Motor Fault Diagnosis Under Strong Background Noise Based on Parameter-Optimized Feature Mode Decomposition and...

Area of Science:

• Mechanical Engineering
• Signal Processing
• Artificial Intelligence

Background:

• Mining motors operate in challenging environments with significant noise interference.
• Extracting weak fault features from complex vibration signals is difficult.
• Existing methods struggle with superimposed working conditions and strong background noise.

Purpose of the Study:

• To develop a high-performance fault diagnosis approach for mining motors under strong noise.
• To improve the efficiency and accuracy of mechanical fault feature extraction and identification.
• To address the limitations of current methods in complex operational scenarios.

Main Methods:

• Integration of parameter-optimized feature mode decomposition (WOA-FMD) with a RepLKNet-BiGRU-Attention dual-channel model.
• Application of the proposed method to noise-added bearing fault datasets and actual mining motor operation data.
• Comparative analysis with existing fault diagnosis techniques.

Main Results:

• Achieved an average accuracy of 97.7% on the noise-added CWRU bearing fault dataset.
• Attained an average accuracy of 93.38% on the actual mining motor operation dataset.
• Demonstrated superior performance compared to similar fault diagnosis methods.

Conclusions:

• The proposed WOA-FMD and RepLKNet-BiGRU-Attention model offers a superior approach for mining motor fault diagnosis.
• The method effectively extracts and identifies mechanical fault features even in strong background noise.
• This approach enhances the reliability and safety of mining operations through accurate fault detection.