Wind Energy Forecasting Using Artificial Intelligence

Wind energy, like all renewable sources, poses the challenge of unpredictability. Making best use of wind energy involves learning the trends in which it generates power. When the power generated is more than the power demand, the power grid often fails, causing disconnection to the dependent parties and sometimes even expensive mishaps. Such incidents can be reduced by power balancing and handling frequency. Being informed of the amount of energy that will be generated enables power balancing. This knowledge will enable power engineers to divert excess energy to storage mechanisms to prevent such accidents. This paper proposes and compares various machine learning and a neural network to forecast exported wind power. Five popular machine learning regressor algorithms and long-short term models are compared with focus on the output target variable. A time series forecasting approach is proposed in the scenario of insufficient weather parameters of wind data.