Adaptive Low-Rank Tensor Estimation Model Based Multichannel Weak Fault Detection for Bearings

Area of Science:

• Mechanical Engineering
• Signal Processing
• Condition Monitoring

Background:

• Multichannel signals offer rich fault information but face challenges in extracting weak fault characteristics due to noise and redundancy.
• Accurate extraction of fault characteristics from multichannel signals is crucial for effective mechanical fault diagnosis and early fault detection.

Purpose of the Study:

• To propose an effective weak fault detection framework for multichannel signals.
• To address the challenges of extracting weak fault characteristics from noisy multichannel signals.
• To enhance early fault detection capabilities in mechanical equipment, specifically rolling bearings.

Main Methods:

• Utilized tensor singular value decomposition to reveal the low-rank property of multichannel fault signals.
• Developed an adaptive low-rank tensor estimation model with an adaptive threshold function for weak fault extraction.
• Implemented a sparsity metric-oriented parameter optimization strategy to refine fault characteristic estimation.

Main Results:

• Demonstrated the effectiveness of the tensor domain in characterizing fault information.
• Successfully extracted weak fault characteristics from multichannel signals with strong background noise.
• Achieved excellent results in multichannel weak fault detection, particularly for early fault identification.

Conclusions:

• The proposed framework effectively addresses the limitations of traditional multichannel signal analysis for fault diagnosis.
• The method provides a practical and effective solution for weak fault detection in rolling bearings using multichannel data.
• The study highlights the potential of advanced tensor-based methods for improving the accuracy and reliability of mechanical fault diagnosis.