基于多率模态智能体集群优化的主动配电网协同控制与故障恢复重构方法研究

This project aims to address three essential scientific problems facing active distribution network (ADN) oriented to distributed clean energy sources, which are designing the system architecture based on multi-mode agent cluster via multi-rate signal communication, the cooperative control in ADN based on multi-mode agent cluster via multi-rate signal communication, and the reconfiguration for service restoration in ADN based on active stabilizing. ① Merging the cluster theory to the cooperative control of multi-mode agent cluster, and considering the optimized communication between the agent cluster, this project aims to present a multiple time scales three-tier architecture based on multi-rate signal communication. This architecture will improve the uniformity of multi-mode agent cluster and weaken the conservativeness. ② In order to improve the adaptive capacity for high permeability distributed generation (DG), this project aims to present a multiple time scales cooperative control strategy of flexible load, DG, and energy storage in ADN. Furthermore, the scheduling algorithm of regional autonomy and global compromise is to be proposed by constructing the relational model of each agent cluster. ③ In order to improve the reliability for high permeability of DG, this project aims to propose different active stabilizing refactoring plans for the different of DG and its distribution. Then, the reconfiguration method for service restoration in AND is to be presented. The project seeks to the goal to develop the admitting ability of DG in ADN, and to ensure the ability of service restoration reconfiguration at the same time, which build the whole prevention and control system of AND when facing the high permeability of DG. The implementation of this project is contributing to the safe operation of large-scale smart grids and consequently holding significant scientific interest and practical value.

本项目拟围绕建立多率模态智能体集群系统架构，基于多率模态智能体集群优化的主动配电网协同控制，以及主动致稳的配电网故障恢复重构方法三个基本科学问题开展研究。①融合集群理论、分析智能体集群多率模态特征，提出基于多时间尺度的智能体集群系统三层架构，完成不同类型通信网络下的智能体集群模型的建立，实现智能体集群间的一致性和弱保守性。②提出多时间尺度下的分布式发电(DG)、柔性负荷及储能间的协调控制策略，构建各智能体集群之间的关系模型，提出主动配电网的区域自治及全局协同调度算法。③为了提高故障时的可靠性，依据不同DG及其分布情况，制定不同主动致稳重构方案，并提出主动配电网的故障恢复重构方法。通过本项目的实施，在有效提高主动配电系统对高渗透率DG接纳能力的同时，还保证了系统具有故障恢复重构的能力，实现高渗透率下主动配电网的全过程防控体系，对电网安全稳定运行具有重要的科学意义和实用价值。

我们承担的国家自然科学基金面上项目“基于多率模态智能体集群优化的主动配电网协同控制与故障恢复重构方法研究”（61773099）按照原定计划顺利完成。没有对原计划做出调整，已经取得了预期的研究成果，圆满完成了所制定的研究计划。在本项目完成过程中，发表论文17篇，SCI检索9篇。同时，我们将获得的理论成果应用于工业生产实际，取得了很好的经济效益和社会效益。申请并授权国家发明专利48项，获得吴文俊人工智能技术发明奖二等奖1项，获中国专利优秀奖1项。. 在此基金的资助下，组织国际学术交流活动4次，指导博士生2名，研究生3名。现已有毕业博士2名，毕业硕士3名。. 本基金完成了原申请预计的建立了三层多时间尺度智能体集群优化框架；定义了配电系统自治区域的划分原则，构建了基于多率模态的智能体集群间通信技术及相应智能体集群系统模型；研究基于多率模态智能体集群的协同控制理论及方法；提出了基于多率模态智能体集群的故障恢复重构的控制策略等4项主要研究内容。该项研究内容的深入突破，将有助于主动配电网协同控制与故障恢复重构，极大提高配网对分布式发电(distributed generation，DG)的接入能力，同时增强其供电可靠性，提高对DG的利用率，增强主动配电网鲁棒性。