A bi-level solution strategy based on distributed proximal policy optimization for transmission and distribution network dispatch with EVs and variable energy

Integrating large-scale wind power and extensive electric vehicle (EV) loads into the power grid impacts the system's safety and economic operations, posing challenges including frequent changes in grid dispatch instructions, unregulated EV charging and discharging behaviors, and increased network losses. Therefore, a bi-level optimization strategy model employing distributed proximal policy optimization for transmission and distribution network dispatch considering large-scale EVs is established, efficiently managing unit outputs and the system's capacity for charging and discharging, allocating these capabilities to individual nodes in real-time. The upper-level model focuses on minimizing the system's total operating costs, optimizing the operational status of thermal units, and regulating the number of EVs charging and discharging in the transmission network. The lower layer seeks to reduce the distribution network's total network loss costs by optimizing EV charging and discharging power, active/reactive power in branch circuits, and voltage levels at node charging stations. The best solutions for the upper-layer and lower-layer models are solved using the distributed proximal policy optimization (DPPO) method. The bi-level optimization model is tested on a modified IEEE-24 and IEEE-33 system and demonstrated by case studies.