基于改进变分模态分解和快速谱峭度图的滚动轴承检测方法  

A Rolling Bearing Detection Method Based on Fast Spectral Kurtogram and Improved Variational Mode Decomposition

作　　者：何凯 廖玉松[1] 胡斌 谭邦俊 HE Kai;LIAO Yusong;HU Bin;TAN Bangjun(Chuzhou Vacationl and Technical College,Chuzhou 239000,Anhui,China)

机构地区：[1]滁州职业技术学院,安徽滁州239000

出　　处：《噪声与振动控制》2020年第2期121-124,197,共5页Noise and Vibration Control

基　　金：安徽省高校学科(专业)拔尖人才学术重点资助项目(gxbjZD51)。

摘　　要：针对滚动轴承早期故障信号微弱、复杂且提取困难的问题,提出一种基于改进变分模态分解(Variational Mode Decomposition,简称VMD)和快速谱峭度图的滚动轴承检测方法。首先利用粒子群算法对VMD最佳影响参数组合进行搜索,采用多尺度模糊熵(Multiscal Enproty,简称MSE)作为适应度函数,并利用优化参数的VMD对原始信号进行分解,得到多个本征模态分量(Intrinsic Mode Function,简称IMF);其次计算原始信号和各模态分量的快速峭度图;再次找出原始信号和各个IMF谱峭度最大值所处的频带区间;然后通过比较原始信号和IMF谱峭度最大值所处频带区间的从属关系来选择最佳IMF;最后,重组最佳IMF并通过共振解调技术求其包络谱图。实验结果表明基于改进变分模态分解和快速谱峭度图的滚动轴承检测方法能更有效诊断出滚动轴承的早期故障。Considering the weakness and complexity of early fault signals of rolling bearings and difficultly of extracting the fault features from these signals,a method for rolling bearing detection based on fast spectral kurtogram and improved variational mode decomposition(VMD)is proposed.Firstly,the particle swarm optimization algorithm is used to search the VMD optimal influence parameters.With the Multiscale enproty(MSE)used as a fitness function,the original signal is decomposed by using the VMD with the optimized parameters.And the multiple intrinsic mode functions(IMF)are obtained.Secondly,the fast kurtogram between the original vibration signal and IMFs are calculated.Then,the frequency band between the original signal and the maximum spectrum kurtosis of each IMF are searched.By comparing the affiliation between the original signal and the reference band of each IMF,the sensitive IMFs can be determined.Finally,the optimal IMFs are reconstructed and their envelope spectrum is obtained by using resonance demodulation technology.The proposed method is applied to analyze the experimental data of a rolling bearing with early faults.The analysis results indicate that the proposed method can recognize the early faults of rolling bearings more effectively.

关 键 词：故障诊断 滚动轴承 快速谱峭度图 改进变分模态分解 本征模态分量 

分 类 号：TH133.3[机械工程—机械制造及自动化]