基于AR模型和支持向量机的转子系统故障诊断方法  

Fault Diagnosis Approach For Rotor Systems based on AR Model and Support Vector Machine

作　　者：于德介[1] 陈淼峰[1] 程军圣[1] 杨宇[1] 

机构地区：[1]湖南大学机械与汽车工程学院,长沙410082

出　　处：《系统工程理论与实践》2007年第5期152-157,共6页Systems Engineering-Theory & Practice

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

摘　　要：提出了基于AR模型和支持向量机的转子系统故障诊断方法.该方法对转子系统的振动信号建立AR模型,以AR模型主要的自回归参数和残差的方差作为特征向量,然后建立支持向量机分类器,进而判断转子系统的工作状态和故障类型.实验结果分析表明,该方法能有效地应用于转子系统的故障诊断.并通过支持向量机与BP神经网络的性能比较,说明了支持向量机的优点.A fault diagnosis approach for rotor systems based on AR model and support vector machine is proposed. Firstly, the AR model of the vibration signal of a rotor system is established. The main auto-regressive parameters and the variances of remnant are regarded as the feature vectors. Then, the support vector machines used as fault classifiers are established to identify the condition and fault pattern of a rotor system. Practical examples demonstrate that the approach based on AR model and support vector machine can be applied to the rotor system fault diagnosis effectively. In comparison, the performances of support vector machine are more excellent than BP neural network.

关 键 词：AR模型 支持向量机 故障诊断 转子系统 

分 类 号：TH113[机械工程—机械设计及理论]