【作者】 欧阳武；

【导师】 张秀彬； 程浩忠；

【作者基本信息】 上海交通大学 ， 电气工程， 2009， 博士

【摘要】 将分布式发电(DG)与集中供电相结合,清洁、高效利用能源,是电力工业重要的技术增长点。在未来智能电网中,改进的互联标准将使不同容量、不同类型的DG更加便利地接入配电网。分布式发电技术的应用必将给传统配电网规划带来深远变革。本文研究了含DG的配电网规划,主要成果包括:(1)结合典型负荷日特性曲线,在变电站规划中考虑了DG削峰作用和不同类型负荷间的错峰效应,避免了传统方法可能出现的主变容量利用率不高和负荷分配不均衡的弊端;基于Voronoi图的遗传算法实现了变电站布点和供电范围的合理划分,为含DG的变电站规划提供了一种有效方法。(2)从不同角度,构建了含DG的配电网规划模型。考虑安全约束的规划模型充分计及DG接入对配电网安全的影响,增强了配电网安全性能;考虑DG出力随机性的规划模型采用随机潮流和机会约束技术,为含多种可再生能源类型DG的配电网规划提供了新方法;考虑DG准入功率的多目标规划模型均衡协调了配电网投资成本与DG准入功率之间的关系,提高了配电网接纳DG的能力;DG与配电网联合规划模型以社会成本最小化为目标,并转换成一个双层规划问题,为在DSM(需求侧管理)和主动管理框架下优化配置DG资源探索了新途径。(3)根据配电网拓扑结构特点,将常规配电网规划遗传算法改进为基于支撑树的遗传算法(St-GA),确保进化过程中每个规划方案均拓扑可行,提高了GA搜索效率;提出将多阶段配电网规划转化为搜索最优目标网架的策略,达到各阶段规划信息在目标网架中的融合,为研究多阶段配电网规划提供了新思路。(4)引入配电网节点边际容量成本(LMCC),反映节点容量的充裕程度。基于节点边际容量成本的配电网扩容规划以DG单位扩容成本小于接入点的边际容量成本作为DG接入配电网的条件,既发挥了DG扩容作用,又避免了原有供电设备利用率降低,为评价DG接入的经济性提供了新工具。更多还原

【Abstract】 Distributed generation (DG) is predicted to play an increasing role in the electric power system of the future. A large amount of DG capacity conneted to distribution network will bring about the great influence on operation and planning of conventional distribution network. Therefore, the study on distribution network planning with DG is one of the hotspots in the power engineering research field. This dissertion concentrates on some critical problems in this field and the main innocations are shown as follows:(1) Combined with the typical load characteristic curve, a new method for substation planning with the consideration of DG for peak cutting is proposed to overcome the low utilization and unbalanced loading of main transformers in the conventional methods. Moreover, genetic algorithm based on Voronoi diagram can determine the optimal location, sizing and the service areas and provide an effective design method of substation planning.(2) Planning models of the novel distributution network are established from the different viewpoints. The planning model considering safety constraints takes the risks as the result of DG connection into account and increases the safety capability of distribution netwrok; Based on chance-constrained programming, the planning model considering stochastic character of renewable DG output is presented to evaluate the distribution network investment risk due to DG connected to distribution network; The multi-objective planning model is built to explore the tradeoff between the maximum penetration level of DG and distribution network investment. The iterative algorithm based on linear programming is adopted to maximize the DG capacity in the given network, while the improvedε-constraints approach is used in the optimization decision on the pareto optimal set; The planning model of distribution network integrated with DG, which aims to minimize the social cost, is transformed into a bi-level planning and provide a noval measure to allocate DG source in the mode of demand side management (DSM) and active management.(3) Spanning tree-based genetic algorithm (st-GA) to solve distribution network planning with DG is presented. St-GA with two specialized genetic operators can guarantee the topological feasibility and high heritability of each solution generated and overcome the deficiencies of existing GAs applied for distribution network planning. According to the topological characteristic of distribution network, the multistage distribution network planning can be transformed into a problem of searching the optimal target network with the minimum total costs in the whole planning horizon, which contains the message of each building stage for every candidate branch and its size.(4) Locational marginal capacity cost (LMCC) is introduced to reflect the forward-looking cost of nodal capacity and the degree to which nodal capacity is utilized. On this basis, DG connection criterion that the unit capacity cost of DG is less than LMCC at its candidate node is obtained to avoid the lower utilization of the existing power supply facilities, which provides a new tool to envaluate DG connection from the point of the view of enlarging capacity with DG.Analysis and comprarision of the study cases proves the feasibility, effectivity and practicability of this dissertion.更多还原