Fault diagnosis of gearbox based on wavelet packet transform and CLSPSO-BP

Fault diagnosis of gearbox is difficult due to the complexity and instability of its vibration signal. A fault diagnosis method of gearbox based on WPT-CLSPSO-BP (Wavelet Packet Transform- Chaos Particle Swarm Optimization-Back Propagation Neural Network) is proposed in this study to solve this prob...

Full description

Saved in:
Bibliographic Details
Published inMultimedia tools and applications Vol. 81; no. 8; pp. 11519 - 11535
Main Authors Xiao, Maohua, Zhang, Wei, Zhao, Yuanfang, Xu, Xiaomei, Zhou, Shufang
Format Journal Article
LanguageEnglish
Published New York Springer US 01.03.2022
Springer Nature B.V
Subjects
Algorithms
Back propagation networks
Computer Communication Networks
Computer Science
Data Structures and Information Theory
Energy value
Fault diagnosis
Gearboxes
Local optimization
Multimedia Information Systems
Neural networks
Particle swarm optimization
Special Purpose and Application-Based Systems
Wave propagation
Wavelet transforms
Fault diagnosis
BP neural network
Chaotic particle swarm algorithm
Wavelet packet transform
Online AccessGet full text

Cover

More Information
Summary:Fault diagnosis of gearbox is difficult due to the complexity and instability of its vibration signal. A fault diagnosis method of gearbox based on WPT-CLSPSO-BP (Wavelet Packet Transform- Chaos Particle Swarm Optimization-Back Propagation Neural Network) is proposed in this study to solve this problem. Wavelet packet transform (WPT) is used to decompose and reconstruct the signal, and the energy value of each component is calculated according to the formula, and the energy value is used as a feature input to form a feature sample. Aiming at the problem of slow convergence speed and easy local optimization of traditional BP neural network, a chaotic particle swarm algorithm is proposed to optimize the weight and threshold of the network, and the optimized network performance is verified with the data collected by experiments. Experimental results show that the average diagnosis rate of CLSPSO-BP is above 92%,the trained model not only has a high diagnostic rate, which can be improved by nearly 10%, but also can keep the error value between the actual output and the predicted output below 0.1%, indicating that the optimized network has a higher fault recognition rate.
ISSN:1380-7501
1573-7721
DOI:10.1007/s11042-022-12465-3