Artificial intelligence and machine learning for distributed energy resource management systems: Applications, frameworks, and future directions

The integration of distributed energy resources (DERs) introduces significant operational challenges to conventional power systems due to their decentralized, variable, and bidirectional nature. This paper presents a comprehensive review of artificial intelligence (AI) and machine learning (ML) techniques applied within distributed energy resource management systems (DERMS). Core applications, such as forecasting, optimization, real-time control, demand-side management, energy trading, and cybersecurity, are systematically analyzed using a structured taxonomy that encompasses deep learning, reinforcement learning, federated learning, and explainable AI. A layered framework is proposed to align the AI/ML lifecycle with DERMS functional layers, bridging the gap between theoretical research and practical deployment. Benchmark datasets are designed to standardize algorithm evaluation, while comparative analyses across forecasting, optimization, and fault detection reveal performance differences among techniques. Economic case studies and cost-benefit assessments further underscore the operational and financial benefits of AI-enabled DERMS. Finally, the paper identifies prevailing limitations across edge AI, hybrid modeling, and privacy-preserving learning, and outlines future directions for intelligent, resilient, and scalable distributed energy systems. Overall, this study serves as a foundational reference for researchers, industry stakeholders, and policymakers seeking to advance smart, resilient, and scalable DERMS for future energy infrastructures.