电动汽车集群内的递阶分散最优充电方法  

Hierarchical Decentralized Optimal Charging Algorithm for Electric Vehicles Aggregation

作　　者：李正烁[1] 郭庆来[1] 孙宏斌[1] 辛蜀骏[1] 

机构地区：[1]电力系统及发电设备控制和仿真国家重点实验室(清华大学电机系),北京市100084

出　　处：《电力建设》2015年第7期114-119,共6页Electric Power Construction

基　　金：国家重点基础研究发展计划(973计划)(2013CB228202);国家自然科学基金创新研究群体科学基金资助项目(51321005);国家自然科学基金(51361135703)~~

摘　　要：电动汽车未来有望以集群(例如大型充电站内的电动汽车)形式参与电网调度。收到电网下发的集群优化充电调度指令后,集群代理需优化集群内的电动汽车充电功率以追踪电网指令。该追踪问题是一个大规模的优化问题,难于集中式求解。提出一种递阶分散优化算法求解该问题。集群代理下发协调信息,各电动汽车根据协调信息分布式地优化自身充电功率,并向代理返回迭代信息,通过迭代求得问题的最优解。为进一步提高计算效率,对集群代理和单辆汽车的优化子问题进行了深入研究和算法改进。通过数值实验发现所提方法具有很快的计算速度,尤其适用于求解大规模电动汽车集群内的追踪问题。Electric vehicles（ EVs） are very likely to be dispatched by the power grid in the form of aggregation in the future,e. g.,the EVs in a large charging station being regarded and dispatched as one aggregation. After receiving the dispatch order from the power grid,an aggregator of EV aggregation should optimize each EV＇s charging power to realize the dispatch order with regard to the aggregation,which is defined as a tracking problem. The tracking problem is a largescale optimization problem which is hard to solve in a centralized manner. A hierarchical decentralized optimal algorithm was proposed to solve that problem. In the proposed algorithm,the aggregator sent coordination messages to EVs,while each EV solved a local small-scale optimization problem in a distributed manner and returned messages to the aggregator until the iteration process finished. To further enhance the computational efficiency,the sub-problem regarding the aggregator and each EV was deeply studied and its algorithm was improved. The numerical experiment results showthat the proposed method has fast calculation speed,and is especially suitable for the large-scale tracking problem inside an EV aggregator.

关 键 词：电动汽车 分散优化 集群代理 追踪问题 

分 类 号：TM910.6[电气工程—电力电子与电力传动] U469.72[机械工程—车辆工程]