基于链式出行行为的电动汽车公共充换电设施规划研究

Because of the environmental pollution, instability of the international crude oil market and advances in battery technologies, interests in electric vehicles (EVs) have increased dramatically in recent years. Building public charging infrastructure in the convenient locations of urban areas plays a significant role in promoting EVs’ market acceptance. Previous studies on optimally deploying public charging infrastructure fail to accurately model the EVs’ travelling and recharging behavior, and also ignore the potential impact of building public stations on the entire transportation network. For the first time, this project will investigate the trip-chain-based network equilibrium of EVs, which accurately captures the spontaneous responses of EV drivers to the provision of public charging infrastructure, and comprehensively describes the impact of the charging infrastructure deployment plan on the network traffic distribution. The proposed equilibrium also considers the interaction between the service level of public charging infrastructure and the travel demand of EVs. Based on the network equilibrium of EVs, this project will then formulate the problem of optimally deploying public charging infrastructure for EVs within a given budget as a mixed integer nonlinear program, which simultaneously determines the optimal locations, capacities and types of public charging infrastructure. An efficient solution algorithm will be proposed, followed by the sensitivity analyses. The proposed modeling framework can provide theoretical supports to the policy makers’ decision making, and also serve as reference when companies investigate the public-charging-service market.

建设公共充、换电设施对推广电动汽车至关重要。考虑到资金投入有限和城市土地资源稀缺，设施的建设需要科学地规划与设计。已有的关于公共充、换电设施规划的研究缺乏对电动汽车出行与充电行为的准确刻画，缺少考虑大型公共站点的建设将对整个路网交通的潜在影响。本课题的研究将首次采用出行链的结构对电动汽车的行为进行均衡分析，从而准确描绘出行者在出行与充电时的自我寻优行为，全面评估公共充、换电设施布局方案对电动汽车的发展和整个道路交通的影响，探究设施服务水平与电动汽车出行需求的关系。基于上述链式网络均衡分析，课题进而对电动汽车公共充、换电设施规划问题进行数学建模并提出有效求解算法。在给定的预算范围内，规划模型将灵活地以不同的目标对设施的种类、选址和规模三者同时进行优化。研究成果既可帮助政府最大程度地服务电动汽车用电需求、有效引导和推动电动汽车的发展，又能为企业探索公共充、换电服务市场提供咨询。

近些年来，电动汽车得到了广泛普及，无论是公共交通还是私家汽车，电动车辆的比例日益提升。科学布局和建设公共充换电设施对推广电动汽车至关重要。然而，现有的相关研究较多地依赖假设条件，缺乏相关数据支撑，往往忽略出行者的链式出行特征，造成对电动汽车出行与充电行为的刻画不够准确。此外，现有研究对于所提规划模型的求解算法的探讨较多局限在某一特定类别模型的范畴内，缺乏针对多种模型的全面研究。本项目研究首先对电动汽车的链式出行进行建模，采用模型驱动和数据驱动两种方法，分别通过对链式出行特征的数学建模和对车辆运营大数据的深度挖掘，准确刻画电动汽车出行与充电行为。在此基础上，项目进行了网络均衡分析，并根据真实出行轨迹数据搭建仿真模型，形成了考虑多类型车辆和多类型充电设施的城市电动化交通运行模型。基于所搭建的电动化交通运行模型，项目进而建立双层规划模型，解决面向电动私家车的公共充换电设施规划问题、无线充电车道布设问题和混合车型下区域排放收费策略与充电站选址联合优化问题。除此之外，项目还运用数据驱动的方法解决面向电动出租车的充换电设施规划问题。基于北京市39,053辆出租车的实时轨迹数据，项目对充电设施的数量、规模和类型以及充电诱导策略进行了量化分析，并提出了规划建议。最后，项目针对多种双层规划模型的高效求解，研究和实现了多种启发式算法，包括活跃集合算法、惩罚序贯线性规划法、局部导向的全局优化算法等，项目还开发适用于电动公交车动态调度的近似动态规划框架。综上所述，项目研究所提出的模型与方法具有普适性，可高效地迁移到各种规模与特征的选址问题的建模与仿真分析，为电动汽车的推广与充换电设施的科学布设提供了强有力的理论支撑与实践指导。