Modeling and Forecasting Wind Energy Production by Stochastic Differential Equations

Abstract

Renewable energies are on the rise and their impact on the sustainability of our planet is consensual. Adequate tools for modeling and forecasting production from different sources are needed, so that management of energy resources is automatic and efficient. This work addresses this issue by a first modeling attempt of the wind power production in Continental Portugal using Stochastic Differential Equations (SDEs), based on available hourly observations. We resort to parametric SDE models proposed in the literature on wind energy research (the Ornstein-Uhlenbeck model and a transformed Ornstein-Uhlenbeck model), we estimate the model parameters, we perform the residual analysis and the short-term forecasting. We found that SDEs have produced useful results for the management of wind energy production. However, there would be an interest in evolving toward SDEs models that better explain the data in short periods of time, in order to obtain more reliable forecasts.