A Real-Time Vehicle-to-Grid Bi-Level Coordination Method Integrated With an Electric Vehicle Response-Compensation Mechanism

The increasing penetration of electric vehicles (EVs) brings both challenges and opportunities for modern distribution systems (DSs). This paper proposes a novel bi-level coordination framework integrating a response-compensation mechanism (RCM) to manage real-time vehicle-to-grid (V2G) interactions between the DS operator (DSO) and the EV aggregator (EVA). By integrating the RCM, the EVA can accurately report its real-time response capacity (RC) and the corresponding compensation bid. Consequently, the proposed framework effectively addresses the limitations of conventional coordination methods that fail to simultaneously consider EVs’ interests and real-time V2G demand. However, the integration of RCM introduces a complex coupling that renders conventional solution methods intractable. To address this issue, a lightweight neural network (LWNet) is employed to approximate the RCM mapping. The trained LWNet is subsequently converted into mixed-integer linear constraints and embedded into the upper-level DSO optimization, transforming the coupled bi-level problem into a tractable multi-stage program while reducing communication overhead and preserving data privacy. Simulation results validate the effectiveness of the proposed method, demonstrating accurate RC quantification and compensation calculation, as well as improvements in DS performance, including reductions in network loss and voltage deviation.