【作者】 王峻峰；

【导师】 周家启；

【作者基本信息】 重庆大学 ， 电气工程， 2005， 硕士

【摘要】 本文受国家自然科学基金项目“配电网最优切换免疫模型和算法研究”(5030 7015)等相关课题资助。配电网是电力系统中直接针对用户的环节,对用户供电质量(包括可靠性)有直接的影响。据统计,约80%的停电事故缘于配电系统故障,因此,配电网可靠性评估及结构优化已成为研究的热点。本文主要研究内容如下:(1)实现了中压配电网模糊可靠性评估。可靠性评估中,线路和变压器的故障率及故障排除时间等基础参数对可靠性评估结果有较大影响,而这些基础参数具有明显的不确定性,适合用模糊数来表达。本文分析了论域在模糊数运算中的变化及求解方法,提出不同论域间模糊数的四则运算算法,进而实现了基于模糊参数的中压配电网可靠性评估,其评估结果具有更大的可信度。(2)基于故障扩散法,提出中压配电网可靠性评估的迭代算法。故障扩散算法中存在大量冗余搜索,降低了计算速度,本文进行以下改进:①提出形成开关分层表的算法。故障枚举前,预先形成关于断路器和隔离开关的开关分层表,当前开关所关联的后向负荷等信息可由其下层开关的相关信息参与形成,从而节省了后向搜索时间;②通过判断故障处及其前向线路的开关安装情况,形成不同的故障模式,并给出不同故障模式下负荷分类的迭代算法,不仅避免了前向搜索开关,而且能充分利用已知的故障后果信息和开关分层表信息,提高了计算速度。从理论上说明了,迭代算法的时间复杂度为一阶,故障扩散法的时间复杂度为二阶。(3)提出配电网开关优化配置的改进遗传算法。为提高遗传算法的计算速度和收敛效率,采取以下改进措施:①给出初始种群的生成模型,提高初始种群的适应值;②对染色体进行分段遗传操作,以保证各类开关设备能同步优化;③提出模糊变异的实现方法,以利于优质基因遗传并避免局部最优化;④提出最优染色体的变异机制,以提高算法局部搜索能力。更多还原

【Abstract】 This dissertation is sponsored by the project (“Study on Optimal Switching Model and Algorithm for Electrical Distribution Networks (EDNs) Using Immune-Algorithm”, No.50307015) of the national natural science fund of China.EDNs are connected directly with customers and have great influence on the quality of power supply(including reliability).It was counted that 80% of the total customer interruptions around resulted from EDNs.So It has become a focus to do research on reliability evaluation of EDNS and its configuration optimization.This paper is mainly as follows:(1) Realizing fuzzy reliability evaluation of medium voltage EDNS. In reliability evaluation the component’s reliability parameters such as fault rates and fault repairing time have a big effect on the result of reliability evaluation. However these parameters are typical of imprecision, which makes it reasonable that they should be formulated by fuzzy numbers.This essay analyzes the range’s changes and solutions in fuzzy operation and concludes the calculation algorithm of fuzzy numbers in different ranges so that reliability evaluation based on fuzzy parameters is practical. Thus the result of reliability evaluation will be more credible.(2) Based on the fault-spreading algorithm,the iteration algorithm of reliability evaluation for medium-voltage EDNS is presented. There exists much redundant search in the fault-spreading algorithm,which lowers the calculation speed.In terms of this,some improved measures are taken as follows:①the algorithm for the formation of switch-layer tablet is presented.Switch-layer tablet is established before fault enumeration.In the tablet,the backward information,loads for example,of the present switch can be partially got from the switches in the backward layer,which spare much time for backward search;②according to the switch-fixing information of the fault point and its forward point, various fault modes and loads-classification algorithm based on them are presented,which can avoid forward search and use the information of the known fault sequences and that in the switch-layer.It is explained theoretically that the time complexity degree of iteration algorithm is of 1 rank and that of the fault-spreading algorithm is of 2 ranks.(3) Improved genetic algorithm of optimal switching devices in EDNS is更多还原