Analysis of wind speed and power time series preceding wind ramp events

Wind energy has become one of the fastest growing renewable sources during last few decades. Sudden changes in wind power output, called wind ramps, recently attracted great research interest in wind power forecasting community. Conventional ramp prediction methods derive future ramp estimates from power forecast series. We suggest to analyze real wind power series or other weather parameters, searching for specific patterns and dependencies indicating forthcoming wind ramp events.This paper presents a methodology for parametric analysis of time series preceding wind ramp events. The presented methodology is based on probabilistic data analysis. Sensitivity of the developed algorithm can be adjusted through the size of the time window defined before a ramp event. Power production and other weather data are standardized, averaged and then searched for specific patterns in the form of trend lines. The resulting methodology does not require the use of a numerical weather prediction (NWP) model; this is a significant simplification compared to most conventional methods. The proposed methodology is tested using power production and wind speed data collected between August 2011 and July 2012 at a wind farm located in a wind-rich region of southern Alberta, Canada.