Fault Diagnosis in Printing Press Roller Bearing Based on Spectrum Kurtosis and Improved EEMD

Abstract

According to the problem of the poor rolling bearing signal under the high background noise of the high-speed printing press, the fault diagnosis method is proposed, which is combined with Ensemble Empirical Mode Decomposition (EEMD) and spectrum kurtosis (SK). Due to the traditional EEMD has the defect of directly rejecting the high-frequency section impact fault signal, this paper adopts the three principles of EEMD noise reduction to carry on the improvement processing. The fault vibration signal of rolling bearing was decomposed into several intrinsic mode functions (IMF) under different characteristic scales with EEMD. The false signal of IMF is removed by the three principles of noise reduction to construct a new fault signal. In this paper, the bandwidth and central frequency of band-pass filter are selected by SK method and the envelope spectrum of the filtered signal is analyzed. The result shows that the combination of improved EEMD and SK method are effective than the method of conventional EEMD and envelope analysis. It promotes fault recognition rate of printing press bearings.