一种WTMSST结合自适应参数VMD的滚动轴承故障诊断

Fault diagnosis of rolling bearings based on WTMSST and adaptive parameter VMD

针对现有轴承故障诊断中常用的时频分析方法存在变换系数在时频平面上分布相对离散,时频谱能量模糊等共性问题,提出了一种基于小波变换的时间重分配多同步压缩变换(WTMSST)结合经蜣螂算法(DBO)优化变分模态分解(VMD)的滚动轴承故障诊断方法。该方法采用了一种根据重分配同步压缩变换(WTSST)优化的WTMSST算法,通过固定点迭代减少了在强频率变化下的群延迟,然后通过以最小包络熵为适应度的DBO算法优化VMD输入参数,根据峭度重构信号后,运用WTMSST进行故障特征提取。采用凯斯西储大学的数据集进行测试,验证该法准确描述了信号的冲击特征,并证明其较过往处理方式具有更好的性能。

To address the common problems in the conventional time-frequency analysis methods used in current bearing fault diagnosis,such as relatively discrete transform coefficient distribution on the time-frequency plane and blurry energy in the time-frequency spectrum,this paper proposes a rolling bearing fault diagnosis method based on wavelet transform modulated synchronous squeezing transform(WTMSST)in conjunction with variational mode decomposition(VMD)optimized by dung beetle optimizer(DBO).First,the method adopts a WTMSST algorithm optimized by the weighted time-synchronous squeezing transform(WTSST),reducing the group delay under strong frequency changes through fixed-point iteration.Then,by employing the smallest envelope entropy as the fitness function,the DBO algorithm optimizes the input parameters of VMD.Following the reconstruction of the signal based on kurtosis,the WTMSST time-frequency analysis method is employed for fault feature extraction.Experiments are conducted using the Case Western Reserve University data set.Tests are conducted using the data set of Case Western Reserve University.Our results show the method accurately describes the impact characteristics of the signal and performs better than the previous processing methods.