基于DGA与IGWO-WELM的变压器不平衡故障诊断研究  

Research on transformer unbalance fault diagnosis based on DGA and IGWO-WELM

作　　者：雷家浩 包永强 钱玉军 姜丹琪 王森林 LEI Jiahao;BAO Yongqiang;QIAN Yujun;JIANG Danqi;WANG Senlin(School of Automation,Nanjing Institute of Technology,Nanjing 211167,China)

机构地区：[1]南京工程学院自动化学院,江苏南京211167

出　　处：《现代电子技术》2023年第24期105-108,共4页Modern Electronics Technique

基　　金：教育部中国高校产学研创新基金:面向5G及未来移动通信网络的NOMA系统资源调度算法(2021FNA05002)。

摘　　要：针对变压器故障诊断精度较低的问题,提出一种改进的灰狼算法(IGWO)与加权极限学习机(WELM)的变压器不平衡故障诊断模型。首先,基于油中气体分析(DGA)技术,结合无编码方法将变压器的7种特征量作为可视输入;然后,采用Logistic混沌映射、云模型惯性权重对灰狼算法(GWO)进行改进;最后利用IGWO对WELM的相关参数进行迭代优化,并利用IGWO-WELM故障诊断模型对变压器进行故障诊断。试验结果表明:提出模型的G-mean平均值为96.06%,比GWO-WELM、GA-WELM、PSO-WELM和WELM分别高10.96%、12.92%、1.08%和18.41%;误报率平均值为12.28%,也明显低于其他4种模型。In allusion to the low accuracy of transformer fault diagnosis,an improved grey wolf algorithm(IGWO)and weighted limit learning machine(WELM)model for transformer unbalanced fault diagnosis is proposed.Based on the dissolved gas analysis(DGA)technology,the seven characteristic quantities of the transformer are regarded as visual inputs by combining with the non-coding method.The grey wolf algorithm(GWO)is improved by means of Logistic chaotic mapping and cloud model inertia weight.IGWO is used to iteratively optimize the relevant parameters of WELM,and IGWO-WELM fault diagnosis model is used for the transformer fault diagnosis.The testing results show that the G-mean mean value of the proposed model is 96.06%,which is 10.96%,12.92%,1.08%,and 18.41% higher than that of GWO-WELM,GA-WELM,PSO-WELM,and WELM,respectively,and the mean value of false negative rate is 12.28% lower than GWO-WELM,GA-WELM,PSO-WELM,and WELM.

关 键 词：变压器 不平衡故障诊断 油中气体分析(DGA) IGWO 加权极限学习机(WELM) IGWO-WELM 

分 类 号：TN911.23-34[电子电信—通信与信息系统]