边缘计算架构下多电动汽车与电网互动协同优化控制方法

In the future, there will be a large number of electric vehicles(EVs) integrate to the distribution network. The current centralized scheduling mode of the power grid occupies too much computational resources, it is hard to support the interaction between EVs and the power grid when the number of EVs is huge. To solve the problem, based on the ‘edge computing’ architecture, this project proposes a multi-vehicle cooperative control method for interaction of EVs with the power grid, firstly establishes the edge computation model for multi-electric vehicle V2G scenario, and then carries out the key technology breakthroughs at three levels: optimal scheduling at the top, multi-vehicle coordination at the middle and power flow control at the bottom. At the upper level, An optimal scheduling method for complex distribution network is proposed, which regulates the charging and discharging power of EV groups to complement which with renewable energies. At the middle level, Multiple-edge and single-edge power cooperative allocation strategy are proposed to compensate the forecast error of renewable energies in situ; A coordinated optimization method for control the charging or discharging power of multi-electric vehicles is proposed to optimal distribute charging or discharging power to each EV. At the bottom level, A grid-friendly charging/discharging power control method for EVs is proposed, which executes superior power instruction and help to realize friendly access for EVs to the power grid; And an smart bi-directional charging power decision method for EVs is proposed to achieve autonomous interaction between EVs and the power grid when the communication between them is interrupted. The research of this project can solve some key problems in the interaction between multi electric vehicle and power grid, improve the performance of the distribution grid in terms of efficiency, stability, supply and demand interaction, etc., and help to promote the operation mode of the distribution network from "centralized control" to "distributed autonomy".

未来将有海量电动汽车接入配电网，目前电网的集中调控模式存在占用计算资源过多的问题，难以充分支撑大量电动汽车与电网供需互动。针对这一问题，本项目提出边缘计算架构下多电动汽车协同控制方法，建立多电动汽车与电网互动的边缘计算架构模型，并从上层优化、中层协同、底层控制多个层面开展关键技术突破：在上层，提出配网主站优化调度方法，调控电动汽车群充放电与新能源发电互补；在中层，提出多边缘间和单边缘内功率协同分配策略，实现对新能源预测偏差的就地补偿；提出多电动汽车间功率协调方法，优化分配各车充放电功率；在底层，提出电网友好型电动汽车充放电控制技术，提出通信孤岛下充放电功率自主决策方法，执行中层指令并实现对电网友好接入。项目研究成果可解决海量电动汽车接入配网场景下多电动汽车与电网互动中的关键性难题，满足主动配电网稳定运行、需求响应、供需互动等要求，推动配电网的运行模式从“集中调控”向“分布自治”延伸和发展。

在电网面临未来海量电动汽车终端接入的情景下，目前的电网调度控制方法和机制难以充分支撑削峰填谷、源荷协调、供需互动等主动配电网的核心功能，亟待研究规模化电动汽车与配电网供需互动协同控制新方法，推动配电网的运行模式从“集中调控”向“分布自治”延伸和发展。.本项目围绕“海量分布式接入配电网的多电动汽车与电网互动”这一关键性难题，提出基于边缘计算架构的多电动汽车协同优化控制理论及技术体系，研究了支持多电动汽车与电网互动的电力系统边缘计算网络架构，提出了配网协同规划理论、电动汽车集群充放电协同控制与管理方法、电动汽车去中心化管理和自治控制方法、通信孤岛下电动汽车需求响应自主决策方法等新理论和新方法，研发了电力边缘计算网关。项目研究成果有望解决未来海量电动汽车分布式接入电网场景下多电动汽车与电网互动协同运行控制中的一些关键性科学难题，满足电网的安全稳定运行、需求侧响应、多元用户供需互动用电等要求，确保电动汽车、新能源发电以及未来智能电网的良好可持续发展。