赵梦圆, 荆双喜, 冷军发, 等.CWT-ETVF与SWT结合的齿轮无转速计阶次跟踪及其应用[J].河南理工大学学报（自然科学版）,2023,42(2):98-107.

ZHAO M Y, JING S X, LENG J F, et al.Gear tacho-less order tracking method with CWT-ETVF and SWT and its application[J].Journal of Henan Polytechnic University(Natural Science) ,2023,42(2):98-107.

CWT-ETVF与SWT结合的齿轮无转速计阶次跟踪及其应用

赵梦圆, 荆双喜, 冷军发, 绳飘, 罗晨旭

河南理工大学 机械与动力工程学院，河南 焦作 454000

摘要:变转速齿轮故障振动信号特别微弱时，同步压缩小波变换（synchrosqueezing wavelet transform，SWT）无转速计阶次分析方法的提取效果不佳。基于此，提出一种连续小波变换的椭圆时变滤波（continuous wavelet transform-elliptic time-varying filtering，CWT-ETVF）与SWT相结合的无转速计阶次跟踪方法，用以提取齿轮时变低频故障特征。将CWT-ETVF与SWT结合对振动故障信号进行瞬时频率估计，以获得参考轴相位；再对原信号进行等角度重采样得到角域平稳信号，并作其阶次谱分析和SWT分解；最后，选取SWT重构分量进行阶次谱分析与阶次包络谱分析，以提取齿轮断齿的时变故障特征。仿真及实验结果验证了该方法在齿轮变转速工况下低频微弱故障特征提取的有效性。

关键词:特征提取;连续小波变换;椭圆时变滤波;同步压缩小波变换;阶次跟踪

doi:10.16186/j.cnki.1673-9787.2021070065

基金项目:国家自然科学基金资助项目（51775174，U1804134）；河南省科技攻关项目（222102220037，222102210210）；河南省高等学校重点科研项目（19A440007）；河南理工大学博士基金资助项目（B2017-28）

收稿日期:2021/07/17

修回日期:2021/09/18

出版日期:2023/03/25

Gear tacho-less order tracking method with CWT-ETVF and SWT and its application

ZHAO Mengyuan, JING Shuangxi, LENG Junfa, SHENG Piao, LUO Chenxu

School of Machine and Power Engineering，Henan Polytechnic University，Jiaozuo 454000，Henan，China

Abstract:For the especially weak fault signal from the variable speed gear，the extraction effect of synchrosqueezing wavelet transform（SWT）tacho-less order tracking method is poor.To solve this problem，a tacho-less order tracking method was proposed to extract gear time-varying and low-frequency fault characteristic.This method combined the advantages of continuous wavelet transform-elliptic time-varying filtering（CWT-ETVF）and SWT.Based on CWT-ETVF and SWT，the instantaneous frequency of the vibration fault signal was estimated to obtain the phase of reference axis.Then，the time-varying signal was transformed by equal angle resampling into the angle domain stationary signal，and its order spectrum analysis and SWT decomposition were performed.Finally，the SWT reconstructed component was selected to analyze its order spectrum and order envelope spectrum，from which the time-varying fault characteristic of the fault gear with a missing tooth was extracted.The effectiveness of this method was demonstrated using simulation and experimental datasets，and it was very suitable for the low-frequency weak fault feature extraction under variable speed operation of gear.

Key words:feature extraction;continuous wavelet transform;elliptic time-varying filtering;synchrosqueezing wavelet transform;order tracking

 013_2021070065_赵梦圆_H.pdf