Artificial Intelligence and the Global Energy Transition: A Review of Computational Applications in Renewable Energy Systems

Abstract

Artificial Intelligence (AI) has emerged as a transformative technology offering essential solutions for optimizing sustainable energy systems and accelerating global decarbonization. This scientific literature review systematically examines advanced Machine Learning (ML), Deep Learning (DL), and Reinforcement Learning (RL) techniques deployed across renewable energy infrastructure. The analysis confirms the superiority of DL architectures, such as Hybrid Deep Neural Networks (HDNNs), in complex time-series forecasting, significantly improving the accuracy of solar and wind power predictions. Furthermore, RL algorithms have proven critical for dynamic energy management, achieving substantial operational improvements, including a 23.5% increase in cost savings and a 78.69% reduction in carbon emissions in Battery Energy Storage System (BESS) operation. AI applications also extend to critical operational control, enabling predictive maintenance for wind turbines and enhancing PV energy efficiency by 15–20% through automated fault detection. While the technical feasibility is robust, the review identifies critical systemic challenges, including data standardization, high computational costs, cybersecurity risks , and the necessity of developing standardized protocols and Safe RL frameworks to ensure the resilient and scalable implementation required for the full energy transition.