Exploiting SCADA system data for wind turbine performance monitoring

This paper presents the results of a short study into utilising wind farm supervisory control and data acquisition (SCADA) system data for performance monitoring of large utility-scale wind turbines. The general approach taken is to model the turbine power output of each turbine during fault-free operation and to subsequently use the trained model to identify performance degradation by analysing the residual between the predicted and observed power values for each turbine. Historical data from a large wind farm is used to train and test the turbine models. The trained models are then tested on historical turbine failure examples. The results suggest that the data collected by wind farm SCADA systems, which are typically installed as standard on most modern wind farms, can be exploited for gaining an insight into wind turbine performance and maintenance condition.