A Lightweight and Small Sample Bearing Fault Diagnosis Algorithm Based on Probabilistic Decoupling Knowledge Distillation and Meta-Learning
2024

Lightweight Bearing Fault Diagnosis Algorithm

Sample size: 30 publication 10 minutes Evidence: high

Author Information

Author(s): Luo Hao, Ren Tongli, Zhang Ying, Zhang Li

Primary Institution: College of Information, Liaoning University, Shenyang, China

Hypothesis

Can a lightweight fault diagnosis algorithm effectively diagnose bearing faults using small sample sizes?

Conclusion

The proposed MIX-MPDKD algorithm demonstrates high accuracy and robustness in diagnosing bearing faults, even with limited data.

Supporting Evidence

  • The algorithm achieved an average accuracy of 99.48% across various datasets.
  • Experimental results showed significant improvements over traditional models.
  • The proposed method effectively handles noise and small sample sizes.

Takeaway

This study created a smart way to find problems in bearings using less data, making it easier to use in real-life situations.

Methodology

The study used a combination of meta-learning and knowledge distillation techniques to train a lightweight model for fault diagnosis.

Limitations

The teacher model still requires a large amount of data for training, which may limit its applicability in data-scarce environments.

Statistical Information

P-Value

p<0.05

Statistical Significance

p<0.05

Digital Object Identifier (DOI)

10.3390/s24248157

Want to read the original?

Access the complete publication on the publisher's website

View Original Publication