【作者】 李高望；

【导师】 段献忠； 石东源；

【作者基本信息】 华中科技大学 ， 电力系统及其自动化， 2013， 博士

【摘要】 能源短缺、环境污染等问题的日益加剧，促使动力电气化成为未来交通行业的主要发展趋势。目前，电动汽车的研究与推广已经成为各国政府、汽车制造厂商所关注的焦点。电动汽车的应用在缓解能源、环境等问题的同时，将对电力系统的运行与管理产生深远的影响，其影响涉及到配电、输电、发电以及调度通信系统等多个层面。鉴于此，本文以含电动汽车的电力系统运行问题作为主要研究内容，所做工作和所取得的成果归纳如下：提出利用配电网络重构技术来缓解电动汽车所带来的不利影响，以提高配电网接纳电动汽车的能力。配网重构模型涉及降低有功损耗、改善电压质量和平衡变压器负载等多个优化目标，采用模糊集理论处理多目标之间的权衡问题，并使用量子粒子群优化算法进行求解。此外，构建了以负荷峰谷差最小为目标的电动汽车协调充电模型，对比分析了不控充电模式和协调充电模式下配网负荷曲线的特性。算例结果表明，采用网络重构技术可有效地缓解电动汽车所带来的不利影响，提高配电网接纳电动汽车的能力。提出考虑不确定性因素的配电网鲁棒性综合优化模型，该模型将不确定性环境中网损分布标准差的越限惩罚与网损期望值之和作为目标函数。考虑的不确定性因素包括电动汽车的出行时间、负荷功率以及配网中小型风电场的有功出力，优化手段包括网络重构、投切电容器以及调节有载调压变压器档位。算例分析表明，所提出的配电网鲁棒性综合优化模型可有效减少不可行解的产生，并能保障优化方案在不确定性环境中的鲁棒性。分析了含电动汽车的输电网潮流的概率分布特性。以全美家用车辆调查的数据为蓝本，考察了首次出行时刻、返家时刻和日行驶里程3方面数据之间的相关性，并采用Copula函数对其相关性进行描述。采用蒙特卡洛法计算电动汽车充电负荷的分布，在此基础上，对含有电动汽车的输电网概率潮流进行计算与分析。算例结果表明，Copula函数可有效地拟合车辆行驶数据之间的相关性。协调充电策略将电动汽车的充电负荷安排在晚间负荷低谷时段，但随着电动汽车数量的增多，该时段的充电负荷激增，部分节点的电压分布以及部分线路的功率分布将产生显著的变化。分析了电动汽车的充、放电行为对机组组合优化的影响。在协调充电模型的基础上，对各时段电动汽车的反向放电能力进行估算。算例分析表明，电动汽车的大规模接入将对电力系统的机组组合优化产生影响，其影响主要取决于电动汽车的充电负荷特性和电动汽车的反向放电特性两方面。对电动汽车采取不同的调控方案，将使机组组合的优化结果具有明显的差异性。采用以负荷峰谷差最小化为目标的协调充电方式，将使负荷曲线更加平整，机组的出力调整更为灵活，促使发电成本减少。利用电动汽车的反向放电能力为电网提供旋转备用，可减少传统机组的备用量，从而进一步降低发电成本。建立了电力调度数据网的数据传输模型，定义了数据包拥塞率并将其作为评估信息拥塞程度的指标。模型计及了调度数据网中信息流向的特点，考虑了垂直信息流和随机信息流2种典型的信息流模式。分析了具有星型结构和网状结构的调度数据网的传输特性，并通过模拟核心节点和关键线路遭受攻击的场景，进一步对比了2种结构的调度数据网在遭受攻击时的传输特性变化。仿真结果表明：电力调度数据网的传输特性与拓扑结构有着紧密的关系。正常情况下，星型网络比网状网络具有更优的传输性能，其缓解信息拥塞的能力更强；遭受攻击后，星型网络和网状网络的传输特性呈现不同程度的恶化，其中星型网络恶化得更为严重，表现出较强的脆弱性。更多还原

【Abstract】 The increasing energy shortage and environmental pollution make the electrificationof transportation become major trends. In recently, the research and promotion of electricvehicles has become the focus of governments and automobile manufacturers. Theapplication of electric vehicle will alleviate the problem of energy shortage andenvironmental pollution, and it also significantly affects the operation and management ofthe power grid, which involves many levels such as distribution system, transmissionsystem, generation system and Electric power dispatching data network. In view of this, theoperation of power systems integrating electric vehicles is chosen to be the research topic,and the research work and outputs of the thesis are listed as follows:The strategy of improving the distribution system’s ability to accept electric vehiclebased on network reconfiguration technology is proposed, which is used to mitigate theimpact of electric vehicle on the power system. Multi-objective distribution networkreconfiguration model is proposed, and its objective involves reducing power loss,improving the voltage quality and load balance of the transformers. Fuzzy sets are used tohandle the multi-objective, and the quantum-inspired binary particle swarm algorithm isused to solve the optimal problem. In addition, coordinated charging model aiming tominimize the peak-valley load is proposed, and the load profile of distribution with electricvehicles’ charging load is analyzed. The simulation results show that it is effective toalleviate the adverse effect from electric vehicle by using distribution networkreconfiguration technique, which will be of great helpful for improving the ability ofdistribution network to access electric vehicle.Robust comprehensive optimization for distribution system considering theuncertainties is proposed, which uses the sum of standard deviation punishment and meanvalue of power loss in uncertain environment as objective function. The uncertaintiescontain the travel time of electric vehicle, load power and active power output of the small-sized wind farms in distribution network, and the control means include networkreconfiguration, capacitor switching, regulating positions of on-load transformer changer.The simulation results show that the robust optimization will effectively reduce thegeneration of infeasible solutions, and maintain the robustness of solutions in the uncertainenvironment.The probabilistic power flow of transmission grid with electric vehicles is analyzed.Based on the national household travel survey, the correlation between the first travel time,the returning time and day mileage is investigated, which is described by the copulafunction. Monte Carlo method is used to calculate the distribution of the electric vehicles’charging load, and the probabilistic power flow of transmission system with electricvehicles is analyzed. The simulation results show that Copula function can effectively fitthe correlation between the travel datas. Night coordinated charging strategies can beeffectively arrange the charging load of electric vehicles into the time interval with lowload, however, with the increasing of electric vehicles, the charging load will increasesignificantly, and the voltage distribution of part of nodes and power distribution of partlines will have significant change.The impact of charging and discharging behavior of electric vehicles on the unitcommitment is analyzed. Based on the coordinated charging strategy, the inverse dischargecapacities of electric vehicles are estimated. Simulation results show that applying differentcharging and discharging schemes to electric vehicle will obviously impact theoptimization results of unit commitment, which depends on the characteristics of chargingand discharging of electric vehicle. Applying different charging and discharging schemeswill obviously impact the optimization results of unit commitment. The proposedcoordinated charging strategy makes the load profile become flat, which generates lessgeneration cost. Utilizing inverse discharging capacity of electric vehicle as spinningreserves will reduce the reserve requirement from the traditional units and further lower thegeneration cost.The data transmission model was built, and the packet congestion rate was used asinformation congestion indicator in this paper. Meanwhile, the model took the direction of information flow into account and defined two typical information flow modes, which werevertical information flow and random information flow. The transmission characteristics ofelectric power dispatching data network (EPSDDN) with star and mesh structure wereinvestigated. Furthermore, the transmission characteristics of two EPSDDNs under attackswere investigated with removal of the core nodes and the key links. Simulation resultshows that, there is a close relationship between the transmission characteristic ofEPSDDN and network structure. In the normal case, the star network has betterperformance than the mesh network because of its outstanding capability to relieve theinformation congestion; while under attacks, the transmission performance of twoEPSDDNs show varying degrees of deterioration, and the star network exposes muchhigher vulnerability than the mesh network does.更多还原