Short term wind and energy prediction for offshore wind farms using neural networks

Forecasting short term wind speed is of high importance for wind farm managers. The knowledge of the expected winds helps taking decisions (decision support) as the likes of maintenance and repair jobs or finishing works as health and safety is not guaranteed anymore. There are a number of methods and computations currently being used for forecasts: fuzzy logic, linear prediction or neural networks. For the latter there are also various algorithms and methods, from feed forward up to recurrent neural networks (RNN) and long short-term memory (LSTM). Recurrent neural networks belong to the group of machine learning algorithms and are part of artificial intelligence research. This paper is about forecasting wind speed and pressure using RNN.