Next-Generation PV Power Forecasting: Advances in AI Model Architectures, Sky Imaging, Weather Classification, and Data Validation Strategies

Accurate photovoltaic (PV) power forecasting plays a crucial role in enhancing grid stability and optimizing energy management in renewable power systems. Recent advancements in artificial intelligence (AI) and deep learning have significantly improved forecasting performance by leveraging advanced model architectures and diverse data sources. This review paper provides a comprehensive survey of state-of-the-art deterministic and probabilistic AI-driven PV power forecasting techniques, emphasizing the integration of sky imaging for solar irradiance prediction, clustering and weather classification for improved model generalization, and robust data validation strategies to mitigate overfitting. The latest developments in deep learning models, including hybrid architectures and mathematical models, are also examined. This study serves as a valuable resource for researchers and industry professionals seeking to advanced PV power forecasting methodologies.