Modeling Temporal Correlation for Prediction of Wind Power

Abstract

The use of clean energy increase in the worldwide with an increasing number of people realizing the energy and environmental problems. Wind power is one of most likely to be widely used power, which can produce a huge amount of power by wind power stations. This paper used long short term memory (LSTM) to calculate the wind power curve. By analyzing the principle of the algorithms and establish the model to predict the real data, the feasibility of the model was proved. Furthermore, extensive comparison experiments are conducted to demonstrate the LSTM model has a highest accuracy and efficiency. In summary, the model can be used for the wind power curve forecast system.