Application of empirical mode decomposition on engine fault feature extraction
Duan Chen-dong · Journal of Chang'an University · 2010
An empirical mode decomposition(EMD) method for analyzing crank shaft speed fluctuation signals and extracting the fault features of engine misfiring is investigated in order to solve the problem that an engine's operation will become so complicated that it is difficult to extract the fault features when the engine operates at a high speed.A sifting algorithm is applied in EMD to decompose a signal into some components based on their characteristic time scales,then the inner oscillation modes of different frequencies in the signal are revealed.A number of speed signals of a three-cylinder gasoline engine with the operation of low and intermediate speed,as well as normal and misfiring state are collected in some experiments,and then analyzed by using EMD.The results show that EMD can separate the high frequency components of the speed signals and extract the misfiring fault features of the engine effectively.This method is able to be applied as a pre-processing for artificial intelligence fault diagnosis for engine misfiring.7 figs,9 refs.