Multiple Enhanced Sparse Representation via IACMDSR Model for Bearing Compound Fault Diagnosis

Area of Science:

• Mechanical Engineering
• Signal Processing
• Condition Monitoring

Background:

• Diagnosing compound faults in bearings is challenging due to complex, noise-heavy vibration signals.
• Extracting multiple fault features simultaneously from these signals requires advanced decomposition and representation techniques.

Purpose of the Study:

• To develop a novel model, IACMDSR, for extracting multiple features from noise-heavy vibration signals for bearing compound fault diagnosis.
• To enhance the accuracy and robustness of bearing fault diagnosis in practical engineering applications.

Main Methods:

• Improved Adaptive Chirp Mode Decomposition (IACMD) for simultaneous separation of fault types and extraction of resonance frequencies.
• Construction of an adaptive bilateral wavelet hyper-dictionary to capture fault impulse response characteristics.
• Orthogonal Matching Pursuit (OMP) algorithm for identifying and reconstructing fault-induced features.
• Envelope demodulation analysis for detecting fault characteristic frequencies from the reconstructed signal.

Main Results:

• The IACMDSR model successfully separated distinct fault types and extracted multiple resonance frequencies.
• Simulation and experimental results demonstrated that IACMDSR outperforms leading models like MCKDSR and MCKDMWF.
• The model showed satisfactory capability in practical applications due to the adaptive nature of IACMD and the wavelet dictionary.

Conclusions:

• The developed IACMDSR model is a powerful and versatile tool for bearing compound fault diagnosis.
• IACMD's adaptability and the wavelet dictionary's matching capability contribute to the model's practical effectiveness.
• This approach offers a significant advancement in analyzing complex vibration signals for machinery health monitoring.