提高输电线路缺陷故障识别准确率  

Improving the Accuracy of Fault Identification for Transmission Line Defectsy

作　　者：汪佳 钟月萍 陈阳 王欢欢 蔡光柱 WANG Jia;ZHONG Yueping;CHEN Yang;WANG Huanhuan;CAI Guangzhu(State Grid Information and Telecommunication Group Beijing Guodiantong Network Technology Co.,Ltd.,Beijing fengtai 100070,China)

机构地区：[1]国网信息通信产业集团北京国电通网络技术有限公司,北京丰台100070

出　　处：《农村电气化》2024年第10期72-75,共4页Rural Electrification

摘　　要：输电线路缺陷故障识别率是输电专业向数字化、智能化转型的一项重要考核指标,小组成员经过调查分析得出智能监拍装置通道缺陷故障识别准确率低是影响输电线路缺陷故障识别率的关键症结。针对症结,从软硬件方面分析找到6条末端原因,逐条分析确认出4条要因。针对要因,开展现场勘察与测试,提出提高芯片主频以及NPU算力,更换像素更高的摄像机,改进图片压缩算法以及增加样本库数量4条对策。对策实施后进行效果检查,输电线路智能监拍装置通道缺陷故障识别准确率由86.68%提高至91.50%,症结得以成功解决,活动初见成效。The identification rate of transmission line defects and faults is an important assessment indicator for the transformation of the transmission profession towards digital intelligence.After investigation and analysis,the team members found that the low accuracy of intelligent monitoring device channel defect and fault identification is the key factor affecting the identification rate of transmission line defects and faults.In response to the crux,6 end causes were identified through software and hardware analysis,and 4 main causes were identified through various analyses.In response to the key factors,on-site investigation and testing were carried out one by one,and four strategies were proposed to improve the chip's main frequency and NPU computing power,replace cameras with higher pixels,improve image compression algorithms,and increase the number of sample libraries.After the implementation of the countermeasures,the effectiveness inspection was carried out,and the accuracy of identifying channel defects and faults in the intelligent monitoring device for transmission lines was increased from 86.68%to 91.50%.The crux was successfully resolved,and the activity achieved initial results.

关 键 词：输电线路 缺陷故障识别 智能监拍装置 芯片主频 图片压缩算法 

分 类 号：TM726[电气工程—电力系统及自动化]