Reliability Evaluation of Rolling Bearing Based on Wavelet Packet Energy Entropy

Based on the time-domain analysis of bearing vibration data, the root-mean-square is selected from the 14 time domain indicators as the prediction index and the key threshold is set according to the image. According to the wavelet packet decomposition and reconstruction theory, the vibration signal analysis of bearing is carried out. In this way, the characteristic signals of each frequency band are extracted, and the change of bearing band energy in outer ring wear is analyzed by combining with the wavelet packet energy spectrum. Finally, from the perspective of information entropy and integrated with the state parameter method, the running state of rolling bearing is evaluated and predicted. Compared to the traditional threshold method, the method used in this paper can judge the running state of bearing more effectively. And through the wavelet packet information entropy, the bearing running state is divided into four levels, which can make a more detailed and reasonable evaluation of the bearing’s health and reliability. This provides a basis for reliability assessment in areas such as machining.