规模化电动汽车与电网互动策略研究

发布时间：2018-03-02 18:41

本文选题：电动汽车　切入点：互动策略　出处：《西南交通大学》2017年硕士论文　论文类型：学位论文

【摘要】：电动汽车(Electric Vehicle,EV)具有节能环保的优点,是未来的主要交通工具。电动汽车充放电行为具有波动性,在不受控制的情况下,电动汽车车主倾向于在出行结束后立即将电动汽车接入电网充电,其大规模的无序充电行为会对居民小区的负荷特性、配电网的电压质量以及区域电网的稳定运行造成不同程度的影响。通过电动汽车与电网互动技术(Vehicle to Grid,V2G),使电动汽车有序充放电,在电网侧改善电网负荷特性、电压质量以及频率质量,在用户侧可以降低车主的充电成本,对其展开研究意义重大。论文主要包括以下内容:1)提出了考虑居民小区负荷特性的电动汽车充放电调度策略。综合考虑电动汽车的行驶特性、电池特性等约束条件,建立了计及峰谷差、日负荷率、负荷均方差和用户电费的多目标的电动汽车与居民小区互动的调度策略。采用双层离散粒子群算法优化电动汽车充放电调度模型进行求解,解决智能优化算法难以求解含等式约束方程的问题。第1层优化通过离散粒子群算法求解满足所有约束条件的单辆电动汽车充放电计划可行解,第2层优化采用基本粒子群算法迭代优化电动汽车充放电调度模型。对无序充电、有序充电调度、有序充放电调度模式以及不同用户响应度的电动汽车充放电调度策略进行仿真分析。2)提出了考虑配电网电压越限概率的电动汽车代理商定价策略。首先,通过拉丁超立方采样技术构建电动汽车充电场景,根据主从博弈理论确定各充电场景的充电起始时间,建立电动汽车动态概率负荷模型。其次,应用半不变量动态概率潮流求解各节点的电压幅值概率分布,以此评估电压合格情况,并以代理商期望收益最大为目标,以电价波动限值和电压合格率期望值为约束,建立代理商定价策略。再次,采用双层粒子群算法求解电动汽车代理商定价策略,在底层优化中处理电价约束条件,顶层优化中处理系统约束条件。最后,在IEEE 33节点配电系统中进行仿真分析。3)提出了考虑区域电网有功偏差特性的电动汽车调频控制策略。该控制策略以V2G控制和负荷频率控制的联合优化为研究对象,考虑了区域电网有功偏差分布对非线性频率控制的影响,在两区域互联电网中,建立负荷频率控制模型和V2G控制模型,详细分析了电动汽车的分层调度机制以及有功偏差的电气量和统计特性,通过拉丁超立方采样技术构造了有功偏差波动场景,以各有功偏差波动场景下的时间积分绝对误差之和最小为目标函数,通过带压缩因子的粒子群优化算法对联合优化控制器进行求解。为了验证算法的有效性,分别在单一有功偏差波动场景和长时间序列的有功偏差波动场景中,与无电动汽车参与调频和不进行控制参数优化模式的调频效果进行比较。
[Abstract]:Electric vehicle (EV) has the advantages of energy saving and environmental protection, and is the main vehicle of the future. The charging and discharging behavior of electric vehicle is fluctuating, and it is not under control. Electric vehicle owners tend to charge electric vehicles into the grid immediately after the trip is over, and their large-scale disordered charging behavior will affect the load characteristics of residential areas. The voltage quality of the distribution network and the stable operation of the regional power network have different effects. By means of the interactive technology between electric vehicle and the power network, the electric vehicle can charge and discharge orderly and improve the load characteristics of the power network at the power network side, so that the electric vehicle can charge and discharge in an orderly manner. Voltage quality and frequency quality can reduce the charge cost of the owner on the user side. The thesis mainly includes the following contents: 1) this paper proposes a charging and discharging scheduling strategy for electric vehicles considering the load characteristics of residential areas, considering the driving characteristics of electric vehicles, battery characteristics and other constraints. The scheduling strategy of multi-objective interaction between electric vehicle and residential area considering peak and valley difference, daily load rate, load mean-square deviation and user electricity rate is established. A double-layer discrete particle swarm optimization algorithm is used to optimize the charge and discharge scheduling model of electric vehicle. In order to solve the problem that intelligent optimization algorithm is difficult to solve the problem with equality constraint equation, the first layer optimization solves the feasible solution of charging and discharging plan of single electric vehicle satisfying all constraints by discrete particle swarm optimization (DPSO). The second layer optimizes the charge and discharge scheduling model of electric vehicle by using the basic particle swarm optimization algorithm. Simulation analysis of charge and discharge scheduling strategy of electric vehicle with different user response is carried out. 2) A pricing strategy of electric vehicle agent considering the probability of voltage overrunning in distribution network is proposed. First of all, an agent pricing strategy for electric vehicle is proposed. The charging scene of electric vehicle is constructed by Latin hypercube sampling technology. According to the game theory of master and slave, the charging start time of each charging scenario is determined, and the dynamic probabilistic load model of electric vehicle is established. Secondly, the dynamic probability load model of electric vehicle is established. Using the semi-invariant dynamic probabilistic power flow to solve the probability distribution of voltage amplitude of each node, the voltage qualification is evaluated, and the maximum expected income of the agent is taken as the target, and the limit value of the fluctuation of electricity price and the expected value of the qualified rate of voltage are taken as the constraints. Thirdly, the double-layer particle swarm optimization algorithm is used to solve the agent pricing strategy of electric vehicles, and the power pricing constraints are dealt with in the bottom optimization and the system constraints in the top-level optimization. Finally, Simulation analysis in IEEE 33-bus distribution system is carried out. (3) an FM control strategy for electric vehicles considering the characteristics of active power deviation in regional power network is proposed. The control strategy takes the joint optimization of V2G control and load frequency control as the research object. Considering the influence of active power deviation distribution on nonlinear frequency control, a load frequency control model and a V2G control model are established in two interconnected power networks. The hierarchical scheduling mechanism of electric vehicle and the electrical and statistical characteristics of active power deviation are analyzed in detail. The fluctuating scene of active power deviation is constructed by Latin hypercube sampling technology. Taking the minimum sum of absolute errors of time integral in the fluctuating scene of active power deviation as the objective function, the joint optimization controller is solved by particle swarm optimization algorithm with compression factor. In order to verify the effectiveness of the algorithm, In the single active power deviation fluctuation scene and the active power deviation fluctuation scene in a long time series, the frequency modulation effect of the non-electric vehicle involved in FM and the mode of no control parameter optimization were compared.
【学位授予单位】：西南交通大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TM73

【相似文献】