Short-Term Wind Speed Prediction Based on MI-BI-LSTM

Baolin Ma

Abstract: Wind speed has randomness, volatility and uncontrollable factors, and with accurate wind speed prediction, wind turbines can be more efficiently controlled, and the grid-connection effects of wind generators can be reduced. The bidirectional long short-term memory neural network can provide the output layer with complete contextual information of the past and future of each point in the input sequence, and the wind speed is a time series signal coupled with some simple signals that change with time. Based on the above information, this paper proposes a short-term wind speed prediction method based on Multiple Input Bidirectional Long Short Time Memory (MI-BI-LSTM). Pearson correlation coefficient was used to verify the correlation of different data features, and the multi-dimensional data features with strong correlation were used as the input of the model to establish the wind speed prediction model. The simulation results show that this network can better fit the variation trend of actual wind speed series and has better prediction performance.

Keywords: wind speed prediction, Long and short term memory neural network, Multidimensional features