A novel fault diagnosis method for gearbox based on RVMD and TELM with composite chaotic grey wolf optimizer

Area of Science:

• Mechanical Engineering
• Signal Processing
• Artificial Intelligence

Background:

• Gearbox fault diagnosis is critical for industrial machinery maintenance.
• Traditional methods like Variational Mode Decomposition (VMD) are sensitive to noise and outliers.
• Existing machine learning models may not achieve optimal feature extraction and classification accuracy.

Purpose of the Study:

• To propose a robust and accurate fault diagnosis method for gearboxes.
• To enhance signal decomposition techniques for noisy environments.
• To improve the classification performance of machine learning models in fault detection.

Main Methods:

• Robust Variational Mode Decomposition (RVMD) for noise-resilient signal decomposition into intrinsic mode functions (IMFs).
• Twin Extreme Learning Machine (TELM) for advanced feature extraction and classification.
• Composite Chaotic Grey Wolf Optimizer (CCGWO) to optimize TELM kernel parameters, creating the CCGTELM model.

Main Results:

• The RVMD-CCGTELM method demonstrated higher fault diagnosis accuracy compared to VMD-TELM, VMD-DNN, VMD-CNN, VMD-LSTM, EMD-ELM, and WT-ANN.
• RVMD effectively handles noise and outliers, outperforming standard VMD.
• CCGTELM showed improved feature extraction and classification capabilities.

Conclusions:

• The proposed RVMD-CCGTELM is a highly effective and accurate method for gearbox fault diagnosis.
• RVMD offers significant advantages over VMD in noisy conditions.
• The CCGWO optimization enhances TELM's performance for complex fault diagnosis tasks.