[论文解读] Dynamic Modeling and Real-time Management of a System of EV Fast-charging Stations

Dingtong Yang, Navjyoth J.S. Sarma+2 more|arXiv (Cornell University)|Dec 17, 2020

Electric Vehicles and Infrastructure参考文献 47被引用 5

一句话总结

本文提出了一种用于区域电动汽车快速充电站网络实时管理与战略规划的随机动态仿真框架。该框架整合了多项式Logit选择模型与多服务器排队模型，以模拟电动汽车用户行为与充电站动态，同时测试了动态需求响应定价（DDRPA）方案——特别是二次DDRPA方案——相比静态定价，将平均等待时间减少了26%，社会福利提升了2.7%。

ABSTRACT

Demand for electric vehicles (EVs), and thus EV charging, has steadily increased over the last decade. However, there is limited fast-charging infrastructure in most parts of the world to support EV travel, especially long-distance trips. The goal of this study is to develop a stochastic dynamic simulation modeling framework of a regional system of EV fast-charging stations for real-time management and strategic planning (i.e., capacity allocation) purposes. To model EV user behavior, specifically fast-charging station choices, the framework incorporates a multinomial logit station choice model that considers charging prices, expected wait times, and detour distances. To capture the dynamics of supply and demand at each fast-charging station, the framework incorporates a multi-server queueing model in the simulation. The study assumes that multiple fast-charging stations are managed by a single entity and that the demand for these stations are interrelated. To manage the system of stations, the study proposes and tests dynamic demand-responsive price adjustment (DDRPA) schemes based on station queue lengths. The study applies the modeling framework to a system of EV fast-charging stations in Southern California. The results indicate that DDRPA strategies are an effective mechanism to balance charging demand across fast-charging stations. Specifically, compared to the no DDRPA scheme case, the quadratic DDRPA scheme reduces average wait time by 26%, increases charging station revenue (and user costs) by 5.8%, while, most importantly, increasing social welfare by 2.7% in the base scenario. Moreover, the study also illustrates that the modeling framework can evaluate the allocation of EV fast-charging station capacity, to identify stations that require additional chargers and areas that would benefit from additional fast-charging stations.

研究动机与目标

• 应对在区域部署受限背景下电动汽车快速充电基础设施日益增长的需求。
• 对由单一实体管理的多个快速充电站之间的相互依赖需求进行建模。
• 开发一种实时管理框架，以平衡充电需求并优化系统性能。
• 评估容量分配策略，以确定需要增加充电器或新站点的位置。
• 评估动态定价机制在提升系统效率与社会福利方面的有效性。

提出的方法

• 构建多项式Logit模型，以基于充电价格、预期等待时间及绕行距离来表征电动汽车驾驶员的站点选择行为。
• 在每个站点集成多服务器排队模型，以模拟供需之间的动态交互。
• 实施动态需求响应价格调整（DDRPA）方案，根据站点队列长度实时调整价格。
• 使用二次DDRPA函数通过提高拥堵地点的价格来激励负载在各站点间的均衡分配。
• 利用基于南加州真实数据校准的随机动态仿真框架对整个系统进行仿真。
• 在不同定价策略下评估系统性能指标，包括平均等待时间、收入、用户成本和社会福利。

实验结果

研究问题

• RQ1 充电价格、预期等待时间及绕行距离在多大程度上影响电动汽车驾驶员对快速充电站的选择？
• RQ2 动态定价在多大程度上能够减少快速充电站网络中的平均等待时间？
• RQ3 与静态定价相比，动态需求响应定价（DDRPA）如何影响收入与社会福利？
• RQ4 在区域网络中，哪些站点最可能面临拥堵？应在何处增加额外充电器？
• RQ5 该仿真框架能否有效支持快速充电基础设施扩展的战略规划？

主要发现

• 与无DDRPA基线情景相比，二次DDRPA方案将平均等待时间减少了26%。
• 与无DDRPA情况相比，二次DDRPA策略使充电站收入和用户成本分别提高了5.8%。
• 在基线情景下，二次DDRPA方案使社会福利提升了2.7%，表明系统整体效率更高。
• 建模框架成功识别出需要增加充电器的站点以及通过新建快速充电站可受益的区域。
• 站点间需求的相互依赖性要求实施系统级管理，而非孤立地优化单个站点。
• 基于队列长度的动态定价能有效重新分配充电需求，防止高使用率站点出现拥堵。