A degradation feature extraction technique based on static divided symbol sequence entropy

Area of Science:

• Mechanical Engineering
• Signal Processing
• Condition Monitoring

Background:

• Traditional methods struggle to identify degradation patterns in quay crane gearbox vibration signals due to noise and impact components.
• Accurate feature extraction is crucial for effective condition monitoring and predictive maintenance of critical machinery.

Purpose of the Study:

• To propose a novel degradation feature extraction technique for quay crane gearbox vibration signals.
• To enhance the characterization of signal complexity and performance degradation in the presence of noise.

Main Methods:

• Development of a static divided symbol sequence entropy technique based on basic scale entropy.
• Uniform symbolization standard using the root mean square of the health condition signal and a scale coefficient.
• Expansion of the symbol set to improve information content and complexity characterization, especially in large-value regions.

Main Results:

• The proposed technique accurately and flexibly characterizes performance degradations in vibration signals.
• Analysis using logistic chaotic sequences and hoisting mechanism gearbox lifetime signals demonstrates sensitive description of performance degradation.
• The method effectively characterizes the complexity of nonlinear time series.

Conclusions:

• The static divided symbol sequence entropy technique offers a robust solution for extracting degradation features from noisy vibration data.
• The method is computationally fast and suitable for real-time applications.
• This technique provides a foundation for developing advanced health evaluation methods for quay crane gearboxes.