Research on a vibration signal analysis method for motor bearing

Because the traditional signal analysis methods are unsatisfactory for analyzing the non smooth and strange signal in the high-frequency band. So the vibration signal of motor bearing is selected as the research object, the wavelet transform and approximate entropy are introduced into the vibration signal analysis in order to propose a new vibration signal analysis(TFTFWTAE) method in this paper. In the proposed TFTFWTAE method, the improved threshold function is used to deal with the wavelet coefficients in order to obtain the improved wavelet transform method, which is used to decompose the original vibration signal into multilayer in order to obtain the feature vector of vibration signal energy. Next the approximate entropy is used to judge and display the complexity and irregular of vibration signal energy in the different scale and different frequency band in order to extract the non-stationary characteristics of vibration signal. Finally, the experiment of motor bearing are tested in order to prove the effectiveness of proposed TFTFWTAE method. The experiment results show that the TFTFWTAE method has strong ability to describe the complexity of the vibration signal and effectively judge the running state of motor bearing. It can effectively extract the feature vector from the original vibration signal. (C) 2016 Elsevier GmbH. All rights reserved.