1. A gear fault diagnosis method based on manifold semi-supervised K-means clustering algorithm
- Author
-
Wenjing Liu, Jian Guo Wang, Wen Xing Zhang, and Pengfei Zhao
- Subjects
0209 industrial biotechnology ,Computer science ,business.industry ,k-means clustering ,Pattern recognition ,02 engineering and technology ,Fault (power engineering) ,Manifold ,law.invention ,Wavelet packet decomposition ,ComputingMethodologies_PATTERNRECOGNITION ,020901 industrial engineering & automation ,law ,0202 electrical engineering, electronic engineering, information engineering ,020201 artificial intelligence & image processing ,Artificial intelligence ,Cluster analysis ,business ,Manifold (fluid mechanics) ,Energy (signal processing) - Abstract
Aiming at the complexity and non-stationarity of gearbox vibration signal, a manifold semi-supervised K-means clustering algorithm based on 27-feature fusion is proposed to identify the fault of gear. Vibration signals of gear boxes under different working conditions were collected, and the time-frequency domain features, WPT (Wavelet Packet Transform) energy features and EEMD (Ensemble Empirical Mode Decomposition) energy features were fused into 27-dimensional features, semi-supervised K-means clustering algorithm based on manifold distance was used to classify fault signals. The results show that the fused features can reflect the fault features better, and the improved K-means algorithm has higher efficiency and better fault recognition effect.
- Published
- 2019
- Full Text
- View/download PDF