【作者】 王建；

【导师】 刘志珍；

【作者基本信息】 山东大学 ， 电机与电器， 2013， 博士

【摘要】 发展高效、清洁的电动汽车,是降低化石能源依赖性和缓解大气污染的有效途径。近几年,美国、日本、西欧的电动汽车市场增长势头强劲,进入快速发展通道的迹象愈发明显。2013年以来,国内私人购车补贴政策逐渐落地,市场可选择电动汽车产品逐渐增多,未来几年国内外可能出现电动汽车市场爆发式增长局面,势必会对电网规划运行产生巨大影响。随着电动汽车规模化应用,电网原有装机和输变电容量是否能够应对新增充电负荷需求将成为重要的研究课题。在不扩大电网建设规模的情况下,如何提高原有电网利用率,提高容纳能力,同时最大限度降低充电负荷对电网的负面影响,有必要提前研究电动汽车充电对电网各方面的影响,提出相应对策,对于推动电动汽车产业化进程具有重要意义。本文的研究内容主要包括以下几方面：(1)在谐波分析理论基础上,利用PSCAD软件,建立充电机/站仿真模型,研究充电机谐波电流幅值、相位随输出功率的变化关系,并与实际谐波测试结果对比,验证了模型的科学性、有效性。以此为基础,比较分析目前应用比较广泛的四种充电机在谐波抑制效果、功率因数、经济成本等方面的差异,为充电机的选型提供了参考依据。结合多台充电机的起始工作时间、电池初始充电状态等随机因素,建立N台充电机集中充电时的谐波电流概率分布及衰减特性模型,通过仿真分析多台充电机在不同情景下的谐波电流特性,验证该模型的有效性和实用性,并从工程建设实际出发提出谐波抑制方法。以临沂焦庄电动汽车充／换电站为实例进行了电动汽车充电站APF补偿容量计算分析,为电动汽车充换电站谐波治理提供了依据。(2)从分析电动汽车负荷建模的影响因素(电池特性、充电模式、用户行为习惯等)入手,建立电动汽车充电负荷模型,提出基于优化电网负荷波动方差以降低网损的方法。充电负荷模型中以美国2009NHTS调研报告和焦庄充电站实测数据为基础研究用户行为习惯,运用概率分布理论建立了电动汽车日行驶里程及电动汽车充电时刻的概率密度函数,确保了模型的准确性。以山东某城市10kV双辐射线路为实例进行仿真验证,仿真结果表明不同渗透率情况下,电动汽车无序充电对电网负荷、电压质量、网损等方面造成一定的负面影响,而提出的优化电网负荷波动方差的方法,可有效降低系统网损和电压偏移。(3)提出采用集中式与分布式结合的动态有序充电控制策略。首先,针对电动汽车规模化应用及无序充电对电网规划运行带来的负面影响,依据目前分时电价的划分原则及电动汽车行驶规律,建立电动汽车充电时段控制策略。该策略虽然在一定程度上改善了负荷曲线,但缺乏灵活性。在充分满足用户利益的前提下,进一步提出以平抑负荷、用户充电成本最小化为目标,通过构建中间管理者运行架构模式,采用集中式与分布式结合的优化控制策略,建立了协调控制模型。集中式控制以电网负荷波动最小化作为优化目标。分布式控制利用动态时间窗实时更新电动汽车充电的动态信息,以充电需求与集中式优化控制结果之间的偏差、用户充电成本最小化以及延长电池寿命为优化目标,实现电动汽车充放电的动态优化控制。以山东电网2015、2020、2030年夏季典型日负荷数据为算例,对比分析时段控制和有序充电控制策略对配电网负荷特性的影响,仿真结果表明有序充电控制策略平抑负荷波动、提高电网经济运行水平效果明显。(4)提出基于支持向量机(support vector machine, SVM)的迭代预测方法,对山东2015、2020、2030年的电动汽车保有量进行预测,并与常规预测法进行比较分析,为后续研究电动汽车规模化应用对山东电网负荷特性的影响及不同能源结构下的经济环境效益提供了量化依据。目前国内外研究大部分采用直接假设、弹性系数法或千人保有量法等较为粗略的预测方法,缺乏更为科学准确地预测,而研究电动汽车规模化应用对电网的影响及效益分析的前提是准确预测电动汽车规模数量。用近三年的汽车保有量数据进行验证,基于支持向量机的迭代预测方法预测误差在1.68%以内,具有较高的准确度。(5)充分考虑行业边际效应、能源结构演变等因素,建立经济环境效益模型,对不同能源结构、不同渗透率下电动汽车充电的经济、环境效益进行了量化研究。针对目前社会各界讨论的电动汽车应用能否真正实现低碳的热点,本文全面计算分析了有序充电的短期和长期可以避免的边际供电成本,量化分析了有序充电的经济效益。调研分析山东能源结构发展变化趋势,将能源结构发展趋势划分为三种不同情景：传统能源结构、过渡能源结构和新能源结构。在此基础上,建立电动汽车碳排放模型,计算分析不同能源结构、不同渗透率下2015年、2020年、2030年传统燃油汽车、电动汽车无序充电、电动汽车有序充电的碳排放量,结果表明随着能源结构变化,新能源比例逐渐上升,电动汽车应用日趋规模化,电动汽车有序充电减少碳排放量的效果显著增加。更多还原

【Abstract】 To develop the efficient, clean electric vehicles, is an effective way to reduce fossil fuel and alleviate air pollution. In recent years, in the United States, Japan and Western Europe, electric vehicle market grows strongly and apparently, which enters into a rapid development period. In2013, with the implementation of domestic private EV subsidies, electric vehicle products will gradually increase in the next few years and there would be an explosive growth of EV market at home and abroad. The popularity of EVs will bring a negative impact on planning and operation of the grid. It’s not clear whether the existing installed capacity and the power transmission and transformation system capacity would be able to cope with new demand of charging load, with the large-scale integration of EVs. How we can improve the existing grid utilization and improve carrying capacity and minimize the negative effect brought by the charging load, without expanding the construction scale of power grid? It is necessary for us to study the negative effect and to take appropriate measures in advance, which will have an important significance in the promotion of EVs industrialization process. This thesis mainly includes the following aspects:(1) Based on the harmonic principle, three-phase bridge rectifier charger is analyzed, and models for charger and charger station has been set up, by using PSCAD as a simulation tool. The paper mainly discussed the relationship between the amplitude and phase of charger harmonic, and the output power. Four popular kinds of chargers are analyzed, especially in the power factor, cost benefits and other aspects, which provide a reference for the selection of charger. From the aspects of charging time, initial state of charging, charger characteristics are analyzed to study the probability distribution of the harmonic current and attenuation characteristics while N chargers centrally charging. Harmonic current characteristics of N chargers are simulated by simulation software PSCAD, considering charging time, initial SOC of battery. Finally, take Linyi Jiaozhuang electric vehicle charging/battery swapping station for example, APF compensation capacity has been calculated and analyzed.(2) Beginning with the analysis of influencing factors, such as battery characteristics, charge modes, driver behavior, the probability density functions of total daily trip length and charging time are deduced. The charging model is established and the optimal charging method under minimizing load variance is proposed, based on2009NHTS data and measurement results from Jiaozhuang charging station. The relationship between feeder losses and load variance is deduced, and results are presented to verify that electric vehicle charging under minimizing load variance on the grid can effectively reduce net loss. Take the10kV dual radial line of a city in Shandong for example, the impact of EVs charging on the load, voltage quality and network loss, is analyzed, under different scenarios. Results are presented to verify that electric vehicle charging under minimizing load variance on the grid can effectively reduce net loss, improve power quality and so on.(3) In this paper, a smart charging control strategy is presented, combining centralized and distributed control strategy by an aggregator. According to the negative impact on planning and operation of the power system, as is brought by the popularity of EVs, especially when implemented without control, time-of-use scheduling for EV charging has been put forward, by combining time-delay charging control and valley charging scheduling, especially considering the driver’s behavior. To some extent, the load curve can be improved, but the shortcoming of this method lies in lack of flexibility. On the premise of users’interest, intelligent charging control strategy is proposed and compared with time-of-use control strategy in this paper. The optimal scheduling scheme takes the network load variance and charging cost minimization as the optimization objective. An aggregator framework model is proposed as the means to organize charging, which combining centralized and distributed control strategy by an aggregator. The centralized optimal scheduling scheme takes the network load variance minimization as the optimization objective. The optimization goal of distributed scheduling scheme is to minimize the deviation between charging load demand and the centralized optimization results, within each dynamic sliding window, considering the actual battery constraints and EV owners’ benefits. Therefore, the dynamic control for EV charging and discharging is fulfilled. Then taking Shandong load in2015,2020,2030for example, intelligent charging control strategy, which is compared with time-of-use scheduling, is proved to be effective in reducing net loss, improving power performance.(4) An improved support vector machines prediction method is proposed and compared with conventional prediction method in this paper. The EV holdings of Shandong in2015,2020,2030are predicted by support vector machine method, providing as a quantitative basis of cost-effective analysis under different energy structure of power system. The impact and benefit brought by EV is analyzed on the premise that EV quantity is predicted accurately. Most studies at home and abroad are based on assumption or some rough prediction methods, such as the elastic coefficient method or thousand holdings method, so a more scientific and accurate prediction method is necessary. The improved support vector machines prediction method with1.68%accuracy, through the recent three years comparing data.(5) Fully considering the marginal electricity resulted from generation industry and power structure and other factors, carbon emission model for EV is established. Quantitative analysis of the environmental benefits are conducted and compared between conventional vehicle and EV considering different power structures and various levels of EV penetration. The research whether the EVs is helpful in reducing carbon emission is hot recently. In this paper, the effect of coordinated charging is analyzed from two aspects, which are short-term and long-term marginal operational cost of power supply. Based on energy structure development trends of Shandong, the energy structure is divided into three different scenarios:the traditional energy structure, the transition energy structure and new energy structure. Based on benefits analysis model, quantitative analysis of the enviroment benefits between conventional vehicle and EVs, are compared, under different charging scenarios and charging control strategies, in2015,2020,2030. The results are proved to be effective in reducing carbon emission, with energy structure transition and large-scale intergration of EVs.更多还原