A comprehensive review on sustainable energy management systems for optimal operation of future-generation of solar microgrids

Conventional microgrids face a number of challenges due to intermittency of renewable energy resources and the lack of any effective energy management system. Thus, there is a need to develop a secure, stable, dynamic, microgrid with an AI-based energy management system for sustainable power generation to accomplish the targets of the United Nations Sustainable Goals (SDG7, SDG11 and SDG13). The novelty of the study is that it addresses these challenges, categorizes microgrid problems into optimal power flow, peak-shaving, and optimal network configurations and identifies the most recent metaheuristics, reinforcement learning, and blockchain technology as most promising techniques for microgrid energy management, decentralization, and cyber security. The modeling strategies and formulations, optimization objectives, and solving algorithms, are also described along with metaheuristics (evolutionary, swarm, physics-based, human-based, hybrid, and others), machine learning (model-based control, reinforcement learning, fuzzy logic), and blockchain techniques as reliable decentralized techniques. The results of case studies show that microgrid management systems can be implemented differently considering the size of the system, connectivity with the grid, technology used, capital cost, and automation. Follow-up research areas are also identified to address emerging challenges for the development of future AI applications, for conventional and future generation of renewable energy-based microgrids.