原始信号的峰值特征,同时信号曲线的光滑性有了显著改善。According to the track vibration acceleration signal of metro with the train traveling at low speed, the influence of selection of wavelet de-noising parameters on the de-noising effect is analyzed. With the track acceleration test data of a section of a metro line in Beijing as an example, wavelet decomposition, threshold de-noising and reconstruction of track vibration signal are carried out. The signal-to-noise ratio is used as the evaluation index of the de-noising quality. Kruskal-Wallis non-parametric test method is used to analyze the influences of wavelet series, wavelet order, wavelet decomposition layer number and threshold selections on the de-noising effect. With the maximum signal-to-noise ratio as the target, the best wavelet de-noising parameters of the metro track vibration signal is determined. The results show that the selection of wavelet series and wavelet order has no significant influences on the wavelet de-noising effect of the metro track vibration signal, and the influences of different wavelet decomposition layer number and the threshold on the wavelet de-noising effect of the metro track vibration signal are different. The maximum signal-to-noise ratio is obtained based on sym4 wavelet base, Stein ' s unbiased risk threshold and soft threshold function. The signal after de-noising preserves the peak characteristic of the original signal, and the smoothness of the signal curve has been greatly improved.

关 键 词：振动与波 地铁轨道 低速行驶 小波去噪 参数研究 

分 类 号：TB533[理学—物理] TN911.7[理学—声学]