低压配电台区柔性互联系统多目标DDPG优化调度  

Multi-objective DDPG Optimal Dispatch for Low-voltage Distribution Station Area Flexible Interconnection System

作　　者：刘文军 李帅虎 马瑞[3] 何书耘 LIU Wenjun;LI Shuaihu;MA Rui;HE Shuyun(State Grid Hunan Electric Power Economic&Technology Research Institute,Changsha 410007,China;Hunan Key Laboratory of Energy Internet Supply-demand and Operation,Changsha 410007,China;School of Electrical&Information Engineering,Changsha University of Science and Technology,Changsha 410114,China)

机构地区：[1]国网湖南省电力有限公司经济技术研究院,湖南长沙410007 [2]能源互联网供需运营湖南省重点实验室,湖南长沙410007 [3]长沙理工大学电气与信息工程学院,湖南长沙410114

出　　处：《智慧电力》2024年第6期62-70,共9页Smart Power

基　　金：国家自然科学基金资助项目(52277076);国网湖南省电力有限公司科技项目(5216A520000S)。

摘　　要：针对配电台区柔性互联系统(DSAFIS)中源、荷、设备强不确定性,及台区间功率互济特性,提出基于深度确定性策略梯度(DDPG)的台区运行成本、新能源消纳、负载均衡目标协调优化调度方法。构建系统模型与物理系统自动联动的深度强化学习日前优化调度决策框架,设计考虑多目标奖励和运行约束奖励的优化调度DDPG模型;DDPG采用“在线学习”模式,算法收敛后输出日前调度计划给实际DSAFIS。算例验证了所提方法能自动适应系统的强不确定性,且在降低运行成本的同时兼顾了新能源消纳和台区负载均衡。Aiming at strong uncertainty of source,load and equipment and the power mutual aid characteristics between substation areas in distribution station area flexible interconnection system(DSAFIS),a coordination and optimization dispatch method based on deep deterministic policy gradient(DDPG)for the operation cost,new energy consumption,load balancing is proposed.A deep reinforcement learning day-ahead optimal dispatch decision framework with automatic linkage between system model and physical system is constructed.The optimal dispatch DDPG model considering multi-objective reward and operation constraint reward is designed,DDPG adopts an online-learning mode,the day-ahead dispatch plan is output to the actual DSAFIS after the algorithm converges.An example is given to verify that the proposed method can automatically adapt to the strong system uncertainty,and can reduce the operating cost while taking into account the new energy consumption and the load balance of substation area.

关 键 词：配电台区柔性互联系统 日前优化调度 DDPG 多目标 负载均衡 

分 类 号：TM734[电气工程—电力系统及自动化] TP18[自动化与计算机技术—控制理论与控制工程]