The Research On Hilbert-Huang Transform And Its Mode Mixing Problem

Hilbert-Huang Transform (HHT) is a newly developed signal analysis method.Its most important innovations are the Intrisinc Mode Function (IMF) and EmpricalMode Decomposition (EMD), which gives the instantaneous frequency clearphysical significance. The researche has been made on the theory of HHT and itsMode Mixing problem in this paper.Firstly, the conception of EMD and Hilbert Transform (HT) are introduced indetail, with instantaneous frequency as the main line. Through the numericalsimulation examples, the advantage of using HHT to deal with nonstationary andnonlinear signal is illustrated, and some problems of HHT are indicated too.Secondly, the method of identifying the vibration modal parameters with HHTis introduced. Because the mathematic basis of HHT has not been estabelished, theaccuracy of modal parameters identification will be affected by its own defects.Aimed at mode mixing and end effects, the approcah of using bandpass filter tomode mixing is recommended. The empirical formula of cut-off frequenciesselecting and the criterion of dataset selecting have been given, whose effectivenesshas been domenstrated by numerical simulation result.Finally, the mode mixing of EMD has been deeply researched. Based onsummarizing the reasons which cause mode mixing and the solutions now available,the mathematic mode of signal has been extended, combined with theory system ofHHT and the characteristics of actual engineering signal. The mode mixing ischanged from a global conception to a local one, considered as Instantaneous ModeMixing. The feature of vibration mode parameters identification based oninstantaneous mode mixing is illustrated by the numarical simulation example.