人工神经网络用于发电机定子绝缘局部放电模式识别  

Application of Artificial Neural Networks to the Partial Discharge Patterns Recognition of Generator Stator Windings

作　　者：彭湃[1] 王泽忠[2] 李成榕[2] 贾逸梅[2] 王伟[2] 王景春[2] 

机构地区：[1]中国电力科学研究院,北京100085 [2]华北电力大学(北京),北京102206

出　　处：《现代电力》2001年第2期40-44,共5页Modern Electric Power

基　　金：国家电力公司重大科技项目

摘　　要：讨论了反向传播算法的人工神经网络在局部放电模型类型识别中的应用 ,设计了反映大型发电机定子绝缘局部放电的 5类不同物理仿真绝缘模型。仿真模型包括的局部放电类型有端部放电、槽间放电、在绝缘不同位置的三种内部放电。研究了人工神经网络的识别能力。针对大型发电机定子绝缘局部放电的类型 。The application of the BP neural network to the partial discharge patterns recognition of generator stator windings is discussed Five types of simulation models for the insulation PD in the large generator stator windings are presented The PD simulation models include the surface discharge at end windings,the discharge between two slots and the interior discharge in three different positions of the insulation The recognition ability of the ANN is studied A higher recognition rate for the partial discharge types of the large generator stator insulation is gained

关 键 词：人工神经网络 大型发电机 局部放电 定子绝缘 模型识别 

分 类 号：TM31[电气工程—电机] TP183[自动化与计算机技术—控制理论与控制工程]