Transportation electrification is a critical global policy, and the effective integration of electric vehicles (EVs) is key to ensuring power grid reliability and resilience. This dissertation proposes solutions to address the integration of fleet EVs (FEVs) into distribution grids, using cloud-based and edge-based approaches.
First, a centralized, cloud-based optimization model is developed for the strategic placement of FEV charging stations, designed to enhance grid resilience during high-impact, low-probability events like hurricanes.
Second, a bilevel optimization framework is introduced to aggregate FEV charging with other distributed energy resources (DERs), enabling their participation in the ISO day-ahead energy and reserve markets. This approach models FEVs as part of a Distributed Energy Resource Aggregator (DERA), providing energy and ancillary services.
Third, a novel cloud-edge collaboration framework is proposed to enable decentralized control of FEV charging stations at the grid edge. This framework uses federated reinforcement learning, allowing individual FEVs to coordinate their actions locally while contributing to voltage regulation and grid stability.
These solutions offer a comprehensive approach for optimizing the deployment and control of FEV charging stations, addressing both operational efficiency and market integration in modern power grids.