基于人工神经网络的变压器励磁涌流的鉴别  

Discrimination for Magnetizing Inrush Current of Transformer Based on Artificial Neural Network

作　　者：刘喜梅[1] 魏婉韵[1] 于飞[1] 

机构地区：[1]青岛科技大学自动化与电子工程学院,山东青岛266042

出　　处：《低压电器》2007年第19期57-60,64,共5页Low Voltage Apparatus

基　　金：青岛市自然科学基金项目资助(05-1-jc-88);国家"十五"科技攻关项目资助(j2001-1-jc-88)

摘　　要：研究了基于神经网络方法的电力变压器励磁涌流鉴别。对变压器励磁涌流及故障电流进行了数字仿真,比较了两者在物理特性上的区别。利用Matlab的人工工具箱,分别建立了BP和RBF神经网络模型,对励磁涌流和故障电流的样本进行训练及测试。结果表明,人工神经网络可以正确地区分励磁涌流和故障电流,RBF神经网络能更加快速、准确地判断出变压器的故障。Discrimination for magnetizing inrush current of transformer based on artificial neural network(NN) was studied.First the physical characters of magnetizing inrush currents and internal fault currents were compared through digital simulation under several conditions.The model of BP and RBF NN were established using ANN toolbox in Matlab software.The sample datum of magnetizing inrush currents and internal fault currents were trained and tested.Results show that ANN can distinguish magnetizing inrush current...

关 键 词：变压器 励磁涌流 故障电流 BP神经网络 RBF神经网络 

分 类 号：TM41[电气工程—电器]