Fault diagnosis of localized tooth flaking on planetary wheels of wind turbine gearboxes

• 1、Mingyang Smart Energy Group Co., Zhongshan, Guangdong 528400, China

Abstract：In order to improve the effects of predictive maintenance for wind turbine drive chain, the accurate diagnosis of the initial failure of planetary wheel train is particularly important. In this study, for the extraction and analysis of planetary wheel train fault features, fast spectral gradient and envelope demodulation methods were used to explore effective fault diagnosis pathways. The key fault features were successfully extracted by fast spectral gradient analysis and envelope demodulation of the vibration signal. The analysis results show that the changes of frequency components at different time points reveal the development process of the fault: the fault gradually evolves from the initial characteristics of the gear ring to the characteristics of the planetary wheel, and finally leads to the spalling of the tooth surface of a planetary wheel. In this process, the obvious changes in the localized fault characteristic frequency of the toothed ring and the planetary carrier rotation frequency components provide important clues for the fault development. Based on the evolution of the failure characteristics, an early warning was proposed to be carried out on the 34th day before the failure occurs involving an increase in the distribution of the planetary wheel rotation frequency components. Accordingly, timely maintenance measures can be taken to prevent further deterioration of the failure. In summary, this study provides an effective method for early fault diagnosis of planetary wheel systems and theoretical support for fault early warning.

Keywords： early fault diagnosis planetary wheel system vibration signal analysis fast spectral cliff envelope demodulation warning timing