| 引用本文: | 黄开艺,艾芊,张宇帆,周嘉侃.基于多智能体系统的多区域互联分布式潮流算法[J].电力自动化设备,2020,40(7): |
| HUANG Kaiyi,AI Qian,ZHANG Yufan,ZHOU Jiakan.Multi-area interconnected distributed power flow algorithm based on multi agent system[J].Electric Power Automation Equipment,2020,40(7): |
|
| 摘要: |
| 针对多区域间潮流计算存在数据难以共享及数据隐私问题，提出一种基于多智能体系统(MAS)的多区域互联分布式潮流算法。首先，介绍基于MAS的多区域分布式潮流计算框架，并给出一种考虑数据保护的各级多智能体协调互动方式。然后基于主网、区域管理(包括外层管理智能体和内层管理智能体)两级多智能体结构，利用改进迭代投影搜索法构建分布式潮流内外层双重迭代算法，以实现多区域间通信少量非重要信息，各区域内指数快速收敛的目标。最后，分别以6节点、30节点、118节点系统为测试系统，并与集中式算法进行比较分析，验证所提算法的有效性和可拓展性。 |
| 关键词: 分布式算法 潮流计算 多智能体系统 改进投影搜索法 隐私保护 |
| DOI：10.16081/j.epae.202005020 |
| 分类号:TM711 |
| 基金项目:国家自然科学联合基金资助项目(U1866206) |
|
| Multi-area interconnected distributed power flow algorithm based on multi agent system |
|
HUANG Kaiyi1, AI Qian1, ZHANG Yufan1, ZHOU Jiakan2
|
|
1.Key Laboratory of Control of Power Transmission and Conversion, Ministry of Education, Shanghai Jiao Tong University, Shanghai 200240, China;2.State Key Laboratory of Ocean Engineering, Shanghai Jiao Tong University, Shanghai 200240, China
|
| Abstract: |
| To address the issue that the private information cannot be fully shared in a multi-area power flow calculation problem, a multi-area interconnected distributed power flow algorithm based on MAS(Multi Agent System) is proposed. Firstly, the framework of multi-area distributed power flow calculation based on MAS is introduced, and a coordinated interaction way of each agent at all levels considering data protec-tion is given. Secondly, a two-level multi agent structure of the main grid and regional-area management that includes both outer and inner management agent is developed. Meanwhile, an improved iteration pro-jection search approach is employed to establish a double-iteration algorithm for distributed power flow calculation, which can enable a very limited amount of non-important information in multi-area communication and achieve exponential fast convergence in each area. Finally, test results of the 6-bus, 30-bus and 118-bus systems verify the effectiveness and flexibility of the proposed algorithm by comparing with the centralized algorithm. |
| Key words: distributed algorithm power flow calculation multi agent systems improved projection search me-thod privacy protection |
