Planetary Gear Fault Diagnosis via Feature Image Extraction Based on Multi Central Frequencies and Vibration Signal...

Area of Science:

• Mechanical Engineering
• Signal Processing
• Artificial Intelligence

Background:

• Planetary gear trains are prone to failures due to poor working environments.
• Non-linear and non-stationary vibration signals complicate fault diagnosis.

Purpose of the Study:

• To propose an effective planetary gear fault diagnosis method.
• To address challenges posed by complex vibration signals in fault detection.

Main Methods:

• Variational Mode Decomposition (VMD) to decompose vibration signals.
• Feature image extraction based on multi-central frequencies and spectrum analysis.
• Convolutional Neural Network (CNN) for fault signal identification.

Main Results:

• Achieved an overall fault recognition rate of 98.75%.
• The proposed feature band extraction method requires fewer iterations than direct component spectrum methods.
• Demonstrated the effectiveness of the VMD and CNN integrated approach.

Conclusions:

• The proposed method provides an effective solution for planetary gear fault diagnosis.
• Feature image extraction based on multi-central frequencies enhances diagnostic accuracy.
• The approach offers improved efficiency in fault identification for complex machinery.