DCA-BiGRU

The pytorch implementation of the paper Fault diagnosis for small samples based on attention mechanism

However, in fact, the title Fault diagnosis for small samples based on interpretable improved space-channel attention mechanism and improved regularization algorithms fits the research content of the paper better.

The dataset comes from 12khz, 1hp

Contributions:

1. 1D-signal attention mechanism [code]
2. AMSGradP [code]
3. 1D-Meta-ACON [code]
4. At the beginning, I found that many model designs did not connect GAP operation after BiGRU/BiLSTM, which is the basically routine operation. I found that GAP works very well. [code]
5. 1D-Grad-CAM++ [code]
6. AdaBN [code]

Attention Block（SCA）

How does it work?

If it is helpful for your research, please kindly cite this work:

@article{He_2024, 
title={Interpretable modulated differentiable STFT and physics-informed balanced spectrum metric for freight train wheelset bearing cross-machine transfer fault diagnosis under speed fluctuations}, 
volume={62}, 
DOI={10.1016/j.aei.2024.102568}, 
journal={Advanced Engineering Informatics}, 
author={He, Chao and Shi, Hongmei and Li, Ruixin and Li, Jianbo and Yu, ZuJun}, 
year={2024}, 
pages={102568} 
}

Our other works

@article{HE,  
title = {Physics-informed interpretable wavelet weight initialization and balanced dynamic adaptive threshold for intelligent fault diagnosis of rolling bearings},  
journal = {Journal of Manufacturing Systems},  
volume = {70},  
pages = {579-592},  
year = {2023},  
issn = {1878-6642},  
doi = {https://doi.org/10.1016/j.jmsy.2023.08.014},  
author = {Chao He and Hongmei Shi and Jin Si and Jianbo Li}   
}

Environment

pytorch == 1.10.0
python == 3.8
cuda == 10.2

Contact

• Chao He
• chaohe#bjtu.edu.cn (please replace # by @)

Views