摘要
气体绝缘组合电器(GIS)作为重要输电设备,由局部放电导致GIS绝缘劣化是常见的故障,需对其进行诊断。从不同放电故障下GIS振动信号的特点出发进行GIS放电故障的识别,搭建GIS典型放电故障实验模拟平台,采集GIS正常运行与尖刺放电、悬浮放电和沿面放电3种典型放电故障的振动信号,采用改进集总经验模态分解法(MEEMD)提取不同运行状态下振动信号模态函数,计算模态函数的近似熵值,将差异较为明显的近似熵值作为特征量输入相关向量机(RVM)进行训练,建立二叉树放电模型,对测试样本进行分类,验证此方法的正确性,实现放电故障诊断。结果表明,改进集总经验模态分解可以较好提取信号中不同的频率成分,近似熵作为GIS振动信号模态函数的特征量能有效表征放电故障下信号的特点,实现对GIS放电故障进行诊断。
As an important transmission equipment,GIS insulation degradation caused by partial discharge is a com⁃mon fault needing to be diagnosed.GIS discharge faults are identified based on the characteristics of GIS vibration signals under different discharge faults.An experimental simulation platform for its typical discharge faults of GIS is built to collecting vibration signal simples under four states:the normal state,the spike discharge,suspended dis⁃charge and surface discharge.The method of modified ensemble empirical mode decomposition(MEEMD)is pro⁃posed to extracting the modal features in various states,and then calculated the approximate entropy of each modal state,input the approximate entropy values which have significant difference to relevance vector machine(RVM)for training and established different discharge fault models to classify the test samples.The results show that the approx⁃imate entropy extracted from the MEEMD can effectively show the characteristics of the signal under discharge faults also can be used as GIS vibration signal features to diagnose the discharge faults of GIS effectively.
作者
臧旭
马宏忠
吴金利
李呈营
陈明
张利
屈斌
ZANG Xu;MA Hongzhong;WU Jinli;LI Chengying;CHEN Ming;ZHANG Li;QU Bin(College of Energy and Electrical Engineering,Hohai University,Nanjing 211100,China;Electric Power Research Institute of State Grid Tianjin Electric Power Company,Tianjin 300041,China)
出处
《高压电器》
CAS
CSCD
北大核心
2020年第6期129-137,共9页
High Voltage Apparatus
基金
国网天津市电力公司重点科技项目资助(KJ19⁃1⁃19)。
关键词
GIS
局部放电
振动信号
MEEMD
近似熵
相关向量机
gas insulated switchgear(GIS)
partial discharge
vibration signal
MEEMD
approximate entropy
relevance vector machine
