基于K-SVD联合参数自适应TQWT的齿轮箱故障诊断  

Gearbox Fault Diagnosis Based on K-SVD Joint Parameter Adaptive TQWT

作　　者：刘庆友 娄志宁 赵新维 LIU Qingyou;LOU Zhining;ZHAO Xinwei(School of Mechanical Engineering,Jiangnan University,Wuxi,Jiangsu 214122,China)

机构地区：[1]江南大学机械工程学院,江苏无锡214122

出　　处：《轻工机械》2024年第6期65-72,共8页Light Industry Machinery

摘　　要：针对齿轮箱信号在强噪声背景下故障特征提取难的问题,课题组提出了一种K-均值奇异值分解(K-means singular value decomposition,K-SVD)联合参数自适应可调品质因子小波变换(tunable Q-factor wavelet transform,TQWT)的齿轮箱故障诊断方法。利用K-SVD稀疏表示齿轮箱故障信号,重构信号后去除噪声;针对TQWT对噪声鲁棒性不强且参数过度依赖人为选择的问题,结合齿轮箱早期故障信号的冲击性与周期性特征,提出了自相关峭谱积指标,以自相关峭谱积为优化指标对TQWT参数进行自适应选择;根据自相关峭谱积指标对子带进行筛选,对选出的子带进行重构,通过包络谱分析得到齿轮箱故障特征信息。仿真和试验结果表明所提出的诊断方法能有效提取低转速、强噪声背景下的齿轮箱故障特征。A gear box fault diagnosis method based on K-means singular value decomposition(K-SVD)combined with parameter adaptive tunable Q-factor wavelet transform(TQWT)is proposed to address the difficulty of extracting fault features from gear box signals in strong noise backgrounds.Using K-SVD to sparsely represent gearbox fault signals,reconstructing the signals and removing noise,in response to the problem of weak noise robustness and excessive reliance on manual parameter selection in TQWT,combined with the impact and periodic characteristics of early gearbox fault signals,an autocorrelation kurtosis product index is proposed.The autocorrelation kurtosis product is used as the optimization index to adaptively select TQWT parameters.Based on the autocorrelation kurtosis product index,the selected sub bands are screened,reconstructed,and the fault feature information of the gearbox is obtained through envelope spectrum analysis.The simulation and experimental results show that the proposed diagnostic method can effectively extract the fault characteristics of gearbox under low speed and strong noise background.

关 键 词：故障诊断 齿轮箱 K-均值奇异值分解 可调品质因子小波变换 自相关峭谱积 

分 类 号：TH165.3[机械工程—机械制造及自动化] TH133.3