Rolling bearing fault diagnosis method based on multistage abstract time feature fusion

The invention provides a rolling bearing fault diagnosis method based on multistage abstract time feature fusion, and the algorithm can extract multistage abstract features from a bearing vibration signal and fuse the multistage abstract features, and capture time dependence hidden in the signal, thereby overcoming the defect that most methods at present are not comprehensive enough in feature extraction, and improving the fault diagnosis accuracy. Dependence on fault diagnosis priori knowledge in a feature extraction stage is eliminated, and generalization ability of the fault diagnosis method is improved. According to the fault diagnosis method, the fault damage degree of the bearing can be accurately identified and accurately positioned, good load adaptability and anti-interference capability are achieved, and bearing fault diagnosis can still be effectively carried out under the scenes of variable loads and strong noise interference. 本发明提出了一种基于多级抽象时间特征融合的滚动轴承故障诊断方法，所提出的算法能够从轴承振动信号中提取多级抽象特征并融合，同时捕捉隐藏在信号中的时