航天装备机械元件早期故障特征提取技术研究  

Research on early fault feature extraction of mechanical components in space equipment

作　　者：马伦 陈铭 孙丽慧 刘宁宁 吕艳军 MA Lun;CHEN Ming;SUN Lihui;LIU Ningning;LYU Yanjun(School of Non-Commissioned Officer,Space Engineering University,Beijing 102249,China)

机构地区：[1]航天工程大学士官学校,北京102249

出　　处：《航天工程大学学报》2024年第2期42-50,共9页

基　　金：试验技术研究项目(SY1800050395);军队重点学科建设项目(SZ1905)。

摘　　要：针对航天装备机械元件故障初期振动信号中的微弱特征成分难以检测的问题,结合Morlet小波变换降噪原理,在确定故障特征的滤波频带时,引入联合自信息量概念,提出一种定量选择最优尺度的方法。以最小Shannon熵优化小波的形状参数,实现母小波与故障特征的最佳匹配;将不同变换尺度下的小波系数绘制尺度-能量谱,利用故障特征能量聚集的特性,选择谱图中凸点对应尺度计算3个指标,即相关系数、小波系数能量熵和谱峭度,再比较以上指标的联合自信息量值确定最优尺度。案例应用表明:该方法能够从轴承数据中提取微弱故障特征,可为轴承早期故障特征提取提供参考。At the early fault stage of mechanical components in space equipment,it is hard to detect corresponding weakness from vibration data.Referring to de-noising principle through Morlet wavelet transform,a method to select the optimized scale determining the filtering frequency range based on joint self-information concept is presented.The main filtering procedure include as follows:one relates to optimizing shape factor controlled by the minimum Shannon entropy in order to reach the best match between mother wave and impact component;according to the accumulative character for fault characteristic power,the other is to select scale with the best filtering effect from the corresponding salient points in scale-power spectrum.Then,the optimized scale is determined by the joint self-information value from three indexes such as correlation coefficient,wavelet coefficient energy entropy and spectral kurtosis.Viewed from processing result for bearing vibration datasets,the proposed method can extract the weak feature component.So the reference for early fault feature extraction of bearings is provided.

关 键 词：联合自信息量 Morlet小波变换 最优尺度参数 Shannon熵 尺度-能量谱 

分 类 号：TH165.3[机械工程—机械制造及自动化] TN911.2[电子电信—通信与信息系统]