基于粗糙集和证据理论的智能变电站故障诊断

发布时间：2018-06-22 05:34

本文选题：智能变电站 + 故障诊断　；参考：《天津大学》2014年硕士论文

【摘要】：变电站的故障诊断系统对提高电力系统的安全性、经济性有着非常重要的作用。伴随着智能变电站的建设和IEC61850体系标准的飞速发展,变电站内部可以非常方便地进行信息互享和数据交互。针对变电站故障诊断中告警信息的不确定性和多源性,本文提出了一种基于粗糙集和信息融合技术的智能变电站故障诊断方法。论文介绍了智能变电站的现状,阐述了智能变电站的基本结构、主要技术特征以及新技术对智能变电站发展的影响,并且较为详细地研究了智能变电站的通信方式,包括IEC61850标准体系的技术特点、IED配置描述语言以及采样值报文和GOOSE报文的内容和传输特性。与粗糙集方法进行结合,阐述了粗糙集理论的基本概念、信息系统、近似空间及不确定性表述方式,详细地研究了离散化算法和约简算法,着重考虑基于关联度的Naive Scaler离散算法以及基于Apriori算法的约简枚举方法。为了将分析得到的信息结果进行总结和统一,研究了证据理论中的信息融合技术,包括概率分配函数、信任函数和似然函数的应用情况,并阐述了不确定性的表示关系,特别考虑了Dempster合成法则的概念以及应用方法。通过对一个220/110kV的变电站系统进行建模与仿真,利用GOOSE报文和采样值报文得到故障告警信息,将故障区域和故障装置作为决策对象,通过采用粗糙集和证据理论中信息融合的方法,并且结合层次分析法实施故障诊断,在信息传输有误的情况下也有较好的诊断效果,综合研究并验证本文算法的正确性和有效性。
[Abstract]:Substation fault diagnosis system plays an important role in improving the safety and economy of power system. With the construction of intelligent substation and the rapid development of IEC61850 system standard, it is very convenient to exchange information and data within the substation. Aiming at the uncertainty and multi-source of alarm information in substation fault diagnosis, this paper presents a method of intelligent substation fault diagnosis based on rough set and information fusion technology. This paper introduces the present situation of intelligent substation, expounds the basic structure of intelligent substation, the main technical characteristics and the influence of new technology on the development of intelligent substation, and studies the communication mode of intelligent substation in detail. It includes the technical features of IEC61850 standard system and the IED configuration description language as well as the contents and transmission characteristics of sample value message and goose message. Combining with rough set method, the basic concepts of rough set theory, information system, approximate space and uncertainty expression are expounded. The discretization algorithm and reduction algorithm are studied in detail. The naive Scaler discrete algorithm based on correlation degree and the reduction enumeration method based on Apriori algorithm are considered. In order to summarize and unify the information obtained from the analysis, this paper studies the information fusion technology in evidence theory, including the application of probability distribution function, trust function and likelihood function, and expounds the representation of uncertainty. In particular, the concept and application of Dempster's law of composition are considered. Through modeling and simulation of a 220 / 110kV substation system, fault warning information is obtained by using goose message and sample value message. Fault area and fault device are taken as decision objects. By adopting the method of information fusion in rough set and evidence theory, and combining with analytic hierarchy process to implement fault diagnosis, it has better diagnosis effect in the case of information transmission error. The correctness and validity of this algorithm are comprehensively studied and verified.
【学位授予单位】：天津大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TM63;TM76

【相似文献】