石墨烯材料与传感器技术在变压器故障诊断中的应用  

Application of graphene materials and sensor technology in transformer fault diagnosis

作　　者：袁捷 钱俊凤 甘润东 YUAN Jie;QIAN Junfeng;GAN Rundong(Information Center of Guizhou Power Grid Co.,Ltd.,Guiyang 550000,China)

机构地区：[1]贵州电网有限责任公司信息中心,贵州贵阳550000

出　　处：《粘接》2024年第11期81-84,共4页Adhesion

摘　　要：为了实现精度更高的变压器故障诊断,基于石墨烯材料传感器技术进行变压器的故障诊断研究。以NMOs修饰石墨烯材料为原材料进行4种薄膜气敏传感器的制备,并将其用于变压器的故障气体采集;使用基于GA-BP神经网络的预测模型作为变压器故障诊断模型,并通过数字孪生可视化技术进行展示。试验结果表明,分别以ZnO-rGO、CuO-GO-CuO、SnO_(2)-GO-SnO_(2)-Ag为原材料制备的气敏薄膜传感器对变压器故障气体中的CH4、H2和C2H2具有良好的敏感性,分别达到了12.11%、10.05%和16.14%;使用设计的基于GA-BP神经网络的预测模型进行故障气体浓度预测时精度较高。In order to achieve higher accuracy in transformer fault diagnosis,the fault diagnosis of transformer was studied based on graphene material sensor technology.Four kinds of thin-film gas sensors were prepared using NMOs-modified graphene materials as raw materials,and they were used for transformer fault gas collection.The prediction model based on GA-BP neural network was used as the transformer fault diagnosis model,and was demonstrated by digital twin visualization technology.The experimental results showed that the gas sensitive thin film sensors prepared with ZnO-rGO,CuO-GO-CuO and SnO_(2)-GO-SnO_(2)-Ag as raw materials had good sensitivity to CH4,H2 and C2H2 in transformer fault gases,reaching 12.11%、10.05%,and 16.14%,respectively.The prediction model based on GA-BP neural network designed for fault gas concentration prediction has high accuracy.

关 键 词：数字孪生 变压器故障 薄膜传感器 GA-BP 

分 类 号：TQ317[化学工程—高聚物工业] TM41[电气工程—电器]