Rolling Bearing Fault Diagnosis Using GADF and Dynamic Self-Calibrated Convolution
Author Information
Author(s): Liu Chunli, Bai Jiarui, Xue Linlin, Xue Zhengkun
Primary Institution: University of Science and Technology Liaoning
Hypothesis
Can a rolling bearing fault diagnosis method based on GADF and DSC improve feature extraction under noise and small sample conditions?
Conclusion
The proposed method shows excellent diagnostic performance with over 90% accuracy even under strong noise and small training samples.
Supporting Evidence
- The proposed method achieved classification accuracy above 90% under strong noise conditions.
- DSCNN outperformed other advanced models by at least 6% in classification accuracy.
- The method effectively captures nonlinear relationships in time-series data.
- Dynamic weight adjustment in the DSC module enhances feature extraction.
- Experimental validation was conducted using datasets from HUST and HIT.
Takeaway
This study created a new way to find problems in rolling bearings, even when there's a lot of noise and not many examples to learn from.
Methodology
The study used GADF to convert vibration signals into images and a DSC module to enhance feature extraction.
Potential Biases
Potential bias in dataset selection and noise conditions may affect generalizability.
Limitations
The study may not cover all types of noise or all operational conditions of rolling bearings.
Statistical Information
P-Value
p<0.05
Statistical Significance
p<0.05
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website