STATISTICAL MODELING AND MONITORING FOR OPERATIONS IN WIND ENERGY

Abstract

The wind industry faces great operational challenges caused by the fluctuated power generation in response to the naturally intermittent weather conditions. It is therefore of great interest to model the operations of wind energy systems to ensure better system functionality, reliability, and flexibility, thus enhancing the system resiliency. This dissertation focuses on developing advanced statistical modelling and monitoring methods on wind turbine level and wind farm level. For performance assessment, the first research work developed a novel conditional density estimation considering autocorrelation for wind power generation distributions. For anomaly detection, the second research work did a retrospective analysis of the historical wind turbine operation data, while the third research work conducted monitoring in real-time and tackles the irregular and typically sparse observations. For degradation modeling, the last research work developed a novel Wiener process model based on Kriging to better characterize the heterogeneity of multiple wind farms.