Wind Energy Forecasting Using Recurrent Neural Networks

Abstract

Wind energy forecasting is a very challenging task as it involves many variable factors from wind speed, weather season, location and many other factors. Its accurate prediction can be quite helpful in maintaining the balance between demand and supply, and issues related to the reliability of a power system. In this article, the Recurrent Neural Network (RNN) based forecasting algorithm is used for the three day-ahead predictions of energy generation from wind sources in Estonia. Then a comparison is made between the predicted energy generation of Estonian energy regulatory authority's algorithm and this RNN based algorithm. The simulation results show that our proposed algorithm has lower Root Mean Square Error (RMSE) value and it gives better forecasting.