Intelligent fault diagnosis algorithm of rolling bearing based on optimization algorithm fusion convolutional neural...

Area of Science:

• Mechanical Engineering
• Artificial Intelligence
• Signal Processing

Background:

• Rolling bearing faults threaten mechanical equipment operation and cause financial losses.
• Traditional Convolutional Neural Networks (CNNs) have limitations in feature representation and hyperparameter determination for fault diagnosis.

Purpose of the Study:

• To propose an improved fault diagnosis method for rolling bearings.
• To enhance diagnostic accuracy and feature representation capabilities of CNNs.
• To optimize CNN hyperparameters efficiently.

Main Methods:

• An attention mechanism was integrated with a 1D-CNN to improve fault feature representation.
• A swarm intelligence algorithm combining Differential Evolution (DE) and Grey Wolf Optimization (GWO) was developed for hyperparameter optimization.
• The DE-GWO-CNN model was applied to rolling bearing fault diagnosis using CWRU and JNU datasets.

Main Results:

• The proposed DE-GWO-CNN method demonstrated increased diagnostic accuracy compared to traditional models.
• The enhanced model showed superior anti-noise capabilities.
• The optimization algorithm improved convergence accuracy and search efficiency.

Conclusions:

• The DE-GWO-CNN algorithm is a dependable and effective method for rolling bearing fault diagnosis.
• The methodology accurately identifies and classifies faults, aiding in prompt maintenance.
• This approach helps minimize equipment failures and operational instabilities.