基于异构数据的变压器远程状态孪生与寿命评估技术  

Remote Status Digital Twin and Life Assessment of Transformer Based on Heterogeneous Data

作　　者：范桂有 濮卫兴[2] 田存建 郑穗生[2] 沈龙山 FAN Guiyou;PU Weixing;TIAN Cunjian;ZHENG Suisheng;SHEN Longshan(Ultra High Voltage Company,State Grid Fujian Electric Power Co.,Ltd.,Fuzhu Fujian 350001,China;NARI Technology Co.,Ltd.,Nanjing Jiangsu 210000,China)

机构地区：[1]国网福建省电力有限公司超高压分公司,福建福州350001 [2]国电南瑞科技股份有限公司,江苏南京210000

出　　处：《电子器件》2024年第3期796-803,共8页Chinese Journal of Electron Devices

基　　金：福建省超高压公司宁德等变电站远程智能巡视加装项目。

摘　　要：目前以机器人巡视为主的远程运维方式,难以构建全面的数字孪生运维体系,而以电力变压器为代表的变电站一次设备,是集合电气与多种物理特性的状态集合,其远程精准运维与状态分析,必然建立在海量异构数据基础之上。以变压器为例,研究变压器数字孪生模型构建方法和数字孪生数据层技术,建立传感器数据下的数字孪生变压器模型,从物理模型类、基于参数类和基于认识模型类三个方面探索多源异构数据融合方法。依照所构建的寿命模型,在2年的运行阶段中,由于温度导致的变压器老化效应,不断在工作状态中进行累积,最终其热寿命损失累计超过了360 h,与实际情况相符合。At present,the remote operation and maintenance method based on robot inspection is difficult to build a comprehensive digital twin operation and maintenance system.The primary equipment of the substation,represented by the power transformer,is a state collection of electrical and various physical characteristics,and its accurate remote operation and maintenance and state analysis must be based on massive heterogeneous data.Taking transformer as an example,the construction method of transformer digital twin model and digital twin data layer technology are studied,the digital twin transformer model under sensor data is established,and the multi-source heterogeneous data fusion method is explored from three aspects,i.e.,physical model class,parameter class and cognitive model class.According to the life model constructed,in the 2-year operation stage,the aging effect of the transformer caused by temperature continuously accumulates in the working state,and finally its thermal life loss accumulates more than 360 h,which is consistent with the actual situation.

关 键 词：远程运维 异构数据 数据融合 变压器老化 

分 类 号：TM401.1[电气工程—电器] TM743