A noise reduction method for rolling bearing based on improved Wiener filtering.

Area of Science:

• Mechanical Engineering
• Vibration Analysis
• Fault Diagnosis

Background:

• Accurate identification of compound bearing faults is crucial for machinery maintenance.
• Traditional noise reduction methods struggle with complex fault signatures in vibration signals.
• Feature Mode Decomposition (FMD) offers a potential avenue for signal separation.

Purpose of the Study:

• To develop a novel noise reduction method for accurate compound bearing fault identification.
• To adaptively determine Wiener filtering parameters based on signal characteristics.
• To enhance the reliability of fault diagnosis in rotating machinery.

Main Methods:

• Vibration signals are decomposed into modal components using Feature Mode Decomposition (FMD).
• Signal evaluation indices (kurtosis, RMS, variance) and Euclidean distance are used to select optimal signals for Wiener filtering.
• The order of Wiener filtering is adaptively determined using maximum kurtosis as the criterion.

Main Results:

• The proposed method adaptively determines the input signals and Wiener filtering order.
• It effectively separates fault-related frequency components from noise.
• Comparison with classical methods shows superior noise restriction and more accurate compound fault identification.

Conclusions:

• The presented adaptive noise reduction method significantly improves the accuracy of compound bearing fault diagnosis.
• The technique offers a robust solution for analyzing complex vibration signals in machinery.
• This approach enhances diagnostic capabilities for bearing health monitoring.