A GAN-based approach to solar radiation prediction: data augmentation and model optimization for Saudi Arabia

Area of Science:

• Renewable Energy Systems
• Artificial Intelligence
• Data Science

Background:

• Accurate solar radiation prediction is vital for renewable energy optimization but hindered by data scarcity and variability.
• Generative Adversarial Networks (GANs) are employed to generate high-quality synthetic solar radiation data, addressing data limitations.

Purpose of the Study:

• To develop a novel framework integrating GAN-generated synthetic data with machine learning and deep learning models.
• To improve the accuracy and adaptability of solar radiation prediction models across diverse climatic zones.

Main Methods:

• A framework integrating GAN-generated synthetic data with CNN-LSTM architectures was developed.
• Models were trained and evaluated using augmented datasets, enhancing predictive accuracy and generalization.

Main Results:

• Models trained on augmented datasets showed significant improvements: Root Mean Square Error (RMSE) reduced by 15.2% and Mean Absolute Error (MAE) decreased by 19.9%.
• The framework effectively bridged data gaps and enhanced model generalization for various climatic regions in Saudi Arabia.

Conclusions:

• The proposed framework supports practical applications like photovoltaic system optimization and grid stability.
• This scalable and adaptable approach aligns with Saudi Arabia's Vision 2030 and global renewable energy goals.
• Further research into computational complexity and hyperparameter sensitivity is recommended for advancing sustainable energy solutions.