Novel Fault Diagnosis Technology Based on Integrated Spectral Kurtosis for Gearboxes

Area of Science:

• Mechanical Engineering
• Signal Processing
• Condition Monitoring

Background:

• Gearbox failures pose significant risks to industrial operations.
• Accurate fault diagnosis is crucial for predictive maintenance and operational efficiency.
• Conventional spectral kurtosis (SK) has limitations in diagnosing gearbox defects.

Purpose of the Study:

• To introduce and evaluate a novel Integrated Spectral Kurtosis (ISK) technology for gearbox fault diagnosis.
• To compare the diagnostic performance of ISK with conventional SK and motor current analysis.
• To assess the effectiveness of ISK in distinguishing between healthy and defected gearboxes.

Main Methods:

• Vibration signals processed using Time Synchronous Averaging (TSA).
• Analysis performed using Spectral Kurtosis (SK) and Integrated Spectral Kurtosis (ISK).
• ISK feature estimation across the frequency domain; envelope obtained via SK-based filtering and Hilbert demodulation.
• Motor current signals analyzed using harmonic amplitude tracking.

Main Results:

• ISK achieved a Total Probability of Correct Diagnosis (TPCD) of 91.5% for pinions and 96.1% for gears.
• Conventional SK-based squared envelope technology showed limited effectiveness (max TPCD of 80%).
• Motor current analysis reached a maximum TPCD of 90% for gears, lower than ISK.

Conclusions:

• Integrated Spectral Kurtosis (ISK) technology demonstrates significant superiority for gearbox fault diagnosis.
• ISK provides a more effective and reliable method for machinery health monitoring compared to SK and motor current analysis.
• The proposed ISK method enhances diagnostic accuracy for critical gearbox components.