基于VMD-SVD能量标准谱-Teager能量算子的轴承故障诊断方法

Bearing Fault Diagnosis Method based on VMD-SVD Energy Standard Spectrum and Teager Energy Operator

对难以提取处于微弱故障状态的滚动轴承非线性、非平稳时变特性振动信号中故障特征频率的问题,提出基于VMD-SVD能量标准谱-Teager能量算子联合诊断方法。首先,对预处理后轴承微弱故障信号进行VMD分解,根据各模态分量(IMF)中心频率确定最优模态数K,再由各IMF分量峭度和相关系数指标确定包含故障信号的敏感IMF。然后,对选取模态分量的Hankel矩阵进行SVD分解,由奇异值能量标准谱确定有效奇异值数量,实现对信号的降噪重构。最后,利用瞬时Teager能量算子及其频谱分析识别微弱故障产生的周期性冲击特征频率。运用该方法处理滚动轴承微弱故障信号,能准确提取故障特征频率及倍频,文中证明了其准确性和有效性。

To solve the problem that it is difficult to extract the fault characteristic frequencies from vibration signals with non-linear and non-stationary time-varying characteristics for rolling bearings in weak fault state, a combined diagnosis method based on VMD-SVD energy standard spectrum and Teager energy operator is proposed. Firstly, VMD of the pretreated weak bearing fault signal is carried out, the optimal mode number K is determined according to the central frequency of each IMF component, and the sensitive IMF containing fault signal is selected by the kurtosis and correlation coefficient. Then, the Hankel matrix of the selected IMF is decomposed by SVD, and the number of effective singular values is determined by the standard spectrum of the singular value energy. And the de-noising and reconstruction of the signal are realized. Finally, the instantaneous Teager energy operator and its spectrum are used to extract the characteristic frequencies of the periodic shocks caused by the faults. Through this method, the weak fault signal of the rolling bearing can be processed, the fault characteristic frequency and its octave can be accurately extracted, which proves the accuracy and effectiveness of this method.