ICEEMDAN-Informer-GWO: a hybrid model for accurate wind speed prediction

Area of Science:

• Renewable Energy Systems
• Computational Intelligence
• Environmental Engineering

Background:

• Growing global energy demand and environmental concerns necessitate advancements in wind energy technologies.
• Accurate wind speed prediction is critical for the efficient integration and operation of large-scale wind power systems.
• Existing prediction models often face challenges in accuracy and computational efficiency.

Purpose of the Study:

• To introduce and evaluate a novel hybrid framework, ICEEMDAN-Informer-GWO, for enhancing wind speed prediction accuracy.
• To leverage advanced signal decomposition and optimization techniques for improved forecasting.
• To assess the framework's performance across different wind farm datasets and prediction horizons.

Main Methods:

• Utilizing Improved Complete Ensemble Empirical Mode Decomposition with Adaptive Noise (ICEEMDAN) for robust wind speed data decomposition.
• Employing the Informer model for computationally efficient wind speed predictions.
• Integrating Grey Wolf Optimization (GWO) algorithm to fine-tune Informer model parameters for superior performance.

Main Results:

• The proposed ICEEMDAN-Informer-GWO framework demonstrated superior performance in wind speed prediction.
• Statistically significant improvements in accuracy were observed across all tested time horizons (5 minutes, 30 minutes, and 1 hour ahead).
• Consistent high performance was validated using diverse wind farm datasets from Block Island, Gulf Coast, and Garden City.

Conclusions:

• The hybrid ICEEMDAN-Informer-GWO framework offers a highly effective solution for accurate wind speed forecasting.
• This approach significantly enhances the reliability of wind energy integration into power grids.
• The study confirms the framework's robustness and applicability in real-world wind energy scenarios.