Abstract:

With the development of battery technology, the range of electric vehicles (EVs) has greatly increased. As a result, EV users choose to charge near their destination instead of on-route. It can be seen that the charging demand of EVs has formed a kind of node-demand-based, flow-demand-assisted comprehensive demand pattern. Therefore, it is very necessary to take both node-based and flow-based demand into account in the deployment of the fast-charging station, so that the facility layout can fully match the charging demand pattern and improve their utilization. By combining the set covering location model and flow capturing location model (FCLM), this paper proposes a node-flow fusion location model that can serve the comprehensive charging demand. Considering the uncertainty of charging demand, it uses robust optimization to minimize the total planning cost of charging facilities and takes into account the grid restrictions and capacity expansion costs of power equipment. It establishes a robust location model under uncertain demand and solves it by equivalent dual transformation.  Finally, it conducts numerical analysis on two numerical experiments. The result demonstrates that compared with classic models, the node-flow fusion model proposed can effectively reduce the overall planning cost of charging facilities while fully covering different types of charging needs.

Key words:

electric vehicles, charging station, facility location, uncertain demand, robust optimization

摘要：

随着电池技术的发展，电动汽车续航里程日益增长，用户充电行为从在途充电逐渐向目的地充电转变，形成了以点需求为主、流需求为辅的综合需求模式。因此，规划充电设施选址需同时考虑点需求和流需求，以便设施布局与充电需求模式能够充分匹配，提高设施利用率。将经典选址理论中的集合覆盖模型与截流模型相结合，提出了一种能够服务综合充电需求的点流融合选址模型。考虑到充电需求的不确定性，采用鲁棒优化方法，以最小化充电设施建设和电力设备扩容总成本为目标，建立了需求不确定情形下的鲁棒选址模型，并通过对模型做等价对偶转换进行求解。最后，针对两个路网算例开展了数值分析，结果表明，与传统单一需求选址模型相比，本文提出的点流融合模型能在充分覆盖不同类型充电需求的情形下有效降低充电设施的总体规划成本。

关键词:

电动汽车, 充电站, 设施选址, 需求不确定, 鲁棒优化