基于图计算的电力变压器状态评估及溯源研究  

Research on State Evaluation and Traceability of Power Transformers Based on Graph Computing

作　　者：郭松 党建[2] 王勇 吕挺 赵宁波 李昳溦 韦江瑶 霍浩伟 GUO Song;DANG Jian;WANG Yong;LYU Ting;ZHAO Ningbo;LI Yiwei;WEI Jiangyao;HUO Haowei(State Grid Shaanxi Procurement Company,Xi'an 710054,Shannxi Province,China;School of Electrical Engineering,Xi'an University of Technology,Xi'an 710048,Shannxi Province,China;State Grid Shaanxi Electric Power Co.,Ltd.,Xi'an 710054,Shannxi Province,China)

机构地区：[1]国网陕西省电力有限公司物资公司,陕西省西安市710054 [2]西安理工大学电气工程学院,陕西省西安市710048 [3]国网陕西省电力有限公司,陕西省西安市710054

出　　处：《电力信息与通信技术》2025年第4期18-24,共7页Electric Power Information and Communication Technology

基　　金：国家自然科学基金项目(52009106);国网陕西省电力有限公司科技项目(5226WZ230001)。

摘　　要：电力变压器的高效运维对电网的安全稳定运行意义重大,文章考虑全寿命周期开展基于图计算的电力变压器状态评估及故障溯源研究。首先,围绕电力变压器生产信息和设备运行2个方面搭建了多维度指标体系;其次,利用分位数理论实现评估指标差异化赋分,提出基于图模型的电力变压器客观量化评估策略;最后,在状态劣化评估基础上建立了基于图卷积神经网络的故障诊断模型,为变压器状态劣化溯源提供了指导建议。实例验证结果表明,所提方法通过多维度数据联动图模型客观量化分析,有效降低评估过程中主观裁量权,提高小样本下故障识别准确性,为电力变压器高效运维提供有力支撑。Accurately evaluating the operating status of power transformers is of great significance for ensuring the safe and stable operation.This paper proposes to conduct research on the state evaluation and fault tracing of power transformers based on graph calculation throughout entire life cycle.Firstly,a power transformer evaluation index system was established based on supplier production information and equipment operation;Secondly,differential scoring of evaluation indicators was achieved using quantile theory,and an objective quantitative evaluation strategy for power transformers based on graph models was proposed.Finally,on the basis of state evaluation,a fault diagnosis model based on graph convolutional neural networks was established,providing guidance and suggestions for tracing the source of transformer state degradation.The results show that the work done in this paper objectively quantifies the analysis through a multi-dimensional data linkage graph model,effectively reducing human subjective discretion in the evaluation process.At the same time,compared with deep neural networks,it improves the accuracy of fault identification under few samples,providing support for efficient operation and maintenance,fault handling of power transformers.

关 键 词：电力变压器 图计算 客观量化评估 故障溯源 图卷积神经网络 

分 类 号：TN915.853[电子电信—通信与信息系统]