A taxonomical review on recent artificial intelligence applications to PV integration into power grids

Abstract The exponential growth of solar power has been witnessed in the past decade and is projected by the ambitious policy targets. Nevertheless, the proliferation of solar energy poses challenges to power system operations, mostly due to its uncertainty, locational specificity, and variability. The prevalence of smart grids enables artificial intelligence (AI) techniques to mitigate solar integration problems with massive amounts of solar energy data. Different AI subfields (e.g., machine learning, deep learning, ensemble learning, and metaheuristic learning) have brought breakthroughs in solar energy, especially in its grid integration. However, AI research in solar integration is still at the preliminary stage, and is lagging behind the AI mainstream. Aiming to inspire deep AI involvement in the solar energy domain, this paper presents a taxonomical overview of AI applications in solar photovoltaic (PV) systems. Text mining techniques are first used as an assistive tool to collect, analyze, and categorize a large volume of literature in this field. Then, based on the constructed literature infrastructure, recent advancements in AI applications to solar forecasting, PV array detection, PV system fault detection, design optimization, and maximum power point tracking control problems are comprehensively reviewed. Current challenges and future trends of AI applications in solar integration are also discussed for each application theme.