基于红外图像特征参量和GK-SVM算法的复合绝缘子劣化诊断方法研究  

Research on Composite Insulator Degradation Diagnosis Method Based on Infrared Image Feature Parameters and GK-SVM

作　　者：武永泉 周晖 赵轩 张四维 祝仁杰 黄昭 王艺蛟 乔新涵 张东东 WU Yongquan;ZHOU Hui;ZHAO Xuan;ZHANG Siwei;ZHU Renjie;HUANG Zhao;WANG Yijiao;QIAO Xinhan;ZHANG Dongdong(Nanjing Power Supply Company,State Grid Jiangsu Electric Power Co.Ltd.,Nanjing 210024,China;Institute of Transmission and Distribution Equipment and New Technologies,Nanjing Institute of Technology,Nanjing 211167,China)

机构地区：[1]国网江苏省电力有限公司南京供电分公司,南京210024 [2]南京工程学院输配电装备与新技术研究所,南京211167

出　　处：《西南大学学报（自然科学版）》2025年第2期184-194,共11页Journal of Southwest University(Natural Science Edition)

基　　金：国网江苏省电力有限公司科技项目(J2023087);江苏省自然科学基金项目(BK20181021)。

摘　　要：复合绝缘子劣化、局部温升已成为输电系统运行过程中的严重问题,红外图像可用于复合绝缘子发热的识别,但复合绝缘子发热影响因素复杂、缺乏有效识别方法等问题制约了红外图像的有效应用。考虑风速、湿度、碳化通道长度、水分侵入等条件,建立了复合绝缘子碳化通道电—热—流多物理场仿真模型,系统研究了劣化复合绝缘子发热的影响因素,并提出基于环境输入参量及红外图像特征参量的粒度高斯核支持向量机(GK-SVM)故障识别方法。研究结果表明:电压等级、风速、湿度、碳化通道长度、水分侵入等级均对复合绝缘子缺陷处温度产生影响,相同条件下,220 kV复合绝缘子缺陷处的相对温升最大,可达13.2℃。其中碳化通道缺陷长度和水分侵入等级对复合绝缘子缺陷处温度产生较大影响,当缺陷长度从0 mm升至200 mm时,缺陷处温度相对于环境温升增加了4.3倍;而水分侵入等级从0级升至3级时,缺陷处温度相对于环境温升增加了0.7倍。湿度和风速均对缺陷处温度产生负向影响,当环境湿度从30%升至80%时,缺陷处温度相对于环境温升降低了20.6%;而环境风速从0 m/s升至6 m/s时,缺陷处温度相对于环境温升降低了24.2%。基于环境输入参量及红外图像特征参量,应用GK-SVM算法的复合绝缘子发热故障识别方法对110 kV和220 kV线路故障绝缘子的故障识别率分别可达93.98%和94.67%,误报率仅为7.14%和5.33%。研究结果对于劣化复合绝缘子诊断方法的应用具有重要的参考价值。The degradation and local temperature rise of composite insulators have become a serious problem in the operation of transmission systems.Infrared image can be used for identifying the heating of composite insulators,however,the complex factors affecting the heating of composite insulators and the lack of effective identification methods have hindered the effective application of infrared image.This research considered conditions such as wind speed,humidity,carbonization channel length,and moisture intrusion to establish a electricity-heat-flow multi-physical field simulation model for composite insulator carbonization channels.The factors affecting the heating of deteriorated composite insulators were systematically studied,and a Granular Gaussian Kernel Support Vector Machine(GK-SVM)fault identification method based on environmental input parameters and infrared image characteristic parameters was proposed.Research had shown that voltage level,wind speed,humidity,carbonization channel length,and mositure intrusion all had an impact on the temperature at the defect location of composite insulators.Under the same conditions,the relative temperature rise at the defect location of 220 kV composite insulators was the highest,reaching 13.2℃.The length of carbonization channel defects and the level of moisture intrusion had a significant impact on the temperature at the defect site of composite insulators.When the defect length increased from 0 to 200 mm,the temperature at the defect site increased by 4.3 times relative to the environmental temperature rise.When the moisture intrusion level increased from 0 to 3,the temperature at the defect site increased by 0.7 times relative to the environmental temperature rise.Both humidity and wind speed had a negative impact on the temperature at the defect site.When the environmental humidity increased from 30%to 80%,the temperature at the defect site decreased by 20.6%relative to the environmental temperature rise.When the environmental wind speed increased from 0 to 6 m/s,

关 键 词：红外图像特征 复合绝缘子 劣化 发热 识别方法 粒度高斯核 

分 类 号：TM216[一般工业技术—材料科学与工程]