摘要
针对新能源出力存在不确定性的情况,提出了一种基于多代理技术的微网社区双阶段能量调度策略,分为日前调度和实时调度两个阶段。在日前调度阶段,微网间调度代理(Dispatch Agent)运用蒙特卡洛模拟技术生成场景,通过引入微网之间能量互传模式,计算对应的能量调度矩阵,运用粒子群算法代理(PSOAgent)求解该场景下最优日前调度计划。在实时调度阶段,微网i代理(MGi Agent)基于日前最优期望调度计划,接收新能源t时刻出力数据并与预测期望值进行比较,快速制定t时刻实时调度计划,满足微网i的t时刻供需平衡且运行成本最小,实现微网社区的经济运行。通过与传统多时间尺度方法的对比实验,仿真结果表明所提出的方法在实时调度环节比传统方法具有更小的调度偏差,并且可以有效减少微网社区的弃风弃光现象。
A double time scale scheduling strategy based on multi-agent technology is proposed for microgrid community system with renewable energy uncertainties,which includes day-ahead stage and real-time scheduling.In each particular scene of day-ahead stage,DispatchAgent uses the Monte Carlo simulation to generate scenario.By introducing the mutual energy transfer model between microgrids,the corresponding energy scheduling matrix is calculated.According to the economic operation of microgrid community,PSOAgent uses particle swarm algorithm to get the day-ahead scheduling under the scenario.In real-time scheduling,the output data of t moment of new energy are received by MGiAgent and compared with forecasted expected value,and it works out real time scheduling scheme of t moment quickly to meet the demand and supply balance of t moment of microgrid i and achieve cost optimum based on the expected day-ahead scheduling scheme,thereby realizing the economic operation of microgrid community.Compared with traditional time scale method,the proposed method has smaller scheduling bias than the traditional method in real-time scheduling,and can effectively reduce the abandon and exhaust phenomena in microgrid communities.
作者
王思明
牛玉刚
方磊
贾廷纲
WANG Siming;NIU Yugang;FANG Lei;JIA Tinggang(Key Lab of Advanced Control and Optimization for Chemical Process,Ministry of Education,East China University of Science&Technology,Shanghai 200237,China;Shanghai Electric Automation Group,Shanghai 200070,China)
出处
《电力系统保护与控制》
EI
CSCD
北大核心
2018年第17期89-98,共10页
Power System Protection and Control
基金
国家自然科学基金项目资助(61673174)
高等学校学科创新引智计划资助(B17017)~~
关键词
微网社区
不确定性
合作环境
蒙特卡洛模拟
多代理
实时调度
microgrid community
uncertainty
cooperation environment
Monte Carlo simulation
multi-agent
real-time scheduling
