Simulating scenarios and prediction intervals in wind power forecasting with the Beta distribution

Abstract

A methodology for the simulation of the wind power scenario for a short term horizon (one or two days in advance) is proposed. The covariance of the historical errors and the wind power forecast are use to generate a conditional random variable that represent the power wind production as a scenario. With the information provided by the scenario simulation, the energy deviation during a period and the prediction interval for each hour are obtained. The Beta distribution is used to represent the behaviour of the wind power production due to its better performance. With the results, it is possible to quantify the uncertainty of wind energy production. Finally, comparing the covariance and correlation of the simulated errors with historical errors, the procedure of the methodology is validated.