A Fast Sparse Decomposition Based on the Teager Energy Operator in Extraction of Weak Fault Signals

Area of Science:

• Mechanical Engineering
• Signal Processing
• Condition Monitoring

Background:

• Diagnosing incipient faults in rotating machinery is challenging due to complex operating conditions and noise interference.
• Traditional methods may struggle with early fault detection, especially under low signal-to-noise ratio (SNR) conditions.

Purpose of the Study:

• To propose a fast sparse decomposition method enhanced by the Teager energy operator (TEO) for improved incipient fault diagnosis in rotating machinery.
• To enhance the sensitivity to fault impulses while suppressing noise and harmonic components.

Main Methods:

• Utilizing the Teager energy operator (TEO) to enhance impulse envelopes, making them more sensitive to frequency changes and noise.
• Applying a smoothing filter to the TEO envelope to further suppress noise.
• Reconstructing the fault signal by multiplying the filtered TEO envelope with the original signal.
• Employing sparse decomposition based on the generalized S-transform (GST) for efficient signal representation.

Main Results:

• The proposed preprocessing method effectively overcomes high-frequency noise interference while preserving critical fault impulse structures.
• Sparse decomposition based on GST provides a faster and more accurate representation of the fault signal.
• The method demonstrated high accuracy and efficiency, successfully detecting impulses with an SNR of -8.75 dB submerged in noise.

Conclusions:

• The developed TEO-enhanced sparse decomposition method offers a robust solution for incipient fault diagnosis in rotating machinery.
• The technique is particularly effective under challenging conditions with very low SNR, improving machinery health monitoring.