试验发电机定子内部故障检测的仿真研究  

Simulation Research on Stator Internal Fault Detection of Test Generator

作　　者：孟伟[1] 张健欣[1] 宋红喜 

机构地区：[1]内蒙古工业大学电力学院,内蒙古呼和浩特010080

出　　处：《计算机仿真》2014年第6期142-146,共5页Computer Simulation

基　　金：国家自然科学基金(61364009);内蒙古自然科学基金项目(2012MS0906);内蒙古自治区高等学校科学基金资助项目(NJZY11081)

摘　　要：发电机定子匝间短路使气隙磁场产生大量的谐波,导致故障检测精度降低。传统的方法主要通过简化参数进行理论分析或建立数学仿真模型,不能准确反映真实系统磁场的状况。为实现实时检测故障,将主元分析和核主元分析应用到定子短路故障检测中,这两种方法都是在缩小了的实际系统--动模实验室的基础上,实测了一台容量为15kVA的发电机定子故障时机端的电气量信息,实验数据能准确反应系统的真实磁场变化。为比较两种方法的优劣,提出一种错误率的计算方法,通过比较错误率和仿真结果,说明核主元分析方法更适合于定子匝间短路的故障检测。A large amount of harmonics of gap magnetic is produced when inter - turn short circuit happens on generator stator winding, then the accuracy of fault detection is decreased. To detect the faults in real time, the principal component analysis and kernel principal component analysis were used for detecting stator fault in the paper. The two methods are based on dynamic model laboratory -a reduced actual system. In the experiment, the electric parameters of generator were measured under the status of stator fault, and the capacity of generator is 15 KVA. Then the real magnetic field was reflected by these data. In this paper, a new way of calculating error rate was proposed for comparing the two methods. The results show that kernel principal component analysis method is more suitable for inter-turn short circuit detection.

关 键 词：匝间短路 主元分析 核主元分析 动模实验室 故障检测 

分 类 号：TM341[电气工程—电机]