Wind power generation prediction using LSTM model optimized by sparrow search algorithm and firefly algorithm

Abstract

As an important renewable energy source, wind power generation is highly stochastic and uncertain due to various environmental factors affecting its output. To raise the accuracy of wind power generation prediction, a bidirectional long short-term memory network combination model based on sparrow search algorithm and firefly algorithm optimization is designed. The model first employs a bidirectional long short-term memory network to capture the long-term dependency features of time series, and uses random forests for nonlinear modeling and feature selection. Then, the sparrow search algorithm and firefly algorithm are combined to optimize the hyperparameter configuration, improving the predictive performance and global search ability of the model. The findings denote that the accuracy of the designed model reaches 98.5%, with a mean square error as low as 0.005 and a prediction time as short as 0.18 s. The simulation analysis results show that the predicted values of the developed model almost coincide with the actual values, with small errors. The research outcomes denote that the optimized model greatly raises the accuracy and efficiency of wind power generation prediction, and has good application prospects.