Amplitude Normalization for Speed-Induced Modulation in Rotating Machinery Measurements

Area of Science:

• Mechanical Engineering
• Signal Processing
• Condition Monitoring

Background:

• Variable-speed rotating machinery experiences amplitude modulation (AM) due to speed fluctuations, hindering precise fault detection.
• Existing methods struggle to effectively mitigate these speed-induced AM effects, impacting diagnostic accuracy.

Purpose of the Study:

• To develop and validate a robust amplitude normalization technique for mitigating speed-induced AM effects in rotating machinery.
• To enhance the clarity of fault signals for improved diagnostic performance.

Main Methods:

• A correlation-based feature selection strategy was used to identify speed-related features.
• Support vector regression (SVR) was employed to model and remove the estimated AM effects.
• The proposed method was validated using two case studies and compared against existing techniques.

Main Results:

• The proposed SVR-based method accurately estimates and removes speed-induced AM effects.
• The technique demonstrated superior accuracy, robustness, and reliability compared to advanced normalization methods.
• Fault diagnosis accuracy was significantly improved, showing an enhancement of up to 34.7%.

Conclusions:

• The developed amplitude normalization method effectively addresses the challenges of variable-speed operation in rotating machinery.
• This approach offers a reliable solution for enhancing fault detection accuracy in condition monitoring applications.
• The study highlights the potential of SVR in signal processing for mechanical fault diagnosis.