Early Waming Method for Bearing Over-temperature Fault of Wind Turbine Based on Sparse Bayesian Learning

The bearing of the wind turbine is the carrier of energy transfer and plays a very important role in the process of converting wind energy into electrical energy.This paper takes the bearing over-temperature fault of wind turbine as the research object, and proposes a fault early warning model based on sparse Bayesian learning(SBL). Firstly, using the data collected by the SCADA(Supervisory Control And Data Acquisition)system, the relevant variables of bearing overtemperature fault are filtered and correlated to determine the input and output variables of the SBL model. After that, the SBL model is used to form the probability interval of the expected value distribution of the output variable; Finally, the result of the hypothesis test is used as a judgment index to judge whether the current operating state of the wind turbine is abnormal. Through the analysis of the actual wind turbine bearing over-temperature faults, it is proved that the bearing over-temperature fault early warning method based on the SBL model has strong practicability and can warn the bearing over-temperature faults accurately.