齿轮箱振动信号的多小波阈值去噪方法研究  

Research on Multiple Wavelet Threshold Denoising Methodfor Gearbox Vibration Signal

作　　者：王旭升[1] 吴志光[1] WANG Xu-sheng;WU Zhi-guang(Department of Electrical and Mechanical Services,Chizhou Vocational and Technical College,Chizhou Anhui 247000,China)

机构地区：[1]池州职业技术学院机电与汽车系,安徽池州247000

出　　处：《佳木斯大学学报（自然科学版）》2020年第5期89-92,共4页Journal of Jiamusi University：Natural Science Edition

基　　金：安徽省2019年度高校优秀青年骨干人才国内访问研修项目(gxgnfx2019131);安徽省高校2018年度重点自然科学研究项目(KJ2018A0187);池州职业技术学院科研自然重点项目(ZR2018Z01)。

摘　　要：人们在获取振动信号的过程中会受到噪声的影响,为了能够从振动信号中提取特征信息,就必须要消除振动信号中的背景噪声。对振动信号采用GHM多小波自适应阈值去噪方法,即根据各层高频系数的噪声水平,估计噪声的标准差,利用每层高频系数的能量比自适应的选择阈值,避免了对每层高频系数使用相同阈值的缺点。最后,选用获取的齿轮箱振动信号进行降噪实验,对比Db4小波固定阈值降噪方法,证实了该方法在振动信号去噪中的优势。People will be affected by noise in the process of obtaining vibration signals.In order to be able to extract feature information from vibration signals,we must eliminate the background noise in vibration signals.A GHM multi-wavelet adaptive threshold denoising method is used for vibration signals,that is,the standard deviation of noise is estimated according to the noise level of the high frequency coefficients of each layer.Using the energy ratio adaptive selection threshold of each layer high frequency coefficient,the disadvantage of using the same threshold for each layer high frequency coefficient is avoided.Finally,the vibration signal of the gear box is selected for the noise reduction experiment,and the Db4 wavelet fixed threshold noise reduction method is compared,which proves the advantage of the method in the vibration signal denoising.

关 键 词：振动信号 多小波变换 自适应阈值 阈值去噪 

分 类 号：TN911[电子电信—通信与信息系统]