基于特征提取和分类算法的微电网孤岛检测模型

发布时间：2018-07-01 20:04

本文选题：微电网 + 孤岛检测　；参考：《华北电力大学》2014年硕士论文

【摘要】：近年来,微电网成为电力系统研究的热点。目前,分布式电源(Distributed Generation, DG)大规模并网过程中仍存在很多问题,DG的发电能力不能充分发挥,微电网灵活、智能的控制方式为这一问题提供了新的解决思路,可以提高供电可靠性；微电网中DG多采用可再生能源,清洁无污染,符合保护环境的理念；微电网还可作为超高压、远距离电力网络的补充供电方式。但是,微电网发挥上述优势有赖于孤岛检测问题的有效解决,微电网孤岛检测问题的研究具有重要的理论和现实意义。目前学者们提出的孤岛检测方法主要针对单个逆变器,主要包括基于通信手段的孤岛检测方法、主动式孤岛检测法和被动式孤岛检测法。基于通信的检测方法准确可靠,主要限制在于成本高,目前实用困难,随着智能电网的建设,特别是智能配电网的发展,有一定潜力；主动法检测盲区小、灵敏性较高,但是会引起电能质量下降,且不适用于多个逆变器；被动法原理简单、实现容易,对电能质量无不良影响,但是无法避免较大的检测盲区。本文介绍了微电网的概念及世界各国的发展现状,研究了微电网并网运行和孤岛运行这两种运行模式,分析了计划孤岛运行的优点和非计划孤岛运行带来的危害；指出了目前对孤岛效应的两种原则立场：反孤岛策略和孤岛运行的利用。研究了孤岛检测的原理及特性,提出了孤岛检测的主要标准和指标,分析了现有的孤岛检测方法和其优缺点,指出了孤岛不易检测出的主要原因,提出了新的基于特征提取和分类算法的孤岛检测方法。研究了ELM的基本理论,建立了基于特征提取和分类算法的微电网孤岛检测模型,实现了微电网孤岛检测模型的离线参数寻优和在线检测,利用PSCAD仿真软件搭建了微电网模型,利用Matlab仿真软件搭建了孤岛检测模型,通过仿真计算验证了孤岛检测方法的有效性。
[Abstract]:In recent years, microgrid has become a hot spot in power system research. At present, there are still many problems in the process of large-scale grid connection of distributed Generation (DG). The generation capacity of DG can not be brought into full play. The microgrid is flexible, and the intelligent control mode provides a new way to solve this problem, which can improve the reliability of power supply. In microgrid, DG mostly adopts renewable energy, clean and pollution-free, which is in line with the idea of protecting the environment. Microgrid can also be used as a supplementary power supply mode for ultra-high voltage and long-distance power network. However, the advantage of microgrid depends on the effective solution of islanding detection problem, and the research of islanding detection problem in micro-grid has important theoretical and practical significance. At present, the islanding detection methods proposed by scholars are mainly aimed at single inverter, including islanding detection method based on communication means, active islanding detection method and passive islanding detection method. The detection method based on communication is accurate and reliable, the main limitation lies in the high cost and the practical difficulty at present. With the construction of smart grid, especially the development of smart distribution network, it has a certain potential, and the active method has small blind area and high sensitivity. The passive method is simple in principle, easy to realize and has no adverse effect on the power quality, but it can not avoid a large blind spot. This paper introduces the concept of microgrid and the development status of countries in the world, studies the two operation modes of microgrid grid-connected operation and islanding operation, and analyzes the advantages of planned islanding operation and the harm caused by unplanned islanding operation. This paper points out two kinds of principle positions on islanding effect: anti-islanding strategy and utilization of islanding operation. This paper studies the principle and characteristics of islanding detection, puts forward the main standards and indexes of islanding detection, analyzes the existing detection methods and their advantages and disadvantages, and points out the main reasons for the difficulty of islanding detection. A new islanding detection method based on feature extraction and classification algorithm is proposed. In this paper, the basic theory of ELM is studied, and the detection model of isolated island in microgrid based on feature extraction and classification algorithm is established. The off-line parameter optimization and on-line detection of the detection model of isolated island in micro-grid are realized. The micro-grid model is built by using PSCAD simulation software. The model of islanding detection is built with Matlab simulation software, and the effectiveness of the detection method is verified by simulation calculation.
【学位授予单位】：华北电力大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TM73

【参考文献】