基于直流局部放电的高压储能电容器缺陷的模糊识别  

Fuzzy Recognition of the High-voltage DC Energy Storage Capacitor Defects Based on Partial Discharge

作　　者：王鹏[1] 吴广宁[1] 张依强[1] 张血琴[1] 

机构地区：[1]西南交通大学电气工程学院,成都610031

出　　处：《高压电器》2011年第3期64-68,共5页High Voltage Apparatus

基　　金：国家自然基金联合资助项目(10476022)~~

摘　　要：为提高脉冲功率系统中高压储能电容器的可靠性,笔者利用直流下局部放电对电容器进行缺陷检测,得到直流局部放电测量数据的统计谱图,提出基于模糊数学理论识别电容器缺陷类型的方法。首先从22个统计算子中挑选出12个具有典型特征的统计算子组成特征向量,然后通过内部缺陷、接触缺陷、重叠缺陷、油质缺陷4种典型缺陷的测量,计算出每种缺陷特征向量的正态分布隶属函数。最后对未知缺陷类型的电容器进行了模糊检测和解剖验证,对含有可视气泡已知缺陷的电容器进行了检测和模糊识别,证明了系统的实用性,系统可用于脉冲功率系统组装前电容器的可靠性检测。Defect detection via partial discharge under DC voltage was conducted to enhance the reliability of storage capacitor in high-voltage pulsed power system,and DC partial discharge measurement data were gained.Fuzzy mathematics theory was adopted for capacitor defect type recognition.Firstly,12 statistical operators from 22 were selected to form the statistical operator characteristic vector.Then each type of defect＇s characteristic vector membership function with normal distribution was calculated through the measurement of internal defect,contact defect,overlap defect,and oil quality defect,respectively.Finally,two capacitors with respective an unknown defect and a known defect were detected and anatomized to validate the utility of the method.This proposed method can be used for capacitors reliability detection for pulsed power system prior to assembly.

关 键 词：绝缘缺陷 模糊数学 直流局部放电 模式识别 统计算子 

分 类 号：TM835[电气工程—高电压与绝缘技术] TM531