Research on reliability of centrifugal compressor unit based on dynamic Bayesian network of fault tree mapping

Yuan, Gao; Zhang, Liang; Jiawei, Zhou; Bojia, Wei; Zhongchao, Yan

doi:10.1007/s12206-023-0440-7

Cited by 5 publications

(1 citation statement)

References 20 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…To lower the model's complexity, many scholars first use traditional feature transformation to extract shallow features from the original data, and then use DL technology to extract deep-level features and conduct pattern recognition [4]. Fault diagnosis methods based on DL often use convolutional neural networks [5] (CNN), recurrent neural networks [6] (RNN), generative adversarial networks [7] (GAN), deep belief networks [8] (DBN), stacked autoencoders [9] (SAE) and other models. Islam et al [4] collected the signal of the bearing through the acoustic transmitter, converted the information into wavelet spectrum by wavelet packet transform, then selected the band signal with significant characteristics through the defect rate index, and finally input the band signal into the adaptive CNN to diagnose the fault.…”

Section: Introductionmentioning

confidence: 99%

A lightweight fault diagnosis model for planetary gearbox using domain adaptation and model compression

Song,

Xiong,

Zhang

et al. 2024

J. vibroeng.

View full text Add to dashboard Cite

This article proposes a novel lightweight attention spatiotemporal joint distribution adaptation network fault diagnosis model to address the key challenges of domain transfer and high model complexity in traditional methods. The novelty lies in 1. Using model compression techniques to reduce the complexity of the network model and improve its computational efficiency; 2. Introducing new domain adaptation and adversarial methods to solve the domain transfer problem. The effectiveness of the proposed model is verified through a transfer experiment of planetary gearbox vibration data. The experimental results show that the proposed model reduces the parameters and computational complexity to 18 % and 15 % of the original model, respectively, and has a diagnostic accuracy of over 98 % in cross-condition transfer tasks, and still maintains an accuracy of over 88 % even under high noise levels. This indicates that the proposed model is an efficient and accurate fault diagnosis model.

show abstract