Modeling High Wind Speed Shut-Down Events Using Meso-Scale Wind Profiles and Survival Analysis

Optimal siting is of paramount importance for managing the grid impacts of variable wind energy resources, especially in view of the increasing penetration of wind energy as part of the renewable energy mix. High variability of the renewable energy generation component requires that the system operator respond to generation deficits through frequent cycling of conventional generation and increased use of expensive peaking generation plants. Limited research has been conducted into methodologies whereby which the frequency, severity and relative risk of the loss of generation events can be quantified. This paper proposes a technique to empirically evaluate the risk of loss of generation events introduced by a singular site based on its simulated, or measured, historicity. This is accomplished through the statistical translation of meso-scale wind speeds into a number of wind speeds representative of those experienced by the turbines at a wind farm. An event series is derived describing the potential number of turbines in high wind speed shut down mode at the wind farm. Once the event series is obtained, survival analysis and the Cox proportional hazards model are utilized to characterize a site and contrast multiple sites in terms of relative risk. The methodology is demonstrated for high wind speed shutdown events, but can be applied for any event of interest, such as ramp-rate exceedances and drop below cut-in.