基于EMD和分形维数的转子系统故障诊断  

Fault Diagnosis for Rotor System Based on EMD and Fractal Dimension

作　　者：程军圣[1] 于德介[1] 杨宇[1] 

机构地区：[1]湖南大学,长沙410082

出　　处：《中国机械工程》2005年第12期1088-1091,共4页China Mechanical Engineering

基　　金：国家自然科学基金资助项目(50275050);高等学校博士学科点专项科研基金资助项目(20020532024)

摘　　要：提出了一种基于EMD方法和分形维数的转子系统故障诊断方法。利用EMD方法将转子振动信号进行分解,得到若干个基本模式分量,然后将包含主要故障信息的几个基本模式分量相加得到降噪后的转子振动信号,求得降噪后的转子振动信号的分形维数。试验数据的分析结果表明,在不同的故障状态下,采用EMD方法对转子振动信号降噪后求得的分形维数是不同的,从而可以通过分形维数的大小有效地判断转子系统的工作状态和故障类型。A fault diagnosis approach for rotor system based on EMD (Empirical Mode Decomposition) and fractal dimension was proposed. Firstly, EMD was used to decompose the rotor vibration signal into a number of intrinsic mode functions (IMFs). then, the sum of some IMF components which contain main fault information was computed to obtain the noise-reducing rotor vibration signals. Finally, the fractal dimension of the noise-reducing rotor vibration signal under proper embed dimension is computed. The experimental results show that the fractal dimensions which are obtained after EMD is used to reduce the noise of the rotor vibration signals are different under different fault conditions and can be used to identify the work state and fault pattern for the rotor system efficiently.

关 键 词：EMD方法 分形维数 转子系统 故障诊断 

分 类 号：TH165[机械工程—机械制造及自动化]