【作者】 丰保民；

【导师】 王哈力；

【作者基本信息】 哈尔滨理工大学 ， 电工理论与新技术， 2003， 硕士

【摘要】 随着国家现代化建设的发展，社会各方面承受停电的能力越来越弱，对配电网和供电设备提出了更高的要求。因此开发出可进行在线监测与控制的供电自动化系统，提高配电网的供电质量，已成为提高社会发展水平的重要内容。 本文正是在这样的背景之下，将供配电自动化系统的重要组成部分--箱式变电站作为研究对象，对用神经网络方法进行故障诊断进行了探讨，对箱式变电站的典型故障模式进行了识别研究。 本文进行了对人工神经网络理论的研究，尤其是对人工神经网络在电气设备故障诊断中的模型和算法的研究，使人工神经网络的模型和算法能够应用到箱式变电站故障诊断中去。同时，对神经网络的主要学习算法之一的反向传播算法(BP算法)进行了深入的研究，通过对几种改进算法的模拟实现和对比综合，构造了适合于对箱式变电站典型故障进行识别的学习算法。 本文确定了应用神经网络理论对箱式变电站进行故障诊断的总体设计思想和步骤：确定了监测数据的预处理模糊化方法；建立了箱式变电站典型故障集和典型故障征兆集；确定了学习样本的格式，完成了学习样本的生成；确定了神经网络结构和参数，并对学习样本应用本文的学习算法进行了学习训练，使误差控制在给定范围内；以集散监测诊断系统的思想，提出了由多个神经网络协同构成的多神经网络故障诊断模型，并论述了其诊断原理。 为了完成设计，学习了C++语言和数据库编程的基本方法，并以Visual C++6.0和Access2000为开发工具，进行了箱式变电站故障诊断系统软件的编写。在本文研究的诊断原理的基础上，开发了一套实用的箱式变电站故障诊断系统软件包。 本文对箱式变电站典型故障的识别进行了较为深入的研究，论文工作有新意，为今后的工作打下了良好基础。更多还原

【Abstract】 With the fast development of modernization construction, it is weaker and weaker for social different fileds to bear the ability to have a power failure, having put forward higher request to the distribution network and power-supply unit. So, it has already become the important content of improving the social development level that developing the power-supply automated system which can monitor online and control to improve the power-supply quality of the distribution network.Just under such a background, the important part in distributionautomated system-Intelligent Prefabricated Substation-isregarded as the research object in this paper. The method diagnosing the fault of Electric equipments in the prefabricated substation with Artificial Neural Network(ANN) theory is probed into and the key problem such as the identification of typical fault mode of prefabricated substation is also studied.The fundamental theory of ANN, especially its model and algorithm which can be used better for the fault diagnosis of prefabricated substation is studied.As one of the main learning algorithms .Back Propagation(BP) algorithm is studied emphatically. Some developed BP algorithms which were simulated and compared using computer seperately are also presented, ANN learning algorithm in this paper is constructed on the basis of the goodness of all above algorithms which is suitable for the identification of typical fault mode.In this paper,overall design philosophy and measure while diagonose the prefabricated substation using ANN theory are defined, including the definition of fuzzy expression method for fault symptoms, the definition of typical fault collection and typical fault sign collection,thedefinition of the format of the learning sample and test sample, and the definition of fault diagnosis model formed in coordination by multi ANN whose diagnosis principle are also described.A practical software using Visual C++6. 0 and Access2000 as developing instrument are developed on the basis of diagnosis principle put forward by this paper.The identification of the typical fault mode of prefabricated substation is researched in detail, which make a good fundament for further work.更多还原