Generative Adversarial Learning Enhanced Fault Diagnosis for Planetary Gearbox under Varying Working Conditions

Area of Science:

• Mechanical Engineering
• Artificial Intelligence
• Signal Processing

Background:

• Planetary gearboxes are crucial for mechanical systems, requiring continuous health monitoring.
• Deep learning enhances mechanical fault diagnosis, but varying working conditions pose a significant challenge (cross-domain issue).

Purpose of the Study:

• To develop an intelligent fault diagnosis scheme for planetary gearboxes under varying speed conditions.
• To address the cross-domain issue in deep learning-based fault diagnosis.

Main Methods:

• A novel generative adversarial learning method was developed using deep convolution neural networks.
• Experiments were conducted using measured vibration signals from a planetary gearbox.

Main Results:

• The proposed method demonstrated enhanced capability in diagnosing planetary gear faults under varying speed conditions.
• Experimental validation confirmed the effectiveness and efficiency of the developed fault diagnostic scheme.

Conclusions:

• The novel generative adversarial learning approach effectively overcomes the cross-domain issue in planetary gearbox fault diagnosis.
• The intelligent fault diagnostic scheme significantly improves the accuracy of fault detection under dynamic operating conditions.