Analysis of wind ramp events using their conditional probabilities from PCA 3D tables

Abstract

Wind is one of the fastest growing sources of renewable energy. However, its intermittency and stochastic nature often hamper integration of wind-generated energy in electric power networks. The most significant obstacles to efficient wind integration are the wind ramps, sudden changes of wind power output caused by dramatic changes in wind speed. Most ramp prediction methods derive future ramp estimates from forecast power series. This contribution suggests a different approach that would directly analyze wind power time series and other weather parameters to identify specific patterns and dependencies signaling approaching ramp events. In this paper, the 3D tables of ramp events' conditional probabilities are evaluated using principal component analysis. Two different datasets are analyzed. First one contains power production hourly collected from August 2011 to July 2012 on sample wind farm located close to Lethbridge, AB Canada. The other consists of total BPA wind power production in 2013 with 5 minute resolution. Frequency tables for the first three principal components are used to evaluate the conditional probability of a forthcoming wind ramp event. Utilization of resulting 3D tables for wind ramp event forecast is tested on the last quarter of the data. Advantage of proposed methodology compared to conventional methods is that numerical weather prediction (NWP) model producing wind forecasts is not required.