基于EMD与关联维数的故障诊断AR模型  

Application of correlation dimension and EMD-based AR model in the fault diagnosis

作　　者：程军圣[1] 于德介[1] 唐驾时[2] 杨宇[1] 

机构地区：[1]湖南大学机械与汽车工程学院,湖南长沙410082 [2]湖南大学力学与航空航天学院,湖南长沙410082

出　　处：《系统工程与电子技术》2007年第9期1589-1592,共4页Systems Engineering and Electronics

基　　金：国家自然科学基金(50275050);湖南省自然科学基金(05JJ40079);中国博士后科学基金资助课题

摘　　要：提出了一种基于EMD和关联维数的转子系统故障诊断方法。该方法用EMD将转子系统振动信号分解成若干个基本模式分量IMF,对包含主要故障信息的IMF分量建立AR模型,AR模型自回归参数的关联维数作为特征量神经网络识别转子系统的工作状态和故障类型。对实验数据分析的结果表明,该方法能有效地应用于转子系统的故障诊断。A fault diagnosis approach for rotor systems based on correlation dimension is proposed. The EMD method is used to decompose the rotor vibration signal into several intrinsic mode functions （IMFs）. The AR models of some IMF components which contain main fault information are constructed. Finally, the correlation dimensions of auto-regressive parameters in AR model are calculated and input into the neural network as feature vectors. The working state and fault pattern of the rotor system can be identified by the neural network. The analysis results of the experimental data show that the proposed approach can be applied to the fault diagnosis for rotor system efficiently.

关 键 词：经验模态分解方法 自回归模型 关联维数 转子系统 故障诊断 神经网络 

分 类 号：TH113[机械工程—机械设计及理论]