基于局部均值分解的边际谱在滚动轴承故障诊断中的应用

局部均值分解（Local Mean Decomposition,简称LMD）将复杂的多分量信号自适应地分解为有限个乘积函数（PF）的和，在计算了各个分量的瞬时幅值（IA）和瞬时频率（IF）后，可以计算出基于LMD的边际谱。针对直接法求取瞬时频率存在端点误差大问题，提出了一种改进的直接求取瞬时频率的方法；提出了基于LMD的边际谱的滚动轴承故障诊断方法，将该方法应用于实际滚动轴承故障诊断中，结果表明该方法能有效地提取出滚动轴承的故障特征频率，从而确定故障部位。

Application of marginal spectrum based on local mean decomposition in rolling bearing fault diagnosis

Local mean decomposition（LMD）can decompose complex multi-component signal into a linear combination of a finite set of product functions(PFs), and after obtaining the instantaneous amplitudes and instantaneous frequencies of all PF components the marginal spectrum based on LMD can be calculated. Aiming at the big end-point error problem of the direct instantaneous frequency extraction method, an improved direct method was put forward. The new rolling bearing fault diagnosis method named marginal spectrum based on LMD was proposed, and was applied in actual rolling bearing fault diagnosis. The analysis results show that the fault characteristic frequency can be extracted effectively, and the fault position can be determined.