Pang Songling, Fan Kaidi, Huo Meiyi
Electric Power Research Institute of Hainan Power Grid Co., Ltd., Haikou, 570311, China.
Smart Grid and Island Microgrid Joint Laboratory, Haikou, 570311, China.
Sci Rep. 2025 May 9;15(1):16202. doi: 10.1038/s41598-025-00265-2.
This paper addresses the challenge of charging and discharging scheduling for large-scale electric vehicles (EVs) in the Vehicle-to-Grid (V2G) mode by proposing a user-oriented scheduling algorithm. First, a large-scale EV charging and discharging scheduling model grounded in the V2G mode is developed, where the objective function mainly focuses on the load variance at the user side and the charging and discharging costs for EV owners, and constraints such as the available time of EVs, charging and discharging power limits, available state of charge values, and upper and lower bounds of real-time prices are incorporated to make the model more applicable to practical engineering scenarios. Based on this model, a multi-level grouping based competitive swarm optimizer (MLGCSO) is put forward. Compared with traditional methods, the diversity and convergence of particle swarm learning are enhanced, and the optimization performance is improved. Simulation results indicate that when compared with three state-of-the-art optimizers, the optimization accuracy of the proposed algorithm is increased by at least 34% and the total cost is reduced by 3.14% and 1.62% respectively, demonstrating that the MLGCSO exhibits high optimization performance and remarkable optimization effects.
本文通过提出一种面向用户的调度算法，解决了车辆到电网（V2G）模式下大规模电动汽车（EV）充放电调度的挑战。首先，建立了基于V2G模式的大规模电动汽车充放电调度模型，其目标函数主要关注用户侧的负荷变化以及电动汽车车主的充放电成本，并纳入了电动汽车的可用时间、充放电功率限制、可用荷电状态值以及实时电价上下限等约束条件，以使模型更适用于实际工程场景。基于该模型，提出了一种基于多级分组的竞争群优化器（MLGCSO）。与传统方法相比，增强了粒子群学习的多样性和收敛性，提高了优化性能。仿真结果表明，与三种先进的优化器相比，所提算法的优化精度至少提高了34%，总成本分别降低了3.14%和1.62%，表明MLGCSO具有较高的优化性能和显著的优化效果。
