新型的RBF神经网络在水轮发电机组故障诊断中的应用  

Application of a novel RBF algorithm to fault diagnosis of hydro-turbine generating unit

作　　者：张利平[1] 孙美凤[1] 王铁生[1] 

机构地区：[1]华北水利水电学院,郑州450011

出　　处：《水力发电学报》2009年第6期219-223,共5页Journal of Hydroelectric Engineering

摘　　要：本文应用免疫遗传系统的调节原理对RBF神经网络隐层中心数量和位置的进行选择,同时采用递推最小二乘法来确定网络输出层的权值,从而建立了一种新型的RBF神经网络模型,并将该模型应用于水轮机发电机组的故障诊断。诊断结果表明,该模型收敛速度快,精度高并有较好的泛化能力,为水轮发电机组状态监测及故障诊断提供了一种新途径。This paper presents a novel RBF neural network model. In are introduced to select the number and positions of hidden layer RBF decided with the recursive least squares algorithm. This is an effective unit. The application to a practical generator set shows advantages of easiness in generalization. This approach is provided for monitoring and this model, the immune genetic principles centers, and the output layer weights are model for fault diagnosis of hydro turbine the model, swiftness, high precision and fault diagnosis of hydro-turbine unit.

关 键 词：水轮发电机 故障诊断 RBF神经网络 免疫. 

分 类 号：TM312[电气工程—电机]