基于稳态故障特征量的大型汽轮发电机定子绕组匝间短路智能在线识别研究  

Study of On-line Identification of Large Turbogenerator Stator Windings’ Faults Based on Stable Fault Characters by Use of Intelligent Method

作　　者：党晓强[1] 刘华[1] 桂林[1] 

机构地区：[1]四川大学水利水电学院,成都610065

出　　处：《大电机技术》2007年第6期11-15,共5页Large Electric Machine and Hydraulic Turbine

基　　金：四川大学青年科学基金资助项目(06019)

摘　　要：定子绕组匝间短路是汽轮发电机破坏性很强的内部故障,其缺乏有效的针对性保护装置为系统安全运行留下隐患。本文提出基于在线故障特征量,通过智能技术的非线性映射来及时识别匝间短路。文中分析定子绕组匝间短路过程中故障相电流的变化和纵向零序电压的产生,建立这两者作为特征量的数学模型,提出应用动态Elman神经网络对匝间短路进行在线识别。依据一台大型汽轮发电机的典型参数,计算其在线时的匝间短路故障量,具体特征数据输入Elman网络进行识别。算例结果表明,基于合理的稳态故障特征量,定子绕组匝间短路是可以被Elman网络有效在线识别的。Turbo-generator stator＇s inter-turn short-circuit is a usual serious fault, there would be hidden big trouble for electric power system＇s safety due to lack of efficient protection. On-line monitoring generator＇s operation condition combined with intelligent non-linear identification technology were presented to observe fault in time. Longitudinal zero-sequence voltage and fault phase current were analyzed as stator winding＇s inter-turn short-circuit stable fault characters, mathematical model of which were built, dynamic Elman neural network were introduced to identify the fault. General parameters of a large turbogenerator were used for calculate its stator winding＇s inter-turn short-circuit fault characters, and identification were performed by trained Elman neural network followed. Example indicated that the Elman network could efficiently identify generator stator＇s inter-turn short-circuit based on rational fault characters combination.

关 键 词：定子绕组匝间短路 稳态故障特征量 数学模型 ELMAN神经网络 在线识别 

分 类 号：TM623.3[电气工程—电力系统及自动化]