Abstract: For gearbox fault feature extraction, a novel method based on ensemble empirical mode decomposition and Choi-Williams distribution for gearbox vibration signal extraction is proposed. Firstly, vibration data are decomposed into several intrinsic mode functions (IMF) with EEMD, and IMFs are sorted by kurtosis criterion, then CWD is applied to the selected IMF which kurtosis is larger than others, the Choi-Williams distribution features in time, frequency and amplitude domains of the original signal can be extracted. On the basis of discussing teeth break and crack vibration fault mechanism of gearbox, the proposed method is used to analyze the vibration signal of the actual fault gearbox. The result shows that this method can efficiently extract the fault information and have great importance for condition recognition and fault diagnosis of gearbox.

Key words: ordnance science and technology, gearbox, ensemble empirical mode decomposition, kurtosis, Choi-Williams distribution, feature extraction