Feature extraction on connecting rod bearing fault based on EMD and Gabor transformation

For the non-stationary characteristics of diesel vibration signals, a method of connecting rod bearing based on EMD and Gabor transformation is proposed for fault feature extraction. The EMD method was used to decompose engine acceleration vibration signals into multiple empirical mode components > than using Gabor transformation to the first to the fourth correlation IMF. Effective fault features were extracted from the curves of the frequency band energy accumulation of the time/frequency distribution. The experiment indicates that this method can effectively extract the faults feature of diesel connecting rod bearing.