基于希尔伯特黄变换的刀具磨损特征提取

Tool wear feature extraction based on Hilbert-Huang transformation

概述了希尔伯特黄变换(HHT)的基本理论和算法,对信号经过经验模态分解(EMD)后得到的固有模态函数(IMF)求取振幅均值,差值筛选出与刀具磨损相关的IMF分量,并对单分量固有模态函数求取边际谱,获取边际谱最大幅值点,建立他们与刀具磨损之间的映射关系,进行特征提取,将其作为神经网络的输入特征向量,结合希尔伯特三维时频谱进行刀具磨损状态的判断。研究结果证明,该方法可以作为刀具磨损监测中信号特征提取的一种简单和可靠的方法。

After presenting the basic theory and algorithm of Hilbert-Huang transformation （ HHT）, a tool signal was decomposed with the empirical mode decomposition （EMD） method and its intrinsic mode functions （IMFs） were gained to obtain their average amplitude. The IMF components related to tool wear were chosen using a difference screen. Meanwhile, the marginal spectrum of a single intrinsic mode function was obtained and its maximum amplitude point was then found. By establishing the mapping relationship between maximum amplitude points and tool wear, the features of tool wear were extracted. Regarding them as input eigen-vectors of a neural network, and combined with Hilbert spectra, the tool wear status judgment was implemented. The study results showed that this approach is a simple and reliable method to detect the level of tool wear.