Short-term Wind Power Prediction based on FFA-BiLSTM Algorithm

Ensuring the effective and safe operation of wind power facilities hinges critically on accurate short-term wind power forecasting. However, the inherent volatility and unpredictability of wind power pose significant challenges to the stability and economic efficiency of power systems. To enhance the precision of short-term wind power forecasting and address these challenges, this study proposes an innovative hybrid prediction model. The model integrates Pearson correlation coefficient analysis, which helps in identifying key factors influencing wind power generation, with a bidirectional long short-term memory network (BiLSTM) whose hyperparameters have been optimized using the fennec fox algorithm (FFA). The FFA is known for its remarkable global search capability, making it ideal for finding the optimal set of hyperparameters for the BiLSTM. By leveraging the combined strengths of the FFA and BiLSTM, this model creates a prediction tool that excels in adaptability and can precisely capture subtle variations in wind power data. The BiLSTM's ability to learn and adapt to new patterns in the data, combined with the FFA's optimization capabilities, allows the model to generate more accurate and reliable forecasts. After training the model and applying it to short-term wind power forecasting tasks, we conducted comprehensive comparative experiments with existing models such as long short-term memory (LSTM), gated recurrent unit (GRU), and standard bidirectional long short-term memory networks. The experimental results demonstrate that the proposed FFA-BiLSTM model outperforms these models, exhibiting significant advantages in prediction accuracy and reliability.