【作者】 张云；

【导师】 王艳君；

【作者基本信息】 河北农业大学 ， 农业电气化与自动化， 2008， 硕士

【摘要】 随着发电技术的进步和日益增长的负荷需求,发电容量和电力系统规模不断扩大,大电网难以灵活跟踪负荷变化、系统可靠性降低、输配电网络损耗较大以及严重的环境污染等问题逐渐暴露出来,已经不能满足当今社会对电力供应质量与安全可靠性的要求。分布式发电(Distributed Generation,DG)具有减轻环境污染、降低终端用户费用、改善电能质量、提高供电可靠性、灵活地跟踪负荷变化、能满足能源可持续发展等众多优点,因此分布式发电与大电网供电相互补充、协调,是综合利用现有资源和设备、为用户提供可靠和优质电能的最佳方式。分布式发电除了在偏远或特殊的地区作为唯一的供电电源外,大部分用户希望既能使用分布式发电供电又可以由当地电网供电,或由它们同时供电。因此配电网络必须考虑与分布式发电的配合,当大量的分布式发电出现在规划方案中时,大量的随机变化使得系统的复杂性大大地增加。传统的规划方法没有足够的能力解决含分布式发电的配电网规划问题,这主要是因为传统的规划方法都不同程度地将规划问题进行了简化,对于规划中客观存在的难以定量表达的不确定性因素缺乏较好的处理方法。本文对含分布式发电的配电网规划问题进行了较深入的分析和探讨。首先,介绍了分布式发电的概念、分类和国内外发展现状,并列举了几种分布式发电技术及其优越性;分析了分布式发电接入配电网后对其运行和规划方面的影响。其次,针对配电系统中计及分布式发电的单一规划问题,在考虑分布式发电的经济性和安全性的基础上,建立了以分布式发电投资成本最小、系统网损最小和静态电压稳定裕度最大为优化子目标的多目标规划模型。为了获得全局最优方案,应用模糊理论制定了总体满意度,将多目标优化转化为单一目标的总体优化。最后,采用遗传算法求解模型,并针对遗传算法在求解过程中暴露的缺陷,如收敛速度较慢,易早熟等现象,提出了改进自适应遗传算法。采用算例验证了规划模型制定的正确性,还证明了改进自适应遗传算法比遗传算法收敛更快,全局优化能力更强。更多还原

【Abstract】 With the development of the power technology, the daily increasing load demand and the expansion of the generation size, problems of the big electric network gradually emerge, such as the hard tracking of power system to the load’s change, low system reliability, large line loss of transmission system and distribution system, serious environment pollution, so it has been unable to meet the quality and security requirements of electric power supply. Distributed generation (DG) has many advantages, such as reducing the pollution to environment and lowing charge to users, improving quality of the electric energy and security of the power supply, easy tracking of power system to the load’s change, be able to meet the requirements of sustainable energy development, so supplement and corresponding with the big electric network, distributed generation is the best way to supply the reliable and high-quality electric power to users by making good use of existed resource and equipment.Most users would like to use electric power supplied by both DG and the big electric network, except some remote or special areas which must use DG only. So we must consider that how DG to operate with the big electric network, when more and more DG have appeared in distribution network planning project and the systemic complexity has been increased consumedly by introducing a mass of random changes. Because the traditional planning methods have simplified the planning problem more or less and lack good treatment for the uncertainties which is existent impersonality but hard to express, it is unable to solve the problem of distribution network planning including DG.The paper makes some analysis and research about the distribution network planning including DG. Firstly, the paper introduces conception and class of DG and its development situation at home and abroad, also introduces several technologies of DG and its advantages. Secondly, for a single planning of distribution network including DG, based on considering economy and security of distributed generation, constructing a multi-objective optimal model, which including the minimum investment cost of distributed generation and the minimum power loss of distribution network and the maximum static voltage stability margin. The paper introduces total satisfied degree by employing the fuzzy optimal theory, which makes a good way to transform multi-objective into single objective, in order to get the best project in all. Lastly, the genetic algorithm has some disadvantages when resolving the planning model such as slow convergence and easy to precocity. In order to overcome the shortcomings of genetic algorithm, a new improved self-adaptive genetic algorithm was proposed in this paper. The paper use a system simulation results to verify the proposed model correction, prove the improved self-adaptive genetic algorithm is more effectively in seeking the best result of overall situation and increasing the convergence speed than genetic algorithm.更多还原