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
余彬,胡洛娜,王吉哲
YU Bin,HU Luona,WANG Jizhe
摘要(Abstract)：
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 division 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 requirements of substation equipment overheat.
关键词(KeyWords)：
变电站;过热故障;在线监测;支持向量分类器
substation;overheat fault;online monitoring;support vector classifier
基金项目(Foundation):
作者(Author):
余彬,胡洛娜,王吉哲
YU Bin,HU Luona,WANG Jizhe
DOI: 10.19585/j.zjdl.2016.01.008
参考文献(References)：
- [1]韩蓓,盛戈嗥,蒋秀臣,等.基于Zig Bee无线传感器网络的导线接头在线测温系统[J].电力系统自动化,2008,32(16):72-77.
- [2]曹楠,李刚,王冬青.智能变电站关键技术及其构建方式的探讨[J].电力系统保护与控制,2011,39(5):63-68.
- [3]KAYANO PD,SILVA MS,MAGRINI LC.Distribution substation transformer and circuit breaker diagnoses with the assistance of real-time monitoring[C].IEEE Transmission&Distribution Conference&Exposition,2004.
- [4]梁湖辉,张峰,常冲,等.基于Zig Bee的变电站监测报警系统[J].电力系统保护与控制,2010,38(12):121-124.
- [5]邬春明,程亮.变电站监测数据采集系统[J].电测与仪表,2014,51(3):64-67.
- [6]李勇,刘哲.基于LS-SVM的汽轮机组加热器运行经济性评价方法[J].热力发电,2014,43(2):9-15.
- [7]FEDERICO SFORZA,VITTORIO LIPPI.Support vector machine classification on a biased training set:Multijet background rejection at hadron colliders[J].Nuclear Instruments and Methods in Physics Research A,2013(722):11-19.
- [8]余海涛.高压变电站无线自动测温控制系统研究[D].天津:天津大学,2007.
- [9]张涛,徐晓苏.基于C-SVM的组合导航系统故障诊断算法[J].中国惯性技术学报,2011,19(2):239-242.