气体绝缘组合开关电器绝缘子机械故障智能检测研究  

Research on intelligent detection of mechanical faults of gas insulated switchgear insulators

作　　者：潘静娟 PAN Jingjuan(State Grid Nantong Power Supply Company,Nantong 226007,Jiangsu China)

机构地区：[1]国网江苏省电力有限公司南通供电分公司,江苏南通226007

出　　处：《粘接》2025年第4期155-158,共4页Adhesion

摘　　要：为提高变电站气体绝缘组合开关电器绝缘子机械松动检测精度和检测速度,提出一种改进CNN网络的智能检测方法。方法以盆式绝缘子机械松动类型检测为研究对象,通过将CNN网络卷积核尺寸大小调整为1*3,并在标准CNN网络中增加2层卷积层和全连接层,在训练过程中引入了批量归一化操作,增强CNN网络的非线性表达能力,缩短网络训练时间,实现了盆式绝缘子机械松动智能检测。仿真结果表明,所提方法具有较高的检测准确率,平均检测准确率为99%,平均损失值为0.0011,且具有更快的检测速度,检测时长为76 s。相较于标准CNN模型、SVM模型、KNN模型,所提方法在检测准确率和检测速度具有明显优势,为变电站气体绝缘组合开关电器绝缘子机械松动智能检测提供了参考。In order to improve the accuracy and detection speed of mechanical looseness detection of gas insulated switchgear insulators in substations,an improved CNN network intelligent detection method is proposed.The method focuses on the detection of mechanical looseness types in basin insulators.By adjusting the size of the convolutional kernel of the CNN network to 1*3 and adding two layers of convolutional layers and fully connected layers to the standard CNN network,batch normalization operations are introduced during the training process to enhance the non-linear expression ability of the CNN network,shorten the training time,and achieve intelligent detection of mechanical looseness in basin insulators.The simulation results show that the proposed method has a high detection accuracy,with an average detection accuracy of 99%and an average loss value of 0.0011.It also has a faster detection speed and a detection time of 76 seconds.Compared with standard CNN models,SVM models,and KNN models,the proposed method has significant advantages in detection accuracy and speed,providing a reference for inteligent detection of mechanical looseness of gas insulated switchgear insulators in substations.

关 键 词：气体绝缘组合开关电器 绝缘子机械松动 智能检测 CNN网络 

分 类 号：TM216[一般工业技术—材料科学与工程] TQ116[电气工程—电工理论与新技术]