A CEEMD Method for Diesel Engine Misfire Fault Diagnosis based on Vibration Signals

Aiming at the characteristics of diesel engine fault vibration signals which are generally nonlinear and non-stationary, and the difficulty in extracting fault frequencies, a diesel engine fault diagnosis method based on Complementary Ensemble Empirical Mode Decomposition (CEEMD) and Least Square Support Vector Machine (LSSVM) was proposed. CEEMD was used to decompose the original signals, and a number of inherent mode functions (IMF) were obtained. The IMF components were screened by the correlation coefficient method. In order to extract features from vibration signals, we made normalized energy as the features which were inputted into LSSVM for training and testing. Finally we realize the identification and diagnosis of diesel engine misfire fault.