飞行器液压舵面作动机构故障诊断研究

Research on Fault Diagnosis of Hydraulic Rudder Turning Shaft Machine Based on Wavelet and SVM

针对液压舵面伺服控制机构的结构部分的故障,找到一种较为成熟,可供工程实用的理论,为此提出采用小波包和支持向量机相结合的诊断方法;为了验证诊断方法的有效性,搭设物理试验平台模拟液压舵面伺服控制系统,通过试验手段,对舵面机构的转轴结构损伤特征以及对轴承卡阻故障特征应用小波包分解和计算小波包能量的方法进行有效的特征提取,针对结构损伤和故障识别中实际样本较少的问题,提出了基于支持向量机的识别诊断方法,结果显示该方法实用有效。

This research is mainly about an analysis of effective characteristic pick--up of hydraulic rudder turning shaft＇s fault. Two fault are set for research artificially, damage of turning shaft and turning prevented fault of bearing. Wavele＇t Analysis and Empirical Mode Decomposition are two pop methods for signal analysis at present. In this thesis, the two methods are used for the issue of fault diagnosis. Fault signal samples are rare ones in practicality for a complex system such as an aircraft. Support Vector Machine （SVM） is a good measure for rare samples＇ identification.