As an important part of wind turbine health monitoring and management

the identification of operating condition is often affected by uncertainty information and high-speed real-time data streams

which causes it difficult to carry out a health assessment effectively. In this context

this paper proposes a real-time health assessment method based on Spark streaming

First

the big data analysis technology is used to realize the spatial division of wind turbine operating condition. And then

taking into consideration the uncertainty of monitoring information

the Gaussian cloud model and Gaussian cloud transformation-based assessment model is trained according to the historical data of supervisory control and data acquisition(SCADA). In this process

the health index is introduced and can be viewed as an indicator of wind turbines health assessment. Finally

the method is applied in order to assess the health condition of a 1. 5 MW wind turbine located in north of China. Results show that the method can monitor the changing of the health of wind turbines and the accuracy improves compared with the traditional method.