Gear Fault Diagnosis Method Based on Multi-Sensor Information Fusion and VGG
Dongyue Huo1, Yuyun Kang2, Baiyang Wang1
1School of Information Science and Engineering, Linyi University, Linyi 276000, China.
Entropy (Basel, Switzerland)
|November 11, 2022
Summary
This study introduces a novel gear fault diagnosis method using multi-sensor data fusion and the Visual Geometry Group (VGG) network. The technique achieves 100% accuracy in identifying gear failures under various conditions.
Area of Science:
- Mechanical Engineering
- Artificial Intelligence
- Signal Processing
Background:
- Gearboxes are critical in mechanical transmission systems for aerospace and wind power.
- Gear failure is a primary cause of gearbox malfunction, necessitating accurate fault diagnosis.
- Traditional methods struggle with identifying gear faults under complex operating conditions.
Purpose of the Study:
- To develop an effective gear fault diagnosis method for complex working conditions.
- To improve the accuracy and reliability of gear fault identification.
- To leverage multi-sensor information fusion and deep learning for enhanced diagnosis.
Main Methods:
- Calculated power spectral density from multi-sensor frequency domain signals.
- Fused sensor data into a power spectral density energy map.
- Employed the Visual Geometry Group (VGG) network for fault diagnosis model creation.
Main Results:
- The proposed method achieved up to 100% accuracy on two independent datasets.
- Demonstrated high effectiveness and generalization ability in gear fault diagnosis.
- Successfully identified various gear fault types under diverse operating conditions.
Conclusions:
- The multi-sensor fusion and VGG-based method offers a robust solution for gear fault diagnosis.
- This approach significantly outperforms traditional methods in complex environments.
- The findings have strong implications for predictive maintenance in critical industries.
