低温下油浸纸式套管绝缘局部放电发展规律及识别方法  

Development Law and Identification Method of PD of Oil-impregnated Paper Bushing Insulation at Low Temperature

作　　者：曾中伟 康吉义 ZENG Zhong-wei;KANG Ji-yi(ALSHA Electric Power Bureau of Inner Mongolia Power(Group)Co.,Ltd.,Alsha 750306,China)

机构地区：[1]内蒙古电力(集团)有限责任公司阿拉善电业局,内蒙古阿拉善750306

出　　处：《变压器》2022年第5期43-51,共9页Transformer

摘　　要：为了保障高寒低温环境下油浸式高压套管的安全可靠运行,本文中作者开展了低温下套管油纸绝缘局部放电发展规律研究和局部放电程度评估。搭建了油纸绝缘试样和套管模型的低温局部放电试验平台,将放电全过程划分为3个阶段从平均放电量、最大放电量、脉冲重复率三方面分析油纸绝缘不同温度下局部放电的发展特性,根据灰度梯度共生矩阵和颜色矩对演变过程中PRPD谱图提取的特征参量,提出了一种基于鲸鱼优化极限学习机的局部放电阶段识别方法,并与极限学习机、支持向量机和反向传播神经网络进行对比。研究发现提出方法的不同缺陷识别准确率均高于其他方法,可有效评估高寒地区高压套管的局部放电严重程度,为后期套管维护计划提供了建议。In order to ensure the safe and reliable operation of oil-immersed high-voltage bushing in low-temperature environments,the development law and levels of oil-paper insulation partial discharge of bushing at low temperatures were carried out.A low-temperature partial discharge test platform for oil-paper insulation samples and bushing models were built,and the entire discharge process was divided into 3 stages.The average discharge capacity,maximum discharge capacity,and pulse repetition rate were analyzed for partial discharge at different temperatures for oil-paper insulation.Based on the characteristic parameters extracted from the PRPD spectrum during the evolution of the gray-gradient co-occurrence matrix and the color moment,a partial discharge stage recognition method based on the whale-optimized extreme learning machine is proposed,and it is combined with the extreme learning machine and support vector machine and back-propagation neural network.The study found that the accuracy of different defect identification of the proposed method is higher than that of other methods,which can effectively assess the severity of partial discharge of high-voltage bushings in high-cold areas,and provide suggestions for later maintenance plans.

关 键 词：油浸纸式套管 局部放电 低温 典型缺陷 

分 类 号：TM406[电气工程—电器]