电力变压器绕组故障智能诊断优化方法研究  

Research on Intelligent Diagnosis Optimization Method of Power Transformer Winding Fault

作　　者：伍芳 WU Fang(Changzhou Haineng Electric Appliance Co.,Ltd.,Changzhou,Jiangsu 213002,China)

机构地区：[1]常州海能电器有限公司,江苏常州213002

出　　处：《自动化应用》2024年第17期179-181,共3页Automation Application

摘　　要：由于电力变压器绕组不同故障状态具有不同的特征,难以保障诊断效果,为此,提出电力变压器绕组故障智能诊断优化方法。以电力变压器绕组振动信号为对象,通过缩放和平移的方式将其分解为一组小波包基函数,并将每个分解得到的频带能量作为特征值,形成电力变压器绕组故障信号的特征向量。在诊断阶段,引入了核极限学习机,实现对特征对应电力变压器绕组故障状态的诊断。结果表明,该方法对不同类型绕组故障状态的有效诊断率均达到了85.0%以上,与对照组相比具有明显优势。Due to the different characteristics of different fault states in power transformer windings,it is difficult to ensure diagnostic effectiveness.Therefore,an intelligent diagnosis optimization method for power transformer winding faults is proposed.Taking the vibration signal of power transformer winding as the object,it is decomposed into a set of wavelet packet basis functions through scaling and translation,and the energy of each decomposed frequency band is used as the eigenvalue to form the eigenvector of power transformer winding fault signal.In the diagnostic stage,a kernel extreme learning machine was introduced to diagnose the fault state of the power transformer winding corresponding to the features.The results show that the effective diagnostic rate of this method for different types of winding fault states has reached over 85.0%,which has significant advantages compared to the control group.

关 键 词：电力变压器 绕组故障 小波包基函数 核极限学习机 

分 类 号：TM41[电气工程—电器]