摘要
对经验模式分解(EMD,Empirical Mode Decomposition)的滤波器特性进行了研究。事实上,EMD的倍频程滤波器特性在几年前就被发现了,在此基础上,针对EMD滤波器的其他特性展开了研究。首先,对EMD滤波器的自适应特性展开了研究,其自适应特性除了表现为倍频程滤波器的形式,还有两个表现:其一是滤波器的上限频率随信号自身的上限频率而定,另一表现是每一个IMF滤波器的形式不仅与信号频率有关,还与信号的结构有关。利用EMD的自适应特性,建立了一种新的信号表示方法,与已有的EMD方法相比,该方法可以提供更多的关于信号的信息。还对该方法进行了相关的讨论。
The filtering properties of empirical mode decomposition （EMD） are investigated in this paper. The property that EMD acts as an octave filter bank has been discovered a few years before. Based on that, some other properties of EMD are investigated in this paper. First of all, the adaptability of EMD is investigated. Besides the octave filtering property, the adaptability lies in two points. One is the upper bound frequency of the filter bank depends on its own upper bound frequency. The other is that the filter type of each IMF not only depends on the frequency spectrum of the analyzed signal, but also on the temporal structural of the signal. Then, the adaptability of EMD is utilized to build a new representation for signals, which can provide more detailed information for signals than original EMD does. The concerned Discussions are given on this method.
出处
《洛阳理工学院学报（自然科学版）》
2008年第1期5-8,52,共5页
Journal of Luoyang Institute of Science and Technology：Natural Science Edition
基金
国家自然科学基金(10674149).
关键词
经验模式分解
故障诊断
特征提取
滚动轴承
Empirical Mode decomposition
Fault diagnosis
Feature extraction
Roller bearing
