基于神经网络和小波分析的机组振动故障诊断  

Research on fault diagnosis of hydro-generating unit vibration based on neural network and wavelet analysis

作　　者：李郁侠 陈继尧[1] 刘立峰[1] 

机构地区：[1]西安理工大学水利水电学院,陕西西安710048

出　　处：《武汉大学学报（工学版）》2006年第5期22-25,共4页Engineering Journal of Wuhan University

基　　金：陕西省教育厅专项科研计划项目(编号:05JK266)

摘　　要：对于水轮发电机组而言,尽快实施其故障诊断是十分必要的.小波包分析能有效地提取机组振动信号中的有用成分,采用小波包分解方法提取特殊频段上的能量特征值作为神经网络输入向量,针对南桠河水电厂实测数据采用单隐层BP网络进行训练并对不同故障模式进行识别,取得良好效果.For hydro-generators, it is necessary to carry out fault diagnosis. Wavelet packet analysis can pick up the useful information of the hydro-generators, which is regarded as evidence to diagnose fault. The characteristic vector is picked up by using the wavelet packet and energy method as the ANN input neural vector. The ANN uses BP algorithm with training by teacher for data of Nanyahe Hydropower Station. The good effectiveness is obtained.

关 键 词：特征向量提取 水电机组 小波包 神经网络 故障诊断 

分 类 号：TM312.07[电气工程—电机]