Variable Filtered-Waveform Variational Mode Decomposition and Its Application in Rolling Bearing Fault Feature...

Area of Science:

• Signal Processing
• Mechanical Engineering
• Fault Diagnosis

Background:

• Variational Mode Decomposition (VMD) effectively decomposes signals into narrowband components.
• Classical VMD's reliance on Wiener filters limits adaptability for broadband signals, causing issues like mode mixing and over-smoothing.

Purpose of the Study:

• To introduce a novel Variable Filtered-Waveform Variational Mode Decomposition (VFW-VMD) method.
• To address the limitations of traditional VMD in processing broadband and chirp signals.
• To enhance the accuracy and effectiveness of bearing fault diagnosis.

Main Methods:

• Developed the Variable Filtered-Waveform Variational Mode Decomposition (VFW-VMD) algorithm.
• Incorporated fractional-order constraints and dynamic filter waveform adjustments.
• Validated the VFW-VMD framework through simulations with synthetic and real-world signals.

Main Results:

• VFW-VMD demonstrated superior adaptability in extracting broadband signals compared to classical VMD.
• The proposed method effectively mitigated mode mixing and over-smoothing issues.
• Enhanced capture of rolling bearing fault features was observed, improving diagnostic capabilities.

Conclusions:

• VFW-VMD offers significant advancements in signal decomposition for complex signals.
• The method provides enhanced performance for practical bearing fault diagnostic applications.
• This work contributes to improved signal processing techniques for machinery health monitoring.