A study on Fault Diagnosis Method of Rolling Bearing Based on Wavelet Packet and Improved BP Neural Network

1st International Conference on Frontiers of Materials Synthesis and Processing

材料科学;化学

Song, Mengmeng^1 ; Song, Haixia^2 ; Xiao, Shungen^1,3

School of Information, Mechanical and Electrical Engineering, Ningde Normal University, Ningde, China^1

School of Information Science and Engineering, Huaqiao University, Xiamen, China^2

School of Mechatronic Engineering and Automation, Shanghai University, Shanghai, China^3

关键词: BP neural networks;    De-noised signals;    Energy characteristics;    Fault diagnosis method;    Improved BP neural network;    Rolling bearings;    Threshold de-noising;    Wavelet Packet;   

Others  :  https://iopscience.iop.org/article/10.1088/1757-899X/274/1/012133/pdf
DOI  :  10.1088/1757-899X/274/1/012133

学科分类：材料科学（综合）

来源: IOP

【 摘 要 】

In this paper, rolling bearing fault diagnosis method is proposed based on wavelet packet threshold de-noising and improved BP neural network. It achieves the goal of signal de-noising by setting the appropriate threshold, and then the denoised signal is decomposed into three layers by wavelet packet. The energy characteristics of the 8 frequency bands are calculated respectively. Levenberg-Maquardt algorithm which is improved the traditional BP neural network to improve the diagnosis efficiency of BP neural network, is proposed. Taking the outer ring fault of rolling bearings as an example, the experimental results show that the wavelet packet threshold de-noising can effectively improve the signal-to-noise ratio. Compared with the traditional BP neural network, the improved BP neural network has better diagnosis efficiency.