Fault diagnosis method based on minimum entropy deconvolution and fruit fly optimization algorithm

Abstract

Aiming at the problems of fault diagnosis for rotating machinery, this paper proposed a fault diagnosis method combining minimum entropy deconvolution (MED) with fruit fly optimization algorithm (FOA). In the MED method, the objective function method (OFM) is used to find the set of filter coefficients under the condition of maximal kurtosis. Given that the filter coefficients obtained by OFM are local optima not global optima and MED is difficult in parameter selection, FOA is applied instead. A filtered signal is obtained by FOA and MED, and envelope demodulation is carried on it for fault diagnosis. Results from rolling bearing fault simulation experimental system show that the proposed method has better noise reduction performance and is able to extract fault features of rolling bearings, and it is better adapted to engineering applications as compared with prior MED method.