Short-Term Generation Forecasting Against the High Penetration of the Wind Energy

摘要

The current renewable energy sources penetration increase requires adequate planning for the safe and reliable system operation, due to this energy to be an intermittent source. Considering that, this study aims to propose a modeling structure and simulation in the short-term horizon, applied to forecasting generation capacity from wind farms located in the southern region of Brazil. In view of the stochastic characteristics of wind energy forecasting were generated multi-scenarios by the Monte Carlo (MC) method. Besides, wind generation forecasting was modeled by a structure of Multilayer Perceptron Artificial Neural Networks (MLP NNs) due to its learning capacity of complex non-linear relations between input and output variables from a database. Taking into account that from a detailed planning begins the process of expansion the new enterprises electric power generating, this study brings an interesting tool for to predict the availability of renewable energy generation, like wind source.