Ultra-short-term wind power prediction based on hybrid denoising with improved CEEMD decomposition

Abstract

Ultra-short-term wind power prediction plays a crucial role in grid stability and energy management. To tackle the complexity of wind power data, this paper presents a Grey Relation Analysis-Improved Complete Ensemble Empirical Mode Decomposition (GCEEMD) method, integrating advanced denoising techniques and improvements. First, the Discrete Fourier Transform (DFT) and Discrete Wavelet Transform (DWT) are applied to analyze the data across multiple scales in both the frequency and time domains. This effectively extracts periodic features, removes noise, and preserves the dynamic characteristics of wind power. Next, Grey Relation Analysis (GRA) is introduced to optimize the CEEMD decomposition process by prioritizing key variables, which enhances the model's responsiveness to major trends and patterns. Finally, a Generative Adversarial Network (GAN) is combined with the NSGA-II algorithm to optimize component integration, significantly enhancing prediction accuracy and stability. Experimental results confirm the effectiveness of the proposed method when applied to real wind farm data.