A Fault Diagnosis Method for Rolling Bearings Based on Enhanced Sparrow Search Algorithm-Optimized VMD and...

Area of Science:

• Mechanical Engineering
• Signal Processing
• Artificial Intelligence

Background:

• Rolling bearing fault diagnosis is challenging due to nonlinear and non-stationary vibration signals.
• Accurate feature extraction is crucial for reliable fault detection.

Purpose of the Study:

• To propose a novel and effective method for rolling bearing fault diagnosis.
• To enhance feature extraction from complex vibration signals using optimized decomposition techniques.

Main Methods:

• Levy-Cauchy Optimized Sparrow Search Algorithm (LOCSSA) optimized Variational Mode Decomposition (VMD) for signal decomposition.
• Convolutional Neural Network (CNN) for local feature extraction from intrinsic mode functions (IMFs).
• Bidirectional Long Short-Term Memory (BiLSTM) network for modeling temporal dependencies and final fault classification.

Main Results:

• The proposed method achieved high accuracy (96.33%), recall (96.67%), and F1-score (96.54%).
• Outperformed five benchmark models in comparative experiments.
• Ablation studies confirmed the significant contributions of LOCSSA-optimized VMD and BiLSTM (p < 0.05).

Conclusions:

• The developed LOCSSA-VMD-CNN-BiLSTM framework offers a robust solution for rolling bearing fault diagnosis.
• The method effectively addresses challenges posed by nonlinear and non-stationary vibration data.
• The findings validate the proposed approach's superiority and practical applicability.