智能变电站状态监测系统的设计研究及应用

发布时间：2018-01-08 06:27

本文关键词：智能变电站状态监测系统的设计研究及应用　出处：《华东理工大学》2017年硕士论文　论文类型：学位论文

【摘要】：智能变电站是实现电能集中分配与电压电流变换的场所,在输配电系统中具有极其重要的地位。因此当变电站发生故障时,需要相关运行人员快速准确地定位故障区域,识别真正的故障元件并将其隔离,从而快速恢复非故障区域的正常运行。但是,由于系统发生复杂的多重故障、开关或保护在上传信息时存在信道干扰造成的信息丢失等因素的影响,其后台报警信息往往存在不确定性和不完整性的问题,给状态监测和故障诊断工作造成很多困难。本文在综合分析数据挖掘算法的基本原理的基础上,结合智能变电站的数据类型、网络结构和过程层的配置,并根据数据采集的特征和智能变电站状态监测信息的特点以及故障诊断信息,设计了数据挖掘算法在智能变电站中系统架构,仿真了的数据挖掘算法在智能变电站中的应用的可行性。并以实际的110kV变电站为例,仿真了数据挖掘方法对智能变电站的整体与局部采取不同状态监测方法的影响。利用基于遗传算法的粗糙集方法对110kV区域为整体进行故障诊断仿真实验。结果表明该方法确是一种快速准确、容错性强、适应性好的智能变电站状态监测策略方法,对实现高效的在线智能变电站状态监测具有重要的意义。以变压器为局部对象研究数据挖掘技术在变压器状态监测中的应用。实验表明基于粒子群的多核支持向量机算法可以充分保证计算速度和较高故障判断精度,并且该模型能确定变压器故障种类,并具有较高的正确率。
[Abstract]:Intelligent substation is the place to realize the centralized distribution of electric energy and the conversion of voltage and current, which plays an extremely important role in the transmission and distribution system. It is necessary for the relevant operators to locate the fault area quickly and accurately, identify and isolate the real fault elements, so as to quickly restore the normal operation of the non-fault area. However, complex multiple faults occur in the system. The information loss caused by channel interference exists in the switching or protection of uploading information, and the background alarm information often has the problem of uncertainty and incompleteness. On the basis of comprehensive analysis of the basic principles of data mining algorithm, combined with the data types of intelligent substation, network structure and process layer configuration. According to the characteristics of data acquisition, the characteristics of intelligent substation state monitoring information and fault diagnosis information, the system architecture of data mining algorithm in intelligent substation is designed. The feasibility of the application of the simulated data mining algorithm in the intelligent substation is presented, and the actual 110 kV substation is taken as an example. The influence of data mining method on the whole and local state monitoring methods of intelligent substation is simulated. The fault diagnosis simulation experiment of 110kV region is carried out using rough set method based on genetic algorithm. The results show that the method is fast and accurate. An intelligent substation condition monitoring strategy with strong fault tolerance and good adaptability. The application of data mining technology in transformer condition monitoring based on PSO is studied. The experiment shows that the multi-core support direction based on particle swarm optimization (PSO) is very important to the realization of on-line intelligent substation state monitoring. The calculation speed and the accuracy of fault determination can be fully guaranteed by the metering algorithm. And the model can determine the type of transformer fault, and has a high accuracy.
【学位授予单位】：华东理工大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TM76;TM63;TP274

【相似文献】