Jixie qiangdu (Jan 2022)
RESEARCH ABOUT FAULT DIAGNOSIS OF BEARING BASED ON INSTRINSIC TIME SCALE DECOMPOSITION AND CONVOLUTIONAL NEURAL NETWORK
Abstract
The working environment of rolling bearing is complex, the nonlinear vibration signal and the interference of environmental noise lead to the difficulty of fault diagnosis. Therefore, based on the experimental data of bearing damage and the fractal theory, the Intrinsic Time scale Decomposition（ITD） was used to extract the nonlinear features of vibration signals, and the effective fault feature components were selected. The intelligent fault diagnosis of bearings was realized through Convolutional Neural Network（CNN）. The results show that compared with the existing methods, ITD-CNN has higher accuracy under different SNR. At-4 dB signal to noise ratio, the accuracy is still 2.57%～13.35% higher than the existing methods, which indicates that the proposed method has good recognition ability and generalization performance.
