• CSCD核心库收录期刊
  • 中文核心期刊
  • 中国科技核心期刊

电力建设 ›› 2020, Vol. 41 ›› Issue (6): 60-68.doi: 10.12204/j.issn.1000-7229.2020.06.008

• 电动汽车 ·栏目主持 孙英云教授、胡俊杰副教授· • 上一篇    下一篇

基于局部电压幅值与云边协同的分散式充电桩充电协调方法

陈明庆1,3,李建林2,3   

  1. 1.先进输电技术国家重点实验室(全球能源互联网研究院有限公司),北京市 102211;2.北京市变频技术研究中心(北方工业大学),北京市 100144;3.中国电力科学研究院有限公司,北京市 100192
  • 出版日期:2020-06-01
  • 作者简介:陈明庆(1984),女,硕士研究生,主要研究方向为电力电子、电力系统控制及自动化; 李建林(1976),男,博士,教授级高级工程师,主要研究方向为大规模储能系统在电力系统中的应用。
  • 基金资助:
    国家电网公司科技项目“‘互联网+’业态下面向多元应用场景的充换电设施关键技术及运营评价研究与示范”(SGGR0000DLJS1800118)

A Charging Coordination Method for Distributed Charging Piles Considering Local Voltage Amplitude and Cloud-Edge Collaboration

CHEN Mingqing1,3,LI Jianlin2,3   

  1. 1.State Key Laboratory of Advanced Power Transmission Technology (Global Energy Interconnection Research Institute),  Beijing 102211, China;2. Inverter Technologies Engineering Research Center of Beijing (North China University of Techology), Beijing 100144, China; 3. China Electric Power Research Institute, Beijing 100192, China
  • Online:2020-06-01
  • Supported by:
    This work is supported by State Grid Corporation of China Research Program (No.SGGR0000DLJS1800118).

摘要: 针对配电系统中大量充电桩集中充电带来的谐波、负荷峰谷差和网络损耗增加等问题,文章提出了一种基于局部电压幅值与云边协同的分散式充电桩充电协调方法。在“云边端”电力物联网架构下,首先使用每个端设备(即充电桩)的电压幅值历史信息,分析各充电桩最大消耗功率等信息;然后通过电力物联网中边缘节点(配变终端)的边缘计算能力,收集充电桩数据并结合用户偏好、能源成本和电动汽车的历史路线等信息,分析过去充电桩的充电负荷数据曲线;最后,在Hadoop云平台采用深度信念网络(deep belief network,DBN)算法训练和预测未来负荷数据,建立目标优化模型。此外,基于蒙特卡洛模拟分析所提出的方法在不同的场景下的性能,以及与集中式充电协调方法进行比较。实验结果表明,所提方法实现了大量分散式充电桩的协同有序充电,减少负荷波动,降低充电负荷对电网的影响。

关键词: 局部电压幅值, 云边协同, 分散式充电桩, 深度信念网络(DBN), 负荷波动

Abstract:  In order to solve the problems of harmonics, load peak-valley difference and network loss caused by centralized charging of a large number of charging piles in the distribution system, a coordination method for decentralized charging piles on the basis of local voltage amplitude and cloud-edge collaboration is proposed. Under the cloud-edge-end power Internet of Things architecture, firstly, using the history information on voltage amplitude of each end device (charging pile) to analyze the maximum power consumption information of each charging pile; then, through the edge computing ability of edge nodes (distribution and transformation terminals) in the power Internet of Things, collecting the data of charging piles and combining the user preference, energy cost and the historical route of electric vehicles. Finally, in the Hadoop cloud platform, the deep belief network (DBN) algorithm is used to train and predict the future load data, and the target optimization model is established. In addition, the performance of the proposed method in different scenarios is compared with that of the centralized charging coordination method. The experimental results show that the proposed method realizes the coordinated and orderly charging of a large number of decentralized charging piles, reduces the load fluctuation and the impact of charging load on the power grid.

Key words: local voltage amplitude, cloud edge coordination, decentralized charging pile, deep belief network(DBN), load fluctuation

中图分类号: