基于SWT与IEWT的齿轮无转速计阶次跟踪  

Gear Tacho-less Order Tracking Method Based on SWT and IEWT

作　　者：刘奇 田辈辈 冷军发[2] 罗晨旭[2] 荆双喜[2] LIU Qi;TIAN Beibei;LENG Junfa;LUO Chenxu;JING Shuangxi(College of Mechanical and Electrical Engineering,Jiaozuo University,Jiaozuo Henan 454000,China;School of mechanical and power engineering,Henan Polytechnic University,Jiaozuo Henan 454000,China)

机构地区：[1]焦作大学机电工程学院,河南焦作454000 [2]河南理工大学机械与动力工程学院,河南焦作454000

出　　处：《机械设计与研究》2024年第2期151-157,共7页Machine Design And Research

基　　金：国家自然科学基金资助项目(U1804134);河南省科技攻关项目(222102220037,232102221038)。

摘　　要：针对变转速工况下,多级齿轮传动低速级齿轮故障信号易受背景噪声干扰,导致频谱特征模糊,微弱故障特征难以提取的问题,提出一种基于同步压缩小波变换(Synchrosqueezing Wavelet Transform, SWT)与改进经验小波变换(Improved Empirical Wavelet Transform, IEWT)相结合的齿轮无转速计阶次跟踪方法。首先为提高无转速计阶次跟踪瞬时频率估计精度,设计连续小波变换-椭圆时变滤波器(Continue Wavelet Transform-Elliptic Time-Varying Filtering, CWT-ETVF)对齿轮振动信号滤波降噪,依据滤波所得单分量的SWT时频分布进行峰值搜索,以实现高精度的瞬时频率估计,然后对时变故障信号等角度重采样获得角域平稳信号。针对EWT方法频谱分割不合理的问题,提出一种依据频谱包络趋势进行边界划分的改进经验小波变换方法对角域平稳信号自适应分解。最后选择合适分量自相关去噪,并通过阶次解调分析识别故障特征。仿真及实测局部断齿数据分析表明,该方法可以准确提取变转速齿轮时变微弱故障特征。Aiming at the problem that the low speed gear fault signal of multistage gear transmission is susceptible to background noise interference under variable rotational speed conditions,which makes it difficult to extract weak fault features,a gear tacho-less order tracking method based on synchrosqueezing wavelet transform(SWT)and improved empirical wavelet transform(IEWT)is proposed.Firstly,in order to improve the accuracy of order tracking instantaneous frequency estimation without a tachometer,a continuous wavelet transform-elliptic time varying filter is designed to filter and denoise the gear vibration signal,and then a peak search is performed based on the filtered single component SWT time frequency distribution to achieve high-precision instantaneous frequency estimation;Then,according to the IFE curve,the time-varying fault signal is resampled at equal angles to obtain the angular domain stationary signal.An improved empirical wavelet transform method for adaptive decomposition of stationary signals in diagonal domain is proposed to solve the problem of unreasonable spectral segmentation in empirical wavelet transform.Finally,appropriate components are selected for autocorrelation denoising,and fault features are identified through order demodulation analysis.Simulation and analysis of measured partial tooth breakage data show that the method can accurately extract low-frequency weak fault features of variable speed gearbox.

关 键 词：齿轮 同步压缩小波变换 改进经验小波变换 阶次跟踪 故障特征提取 

分 类 号：TH17[机械工程—机械制造及自动化] TH132.41