Short and medium-term wind power forecasting using a novel time series dynamic context-aware GRU model

Wind energy is a low-carbon source that has experienced significant technological advancements in recent decades. The random and unpredictable nature of wind poses distinct challenges for accurate wind power prediction. Accurate forecasts of wind power generation are crucial for the effective management of energy grids. This study introduces a novel, transformer-based, dynamic context-aware power forecasting gated recurrent unit (GRU) hybrid model that adapts the transformer architecture, incorporating innovative modifications to address the complications of wind power forecasting. It improves the extraction of contextual features while simplifying the model's structure. The results for the Jhimpir dataset demonstrate normalised mean absolute error (NMAE) values of 0.636, 1.747, 1.581, and 1.640 for the hybrid dynamic context-aware GRU model, transformer, long short-term memory (LSTM), and single GRU model, respectively. The proposed model outperforms other models, achieving the lowest prediction error, and provides comprehensive knowledge to decision-makers interested in wind turbines and energy optimisation.