Real-Time Bearing Faults Diagnosis with ARIMA-based Denoising and Residual Attention
Detection and diagnosis of bearing faults is crucial to prevent the unexpected failure of industrial machinery. Bearings are usually applied in harsh conditions which are more susceptible to failure. In the industry applications, bearing condition monitoring is a big challenge due to the complex har...
Saved in:
| Main Authors: | , , , , |
|---|---|
| Format: | Proceeding |
| Language: | en |
| Published: |
2025
|
| Subjects: | |
| Online Access: | http://ir.unimas.my/id/eprint/51575/3/Real-Time%20Bearing.pdf http://ir.unimas.my/id/eprint/51575/ https://ieeexplore.ieee.org/abstract/document/11233769 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | Detection and diagnosis of bearing faults is crucial to prevent the unexpected failure of industrial machinery. Bearings are usually applied in harsh conditions which are more susceptible to failure. In the industry applications, bearing condition monitoring is a big challenge due to the complex harmonic interference from the working system and also the machinery noise. Thus, an autoregressive integrated moving average (ARIMA) model is implemented with a residual attention model in the bearing fault diagnosis. The ARIMA model is applied to preprocess the bearing vibration signals to minimize the interference of the noise components. The permutation entropy is applied to select the most appropriate ARIMA model. The processed signal is then passed to the proposed residual attention model for the fault diagnosis. The analysis is conducted based on a real-time simulation to analyze the model’s responsiveness in real-time conditions. By comparing the prediction performance between original signals and signals processed by the ARIMA model, the processed signals proved to have a significant improvement of 9.13% average performance, with 23.70 ± 2.40 ms for each second of data prediction. The application of ARIMA models effectively cleans the vibration signals and improves the diagnostic capabilities of the model in bearing faults diagnosis. |
|---|