Skip to Main Content
A probabilistic forecast, in contrast to a point forecast, provides to the end-user more and valuable information for decision-making problems such as wind power bidding into the electricity market or setting adequate operating reserve levels in the power system. One important requirement is to have flexible representations of wind power forecast (WPF) uncertainty, in order to facilitate their inclusion in several problems. This paper reports results of using the quantile-copula conditional Kernel density estimator in the WPF problem, and how to select the adequate kernels for modeling the different variables of the problem. The method was compared with splines quantile regression for a real wind farm located in the U.S. Midwest.