基于BP神经网络的绝缘软梯缺陷识别模型在带电工作中的应用研究  

Application of BP neural network based on defect identification model of insulated rope ladder in live working

作　　者：代盛熙 俞晓鹏 宣梦真 叶卫忠 黄晔凯 Dai Shengxi;Yu Xiaopeng;Xuan Mengzhen;Ye Weizhong;Huang Yekai(Jinhua Power Supply Company,State Grid Zhejiang Electric Power Co.,Ltd.,Zhejiang Jinhua,321000,China)

机构地区：[1]国网浙江省电力有限公司金华供电公司,浙江金华321000

出　　处：《机械设计与制造工程》2023年第3期83-86,共4页Machine Design and Manufacturing Engineering

摘　　要：带电作业是输变电设备测试、检修、改造的重要手段,为了确保工人安全带电作业,以绝缘软梯的绝缘状态为研究对象,利用BP神经网络算法对其进行故障分类辨别。首先分析了各种传感器的优缺点;然后利用传感器对泄漏电流的信号数据进行采集,分析不同情况下泄漏电流的特性;最后建立基于BP神经网络算法的故障缺陷识别系统,将泄漏电流特性信号作为故障缺陷识别系统的输入、故障类型信号作为故障缺陷识别系统的输出,训练后的绝缘软梯缺陷识别系统识别准确率为96%。研究表明,绝缘软梯缺陷识别系统可以实现绝缘软梯的缺陷识别。Live working is an important means of testing,overhauling and reforming power transmission and transformation equipment.In order to ensure workers'safe live working,it takes the insulation state of insulation ladder as the research object,uses BP neural network algorithm to classify and identify its faults.Firstly,the advantages and disadvantages of various sensors are analyzed.Then the signal data of leakage current are collected by sensors,and the characteristics of leakage current under different conditions are analyzed.Finally,a fault defect identification system based on BP neural network algorithm is established.The leakage current characteristic signal is used as the input of the fault defect identification system,and the fault type signal is used as the output of the fault defect identification system.After training,the identification accuracy of the insulation ladder defect identification system is 96%.The research shows that the defect identification system of insulation ladder can realize the defect identification of insulation ladder.

关 键 词：BP神经网络 带电作业 绝缘软梯 泄漏电流 缺陷识别 

分 类 号：TM50[电气工程—电器]