A hybrid prediction framework based on deep learning for wind power

A wind power prediction model based on ensemble empirical mode decomposition and deep learning is proposed. Since wind power is closely related to wind speed, wind speed is highly uncertain, which brings volatility and non-stationarity to wind power sequential signal. In order to reduce the interference of non-stationary sequential signals and reduce the error of wind power prediction, a multi-branch neural network prediction model based on classical mode decomposition is proposed. In this model, the four-point method is used to eliminate outliers. In addition, convolutional neural network and long and short and term memory network are used to extract the time sequence of power. Finally, wind power prediction is realized through aggregation and integration.