Vibration data-driven fault detection for wind turbine gearbox

Abstract

Wind turbine gearbox is a very important transmission component. Therefore, it is very important to diagnose the wind turbine gearbox fault. This paper proposed a fault detection method by analyzing the vibration signals of wind turbine gearboxes to identify faults. In this paper, the vibration signals are turned into the angular domain by using equal angle resampling, and then Time-frequency diagram of obtaining vibration signals by continuous wavelet transform. And the difference between the health and the faulty gearbox is compared by extracted some fault features. The proposed method select the best fault features for fault diagnosis using the ROC curve.