基于YOLOv3算法的变电站保护硬压板检测技术研究  

Research on Protection Hard Platens Detection Technology in Substations Based on YOLOv3 Algorithm

作　　者：王磊 黄照厅 张礼波 晏丽丽 谭小龙 WANG Lei;HUANG Zhaoting;ZHANG Libo;YAN Lili;TAN Xiaolong(Liupanshui Power Supply Bureau of Guizhou Power Grid Co.,Ltd.,Liupanshui 553000,China;Wuhan Yingrui Power Technology Co.,Ltd.,Wuhan 430074,China)

机构地区：[1]贵州电网有限责任公司六盘水供电局,贵州六盘水553000 [2]武汉映瑞电力科技有限公司,湖北武汉430074

出　　处：《机械与电子》2024年第5期46-49,56,共5页Machinery & Electronics

摘　　要：针对目前变电站保护硬压板巡检多由人工进行读取和核对操作,费时费力且易出错,难以满足变电站二次设备智能化的要求的问题,研究了基于YOLOv3算法的变电站保护硬压板检测与状态识别技术。首先,通过对变电站保护硬压板进行数据采集和标注,建立了一个包含大量样本的数据集。然后,使用YOLOv3算法进行目标检测,实现了对变电站保护硬压板的准确定位和识别。针对硬压板的不同状态,设计了相应的特征提取和分类模块,通过对检测到的硬压板进行状态识别,实现了对硬压板状态的准确判别。实验结果表明,该方法在变电站保护硬压板检测和状态识别方面具有较高的准确性,对压板作业操作票确认、压板状态校核等业务实现具有一定参考价值,可为变电站设备的状态监测和故障的智能诊断提供有效的技术支持。Currently,the inspection of substation protection hard platens is mostly carried out manually,involving time-consuming and error-prone tasks,which does not meet the requirements for intelligent secondary equipment in substations.This paper presents a study on the detection and state recognition technology of substation protection hard platens based on the YOLOv3 algorithm.Firstly,a dataset containing large samples is established through data collection and annotation of substation protection hard platens.Then,the YOLOv3 algorithm is employed for object detection,achieving accurate localization and recognition of the substation protection hard platens.Different feature extraction and classification modules are designed for different states of the hard platens,enabling accurate state recognition of the detected platens.Experimental results demonstrate that this method exhibits high accuracy.This paper provides valuable insights for implementing tasks such as the confirmation of platen operation tickets and platen status verification,offering effective technical support for substation equipment intelligent monitoring and fault diagnosis.

关 键 词：变电站 二次保护 YOLOv3算法 压板状态识别 

分 类 号：TM76[电气工程—电力系统及自动化]