Learning-Based Coordination of Distributed Energy Resources in Distribution Networks

This thesis investigates the intelligent coordination of distributed energy resources, focusing on electric vehicles as a representative mobile distributed energy resource type. Addressing challenges such as system uncertainty, battery degradation, and distribution network constraints, the study proposes a suite of reinforcement learning-based frameworks. These include decentralized EV coordination via multi-agent learning, safety-aware charging under voltage constraints, and a learning-based dynamic operating envelop approach for indirect control. Collectively, these methods enable scalable, reliable, and sustainable distributed energy resource integration, bridging the gap between user-centric flexibility and grid-level safety.