Abstract: Targeting at the problem of complex transmission path of inter-shaft bearing fault signal, weak fault signal characteristics and difficulty in fault feature extraction, a fault diagnosis method based on Multi-scale Quantum Entropy (MQE), Locally Linear Embedding (LLE) algorithm and Probabilistic Neural Network (PNN) is proposed in this paper. Firstly,the inter-shaft bearing fault signals are denoised through the spatial correlation noise reduction method to improve the signal to noise ratio. Secondly, MQE is utilized to extract the features of inter-shaft bearings. Then, LLE is utilized to reduce and fuse high-dimensional fault features of multi-sensor to construct fault samples. Finally, the low-dimensional fault features are input into PNN multi-fault classifier for fault identification. The fault simulation test bench of the inter-shaft bearing is built to simulate the normal bearing, inner ring fault, outer ring fault and rolling element fault, and the data were collected to verify the MQE-LLE-PNN inter-shaft bearing fault diagnosis algorithm established in this paper. The experimental results validate that the proposed method can effectively identify the inter-shaft bearing fault, and shows good generalization ability without any over-fitting phenomenon.

Key words: inter-shaft bearing, spatial correlation, multi-scale quantum entropy, locally linear embedding, fault diagnosis