基于数字孪生驱动的智慧微电网多智能体协调优化控制策略  

Multi Agent Coordinated Optimal Control Strategy for Smart Microgrid Based on Digital Twin Drive

作　　者：高扬[1] 贺兴[1] 艾芊[1] GAO Yang;HE Xing;AI Qian(Key Laboratory of Control of Power Transmission and Conversion(Shanghai Jiao Tong University),Ministry of Education,Minhang District,Shanghai 200240,China)

机构地区：[1]电力传输与功率变换控制教育部重点实验室(上海交通大学),上海市闵行区200240

出　　处：《电网技术》2021年第7期2483-2491,共9页Power System Technology

基　　金：上海市青年科技英才扬帆计划资助项目(20YF1418800)。

摘　　要：随着电力系统运行中负荷、气象、设备、电/气/热多能流等全景信息的日益增多,传统的电力系统建模仿真技术已很难适应未来智慧能源系统运行控制的需要,而结合机器学习、通信网络、高性能分析计算、物联网的数字孪生技术可以有效解决这一技术难题。首先,针对国内外不同领域对于数字孪生技术的基本定义和应用场景,讨论了面向智慧能源系统的数字孪生构建方式和未来发展思路,设计了考虑多通信协议交互的智慧微电网数字孪生和多智能体控制架构。考虑到物理空间、数字空间及耦合关系,对智慧微电网的全生命周期管理过程进行任务分解,构建调控资源智能体、调控过程智能体和调控服务智能体;然后,构建系统的数字和物理孪生体,利用传感器采集的分布式元件和周围环境的各项孪生数据,对数据库管理模块进行更新,进行组合功率预测和多目标优化决策;最后,构建基于Opal-RT半实物仿真平台的智慧微电网测试模型,利用用户数据报协议(user datagram protocol,UDP)通信协议实时感知试验数据,在微电网并网和孤岛2种工作场景下验证所提控制策略,结果表明:智慧微电网能够按照数字孪生优化后的试验方案在实际系统中执行,同时将执行情况与数字孪生系统实时交互,为全生命周期管理提供运行指导。With the increasing panoramic information of load,weather,equipment,electricity-gas-heat multi-energy flow in the power system operation,it is difficult for the traditional power system modeling and simulation technology to adapt to the needs of future intelligent energy system operation control.The digital twin technology combined with machine learning,communication network,high-performance analysis and calculation,and internet of things can effectively solve this technical problem.Firstly,according to the basic definition and application scenarios of digital twin technology in different fields at home and abroad,this paper discusses the construction modes and the future development ideas of digital twin technology for a smart energy system,and designs the digital twin and multi-agent control architecture of the smart microgrid considering multi communication protocol interaction.Considering the physical space,the digital space and their coupling relationship,this paper decomposes the whole life cycle management process of the smart microgrid,and constructs the regulation resource agent,the regulation process agent and the regulation service agent.Then,it constructs the digital and physical twin of the system,and uses the twin data of the distributed components and the surrounding environment collected by the sensors to update the database.Finally,the smart microgrid test model based on the Opal-RT hardware in the loop simulation platform is constructed,and the UDP(user datagram protocol)is used to sense the test data in real time to verify the control strategy proposed in this paper in the two working scenarios of microgrid grid connection and island.The results show that the smart microgrid is implemented in the actual system according to the optimized test scheme of the digital twin,and the implementation interacts with the digital twin system in real time to provide the operation guidance for the whole life cycle management.

关 键 词：组合功率预测 多目标优化决策 多智能体 UDP通信 半实物仿真 

分 类 号：TM721[电气工程—电力系统及自动化]