| 引用本文: | 李亚平,杨胜春,毛文博,高冠中,陆亚楠,黄展鸿.基于群体智能的分布式柔性资源有功平衡调度架构及策略[J].电力自动化设备,2022,42(7): |
| LI Yaping,YANG Shengchun,MAO Wenbo,GAO Guanzhong,LU Yanan,HUANG Zhanhong.Active power balance scheduling architecture and strategy of distributed flexible resource based on collective intelligence[J].Electric Power Automation Equipment,2022,42(7): |
|
| 摘要: |
| 针对电力系统中分布式柔性资源数量众多、分散分布、不确定性强给调度运行带来的挑战，引入新一代人工智能中的群体智能思想，提出了基于群体智能的分布式柔性资源有功平衡调度架构。按照分层分布集群控制模式将海量柔性资源的组织与调控运行分为终端→用户→子群→群体4层。在该架构下，用户、子群、群体每层看作是不同的智能体，分别提出了群体的外特性建模、群内自治决策和群间交互协同策略，实现了“弱中心化”的群体自律运行。仿真结果验证了分布式架构的合理性和智能策略的有效性。借助分布式调控架构和新一代人工智能技术是实现海量分布式柔性资源“群调群控”的有效手段。 |
| 关键词: 海量柔性资源 分布式架构 群体智能 自治决策 交互协同 |
| DOI：10.16081/j.epae.202206005 |
| 分类号:TM73 |
| 基金项目:国家电网有限公司总部科技项目(能源互联网环境下基于群体智能的多类型可调控资源协同调度策略研究) |
|
| Active power balance scheduling architecture and strategy of distributed flexible resource based on collective intelligence |
|
LI Yaping1, YANG Shengchun1, MAO Wenbo1, GAO Guanzhong1, LU Yanan1, HUANG Zhanhong2
|
|
1.China Electric Power Research Institute(Nanjing),Nanjing 210003, China;2.School of Electric Power, South China University of Technology, Guangzhou 510641, China
|
| Abstract: |
| Aiming at the challenges of scheduling operation brought by the large number, distributed location and strong uncertainty of distributed flexible resources in power system, the collective intelligence idea of the new generation’s artificial intelligence is introduced, and the active power balance scheduling architecture of distributed flexible resource based on collective intelligence is proposed. According to the hierarchical and distributed cluster control mode, the organization and regulation operation of massive flexible resources is divided into four layers: terminal → user → sub-cluster → cluster. Under this architecture, user layer, sub-cluster layer and cluster layer are regarded as different agents, and the external characteristic modeling of cluster, the autonomous decision-making within cluster and the interactive collaboration strategy between clusters are proposed respectively, thus realizing the “weakly centralized” cluster self-discipline operation. Simulative results verify the rationality of distributed architecture and the effectiveness of intelligent strategy. With the help of distributed regulation architecture and the new generation’s artificial intelligence techno-logy, it is an effective means to realize the “cluster scheduling and control” of massive distributed flexible resources. |
| Key words: massive flexible resources distributed architecture collective intelligence autonomous decision-making interactive collaboration |
