Denoising method of machine tool vibration signal based on variational mode decomposition and Whale-Tabu optimization...

Area of Science:

• Engineering
• Signal Processing
• Machine Learning

Background:

• Vibration signals from CNC machine tools are often contaminated with noise from various sources.
• Accurate vibration analysis is crucial for machine health monitoring and performance optimization.

Purpose of the Study:

• To propose and validate a novel joint analysis denoising method for CNC machine tool vibration signals.
• To enhance the accuracy and reliability of vibration data by effectively removing noise.

Main Methods:

• Variational Mode Decomposition (VMD) to decompose signals into intrinsic mode functions (IMFs).
• Whale Optimization Algorithm (WOA) and Tabu Search (TS) for optimizing VMD parameters, using minimum permutation entropy as the fitness function.
• Correlation Analysis (CA) to categorize IMFs into pure components, noisy signals, and noise.
• Wavelet Threshold (WT) denoising applied to noisy components, followed by signal reconstruction using pure components.

Main Results:

• The proposed TS-WOA-VMD-CA-WT method demonstrates superior performance compared to single optimization algorithms in denoising simulations.
• The method effectively denoises actual machine tool vibration signals, validating its practical applicability.
• The joint analysis approach significantly improves signal quality over existing denoising techniques.

Conclusions:

• The TS-WOA-VMD-CA-WT method offers a robust and generalizable solution for denoising CNC machine tool vibration signals.
• This advanced technique holds significant potential for widespread adoption in industrial applications for improved machine monitoring and maintenance.