Fault diagnosis of wind turbine generator bearings using fast spectral correlation

Abstract

Bearings are the critical parts that support the rotating of rotor of wind turbine generators. Due to high speed revolution and affected by potential misalignment between rotor and the high speed shaft in wind turbine gearbox, the fault ratio in wind turbine generator bearings is high. Once the bearings fail, it will cause gap eccentricity, even rub, or sweeping chamber between rotor and stator. Under fault conditions, the vibration signals from rotating machinery exhibits distinct second cyclostationarity. In the light of this, the fast spectral correlation based method is applied to the fault extraction of bearings in wind turbine generators. Through converting conventional correlation into summation algorithm, the computational cost is reduced largely, meanwhile, the diagnosis accuracy is guaranteed. The effectiveness of the method in this paper is verified by two fault cases from on-site wind turbines.