Application of Wavelet synchrosqueezing transform for wind turbine gearbox fault diagnosis

Abstract

In Wind turbine condition monitoring systems, the vibration signal analysis is used as the effective tool for fault diagnosis. Gear transmission systems are complex and expensive among the drive train components. Faults developed in the gearbox will lead to unscheduled down time of wind turbine. Vibration signals are non stationary in nature and the usual spectral analysis tools are not sufficient for the effective fault diagnosis. In this paper, A novel joint time-frequency Wavelet synchrosqueezing transform (WSST) is proposed for the effective intrinsic modes extraction. WSST provides fine Time-Frequency resolution. To improve the performance of WSST towards noise sensitivity, Wavelet Denoising with soft thresholding is employed. The efficacy of the proposed methodology is tested with the NERL GRC Wind Turbine condition monitoring benchmark datasets. From the simulation results, it has been shown that the proposed method extracts the intrinsic mode functions with more precise time frequency resolution.