中文摘要：

滚动轴承发生故障时，其振动信号往往包含多种振动分量，主要由轴承自身固有振动引起的谐振分量、点蚀或裂纹等故障产生的冲击分量和其他的干扰分量组成。实现故障信号中各分量的有效分离非常有利于轴承的故障诊断。针对此问题，提出形态分量分析和谱峭度相结合的故障诊断方法，首先用形态分量分析处理轴承故障信号，使信号中的冲击分量与谐振分量分离，再以谐振分量为对象，利用谱峭度方法对谐振分量进行滤波，对滤波结果进行Hilbert包络解调分析，然后根据包络谱诊断滚动轴承发生的故障。实验结果表明，这种方法可以提取到明显的故障特征频率，从而验证了该方法的有效性。

英文摘要：

When a rolling bearing is faulted, its vibration signal is usually composed of many components generated by different vibration sources. These components are dominated by the harmonic components induced by the natural vibration of the bearing, the impulsive components resulted from fitting or crack faults and some other perturbation components. It is helpful to separate these components from the original vibration signals effectively for fault diagnosis of the rolling bearing.In this paper, a new fault diagnosis method is proposed based on morphological component analysis (MCA) and fast spectral kurtosis (FSK). Firstly, the impulsive components and harmonic components are separated from the vibration signals of the faulted rolling bearing by using MCA. Then, the harmonic components are analyzed by using FSK filtering analysis. Finally,the signals obtained from the previous steps are analyzed by using Hilbert envelope demodulation analysis. The fault diagnosis of a rolling bearing is carried out according to the envelope spectrum method. The feasibility and effectiveness of this method are verified by the experiments.