基于支持向量机的变电站过热故障的在线监测  

On-line Monitoring of Substation Overheat Fault Based on Support Vector Machine

作　　者：余彬 胡洛娜[2] 王吉哲[2] 

机构地区：[1]国网浙江杭州市萧山区供电公司,杭州311200 [2]国网重庆市电力公司信息通信分公司,重庆400014

出　　处：《浙江电力》2016年第1期31-34,共4页Zhejiang Electric Power

摘　　要：作为电力传输的重要组成部分,变电站的安全运行对电网具有十分重要的意义。根据变电站设备过热故障的特点,基于支持向量分类器的数学理论模型,建立了设备过热故障的判断机制。该方法利用核函数实现向高维空间的非线性映射,并通过最优超平面对特征空间进行划分,进而完成对数据样本的分类。以某变电站为例,验证了支持向量分类器的强大功能,证明其能够出色的完成故障诊断,完全满足变电站设备过热故障的在线监测要求。As an important part of power transmission, the safe operation of substation is key to power grid. In this paper, according to the characteristics of the substation equipment overheat fault and mathematical theory model of support vector classifier, equipment overheat fault diagnosis mechanism is established. The method uses kernel function to realize nonlinear mapping to high dimensional space and completes feature space divi- sion through optimal hyperplane to classify data sampling. The powerful functions of support vector classifier is proved by its application in a substation that it can well diagnose faults and meet the online monitoring re- quirements of substation equipment overheat.

关 键 词：变电站 过热故障 在线监测 支持向量分类器 

分 类 号：TM63[电气工程—电力系统及自动化] TP18[自动化与计算机技术—控制理论与控制工程]