【作者】 刘云；

【导师】 陈金富； 段献忠；

【作者基本信息】 华中科技大学 ， 电力系统及其自动化， 2007， 硕士

【摘要】 配电网是电力系统到用户的最后一环,直接地反映着用户在安全、优质、经济等方面的要求,配电网的供电可靠性与供电质量和国民经济的发展及人们的日常生活息息相关,因此建构一个结构合理同时满足经济性、可靠性要求的配电网络具有及其重要的意义。本文对配电网规划问题及求解方法进行了一定的研究,首先总结了电力系统配电网规划问题的研究现状,根据不同时期的发展状况,从规划的模型和算法两个方面分别进行了综述,并针对各个模型和算法的特点进行了总结和比较。建立了以配电网中变电站及馈线的投资费用、年维护折旧费用以及运行费用之和为目标函数,综合考虑了变电站及馈线的容量约束、电压降落约束以及辐射状网架结构约束等条件限制的配电网规划模型。针对配电网网络结构特点,采用了前推回代法进行配电网潮流计算,避免了矩阵运算,计算效率高、处理方法简单。引入了序优化理论进行配电网规划模型的求解,在避免维数灾的前提下,能够确保以足够高的概率求解到足够好解,同时使最终规划结果的优劣程度得以量化,具有较高的工程实用价值。更进一步的,将序优化理论与传统的遗传算法进行结合,利用序优化思想的指导来进行遗传算法的操作,通过算法混合将两者的优良特性进行集成来弥补各自的缺陷,能够更有效的避免遗传算法存在的过早收敛、终止条件难以确定等不足,高效地保证优化质量。更多还原

【Abstract】 Distribution network is the last link to the users in power system, it sensitively reflects the request of the users in the security, quality and economic aspects. The reliability and quality of the power supply of the distribution network are directly related to the development of the national economy and people’s daily life, so that an optimal distribution network which satisfying the economic and reliability requirements has a very important significance. This paper concentrates on the distribution network planning modeling and its solution method. The main contents are as follows:Firstly, the study status of the distribution network planning are summed up, based on different periods of development, both in terms of planning models and algorithms are reviewed, and the features and difference of the models and algorithms are analyzed.A model on distribution network is presented, and the objective of the model is to minimize the total cost including the fixed investment, maintenance costs and feeder resistive loss, while the node voltage and feeder and transformer capacities are kept within the specified limits. In view of structural features of the distribution network, used the distflow method to calculate the power flow to avoid the matrix operation, this method is simple and can get a high efficiency.Ordinal Optimization theory is used to solve the model, it can overcome the dimension problems and can guarantee to get good enough planning schemes with high probability, so that the final outcome of the planning merits can be quantified, and the result has a high practical value.Furthermore, the ordinal optimization theory was combined with the genetic algorithm optimization (GOO) to solve the limitations of each single algorithm. Use the idea of ordinal optimization to guide the evolutionary operators of genetic algorithm. This method can overcome the troubles of the GA such as premature convergence、the hardness to determine stopping criterion, and it guarantees to get the global optimal solution with a high probability.更多还原