An Intelligent Gear Fault Diagnosis Methodology Using a Complex Wavelet Enhanced Convolutional Neural Network.

Area of Science:

• Mechanical Engineering
• Signal Processing
• Artificial Intelligence

Background:

• Rotating machinery is susceptible to various mechanical faults, particularly within gear transmission systems.
• Identifying fault signatures in vibration signals is challenging due to noise and interference.
• Weak fault features are often difficult to detect and diagnose accurately.

Purpose of the Study:

• To develop a novel intelligent fault diagnosis method for rotating machinery.
• To enhance the recognition accuracy of characteristic fault signals, especially weak ones.
• To address the challenges of submerged fault signatures in complex mechanical systems.

Main Methods:

• Utilized dual-tree complex wavelet transform (DTCWT) for multiscale feature extraction from vibration signals.
• Employed a convolutional neural network (CNN) for automatic fault feature recognition.
• Integrated DTCWT and CNN to create a robust intelligent fault diagnosis system.

Main Results:

• The proposed method demonstrated feasibility and effectiveness in recognizing gear faults.
• Successfully identified weak fault features that are typically challenging to detect.
• Achieved high accuracy in fault diagnosis for rotating machinery components.

Conclusions:

• The combined DTCWT and CNN approach offers a powerful solution for intelligent fault diagnosis.
• This method significantly improves the detection and recognition of weak gear fault signals.
• The study validates the effectiveness of the proposed intelligent system for rotating machinery health monitoring.