Wavelet coherence-aware multi-branch deep ensemble for fault identification in centrifugal pumps

Area of Science:

• Mechanical Engineering
• Signal Processing
• Artificial Intelligence

Background:

• Centrifugal pump fault diagnosis is difficult due to complex vibration signals.
• Early fault signatures are often weak and overlap with other defects.

Purpose of the Study:

• To develop a robust framework for reliable centrifugal pump fault diagnosis.
• To integrate time-frequency analysis with deep learning for enhanced feature extraction.

Main Methods:

• A wavelet coherence-aware multi-branch deep ensemble framework was proposed.
• Multi-channel vibration signals were converted into 2D wavelet coherence maps.
• Three diverse convolutional neural networks were trained in parallel and their outputs fused via soft-voting.

Main Results:

• The framework achieved consistent and reliable fault discrimination across different operating pressures.
• Strong class separability was confirmed using Receiver Operating Characteristic analysis and feature-space visualization.
• The proposed method effectively addresses challenges in diagnosing centrifugal pump faults.

Conclusions:

• The wavelet coherence-aware deep ensemble framework demonstrates high effectiveness for centrifugal pump fault diagnosis.
• The integration of physically meaningful features with deep learning improves diagnostic accuracy and stability.
• This approach offers a promising solution for real-world industrial applications.