Wind speed prediction and reconstruction based on improved grey wolf optimization algorithm and deep learning networks

The random and intermittent of wind speeds can affect the secure and stable functioning of wind turbines (WTs). To enhance the security and stability of WTs, accurate and effective wind speed prediction, and abnormal wind speed reconstruction are essential. This study proposes a wind speed prediction and reconstruction approach that combines an improved grey wolf optimization algorithm with an adaptive search strategy (SAGWO) and long- and short-term memory network (LSTM). Firstly, the wind speed data collected by the anemometer is analyzed to determine the corresponding correlation threshold and to select WTs with high similarity. Then, the wind speed data of highly similar WTs are taken as input variables and the LSTM is optimized using the grey wolf optimization algorithm with an adaptive search strategy to enhance the convergence speed and prediction ability. Finally, the SAGWO-LSTM model is built to reconstruct the wind speed data with added noise. The experimental outcomes indicate that the developed approach has a favorable performance concerning wind speed prediction and reconstruction compared to the baseline model.