桥梁健康监测采样信号EMD小波相关降噪研究  

Study on EMD Wavelet Correlation De-noising of Bridge Health Monitoring Sampling Signals

作　　者：严鹏 YAN Peng(Sichuan Railway Development Co.,Ltd.,Southwest Jiaotong University,Chengdu 610073,China)

机构地区：[1]四川西南交大铁路发展股份有限公司

出　　处：《噪声与振动控制》2019年第3期204-209,共6页Noise and Vibration Control

摘　　要：为了进一步降低桥梁健康监测采样信号的噪声水平,根据其信号特征,在传统EMD小波阈值降噪算法的基础上,提出一种改进的降噪算法,称为EMD小波相关降噪算法。该算法综合了EMD、小波变换和相关检测3种方法的优点,首先对各阶IMF小波降噪前后的分量分别进行整体和局部相关检测,并以前3阶整体相关系数的均值作为降噪阈值,最后进行局部相关阈值降噪,得到降噪后的信号。将该算法与EMD小波阈值降噪和小波默认阈值降噪算法进行数值仿真和有限元仿真试验对比。结果表明,提出的EMD小波相关降噪算法具有更好的降噪效果,能够用于桥梁健康监测采样信号降噪处理。In order to further reduce the noise level of bridge health monitoring sampling signals,an improved denosing algorithm named EMD wavelet correlation de-nosing algorithm is proposed on the basis of traditional EMD wavelet threshold de-nosing algorithm.This improved algorithm combines the advantages of EMD,wavelet transform and correlation detection.Firstly,whole and local correlation detections are performed for the original IMF and the corresponding de-noised IMF.Then,the average of the first three whole correlation coefficients is chosen as the de-nosing threshold.Finally,the local correlation threshold de-noising is proceeded and the de-noised signal is obtained.The improved algorithm is compared with EMD wavelet threshold de-nosing and wavelet default threshold de-nosing algorithm through numerical simulation and finite element simulation.The results confirm that the proposed EMD wavelet correlation denosing algorithm has better de-nosing effect,and can be used in de-nosing processing of bridge health monitoring sampling signals.

关 键 词：振动与波 桥梁健康监测 经验模式分解(EMD) 小波变换 相关检测 降噪 

分 类 号：U446[建筑科学—桥梁与隧道工程]