City Readiness System Assessment of Electric Vehicle Charging Infrastructure*

Ning Wang, Yafei Liu, Gangzhan Fu

Abstract—The development of charging infrastructure not only affects the operating results of electric vehicles in demonstration cities, but also affects the purchase and use behavior of users. Based on literature and summary reports of demonstration cities, an assessment indicator system of 4 major factors and 13 observation indicators is developed on city readiness of charging infrastructure. By combining the advantages of partial least squares (PLS) path model, technique for order preference by similarity to ideal solution (TOPSIS) model and cluster model, a comprehensive city readiness assessment of charging infrastructure is made based on the empirical research of 25 demonstration cities in China. The results indicate that, some cities are doing better in the four areas of policy support and investment, charging infrastructure layout and planning, operation scope and effect, operation service and security. There are 6 indicators of the quantity of charging piles, quantity of cumulative charging times, vehicle types in service, quantity of cumulative charge for electricity, quantity of training people and total investment that have greater impacts on city readiness. The demonstration cities need to take the above factors and indicators into consideration when making policies to promote the development of charging infrastructure.

Keywords—Charging Infrastructure; City Readiness; Partial Least Square Path Modelling; Technique for Order Preference by Similarity to Ideal Solution; Cluster Model.

I. INTRODUCTION

In 2009，the Chinese government initiated the “Ten Cities, Thousand Vehicles” program for new energy vehicles in 13 cities. And in December 2012, the first three-year demonstration operation of 25 cities has been finished. The research and summary of these 25 cities indicated that they have only achieved 26% of the whole deployment goals. One of the most important reasons that restricts the implementation of the demonstration operation plan is that the planning and layout of the charging infrastructure is not reasonable, which directly affects the effect of the demonstration operation and the usersrsquo; purchase and usage behavior. In September 2013, the central four ministries and commissions introduced the development plan of new energy vehicles industry to realize the production and sales quantity of electric vehicles in 2015 and 2020, respectively up to 5 and 50 million vehicles. To achieve this goal, the central four ministries and commissions have selected 88 demonstration cities to launch a new round of demonstration operation and introduced incentives to encourage the purchase of electric vehicles. However, there are no specific policies for charging infrastructure which has critical impacts on the implementation of the new demonstration operation. And as the construction and operation of the charging infrastructure involves many stakeholders, the city readiness system assessment of electric vehicle charging infrastructure is particularly necessary. On the one hand, it can provide decision-making basis for developing specific subsidies or non-financial policies of charging infrastructure. On the other hand, it can also guide the demonstration cities to develop specific construction and operation strategies according to their various differences and local conditions. As a result, the new round of the demonstration operation can be effectively guaranteed to promote the development of electric vehicles industrialization.

There have been a lot of studies for the city readiness system assessment of electric vehicle charging infrastructure. Some studies incorporate two aspects into the charging infrastructure assessment: the charging infrastructure layout planning and the reliability of electrical energy. The reasonable charging infrastructure layout planning helps to reduce the 'range anxiety' of users. And the impacts of charging infrastructure distribution and capacity on the power grid load should be considered in the construction of charging infrastructure. Jian Liu [1] proposes an assignment model to distribute charging infrastructure and discusses the electrical load on the Beijingrsquo;s power grid. Galus et al. [2] assess the EV impacts based on the integration of power systems, transport systems and vehicle technology. The comprehensive evaluation methods are applied in the charging infrastructure assessment. Xizheng Liu [3] selects four aspects of technical indicators, economic indicators, environmental indicators and safety indicators in the evaluation indicator system to assess the overall performance of electric vehicle charging infrastructure. The uncertain analytic hierarchy process (AHP) is applied with the test data and expert opinions to analyze the charging infrastructure. Liqiao Ma [4] discusses the operation evaluation of the pure electric bus charging stations. With the methods of qualitative analysis and quantitative analysis, external social effects and environmental effects of the charging station operations in Xian are evaluated. The multi-indicator comprehensive evaluation methods are used for charging infrastructure assessment, which include the principal component analysis (PCA) and fuzzy evaluation method and analytic hierarchy process (AHP) [5]. The common problem of these methods is that the correlation between indicator variables is not considered. When there are serious problems of multicollinearity with the indicator variables, the role of certain variables in the evaluation system will be exaggerated [6]. In order to solve the problem, this paper employs the PLS path model method to establish comprehensive evaluation system, which can solve the problem of multicollinearity with the indicators [7]. And it is easy to study th

剩余内容已隐藏，支付完成后下载完整资料

电动汽车充电基础设施的城市准备系统评估

王宁，刘亚飞，付刚站

摘要：充电基础设施的发展不仅影响电动汽车在示范城市的经营成果,还会影响用户的购买和使用行为。基于文献和示范城市的总结报告, 在充电设施基本完善的发达城市，评估指标体系有4个主要因素和13个观测指标。通过结合偏最小二乘(PLS)路径模型和逼近理想解排序法(TOPSIS)模型以及集团模型的优点,基于在中国25个示范城市的实证研究，建立了一个全面的城市准备充电设施的评估。结果表明,一些城市在政策支持和投资、充电设施布局和规划、操作范围和效果、操作和安全服务这四个方面做得很好。其中，6个充电桩的数量指标为：累积数量的充电时间、车辆类型服务、累积量收费电力、训练人员的数量和总投资，这六个指标对城市完善有更大的影响。因此，示范城市在促进充电基础设施发展做决策时需要考虑上述因素及指标。

关键词：充电基础设施、城市发展、偏最小二乘路径模型、逼近理想解排序法、集团模型。

一、介绍

2009年,中国政府在13个城市启动了新能源汽车 “十城,千车”项目。2012年12月,第一个为期三年的25个城市的示范操作已经完成，根据25个城市的研究和总结表示,他们仅达到了26%的整体部署目标。限制演示操作的实现计划最重要的一个原因是充电基础设施的规划和布局不合理,这直接影响示范运行的效果和用户的购买和使用行为。2013年9月,中央四部委介绍了新能源汽车产业的发展规划,预计电动汽车的生产和销售数量在2015年和2020年分别为5万和5000万辆。为了实现这一目标,中央四部委选择88个示范城市推出新一轮的示范操作并且引入激励机制,鼓励购买电动车。然而,重要影响实现新示范操作的充电基础实施仍没有具体的政策。由于充电基础设施的建设和运营涉及许多利益相关者,因此电动汽车充电基础设施的城市准备系统评估尤为必要，一方面,它可以提供决策依据开发特定补贴或非金融充电基础设施的政策；另一方面,它还可以根据他们的各种差异和当地条件，指导示范城市发展特定的建设和运营策略。因此,新一轮的示范操作可以有效地保证并促进电动汽车产业化的发展。

电动汽车充电基础设施的城市准备系统评估现阶段已有很多研究。一些充电设施评估研究包含两个方面:充电基础设施布局规划和电能的可靠性。合理的充电基础设施布局规划有助于减少“里程焦虑”的用户；充电设施的影响分布和容量在电网负荷应考虑充电基础设施的建设。刘健[1]提出了一种分配模型分发充电设施和讨论北京电网的电力负荷；盖鲁斯等人[2] 基于集成电力系统,运输系统和汽车技术评估电动汽车影响—综合评价方法应用于充电设施的评估；刘希正[3]从四个方面：技术指标、经济指标、环境指标、安全指标的评价指标体系来评估电动汽车充电基础设施的总体性能——不确定层次分析法(AHP)应用与测试数据和专家意见分析充电设施。马丽桥[4]讨论了纯电动汽车充电站的运营评价。运用定性分析与定量分析的方法, 在西安充电站业务评估外部的社会影响和环境影响。多指标综合评价方法用于充电基础设施评估,包括主成分分析(PCA)和模糊评价方法和层次分析法(AHP)[5]。这些方法的共同问题是,指标变量之间的相关性是不需要考虑的。当存在严重的多重共线性问题的指标变量时评价体系中的某些变量的作用将会夸大[6]。为了解决这个问题,本文采用PLS路径模型方法建立综合评价体系,从而解决多重共线性的问题指标[7]。并且很容易研究每个潜变量之间的关系和显化变量集，一个综合变量,整合了所有潜在的变量和变量可以代表所有指标[8]。

在现状分析的基础上充电基础设施的容量在25个示范城市配套服务, 4个主要因素和13个观测指标的评估指标体系被建立在准备充电基础设施的城市。将PLS路径模型“降维,避免重复的信息和客观权重”的特征结合TOPSIS模型“限制条件少,操作简单”的特征开发PLS路径加权TOPSIS评价模型。25个城市总结报告的数据显示，评估电动汽车充电基础设施的城市准备利用PLS路径模型和PLS路径权重TOPSIS评价模型。然后两个模型的评价结果用于聚类分析。在此基础上,不同城市的发展对策和建议提出应结合城市充电基础设施建设和运营的实际问题。

二、评价指标体系

从文献综述和25个城市的示范总结报告中不难看出，充电基础设施建设与运营所需的财政和税收政策指导与支持包括土地,鼓励社会资本引入,电力安全应急措施与服务价格以及比例的补贴资金,总投资规模等[9]。充电基础设施的运营规模和服务能力反映在充电桩的数量,累计收费电力以及累计充电汽车[10]。监控中心的数量反映了安全操作和管理水平，车辆类型服务反映了基础设施类型的多样化 [11]，运行里程反映了服务效果，运营商反映了操作能力[12]。安全措施,包括监测、应急处理、安全知识培训,定期检查和维护,安全主体责任制度和相关的标准和规范为基础设施提供安全操作[13]。基于上述具体指标和内容、充电基础设施的城市准备可以表示的四个主要因素“政策支持与投资,充电设施布局与规划、经营范围与效果,操作服务与安全”。 根据全面性和可行性的原则, 4个主要因素和13个观测指标的评估指标体系在充电设施的发达城市已经建立，其潜变量和清单变量如表1所示。

三、PLS路径模型

1、PLS路径模型的建立

PLS路径模型由测量模型和结构模型两部分组成。测量模型描述潜变量之间的关系和显化变量,而结构模型描述了潜变量之间的关系。根据格威特提出的复杂的数据分析方法[14]，充电基础设施的城市准备的评价模型如图1所示。左边的图是显示变量组,代表了4潜在变量。右边的图是一个变量组所有显化变量组成。相应的潜变量表示为城市准备充电的基础设施。

2、PLS路径模型的估计与检验

根据PLS路径模型的需求，分别分析4组变量的主成分是一维测试（运用SPSS）的主要工作。从表二中可以看出，四个变量组的第一主成分特征值都大于1, 第二个主成分特征值都小于1, 因而所有的四个变量组均通过一维测试。

PLS路径模型主要采用平均变化量(AVE)和合成可靠性(CR)来评估模型的信度和效度。其一般标准为：AVE的值应大于0.5,CR的值应大于0.7[15]。从表3可见,每个潜变量的所有AVE和CR值均满足要求,因此PLS路径模型具有良好的有效性和可靠性水平。多重回归方程的拟合优度系数R2 = 0.9997表明充电基础设施的城市准备可以表示的最大级别的原始变量的信息。

3、PLS路径模型的结果分析

所有的显性变项均按照标准（运用SPSS）处理，运用SmartPLS软件计算外部重量和相关系数与相应的潜变量，其结果如表4所示。由表5和表6可以看出25个示范城市的评价结果。模型表明,充电设施布局和规划、经营范围和效果是影响充电基础设施的服务能力两个因素。与充电基础设施的服务能力最相关的6个显示变量是：充电桩的数量、车辆累积充电的数量、车辆类型服务、电力累积充电的数量、训练人员的数量和总投资，这6个显示变量在提高充电基础设施的服务能力时应该被予以考虑。

表1. 城市准备充电基础设施的评价指标体系

图1. 充电基础设施的城市准备PLS 路径模型

表2.一维测试结果

表 4. 外部权重和PLSPATH模型的相关系数