Empirical Mode Decomposition for Improved Wind Power Forecasting with Boosted GRU Model

This paper addresses the challenges related to forecasting wind energy, particularly focusing on short-term predictions essential for maintaining grid stability. With an increasing demand for electricity globally and concerns regarding reliance on fossil fuels, clean energy sources such as wind energy are gaining prominence. However, the unpredictable nature of wind poses technical obstacles for ensuring power system reliability. To tackle these challenges, the study proposes an innovative forecasting method that integrates decomposition and deep learning techniques. The proposed approach seeks to improve wind power forecasting(WPF) accuracy by employing techniques such as Empirical Mode Decomposition (EMD), Complete Ensemble Empirical Mode Decomposition with Adaptive Noise (CEEMDAN), and sophisticated deep learning models like Gated Recurrent Unit (GRU). The proposed hybrid approach integrating the CEEMDAN and GRU model outperforms the single model by an improvement of 62.5% in RMSE. Experimental outcomes demonstrate the efficacy of the approach in capturing intricate dynamics of wind power generation.