基于DQN算法的多微网非计划孤岛切换策略  

Strategy of Unintentional Islanding Transition for Multi-Microgrids Based on DQN Algorithm

作　　者：贺旭辉 褚四虎 张羽 张雪菲 HE Xuhui;CHU Sihu;ZHANG Yu;ZHANG Xuefei(Hubei New Energy Microgrid Collaborative Innovation Center,Yichang,Hubei 443002,China;College of Electrical Engineering&New Energy,Three Gorges University,Yichang,Hubei 443002,China)

机构地区：[1]湖北省新能源微电网协同创新中心,湖北宜昌443002 [2]三峡大学电气与新能源学院,湖北宜昌443002

出　　处：《东北电力技术》2023年第8期14-18,共5页Northeast Electric Power Technology

摘　　要：配电网发生故障时,多微网会发生非计划孤岛对内部负荷供电产生较大冲击。针对这一问题,提出一种基于深度Q网络(deep Q network,DQN)算法的多微网非计划孤岛切换策略。首先,利用多微网非计划孤岛过程中的源荷储信息进行马尔科夫决策过程建模,并利用DQN算法对多微网的运行环境进行探索式学习,以找到最佳的减载策略。其次,执行该减载策略弥补由于非计划孤岛造成的多微网系统内部功率缺额,以保证多微网频率恢复正常并使减载损失最小。最后,基于改进IEEE-33节点的多微网模型对所提策略性能进行测试,测试结果表明了该策略的可行性和有效性。When the distribution network fails,multi-microgrids can cause unintentional islanding which has a great impact on the internal load power supply.In response to this problem,a strategy of unintentional islanding transition for multi-microgrids based on Deep Q Network(DQN)algorithm is proposed.Firstly,the Markov decision process is modeled by using the source-load-storage information in the multi-microgrid unintentional islanding process,and DQN algorithm is used to exploratory study for multi-microgrid operation environment,which finds the optimal load shedding strategy.Secondly,the load shedding strategy is implemented to make up for the internal power shortage of the multi-microgrid system caused by unintentional islanding,so as to ensure that the multi-microgrid frequency returns to normal and minimizes the load shedding loss.Finally,the performance of the proposed strategy is tested for an multi-microgrids model based on a modified IEEE 33-bus system.The test results verify the practicality and validity of the proposed strategy.

关 键 词：多微网 非计划孤岛 DQN算法 低频减载 

分 类 号：TM732[电气工程—电力系统及自动化]