基于改进经验模态分解的HHT密集模态识别方法

HHT dense mode identification method based on improved empirical mode decomposition

针对传统HHT方法不能有效识别密集模态的问题,提出基于改进经验模态分解(EMD)的HHT密集模态识别方法。EMD密频信号分解能力不足是限制HHT法识别密集模态的主要原因,因此在EMD分解过程中嵌入信号调频(FM)和模态解相关操作提升其分解密频信号的能力,称改进后的方法为调频—解相关模态分解(FM-DEMD)。以FM-DEMD分解取代传统HHT法中的EMD分解,得到改进HHT模态识别方法。仿真实验证明,传统HHT法和ITD法密集模态识别失效时,改进HHT法仍能准确地识别密集模态信息。

In view of traditional HHT method which can＇t effectively identify dense modes,this paper proposed a modified HHT method based on improved empirical mode decomposition. The lack of decomposition ability of EMD was the main reason to limit HHT method to identify dense modes. Therefore,it would enhance the EMD decomposition capability by embedding signal frequency modulation（ FM） and mode decorrelation operation into EMD decomposition process,and the combined mode decomposition method was called FM-DEMD. By replacing EMD in the traditional HHT method with FM-DEMD,then it obtained the modified HHT method. Simulation results show that the modified HHT method can accurately identify dense modes even if traditional HHT and ITD fail.