LMD-KICA滚动轴承信号特征提取方法研究

Research on Signal Feature Extraction Method of LMD-KICA Rolling Bearing

在实际工程机械所产生的滚动轴承故障信号十分复杂,有效的故障信号和特征信息极易被噪声信号所干扰,针对工程中振动信号的特点,提出了一种基于局部均值分解(LMD)结合核独立分量分析(KICA)的方法提取故障信号特征。首先将源信号进行LMD分解,根据相关系数准则提取相关程度高的PF分量并构建新信号,对新构建的故障信号运用KICA进行噪声分离,进而获得故障信号特征。先通过构造信号仿真分析方法有效性,再通过西储大学轴承实验数据对比分析,验证该方法适用于提取滚动轴承的故障特征。

The rolling bearing fault signal generated by the actual engineering machinery is very complicated,and the effective fault signal and characteristic information are easily interfered by the noise signal.According to the characteristics of the vibration signal in engineering,a method based on Local Mean Decomposition(LMD)and Kernel Independent Component Analysis(KICA)is proposed to extracts fault signal characteristics.Firstly,the source signal is decomposed by LMD,and the PF component with high correlation is extracted according to the correlation coefficient criterion to construct a new signal.The KICA is used to separate noise from newly constructed fault signal,and then the fault signal characteristics are obtained.Firstly,the validity of the method is analyzed by constructing the signal.Then,through the comparison and analysis of the bearing data of the Western Reserve University,the method is validated to extract the fault characteristics of the rolling bearing.