Feature Extraction of Helicopter Fault Signal Based on Wavelet Packet Analysis

Abstract

Feature extraction is one of the key links of helicopter fault diagnosis. Wavelet analysis is a new time-frequency analysis method. Wavelet packet analysis is a more precise orthogonal decomposition method based on wavelet analysis. It makes multi-level frequency band division for signals in the whole frequency band. In this paper, based on wavelet packet analysis, the signal with fault information is decomposed into different frequency bands, then the signal in each frequency band is analyzed statistically, and the energy value of the frequency band containing the fault information of the component to be diagnosed is formed as the characteristic index for fault diagnosis. Simulation results show the effectiveness of this method.