基于支持向量机和有限元分析的变压器绕组变形分类方法  

Classification Method of Transformer Winding Deformation Based on SVM and Finite Element Analysis

作　　者：邓祥力[1] 谢海远 熊小伏[2] 高亮[1] 

机构地区：[1]上海电力学院电气工程学院,上海市杨浦区200090 [2]重庆大学电气工程学院,重庆市沙坪坝区400044

出　　处：《中国电机工程学报》2015年第22期5778-5786,共9页Proceedings of the CSEE

基　　金：上海绿色能源并网工程技术研究中心项目(13DZ2251900)~~

摘　　要：电力变压器绕组变形导致内部匝间故障的事故率已升至首位,成为影响变压器安全稳定运行的重要因素。变压器运行过程中绕组变形导致绕组漏磁场发生变化,进而引起漏感参数发生变化。该文根据这一基本特征,研究绕组参数和变形程度之间的关系,利用有限元模型计算变压器的磁场能量和电感参数,采用工程计算法对该模型的正确性进行校验,继而提出利用支持向量机模型对绕组变形进行分类,并通过优化的网格搜索法对参数进行优化。仿真结果表明,所提方法具有较高的分类准确度,适用于变压器绕组变形的在线监测。Internal turn-to-turn fault caused by power transformer winding deformation has risen to the first of all the faults, which have a significant effect on the safety and stable operation of transformer. The winding deformation will cause winding leakage magnetic field to change during the running of the transformer, then lead to the change of leakage inductance. According to the basic characteristics, relationship between the winding parameters and deformation was studied in this paper. After that, the energy of magnetic field and inductance parameters was calculated by the finite element model of transformer, which was verified by the method of engineering calculation. Support vector machine（SVM） model was proposed to classify the winding deformation, which was optimized by grid search algorithm in this paper. The simulation results show that the classification method has a high reliability, and is suitable for online monitoring in transformer winding deformation.

关 键 词：变压器 绕组变形 有限元方法 支持向量机 

分 类 号：TM41[电气工程—电器]